Abstract
Background
This review is the third update of the Cochrane review "Selenium for preventing cancer". Selenium is a naturally occurring element with both nutritional and toxicological properties. Higher selenium exposure and selenium supplements have been suggested to protect against several types of cancer.
Objectives
To gather and present evidence needed to address two research questions:
1. What is the aetiological relationship between selenium exposure and cancer risk in humans? 2. Describe the efficacy of selenium supplementation for cancer prevention in humans.
Search methods
We updated electronic searches of the Cochrane Central Register of Controlled Trials (CENTRAL; 2017, Issue 2), MEDLINE (Ovid, 2013 to January 2017, week 4), and Embase (2013 to 2017, week 6), as well as searches of clinical trial registries.
Selection criteria
We included randomised controlled trials (RCTs) and longitudinal observational studies that enrolled adult participants.
Data collection and analysis
We performed random‐effects (RE) meta‐analyses when two or more RCTs were available for a specific outcome. We conducted RE meta‐analyses when five or more observational studies were available for a specific outcome. We assessed risk of bias in RCTs and in observational studies using Cochrane's risk assessment tool and the Newcastle‐Ottawa Scale, respectively. We considered in the primary analysis data pooled from RCTs with low risk of bias. We assessed the certainty of evidence by using the GRADE approach.
Main results
We included 83 studies in this updated review: two additional RCTs (10 in total) and a few additional trial reports for previously included studies. RCTs involved 27,232 participants allocated to either selenium supplements or placebo. For analyses of RCTs with low risk of bias, the summary risk ratio (RR) for any cancer incidence was 1.01 (95% confidence interval (CI) 0.93 to 1.10; 3 studies, 19,475 participants; high‐certainty evidence). The RR for estimated cancer mortality was 1.02 (95% CI 0.80 to 1.30; 1 study, 17,448 participants). For the most frequently investigated site‐specific cancers, investigators provided little evidence of any effect of selenium supplementation. Two RCTs with 19,009 participants indicated that colorectal cancer was unaffected by selenium administration (RR 0.99, 95% CI 0.69 to 1.43), as were non‐melanoma skin cancer (RR 1.16, 95% CI 0.30 to 4.42; 2 studies, 2027 participants), lung cancer (RR 1.16, 95% CI 0.89 to 1.50; 2 studies, 19,009 participants), breast cancer (RR 2.04, 95% CI 0.44 to 9.55; 1 study, 802 participants), bladder cancer (RR 1.07, 95% CI 0.76 to 1.52; 2 studies, 19,009 participants), and prostate cancer (RR 1.01, 95% CI 0.90 to 1.14; 4 studies, 18,942 participants). Certainty of the evidence was high for all of these cancer sites, except for breast cancer, which was of moderate certainty owing to imprecision, and non‐melanoma skin cancer, which we judged as moderate certainty owing to high heterogeneity. RCTs with low risk of bias suggested increased melanoma risk.
Results for most outcomes were similar when we included all RCTs in the meta‐analysis, regardless of risk of bias. Selenium supplementation did not reduce overall cancer incidence (RR 0.99, 95% CI 0.86 to 1.14; 5 studies, 21,860 participants) nor mortality (RR 0.81, 95% CI 0.49 to 1.32; 2 studies, 18,698 participants). Summary RRs for site‐specific cancers showed limited changes compared with estimates from high‐quality studies alone, except for liver cancer, for which results were reversed.
In the largest trial, the Selenium and Vitamin E Cancer Trial, selenium supplementation increased risks of alopecia and dermatitis, and for participants with highest background selenium status, supplementation also increased risk of high‐grade prostate cancer. RCTs showed a slightly increased risk of type 2 diabetes associated with supplementation. A hypothesis generated by the Nutritional Prevention of Cancer Trial ‐ that individuals with low blood selenium levels could reduce their risk of cancer (particularly prostate cancer) by increasing selenium intake ‐ has not been confirmed. As RCT participants have been overwhelmingly male (88%), we could not assess the potential influence of sex or gender.
We included 15 additional observational cohort studies (70 in total; over 2,360,000 participants). We found that lower cancer incidence (summary odds ratio (OR) 0.72, 95% CI 0.55 to 0.93; 7 studies, 76,239 participants) and lower cancer mortality (OR 0.76, 95% CI 0.59 to 0.97; 7 studies, 183,863 participants) were associated with the highest category of selenium exposure compared with the lowest. Cancer incidence was lower in men (OR 0.72, 95% CI 0.46 to 1.14, 4 studies, 29,365 men) than in women (OR 0.90, 95% CI 0.45 to 1.77, 2 studies, 18,244 women). Data show a decrease in risk of site‐specific cancers for stomach, colorectal, lung, breast, bladder, and prostate cancers. However, these studies have major weaknesses due to study design, exposure misclassification, and potential unmeasured confounding due to lifestyle or nutritional factors covarying with selenium exposure beyond those taken into account in multi‐variable analyses. In addition, no evidence of a dose‐response relation between selenium status and cancer risk emerged. Certainty of evidence was very low for each outcome. Some studies suggested that genetic factors might modify the relation between selenium and cancer risk ‐ an issue that merits further investigation.
Authors' conclusions
Well‐designed and well‐conducted RCTs have shown no beneficial effect of selenium supplements in reducing cancer risk (high certainty of evidence). Some RCTs have raised concerns by reporting a higher incidence of high‐grade prostate cancer and type 2 diabetes in participants with selenium supplementation. No clear evidence of an influence of baseline participant selenium status on outcomes has emerged in these studies.
Observational longitudinal studies have shown an inverse association between selenium exposure and risk of some cancer types, but null and direct relations have also been reported, and no systematic pattern suggesting dose‐response relations has emerged. These studies suffer from limitations inherent to the observational design, including exposure misclassification and unmeasured confounding.
Overall, there is no evidence to suggest that increasing selenium intake through diet or supplementation prevents cancer in humans. However, more research is needed to assess whether selenium may modify the risk of cancer in individuals with a specific genetic background or nutritional status, and to investigate possible differential effects of various forms of selenium.
Plain language summary
Selenium for preventing cancer
Review question We reviewed the evidence investigating the relation between selenium intake and cancer prevention. This review updates the most recent Cochrane review on this topic (Vinceti 2014), which was an update of Dennert 2011.
Background Selenium is a naturally occurring element that individuals are exposed to mainly through food consumption, although exposure can also occur through air, drinking water, and dietary supplements. Small amounts of selenium are essential for certain biological functions in humans, but slightly higher amounts can pose a toxicity risk, making selenium an element with a narrow, but as yet not well‐defined, safe range of exposure. Selenium occurs in many different chemical forms with different biological activity. From the late 1960s, a few observational studies reported that people with high levels of selenium in their diet or in their body tissues had lower risk of cancer, and some laboratory studies showed that selenium could inhibit the growth of cancer cells. This led to widespread interest in selenium supplements and claims that taking such supplements could prevent cancer. Since that time, many more observational studies have been conducted to compare cancer rates among individuals with high and low selenium exposure. More recently, several randomised controlled trials designed to assess whether selenium supplementation can prevent cancer have been carried out. These trials played a major role in enhancing our understanding of the relation between selenium and cancer risk as a result of their stronger study design as compared with observational studies. The most recent trials in particular have shown high methodological quality and statistical power. Several trials focused on whether selenium could prevent prostate cancer.
Study characteristics This review includes 10 trials in which adults were randomly assigned to receive selenium supplements or placebo, and 70 observational studies in which adults were followed over time to determine whether their baseline selenium status was associated with their risk of cancer. The evidence is current to January 2017.
Key results All of the high‐quality randomised trials reported no effect of selenium on reducing overall risk of cancer or risk of particular cancers, including the most investigated outcome ‐ prostate cancer. Some trials unexpectedly suggested that selenium may increase risks of high‐grade prostate cancer, type 2 diabetes, and dermatological abnormalities.
Observational studies have yielded inconsistent evidence of a possible effect of selenium exposure on cancer risk, with no evidence of a dose‐response relation. When we pooled results of these studies, overall they suggested an inverse relation between cancer exposure and subsequent incidence of any cancer or some specific cancers, such as colon and prostate cancer. However, observational studies have major weaknesses. The selenium exposure status of participants could have been misclassified owing to limitations of the indicators of selenium exposure used, as well as to uncertainty regarding the particular selenium species contributing to overall exposure. In addition, unmeasured confounding from lifestyle or nutritional factors ‐ a major and well‐known source of bias in nutritional epidemiology studies of observational design ‐ could have been present. Therefore, the internal validity of these studies is limited.
Currently, the hypothesis that increasing selenium intake may reduce cancer risk is not supported by epidemiological evidence. Additional research is needed to assess whether selenium may affect the risk of cancer in individuals with specific genetic backgrounds or nutritional status, and to determine how the various chemical forms of selenium compounds may have different effects on cancer risk.
Summary of findings
Summary of findings for the main comparison. Highest compared with lowest selenium exposure for preventing cancer in randomised controlled studies with low risk of bias.
| Highest compared with lowest selenium exposure for preventing cancer in randomised controlled studies with low risk of bias | ||||||
| Patient or population: Participants in trials with low risk of bias Setting: out‐patient Intervention: highest selenium exposure Comparison: lowest selenium exposure | ||||||
| Outcomes | Relative effect (95% CI) | Anticipated absolute effects* (95% CI) | Quality of the evidence (GRADE) | Comments | ||
| Without highest | With highest | Difference | ||||
| Any cancer risk No. of participants: 19,475 (3 RCTs) | RR 1.01 (0.93 to 1.10) | Study population | ⊕⊕⊕⊕ HIGH | SELECT study had the strongest influence on the effect estimate. The RR in all RCTs is 0.99 (95% CI 0.86 to 1.14). | ||
| 10.0% | 10.1% (9.3 to 11.0) | 0.1% more (0.7 fewer to 1 more) | ||||
| Cancer mortality risk No. of participants: 17,448 (1 RCT) | RR 1.02 (0.80 to 1.30) | Study population | ⊕⊕⊕⊕ HIGH | The effect is led from the study SELECT. The RR in all RCTs is 0.81 (95% CI 0.49 to 1.32). | ||
| 1.4% | 1.5% (1.1 to 1.9) | 0.0% more (0.3 fewer to 0.4 more) | ||||
| Colorectal cancer risk No. of participants: 19,009 (2 RCTs) | RR 0.99 (0.69 to 1.43) | Study population | ⊕⊕⊕⊕ HIGH | SELECT study had the strongest influence on the effect estimate. The RR in all RCTs is 0.74 (95% CI 0.41 to 1.33). | ||
| 0.7% | 0.7% (0.5 to 1.0) | 0.0% fewer (0.2 fewer to 0.3 more) | ||||
| Non‐melanoma skin cancer risk No. of participants: 2027 (2 RCTs) | RR 1.16 (0.30 to 4.42) | Study population | ⊕⊕⊕⊝ MODERATEa | Pooled estimate is imprecise owing to high heterogeneity. The RR in all RCTs is 1.23 (95% CI 0.73 to 2.08). | ||
| 2.9% | 3.4% (0.9 to 12.9) | 0.5% more (2 fewer to 10 more) | ||||
| Lung cancer risk No. of participants: 19,009 (2 RCTs) | RR 1.16 (0.89 to 1.50) | Study population | ⊕⊕⊕⊕ HIGH | The RR in all RCTs is 1.03 (95% CI 0.78 to 1.37). | ||
| 1.0% | 1.2% (0.9 to 1.5) | 0.2% more (0.1 fewer to 0.5 more) | ||||
| Breast cancer risk No. of participants: 802 (1 RCT) | RR 2.04 (0.44 to 9.55) | Study population | ⊕⊕⊕⊝ MODERATEb | The RR in all RCTs is 1.44 (95% CI 0.96 to 2.17). | ||
| 0.7% | 1.5% (0.3 to 7.0) | 0.8% more (0.4 fewer to 6.3 more) | ||||
| Bladder cancer risk No. of participants: 19,009 (2 RCTs) | RR 1.07 (0.76 to 1.52) | Study population | ⊕⊕⊕⊕ HIGH | SELECT study had the strongest influence on the effect estimate. The RR in all RCTs is 1.10 (95% CI 0.79 to 1.52). | ||
| 0.6% | 0.7% (0.5 to 1.0) | 0.0% fewer (0.2 fewer to 0.3 more) | ||||
| Prostate cancer risk No. of participants: 18,942 (4 RCTs) | RR 1.01 (0.90 to 1.14) | Study population | ⊕⊕⊕⊕ HIGH | SELECT study had the strongest influence on the effect estimate. The RR in all RCTs is 0.91 (95% CI 0.75 to 1.12). | ||
| 5.4% | 5.4% (4.8 to 6.1) | 0.1% more (0.5 fewer to 0.8 more) | ||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; OR: odds ratio; RCT: randomised controlled trial; RR: risk ratio; SELECT: Selenium and Vitamin E Cancer Prevention Trial. | ||||||
| GRADE Working Group grades of evidence. High quality: We are very confident that the true effect lies close to that of the estimate of the effect. Moderate quality: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low quality: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect. Very low quality: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect. | ||||||
aDowngraded one level for moderate heterogeneity (tau² = 0.69, I² = 72%, P = 0.06) not explained. bDowngraded one level owing to imprecision.
Summary of findings 2. Highest compared with lowest selenium exposure for preventing cancer in observational studies.
| Highest compared with lowest selenium exposure for preventing cancer in observational studies | ||
| Patient or population: Participants in non experimental cohort studies on selenium and cancer Setting: out‐patient Intervention: highest selenium exposure Comparison: lowest selenium exposure | ||
| Outcomes | Relative effect (95% CI) | Certainty of the evidence (GRADE) |
| Any cancer risk No. of participants: 76,239 (7 observational studies) |
OR 0.72 (0.55 to 0.93) | ⊕⊝⊝⊝ VERY LOWa |
| Cancer mortality risk No. of participants: 183,863 (7 observational studies) |
OR 0.76 (0.59 to 0.97) |
⊕⊝⊝⊝ VERY LOWa |
| Colorectal cancer risk No. of participants: 712,746 (6 observational studies) |
OR 0.82 (0.72 to 0.94) | ⊕⊝⊝⊝ VERY LOWa |
| Lung cancer risk No. of participants: 371,067 (11 observational studies) |
OR 0.82 (0.59 to 1.14) | ⊕⊝⊝⊝ VERY LOWa,b,c |
| Breast cancer risk (women) No. of participants: 169,028 (8 observational studies) |
OR 1.09 (0.87 to 1.37) | ⊕⊝⊝⊝ VERY LOWa,c |
| Bladder cancer risk No. of participants: 279,100 (5 observational studies) |
OR 0.67 (0.46 to 0.97) | ⊕⊝⊝⊝ VERY LOWa,c |
| Prostate cancer risk No. of participants: 576,667 (21 observational studies) |
OR 0.84 (0.75 to 0.95) | ⊕⊝⊝⊝ VERY LOWa,d |
| CI: confidence interval; OR: odds ratio. | ||
| GRADE Working Group grades of evidence. High certainty: We are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect. Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect. | ||
aDowngraded one level owing to risk of bias, which we deemed as serious because of inability to rule out unmeasured confounding, particularly from lifestyle or nutritional factors that might covary with selenium exposure beyond those factors taken into account in the multi‐variable analyses. bDowngraded one level for moderate heterogeneity (tau² = 0.19, I² = 66%, P = 0.0008) not explained. cDowngraded one level owing to imprecision. dDowngraded one level owing to potential presence of publication bias suggested by the funnel plot.
Background
This review is the third update of the Cochrane review titled "Selenium for preventing cancer" (Dennert 2011; Vinceti 2014).
Description of the condition
Cancer is a leading cause of death worldwide (WHO 2017). According to estimates of the International Agency for Cancer Research, 14.1 million people developed and 8.2 million died of cancer in 2012, and more than half of all new cases occurred in less developed regions of the world (IARC 2014).
The role of diet and nutrition in carcinogenesis and cancer prevention and the identification of nutritional factors and supplements with cancer preventive properties have been areas of active research for decades. Dietary factors that reduce cancer risk would clearly have major public health implications, but unfortunately, investigations into supplementation of various vitamins, trace elements, and other dietary constituents have typically yielded disappointing and even troubling results (Bjelakovic 2014; Fortmann 2013; Guallar 2013; Rocourt 2013; Schwingshackl 2017). Selenium is one of these nutritional factors (Vinceti 2013b).
Description of the intervention
The element selenium has received considerable attention as a potential cancer preventive agent, at least in populations with low intake. Selenium is recognised as nutritionally essential for humans, but it is toxic at levels slightly higher than those required for health, with a narrow and still not well‐defined safe range of intake (Jablonska 2015a; Vinceti 2017a). Whether selenium provides various health benefits (including a cancer preventive effect) beyond its essential nutritional role continues to be a matter of debate (Allingstrup 2015; Bodnar 2012; Brigelius‐Flohe 2017; Fortmann 2013; Karp 2013; Lippman 2009, in: SELECT 2009; Rayman 2012; Stranges 2010;Vinceti 2013a; Vinceti 2013b; Vinceti 2014a; Vinceti 2017a; Visser 2017; Wichman 2016). Humans usually ingest this trace element with crop, animal products, fish, and seafood, and sometimes in supplements (Hurst 2013a; Vinceti 2017a).
Chemical forms and concentrations of selenium in environmental matrices, foods, drinking water, and other sources of exposure vary considerably (Fairweather‐Tait 2011). Selenium species can be classified into organically bound selenium forms (e.g. selenomethionine, selenocysteine) and inorganic forms (e.g. selenate, selenite) (Gammelgaard 2011; Weekley 2013). Organically bound selenium is present in the large number of selenoproteins identified in living organisms including humans, although the exact activity of some of these proteins remains to be identified (Brigelius‐Flohe 2017; Hatfield 2014; Labunskyy 2014). Selenium yeast refers to a selenium‐enriched yeast medium that usually contains selenium that is almost entirely organically bound, along with a high proportion of selenomethionine (Block 2004; Rayman 2004).
Recommended intake of selenium varies considerably among different regulatory agencies and scientific authorities (Vinceti 2017a). For example, the USA Institute of Medicine recommends daily intake of 55 µg/d for adults (Institute of Medicine 2009), whereas the World Health Organization (WHO) recommends amounts ranging from 25 to 34 µg/d, depending on age and sex (WHO 2004). More generally, international bodies have recommended amounts ranging from 25 to 70 µg/d for the adult population (Vinceti 2017a). The main reason for these differences in recommendations is the differing value and weight given to the proteomic effects of selenium, in particular whether or not selenoproteins sensitive to selenium supply must be up regulated to their maximal level, and whether any adverse health effects may arise at lower selenium intakes than those required to maximise selenoprotein expression (Jablonska 2015a; Vinceti 2017a). In addition, these standards generally do not take into account the chemical forms nor the source of selenium (diet, drinking water, air, etc.), despite established relevance of selenium speciation in addressing and assessing the health effects of this element (Vinceti 2013a; Vinceti 2013c; Weekley 2013; Vinceti 2017d).
To prevent adverse effects due to excessive selenium intake, the USA Institute of Medicine has set the tolerable upper intake level at 400 µg/d for adults (Office of Dietary Supplements 2009). However, recent epidemiological studies suggest overt human toxicity at lower intake levels (Lippman 2009, in: SELECT 2009; Stranges 2007; Vinceti 2017a), and lower upper safe levels have already been proposed (Tsubota‐Utsugi 2012). In addition to the acute and chronic toxicity of high selenium exposure, possible harmful effects of long‐term overexposure to lower dosages have been a matter of concern. However, these effects, such as those affecting the endocrine system, remain inadequately investigated (Vinceti 2001; Vinceti 2017a). Furthermore, evidence shows different biological activities of the various organic and inorganic forms of selenium (Hazane‐Puch 2013; Mandrioli 2017; Vinceti 2013c; Vinceti 2017d; Weekley 2013), emphasising the need to better characterise the specific toxicological and nutritional properties of each selenium species in humans, in animals, and in the environment. Recent publications have questioned the adequacy of the current upper safe limit of intake (Jablonska 2015a; Jerome‐Morais 2011; Marschall 2017; Morris 2013; Moyad 2012; Rocourt 2013; Sacco 2013; Vinceti 2013b; Vinceti 2017a) and have espoused the need to set different limits for the many different sources of organic and inorganic selenium. On the other hand, other investigators have described claims of widespread deficient intake of selenium (Hughes 2016).
Accurate estimation of selenium exposure in epidemiological studies presents several challenges. Individual exposure is typically assessed by using peripheral biomarkers of exposure, such as blood (usually plasma or serum) or nail concentrations, or by estimating dietary intake (Ashton 2009). Each of these methods has strengths and limitations and has had its validity questioned (Ashton 2009; Haldimann 1996; Vinceti 2013b). However, levels of selenium in peripheral biomarkers such as blood, toenail, and hair have been found to correlate to a moderate degree with dietary intake as assessed through self‐reported consumption of supplements, food frequency questionnaires, and dietary records (Hurst 2013a; Longnecker 1996; Ovaskainen 1993; Pestitschek 2013; van den Brandt 1993). Stronger correlation has been seen at high intake levels (Morris 2013), although results of some studies were not consistent (Hunter 1990; Karita 2003; Satia 2006; Vinceti 2012). Assessment of selenium levels in specific body tissues is extremely complex, as these levels are not necessarily homogeneously reflected by all biomarkers because overall selenium exposure, as well as its chemical forms and other factors, influences distribution of the metalloid into various body compartments (Behne 1996; Behne 2010; Panter 1996; Vinceti 2000; Vinceti 2013c). For example, circulating levels of some selenium species and of total selenium did not correlate with selenium content in the central nervous system as assessed by cerebrospinal fluid concentrations (Solovyev 2013; Vinceti 2013c), indicating both the tissue‐specific significance of biomarkers and the importance of selenium speciation when the distribution of selenium in different body compartments, representing target organs for different diseases, is assessed.
Selenium levels found in human specimens and characterising intake of selenium show high global variability due to variation in factors such as dietary habits, food and soil selenium content, ethnicity, sex, age, individual metabolism, occupational exposure, exposure to coal and other sources of combustion, and smoking (Fairweather‐Tait 2011; Haldimann 1996; Jablonska 2013; Rayman 2008). It is interesting to note that smoking tends to lower selenium biomarker concentrations, even though smoking is a source of selenium exposure ‐ a phenomenon that might be related to increased excretion of the metalloid due to interaction with cadmium or other heavy metals (Jossa 1991; Kafai 2003). Globally, inconsistencies have been noted as to how these factors are associated with selenium levels (Haldimann 1996; Vinceti 2000). For example, selenium levels increased with age in women, but not in men, in the French SU.VI.M.AX cohort study (Arnaud 2007), but decreased with age in a female population in Ohio (Smith 2000); however, two studies in Switzerland and Austria could not find an association between age and selenium status among individuals of either sex (Burri 2008; Gundacker 2006). Sex‐specific nutritional and health behaviours, as well as sex‐specific differences in selenium metabolism and distribution across various body compartments, may contribute to observed discrepancies in selenium levels between men and women (Combs 2012; Rodriguez 1995).
How the intervention might work
The ability of selenium to counteract cancer cell growth as observed in a large number of laboratory studies may be due to its effects on DNA stability, cell proliferation, necrotic and apoptotic cell death in healthy and malignant cells, and/or regulation of oxidative stress and the immune system (for reviews, see: Fernandes 2015; Misra 2015). These abilities have suggested the possible utility of selenium compounds not only for cancer prevention but also for cancer therapy ‐ a hypothesis that has been under active investigation (Bhattacharjee 2017; Shigemi 2017; Vinceti 2017b). Selenium may be involved in cancer prevention through the antioxidant properties of selenoproteins (Hatfield 2014; Labunskyy 2014), as well as through several other mechanisms (Fernandes 2015; Misra 2015;Weekley 2013). However, laboratory studies have shown that selenium can promote malignant cell transformation and progression (Chen 2000; Kandas 2009; Kasaikina 2013; National Toxicology Program 2011; Novoselov 2005; Rose 2014; Su 2005; Tsuji 2015), thus confirming the complex ‘dual personality’ of both this Janus‐faced element and selenoproteins in preventing and promoting cancer (Hatfield 2014).
In addition, numerous epidemiological studies of observational design, which have reported an inverse association between selenium exposure and cancer risk (Vinceti 2017b), have provided support for the potential of selenium in cancer prevention. The first of these studies used an ecological study design (Schrauzer 1977; Shamberger 1969). These were followed by case‐control and cohort observational studies, then by randomised trials, some of which received substantial attention from both the general public and the scientific community (Brinkman 2006; Fortmann 2013; Steinbrenner 2013; Vinceti 2013b). Some observational and experimental human studies have suggested that sex‐related differences regarding effects of selenium on cancer risk, as well as differences in selenium tissue distribution, tumour biology, and other factors, may explain the possibly greater beneficial effect of selenium for men than for women in the earliest studies (NPCT 2002; Waters 2004).
Why it is important to do this review
Findings of laboratory studies and early epidemiological studies have led to the suggestion that selenium may be involved in central anticarcinogenic processes. This has resulted in widespread marketing of selenium supplements with associated health claims, particularly claims for prevention of cancer (Dennert 2011; Vinceti 2013b), as well as prevention of cardiovascular disease (Rees 2013). However, accumulating evidence suggests that this early optimism may have been unwarranted (Kryscio 2017; Lance 2017; Lu 2016; Ramamoorthy 2015; Vinceti 2017a; Vinceti 2017b). In particular, additional evidence on selenium and cancer risk gathered by high‐quality randomised controlled trials (RCTs) has become available in recent years, and a few observational studies have been published, thus justifying an update on epidemiological evidence regarding selenium exposure and cancer risk. We undertook this updated review to perform a comprehensive synthesis of current epidemiological evidence.
Objectives
To gather and present evidence needed to address two research questions:.
What is the aetiological relationship between selenium exposure and cancer risk in humans?
Which is the efficacy of selenium supplementation for cancer prevention in humans?
Methods
Criteria for considering studies for this review
Types of studies
We included published randomised controlled trials (RCTs) and observational studies of longitudinal design (i.e. cohort studies and nested case‐control studies), irrespective of publication status or language, provided they were published in extenso. We also included conference abstracts in this review when we were able to retrieve them through citation chasing (Vinceti 2017c).
Types of participants
Adult participants (18 years of age and older).
Types of interventions
We considered RCTs for inclusion if they used selenium supplementation at any dose or route of administration for a minimum of four weeks versus placebo or no intervention. We excluded trials using selenium supplementation as part of a multi‐component preparation if they did not include a study arm using selenium monotherapy supplementation.
We considered prospective observational studies (cohort studies and cohort‐nested and nested case‐control studies) for inclusion if they assessed baseline exposure to selenium in apparently cancer‐free individuals as a biomarker of selenium status or as dietary assessment of selenium intake at study entry, provided that such assessment was based on exposure categories ‐ not just on continuous values.
Types of outcome measures
We systematically analysed all (primary and secondary) outcomes.
Primary outcomes
Incidence of any cancer and of site‐specific cancers, assessed as proportions of participants developing cancers during the study period.
Mortality from any cancer and from site‐specific cancer, assessed as proportions of participants dying from cancers during the study period.
Secondary outcomes
Incidence of selected adverse effects, assessed as proportions of participants developing adverse health conditions (RCTs only).
Search methods for identification of studies
Using the search strategies described previously, we conducted updated electronic searches of the Cochrane Central Register of Controlled Trials (CENTRAL; 2017, Issue 2), MEDLINE (Ovid, 2013 to January 2017, week 4), Embase (2013 to 2017, week 6), CancerLit (cancer literature database; February 2004), and Clinical Contents in Medicine (CCMed; February 2011). We conducted the initial search in 2004 and updated searches in July 2007, January 2009, October 2009, February 2011, February 2013, and February 2017. As MEDLINE now includes the journals indexed in CancerLit no further searches of this database were made after 2004.
We also searched the following online clinical trials databases as in the previous review Vinceti 2014.
Clinical Trials of the American Cancer Society (http://www.cancer.gov; February 2011).
metaRegister of Controlled Trials (http://www.controlled‐trials.com; February 2011).
German Cancer Study Register (http://www.studien.de; February 2011).
System for Information on Grey Literature in Europe (SIGLE) (February 2004, discontinued in 2005).
International Standard Registered Clinical/Social Study Number (ISRCTN) registry (http://www.isrctn.com; February 2017).
ClinicalTrials.gov registry (https://clinicaltrials.gov; February 2017).
We have provided the search strategies in Appendix 1.
Data collection and analysis
Selection of studies
Two review authors independently checked all electronic search results for eligibility. When search results could not be rejected with certainty on the basis of title, abstract, or both, we obtained full‐text material.
We scanned bibliographies of papers retrieved using the described search strategy to identify additional studies. When additional information was needed, we contacted the correspondent authors of included studies; we also asked investigators for information about unpublished RCTs.
Two review authors (MV and TF) independently applied the inclusion and exclusion criteria, if necessary with the assistance of a translator. We resolved disagreements by discussion and with involvement of a third review author (CDG).
Data extraction and management
We used piloted extraction forms for epidemiological studies and RCTs to document data from the original material and to assess the quality of studies. One review author (TF) extracted data, and two review authors (MV and CDG) checked extracted data for discrepancies, which the three review authors (TF, MV, and CDG) then discussed. If several reports from the same study were available, we considered as primary publications studies reporting the entire period of follow‐up with active selenium supplementation, when available, but we also extracted study details and results available from other publications, if they were not reported in the primary study reference.
For comparison of selenium exposure measured in serum and plasma specimens, we converted all data into the unit µg/L. We converted results provided as ppm (parts per million) or µg/g by using the factor 1.026 g/mL (density of blood plasma), and we converted data provided as µmol/L using the factor 78.96 (atomic weight of selenium).
For inclusion, prospective observational studies had to report estimates of risk ratio (RR), such as hazard ratio (HR) or odds ratio (OR), for various selenium category exposure levels. We did not include in the analysis studies reporting only the RR for a one‐unit increase in selenium exposure on a continuous scale.
Assessment of risk of bias in included studies
Randomised controlled trials
We categorised generation of allocation sequence, allocation concealment, blinding, and completeness of outcome data as adequate (low risk of bias), inadequate (high risk of bias), or unclear, according to the criteria specified in the Cochrane Handbook for Systematic Reviews of Interventions and Higgins et al (Higgins 2011a; Higgins 2011b). We considered these four items to be key domains for risk of bias assessment. We considered studies that were categorised as 'adequate' in all four domains to have low risk of bias; and studies with 'inadequate' procedures in one or more key domains to have high risk of bias. We considered studies with 'unclear' procedures in one or more key domains to have unclear risk of bias.
We assessed fulfilment of ethical standards as follows.
Was informed consent obtained from participants? (yes/no/unclear).
Was approval obtained from an ethics board? (yes/no/unclear).
Observational studies
We assessed risk of bias in observational studies by using assessment forms adapted from the Newcastle‐Ottawa Quality Assessment Scale (NOS) for cohort and case‐control studies (Wells 2004). We used the NOS form for cohort studies for all included observational studies, and the NOS case‐control form for nested case‐control studies. Both forms must be adapted a priori for use in a systematic review according to the research questions examined and the review topic explored. The NOS uses a star system by which studies are judged on key domains pertaining to selection and comparability of study groups, ascertainment of exposure and outcomes, and duration of follow‐up. For each domain, we assigned either a 'star' or 'no star', with a star indicating that study design element was considered adequate and was less likely to introduce bias. A study could receive a maximum of nine stars during the cohort assessment (Appendix 2) and nine stars during assessment of the case‐control portion (Appendix 3).
The risk of bias assessment was based on data provided in the included publications. When relevant data for such assessment were missing, we tried to contact the trial authors to ask that they provide them.
Measures of treatment effect
This review includes only the binary outcome of cancer diagnosis (i.e. cancer incidence) or death from cancer (i.e. cancer mortality), or a combination of both. We used the term 'cancer risk' in this paper as a generic term that refers generally to cancer incidence, cancer mortality, and combined incidence/mortality data.
For RCTs, we used risk ratios (RRs) and their 95% confidence intervals (95% CIs). When hazard ratios (HRs) rather than RRs were reported in the original study, we reported individual study results as HRs along with their 95% CIs.
For observational studies, we used odds ratios (ORs), risk ratios (RRs), or hazard ratios (HRs) and their corresponding 95% CIs as measures of association between cancer risk and selenium exposure. When adjusted estimates were reported, we used those with the most extensive covariate adjustment reported in the publication.
Dealing with missing data
When data were missing or when discrepancies in study publications were found, we tried to contact the study investigators to request further information. In most cases, review authors resolved the issues through collaboration; when no reply came from the trial authors, we did not use the corresponding data.
When a study combined subgroups, only some of which fulfilled our eligibility criteria (e.g. including individuals not affected by cancer), or did not report enough information to be included in this update, we systematically contacted trial authors to ask that they provide the additional information. We are grateful to the several trial authors who agreed to provide these additional data.
Assessment of heterogeneity
We used the Chi² test for heterogeneity and I² statistics to quantify heterogeneity of study results (Higgins 2003).
Assessment of reporting biases
We evaluated the possibility of reporting bias by using funnel plots.
Data synthesis
We performed data synthesis and analysis separately for RCTs and observational studies.
For RCTs, we performed meta‐analyses for all cancers or site‐specific cancers when at least two trials could provide data, given their fundamental importance in epidemiological research. When more than one publication from the same trial was available and reported different periods of follow‐up for the same cancer site, we included in the meta‐analysis only the longest period of follow‐up, provided that the experimental protocol was ongoing at the time of follow‐up (i.e. that selenium supplementation was still actively supplied). We assessed the stability of effect estimates through their 95% or 99% confidence intervals. We included lack of precision of effect estimates among the factors used to downgrade the certainty (quality) of evidence generated by studies via the GRADE approach (www.gradeworkinggroup.org). For RCTs, we considered pooled data from studies with low risk of bias as the primary analysis.
For observational studies, the minimum number of studies for inclusion in the meta‐analysis was five, as in the previous version of the review. We applied this latter restriction not only to limit the number of analyses performed, but also because results were largely expected to be heterogeneous, and heterogeneity cannot be described and quantified adequately if too few studies are available (Higgins 2009).
We calculated RRs and 95% CIs using numbers of participants and cases when these were provided in the publication and the meta‐analysis tool provided by Review Manager 2014; otherwise, we used RRs reported in the original publication, and, in particular, we selected RRs with the least adjustment for potential confounders. We used the same approach in calculating the RRs of adverse outcomes. We conducted random‐effects meta‐analyses of summary statistics for both observational studies and RCTs. For observational studies, we used the OR or RR comparing highest and lowest selenium exposure categories. We entered effect estimates as the natural logarithm of the OR or RR, and we used the squared standard error of the natural logarithm of the OR or RR as a weight. We calculated the latter from reported upper and lower boundaries of the 95% CI of the OR or RR. If a 95% CI was not reported, we used the total number of cases and the total number of controls, as well as the number of categories of selenium exposure, to estimate numbers of cases and controls per exposure category. We then used the standard normal approximation formula to calculate the standard error of the OR, comparing the highest versus the lowest exposure category (lnOR = (1/a + 1/b + 1/c + 1/d), where a, b, c, and d are the four counts needed to calculate the OR via (a*d)/(b*c)). For experimental studies, we computed the RR of cancer in the intervention group compared with that in the placebo group. For one study, which included more than one treatment (Algotar 2013), we used only results for the lowest dose (200 µg/d) for consistency with other studies. We conducted all meta‐analyses by using Review Manager 5.3.5 and Stata‐15 statistical tools. To do this, we copied logarithmic data for the OR and the standard error from Stata into Review Manager, then double‐checked results for errors.
Subgroup analysis and investigation of heterogeneity
We carried out a subgroup meta‐analysis for high‐quality RCTs while excluding from analysis all trials showing high or uncertain risk of bias.
For observational studies, we used sex‐disaggregated data from mixed‐sex studies, together with data from single‐sex cohorts, to conduct subgroup analyses by sex. We also carried out subgroup analyses specific for baseline selenium status. For these analyses, we assessed the evidence for an exposure‐response relation by examining studies in ascending order from the bottom category of selenium exposure and by examining differences between highest and lowest exposure categories.
Sensitivity analysis
For RCTs, we considered risk estimates derived by pooling data from all studies, regardless of risk of bias, as part of a sensitivity analysis.
For observational studies, we conducted sensitivity analyses to assess the effects of different methods used to assess selenium status (i.e. assessment of intake via dietary assessment methods or measurement of exposure biomarkers such as blood and toenail selenium content).
'Summary of findings' table
We presented the overall certainty (quality) of evidence for the risk of any cancer, cancer mortality, colorectal cancer, lung cancer, non‐melanoma skin cancer, breast cancer, bladder cancer, and prostate cancer from RCTs with low risk of bias. We also presented the overall certainty of evidence for these outcomes from observational studies, with the exception of non‐melanoma skin cancer.
We evaluated the overall certainty of evidence according to the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach (GRADE Working Group 2004), which takes into account issues related not only to internal validity (risk of bias, inconsistency, imprecision, publication bias) but also to external validity, such as directness of results (Langendam 2013). We created two 'Summary of findings' tables (Table 1; Table 2) while adhering to the methods described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011a) and using GRADEpro GDT We used the GRADE checklist and GRADE Working Group certainty (quality) of evidence definitions (Meader 2014), as follows.
High quality: We are very confident that the true effect lies close to that of the estimate of the effect.
Moderate quality: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different.
Low quality: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect.
Very low quality: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect.
When possible, for each outcome in RCTs, we based the assumed risk in the control group on the proportion of events in the included studies. In accordance with GRADE methodological criteria, we based our assessment of the certainty (quality) of evidence on RCTs with low risk of bias (Guyatt 2011). We downgraded the evidence from 'high' quality by one level for serious (or by two levels for very serious) concerns regarding each of the validity issues.
Results
Description of studies
Citation style: Please note that we reference the sources of relevant information in a certain way to enhance traceability of our results for interested readers. When the source of information is not the primary publication of an included study, we also reference the specific publication of interest. For example "Hakama 1990, in: Knekt 1990" indicates that the cited paper is "Hakama 1990" as part of the mentioned study.
We could not access three full‐text theses published in the United States (Coates 1987, in: Coates 1988; Menkes 1986a, in: Menkes 1986; Schober 1986, in: Menkes 1986). However, later journal publications were available, and we included them in this review as main study publications (Coates 1988, in: Coates 1988; Menkes 1986b, in: Menkes 1986; Schober 1987, in: Menkes 1986). Thus we considered retrieval of the full‐text theses to be unnecessary.
Results of the search
In the previous Cochrane review, of 4082 hits of potential relevance, we retrieved 268 publications in full text. Of these, we considered 137 papers as relevant (see the flow chart of the literature search in Dennert 2011).
In our first updated search, after we excluded internal duplicates and duplicates against the database of the literature search conducted in January 2011, we retrieved 766 hits. Of these, we excluded 744 references as clearly irrelevant on the basis of title and abstract review (see the flow chart of the literature search in Vinceti 2014).
In the second updated search process, conducted in February 2017, including online database searches and searches within grey literature, study references, and trial databases, we identified 859 new hits after de‐duplication. Of these, we excluded 831 references as clearly irrelevant on the basis of the title and abstract review (see the flow chart of the literature search in Figure 1). We considered the remaining 28 publications of possible relevance and re‐evaluated and retrieved them in full text from this updated search. Upon further review, we considered 20 of these publications relevant.
1.
Flow chart.
Included studies
In total, from the previous Cochrane review and from our updates, we identified 168 papers for inclusion in this review: 105 papers referred to 70 completed observational studies, and 63 papers referred to one ongoing and 10 completed RCTs (Figure 1). (The previous version of the review was based on 148 papers; 89 referred to one ongoing and 55 completed observational studies, and 59 papers referred to four ongoing and eight completed RCTs.)
We have provided a detailed description of the included studies in the Characteristics of included studies table.
Randomised controlled trials
We included in this review 11 randomised controlled trials (RCTs) with a total of 44,743 participants (94% men). All used parallel‐group designs with two arms (Dreno 2007; Karp 2013; Li 2000; Lubinski 2011; Marshall 2011; NPCT 2002; Reid 2008; Yu 1991; Yu 1997), three arms (Algotar 2013), or four arms (SELECT 2009). Three were conducted in China (Li 2000; Yu 1991; Yu 1997), four in the United States (Karp 2013; Marshall 2011; NPCT 2002; Reid 2008), one in the United States/New Zealand (Algotar 2013), one in the United States/Canada/Puerto Rico (SELECT 2009), and one in Europe (Lubinski 2011).
Investigators administered selenium supplements and placebos daily. As an active intervention, trials used selenium 200 µg/d (Dreno 2007; Karp 2013; Marshall 2011; NPCT 2002; Yu 1991; Yu 1997), or 400 µg/d (Reid 2008), in the form of selenised yeast tablets, composed almost entirely of organic selenium and particularly of selenomethionine (Block 2004). Algotar 2013 used 200 µg and 400 µg as different arms. Li 2000 used 500 µg sodium selenite, and SELECT 2009 used 200 µg/d of selenomethionine. Lubinski 2011 used 250 µg/d of inorganic selenite.
Three Chinese trials investigated the preventive efficacy of selenium supplementation against primary liver cancer for different high‐risk populations. Participants were carriers of the hepatitis B surface antigen (HBs‐Ag) with normal liver function, or they were first‐degree relatives of patients with liver cancer. Two trials used selenised yeast (Yu 1991; Yu 1997), and one used sodium selenite (Li 2000).
The Nutritional Prevention of Cancer Trial (NPCT) investigated the influence of selenium on the development of non‐melanoma skin cancer (basal and squamous cell carcinoma) in a population considered at high risk of the disease, namely, patients with a history of non‐melanoma skin cancer (NPCT 2002). Participants consisted of 1312 men and women from the eastern United States 18 to 80 years of age, with a history of two or more basal cell carcinomas or of one squamous cell carcinoma. Investigators reported RR estimates for basal cell carcinoma, squamous cell carcinoma, and overall non‐melanoma skin cancer for two periods of follow‐up: an intermediate study period (from 15 September 1983 to 31 December 1993: Clark 1996, in: NPCT 2002), and the entire blinded intervention period (from 15 September 1983 to 31 January 1996: Duffield‐Lillico 2002 for secondary outcomes; Duffield‐Lillico 2003a for the primary outcome, i.e. non‐melanoma skin cancer; and Duffield‐Lillico 2003b for an in‐depth analysis of prostate cancer risk; see NPCT 2002). In the present analysis, we used only final reports concerning the entire period of blinded follow‐up, which was characterised by active administration of selenium supplements.
In 1990, NPCT 2002 identified additional secondary endpoints post hoc (i.e. total cancer mortality; total cancer incidence; incidence of lung, prostate, and colorectal cancers). Trial publications also reported incidences of female breast cancer, bladder cancer, oesophageal cancer, melanoma, haematological cancer, and cancers of the head and neck (NPCT 2002).
A substudy of the NPCT investigated the efficacy of a higher selenium dose, supplied as selenised yeast orally, for prevention of non‐melanoma skin cancer at one of the NPCT study sites (Reid 2008). Study design was similar to that of the NPCT study, except that investigators randomly assigned 423 participants at this site to placebo or intervention with 400 µg/d of selenium. Reid 2008 also reported the incidence of internal cancers.
Dreno 2007 evaluated the incidence of skin cancer as a secondary outcome in a group of 184 organ transplant recipients who received 200 µg/d of selenium for three years, then were followed up for an additional two years. In this multi‐centre, randomised, placebo‐controlled trial, investigators monitored 91 selenium‐supplemented participants and 93 non‐supplemented participants for development of both non‐malignant (warts and various keratoses) and malignant skin lesions.
The Selenium and Vitamin E Cancer Prevention Trial (SELECT 2009) investigated the effect of selenium as L‐selenomethionine and/or vitamin E supplementation in men of diverse ethnic backgrounds against the development of prostate cancer and other 'secondary' outcomes (i.e. risk of all cancers, lung cancer, colorectal cancer, and bladder cancer). This study was a very large phase 3 randomised, placebo‐controlled trial, activated in June 2001 and originally designed for a 7‐ to 12‐year period of follow‐up, carried out at 427 sites in the United States, Canada, and Puerto Rico. However, the independent Data and Safety Monitoring Committee (DSMC) recommended on 15 September 2008, discontinuation of study supplements based on absence of benefit from vitamin E or selenium and no possibility of benefit to the planned degree with additional follow‐up (SELECT 2009). The Committee also expressed concern about increased prostate cancer risk among vitamin E–treated participants and increased diabetes risk among selenium‐supplemented participants (SELECT 2009) (RR 1.07, 99% CI 0.94 to 1.22). Therefore, investigators discontinued administration of these supplements on 23 October 2008, in spite of the planned supplementation period of 12 years. Results of SELECT are based on follow‐up provided at the end of the blinded supplementation period, which included 117,660 person‐years of follow‐up ‐ not on an extended period of follow‐up, which encompassed an additional 32 months of surveillance (144,846 person‐years in total) after the end of the supplementation period (Klein 2011, in: SELECT 2009). Endpoints were prostate cancer (the 'primary' endpoint) and colorectal cancer, lung cancer, all other cancers, and all cancers overall. A subsequent study from SELECT also evaluated the risk of bladder cancer, adding to standard follow‐up an additional post supplementation period of 32 months (Lotan 2012, in: SELECT 2009).
Three phase III trials published in 2011 ‐ Marshall 2011 ‐ and in 2013 ‐ Algotar 2013; Karp 2013 ‐ also evaluated the effect of selenium supplementation on prostate cancer. In Marshall 2011 (trial code SWOG S9917), investigators randomly assigned 423 men with high‐grade prostatic intraepithelial neoplasia, and therefore considered to be at very high risk of prostate cancer, to selenium (200 µg/d as selenomethionine) or placebo. Algotar 2013 evaluated whether supplementation with 200 or 400 µg/d of selenium as selenised yeast reduced the risk of prostate cancer among men at high risk of the disease, based on a prostate‐specific antigen (PSA) level exceeding 4 ng/L, suspicious digital rectal examination. and PSA velocity greater than 0.75 ng/mL/y. This trial, called the Negative Biopsy Trial (NBT), followed study participants in the United States (where both supplementation and follow‐up were completed for such period) for five years, and in New Zealand for no longer than three years, and was discontinued after an external DSMC issued a recommendation to stop the trial. Karp 2013 investigated the effect of supplementation of 200 µg/d selenium as selenised yeast in 1561 individuals with resected stage I non–small‐cell lung cancer (trial code ECOG 5597). The primary outcome was the incidence of second primary tumours. Investigators enrolled both men and women in the study and investigated all cancer types and a few major side effects during follow‐up. Follow‐up included the period of active supplementation and some additional follow‐up after the trial anticipated discontinuation. This decision was made by the trial DSMC, which, on October 21, 2009, reviewed the first planned interim analysis of the primary endpoint and recommended that the study should be terminated for futility. Based on that DSMC recommendation, on November 5, 2009, accrual for the Eastern Cooperative Oncology Group (ECOG) trial was interrupted, and all current participants were invited to discontinue selenium/placebo tablets and were monitored only for follow‐up of cancer incidence and survival. In accordance with recommendations by the trial DSMC concerning possible adverse effects of selenium supplementation, the incidence of basal and squamous cell skin cancers, as well as type 2 diabetes, was monitored. The main paper reported follow‐up until June 2011 (Karp 2013), and results for only second primary lung tumours were updated as of January 2014, including a longer post supplementation period of follow‐up (Pillai 2014, in: Karp 2013).
Investigators conducted a trial in Poland that included a female population of carriers of a breast cancer‐related mutation, BRCA1 (Lubinski 2011). Trial authors randomised 1135 women carrying that mutation to 250 µg/d of selenium in its inorganic tetravalent form (selenite), or to placebo, in a double‐blind trial. Median follow‐up lasted 35 months (ranging from 6 to 62 months), and final analysis was based on 105 incident cases diagnosed during follow‐up ‐ 60 cases in the selenium‐supplemented arm and 45 cases in the placebo arm.
Observational studies
We included in this review 70 completed observational studies. Forty‐five studies were nested case‐control studies, the others were subcohort‐controlled or cohort studies, and one study used a cohort together with a nested case‐control design. Subcohort‐controlled studies used (random) samples of the cohort as controls. The original papers were published between 1983 and 2017. Eight studies were conducted in Asia (China, Iran, Japan, and Taiwan), one in Australia, 30 in Europe (Belgium, Denmark, Germany, Greece, Italy, Netherlands, Norway, Spain, Sweden, Channel Islands, Finland, France, and UK), 30 in the United States, and one in Canada. Overall, studies included more than 2,300,000 participants. Study populations in Europe made up 42.9%, North America 44.3%, Asia 11.4%, and Australia 1.4% of all study participants. The median size of study populations was 11,457. Forty‐one studies included men and women, one did not report sex, 22 included only men, and six included only women. Eleven studies with mixed‐sex populations reported results stratified by sex. Study populations were derived from 55 different cohorts. Twenty‐four cohorts were non‐randomly recruited (e.g. included volunteers), and 31 cohorts consisted of a random sample of the population of interest. Fifty‐two studies reported mean or median age, 12 studies reported only age range, and six studies did not report this information on study participants. Most studies included adults older than 40 years of age.
Sixteen studies investigated nutritional and/or supplemental selenium intake by using food frequency questionnaires or interviews. Fifty‐four studies assessed biochemical selenium status whereby:
9 used toenail specimens;
14 used plasma specimens;
29 used serum specimens;
1 used both serum and plasma specimens; and
1 measured both serum selenium levels and intake.
The mean follow‐up period lasted up to three years in five studies, and longer than three years in the remaining studies. Generally, study authors grouped cases according to the version of the International Classification of Diseases (ICD) that was up‐to‐date at the inception of the cohort observation. The level of disaggregation of data varied markedly between studies. Although some studies reported cancer risk according to organ system (e.g. urinary tract, respiratory tract), others reported cancer risk for one or two organs (e.g. female breast, urinary bladder). Only in the case of skin cancer did studies also differentiate according to histological type (e.g. melanoma, basal cell carcinoma).
For the following outcomes, we included five or more studies in the review and meta‐analysed observational data.
Any cancer (16 studies).
Female breast cancer (8 studies).
Urinary bladder cancer (6 studies).
Lung cancer (15 studies).
Prostate cancer (21 studies).
Stomach cancer (5 studies).
Colorectal cancer (6 studies) and colon cancer (5 studies).
Goyal 2013 updated results of Bleys 2008, which reported longer follow‐up for the same population.
Table 3 provides an overview of the studies examining each outcome. Five studies provided data for the group of 'other' cancers, which encompassed any type of cancer not reported separately in study publications. The definition of 'other' cancers varied between studies, including rare cancers but also cancers of unknown origin. We have mentioned results of studies within the category 'other cancers' for the sake of completeness; however, because of the diversity of outcomes, we have not included these results in further analysis or discussion of this review.
1. Included observational studies by outcome.
| Organ system | Outcome | Number of studies/case definitions | Meta‐ analysis | Countries | Number of participants | Number of cases | Selenium assessment | Reporting study |
| Any cancer | Any cancer | total: 16 incidence: 7 mortality: 7 incidence and mortality combined: 1 |
✓ yes | USA
Finland
Netherlands
Sweden
Norway
Belgium
France China Japan |
total: ˜ 276,000 | total: 6488 male: 3196 female: 1541 |
serum: 12 plasma: 2 serum + plasma: 1 dietary intake: 1 |
Willett 1983 Salonen 1984 Peleg 1985 Salonen 1985 Nomura 1987 Virtamo 1987 Coates 1988 Fex 1987 Kok 1987a Ringstad 1988 Knekt 1990 Kornitzer 2004 Akbaraly 2005 Bleys 2008 Fujishima 2011 Sun 2016 |
| Gynaecological cancer | Female breast cancer | total: 8 incidence: 8 mortality: 0 incidence and mortality combined: 0 |
✓ yes | USA Finland Netherlands Channel Islands | total/female: 169,028 | total/female: 1277 | serum: 2 plasma: 1 serum + plasma: 1 toenail: 3 intake: 1 |
van Noord 1987
Coates 1988 Knekt 1990 Overvad 1991 van den Brandt 1993 Garland 1995 Dorgan 1998 Pantavos 2015 |
| Cervical cancer | total: 2 incidence: 2 mortality: 0 incidence and mortality combined: 0 |
✗ no | USA | total/female: > 15,161 (1 study did not report cohort size by sex) |
total/female: 62 | serum: 2 | Menkes 1986 Coates 1988 | |
| Uterine cancer | total: 1 incidence: 1 mortality: 0 incidence and mortality combined: 0 |
✗ no | USA | total/female: 62,641 | total/female: 91 | toenail: 1 | Garland 1995 | |
| Ovarian cancer | total: 4 incidence: 4 mortality: 0 incidence and mortality combined: 0 |
✗ no | USA Finland | total/female: ˜ 214,000 | total/female: 568 | serum: 2 toenail: 1 supplemental intake: 1 |
Menkes 1986 Knekt 1990 Garland 1995 Thomson 2008 |
|
| Gynaecological cancer (without breast cancer) | total: 1 incidence: 1 mortality: 0 incidence and mortality combined: 0 |
✗ no | Finland | total/female: 18,096 | total/female: 86 | serum: 1 | Knekt 1990 | |
| Urological cancers | Renal cancer | total: 1 incidence: 1 mortality: 0 incidence and mortality combined: 0 |
✗ no | United Kindom | total: 23,658 | total: 65 | dietary intake: 1 | Banim 2013 |
| Urinary bladder cancer | total: 6 incidence: 6 mortality: 0 incidence & mortality combined: 0 |
✓ yes | USA/Hawaii Finland Netherlands | total: 279,100 female: 130,786 male: 128,009 |
total: 1295 female: 175 male 755 |
serum: 3 toenail: 3 |
Menkes 1986
Nomura 1987
van den Brandt 1993
Michaud 2002 Michaud 2005 Hotaling 2011 |
|
| Urinary tract cancer | total: 1 incidence: 1 mortality: 0 incidence & mortality combined: 0 |
✗ no | Netherlands | total: 38,500 | total: 47 male: 34 female: 13 |
serum: 1 | Knekt 1990 | |
| Respiratory tract cancers | Lung cancer | total: 15 incidence: 13 mortality: 2 incidence and mortality combined: 0 |
✓ yes | China
Japan
USA
Finland
Netherlands Denmark |
total: 371,067 male: 125,341 female: 181,895 |
total: 2223 male: 1384 female: 416 |
serum: 9 serum + plasma: 2 toenail: 2 dietary intake: 2 (1 study reported both serum levels and food intake) |
Menkes 1986
Kromhout 1987
Nomura 1987 Coates 1988 Knekt 1990 van den Brandt 1993 Kabuto 1994 Garland 1995 Comstock 1997 Knekt 1998 Ratnasinghe 2000 Goodman 2001 Epplein 2009 Suadicani 2012 Muka 2017 |
| Oral/pharyngeal cancer | total: 1 incidence: 1 mortality: 0 incidence and mortality combined: 0 |
✗ no | USA | total: 20,305 | total: 28 | serum: 1 | Menkes 1986 | |
| Andrological cancers | Prostate cancer | total: 21 incidence: 21 mortality: 0 incidence and mortality combined: 0 |
✓ yes | USA Canada Puerto Rico Europe |
total/male: 576,667 | total/male: 14,950 | serum: 8 plasma: 5 toenail: 4 dietary intake: 4 |
Coates 1988
van den Brandt 1993 Hartman 1998 Yoshizawa 1998 Helzlsouer 2000 Nomura 2000 Brooks 2001 Goodman 2001 Li 2004a Peters 2007 Allen 2008 Peters 2008 Epplein 2009 Kristal 2014 Park 2015 Outzen 2016 Graff 2017 |
| Gastrointestinal cancers | Oesophageal cancer | total: 2 incidence: 2 mortality: 1 incidence and mortality combined: 0 |
✗ no | China USA | total: 29,923 | total: > 959 | serum: 1 supplemental intake: 1 |
Wei 2004 Dong 2008 |
| Oesophageal squamous cell carcinoma | total:2 incidence: 2 mortality: 0 incidence and mortality combined: 0 |
✗ no | Netherlands Iran |
total: 168,257 | total: 265 | toenail: 1 intake: 1 |
Steinbrecher 2010 Hashemian 2015 |
|
| Oesophageal adenocarcinoma | total:1 incidence:1 mortality:0 incidence and mortality combined: 0 |
✗ no | Netherlands | total: 120,852 | total: 112 | toenail: 1 | Steinbrecher 2010 | |
| Oesophageal/stomach cancer | total: 1 incidence: 1 mortality: 0 incidence and mortality combined: 0 |
✗ no | Netherlands | total: 36,265 | total: 86 male: 51 female: 35 |
serum: 1 | Knekt 1998 | |
| Gastric cardia adenocarcinoma | total:1 incidence:1 mortality:0 incidence and mortality combined: 0 |
✗ no | Netherlands | total: 120,852 | total:114 | toenail: 1 | Steinbrecher 2010 | |
| Stomach cancer | total: 5 incidence: 5 mortality: 1 incidence and mortality combined: 0 |
✓ yes | China Japan USA/Hawaii Finland Netherlands | total: ˜ 197,000 male: 86,311 female: 80,669 |
total: 955 male: 626 female: 329 |
serum: 4 toenail: 1 |
Nomura 1987 Knekt 1990 van den Brandt 1993 Kabuto 1994 Wei 2004 |
|
| Primary liver cancer | total: 4 incidence: 3 mortality: 1 incidence and mortality combined: 0 |
✗ no | China Europe Taiwan |
total: 701,809 male: 61,470 female: 74,941 |
total: 877 male: 567 female: 204 |
plasma: 1 serum: 1 toenail: 1 intake: 1 |
Yu 1999
Sakoda 2005 Hughes 2016 Ma 2017 |
|
| Pancreatic cancer | total: 4 incidence: 4 mortality: 0 incidence and mortality combined: 0 |
✗ no | USA
Finland UK |
total: 159,062 | total: 311 male: 69 female: 84 |
serum: 2 intake: 1 supplemental intake: 1 |
Menkes 1986
Knekt 1990 Banim 2013 Han 2013 |
|
| Colorectal cancer | total: 6 incidence: 6 mortality: 0 incidence and mortality combined: 0 |
✓ yes | USA/Hawaii Europe | total: 712,746 male: 216,272 female: 442,266 |
total: 2627 male: 810 female: 797 |
serum: 3 toenail: 2 supplement use: 1 |
Nomura 1987 Knekt 1990 van den Brandt 1993 Garland 1995 Hansen 2013 Hughes 2015 |
|
| Colon cancer | total: 5 incidence: 5 mortality: 0 incidence and mortality combined: 0 |
✓ yes | USA/Hawaii Europe |
total: 636,641 male: 195,100 female: 361,529 |
total: 1677 male: 525 female: 510 |
serum: 3 toenail: 1 supplement use: 1 |
Menkes 1986 Nomura 1987 van den Brandt 1993 Hansen 2013 Hughes 2015 |
|
| Rectal cancer | total: 4 incidence: 4 mortality: 0 incidence and mortality combined: 0 |
✗ no | USA/Hawaii Europe | total: 610,837 male: 195,100 female: 361,529 |
total: 861 male: 303 female: 210 |
serum: 2 toenail: 1 supplement use:1 |
Nomura 1987 van den Brandt 1993 Hansen 2013 Hughes 2015 |
|
| All gastrointestinal cancers | total: 1 incidence: 1 mortality: 0 incidence and mortality combined: 0 |
✗ no | USA | total: 6,167 | total: 143 | plasma and serum: 1 | Coates 1988 | |
| Skin cancer | Melanoma | total: 3 incidence: 3 mortality: 0 incidence and mortality combined: 0 |
✗ no | USA | total: ˜ 158,000 | total: 547 | serum: 1 toenail: 1 supplemental intake: 1 |
Menkes 1986 Garland 1995 Peters 2008 |
| Basal cell carcinoma | total: 3 incidence: 3 mortality: 0 incidence and mortality combined: 0 |
✗ no | Australia USA Finland | total: > 66,000 | total: 292 | serum: 3 dietary intake: 1 |
Menkes 1986 Knekt 1990 McNaughton 2005 |
|
| Squamous cell carcinoma | total: 4 incidence: 4 mortality: 0 incidence and mortality combined: 0 |
✗ no | Australia USA | total: ˜ 30,000 | total: 488 | serum: 2 plasma: 1 dietary intake: 1 |
Menkes 1986 Combs 1993 Karagas 1997 McNaughton 2005 |
|
| Total non‐melanoma skin cancer | total: 1 incidence: 1 mortality: 0 incidence and mortality combined: 0 |
✗ no | USA | total: 117 | total: 19 | plasma: 1 | Clark 1985 | |
| Rare and other cancers | Haematological cancers | total: 1 incidence: 1 mortality: 0 incidence and mortality combined: 0 |
✗ no | USA | total: 6167 | total: 12 | serum + plasma: 1 | Coates 1988 |
| Thyroid cancer | total: 2 incidence: 2 mortality: 0 incidence and mortality combined: 0 |
✗ no | Norway | total: 582,807 male: 287,944 female: 194,863 |
total: 635 male: 269 female: 366 |
serum: 1 intake:1 |
Glattre 1989 O'Grady 2014 |
|
| Other cancers | total: 4 incidence: 3 mortality: 1 incidence and mortality combined: 0 |
✗ no | China USA Finland | total: 109,179 male: 21,172 female: 80,737 |
total: 512 male: 169 female: 285 |
serum: 2 serum + plasma: 1 toenail: 1 |
Coates 1988
Knekt 1990 Garland 1995 Wei 2004 |
Some studies did not report the sex of participants or cancer cases; consequently, figures for women and men do not always sum up to the total number of participants or cancer cases.
Excluded studies
Of 28 potentially relevant papers retrieved in the updated search, eight papers did not fulfil the inclusion criteria. We rejected six of these publications as investigators did not report results according to inclusion criteria; one paper because trial authors reported duplicated data from an already included study; and another paper because the trial was carried out in patients with cancer. The Characteristics of excluded studies table describes the reasons for exclusion of trials from the previous Cochrane review and from this update.
Risk of bias in included studies
Randomised controlled trials
We assessed risk of bias of the included RCTs according to Cochrane criteria (Higgins 2011a; Higgins 2011b). We presented judgements about each risk of bias item as percentages across all included RCTs, and we provided a summary of the risk of bias assessment in Figure 2. We provided details on the judgement for each RCT and the reason for that judgement in Characteristics of included studies.
2.
Review authors’ judgements about each risk of bias item presented as percentages across all included RCTs and summary of review authors’ judgements about each risk of bias item for the included RCTs.
We considered all three trials on liver cancer risk (Li 2000; Yu 1991; Yu 1997), as well as the trial on breast cancer (Lubinski 2011), to have unclear risk of bias. These trials did not report generation of allocation sequence and allocation concealment. One study mentioned that the dropout rate was similar in intervention and control groups; the remaining three studies did not report the completeness of outcome data. We judged blinding as adequate in three studies, as investigators reported the use of placebo supplements. We inferred from this procedure that at least the study participants and the physicians directly involved were blinded towards treatment status.
In addition, it is unclear whether Li 2000 was an individually randomised controlled trial. Study investigators used the phrase "randomisation based on the residence area" and did not describe the randomisation procedure any further. As participants were recruited from 17 villages, these villages ‐ not individual participants ‐ may have been randomly assigned to intervention and control groups. However, we could not make contact with study investigators to clarify these questions. Randomisation of villages instead of individuals could have introduced bias into the study results, as the incidence of liver cancer is known to differ between geographical areas as a result of lifestyle and environmental factors.
It has been found that RCTs with inadequate or unclear allocation concealment, especially those with subjective outcomes, may overestimate the benefit of interventions (Pildal 2007; Wood 2008). All three RCTs on liver cancer did not report follow‐up and case detection procedures, so the influence of subjective factors on case detection, such as interpretation of bodily symptoms as triggers of further diagnostic tests, is unknown. Although we judged blinding as 'adequate' in all three liver cancer trials, we do not know whether blinding was successful in practice for participants, healthcare providers, and outcome assessors.
These uncertainties about study methods seriously weaken our confidence in reported RCT results on liver cancer risk.
We considered Algotar 2013, Karp 2013, Marshall 2011, and SELECT 2009 to have low risk of bias because they reported adequate generation of allocation sequence, allocation concealment, blinding, and completeness of outcome data.
We judged Dreno 2007 and Duffield‐Lillico 2002 to 2003, in: NPCT 2002 to have unclear risk of bias. Dreno 2007 provided unclear generation of allocation sequence, allocation concealment, and blinding; only completeness of outcome data was adequate. We considered NPCT to be at unclear risk of bias because of exposure‐related detection bias for its primary outcome, as the percentage of study participants with an abnormal PSA (> 4 ng/mL) who underwent biopsy varied according to selenium treatment group, at 35% in the placebo group and 14% in the selenium‐treated group (Duffield‐Lillico 2003b, in: NPCT 2002; Marshall 2011). As reported by the trial authors themselves in analyses stratified by baseline selenium concentration, the difference was greatest among participants in the lowest tertile, in whom the inverse association between selenium administration and prostate cancer risk was strongest. The difference in biopsy rates could not be accounted for by factors such as PSA concentration, age at which abnormal PSA was detected, or alternative diagnostic procedures. Although a difference this large could have occurred by chance, this finding raises concerns about possible disruption of blinding. Investigators provided no information as to the prostate biopsy rate among participants with lower PSA levels or biopsy rates for the primary outcome of non‐melanoma skin cancer, which also requires pathological confirmation, nor for the secondary outcomes examined in this trial.
Observational studies
We presented in Table 4 a summary of study ratings according to the Newcastle‐Ottawa Scale (NOS). The median number of assigned stars was eight for both (nested) case‐control and cohort study assessments, out of a maximum of nine stars each.
2. Risk of bias: observational studies.
| Study | Publication | Newcastle‐Ottawa Scale (cohort) | Newcastle‐Ottawa Scale (case‐control) | ||||||
| Selection | Comparability | Outcome | Total | Selection | Comparability | Exposure | Total | ||
| Agalliu 2011 | Agalliu 2011 | 0‐1‐0‐1 | 1 | 1‐1‐0 | 5 | 0‐1‐0‐1 | 1 | 1‐1‐0 | 5 |
| Akbaraly 2005 | Akbaraly 2005 | 0‐1‐1‐1 | 2 | 0‐1‐0 | 6 | .‐.‐.‐. | . | .‐.‐. | . |
| Allen 2008 | Allen 2008 | 1‐1‐1‐1 | 2 | 1‐1‐0 | 8 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 |
| Banim 2013 | Banim 2013 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 |
| Barrass 2013 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 | |
| Bleys 2008 | Bleys 2008 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 | .‐.‐.‐. | . | .‐.‐. | . |
| Goyal 2013 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 | .‐.‐.‐. | . | .‐.‐. | . | |
| Brooks 2001 | Brooks 2001 | 0‐1‐1‐0 | 2 | 1‐0‐0 | 5 | 1‐0‐1‐1 | 2 | 1‐1‐0 | 7 |
| Clark 1985 | Clark 1985 | 0‐1‐1‐0 | 0 | 0‐0‐0 | 2 | .‐.‐.‐. | . | .‐.‐. | . |
| Coates 1988 | Coates 1988 | 0‐1‐1‐0 | 1 | 1‐1‐0 | 5 | 1‐0‐1‐0 | 1 | 1‐1‐1 | 6 |
| Coates 1987 | .‐.‐.‐. | . | .‐.‐. | . | .‐.‐.‐. | . | .‐.‐. | . | |
| Combs 1993 | Combs 1993 | 0‐1‐1‐0 | 2 | 1‐0‐0 | 5 | .‐.‐.‐. | . | .‐.‐. | . |
| Comstock 1997 | Comstock 1997 | 0‐1‐1‐0 | 2 | 1‐1‐0 | 6 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 |
| Dong 2008 | Dong 2008 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 | .‐.‐.‐. | . | .‐.‐. | . |
| Dorgan 1998 | Dorgan 1998 | 0‐1‐1‐1 | 2 | 0‐1‐0 | 6 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 |
| Epplein 2009 | Epplein 2009 | 0‐1‐1‐1 | 2 | 1‐1‐0 | 7 | 0‐1‐1‐1 | 2 | 1‐1‐1 | 8 |
| Gill 2009 | 0‐1‐1‐1 | 1 | 1‐1‐0 | 6 | 0‐1‐1‐1 | 1 | 1‐1‐1 | 7 | |
| Fex 1987 | Fex 1987 | 1‐1‐1‐0 | 2 | 1‐1‐1 | 8 | 1‐0‐1‐1 | 2 | 1‐1‐1 | 8 |
| Fujishima 2011 | Fujishima 2011 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 | .‐.‐.‐. | . | .‐.‐. | . |
| Garland 1995 | Garland 1995 | 0‐1‐1‐1 | 2 | 1‐1‐1 | 8 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 |
| Hunter 1990 | 0‐1‐1‐1 | 2 | 1‐1‐1 | 8 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 | |
| Glattre 1989 | Glattre 1989 | 0‐1‐1‐0 | 1 | 1‐1‐1 | 6 | 1‐1‐1‐1 | 1 | 1‐1‐1 | 8 |
| Goodman 2001 | Goodman 2001 | 0‐1‐1‐0 | 2 | 1‐1‐0 | 6 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 |
| Graff 2017 | Graff 2017 | 0‐1‐1‐1 | 2 | 1‐1‐0 | 7 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 |
| Grundmark 2011 | Grundmark 2011 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 | .‐.‐.‐. | . | .‐.‐. | . |
| Han 2013 | Han 2013 | 0‐1‐0‐1 | 2 | 1‐1‐0 | 7 | .‐.‐.‐. | . | .‐.‐. | . |
| Hansen 2013 | Hansen 2013 | 0‐1‐1‐1 | 1 | 1‐1‐1 | 7 | .‐.‐.‐. | . | .‐.‐. | . |
| Hartman 1998 | Hartman 1998 | 1‐1‐0‐1 | 2 | 1‐1‐0 | 7 | .‐.‐.‐. | . | .‐.‐. | . |
| Hashemian 2015 | Hashemian 2015 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 | .‐.‐.‐. | . | .‐.‐. | . |
| Helzlsouer 2000 | Helzlsouer 2000 | 0‐1‐1‐1 | 1 | 1‐1‐0 | 6 | 1‐1‐1‐1 | 1 | 1‐1‐1 | 8 |
| Hughes 2015 | Hughes 2015 | 1‐1‐1‐1 | 2 | 0‐1‐0 | 7 | 0‐1‐1‐1 | 2 | 1‐1‐1 | 8 |
| Hughes 2016 | Hughes 2016 | 1‐1‐1‐1 | 2 | 0‐1‐1 | 8 | 0‐1‐1‐1 | 2 | 1‐1‐1 | 8 |
| Kabuto 1994 | Kabuto 1994 | 0‐1‐1‐1 | 2 | 1‐1‐0 | 7 | 0‐1‐1‐1 | 2 | 1‐1‐1 | 8 |
| Karagas 1997 | Karagas 1997 | 0‐1‐1‐1 | 2 | 1‐1‐1 | 8 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 |
| Knekt 1990 | Knekt 1990 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 | 0‐1‐1‐1 | 2 | 1‐1‐1 | 8 |
| Hakama 1990 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 | 0‐1‐1‐1 | 2 | 1‐1‐1 | 8 | |
| Knekt 1988 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 | 0‐0‐1‐1 | 2 | 1‐1‐1 | 7 | |
| Knekt 1996 | 1‐1‐1‐1 | 1 | 1‐1‐1 | 8 | 0‐1‐1‐1 | 1 | 1‐1‐1 | 7 | |
| Knekt 1991 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 | 0‐1‐1‐1 | 2 | 1‐1‐1 | 8 | |
| Knekt 1998 | Knekt 1998 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 | 0‐1‐1‐1 | 2 | 1‐1‐1 | 8 |
| Kok 1987a | Kok 1987b | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 | 1‐0‐1‐1 | 2 | 1‐1‐1 | 8 |
| Kok 1987a | .‐.‐.‐. | . | .‐.‐. | . | .‐.‐.‐. | . | .‐.‐. | . | |
| Kornitzer 2004 | Kornitzer 2004 | 1‐1‐1‐0 | 1 | 1‐1‐1 | 7 | 1‐1‐1‐1 | 1 | 1‐1‐1 | 8 |
| Kristal 2014 | Kristal 2014 | 1‐1‐1‐1 | 1 | 1‐1‐1 | 8 | 1‐1‐1‐1 | 1 | 1‐1‐1 | 8 |
| Kromhout 1987 | Kromhout 1987 | 1‐1‐1‐0 | 2 | 1‐1‐1 | 8 | .‐.‐.‐. | . | .‐.‐. | . |
| Li 2004a | Li 2004a | 0‐1‐1‐1 | 2 | 0‐1‐1 | 7 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 |
| Ma 2017 | Ma 2017 | 1‐1‐1‐1 | 2 | 1‐1‐0 | 8 | .‐.‐.‐. | . | .‐.‐. | . |
| McNaughton 2005 | McNaughton 2005 | 1‐1‐1‐1 | 1 | 1‐1‐0 | 7 | 1‐1‐1‐1 | 1 | 1‐1‐1 | 8 |
| Heinen 2007 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 | .‐.‐.‐. | . | .‐.‐. | . | |
| van der Pols 2009 | 1‐1‐1‐1 | 2 | 1‐1‐0 | 8 | .‐.‐.‐. | . | .‐.‐. | . | |
| Menkes 1986 | Menkes 1986 | 0‐1‐1‐1 | 2 | 1‐1‐0 | 7 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 |
| Batieha 1993 | 0‐1‐1‐1 | 2 | 1‐1‐0 | 7 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 | |
| Breslow 1995 | 0‐1‐1‐1 | 2 | 1‐1‐0 | 7 | 1‐0‐1‐1 | 2 | 1‐1‐1 | 8 | |
| Burney 1989 | 0‐1‐1‐1 | 2 | 1‐1‐0 | 7 | 0‐1‐1‐1 | 2 | 1‐1‐1 | 8 | |
| Helzlsouer 1996 | 0‐1‐1‐1 | 2 | 1‐1‐0 | 7 | 0‐1‐1‐1 | 2 | 1‐1‐1 | 8 | |
| Helzlsouer 1989 | 0‐1‐1‐1 | 2 | 1‐1‐0 | 7 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 | |
| Ko 1994 | 0‐1‐1‐0 | 2 | 1‐1‐0 | 6 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 | |
| Menkes 1986 | .‐.‐.‐. | . | .‐.‐. | . | .‐.‐.‐. | . | .‐.‐. | . | |
| Schober 1987 | 0‐1‐1‐1 | 1 | 1‐1‐0 | 6 | 0‐1‐1‐1 | 1 | 1‐1‐1 | 7 | |
| Schober 1986 | .‐.‐.‐. | . | .‐.‐. | . | .‐.‐.‐. | . | .‐.‐. | . | |
| Zheng 1993 | 0‐1‐1‐1 | 2 | 1‐1‐0 | 7 | 0‐1‐1‐1 | 2 | 1‐1‐1 | 8 | |
| Michaud 2002 | Michaud 2002 | 1‐1‐1‐1 | 2 | 1‐1‐0 | 8 | 0‐1‐1‐1 | 2 | 1‐1‐1 | 8 |
| Michaud 2005 | Michaud 2005 | 0‐1‐1‐1 | 2 | 0‐1‐0 | 6 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 |
| Muka 2017 | Muka 2017 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 | .‐.‐.‐. | . | .‐.‐. | . |
| Nomura 1987 | Nomura 1987 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 |
| Nomura 2000 | Nomura 2000 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 |
| O'Grady 2014 | O'Grady 2014 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 | .‐.‐.‐. | . | .‐.‐. | . |
| Outzen 2016 | Outzen 2016 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 | 1‐0‐1‐1 | 2 | 1‐1‐1 | 8 |
| Overvad 1991 | Overvad 1991 | 1‐1‐1‐0 | 1 | 1‐1‐0 | 6 | .‐.‐.‐. | . | .‐.‐. | . |
| Pantavos 2015 | Pantavos 2015 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 | .‐.‐.‐. | . | .‐.‐. | . |
| Park 2015 | Park 2015 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 | .‐.‐.‐. | . | .‐.‐. | . |
| Peleg 1985 | Peleg 1985 | 1‐1‐1‐1 | 1 | 1‐1‐0 | 7 | 1‐1‐1‐1 | 1 | 1‐1‐1 | 8 |
| Peters 2007 | Peters 2007 | 0‐1‐1‐1 | 2 | 1‐1‐0 | 7 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 |
| Peters 2008 | Peters 2008 | 0‐1‐1‐1 | 1 | 1‐1‐1 | 7 | .‐.‐.‐. | . | .‐.‐. | . |
| Asgari 2009 | 0‐1‐1‐1 | 1 | 1‐1‐0 | 6 | .‐.‐.‐. | . | .‐.‐. | . | |
| Hotaling 2011 | 0‐1‐0‐1 | 0 | 1‐1‐1 | 5 | .‐.‐.‐. | . | .‐.‐. | . | |
| Walter 2011 | 0‐1‐0‐1 | 2 | 1‐1‐1 | 7 | .‐.‐.‐. | . | .‐.‐. | . | |
| Ratnasinghe 2000 | Ratnasinghe 2000 | 1‐1‐1‐1 | 2 | 1‐0‐0 | 7 | 0‐0‐1‐1 | 2 | 1‐1‐1 | 7 |
| Ringstad 1988 | Ringstad 1988 | 1‐1‐1‐1 | 2 | 1‐1‐0 | 8 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 |
| Thomson 2008 | Thomson 2008 | 0‐1‐1‐1 | 2 | 0‐1‐0 | 6 | .‐.‐.‐. | . | .‐.‐. | . |
| Sakoda 2005 | Sakoda 2005 | 0‐1‐1‐0 | 1 | 1‐1‐0 | 5 | 1‐1‐1‐1 | 1 | 1‐1‐1 | 8 |
| Salonen 1984 | Salonen 1984 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 | 0‐1‐1‐1 | 2 | 1‐1‐1 | 8 |
| Salonen 1985 | Salonen 1985 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 |
| Steinbrecher 2010 | Steinbrecher 2010 | 1‐1‐1‐1 | 2 | 0‐1‐0 | 7 | 1‐1‐1‐1 | 2 | 0‐1‐1 | 8 |
| Suadicani 2012 | Suadicani 2012 | 0‐1‐1‐1 | 2 | 1‐1‐1 | 8 | .‐.‐.‐. | . | .‐.‐. | . |
| Sun 2016 | Sun 2016 | 1‐1‐1‐1 | 2 | 1‐1‐0 | 8 | .‐.‐.‐. | . | .‐.‐. | . |
| van den Brandt 1993 | van den Brandt 1993 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 | .‐.‐.‐. | . | .‐.‐. | . |
| van den Brandt 1994 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 | .‐.‐.‐. | . | .‐.‐. | . | |
| van den Brandt 1993 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 | .‐.‐.‐. | . | .‐.‐. | . | |
| van den Brandt 2003 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 | .‐.‐.‐. | . | .‐.‐. | . | |
| Zeegers 2002 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 | .‐.‐.‐. | . | .‐.‐. | . | |
| Steevens 2010 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 | 0‐1‐1‐1 | 2 | 1‐0 | 6 | |
| van Noord 1987 | van Noord 1987 | 1‐1‐1‐0 | 1 | 1‐0‐1 | 6 | 1‐1‐1‐0 | 1 | 1‐1‐1 | 7 |
| Virtamo 1987 | Virtamo 1987 | 0‐1‐1‐1 | 2 | 1‐1‐1 | 8 | .‐.‐.‐. | . | .‐.‐. | . |
| Wei 2004 | Wei 2004 | 1‐1‐1‐1 | 1 | 1‐1‐1 | 8 | .‐.‐.‐. | . | .‐.‐. | . |
| Mark 2000 | 1‐1‐1‐1 | 1 | 1‐1‐1 | 8 | .‐.‐.‐. | . | .‐.‐. | . | |
| Willett 1983 | Willett 1983 | 1‐1‐1‐0 | 2 | 1‐1‐0 | 7 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 |
| Yoshizawa 1998 | Yoshizawa 1998 | 0‐1‐1‐1 | 2 | 1‐1‐1 | 8 | 1‐0‐1‐1 | 2 | 1‐1‐1 | 8 |
| Yu 1999 | Yu 1999 | 0‐1‐1‐1 | 2 | 1‐1‐0 | 7 | 1‐1‐1‐1 | 2 | 1‐1‐1 | 9 |
All but one cohort study received five to nine stars on the NOS. The exception (two stars) was an early investigation that was available only in abstract form for assessment (Clark 1985). In the NOS cohort assessment, we considered representativeness of the cohort for the target population to be adequate in 59% of studies, which received a star; 79% of studies provided evidence that cancer was not present at study commencement; we considered completeness of follow‐up (≥ 95%) data to be adequate in 93% of studies. The representativeness of the cohort for the target population is a matter of external validity and generalisability of study results, but a systematic deviation of participants from the target population might also introduce bias into study results. The target population of included studies varied with study objectives and could have been the general population, as well as special occupational groups. We did not assign a star for this question to studies that did not identify their target population or to studies that recruited volunteers. Differential selection of study participants (e.g. volunteers) from the target population can lead to confounding by factors associated with selenium status and cancer incidence (e.g. nutritional behaviour, socioeconomic position). All included studies chose comparison groups (cases/controls or exposed/non‐exposed) from the same study population. This approach enhanced comparability between groups.
We considered follow‐up data as complete or as missing data unlikely to introduce bias to study results in 47% of included observational studies. In the other cohorts, losses to follow‐up were greater than 5% and trial authors did not provide a description of losses to follow‐up. A high attrition rate may alter the characteristics of the population under investigation and may impede the generalisability of study results to the intended target population (external validity). The presence of attrition does not necessarily mean that study results are biased. However, given the possibility that selenium status may be linked to sociodemographic variables and socioeconomic position, which may also influence participation in follow‐up procedures, a differential effect of attrition may introduce bias towards underestimation or overestimation of the true exposure effect.
Forty‐five included observational studies were nested case‐control studies; therefore we assessed them by using the NOS case‐control form. The number of stars in the NOS assessment of case‐control studies ranged from five to nine, with 87% of studies receiving eight or nine stars. Although we generally assessed included prospective case‐control studies as having low risk of bias, we had concerns regarding case definition and the question of the representativeness of cases in some studies.
We considered the definition of cases as inadequate in 24% of nested case‐control studies, as cases were identified by self‐reporting; investigators did not describe linkage to databases with unclear validity or procedures. The magnitude and direction of bias that might have been introduced to the study results remain unclear.
In 16% of studies, investigators did not include all identified cases (or an appropriate sample of them) in the trial analyses, or they did not report selection procedures for analysed cases. Some studies lost blood specimens as the result of technical problems (e.g. cooler breakdown at one study centre); other studies reported that material available for analysis was insufficient; and others selected cases for analysis in a non‐random manner. This might bias the estimates of association in either direction.
We noted no obvious asymmetry (as an indicator of publication bias) in the funnel plots of studies on total cancer risk (Figure 3) and selected cancer types (Figure 4; Figure 5; Figure 6).
3.
Funnel plot of comparison: 1 Highest versus lowest selenium exposure, outcome: 2.1 Total cancer incidence and mortality.
4.
Funnel plot of comparison: 1 Observational studies: highest versus lowest selenium exposure, outcome: 2.8 Colorectal cancer risk.
5.
Funnel plot of comparison: 1 Observational studies: highest versus lowest selenium exposure, outcome: 2.12 Lung cancer risk incidence and mortality
6.
Funnel plot of comparison: 1 Highest versus lowest selenium exposure, outcome: 2.19 Prostate cancer risk.
Ethical criteria
All trials fulfilled informed consent and ethics board approval criteria (Algotar 2013; Dreno 2007; Karp 2013; Marshall 2011; NPCT 2002; Reid 2008; SELECT 2009), except for Li 2000, Yu 1991, Yu 1997, and Lubinski 2011, which did not mention these criteria.
Effects of interventions
1. Randomised controlled trials
We reported results from Duffield‐Lillico 2002 for all outcomes evaluated in the NPCT study (NPCT 2002) (prostate cancer, lung cancer, bladder cancer, colorectal and breast cancer, any cancer, and death from cancer), except for prostate cancer, for which we also used Duffield‐Lillico 2003a, in: NPCT 2002, and for the primary outcome, non‐melanoma skin cancer, whose results were reported in Duffield‐Lillico 2003b, in: NPCT 2002. For the SELECT study (SELECT 2009), we included only results from Lippman 2009, in: SELECT 2009, which reported on the blinded period of follow‐up with continuing selenium supplementation ‐ not from Klein 2011, in: SELECT 2009, which reported a longer period of follow‐up, including a subsequent period without selenium supplementation, and was discontinued in 2008 in compliance with the recommendation of the trial's independent DSMC (Lippman 2009 and Klein 2011, in: SELECT 2009). This second report by Klein et al included an additional period of 32 months (23% person‐time increase), along with the first follow‐up period, and results were essentially similar to those of Lippman et al 2009. For bladder cancer risk in SELECT, we used data from Lotan 2012, in: SELECT 2009, which encompassed the same extended period of follow‐up as Klein 2011, in: SELECT 2009, but was the only report available from the SELECT trial on this cancer type. For prostate cancer in SELECT, we also evaluated three reports published in 2014 that addressed specific population subgroups and cancer subtypes (Albanes 2014; Kristal 2014; Martinez 2014). For the ECOG trial, we used the 2013 report for all cancer types (Karp 2013).
1.1. Preventive efficacy outcomes
1.1.1. Any cancer incidence and mortality
Five studies evaluated the outcome of any cancer incidence (Algotar 2013; Karp 2013; Lubinski 2011; NPCT 2002; SELECT 2009); we assessed three of these trials as having low risk of bias (Algotar 2013; Karp 2013; SELECT 2009). Risk ratios (RRs) were based on detection of 1043 cases among 10,026 participants receiving supplemental selenium and 942 cases among 9449 participants allocated to placebo. We found no evidence of reduced incident cancer risk in studies at low risk of bias (RR 1.01, 95% confidence interval (CI) 0.93 to 1.10), nor in the analysis including all studies (RR 0.99, 95% CI 0.86 to 1.14) (Analysis 1.1).
1.1. Analysis.
Comparison 1 Randomised controlled trials: highest versus lowest selenium exposure, Outcome 1 Any cancer risk.
When we evaluated mortality from all cancers as an outcome, we could include only two studies in the analysis (NPCT 2002; SELECT 2009), one of which was at low risk of bias (SELECT 2009). When we considered only this latter trial, no difference in mortality rates between selenium and placebo arms emerged (RR 1.02, 95% CI 0.80 to 1.30). However, when we considered all studies, risk in the selenium group was lower than risk in the placebo group (RR 0.81, 95% CI 0.49 to 1.32) (Analysis 1.2).
1.2. Analysis.
Comparison 1 Randomised controlled trials: highest versus lowest selenium exposure, Outcome 2 Cancer mortality.
1.1.2. Head and neck cancer
Two trials investigated effects of selenium supplementation on risk of head and neck cancer (Karp 2013; NPCT 2002), but only one was at low risk of bias (Karp 2013). In analysis restricted to the study having low risk of bias, no relation emerged for the risk of this cancer type, with a summary RR of 1.00 (95% CI 0.18 to 5.45), and analysis pooling both studies yielded statistically unstable risk estimates (RR 1.22, 95% CI 0.52 to 2.85), based on 13 cases in the selenium arms and 9 cases in the placebo arms (Analysis 1.3).
1.3. Analysis.
Comparison 1 Randomised controlled trials: highest versus lowest selenium exposure, Outcome 3 Head and neck cancer risk.
1.1.3. Esophageal cancer
Two RCTs investigated the risk of oesophageal cancer associated with selenium supplementation (Karp 2013; NPCT 2002), but only one was at low risk of bias (Karp 2013). The number of cases in these studies was very low (3 in the selenium arms and 5 in the placebo arms), thus yielding very imprecise RR estimates. The summary RR for oesophageal cancer was 1.50 (95% 0.06 to 36.86) in the only study with low risk of bias, and 0.53 (95% CI 0.12 to 2.28) in overall studies (Analysis 1.4).
1.4. Analysis.
Comparison 1 Randomised controlled trials: highest versus lowest selenium exposure, Outcome 4 Oesophageal cancer risk.
1.1.4. Colorectal cancer
Three randomised controlled trials investigated the risk of colorectal cancer following selenium supplementation. These studies reported 76 cases in the selenium arms and 83 cases in the placebo arms (Karp 2013; NPCT 2002; SELECT 2009); two were at low risk of bias (Karp 2013; SELECT 2009). The summary RR of colorectal cancer was 0.99 (95% CI 0.69 to 1.43) in the two studies with low risk of bias, and 0.74 (95% CI 0.41 to 1.33) in all studies (Analysis 1.5).
1.5. Analysis.
Comparison 1 Randomised controlled trials: highest versus lowest selenium exposure, Outcome 5 Colorectal cancer risk.
1.1.5. Liver cancer
Four RCTs investigated the efficacy of selenium supplementation for liver cancer prevention, three of which were conducted in China with participants of different high‐risk groups in Qidong province, and one in the United States among individuals with resected non‐small‐cell lung cancer (Karp 2013; Li 2000; Yu 1991; Yu 1997). Yu 1991 reported on a trial with 2474 male and female first‐degree relatives of patients with liver cancer. During the study period of two years, investigators observed 10 participants in the selenium group, who received 200 µg selenium yeast/d, and 13 cases in the placebo group (RR 0.55, 95% CI 0.24 to 1.25). Yu 1997 investigated a four‐year supplementation period with 200 µg selenium yeast/d in 226 male and female hepatitis B‐surface antigen (HBs‐Ag) carriers. Investigators detected 11 cases (person‐time incidence rate: 1573.03/100,000) in the placebo group and four cases in the selenium group (RR 0.36, 95% CI 0.12 to 1.11) during the eight‐year follow‐up period. The mean blood selenium level during the intervention period was 152 µg/L in the intervention group and 107 µg/L in the control group. Li 2000 randomly assigned 2065 male HBs‐Ag carriers to receive 0.5 mg sodium selenite or placebo daily for three years. Thirty‐four cases of liver cancer occurred among 1112 participants receiving selenium, and 57 cases occurred among 953 placebo participants (RR 0.51, 95% CI 0.34 to 0.77).
Karp 2013 allocated 521 individuals with history of resected non‐small‐cell lung cancer to 200 µg/d selenium as selenised yeast or to placebo. During follow‐up, investigators diagnosed six new cases of liver cancer (actually coded as occurring to the ‘liver, gallbladder and bile duct’)‐ all in the selenium arm. We deemed this study to have low risk of bias.
The three Chinese studies had unclear risk of bias owing to lack of clear reporting of generation of allocation sequence or allocation concealment, and/or completeness of outcome data. Limiting analysis to the only study not downgraded owing to risk of bias yielded an RR of 6.52 (95% CI 0.37 to 115.49) (Analysis 1.6). The overall RR of the four studies was 0.52 (95% CI 0.35 to 0.79).
1.6. Analysis.
Comparison 1 Randomised controlled trials: highest versus lowest selenium exposure, Outcome 6 Liver cancer risk.
1.1.6. Melanoma
Three RCTs investigated the risk of melanoma following selenium supplementation (Algotar 2013; Karp 2013; NPCT 2002), but we judged only two of them to have low risk of bias (Algotar 2013; Karp 2013). For eight cases in the selenium arms and four cases in the placebo arms, the summary RR estimate was 1.35 (95% CI 0.41 to 4.52) in RCTs at low risk of bias. The RR estimate was slightly lower when all studies were considered (RR 1.28, 95% CI 0.63 to 2.59) (Analysis 1.7).
1.7. Analysis.
Comparison 1 Randomised controlled trials: highest versus lowest selenium exposure, Outcome 7 Melanoma risk.
1.1.7. Non‐melanoma skin cancer
1.1.7.1. Total non‐melanoma skin cancer
Risk of non‐melanoma skin cancer was the primary outcome of the NPCT,which reported higher risk in the selenium‐supplemented group than in the placebo group (unadjusted RR 1.27, 95% CI 1.11 to 1.45) (Duffield‐Lillico 2003a, in: NPCT 2002). This increase was confirmed by multi‐variable analysis after adjustment for confounders (hazard ratio (HR) 1.17, 95% CI 1.02 to 1.34) and was concentrated among participants in the highest two tertiles of baseline plasma selenium (≥ 105.6 µg/L), although increased risk for total non‐melanoma skin cancer was seen in all tertiles of baseline plasma selenium levels (Reid 2008). No variation in this effect appeared to be induced by age, sex, or smoking habits, and eliminating cases that occurred during the first period of selenium supplementation (one to two years) induced a slight decline in RRs. The mean selenium plasma concentration for participants was 114 µg/L at the time of randomisation. In the arm of the NPCT that was carried out in a single location ‐ Macon, Georgia, USA ‐ and included both 200 and 400 µg/d selenium supplementation (Reid 2008), non‐melanoma skin cancer risk increased in the 200‐µg/d arm after adjustment for age, sex, and smoking (unadjusted RR 1.49, 95% CI 1.10 to 2.03; adjusted HR 1.50, 95% CI 1.13 to 2.04) but not in the 400‐µg/d arm (unadjusted RR 0.88, 95% CI 0.66 to 1.16; adjusted HR 0.91, 95% CI 0.69 to 1.20). At the remaining sites, where only 200 µg/d of supplemental selenium was given, the RR was 1.24 (95% CI 1.07 to 1.45) and the HR was 1.18 (95% CI 1.02 to 1.37). Distribution of baseline plasma selenium levels was similar in this substudy to that in the NPCT main study, and no evidence of effect modification according to baseline selenium exposure emerged. Overall, the NPCT did not support preventive efficacy of selenium yeast supplementation against non‐melanoma skin cancer in these populations; on the contrary, investigators reported a cancer‐promoting effect of selenium for this cancer type, which was the primary trial endpoint, raising concern about potentially harmful effects of such selenium supplementation (NPCT 2002).
SELECT, which is the largest selenium supplementation trial conducted to date (Lippman 2009 and Klein 2011, in: SELECT 2009), thus far has not investigated the incidence of non‐melanoma skin cancer. A small trial in a French population of 184 organ graft recipients who were considered to be at high risk of premalignant and malignant epithelial lesions (Dreno 2007) investigated non‐melanoma skin cancer. This trial detected a higher incidence of skin cancer among 91 selenium‐supplemented participants (six cases; 6.6%) compared with 93 placebo‐supplemented participants (two cases; 2.2%; P = 0.15) during a five‐year follow‐up, which in its first three years comprised daily supplementation with selenised yeast containing 200 µg selenium.
A small trial among participants at high risk for prostate cancer also investigated the effects of using selenium supplements of 200 and 400 µg/d on risk of non‐melanoma skin cancer, with a median follow‐up of three years (Algotar 2013). Results for non‐melanoma skin cancer from this study showed the occurrence of three cases among 232 placebo‐treated participants and 11 cases among 467 selenium‐supplemented participants (eight cases among 234 individuals receiving 200 µg/d of selenium, and three cases among 233 individuals receiving 400 µg/d), with increased risk after overall selenium supplementation (incidence rate ratio from our calculation 1.8, 95% CI 0.5 to 10.2) but no evidence of a dose‐response relation.
The ECOG trial investigated non‐melanoma skin cancer and found 19 cases during follow‐up of 521 placebo‐treated participants and 11 cases among 1040 selenium‐allocated participants (Karp 2013). The RR of non‐melanoma skin cancer in this study was computed as 0.66(95% CI 0.37 to 1.19).
Overall, the summary RR for non‐melanoma skin cancer in selenium‐supplemented participants could be computed by pooling RRs from the above trials, rather than by using numbers of participants and cases, because the number of skin cancer cases diagnosed in the NPCT was not reported in the relevant publication (Duffield‐Lillico 2003a, in: NPCT 2002). The estimated RR limited to the only two trials with low risk of bias indicated a statistically unstable increased risk of non‐melanoma skin cancer associated with selenium supplementation of 200 µg/d (RR 1.16, 95% CI 0.30 to 4.42), with similar risk results when analysis was performed on the four trials overall (RR 1.23, 95% CI 0.73 to 2.08) (Analysis 1.8) (Algotar 2013; Karp 2013).
1.8. Analysis.
Comparison 1 Randomised controlled trials: highest versus lowest selenium exposure, Outcome 8 Non‐melanoma skin cancer risk.
1.1.7.2. Basal cell carcinoma (BCC)
Algotar 2013 found in the 200‐ and 400‐µg/d selenium groups an RR of 0.86 (95% CI 0.42 to 1.77) and 0.80 (95% CI 0.38 to 1.66), respectively; and an RR in both treatment groups combined of 0.83 (95% CI 0.45 to 1.54). ECOG 5597 found an RR of 0.54 (95% CI 0.26 to 1.14) (Karp 2013).
At the end of the blinded treatment period in NPCT 2002, the unadjusted RR for basal cell carcinoma in the 200‐µg/d selenium group was 1.17 (95% CI 1.02 to 1.35), and the adjusted HR was 1.09 (95% CI 0.94 to 1.26). Eliminating cases that occurred within the first two years of supplementation had no effect on the RR. Reid 2008 found a crude RR of 0.90 (95% CI 0.65 to 1.24) and an adjusted HR of 0.95 (95% CI 0.69 to 1.29) for this cancer type in the 400‐µg/d selenium substudy. In a small trial with no RR estimates (Dreno 2007), three cases of BCC occurred among 91 selenium‐supplemented participants, along with one case among 93 placebo‐receiving participants.
1.1.7.3. Squamous cell carcinoma (SCC)
Algotar 2013 found an RR of 0.58 (95% CI 0.27 to 1.25) and 0.12 (95% CI 0.03 to 0.50) in the 200‐ and 400‐µg/d trial populations, respectively, and for all participants, the RR was 0.35 (95% CI 0.17 to 0.72). ECOG 5597 found an RR of 0.92 (95% CI 0.34 to 2.47) (Karp 2013).
In NPCT 2002, selenium supplementation increased the risk of SCC (unadjusted RR 1.32, 95% CI 1.09 to 1.60; adjusted HR 1.25, 95% CI 1.03 to 1.51). Adverse effects of selenium supplementation on SCC risk appeared to increase with increasing plasma selenium levels at baseline, in that higher risk was seen only in participants at the highest two tertiles of baseline levels (≥ 105.6 µg/L), suggesting an interaction between supplementation and baseline exposure. In the 400‐µg/d selenium substudy (Reid 2008), investigators reported no change in SCC risk by selenium supplementation (crude RR 1.20, 95% CI 0.85 to 1.68; adjusted HR 1.05, 95% CI 0.71 to 1.56). Dreno 2007, the smaller trial, reported that two among 91 selenium‐supplemented individuals were given a diagnosis of SCC, whereas no cases were observed among placebo participants.
1.1.8. Lung cancer
Three RCTs have investigated lung cancer risk associated with selenium administration (Karp 2013; NPCT 2002; SELECT 2009), with two assessed as having low risk of bias (Karp 2013; SELECT 2009). Summary RR estimates were 1.16 (95% CI 0.89 to 1.50) when we limited the analysis to studies at low risk of bias, and 1.03 (95% CI 0.78 to 1.37) when we included all studies (Analysis 1.9).
1.9. Analysis.
Comparison 1 Randomised controlled trials: highest versus lowest selenium exposure, Outcome 9 Lung cancer risk.
1.1.9. Female breast cancer
Three studies evaluated breast cancer risk associated with selenium supplementation (Karp 2013; Lubinski 2011; NPCT 2002), one of which we judged as having low risk of bias (Karp 2013). The RR from the study with low risk of bias was 2.04 (95% CI 0.44 to 9.55), with statistical imprecision due to the small number of cases (eight in the selenium arm, two in the placebo arm). The pooled RR from all studies was 1.44 (95% CI 0.96 to 2.17) (Analysis 1.10).
1.10. Analysis.
Comparison 1 Randomised controlled trials: highest versus lowest selenium exposure, Outcome 10 Breast cancer risk.
1.1.10. Bladder cancer
Three studies evaluated bladder cancer outcomes (Karp 2013; NPCT 2002; SELECT 2009), two of which we judged as having low risk of bias (Karp 2013; SELECT 2009), The summary RR from the only studies at low risk of bias was 1.07 (95% CI 0.76 to 1.52). The corresponding RR for all studies, encompassing a total of 146 cases ‐ 79 in the selenium arms and 67 in the placebo arms ‐ was 1.10 (95% CI 0.79 to 1.52) (Analysis 1.11).
1.11. Analysis.
Comparison 1 Randomised controlled trials: highest versus lowest selenium exposure, Outcome 11 Bladder cancer risk.
1.1.11. Prostate cancer
Five trials evaluated prostate cancer (Algotar 2013; Karp 2013; Marshall 2011; NPCT 2002; SELECT 2009), all of which we judged as having low risk of bias, except for NPCT 2002. Meta‐analysis for prostate cancer‐based trials at low risk of bias yielded an RR of 1.01 (95% CI 0.90 to 1.14) for the 9630 participants supplemented with selenium (520 cases) compared with the 9312 participants allocated to placebo (500 cases), indicating no effect of intervention (supplementation of organic selenium at 200 µg/d) on prostate cancer risk, with very consistent results and no heterogeneity across these studies (I² = 0.0%). The overall RR was 0.91 (95% CI 0.75 to 1.12) when all studies were considered; moderate heterogeneity (I² = 36%) emerged owing to the addition of the NPCT (Analysis 1.12) (NPCT 2002).
1.12. Analysis.
Comparison 1 Randomised controlled trials: highest versus lowest selenium exposure, Outcome 12 Prostate cancer risk.
The trial that first investigated the relation between selenium exposure and prostate cancer risk (Duffield‐Lillico 2002 and Duffield‐Lillico 2003b, in: NPCT 2002) reported a reduction in prostate cancer incidence in the selenium‐treated group, which was particularly strong during the first period of follow‐up (1983 to 1993; adjusted HR 0.35, 95% CI 0.16 to 0.65) and was slightly higher but still much lower than unity during the entire period of follow‐up (1983 to 1996; HR 0.48, 95% CI 0.28 to 0.80). Analyses stratified by baseline plasma selenium category showed greatly reduced risk associated with active treatment among participants with baseline plasma selenium ≤ 106.4 µg/L (HR 0.14, 95% CI 0.03 to 0.61) in the intermediate category (106.8 to 123.2 µg/L; HR 0.33, 95% CI 0.13 to 0.82), while in the upper category (> 123.2 µg/L), the HR was 1.14 (95% CI 0.51 to 2.59). Selenium supplementation in participants with baseline PSA ≤ 4 ng/mL was associated with considerably reduced risk (HR 0.33, 95% CI 0.14 to 0.79) compared with risk in individuals with PSA > 4 ng/mL (HR 0.95, 95% CI 0.42 to 2.14). However, interpretation of these NPCT findings is complicated by a potentially serious source of bias. As reported in 2003 by the study authors, a considerably higher percentage of participants with elevated PSA levels in the placebo group underwent prostatic biopsy as compared with participants in the selenium group (35% vs 14%; P < 0.05; Duffield‐Lillico 2003b, in: NPCT 2002). Differences in biopsy rates were greatest among participants with the lowest baseline selenium concentrations ‐ the subgroup that appeared to derive the greatest beneficial effects of selenium administration. This may have contributed to substantial overestimation of the effects of selenium supplementation in the NPCT.
The SELECT trial found no evidence of benefit derived from selenium supplementation (compared with placebo) over a median of 5.5 years in terms of prostate cancer incidence (HR 1.03, 95% CI 0.90 to 1.18, 99% CI 0.87 to 1.24) (SELECT 2009). The adjusted HR for prostate cancer in the selenium plus vitamin E group compared with the placebo group was 1.05 (95% CI 0.91 to 1.20, 99% CI 0.88 to 1.25). The original report of the trial provided no specific RR estimate according to disease severity, but during an extended follow‐up of this cohort after selenium supplementation had ceased (Klein 2011, in: SELECT 2009), investigators found increased risk of Gleason 7 or greater disease (HR 1.21, 99% CI 0.90 to 1.63). It is interesting to note that the SELECT trial included only participants with PSA ≤ 4 ng/mL ‐ the group in the NPCT that showed greatest apparent benefit. During this further follow‐up of the SELECT cohort, risk of prostate cancer in the selenium arm also slightly increased compared with that described in the first report, which had included only the active supplementation period (Lippman 2009, in: SELECT 2009). In this longer follow‐up based on 575 prostate cancer cases in the selenium arm and 529 in the placebo arm, the RR of prostate cancer was 1.09 (99% CI 0.93 to 1.27).
Three further reports from SELECT on the relation between selenium administration and prostate cancer risk have been published (Albanes 2014; Kristal 2014; Martinez 2014); where investigators looked at more specific associations than were addressed in the two main publications from this trial (Lippman 2009 and Klein 2011, in: SELECT 2009). Kristal 2014 performed a case‐cohort study within the SELECT study by including 1739 total prostate cancer cases (of which 489 showed high‐grade (Gleason 7 to 10) disease) and 3117 randomly selected men composing the control subcohort (Kristal 2014). Administration of selenium (both selenium only and selenium combined with vitamin E) had no effect on prostate cancer risk among men with low baseline selenium status (< 60th percentile of toenail selenium), but among participants in the two upper quintiles of baseline selenium exposure, risk of prostate cancer was increased (HR 1.20, 95% CI 0.85 to 1.81), particularly high‐grade prostate cancer (HR 1.62, 95% CI 0.95 to 2.77). HRs were even higher when any selenium supplementation (alone or with vitamin E) was considered because such supplementation increased the risk of any prostate cancer (RR 1.27, 95% CI 0.92 to 1.74) and high‐grade disease (RR 1.91, 95% CI 1.20 to 3.05).
Martinez 2014 investigated the effect of selenium supplementation on prostate cancer risk among participants in SELECT who had genotypes associated with altered mRNA expression of the androgen‐regulated prostate tumour suppressor protein NKX3.1. The design was still of the case‐cohort type, encompassing 1866 prostate cancer cases and 3135 non‐prostate cancer cases. Trial authors found that selenium administration combined with the CC genotype at rs11781886 increased overall prostate cancer risk (HR 1.68, 95% CI 1.01 to 2.78) and low‐grade prostate cancer risk (HR 1.81, 95% CI 1.02 to 3.23), but they noted no such interaction for the other genotypes.
Finally, in a SELECT subpopulation composed of 1746 prostate cancer cases and a subcohort of 3211 men, Albanes 2014 investigated a possible association between baseline plasma α‐tocopherol and γ‐tocopherol and active supplementation with selenium (and vitamin E as α‐tocopherol) in terms of prostate cancer risk. Trial authors found a strong excess of risk among participants in the highest baseline α‐tocopherol category (fifth quintile) receiving selenium supplementation (HR 2.04, 95% CI, 1.29 to 3.22, P trend 0.005), which was higher with high‐grade (Gleason grade 7 to 10) disease among men receiving selenium (HR 2.12, 95% CI, 1.32 to 3.40, P‐trend 0.0002). These findings suggest a possible biological interaction between α‐tocopherol status and selenium supplementation in increasing high‐grade prostate cancer risk.
In Marshall 2011, prostate cancer incidence was 35.6% versus 36.6% in selenium‐supplemented compared with placebo‐treated participants after three years of follow‐up, respectively. The overall RR was 0.91, with a 95% CI of 0.55 to 1.52 (courtesy of James Marshall, unpublished data). Analysis of RRs according to baseline plasma selenium levels showed no dose‐response effect, with point estimates of 0.82 (95% CI 0.40 to 1.69), 1.38 (95% CI 0.68 to 2.78), 0.98 (95% CI 0.58 to 1.68), and 0.91 (95% CI 0.45 to 1.84), when the quartile of selenium status was increased at baseline.
The NBT reported an HR of prostate cancer of 0.94 (95% CI 0.52 to 1.70) for participants receiving 200 µg/d and 0.90 (95% CI 0.48 to 1.66) for those receiving 400 µg/d, compared with placebo (Algotar 2013). Although average baseline selenium status, as assessed through plasma selenium, was higher than in the NPCT (median value 126.1 vs 115.0 µg/L), the lowest tertile of plasma selenium levels had a median value (101.1 µg/L) well below the apparent threshold of around 120 µg/L, at which a beneficial effect of selenium seemed to occur in the NPCT. Furthermore, as noted by study authors, 45% of participants enrolled in this study had baseline plasma selenium levels < 123 µg/L, which is the upper threshold for a protective effect of selenium supplementation according to results of the NPCT. Trial authors also stated: "None of the baseline variables modified the effect of selenium on the primary endpoint", and plasma selenium concentration at baseline was among these variables (Algotar 2013).
Karp 2013, the ECOG trial, carried out in subjects with resected non‐small‐cell lung cancer, reported nine and 16 cases of newly diagnosed prostate cancer among 250 and 509 male participants in the placebo and selenium groups, respectively. This allowed us to compute an RR of 0.87 (95% CI 0.39 to 1.45) for prostate cancer in the selenium‐supplemented arm.
Following the NPCT, none of the subsequent, high‐quality RCTs provided evidence suggesting that baseline selenium status could modify the effect of selenium supplementation on subsequent prostate cancer occurrence. In the NBT, the bottom category (tertile) of baseline plasma selenium levels in this trial population was 101.1 µg/L, i.e. lower than the upper bound of the bottom category (106.4 µg/L) and the middle category (106.8 to 123.2 µg/L) in the NPCT, both of which had shown a strongly decreased subsequent prostate cancer occurrence (Algotar 2013). In the SWOG S9917 study, results of selenium supplementation were also made available for four categories (quartiles) of baseline plasma selenium and showed no effect of treatment in any categories (Marshall 2011). These categories were < 106, 106–132, 132–162, and > 162 µg/L, and corresponding RRs of prostate cancer in the selenium‐supplemented group were 0.82 (95% CI 0.40 to 1.69), 1.38 (95% CI 0.68 to 2.78), 0.98 (95% CI 0.58 to 1.68), and 0.91 (95% CI 0.45 to 1.84), respectively, versus an overall study RR of 0.97 (95% CI 0.68 to 1.39). Therefore, also in this high‐quality trial, the bottom category of baseline selenium exposure was entirely similar to the corresponding one in the NPCT, but in contrast to NPCT, no effect of selenium supplementation emerged and no evidence showed risk of bias. Finally, a case‐cohort study carried out within SELECT and published in 2014 provided data showing the relation between baseline selenium exposure and effects of selenium supplementation (Kristal 2014). In that study, whose average selenium exposure was higher than that characterising the NPCT and the NBT, investigators reported no effect of selenium supplementation on both overall prostate cancer and low‐grade and high‐grade prostate cancer in the three quintiles of baseline toenail selenium levels, but enhanced risk of high‐grade prostate cancer emerged for the two upper quintiles (alone and combined). Quintile cutoff points for these categories of the trial population were 0.758, 0.832, 0.901, and 1.003 µg/g. Overall, these results clearly indicate that even in subgroups with the lowest baseline selenium status in these Western populations, selenium provided no protective effect for prevention of prostate cancer, although this is the cancer type that once was thought to be most strongly associated with a beneficial effect of selenium supplementation.
1.1.12. Haematological cancers
Two trials evaluated the risk of haematological malignancies associated with selenium administration (Karp 2013; NPCT 2002) using 23 cases only ‐ 14 in the selenium arms and 9 in the placebo arms ‐ but we judged only one trial to be at low risk of bias (Karp 2013). The summary RR was 1.00 (95% CI 0.25 to 3.99) in the study at low risk of bias and 1.21 (95% CI 0.52 to 2.80) when all studies were considered (Analysis 1.13).
1.13. Analysis.
Comparison 1 Randomised controlled trials: highest versus lowest selenium exposure, Outcome 13 Leukaemia and lymphoma risk.
1.2. Adverse effects outcomes
The RCTs on selenium have provided unexpected information about the incidence of adverse effects of selenium supplementation and have unexpectedly become a key source of data for risk assessment of the upper safe level of selenium exposure in humans (Vinceti 2017a; Vinceti 2017b). Thirty‐five participants withdrew from the NPCT because of adverse effects, mainly gastrointestinal upset. The RR for adverse events in the selenium group was 1.51 (95% CI 0.74 to 3.11) (our calculation, based on the number of randomly assigned participants). Reports of increased risk of glaucoma in Marshall 2011 and NPCT 2002 prompted additional studies on this issue (Bruhn 2009), and likely led to inclusion of cataract and glaucoma among the several potential adverse events monitored during subsequent trials in which investigators administered selenium (Algotar 2013).
In the NPCT, a secondary analysis of participants who did not have diabetes at the start of the study unexpectedly revealed an excess risk of type 2 diabetes mellitus in the selenium group (adjusted HR 1.55, 95% CI 1.03 to 2.33) (Stranges 2007). That study found increased risk of developing type 2 diabetes associated with selenium supplementation across all tertiles of baseline plasma selenium levels, although the excess was much greater for the upper category of > 121.6 µg/L (RR 2.70, 95% CI 1.30 to 5.61) than for the lower (RR 1.13, 95% CI 0.58 to 2.18) and intermediate (RR 1.36, 95% CI 0.60 to 3.09) categories. Increased risk of diabetes associated with selenium supplementation was independent of baseline age, sex, smoking status, and body mass index (BMI), with the exception of participants in the top tertile of BMI. SELECT reported a slight increase in the incidence of type 2 diabetes in the selenium‐alone group (RR 1.07, 99% CI 0.94 to 1.22). Any such excess risk decreased over time after selenium supplementation ceased, as is shown by results of the Klein study (Klein 2011, in: SELECT 2009). In this study, the RR of diabetes was 1.04 (99% CI 0.93 to 1.17), thus supporting a short‐term effect of selenium supplementation on diabetes risk.
Although the three trials on liver cancer and Reid 2008 did not mention the occurrence of adverse effects, and Dreno 2007 and Marshall 2011 (the SWOG 2011 trial) apparently performed no assessment of diabetes incidence, three recent phase 3 RCTs have investigated the occurrence of diabetes after selenium supplementation for prevention of malignant and non‐malignant cancer. In the NBT, during five years of follow‐up of 699 participants at high risk for prostate cancer supplemented with 200 or 400 µg/d of selenium or placebo, Algotar 2013 reported the occurrence of diabetes in 12, 12, and 7 participants, respectively. This allowed us to compute an incidence rate ratio of 1.70 (95% CI 0.62 to 5.10) and 1.71 (95% CI 0.62 to 5.12) among 200‐ and 400‐µg/d selenium‐supplemented participants, respectively, compared with those given placebo. The ECOG trial, which was carried out in 1561 participants with resected stage I non–small‐cell lung cancer, trial authors did not explicitly report the RR of diabetes during follow‐up (Karp 2013). However, occurrence during four years of follow‐up (2007 to 2011) was stated as 26 new diagnoses of diabetes in the selenium arm (1040 participants at baseline, of whom 865 underwent toxicity assessment) and 12 new diagnoses among placebo‐treated participants (521/477). On the basis of these numbers, we could compute an RR of 1.09 (95% CI 0.55 to 2.13) or, for participants with toxicity assessment, 1.19 (95% CI 0.61 to 2.35) ‐ values comparable with those observed in the other trials, except for NPCT. Most recently, in an intervention study investigating the effect of selenium supplementation for prevention of colorectal adenoma recurrence compared with placebo (the SELCEL trial), 31 cases of diabetes occurred in the selenium‐treated group and 25 in the placebo group during follow‐up, with an RR of 1.25 (95% CI 0.74 to 2.11) (Thompson 2016). Therefore, an excess incidence of type 2 diabetes systematically emerged in all trials that investigated this adverse effect (Vinceti 2017b).
The SELECT study also looked at other side effects known to be associated with selenium overexposure (Vinceti 2001), finding an association for some of them. Selenium treatment increased the occurrence of alopecia (RR 1.28, 95% CI 1.07 to 1.53, based on 265/206 cases in selenium and placebo arms), dermatitis (RR 1.16, 95% CI 1.03 to 1.29, 619/524), nail changes (RR 1.04, 95% CI 0.96 to 1.13, 1087/1035), and halitosis (RR 1.17, 95% CI 0.99 to 1.38, 503/427).
2. Observational studies
When risks of cancer for higher and lower levels of selenium exposure are compared, a summary risk estimate of one suggests no association between selenium exposure and cancer, and summary risk estimates below and above one suggest a beneficial or harmful effect of higher selenium exposure, respectively. We evaluated the statistical precision of the point estimates by assessing the width of their 95% or 99% confidence intervals.
2.1. Aetiological association: results from meta‐analyses
2.1.1. Any cancer
We meta‐analysed results of 16 prospective observational studies on total cancer risk, including data on more than 276,000 participants. The cohorts of Salonen 1984 and Salonen 1985 overlapped. Hence, we included only data from Salonen 1985 in the meta‐analysis. We had to omit Fex 1987, as the CI value was not reported and could not be calculated from available data.
For participants in the highest category of pre diagnostic selenium exposure, the summary risk estimate was odds ratio (OR) 0.72 (95% CI 0.55 to 0.93) for cancer incidence and OR 0.76 (95% CI 0.59 to 0.97) for cancer mortality for both sexes combined (Analysis 2.1), when compared with participants in the lowest exposure category. We observed moderate to substantial heterogeneity for both incidence (I² = 45%) and mortality (I² = 67%).
2.1. Analysis.
Comparison 2 Observational studies: highest versus lowest selenium exposure, Outcome 1 Total cancer incidence and mortality.
Analyses by sex revealed lower point estimates for men (incidence: OR 0.72, 95% CI 0.46 to 1.14; mortality: OR 0.65, 95% CI 0.45 to 0.94) (Analysis 2.2) than for women (incidence: OR 0.90, 95% CI 0.45 to 1.77; mortality: OR 0.91, 95% CI 0.80 to 1.03) (Analysis 2.3).
2.2. Analysis.
Comparison 2 Observational studies: highest versus lowest selenium exposure, Outcome 2 Total cancer incidence and mortality (men).
2.3. Analysis.
Comparison 2 Observational studies: highest versus lowest selenium exposure, Outcome 3 Total cancer incidence and mortality (women).
All studies but one (Sun 2016) used a circulating biomarker (serum and plasma selenium levels) for assessment of selenium status. Analysis 2.4 shows the results in ascending order of baseline exposure for those studies that reported category borders. The graph does not reveal any systematic pattern of changes in the relation between selenium status and cancer risk according to increasing baseline selenium levels. Analysis 2.5 shows the results in ascending order for differences in selenium levels.
2.4. Analysis.
Comparison 2 Observational studies: highest versus lowest selenium exposure, Outcome 4 Total cancer incidence and mortality (ascending order of selenium levels).
2.5. Analysis.
Comparison 2 Observational studies: highest versus lowest selenium exposure, Outcome 5 Total cancer incidence and mortality (ascending order of differences in selenium levels).
2.1.2. Stomach cancer
No additional cohort studies on stomach cancer and selenium exposure have been published since the last update of this review; therefore meta‐analysis for this cancer type was still based on five studies. The summary risk estimate for both sexes combined was OR 0.66 (95% CI 0.43 to 1.01) in the highest exposure category when compared with the lowest (I² = 51%) (Analysis 2.6). In this meta‐analysis, we included one cohort twice because trial authors reported results stratified according to cardia and non‐cardia gastric cancer (Mark 2000, in: Wei 2004).
2.6. Analysis.
Comparison 2 Observational studies: highest versus lowest selenium exposure, Outcome 6 Stomach cancer risk.
Use of available sex‐stratified results for meta‐analysis yielded a risk estimate for men of OR 0.43 (95% CI 0.14 to 1.32) (I² = 56%), and for women of OR 0.73 (95% CI 0.12 to 4.35) (I² = 62%) (Analysis 2.7).
2.7. Analysis.
Comparison 2 Observational studies: highest versus lowest selenium exposure, Outcome 7 Stomach cancer risk (by sex).
2.1.3. Colorectal/Colon cancer
Six observational studies reported data on the incidence of colorectal cancer. The summary risk estimate was OR 0.82 (95% CI 0.72 to 0.94) for both sexes combined (I² = 0.0%) (Analysis 2.8), with OR 0.86 (95% CI 0.65 to 1.16) for men and OR 0.96 (95% CI 0.61 to 1.50) for women (Analysis 2.9). Five studies reported data stratified or restricted to colon cancer. The summary estimate was OR 0.81 (95% CI 0.69 to 0.96) for both sexes combined (I² = 0.0%) (Analysis 2.10), with OR 0.84 (95% CI 0.56 to 1.25) for men and OR 0.68 (95% CI 0.44 to 1.04) for women (Analysis 2.11).
2.8. Analysis.
Comparison 2 Observational studies: highest versus lowest selenium exposure, Outcome 8 Colorectal cancer risk.
2.9. Analysis.
Comparison 2 Observational studies: highest versus lowest selenium exposure, Outcome 9 Colorectal cancer risk (by sex).
2.10. Analysis.
Comparison 2 Observational studies: highest versus lowest selenium exposure, Outcome 10 Colon cancer risk.
2.11. Analysis.
Comparison 2 Observational studies: highest versus lowest selenium exposure, Outcome 11 Colon cancer risk (by sex).
2.1.4. Lung cancer
We included 13 studies in this meta‐analysis. We did not meta‐analyse data from Menkes 1986 and Knekt 1990, as the study population of the former overlapped with that of Comstock 1997 (another meta‐analysed study) ‐ and results of the latter were presented in insufficient detail.
The summary risk estimate for lung cancer incidence for both sexes combined was 0.82 (95% CI 0.59 to 1.14) (Analysis 2.12). We noted substantial heterogeneity among study results (I² = 66%). We found little difference in summary estimates when results were disaggregated by sex (Analysis 2.13), by indicator of selenium exposure (intake, blood or toenail content) (Analysis 2.14), by baseline serum/plasma bottom exposure category (Analysis 2.15), and by ascending differences in selenium levels (Analysis 2.16). In the latter analyses, we noted no dose‐response relation between baseline selenium and risk.
2.12. Analysis.
Comparison 2 Observational studies: highest versus lowest selenium exposure, Outcome 12 Lung cancer incidence and mortality.
2.13. Analysis.
Comparison 2 Observational studies: highest versus lowest selenium exposure, Outcome 13 Lung cancer risk (sex‐disaggregated data).
2.14. Analysis.
Comparison 2 Observational studies: highest versus lowest selenium exposure, Outcome 14 Lung cancer risk (by exposure assessment).
2.15. Analysis.
Comparison 2 Observational studies: highest versus lowest selenium exposure, Outcome 15 Lung cancer risk (ascending order of selenium levels).
2.16. Analysis.
Comparison 2 Observational studies: highest versus lowest selenium exposure, Outcome 16 Lung cancer risk (ascending order of differences in selenium levels).
2.1.5. Female breast cancer
We included eight studies in this meta‐analysis. Data show little association between baseline selenium levels and breast cancer risk, with a slightly but imprecisely higher risk for higher exposure (OR 1.09, 95% CI 0.87 to 1.37) (Analysis 2.17). The heterogeneity of results was low (I² = 14%).
2.17. Analysis.
Comparison 2 Observational studies: highest versus lowest selenium exposure, Outcome 17 Breast cancer risk (women).
2.1.6. Bladder cancer
Meta‐analysis of bladder cancer incidence in five observational studies revealed an inverse association, with an overall risk estimate of 0.67 (95% CI 0.46 to 0.97) (Analysis 2.18) (heterogeneity: I² = 30%). Sex‐disaggregated data were available only from Michaud 2005 and showed an inverse association between selenium exposure and risk in women, but not in men. Two studies included only male participants (Michaud 2002; Nomura 1987); both found a reduced but imprecisely estimated bladder cancer risk for higher selenium exposure (Analysis 2.18). Heterogeneity was not reduced by sex stratification (I² = 40% in study results for men). No further studies had been published since the last update of this review (Vinceti 2014).
2.18. Analysis.
Comparison 2 Observational studies: highest versus lowest selenium exposure, Outcome 18 Bladder cancer risk.
2.1.7. Prostate cancer
We included 21 epidemiological studies on prostate cancer incidence in the meta‐analysis. The summary risk estimate for higher selenium exposure was OR 0.84 (95% CI 0.75 to 0.95) (heterogeneity: I² = 27%) (Analysis 2.19). Stratification of the analysis by method of selenium assessment revealed an inverse association between baseline selenium and risk when exposure was assessed through blood selenium levels (OR 0.86, 95% CI 0.75 to 0.99) or toenails (OR 0.60, 95% CI 0.44 to 0.82), but not when dietary assessment methods were used (OR 0.99, 95% CI 0.85 to 1.15) (Analysis 2.20). When we stratified analysis according to baseline (blood) selenium exposure or differences in selenium (blood) levels, no specific relation or pattern emerged between selenium and prostate cancer risk across the entire exposure spectrum (Analysis 2.21; Analysis 2.22).
2.19. Analysis.
Comparison 2 Observational studies: highest versus lowest selenium exposure, Outcome 19 Prostate cancer risk.
2.20. Analysis.
Comparison 2 Observational studies: highest versus lowest selenium exposure, Outcome 20 Prostate cancer risk (by exposure assessment).
2.21. Analysis.
Comparison 2 Observational studies: highest versus lowest selenium exposure, Outcome 21 Prostate cancer risk (ascending order of selenium levels).
2.22. Analysis.
Comparison 2 Observational studies: highest versus lowest selenium exposure, Outcome 22 Prostate cancer risk (ascending order of differences in selenium levels).
2.2. Aetiological association: other results
For all other types of cancer, data were available from fewer than five epidemiological studies; thus we did not meta‐analyse the results. We have reported in Table 5 results of observational studies not included in meta‐analyses. None of these study results support an association between selenium exposure and gynaecological cancer risk, and results for cancers of the gastrointestinal, respiratory, or urological tract are inconsistent. For respiratory and urological cancers, studies reported either no association or increased risk for participants with higher selenium exposure. For gastrointestinal cancers including cancer of the liver and other sites not mentioned above, studies found either no association or reduced risk with higher selenium exposure.
3. Results of observational studies not included in meta‐analysis.
| Organ system | Cancer | Case definition | Risk ratio estimate (highest vs lowest exposure category) | 95% CI | Selenium marker | Sex | Study |
| Gynaecological | Cervix | incidence | 0.89 | 0.40 to 2.00 | serum | women | Menkes 1986 (Batieha 1993) |
| 1.10 | n.r. | serum | Coates 1988 | ||||
| Gynaecological (without breast) | incidence | 0.96 | n.r. | serum | Knekt 1990 | ||
| Ovary | incidence | 0.87 | 0.25 to 5.26 | serum | Knekt 1990 (Knekt 1996) | ||
| 1.22 | 0.44 to 3.38 | toenail | Garland 1995 | ||||
| 0.58 | 0.2 to 1.7 | serum | Menkes 1986 (Helzlsour 1996) | ||||
| 1.00 | 0.73 to 1.37 | suppl. intake | Thomson 2008 | ||||
| Uterus | incidence | 1.38 | 0.62 to 3.08 | toenail | Garland 1995 | ||
| Gastrointestinal | Gastrointestinal tract (all) | incidence | 1.00 | n.r. | serum/plasma | both | Coates 1988 |
| Oesophageal squamous cell carcinoma | incidence | 0.37 | 0.16 to 0.86 | toenail | both | Steevens 2010 | |
| 0.67 | 0.53 to 1.30 | intake | both | Hashemian 2015 | |||
| Oesophageal adenocarcinoma | incidence | 0.76 | 0.41 to 1.40 | toenail | both | Steevens 2010 | |
| Oesophagus | incidence | 0.56 | 0.44 to 0.71 | serum | both | Wei 2004 (Mark 2000) | |
| mortality | 0.62 | 0.44 to 0.89 | serum | ||||
| mortality | 0.35 | 0.16 to 0.81 | serum | both | Wei 2004 (Wei 2004) | ||
| incidence | 0.27 | 0.03 to 2.21 | suppl. intake | both | Dong 2008 | ||
| Gastric cardio adenocarcinoma | incidence | 0.52 | 0.27 to 1.02 | toenail | both | Steevens 2010 | |
| Oesophagus and stomach | incidence | 0.45 | n.r. | serum | men | Knekt 1990 (Knekt 1988) | |
| incidence | 0.67 | n.r. | serum | women | |||
| Liver | incidence | 0.62 | 0.21 to 1.86 | plasma | men | Yu 1999 | |
| 0.41 | 0.23 to 0.72 | serum | both | Hughes 2016 | |||
| 0.86 | 0.52 to to 1.43 | intake | both | Ma 2017 | |||
| 0.95 | 0.51 to 1.76 | men | |||||
| 0.70 | 0.26 to 1.90 | women | |||||
| mortality | 0.50 | 0.28 to 0.90 | toenail | both | Sakoda 2005 | ||
| 0.57 | 0.31 to 1.05 | men | |||||
| 0.18 | 0.03 to 1.13 | women | |||||
| Pancreas | incidence | 0.08 | 0.01 to 0.56 | serum | men | Menkes 1986 (Burney 1989) | |
| 0.83 | 0.40 to 1.67 | women | |||||
| 0.58 | n.r. | serum | men | Knekt 1990 | |||
| 3.49 | n.r. | women | |||||
| 0.72 | 0.36 to 1.43 | intake | both | Banim 2013 | |||
| 0.69 | 0.39 to 1.20 | supplemental intake | both | Han 2013 | |||
| Rectum | incidence | 0.625 | n.r. | serum | men | Nomura 1987 | |
| 1.05 | 0.54 to 2.03 | toenail | both |
van den Brandt 1993 |
|||
| 0.91 | 0.41 to 2.00 | men | |||||
| 1.58 | 0.59 to 4.22 | women | |||||
| 0.80 | 0.68 to 0.95 | supplement use | both | Hansen 2013 | |||
| 1.09 | 0.63 to 1.89 | serum | both | Hughes 2015 | |||
| 1.32 | 0.55 to 3.19 | men | |||||
| 0.76 | 0.32 to 1.80 | women | |||||
| Urological cancers | Renal cancer | incidence | 0.40 | 0.17 to 0.98 | dietary intake | both | Banim 2013 |
| Urinary tract (all) | incidence | 0.97 | 0.72 to 1.31 | serum | both | Hotaling 2011 | |
| 0.81 | n.r. | serum | men | Knekt 1990 | |||
| 4.12 | n.r. | women | |||||
| Respiratory tract | Cavum oris/pharynx | incidence | 5.43 | n.r. | serum | both | Menkes 1986 (Zheng 1993) |
| Skin | Melanoma | incidence | 1.66 | 0.71 to 3.85 | toenail | women | Garland 1995 |
| 0.90 | 0.30 to 2.50 | serum | both | Menkes 1986 (Breslow 1995) | |||
| 0.98 | 0.69 to 1.41 | suppl. intake | both | Peters 2008 (Asgari 2009) | |||
| Any non‐melanoma cancer | incidence | 0.77 | n.r. | plasma | both | Clark 1985 | |
| Basal cell carcinoma | incidence | 0.54 | n.r. | serum | men | Knekt 1990 | |
| 1.55 | n.r. | women | |||||
| 0.80 | 0.10 to 4.5 | serum | both | Menkes 1986 (Breslow 1995) | |||
| 0.86 | 0.38 to 1.96 | serum | both | McNaughton 2005 | |||
| 0.95 | 0.59 to 1.50 | intake | |||||
| Squamous cell carcinoma | incidence | 0.69 | 0.51 to 0.92 | plasma | both | Combs 1993 | |
| 0.60 | 0.20 to 1.50 | serum | both | Menkes 1986 (Breslow 1995) | |||
| 0.86 | 0.47 to 1.58 | plasma | both | Karagas 1997 | |||
| 1.30 | 0.77 to 2.3 | intake | both | McNaughton 2005 | |||
| 0.49 | 0.24 to 0.99 | serum | |||||
| Other | Haematological | incidence | 0.60 | n.r. | serum/plasma | both | Coates 1988 |
| incidence | 0.95 | 0.75 to 1.20 | suppl. intake | both | Walter 2011 | ||
| Thyroid | incidence | 0.13 | 0.02 to 0.77 | serum | both | Glattre 1989 | |
| 0.15 | 0.0 to 5.0 | men | |||||
| 0.12 | 0.01 to 1.11 | women | |||||
| 1.35 | 0.99 to 1.84 | intake | both | O'Grady 2014 | |||
| 1.23 | 0.71 to 2.12 | men | |||||
| 1.14 | 1.65 to 2.02 | women |
n.r. = not reported.
Discussion
Summary of main results
The aims of this review were to examine the efficacy of selenium supplementation in preventing cancer and, more generally, to analyse the association between selenium exposure and risk of cancer in men and women.
Randomised controlled trials (RCTs) and preventive efficacy
We aimed to identify all RCTs so far carried out, extending the standard search by using unconventional methods such as citation chasing and scanning of conference proceedings ‐ methods that have proved effective in yielding additional high‐quality evidence for systematic reviews and meta‐analyses for other topics (Greenhalgh 2005; Vinceti 2017c). Using this approach, we identified a total of 10 RCTs that investigated monoselenium supplements for prevention of non‐melanoma skin cancer, prostate cancer, any cancer, and other site‐specific cancers. Overall, clear and consistent evidence indicates that selenium supplementation did not reduce subsequent cancer incidence, whether this endpoint was considered a primary or secondary outcome. Most of these trials raised concerns about possible harmful effects of selenium supplements, including increased incidence of non‐melanoma skin cancer in the Nutritional Prevention of Cancer Trial (NPCT), dermatological effects in the Selenium and Vitamin E Cancer Prevention Trial (SELECT), and type 2 diabetes in all RCTs, although with generally limited and statistically imprecise risk ratios (RRs).
Of the three liver cancer prevention trials, one reported a strongly reduced risk of liver cancer for male carriers of the hepatitis B surface antigen (HBs‐Ag) taking inorganic selenium supplements (sodium selenite) for three years, and the other two studies reported little effect of organic selenium supplements (selenium yeast) for the same cancer site (Li 2000; Yu 1991; Yu 1997). Owing to several methodological concerns related to randomisation and completeness of outcome data, we judged the risk of bias as unclear for all three of these RCTs. Therefore, we could not conclude that we found strong support for selenium supplements as agents for prevention of liver cancer. Unfortunately, the other trials did not include liver cancer among their secondary outcomes, with the exception of ECOG 5597 (Karp 2013). In this RCT, investigators reported new cases of liver, gallbladder, and bile duct cancer only among selenium‐treated participants; however, trial authors observed a total of only six cases, making risk estimates highly statistically unstable. In addition, the population included in this trial, which comprised patients with a history of resected non‐small‐cell lung cancer, was rather different from the general population.
The NPCT (NPCT 2002) reported strongly decreased risk for all cancers (‐22%), and for oesophageal (‐59%), colorectal (‐52%), lung (‐28%), and prostate (‐46%) cancers, showing lesser decreases compared with the ad interim report (Clark 1996, in: NPCT 2002), but still indicative of a strong cancer preventive effect. In addition, when participants were categorised into tertiles according to baseline serum selenium, evidence suggested an inverse relationship between selenium status and effects of supplementation for all cancers and for prostate cancer in the lower two tertiles, and no effect in the upper tertile. However, interpretation of these results is difficult because in 2003, the trial authors acknowledged the occurrence of a detection bias, namely, a considerably higher rate of prostate biopsy in the placebo group, whose cause was not specified. It is unclear whether this detection bias applied only to prostate cancer or applied more generally to other outcomes (as would be the case if the bias was due to unblinding, for example). This major detection bias forced us to downgrade the reliability of this study. Data show an increase in the incidence of its primary outcome ‐ non‐melanoma skin cancer ‐ in selenium‐supplemented participants, as well as in the incidence of five other cancer types, including melanoma, bladder cancer, breast cancer, head and neck cancer, and lymphoma and leukaemia. Trial authors stated: "These results, although non‐significant and based on small case numbers, may indicate potential increased risk with selenium supplementation"; these authors also relied on previous observational studies to provide some support for these positive associations (Duffield‐Lillico 2002, in: NPCT 2002).
The turning point of research on selenium and cancer was the SELECT trial (SELECT 2009), a large, well‐conducted prostate cancer prevention trial carried out in the male general population of North America not at high risk of prostate cancer (≤ 4 ng/mL in serum prostate‐specific antigen (PSA) and digital rectal examination not suspicious for cancer). This trial, widely considered a milestone in cancer prevention and research, found no difference in prostate cancer incidence for selenium–supplemented participants as compared with placebo participants after a median follow‐up of 5.5 years (hazard ratio (HR) 1.04, 95% confidence interval (CI) 0.90 to 1.18), and no effect of selenium on risk of overall cancer or on risk of other cancers (as well as cardiovascular disease). Median selenium at baseline (135 µg/L in serum in the selenium arm vs 137.6 µg/L in the placebo arm) was higher than in the NPCT (average plasma selenium 114 µg/L). The intervention used in this trial was different from that used in the NPCT (selenomethionine in SELECT, and selenised yeast in the former), although this is unlikely to have been responsible for observed differences (Waters 2013); in both cases, the intervention comprised organic selenium species (Block 2004).
In a small study of organ transplant recipients (Dreno 2007), an unexpected increase in non‐melanoma skin cancer incidence emerged; this was a matter of concern in the light of results of the NPCT. In the Polish trial Lubinski 2011, which included 1135 women with high genetic susceptibility to breast cancer due to BRCA1 mutations, evidence was more consistent with increased risk of both all cancers and primary breast cancer than with decreased risk, although with statistically unstable HRs (1.4, 95% CI 0.9 to 2.0; and 1.3, 95% CI 0.7 to 2.5, respectively). In this trial, the intervention consisted of administration of 250 µg/d of inorganic tetravalent selenium (selenite).
More recently, results of three well‐conducted phase 3 trials in participants at higher risk for prostate cancer than the general male population indicated that 200 µg/d of selenium (as selenomethionine in one study ‐ Marshall 2011 ‐ and as selenised yeast in the other two ‐ Algotar 2013; Karp 2013) did not decrease subsequent cancer incidence compared with placebo. The baseline selenium status of populations included in these RCTs was comparable with that in SELECT for Southwest Oncology Group (SWOG) S9917 (135 to 138 µg/L in the two arms) (Marshall 2011), slightly lower in the Negative Biopsy Trial (NBT) (126.1 µg/L) (Algotar 2013), and unfortunately unspecified for Eastern Cooperative Oncology Group (ECOG) 5597 (Karp 2013). Results of these high‐quality RCTs, all characterised by low risk of bias and two of which were discontinued before their planned end for futility, were consistent and showed no beneficial effect of selenium treatment on cancer risk.
Although not eligible for our meta‐analyses because their outcome was non‐malignant neoplasms rather than cancer, two recently published RCTs on colorectal adenoma risk in participants receiving selenium are worth noting. One of these trials was embedded in SELECT (Lance 2017), and the other, the SELCEL trial (an intervention study investigating the effect of selenium supplementation or celecoxib for prevention of colorectal adenoma recurrence), allocated 1374 men and women who had undergone removal of colorectal adenomas to either 200 µg/d selenium as selenised yeast, or placebo (Thompson 2016). Both RCTs did not find a beneficial effect of selenium for prevention of colorectal adenoma.
The RCTs carried out on selenium have generated clear evidence of adverse effects associated with selenium exposure, showing both the health effects related to overexposure and the amount at which these effects become evident, thus providing much more reliable evidence than that generated by environmental studies such as Vinceti 2017a for use in risk assessments of the safe upper limit of selenium exposure in humans. The trial that provided the most evidence about selenium‐associated adverse effects was SELECT. These effects include an excess risk of dermatitis and alopecia, non‐melanoma skin cancer, high‐grade prostate cancer, and type 2 diabetes. The excess risk of dermatological effects was anticipated as a potential side effect based on previous knowledge of health consequences of human overexposure to this element (Vinceti 2001), although such effects had been predicted to occur at higher amounts of selenium exposure than those experienced by SELECT supplemented participants, thus calling for reassessment of the upper limit of selenium exposure. The increased incidence of non‐melanoma skin cancer in NPCT and of advanced prostate cancer in SELECT was extremely disappointing, as they were the primary endpoints in these studies, and the expectation was that they would be reduced. The excess risk of diabetes in selenium‐supplemented NPCT participants, which was also an unanticipated finding, was mostly limited to participants in the two highest tertiles of baseline plasma selenium (> 105.2 µg/L), raising concern about the safety of selenium amounts that thus far had been considered entirely safe (i.e. on the order of 200 µg/d) (Stranges 2007). Therefore, subsequent RCTs added this endpoint to monitored adverse effects that contributed to interruption of the SELECT trial, together with the null effect on cancer mortality and adverse effects of vitamin E on prostate cancer risk (Lippman 2009, in: SELECT 2009). So far, all RCTs that included diabetes among trial endpoints, including trials investigating risk of colorectal adenoma, have shown an increased incidence of type 2 diabetes among selenium‐allocated participants, with RRs ranging from 1.08 to 1.71, although most estimates were statistically imprecise (Vinceti 2017b). In addition, in SELECT, a slight decrease in excess risk of diabetes in the intervention arm followed completion of selenium supplementation, further suggesting a causal relation between selenium administration and the disease (Lippman 2009 and Klein 2011, in: SELECT 2009). Currently, an excess risk of type 2 diabetes appears to be one of the adverse effects of selenium of greatest concern, and its plausibility is supported by the results of observational human studies (cohort, case‐control, and cross‐sectional), as well as by some biological plausibility (Galan‐Chilet 2017; Su 2016; Thompson 2016; Vinceti 2015; Vinceti 2017b; Zhou 2013). These side effects, in addition to the null results of RCTs, particularly of those of the highest quality, make implementation of new trials very unlikely owing to ethical concerns.
Observational studies and aetiological association
From our meta‐analyses of 16 prospective observational studies on overall cancer risk, we found lower cancer risk associated with highest selenium exposure compared with lowest exposure. Risk of cancer was 28% (95% CI 7% to 45%) lower in the highest category of selenium exposure than in the lowest, and risk of death from cancer was 24% (95% CI 3% to 41%) lower. Subgroup analyses by sex yielded increased evidence of this inverse association between selenium exposure and cancer risk in men compared with women.
The inverse association between overall cancer risk and baseline selenium levels was mainly attributable to lower risks of gastrointestinal, lung, and bladder cancer, and for men also prostate cancer. No association was seen between selenium and risk of breast cancer in women. However, when the amount of baseline exposure was taken into consideration, no clear and consistent trend between baseline selenium exposure and risk emerged for any of the major outcomes investigated in observational studies. Lack of lower risk of cancer in the highest versus the lowest selenium category among participants with the lowest baseline exposure levels compared with those with intermediate or high levels, for overall cancer, lung cancer, and prostate cancer, argues against a causal association between selenium exposure and cancer risk. This is supported by lack of a relation between differences in the highest and lowest categories of selenium exposure and the corresponding RR, further suggesting that larger differences in exposure are not associated with large and consistent decreases in RR. Finally, further uncertainty of the evidence generated by observational studies arises from the inconsistent and sometimes sharply conflicting results on the same cancer type that emerged from different studies.
We saw little evidence of any effect of modification on the relation of selenium and cancer by geographical area of residence. It should however be noted that most of the observational cohort studies that we examined were conducted in Europe and in the USA, and none were conducted in Africa or South America. This regional distribution seems to reflect the under representation of non‐Western and resource‐poor countries in epidemiological research (Pearce 2004). Differential regional representation in epidemiological studies is of special interest for this review, as selenium levels in humans around the world vary significantly. Even if selenium levels measured in included cohorts reflect a broad range of naturally occurring selenium exposure, investigators have reported some of the lowest and highest levels of selenium exposure in populations from South America (Jaffé 1992), Africa (Hurst 2013b), China (Li 2012), and India (Chawla 2016) ‐ regions not investigated by any of the reviewed observational studies, with the exception of three Chinese trials. Concerning sex‐related effects, our meta‐analysis of longitudinal studies revealed an inverse association between RR of cancer and selenium status in some cases in men but not in women for the same cancer type. Unfortunately, although more than half of reviewed studies included mixed‐sex populations, most did not report sex‐disaggregated results. In available sex‐specific results, men are over represented ‐ a fact that may potentially hamper assessment of the relation between selenium exposure and cancer risk in women. Theoretically, factors such as variations in body composition between men and women, including lean body mass versus fat composition, or differences in metabolism or in nutritional requirements (e.g. higher antioxidant requirements, particularly for the urological system) between the two sexes might be associated with differential effects of selenium for prevention of cancer.
Concerning the indicator used to assess selenium exposure and its relation with cancer risk, we observed generally null associations when evaluating selenium status through assessment of dietary intake, although some inverse associations at specific cancer sites emerged when we used biomarkers such as blood or toenail selenium levels. We extensively reviewed in the previous version of this review the characteristics and limitations of indicators of selenium exposure, with particular reference to dietary assessment methods and biomarkers, and inconsistencies across studies assessing the validity of different indicators (Ashton 2009; Fairweather‐Tait 2011; Jablonska 2015a; Vinceti 2014). In particular, a large body of literature concerns the limitations of dietary assessment methods, mainly linked to large variations of selenium content in single food types, and the limitations of biomarkers of exposure. Concerning the latter, a major source of exposure misclassification consists of the different behaviours of inorganic and organic selenium species, whose tendency to be retained in the body and to accumulate in specific body tissues greatly varies, although this does not necessarily correlate with their biological activity (Behne 1996; Behne 2010; Kim 2001; Michalke 2017; Panter 1996; Slavik 2008; Solovyev 2013; Steen 2008; Tiwary 2006; Vinceti 2013c). Investigators have frequently proposed that selenoprotein activity may be an indicator of selenium status and may be tested in association with cancer risk (Vinceti 2017b), but this relation has been questioned because different sources of oxidative stress, paradoxically including pro‐oxidant selenium species themselves, may upregulate selenoprotein activity (Jablonska 2015a). Furthermore, intake of heavy metals and other dietary factors such as vitamins, metalloids, and amino acids (e.g. methionine) may modify the health effects of selenium, or the relations between selenium exposure and biomarkers (Jablonska 2015a; Vinceti 2000), owing to metabolic interactions or changes in tissue‐specific deposition and retention of selenium (Behne 1996; Zeng 2005; Zwolak 2012).
Overall, available evidence indicates the potential for exposure misclassification in observational studies on selenium, as well as the pitfalls associated with an approach based on assessment of total selenium content in peripheral biomarkers, suggesting that in some instances, measurements of nutritional intake might provide better exposure estimates than are provided by biomarkers, particularly in the light of relative exposure to inorganic and organic species of the element. In general, observational cohort studies on selenium and cancer are expected to have been characterised by random exposure misclassification, thus shifting RRs towards the unity and reducing the ability to detect real associations. However, some exposure misclassification may have been non‐random, such as that induced by smoking, which although it is a source of selenium exposure also induces lower body selenium levels, possibly owing to an effect of cadmium in increasing selenium excretion (Vinceti 2000). In such cases, exposure misclassification based on biomarkers (serum/plasma selenium levels) may have substantially biased risk estimates and may have been associated with some degree of confounding due to the well‐known effect of smoking on cancer risk, which could not have been adequately captured and controlled for. Inadequate control for smoking has been suggested to be a major confounder inducing spurious associations between low selenium levels and enhanced cancer risk in observational studies (Beane Freeman 2015).
In addition to exposure misclassification, and probably more important than this, a major issue affecting observational studies is unmeasured confounding (Vinceti 2016a). This potential bias is a matter of greater concern than exposure misclassification because it may have systematically biased RRs in one direction, particularly for some cancer types. Moreover, detection (and control) of this bias is extremely difficult and nearly impossible, given the hundreds of nutritional and non‐nutritional lifestyle variables that may be associated with both variations in selenium intake and cancer risk. Among these factors are smoking (Beane Freeman 2015; Vinceti 2013b), socioeconomic status ‐ which appears to be positively associated with socioeconomic position in both men and women (Gundacker 2006; Niskar 2003) ‐and most likely hundreds of nutritional and toxicological factors that may vary in the diet, together with selenium intake. An approach that would reduce the risk of unmeasured confounding in observational studies might include investigation of dietary patterns rather than single nutrients, but these investigations seem not to have made adjustments for diet quality. Finally, it should be noted that most studies did not take into account the role of genetic factors (related to selenoproteins or otherwise) in the relation between selenium exposure and cancer risk, although some studies have suggested the importance of such relations (Jablonska 2016; Meplan 2014); the true relevance of genetic factors has not yet been well defined. Some studies examining selenoprotein‐related single‐nucleotide polymorphisms have suggested a role for genetic variants among genes coding for selenoproteins in modifying cancer risk, or in determining the relation between selenium exposure and subsequent cancer risk, although results have not been consistent (Geybels 2013; Meplan 2012; Penney 2010; Penney 2013; Slattery 2012; Takata 2011).
With awareness of the fundamental limitations of observational studies, even of those of longitudinal design, which may avoid selection bias or reverse causality, investigators designed and carried out in the 1990s and the 2000s several experimental studies as RCTs investigating the effect of selenium supplements on cancer risk. The evidence base from these intervention studies has become so large and complete as to allow a comprehensive evaluation of cancer risk associated with selenium supplementation for some specific cancer types. It is interesting to note that major interest in the cancer preventive activity of selenium originated not just from observational studies (mainly of ecological and cohort design) (Vinceti 2013b), but from a randomised trial ‐ the ad interim analysis of the NPCT, which was published in 1996 and attracted great interest from both the scientific community and the general public because of the apparently large beneficial effect that it reported (Clark 1996, in: NPCT 2002). Null results of the most recent low‐bias RCTs ‐ Algotar 2013; Marshall 2011; SELECT 2009 ‐ also do not suggest a major or strong role of genetic factors in modifying selenium and cancer relations, given their generally null or troubling results. An exception can be seen in recent data from SELECT, which suggest that a genetic variant of the NKX3.1 androgen‐regulated prostate tumour suppressor protein may modify, or increase, the risk of prostate cancer associated with selenium supplementation (Martinez 2014).
From a methodological perspective, we acknowledge that comparison of risks between highest and lowest exposure categories in observational studies, as performed in the present meta‐analysis, is most suitable for identifying an effect when a consistent decrease or increase is seen across absolute exposure levels. Other associations (e.g. threshold effects, U‐shaped relations) may have been missed by this method of meta‐analysis, or their true effect might have been diminished.
Overall completeness and applicability of evidence
RCTs and preventive efficacy
This review investigated a diverse range of cancers, substantially extending the analysis compared with that performed in previous reviews. However, cancer is not a uniform condition, and malignant neoplasms show great differences in tumour biology. Only non‐melanoma skin cancer, liver cancer, and prostate cancer have been investigated as primary outcomes in the included prevention trials, and, regarding these main outcomes, specific characteristics of study populations may limit the generalisability of results. Participants in included RCTs on skin and liver cancer belonged to populations at high risk for the outcome under investigation, and participants in high‐quality prostate cancer trials were at average risk (Karp 2013; SELECT 2009), or at high risk (Algotar 2013; Marshall 2011), for this disease. Most participants in the NPCT were older and white, predominantly male inhabitants of the United States, and the most recent trials were limited to the USA male population.
Average baseline selenium exposure in the NPCT was less than that characterising subsequent trials carried out in the United States, although it was more similar to that seen in some European populations. Although the NPCT suggested that selenium supplementation was beneficial only at the lowest range of baseline selenium exposure, the most recent studies, carried out in populations generally characterised by higher average selenium exposure, did not confirm such an interaction. The NPCT also found an indication of strong effect modification for sex, as demonstrated, for example, by the HR for all cancers associated with selenium supplementation ‐ 0.67 (95% CI 0.50 to 0.89) in men and 1.20 (95% CI 0.66 to 2.20) in women (NPCT 2002).
Participants in the SELECT study on prostate cancer prevention were apparently healthy men over 50 years of age from the general population of North America (SELECT 2009). A large sample size and inclusion of non‐white participants from different socioeconomic backgrounds support the generalisability of study findings to other adequately nourished populations.
Selenium supplements generally contain organic or inorganic species of selenium, or a mixture of both (e.g. in the form of selenised yeast). Different species of selenium may exhibit different effects on human health and more specifically on proteomic endpoints, as also suggested by human controlled randomised trials though with inconsistent results (Ravn‐Haren 2008; Richie 2014). High‐quality RCTs using selenised yeast supplements, almost entirely comprising organic selenium forms (Block 2004; Waters 2013), found no effect of supplementation on the main study outcome and an indication of a harmful effect (i.e. an excess diabetes risk) (Vinceti 2017b). The SELECT trial used supplements of L‐selenomethionine, which is the major component of selenised yeast, and also found no preventive efficacy. The only two RCTs investigating sodium selenite supplements found a protective effect against liver cancer, and null or adverse effects on breast cancer risk, but we considered these trials to have unclear risk of bias. It is unclear how applicable these results are in other settings and in populations with a different nutritional status. Interpretation of the results of clinical trials using selenium supplements should consider the different chemical forms of selenium, as well as their potentially different health effects when used as supplements (Vinceti 2013c; Weekley 2013). Most studies used organic selenium as selenised yeast (Algotar 2013; NPCT 2002), or as selenomethionine (Marshall 2011; SELECT 2009). However, the chemical form used is unlikely to explain the differences in results between NPCT and the other trials (Waters 2013). With reference to this issue, of interest are the results of a 'natural experiment' that occurred in Northern Italy, wherein a small population unintentionally consumed for several years drinking water with an unusually high content of selenium in its inorganic hexavalent form ‐ selenate (Vinceti 2000). Follow‐up of that population revealed increased risk of neurodegenerative disease ‐ a not entirely unexpected finding owing to the potential neurotoxicity of inorganic selenium (Vinceti 2014a), along with a slightly increased risk of cancer, mainly due to excess risk of oropharyngeal cancer, melanoma, kidney cancer, and lymphoid malignancies (Vinceti 2016b).
An important issue is the possibility that participants with low baseline selenium status may experience an inverse association between selenium exposure and cancer risk, as suggested by some trial authors (Lu 2016; Rayman 2009). This has been suggested to explain the different results of SELECT and the NPCT, and could also hypothetically explain, at least in part, the different relations found in experimental as compared with observational studies. NPCT found a strong beneficial effect of selenium supplementation among participants at the lowest tertiles of baseline selenium levels; however, the risk of cancer changed abruptly from an apparently protective effect in the two lower tertiles (HR 0.51 and 0.70) to an excess risk in the highest tertile of plasma selenium (HR 1.20, 95% CI 0.77 to 1.86). This occurred despite a difference of only 16.4 µg/L between lowest and highest tertiles, corresponding to a change in dietary selenium intake as low as around 10 µg. This would imply that such a small a change in selenium dietary intake would change a strongly protective effect of the element on cancer risk into a possibly detrimental effect ‐ an implausible scenario given the wide range of selenium intake (from about 20 to several hundred micrograms) characterising Western populations. Moreover, the intermediate tertile of baseline plasma selenium in the NPCT (105.6 to 122.0 µg/L) appeared to be associated not only with reduced overall cancer risk but also with an excess risk of squamous cell skin carcinoma (HR 1.49, 95% CI 1.05 to 2.12) and overall non‐melanoma skin cancer (NPCT 2002), as well as diabetes (RR 1.36, 95% CI 0.60 to 3.09), whose risk also considerably increased at the highest tertile of baseline selenium (Stranges 2007). Overall, this occurrence of both adverse and beneficial effects is unlikely if the selenium supplementation was serving to remedy a selenium deficiency. In addition, the strongest effect of selenium on overall cancer risk at lower levels of baseline selenium status was due to a considerable decrease in prostate cancer, but this finding was subject to detection bias because of a decreased biopsy rate in selenium‐supplemented participants, particularly in those with lowest baseline selenium status, as recognised by investigators of the NPCT (NPCT 2002).
In addition, after NPCT, three of the four high‐quality RCTs on selenium supplementation for cancer prevention investigated the possible modifying effect of baseline selenium exposure and found no evidence of a beneficial effect of the intervention even in the lowest baseline exposure category. For instance, in NBT (Algotar 2013), the average baseline plasma selenium level at the lowest tertile of the study population was 101.1 µg/L ‐ much lower than the corresponding level at the middle tertile of NPCT (114.6 µg/L), in which the HR of prostate cancer had been as low as 0.33 (95% CI 0.13 to 0.82). However, in this ‘low’ NBT subgroup, investigators found no evidence of a beneficial effect of selenium supplementation on prostate cancer risk. In the SWOG S9917 trial (Marshall 2011), data show no change in the null effect of selenium in the two lowest categories (quartiles) of selenium intake, whose boundaries were < 106 and 106 to132 µg/L ‐ similar to cut points of the two bottom NPCT tertiles and of the bottom category of NBT. In these two subgroups of the SWOG population with the lowest baseline selenium status, the RR of prostate cancer was 0.82 (95% CI 0.40 to 1.69) and 1.38 (95% CI 0.68 to 2.78), and in the third upper quartile, the RR was 0.98 (95% CI 0.58 to 1.68), suggesting no consistent trend of an inverse relation between antecedent selenium exposure and effects of supplementation (as was also shown by analysis for trend in this study). Investigators in SELECT reported no reduction in cancer risk among selenium‐supplemented participants, although they did not provide specific RRs according to baseline selenium status. Calculation of blood selenium content distribution in SELECT, as well as in the three other RCTs (NPCT, NBT, SWOG), showed substantial overlap of plasma and serum selenium levels between this large trial population and the other study populations (Figure 7). In addition, a more recent case‐cohort study carried out within SELECT assessed the effect of selenium supplementation on prostate cancer risk, taking into consideration baseline selenium exposure, as assessed through toenail selenium levels. The study, which involved 1739 prostate cancer cases and 3117 controls, was unable to find a beneficial effect of selenium supplementation in the lowest categories (quintiles) of baseline toenail selenium (Kristal 2014). Actually, a dose‐response effect in that SELECT population emerged, but it favoured an increased risk of (high‐grade) prostate cancer induced by selenium supplementation among participants belonging to the two upper quintiles of baseline selenium exposure (Kristal 2014). Therefore, it seems reasonable to agree with this SELECT statement: “The analysis of our data using lower cut points for baseline toenail Se categories, in an attempt to replicate findings from the NPCT, also showed no evidence of benefit from supplementation among men with low baseline Se status (data given in Results). Given these findings, we believe it reasonable to conclude that Se supplementation of men at the low range of Se intake common in USA men will not reduce PCa risk” (Kristal 2014).
7.
Baseline circulating selenium levels in the NPC trial (Duffield‐Lillico 2003b in: NPCT 2002), the NBT (Algotar 2013), SWOG trial (as plasma selenium) (Marshall 2011), and SELECT (as serum selenium) (Lippman 2009, in: SELECT 2009). When median and interquartile values were reported, we estimated mean and standard deviation according to Cochrane guidelines provided in Higgins 2011a.
Overall, results of recent high‐quality RCTs do not support the hypothesis that differing baseline selenium status may explain conflicting results between NPCT and SELECT (Lu 2016; Rayman 2009). Results of the most recent RCTs seem therefore to be applicable to populations with various degrees of background selenium exposure, with the exception of populations characterised by extremely low (< 20 µg) or high selenium intake.
Observational studies and aetiological association
We reviewed data from prospective observational studies in which investigators measured selenium exposure in populations without evidence of cancer, who were then followed up for a specified period of time. We limited our systematic review to cohort studies to avoid or decrease two major sources of bias in observational investigations, particularly in case‐control and cross‐sectional studies (i.e. selection bias and risk of reverse causality). Data continue to show important differences among included studies in terms of selenium exposure assessment, types of outcomes, and study populations, which may affect their interpretation. The small number of studies that examined most of the meta‐analysed types of cancers prevented a thorough investigation of sources of heterogeneity between study results. In particular, we had limited opportunity to explore the influence of specific sources of bias or the methodological quality of epidemiological studies on heterogeneity.
Participants examined in this review update include more than 2,300,000 individuals, predominantly from Europe and North America, and, to a much lesser extent, from Asia and Australia. We were able to identify no prospective observational studies on selenium and cancer risk from Africa or South America. This regional distribution reflects the under representation of non‐Western and resource‐poor countries in epidemiological research (Pearce 2004). Differential regional representation in epidemiological studies is of special interest for this review, as selenium levels in humans around the world vary significantly. Selenium levels measured in the included cohorts reflect a broad range of naturally occurring selenium exposure, as documented by several epidemiological studies worldwide. However, some of the lowest and highest selenium levels in humans have been reported in populations in South America (Jaffé 1992) ‐ a region not investigated by any of the reviewed observational studies.
More than half of the included studies enrolled mixed‐sex populations, but most did not report sex‐disaggregated results. In available sex‐specific results, men are over represented ‐ a fact that could hamper potential assessment of the relation between selenium exposure and cancer risk in women. Despite this sex imbalance, we systematically saw stronger (inverse) associations with cancer risk among men than among women, for whom such associations with antecedent selenium status was nearly absent. This was true for stomach, colorectal, and lung cancer, and, when added to the inverse association for prostate cancer, led to an impact on overall cancer risk that was clearly lacking in women that could be due to potential confounders (such as smoking, occupational exposures, or other dietary factors) or to a real change in the association between selenium exposure and cancer risk in the two sexes.
The range of selenium exposure experienced by members of cohorts investigated in the observational studies was generally lower than that experienced by participants in RCTs, who added supplemental selenium, generally 200 µg/d and in its organic forms, to their usual background intake, which ranged from about 70 to 90 µg/d as organic selenium, although some RCTs provided no estimate (Jablonska 2015a). It is theoretically possible that a preventive effect of selenium against cancer exists only at low (< 30 to 50 µg/d) intake of the element, and that it disappears at higher intakes, when ‘saturation’ or ‘maximisation’ of selenoprotein expression driven by selenium occurs. Investigations have frequently chosen this proteomic endpoint as a reference point for deriving dietary reference values for selenium (Jablonska 2015a; Vinceti 2017a). Selenium exposure in the range of around 50 to 200 µg of daily selenium intake has not been tested by intervention studies, which have used larger amounts of supplemental selenium, and is unlikely to be tested in RCTs in the future, given the termination of past trials for futility or safety concerns. This possibility must be considered, but within the context of the two fundamental limitations of observational studies ‐ exposure misclassification and unmeasured confounding, which limit the reliability of the evidence they generate and its applicability in terms of cancer prevention.
A few lines of evidence suggest that even at low levels of selenium exposure, it is unlikely that such an inverse association with cancer risk exists. First, inconsistencies in the results found in our meta‐analysis for most cancer sites and lack of a dose‐response relation between cancer risk and selenium at varying levels of background selenium exposure, or of a difference between highest and lowest exposure categories, argue against a real relation between selenium and cancer risk. Limited differences between highest and lowest categories of selenium intake, often amounting to a difference of only 20 to 30 µg per day, compared with large variations in selenium intake worldwide (from 10 to 15 µg in low‐selenium areas up to several hundred µg in seleniferous areas), also argue against a true relation. Finally, as previously described, some recent high‐quality RCTs investigated the effect of baseline selenium status on cancer risk associated with selenium supplementation and found no beneficial effect of selenium supplementation, even among participants with the lowest amounts of baseline exposure. Overall, these findings do not support an association between higher selenium status and lower cancer risk independently from factors such as sex, baseline selenium exposure, and cancer type. One additional observational cohort study, which could not be meta‐analysed in this review because it was released in PubMed in July 2017, appears to confirm these conclusions (Sandsveden 2017).
Quality of the evidence
RCTs and preventive efficacy
SELECT (SELECT 2009), SWOG S9917 (Marshall 2011), NBT (Algotar 2013), and ECOG 5597 (Karp 2013) were the only trials considered to have low risk of bias with adequate sequence generation, allocation concealment, blinding, and reporting of findings, and the consistency of their findings for prostate cancer, as well as for other cancer types for the two trials investigating them (SELECT and ECOG 5597), added to the statistical power of the major trial (SELECT), making their overall results highly reliable and suitable for yielding useful evidence to assess the relation between selenium supplementation and cancer prevention. These trials are also of major importance because (with one exception) they have provided information about baseline selenium exposure and its possible modifying role and about the effect of selenium supplementation on subsequent cancer incidence. Another important feature of these trials has been their ability to address the issue of selenium overexposure and related adverse effects owing to a systematic surveillance system for adverse effects, as well as their ability to extend the monitoring programme to additional effects, if suggested by new analyses targeting previously unplanned secondary endpoints, as was the case for diabetes (Stranges 2007). This is particularly relevant because all of these trials were planned under the hypothesis, later found to be erroneous but at that time endorsed by regulatory agencies, that the supplemental selenium dose administered to intervention arms (200 µg/d in almost all RCTs) was entirely safe and was well below the upper safe limit of the element, even with consideration of background selenium exposure.
These trials may continue to yield important results. Secondary analysis of additional endpoints, or based on genetic and non‐genetic biomarkers of exposure to selenium and other factors, is still possible. For example, major contributions were yielded by SELECT in 2017, concerning outcomes such as prevention of colorectal adenoma and of Alzheimer’s disease by selenium supplementation, in both cases with null results (Kryscio 2017; Lance 2017).
We assessed the certainty of evidence from high‐quality RCTs using the GRADE approach (http://gdt.guidelinedevelopment.org/app/handbook/handbook.html#h.svwngs6pm0f2) and reported the results of this assessment in the 'Summary of findings' table (Table 1). From preliminary assignment to a high level of certainty due to the experimental study design, we did not identify reasons to downgrade trial quality according to standard GRADE guidelines for risk of all cancers, for risk of cancer mortality, or for risk of colorectal, lung, bladder, or prostate cancer. In contrast, meta‐analysis for breast cancer risk yielded a statistically imprecise result mainly reflecting the small number of cases, and meta‐analysis for non‐melanoma skin cancer showed high statistical heterogeneity across studies. When addressing factors possibly increasing the certainty of evidence assessment, we considered as non‐applicable the GRADE item “All plausible confounding would reduce the demonstrated effect or increase the effect if no effect was observed”, neither could we evaluate possible dose‐response gradients because unfortunately they were not tested in these RCTs. We therefore rated the certainty of evidence as ‘high’ if it indicates no effect of selenium supplementation on all cancers overall, on cancer mortality, nor on colorectal, lung, bladder, and prostate cancer, and we considered certainty of the evidence as ‘moderate’ if it indicates no effect on non‐melanoma skin cancer and breast cancer, with downgrades due to heterogeneity and imprecision, respectively. However, stating that the evidence supporting no effect of selenium on cancer prevention at these sites is of moderate rather than high certainty does not mean that the only alternative hypothesis is necessarily that selenium decreases risk of cancer at these sites. Actually, the overall results of high‐quality RCTs, when available, suggest a slight to moderate although statistically imprecise increase in the risk of some of these specific cancers following selenium supplementation.
Concerning the RCTs that we downgraded in our appraisal of risk of bias, we considered the quality of reporting to be an issue in the three trials on liver cancer prevention, thus leading to their classification as having unknown risk of bias. Several papers reported the individual trials, in some cases discrepantly, and essential questions regarding sequence generation, allocation concealment, handling of dropouts and withdrawals, and detection of outcomes remain unanswered. This might be due to inadequate reporting but might also hint at flaws in trial design and implementation. We were uncertain about whether the only trial that reported positive results for selenium supplements in liver cancer prevention randomly assigned participants individually. Cluster randomisation of participants who lived in the same area/village, which may have been the procedure used in this investigation, might have introduced additional bias to the study results (e.g. as the result of different environmental factors contributing to liver cancer development or detection) and might have led to an overestimation of the protective efficacy of selenium. Duplication of results of trials based on a rigorous study design would be necessary to assess the effects of sodium selenite on liver cancer incidence. With regard to the NPCT (NPCT 2002) and the trial of Dreno 2007, indications of serious detection bias for the USA study and of unclear methodological details (such as blinding) for the French investigation led us to consider these experimental studies to be at unclear risk of bias, as discussed in greater detail elsewhere in this review. As far as the trial on breast cancer is concerned (Lubinski 2011), our downgrade of evidence certainty was based on incomplete information provided in the only report that we could retrieve (an abstract), although we acknowledge the relevance of that trial ‐ the only trial specifically targeting breast cancer and a genetically specific population ‐ and the fact that complete reporting of trial procedures may lead to reassessment of trial quality and its upgrade.
Observational studies and aetiological association
The 70 observational studies were heterogeneous, not only in methodology, but also in the quality and level of detail of reporting and in their potential biases. We assessed our confidence in the evidence from these studies using the GRADE approach and reported our findings in Table 2; we reported judgements only for those outcomes evaluated in the 'Summary of findings' table for RCTs with low risk of bias.
Confounding and other biases
Selenium measurement and exposure misclassification
All studies on total cancer risk identified cases by using registry links or a combination of several methods, and losses to follow‐up were generally very low. One study on cancer incidence and two studies on cancer mortality analysed less than 80% of all identified cases (incidence: Coates 1988: 79%; mortality: Kok 1987a: 71%; Kornitzer 2004: 57%). The main reason for this loss of sample was missing selenium measurements. Not all studies that assessed mortality as a measure of cancer risk excluded people with cancer at study inception. This might have led to overestimation of a protective effect if selenium levels were lowered by the presence of cancer. We therefore consider the results for cancer incidence to be more valid than the cancer mortality results.
Concerning the outcome most frequently investigated ‐ prostate cancer ‐ all but two of the included studies identified cases by using links to cancer registries or a combination of personal follow‐up interviews with PSA screening. Two studies with health professionals used self‐reporting for case identification, followed by confirmation through medical records. The number of people lost to follow‐up was low in all included studies. However, two studies included less than 80% of all identified cases in their analyses because samples were not available for selenium measurement, or diagnosis was not confirmed (Brooks 2001: 39%; van den Brandt 2003, in: van den Brandt 1993: 77%). In Brooks 2001, bias might have been introduced to the results to some extent, as demographic variables differed between identified and analysed cases.
Residual confounding and effect modification
Most of the included studies used controls for smoking and age by matching or using multi‐variate techniques. However, the control for self‐declared smoking habits may be inadequate, and this may occur particularly in people with a diagnosis of cancer (Connor Gorber 2009; Gerritsen 2015; Morales 2013). Control for smoking as a known risk factor for several types of cancer is an important issue in epidemiological studies on cancer risk, and inadequate control for this cancer risk factor has been recognised as a major methodological issue affecting observational research on selenium and cancer (Beane Freeman 2015). This possible bias may be particularly relevant for research on selenium biomarkers and cancer. Cigarette smokers tend to have lower selenium biomarker levels, although cigarette smoking in itself is a source of selenium exposure. In addition to this source of non‐random exposure misclassification, it is well recognised that smoking is a powerful cancer risk factor, thus qualifying it also as a major confounder when the selenium and cancer relation is investigated. Therefore, an inverse association between low baseline selenium status and lung cancer risk might be the result of residual confounding and effect modification by smoking, and this may also be true for other cancer types (Beane Freeman 2015). Exposure to environmental and household smoking, which has been shown to be associated with increased risk of cancer (Gorlova 2006; Nishino 2001), might be associated with selenium status due to differential nutritional behaviours or other mechanisms.
Several other factors may act as effect modifiers or confounders. Possible confounding factors could consist of another food nutrient or a certain behaviour that exhibits cancer protective effects and may be associated with higher intake of selenium‐rich foods. The number of candidates for such a role is so large that no observational study can measure all of these factors nor account for them. Furthermore, it is well known that intake of heavy metals (such as arsenic, cadmium, and mercury) and other dietary factors such as methionine may substantially modify selenium health effects or relations between selenium exposure and biomarkers (overview, in: Vinceti 2000; Zeng 2005; Zwolak 2012), and may potentially confound the association between selenium and cancer.
Some potential confounders cluster in population groups according to socioeconomic position (SEP), and this factor has been shown to vary together with selenium status in both men and women (Gundacker 2006; Niskar 2003). Only a few studies attempted to control for indicators of adult SEP as potential confounders (e.g. education, occupation, income). None used a composite index of indicators or considered childhood SEP. Some studies restricted their cohorts to certain subgroups of a population, such as occupational groups, and were likely to include only people of a similar adult socioeconomic background. It has been claimed that associations between vitamins and diseases are the result of confounding by social and behavioural factors acting over the course of a lifetime (Lawlor 2004). Lawlor 2004 argued that divergent results from epidemiological and randomised controlled studies on prevention of cardiovascular disease can be explained by unmeasured confounding due to SEP. Risk of most cancers is known to decrease with higher SEP. Research also indicates a positive association between higher SEP and selenium biomarkers (Barany 2002; Niskar 2003). However, other investigations have not confirmed these findings: Kant 2007, for example, did not find an association between a measure of household poverty and selenium status. The hypothesis of possible confounding due to SEP leading to an indirect association between selenium and cancer would be consistent with results of observational studies for all types of cancers in this review, with the exception of prostate cancer. Dalton 2008 found that prostate cancer has been diagnosed more often in men of a higher SEP, and we saw a protective association of higher selenium exposure with this cancer type. It remains unclear whether the more frequent diagnosis of prostate cancer in men with a higher SEP actually reflects an excess of prostate cancer incidence in this population. It might also result from differential health and screening behaviours leading to detection of otherwise symptom‐free cases, while men with a lower SEP tend to be over represented in diagnoses of the disease at advanced stages (Rapiti 2009). More information on screening and diagnostic behaviours of male cohort participants would be necessary to further elucidate these issues.
Another consideration is genetic factors, which may both confound and modify the role of selenium in cancer prevention and causation. Recent observational studies examining selenoprotein‐related single‐nucleotide polymorphisms have suggested a role for genetic variants in genes coding for selenoproteins or other proteins in modifying cancer risk, or even the relation itself between selenium and cancer risk, although results have not been consistent (Gerstenberger 2015; Geybels 2013; Jablonska 2015b; Meplan 2015). Null results of the most recent low‐bias RCTs do not suggest that at least the most frequent genotypes strongly influence the selenium and cancer relation (Algotar 2013; Marshall 2011; SELECT 2009), although such hypotheses cannot be ruled out for more rare genetic variants of selenoproteins or other proteins. Hypothetically, different genetic factors could increase and decrease the risk of cancer associated with selenium exposure, cancelling each other out and resulting in an overall null effect. Additional data from SELECT based on genotyping of study participants, if available, might be extremely useful for assessing hypotheses regarding genetic variants of selenoenzymes and their interaction with selenium status. So far, the only evidence derived from SELECT indicates that single‐nucleotide polymorphisms related to the prostate tumour suppressor protein NKX3.1 gene (CC genotype at rs11781886) may increase cancer risk following selenium supplementation (Martinez 2014). Recent observational evidence also suggests that polymorphisms of selenoproteins and other antioxidant proteins in men with non‐metastatic prostate cancer may be associated with increased risk of high‐grade disease and subsequent prostate cancer recurrence (Gerstenberger 2015).
Summary
In observational studies, factors that may have accounted for inter‐study heterogeneity and that may have biased study results include type of outcome measure, exposure assessment, sex, incomplete control for confounding (smoking and socioeconomic position), and unmeasured confounding, linked to both dietary and non‐dietary factors. Given the high risk of bias due to these factors, particularly to the unmeasured confounding inherent in observational studies, along with conflicting results of several studies and lack of any modification of the selenium and cancer relation by level of baseline selenium exposure and by the difference between highest and lowest selenium categories, we consider the evidence provided by observational studies to have very low certainty (Table 2); therefore these results must be interpreted with great caution and do not allow firm conclusions about a possible cancer‐preventive effect of selenium intake. Meta‐analyses of spurious findings in observational studies enhance the precision of a summary risk estimate, which does not itself get nearer to the true value and may suggest a non‐existent association (Egger 1998).
Potential biases in the review process
RCTs and preventive efficacy and observational studies and aetiological association
The literature search included major international databases in the English and German languages, and we applied a broad search strategy supplemented by handsearching for references. We assume that we identified all randomised controlled studies and prospective observational studies relevant to our review questions. As we did not search databases in other languages (e.g. Chinese, Russian), we cannot rule out that we might have missed smaller studies that were not published in international journals. However, we consider it unlikely that we could have missed major sources of evidence through our approach. We also might have missed observational studies whose results on selenium exposure and cancer were reported in the body of a paper but were not mentioned in the paper's title or abstract, even if the paper is indexed in the searched databases. However, our systematic use of backward and forward citation chasing and our search for relevant abstracts in conference proceedings or related material should have substantially decreased the risk of missing literature that could have been relevant for our assessment.
When needed because of lack of complete or appropriate participant data (e.g. when cohorts including cancer and non‐cancer participants were mixed in data analysis), we contacted study investigators to ask for data missing from their studies. We also did this when we did not have enough data from published reports to adequately appraise study risk of bias. Sometimes we were unable to obtain answers to questions that we had regarding methods or outcomes, but frequently investigators kindly gave us the information we needed. We were sometimes unable to obtain answers, particularly for earlier epidemiological studies from which primary investigators may have relocated or died, or we found that data were not available in a current electronic format. Similarly, we could not make contact with primary investigators of Chinese RCTs.
We based our risk of bias assessment on information included in the original publications, unless the trial authors that we contacted gave us additional details. This means that in some instances, we may have overestimated the true risk of bias of studies that did not adequately describe their design in the original publications, such as Lubinski 2011.
Another concern, especially with epidemiological studies, is publication bias. Cohort and nested case‐control studies often are not exclusively designed to test for a specific exposure‐outcome association but enable investigators to investigate a range of questions. It is conceivable that unfavourable results were less likely to be published, although we could not find evidence supporting such a hypothesis. Our analysis of this issue through use of a funnel plot gave some support to publication bias for prostate cancer.
We systematically meta‐analysed RCTs even when only two studies were available (Karp 2013; NPCT 2002). Finally, we carried out two meta‐analyses of intervention studies ‐ one on all studies, and another on RCTs assessed through a standard appraisal tool as being at low risk of bias ‐ and we emphasised results of the latter, as derived from high‐quality experimental studies. For observational studies, we decided a priori to conduct meta‐analyses only when five or more studies were available for a study outcome, thus excluding from meta‐analysis the few endpoints for which up to four studies were available (Table 3). Our primary intention was to facilitate the investigation of heterogeneity between studies included in meta‐analyses, to avoid producing more precise, but still unexplainably biased, results. However, our emphasis was clearly given to experimental studies because this trial design is widely recognised as the only one that may provide convincing evidence of an association between a factor and disease risk, or more generally biological endpoints, and this may be particularly true in nutritional epidemiology (Vinceti 2016a).
Finally, the authors of this review, as already noted in the previous version of the review, came from different disciplines and have different areas of focus (e.g. epidemiology, biostatistics, clinical medicine, nutrition). We continue to consider such variety of expertise to be a strength of this review, and we made use of it by applying multiple checking procedures during the entire review process whenever possible.
Agreements and disagreements with other studies or reviews
Recent reviews that have investigated the relation between selenium and cancer prevention have generally concluded that this trace element has no clear beneficial effect(Bjelakovic 2012; Cortes‐Jofre 2012; Cui 2017a; Fortmann 2013; Kushi 2012; Moyer 2014; Posadzki 2013; Schwingshackl 2017), although updated systematic reviews and meta‐analyses on selenium encompassing all of the most recent intervention studies are lacking. These results are true for both all cancers and prostate cancer, and for other specific cancers, such as lung cancer. The turning point in the evaluation of the effect of selenium on cancer risk is generally acknowledged to have been SELECT, and the other trials, although their findings are consistent with SELECT, have received less attention, probably mainly because of their smaller size. It is understandable that most of the selenium trials under way during the 2000s and the 2010s and originally implemented mainly as the result of the promising results of the original NPCT, particularly its ad interim 1996 report, were eventually discontinued owing to the results of SELECT (which was discontinued too) and the null results of ad interim futility analyses (Vinceti 2017b). This seems also to be true for Brodin 2015, Chen 2013, and Vinceti 2017a ‐ planned RCTs on the possible utility of selenium for cancer therapy ‐ and is an issue of considerable interest that has been investigated so far in very few phase 2 and phase 3 trials (Goossens 2016; Karamali 2015; Muecke 2014; Stratton 2010) (although other trials appear to be under way such as Vinceti 2017b).
Concerning observational studies, very few recent reviews have investigated the selenium and cancer relation, and they have focused on only a few cancer sites. These reviews have generally yielded results consistent with ours. For prostate cancer, a recent review found no association between baseline serum selenium and risk in cohort studies (Cui 2017b), as was reported by Allen 2016, which conducted a pooled analysis using individual data from 15 cohort studies. However, in the latter review, baseline serum selenium status was determined to be inversely associated with high‐grade prostate cancer risk, as was toenail selenium and subsequent prostate cancer incidence. Gong 2016 also found reduced risk of gastric cancer among participants in the highest baseline selenium exposure category. Other reviews and meta‐analyses considered other cancer types such as liver, pancreatic, lung, and breast cancer, but these reviews generally incorporated case‐control and cross‐sectional studies in addition to cohort studies, further increasing the risk of bias due to heterogeneity of study designs. Most reviews on observational studies have acknowledged the key methodological issues noted in this type of study, namely, risk of unmeasured confounding and potential biases associated with this limitation.
Authors' conclusions
Implications for practice.
A large body of evidence is now available from high‐quality randomised controlled trials on effects of selenium supplementation on cancer risk, with two new studies published since the last version of this review (Vinceti 2014). None of the new relevant studies have provided information to change the conclusions of the previous version of this review. Overall, results of these studies have consistently shown no effect of selenium in preventing the type of cancer most consistently and strongly associated with antecedent selenium exposure ‐ prostate cancer ‐ or in preventing cancer overall, even when assessment focused on participants with the lowest selenium status at baseline. These intervention studies have suggested that selenium administration on the order of 200 µg/d increased risk of non‐melanoma skin cancer, advanced prostate cancer (in individuals with highest baseline exposure), dermatological abnormalities, and type 2 diabetes. No trial involving administration of low doses of selenium, on the order of 50 to 100 µg/d, has been performed so far.
An update of the meta‐analysis of observational cohort studies continues to show lower risk of cancer and of some specific cancers (colorectal, prostate, and breast) in participants with highest exposure levels at baseline, but these studies are at substantial risk of bias from exposure misclassification and unmeasured confounding. In addition, results of these observational studies are inconsistent and sometimes are strongly conflicting, and no evidence of any dose‐response relation emerged from our analysis when we considered background selenium status or differences in baseline selenium exposure.
Overall, findings of our review do not provide evidence supporting a cancer–preventive effect of selenium in humans.
Implications for research.
Some questions regarding selenium, such as whether selenium might influence cancer risk in individuals with very low or very high baseline exposure to this element, or in individuals with different genotypes, have not been fully resolved, although currently available evidence from randomised trials offers little support for such hypotheses. For ethical reasons, in the light of potential toxicity of selenium supplementation and failure of the most recent and well‐conducted experimental cohort studies to find beneficial effects, new randomised trials on the selenium and cancer relation are unlikely to be undertaken in the future. Therefore expanding results of the SELECT trial and of other high‐quality trials to examine additional outcomes such as liver cancer and non‐melanoma skin cancer, as recently happened for other outcomes (Kryscio 2017; Lance 2017), and to analyse subgroups with specific characteristics (baseline selenium exposure and genetic factors), continues to appear to be the best available option for clarifying these issues. Unfortunately, most of these randomised controlled trials (RCTs), including the Selenium and Vitamin E Cancer Prevention Trial (SELECT), could not address possible sex differences because they enrolled only men.
Finally, when interpreting the results of both intervention and observational studies, it must be taken into account that various chemical forms of selenium have very different nutritional and toxicological properties, and that almost all observational studies have assessed only total selenium exposure. Future observational studies would contribute to a better understanding of the selenium and cancer relation by including selenium speciation among their exposure assessment methods when evaluating cancer risk.
Feedback
Selenium for preventing cancer, 23 November 2011
Summary
Re: Dennert et al., Selenium for preventing cancer, The Cochrane Library, 2011, Issue 5. As selenium scientists with considerable knowledge of the selenium‐cancer field, we wish to draw to the attention of The Cochrane Collaboration the shortcomings of the recent review cited above. We contend that the quality of this review is not up to the expected standard of Cochrane systematic reviews. We are not criticising the way in which the analyses were performed, but rather the ways they were interpreted and summarised, which we believe to be overly negative and rather biased. For these reasons, we find the resulting report to be misleading to the reader. Some of the weaknesses are listed below. Abstract and Plain Language Summary: These sections do not fairly represent the findings of the review. Contrary to the impression given in these summaries, the review itself demonstrates that there is in fact a considerable body of evidence, much of it from prospective observational studies, for a beneficial effect of selenium on a number of cancers. The stated summary of RCT findings is more conclusive than it should be, given the very small number of published clinical trials with selenium alone and the limited trial data that the review authors arbitrarily chose to consider. Furthermore, the NPCT is treated very harshly, and its secondary findings (lung, colorectal and prostate cancers) are more or less discounted. Body of the Paper: 1. Lack of appreciation of the importance of baseline selenium status in influencing trial outcomes (i.e. the fact that only people with a low selenium status profited from supplementation). For example, no acknowledgement was made of the fact that lack of benefit of a 200 μg/d dose of selenium for cancer risk in SELECT occurred in participants with relatively high baseline serum selenium concentrations—well above those found to confer benefit from selenium supplementation in the NPC trial (NPCT). This point was raised by us previously (Rayman et al. JAMA 2010). 2. Lack of discrimination between trials in which supplementation with selenium had the capacity to maximise selenoprotein expression/concentration (e.g., NPCT) and those (e.g., SELECT) in which selenoprotein expression/concentration would already have been maximised at baseline. 3. Lack of appreciation that, despite the high selenium status of SELECT men, the effects of selenium supplementation on type 2 diabetes risk were not significant. 4. Failure to understand that biomarkers of selenium status are considerably more reliable than dietary data, which we know to be much more error‐prone. 5. Frequent failure to distinguish between significant and non‐significant findings. 6. Lack of familiarity with the relevant selenium literature. 7. No mention of oesophageal or gastric cardia cancer results (although RCT results for these are not based on selenium alone) and, in relation to colorectal cancer, no mention of adenoma data. 8. In 'Implications for research', no mention is made of the need to carry out randomised controlled trials in low‐selenium populations, nor to take into consideration selenoprotein genotype, which has been shown to affect selenium metabolism. The relevance of the species of selenium administered in various trials is not mentioned.
Reply
The authors wish to thank the colleagues Doctors Brigelius‐Flohé, Combs, Davis, Green, Hesketh, Köhrle, Kristal, Rayman, Schomburg, Taylor, van den Brandt, Waters and Whanger for their detailed commentary on the selenium review.
Their comments captured some of the same concerns that we had regarding the methodological challenges associated with conducting a systematic review in the field of selenium and cancer.
In response to the commentary, we will first address concerns related to the specific setting of this review as a Cochrane review and will then respond to concerns regarding the content of the review.
We strongly agree with the concerns that it is difficult to capture all differentiations elaborated on by the review in the abstract and summary, which are limited to a certain length. Similarly, length limitations were applied to the background section. We also share the opinion that some headings in the review do not adequately reflect the content of the text that follows. For readers who have not authored Cochrane reviews themselves, we wish to explain that Cochrane reviews are submitted in an electronic format that does not allow for all adaptations authors might wish to make. The headings, for example, cannot be changed. This electronic format is optimised for reviews on intervention studies. Our review included both RCTs and epidemiological studies, and so we encountered several structural challenges throughout the review process. We hope that both the commentary of our colleagues and our experiences will contribute to the continuing work of advancing the structural processes of The Cochrane Collaboration, including the electronic software Review Manager, and to developing a more inclusive format for reviews, which encompasses epidemiological studies.
Has the condensation of information in the abstract and the plain text summary led to a distortion in the presentation of the review results?
The abstract and the plain text summary present to readers the body of evidence that was reviewed as the main results for both study questions. Our aim was to report the answers to our research questions, and although space was a limitation for the abstract and summary results sections, we have endeavoured to provide across the entire review all the best available evidence for the role of selenium in preventing cancer.
We agree with our colleagues that no studies can be found on the association of selenium with cancer in children or on the preventive efficacy of selenium supplements in children. Hence, as stated in the abstract, there is currently no convincing evidence that selenium supplementation may prevent cancer in children. However, we are completely happy not to mention children in the abstract if this may be considered misleading.
We agree with our colleagues that long‐term supplementation is more likely than short‐term supplementation to influence cancer risk, if any effect exists. The minimum of four weeks has been chosen arbitrarily. However, no consistent current agreement has indicated where to draw the line between short‐term and long‐term selenium supplementation, so any cutoff would be arbitrary to some extent. In addition, we wished to avoid making assumptions about supplementation effects in our inclusion criteria and decided rather to address the question of the effect of shorter supplementation periods in the review discussion, if any trial would have been identified.
To our knowledge, there is currently no universal recommended daily allowance for selenium intake or upper tolerable level; therefore recommending a selenium dose or level of safe intake would not be appropriate in this instance. This is clearly an area for further research, taking into account some of the potential influencing factors cited in our review (e.g., baseline levels, gender, population, source). We would like to thank the commentators for the hint to the RNI (reference nutrient intake) values for selenium in the UK, which we are happy to include in a future update of the review. Nevertheless, regarding the RNI, we would like to draw attention to the latest draft of a position paper on selenium by the Scientific Advisory Committee on Nutrition (2011), which notes “that the selenium dietary reference value was set on very limited data and could be set too high” (p74).
Dr Brigelius‐Flohé and colleagues commented that “Quoted recommendations such as 30 and 40 μg/d for men and women (WHO 2004) are no longer credible to anyone with up‐to‐date knowledge of the endpoints and biomarkers (SePP, GPx activity) that we have in 2011. There is no justification for quoting the Vinceti 2009a opinion that 20 μg/day organic selenium should be the maximum safe level.”
The suggestion of an upper safe limit of organic selenium of 20 µg/d was made by Vinceti et al. on the basis of preliminary results of the ORDET study (Vinceti 2009b), published in 2010 (Stranges 2010), and of other studies (please see for a review Vinceti 2009a). The recent availability of new data on endocrine (Lippman 2009; Stranges 2007) and dermatological (Lippman 2009) toxicity of low doses of organic selenium adds new findings supporting the recommendations by the WHO Group. We would like to draw attention to other recent studies on selenium toxicity (reviewed by Vinceti 2009a and Nogueira/Rocha 2011) and the issue of risk assessment of selenium (including the use of uncertainty factors (UF) or alternative approaches) (Aggett 2010; Douron 2010; Renwick 2006; Renwick/Walker 2008).
The diverse recommendations and the controversial discussions clearly underline the need for a systematic review in this field.
To address our research question—What evidence exists on the efficacy of selenium supplementation for cancer prevention?—we restricted our focus to RCTs with mono‐selenium supplementation. Multicomponent interventions, such as those chosen in the SU.VI.MAX, involve several nutritional/antioxidant supplements (e.g., 120 mg of ascorbic acid, 30 mg of vitamin E, 6 mg of beta carotene, 100 µg of selenium, and 20 mg of zinc in SU.VI.MAX), some of which are reportedly thought to have a potentially synergistic effect with selenium (Willett 1983); others may act as antagonists (Schrauzer/White/Schneider 1977) or may have an unknown biological interaction. Although all these factors are important considerations for the overall efficacy of selenium in the long term, we thought that inclusion of these studies in attempts to elucidate an actual anticarcinogenic role for selenium in its own right could potentially conceal the true effects (positive or negative) of selenium. By including the four studies that were mentioned in the commentary, which used multicomponent interventions, we may have gained numbers but lost out in trying to elucidate the actual effects of selenium. Therefore, these RCTs, which use selenium in combination with other nutritional factors, were outside the scope of the current review process but have been addressed in the background and discussions and could be the focus of future valuable investigations.
To avoid any potential preferential and non‐systematic selection of studies and hence results, we established a set of a priori inclusion criteria during the initial stages of the study design. These were outlined in the protocol of the review, which has been available on The Cochrane Library website and for comment since 2005.
The details of all selenium supplementation have been reported for each RCT, including the form of selenium when available, and we emphasised the importance of carefully evaluating the different biological activity and toxicity of each selenium compound. Please refer to the plain language summary: “In general there are two types of selenium supplements: one type uses the salt of selenium as the main ingredient, the other type uses organic selenium. These two types may act differently in the human body when ingested,” and in the RCTs and preventive efficacy section: “Interpretation of the results of clinical trials using selenium supplements should consider the different biological forms as well as their potential differential health effects when supplemented”; and please refer to the table Characteristics of included studies, for details on each RCT.
References are made throughout the review text to the baseline selenium status of study participants and potential interactions with study results. Please refer to Section 2.3. Adverse effects, “The RR for developing type II diabetes mellitus was higher in the participants in the upper two tertiles of plasma selenium levels, indicating a possible interaction with baseline exposure status”, for instance, or page 38 in our review: “SELECT participants had a higher selenium level at randomisation than men in the NPCT. While the mean plasma selenium concentration was 113 to 114 μg/L in the NPCT, median serum concentration was 135 to 138 μg/L in the different study arms in SELECT. Lower prostate cancer incidence in the NPCT trial was confined to men with baseline selenium levels in the lower two thirds (below 121 μg/L). Subgroup analyses of the SELECT trial are underway to investigate a possible modification by pre‐intervention selenium levels“.
Regarding the findings of NPCT and SELECT for type 2 diabetes, we would like to refer our readers to Section 2.3. Adverse effects, “A statistically non‐significant increase in diabetes mellitus type II in the selenium‐alone group (HR 1.07 (99% CI 0.94: 1.22)) was seen. An increased risk for diabetes mellitus type II was also observed in the NPCT (Stranges 2007, in: NPCT 1996). A secondary analysis of participants who did not have diabetes at start of the study revealed an excess risk in the selenium group (adjusted HR 1.55 (95% CI 1.03 to 2.33))”. We have previously outlined the section that referred to the fact that selenium baseline levels were higher in this group and would like to cite the original paper by Stranges et al. (2007), which stated: “Despite the lack of statistically significant interactions between treatment group and baseline co‐variates, the risk for type 2 diabetes was consistently higher in the selenium group within all subgroups of baseline age, sex, smoking
status, and BMI.” (p220). Regarding the issue of a potential diabetogenic effect of selenium supplements and gender, we would like to draw attention to a recent observational cohort study by Stranges (2010), which documented an excess risk of diabetes among a large cohort of women from Varese, Northern Italy. Such a diabetogenic effect of selenium is also supported by suggestive laboratory evidence, recently reviewed by Steinbrenner al. (2011).
Lippman et al. (2009) stated in their publication about the SELECT trial: “The data and safety monitoring committee had some concern over the statistically non‐significant increase in prostate cancer in the vitamin E‐alone group (P=.09 per interim data of August 1, 2008) and over a non‐significant increase in diabetes mellitus associated with selenium (P=.08 per interim data of August 1, 2008)” (p45).
The observation from SELECT (Klein 2011) that the effect diminished over time may suggest exactly the opposite to that hypothesised by Dr Brigelius‐Flohé and colleagues. A decrease in the diabetogenic effect of selenium administration over time after interruption of such administration may well indicate a decreasing adverse effect over time, as expected, of a causal association. This was what occurred in the SU.VI.MAX study, in which administration of selenium/vitamins C‐E/beta‐carotene/zinc led to an excess incidence of skin cancer, including melanoma (Hercberg 2004), which entirely disappeared after interruption of the intervention (Ezzedine 2010). The investigators interpreted such decreasing risk as an indication of the causal effect of the treatment of skin cancer and the origin of melanoma (Ezzedine 2010).
Regarding the interaction of baseline PSA levels with selenium effects in the NPCT, we would like to quote the original publication: “The protective effect of SS [selenium supplements; GD] appeared to be confined to those with a baseline PSA level of <= 4 ng/mL (0.35, 0.13–0.87), although the interaction of baseline PSA and treatment was not statistically significant“ (p608, Duffield‐Lillico 2003a). To summarise, no statistically significant interaction was noted between baseline PSA levels and prostate cancer incidence, as reported by the study authors.
Dr Brigelius‐Flohé highlighted a sentence on page 4 that might be misunderstood if taken out of its context (“risk ratios (RRs) with confidence intervals (CIs) were not calculated because of low numbers”). Our colleagues rightly stated that Hercberg et al. (2004) provided hazard ratios for cancer incidence by gender. However, the sentence our colleagues quoted from our review reads in the context as follows: “In the more recent French SU.VI.M.AX trial (Hercberg 2004), a supplementation with beta‐carotene, vitamin C, vitamin E and 100 μg selenium‐enriched yeast did not alter the incidence of cancer of the digestive tract after a median period of 7.5 years in women. In men, the incidence rate was lower in the intervention group than in the placebo group, but risk ratios (RRs) with confidence intervals (CIs) were not calculated because of low numbers”. The part of the sentence our colleagues cited about the men’s incidence rate refers to cancer of the digestive tract. Site‐specific cancer rates were not calculated or reported by gender: “We were not able to ana lyze differences in site‐specific cancers between men and women because of low statistical power” (p2340, Hercberg 2004).
Our colleagues highlighted another sentence on page 39: “Results from two randomised controlled trials (NPCT and SELECT) have failed to provide evidence that non‐melanoma skin cancer or prostate cancer can be prevented by selenium supplementation in men”. This statement refers to the primary study outcomes of both investigations, which were non‐melanoma skin cancer in NPCT and prostate cancer in SELECT, and is correct. Contrary to what was stated by Dr Brigelius‐Flohé and colleagues, the outcome measures in the NPCT were incident basal cell carcinomas and squamous cell carcinomas, and recurrent skin tumours were excluded from analysis, as summarised in the report of the primary NPCT endpoint by Duffield‐Lillico et al. (2003b). We clearly stated in our review that the NPCT was carried out among non‐melanoma skin cancer participants at baseline.
Our conclusions have been based on the available evidence, and we have highlighted the paucity of literature and data available from RCTs. Please refer to the 'Implications for research' section: “Potential differential effects of sex/gender and the use of selenium supplements in populations with a high burden of specific types of cancer diseases and differing selenium exposure levels, e.g., known low nutritional selenium intake, require further examination”.
Dr Brigelius‐Flohé and colleagues have also expressed concerns regarding our inclusion criteria for epidemiological studies and the ways results of epidemiological studies were included and presented in the systematic review.
In reply to their concern, we might have omitted three relevant studies for gastrointestinal cancers; we would like to refer them to the detailed references to both studies, Mark 2000 and Wei 2004, throughout the review. The Steevens (2010) study has not been included, as it was not available at the time of our review process and submission to The Cochrane Collaboration Group (please refer to Methods section, Search strategy). As reported in Section 1.1.6 of the review, the strength of association varied according to what was included in analyses (e.g., cardia vs non‐cardia cancers, gender), thus preventing any clear and concise conclusion to be drawn between selenium levels and upper gastrointestinal cancers in the observational summary results.
As we understood the publications Wei 2004 and Mark 2000, Wei 2004 reports on a population that was part of the population at risk in Mark 2000. Participants in Wei 2004 were the disease‐free controls for the cases of Mark 2000. Because of this overlap, we decided to report the papers jointly and put emphasis on the detailed description of both papers and their study populations (please refer to the Characteristics of included studies).
Dr Brigelius‐Flohé and colleagues criticised inclusion in the review of observational studies assessing selenium exposure as intake (e.g., with food frequency questionnaires).
Regarding the problems associated with dietary assessment, please refer to the section 'Bias and confounding': “Assessment of total selenium intake from food‐frequency questionnaires (FFQ) or interviews has proven difficult in other investigations because of the lack of food composition data which adequately reflects regional and seasonal variations in selenium concentration”. Additionally, “The FFQ overestimated the mean selenium intake in study participants when compared with laboratory analyses of duplicate meals” and ”Validity problems, possibly leading to misclassification, have also been reported when questionnaires are used to assess supplement use”.
However, studies using dietary assessment add a valuable perspective to the discussion of the relationship between selenium exposure and cancer risk. Furthermore, in addition to the literature cited by Dr Brigelius‐Flohé, other studies (van den Brandt PA et al, 1993; Longnecker et al., 1996; Haldimann et al., 1996) have reported a direct correlation between dietary and body selenium (please also see for a review of this topic Vinceti et al. 2000b and Vinceti et al. in press).
We consider the issue of selenium exposure assessment to be more complex than has been implicated by our colleagues´ comments. Assessment of selenium intake, despite the difficulties associated with its variability and possible individual variability in absorption, in some cases might even yield better estimates of actual exposure compared with biomarkers. This adds an important perspective to the discussion of why several observational studies have suggested a protective effect of higher selenium exposure towards cancer risk and others have not.
With regard to toxicity, animal studies have demonstrated that the intake of equivalent amounts of selenium, when administered in different species, might induce a stronger effect even when retained to a lesser extent (Panter et al., 1996), as shown for the inorganic compounds. The wealth of toxicological data from laboratory studies is clearly and, for obvious ethical reasons, much greater than those yielded by human studies. The same is true for studies investigating tissue distribution and biological activity of the different selenium compounds (see: Hatfield/Berry/Gladyshev 2012). We consider references to laboratory and animal studies as a necessary and valuable contribution to the understanding of selenium effects in humans.
Dr Brigelius‐Flohé and colleagues asked why our summary of the findings of the review of Ashton (2009) on the use of biomarkers for selenium measurement did not mention singular nucleotide polymorphisms (p34 in our review). We summarised the findings of Ashton 2009 that were relevant for the discussion of bias and confounding in our review. Genetic polymorphisms were not included in the analyses of heterogeneity between study results by Ashton (2009). Instead, Ashton et al. proposed singular nucleotide polymorphisms in their discussion as an area for future research and stated: “Also, for all potential biomarkers, more information is needed to understand the limitations of applicability for different population groups, the possible effects of genotype, supplementation doses, duration, baseline status, etc” (p2037S).
The criticism that we failed to distinguish between significant and non‐significant findings in epidemiological studies points to a fundamental difference in the interpretation of epidemiological study results. Indeed, we consider ‘statistical’ significance as an inappropriate approach to data analysis and interpretation with regard to observational studies, as has been long recognised (Rothman KJ 1978; Sterne/Davey Smith 2001; Greenland 2011), with no connection with ‘biological significance’. Pitfalls of statistical significance testing encompass dismissing so called `non‐significant values´ in small studies or putting undue emphasis on ‘statistically significant’ results without attempting to integrate potential biases for a study finding that would affect the estimates from that study (see: e.g., Rothman, Greenland & Lash 2008; Stang/Poole/Kuss 2010). This may lead to confusion between the validity of an investigation and its statistical stability.
Analysis and interpretation of results in biomedical research must be based on a number of considerations, comprising both study design and data analysis. We made a conscious effort in our selenium review to avoid use of an approach that dichotomised study results according to which were statistically significant and which were not. We consider this effort a major strength of our review.
We have attempted to be prudent with our conclusions by highlighting important considerations associated with the results of epidemiological studies that we reported. Both the current literature and our review indicate that although some associations have been noted between selenium levels and risk of cancer at certain body sites (e.g., prostate, bladder), more research and information are clearly required before it can be concluded that these results are “convincing” for a protective effect of selenium. The World Cancer Research Fund’s Second Expert Report (2007) also suggests the possibility of residual confounding between selenium levels and healthy lifestyles (p109).
We admit that the sentence about the marketing situation of selenium in our discussion section expresses a valuation, and we acknowledge that other colleagues might assess the marketing situation differently and as such might disagree with this sentence.
In the last part of our reply, we will address the concerns by Dr Brigelius‐Flohé and colleagues regarding the content of the background section of the review.
The reference Rodriguez 1995, which is listed in the MEDLINE database, in contrast to what our colleagues stated (please refer to PubMed ID 7605824), is an early study that investigated urinary selenium in healthy men and women and addressed the study question of the relationship between factors such as gender/sex, etc., and urinary selenium. It found gender/sex differences in urinary selenium excretion, as well as influences of health behaviours (physical activity), as stated in our background text.
We do not agree that studies investigating primarily the relationship between selenium status, thyroid volume and gland echostructure (Derumeaux 2003) or the relationship between baseline plasma selenium concentration and occurrence of dysglycaemia (Akbaraly 2010) would have been more suitable references for the statement that we made regarding gender differences.
We also would like to recapitulate the Vinceti et al. (2000a) paper because we feel that Dr Brigelius‐Flohé and colleagues misreported the methods and findings of this study. The Vinceti et al. studies in an unusual Northern Italy setting evaluated the health effects of selenium in its inorganic hexavalent form—the one usually found in underground and drinking water—together with the tetravalent species (Vinceti 2010). This study was a ‘natural experiment’, considered to be ‘the paradigm of non‐experimental epidemiologic research’, as in this type of study, ‘nature emulates the sort of experiment the investigator might have conducted, but for ethical and cost constraints’ (p94, Rothman/Greenland/Lash 2008). Study authors assessed the potential for confounding by lifestyle by assessing the socioeconomic status of exposed and unexposed cohorts, and labelling this study as a natural experiment was allowed only after the similarity of the two populations was confirmed. Dr Brigelius‐Flohé stated that Vinceti et al. admitted that their results are consistent with “no effect”, as standardised mortality ratios were generally inconsistent between men and women at most sites, and most site‐specific estimates had limited precision. The citation in the original publication reads: “The results of our study are consistent with either no effect or, particularly among the elderly, unfavourable effects of long‐term exposure to inorganic selenium on cancer mortality”. Then Vinceti et al. analysed the strengths and limitations of their study, both for the melanoma association and more generally for the effects on cancer risk. Excess melanoma risk, despite different study designs and strengths of association, has been documented to be associated with selenium exposure in a number of studies (Garland 1995; Vinceti 1998; Duffield‐Lillico 2002; Vinceti et al., in press) and has been causally associated with administration of selenium in combination with zinc and vitamins in SU.VI.MAX (Hercberg 2007). In general, we would like to propose caution when dealing with the possible selenium‐melanoma association.
In conclusion, we express our appreciation to our commentators for scrutinising our review, offering their criticisms and supporting the scientific endeavour of enclosing epidemiological as well as intervention studies in a Cochrane review. We are hopeful that the review and the commentary of our colleagues will contribute to the important and continuing discussion about the health effects of selenium and selenium supplements globally and in diverse populations.
Aggett PJ. (2010). Toxicity due to excess and deficiency. J Toxicol Environ Health A: 73(2):175‐80.
Akbaraly TN, Arnaud J, Rayman MP, Hininger‐Favier I, Roussel AM, Berr C, Fontbonne A. (2010). Plasma selenium and risk of dysglycemia in an elderly French population: results from the prospective Epidemiology of Vascular Ageing Study. Nutr Metab 18:21. Electronic publication: doi:10.1186/1743‐7075‐7‐2.
Ashton K, Hooper L, Harvey LJ, Hurst R, Casgrain A, Fairweather‐Tait SJ. (2009). Methods of assessment of selenium status in humans: a systematic review. Am J Clin Nutr 89(6):2025S‐39S.
Derumeaux H, Valeix P, Castetbon K, Bensimon M, Boutron‐Ruault MC, Arnaud J, Hercberg S. (2003). Association of selenium with thyroid volume and echostructure in 35‐ to 60‐year‐old French adults. Eur J Endocrinol 148(3):309‐15.
Douron M. (2010). U‐shaped dose‐response curves: implications for risk characterization of essential elements and other chemicals. J Toxicol Environ Health A 73(2):181‐6.
Duffield‐Lillico AJ, Reid ME, Turnbull BW, Combs GF Jr, Slate EH, Fischbach LA, et al. (2002). Baseline characteristics and the effect of selenium supplementation on cancer incidence in a randomised clinical trial: a summary report of the Nutritional Prevention of Cancer Trial. Cancer Epidemiol Biomarkers Prev 11(7):630‐9.
Duffield‐Lillico AJ, Dalkin BL, Reid ME, Turnbull BW, Slate EH, Jacobs ET, et al. (2003a). Selenium supplementation, baseline plasma selenium status and incidence of prostate cancer: an analysis of the complete treatment period of the Nutritional Prevention of Cancer Trial. BJU Int 91(7):608‐12.
Duffield‐Lillico AJ, Slate EH, Reid ME, Turnbull BW, Wilkins PA, Combs GF Jr, et al. (2003b). Selenium supplementation and secondary prevention of nonmelanoma skin cancer in a randomised trial. J Natl Cancer Inst Cancer Spectrum 95(19):1477‐81.
Ezzedine K, Latreille J, Kesse‐Guyot E, Galan P, Hercberg S, Guinot C, Malvy D. (2010). Incidence of skin cancers during 5‐year follow‐up after stopping antioxidant vitamins and mineral supplementation. Eur J Cancer 46(18):3316‐22.
Garland M, Morris JS, Stampfer MJ, Colditz GA, Spate VL, Baskett CK, et al. (1995). Prospective study of toenail selenium levels and cancer among women. J Natl Cancer Inst 87(7):497‐505.
Greenland S. (2011). Null misinterpretation in statistical testing and its impact on health risk assessment. Preventive Med 53:225–228.
Haldimann M, Venner TY, Zimmerli B. (1996). Determination of selenium in the serum of healthy Swiss adults and correlation to dietary intake. J Trace Elem Med Biol 10(1):31‐45.
Hatfield DL, Berry MJ, Gladyshev VN (Editors). (2012). Selenium. Its Molecular Biology and Role in Human Health. 3rd edition. Springer, New York, 2012, and springer ebooks: 10.1007/978‐1‐4614‐1025‐6.
Hercberg S, Galan P, Preziosi P, Bertrais S, Mennen L, Malvy D, Roussel AM, Favier A, Briançon S. (2004). The SU.VI.MAX Study: a randomised, placebo‐controlled trial of the health effects of antioxidant vitamins and minerals. Arch Intern Med 164(21):2335‐42.
Hercberg S, Ezzedine K, Guinot C, Preziosi P, Galan P, Bertrais S, Estaquio C, Briancon S, Favier A, Latreille J, Malvy D. (2007). Antioxidant supplementation increases the risk of skin cancers in women but not in men. J Nutr 137:2098‐2105.
Klein EA, Thompson IM Jr, Tangen CM, Crowley JJ, Lucia MS, Goodman PJ, Minasian LM, Ford LG, Parnes HL, Gaziano JM, Karp DD, Lieber MM, Walther PJ, Klotz L, Parsons JK, Chin JL, Darke AK, Lippman SM, Goodman GE, Meyskens FL Jr, Baker LH. (2011). Vitamin E and the risk of prostate cancer: the Selenium and Vitamin E Cancer Prevention Trial (SELECT). JAMA 306:1549‐56.
Lippman SM, Klein EA, Goodman PJ, Lucia MS, Thompson IM, Ford LG, et al. (2009). Effect of selenium and vitamin E on risk of prostate cancer and other cancers: the Selenium and Vitamin E Cancer Prevention Trial (SELECT). JAMA 301(1):39‐51.
Longnecker MP, Stram DO, Taylor PR, Levander OA, Howe M, Veillon C, et al. (1996). Use of selenium concentration in whole blood, serum, toenails, or urine as a surrogate measure of selenium intake. Epidemiology 7(4):384‐90.
Mark SD, Qiao YL, Dawsey SM, Wu YP, Katki H, Gunter EW, et al. (2000). Prospective study of serum selenium levels and incident esophageal and gastric cancers. J Natl Cancer Inst 92(21):1753‐63.
Nogueira CW, Rocha JB. (2011). Toxicology and pharmacology of selenium: emphasis on synthetic organo‐selenium compounds. Arch Toxicol 85(11):1313‐59.
Panter KE, Hartley WJ, James LF, Mayland HF, Stegelmeier BL, Kechele PO. (1996). Comparative toxicity of selenium from seleno‐DL‐methionine, sodium selenate, and Astragalus bisulcatus in pigs. Appl Toxicol 32(2):217‐23.
Renwick AG. (2006). Toxicology of micronutrients: adverse effects and uncertainty. J Nutr 136:493S‐501S.
Renwick AG, Walker R. (2008). Risk assessment of micronutrients.Toxicol Lett 180(2):123‐30.
Rothman KJ (1978). A show of confidence. N Engl J Med 299:1362‐3.
Rothman KJ, Greenland S, Lash TL. (2008). Modern Epidemiology. Philadelphia: Lippincott Williams & Wilkins.
Schrauzer GN, White DA, Schneider CJ. (1977). Cancer mortality correlation studies—IV: associations with dietary intakes and blood levels of certain trace elements, notably Se‐antagonists. Bioinorg Chem 7(1):35‐56.
Scientific Advisory Committee on Nutrition. (2011). Paper for discussion: Draft Selenium and Health position statement, 19/10/11. http://www.sacn.gov.uk/pdfs/sacn1113_sacn_selenium_and_health_position_paper.pdf (accessed on 20 December 2011).
Stang A, Poole C, Kuss O. (2010). The ongoing tyranny of statistical significance testing in biomedical research. Eur J Epidemiol 25:225‐30.
Steevens J, van den Brandt PA, Goldbohm RA, Schouten LJ. (2010). Selenium status and the risk of esophageal and gastric cancer subtypes: the Netherlands cohort study. Gastroenterology 138(5):1704‐13.
Steinbrenner H, Speckmann B, Pinto A, Sies H. (2011). High selenium intake and increased diabetes risk: experimental evidence for interplay between selenium and carbohydrate metabolism. J Clin Biochem Nutr 48(1):40‐5.
Sterne JA, Davey Smith G. (2001). Sifting the evidence—what's wrong with significance tests? BMJ 322:226‐31.
Stranges S, Marshall JR, Natarajan R, Donahue RP, Trevisan M, Combs GF, et al. (2007). Effects of long‐term selenium supplementation on the incidence of type 2 diabetes: a randomised trial. Ann Intern Med 147:217‐23.
Stranges S, Sieri S, Vinceti M, Grioni S, Guallar E, Laclaustra M, Muti P, Berrino F, Krogh V. (2010). A prospective study of dietary selenium intake and risk of type 2 diabetes. BMC Public Health 10:564. http://www.biomedcentral.com/1471‐2458/10/564.
van den Brandt PA, Goldbohm RA, vant´t Veer P, Bode P, Hermus RJ, Sturmans F. (1993). Predictors of toenail selenium levels in men and women. Cancer Epidemiol Biomarkers Prev 2(2):107‐112.
Vinceti M, Rothman KJ, Bergomi M, Borciani N, Serra L, Vivoli G. (1998). Excess melanoma incidence in a cohort exposed to high levels of environmental selenium. Cancer Epidemiol Biomarkers Prev 7(10):853‐6.
Vinceti M, Nacci G, Rocchi E, Cassinadri T, Vivoli R, Marchesi C, et al. (2000a). Mortality in a population with long‐term exposure to inorganic selenium via drinking water. J Clin Epidemiol 53(10):1062‐8.
Vinceti M, Rovesti S, Bergomi M, Vivoli G. (2000b). The epidemiology of selenium and human cancer. Tumori 86(2):105‐18.
Vinceti M, Maraldi T, Bergomi M, Malagoli C. (2009a). Risk of chronic low‐dose selenium overexposure in humans: insights from epidemiology and biochemistry. Rev Environ Health 24(3):231‐48.
Vinceti M, Stranges S, Sieri S, Grioni S, Malagoli C, Muti P, Berrino F, Krogh V. (2009b). Association between high selenium intake and subsequent increased risk of type 2 diabetes in an Italian population. ISEE 2009 Conference Abstracts Supplement; Epidemiology: 20: S47.
Vinceti M, Bonvicini F, Rothman KJ, Vescovi L, Wang F. (2010). The relation between amyotrophic lateral sclerosis and inorganic selenium in drinking water: a population‐based case‐control study. Environ Health 6:77. Electronic citation: doi 10.1186/1476‐069X‐9‐77.
Vinceti M, Crespi CM, Malagoli C, Bottecchi I, Ferrari A, Sieri S, Krogh V, Alber D, Bergomi M, Seidenari S, Pellacani G. (in press). A case‐control study of the risk of cutaneous melanoma associated with three selenium exposure indicators. Tumori.
Wei WQ, Abnet CC, Qiao YL, Dawsey SM, Dong ZW, Sun XD, et al. (2004). Prospective study of serum selenium concentrations and esophageal and gastric cardia cancer, heart disease, stroke, and total death. Am J Clin Nutri 79(1):80‐5.
Willett WC, Polk BF, Morris JS, Stampfer MJ, Pressel S, Rosner B, Taylor JO, Schneider K, Hames CG. (1983). Prediagnostic serum selenium and risk of cancer. Lancet 2(8342):130‐4.
World Cancer Research Fund International. (2007). Food, Nutrition, Physical Activity, and the Prevention of Cancer. A Global Perspective. Washington, DC: AICR.
World Health Organization (WHO). (2004). Joint FAO/WHO Expert Consultation on Human Vitamin, Mineral Requirements. (1998: Bangkok, Thailand). Vitamin and mineral requirements in human nutrition: report of a joint FAO/WHO expert consultation, Bangkok, Thailand, 21‐30 September 1998. http://whqlibdoc.who.int/publications/2004/9241546123.pdf (accessed on 23 January 2011).
Contributors
Professor Regina Brigelius‐Floh, Professor GF Combs Jr, Dr Cindy D Davis, Dr Fiona R Green, Professor John Hesketh, Professor Josef Köhrle, Dr Alan Kristal, Fred Hutchinson, Professor Margaret P Rayman, Professor Lutz Schomburg, Phil Taylor, Piet van den Brandt, Professor David J. Waters, Professor Phil Whanger.
Maree Brinkman, Gabriele Dennert and Marco Vinceti on behalf of the review authors.
Selenium for preventing cancer, 30 October 2014
Summary
Comment: Selenium for preventing cancer; The Cochrane Library 2014, Issue 3 Vinceti M, Dennert G, Crespi CM, Zwahlen M, Brinkman M, Zeegers MPA, Horneber M, D’Amico R, Del Giovane C
We are pleased to see that a revised version of the review has now been published though it has taken longer than we would have wished. In the updated review, the authors have remedied some of the shortcomings which we pointed out, but not all. I have attached detailed comments on what we think still needs to be changed and hope that these points can be remedied in the very near future.
Comments by section are given below.
Abstract 1. Selection criteria refer to including RCTs with “healthy adult participants”. However, it is clear that SELECT was the only trial that included “healthy adult participants”, all other trials included participants with a high risk of cancer (Li, Yu 1991, Yu 1997, Marshall, Algotar, Dreno) or a previous history of cancer (NPCT 2002, Reid 2008). The word “healthy” should be removed and the statement should be modified to reflect the high proportion of participants at high risk of cancer.
2. The main results of the pooled analysis of RCTs overwhelmingly reflect the results of by far the largest trial, SELECT. However, SELECT was carried out in a population of high selenium status. This needs to be mentioned either under “Main results” or under “Authors’ conclusions”. Not to mention it is to ignore a fact that is likely to be highly relevant to the outcome.
3. The “Authors’ conclusions” assert that there is “little evidence of any influence of baseline selenium status”, but that lack of evidence all relates to trials in populations of much higher baseline selenium status than the NPCT where such an effect was seen: baseline plasma Se was 114 µg/L in the NPCT compared to 126.1 µg/L in Algotar and 135.2‐138.1 µg/L in Marshall. [No such effect was seen in SELECT, but baseline selenium status was also high ‐ 136 µg/L (Kristal et al. 2014).]
Plain language summary The sentence that begins “Recent trials that were judged to be well conducted and reliable… “ should be modified to read “Recent trials that were judged to be well conducted and reliable, though conducted in high‐selenium populations, have found no effects of selenium supplementation on reducing the overall risk of cancer or on reducing the risk of particular cancers, including prostate cancer”.
Main text Page 5 column 2: We previously pointed out that having inclusion criteria that allowed RCTs of only four‐weeks’ length to be included is unjustifiable. While no studies as short as that were included, clearly a four‐week intervention with Se is insufficient to alter cancer risk so what is the justification retaining this inclusion criterion?
Page 21 column 1: We previously objected to the description of an increased risk of diabetes mellitus type 2 being found in SELECT yet such a description is there again: “An increase in diabetes mellitus type 2 was seen in the selenium‐alone group (RR 1.07, 99% CI 0.94 to 1.22)”, despite the confidence interval spanning 1. The only trial in which an increased risk of type‐2 diabetes was seen was the NPCT. The authors also refer to a short‐term effect of selenium supplementation on type‐2 diabetes risk. However, there is no mention, either here or elsewhere, of our RCT that found no increased risk of type‐2 diabetes in 500 people treated with 100, 200 or 300 µg selenium or placebo for a period of six months (Rayman et al. A randomised trial of selenium supplementation and risk of type‐2 diabetes, as assessed by plasma adiponectin. PLoS One. 2012;7:e45269).
Page 20‐21: There should have been some mention of baseline selenium status in this section. Clearly SELECT was showing evidence of toxicity, which is unsurprising given the high baseline status and substantial level of supplementation.
Page 23 column 2: In discussing the change from a protective to a possibly detrimental effect, the authors should be aware of the possibility of a threshold effect that may relate to a mechanism dependent on selenoprotein concentration/activity. Furthermore, discussing the relationship between selenium status and the risk of non‐melanoma skin cancer and type‐2 diabetes in the same breath ignores the likelihood of totally different mechanisms applying.
Page 23 column 2: The sentence “Little evidence of a beneficial effect of selenium supplementation was noted among participants with the lowest baseline selenium exposure (plasma selenium < 106 μg/L) in either the prostate cancer trial of Marshall et al. (Marshall 2011) or the prostate cancer trial of Algotar et al. (Algotar 2013), despite the fact that 45% of the participants in that study had baseline plasma selenium levels < 123 μg/L – the suggested threshold for beneficial effects of selenium supplementation according to the NPCT (NPCT 2002)”, should be qualified by pointing out that both the Marshall and Algotar trials were in men at high risk for prostate cancer and in whom prostate cancer was probably already initiated. Thus this is not an appropriate test for evidence of benefit of selenium supplementation for primary prevention in those with low selenium status.
Page 24 column 2: SNPs could be mentioned as a potential explanation of “the … unexplained heterogeneity in the reaction of participants’ plasma selenium levels to selenium supplementation”.
Page 25 column 1: As explained in our criticisms of the primary review, we and others profoundly disagree with the statement that “measurements of nutritional intake might provide better exposure estimates than do biomarkers, which may considerably mis‐classify the exposure to inorganic and organic selenium sources”. This is particularly true of exposure to selenium where food concentration data differ very considerably from one part of the world to another and many countries have no such data.
Page 28 column 1: The paragraph that contains the sentence “These ideas stimulated the largest ever cancer prevention trial, SELECT, which failed to provide support for this hypothesis, and two additional prostate cancer trials (Algotar 2013; Marshall 2011),whose results were in line with the SELECT findings in failing to find a beneficial effect of selenium”, needs to point out that SELECT, Algotar 2013 and Marshall 2011 were all carried out in high‐selenium populations and that Algotar 2013 and Marshall 2011 were both in men at high risk of prostate cancer.
Page 29 column 2: It is not especially accurate or informative to say that the Blot and Hercberg trials produced divergent results. Although they were both RCTs, they used very different designs in hugely different populations with different baseline selenium levels. It could equally fairly be said that they produced comparable results in that they both saw beneficial effects (of one sort or another).
Page 30 column 1: Karp was a secondary prevention trial in lung‐cancer patients. In relation to that trial, there should be some mention of the likely difference in mechanisms of primary prevention and those relevant to prevention of secondary tumours in already initiated patients.
Page 30 column 1: the previous RCT that found no increased risk of type‐2 diabetes in 500 people treated with 100, 200 or 300 µg selenium or placebo for a period of six months should be mentioned (Rayman et al. A randomised trial of selenium supplementation and risk of type‐2 diabetes, as assessed by plasma adiponectin. PLoS One. 2012;7:e45269).
Page 30 column 1: Under the heading, “Implications for practice”, it should be made clear that the “Results from the most recent randomised controlled trials, which were carried out in men and had a low risk of bias” were all in men of high selenium status.
Page 30 column 2: Under “Implications for research”, there is a statement that needs qualifying, “whether selenium might influence cancer risk in individuals with very low or very high baseline exposure to this element …….. have not been fully resolved, although currently available evidence from randomised trials offers little support for such hypotheses”. It needs to be acknowledged that there are no cancer trials of selenium as a single nutrient in people with low baseline selenium status.
Even if the results of SELECT are expanded to look at other endpoints, they will still not apply to low‐selenium populations and cannot compare truly low to higher levels; this also needs to be specified.
A question that remains ignored by this review, by design, is whether selenium in combination with other agents may be beneficial in cancer. This deserves some sort of comment under “Implications for research”.
Errors Page 4, column 2: Though we pointed out in our previous set of comments that SU.VI.M.AX was incorrect, it has not been corrected.
Page 6 column 2: 78.96 is described as the molecular weight of selenium; it should be atomic weight.
Page 28 column 2: we have previously pointed out that selenium supplement are not aggressively marketed to women with regard to breast cancer prevention and treatment.
Contributors (in alphabetical order): Professor Regina Brigelius‐Flohé, University of Potsdam, German Institute of Human Nutrition Professor GF Combs Jr, Grand Forks Human Nutrition Research Center, ARS/USDA, USA Dr Cindy D Davis, Office of Dietary Supplements, NIH, USA Dr Fiona R Green, Reader in Functional Genomics, University of Surrey, UK Professor John Hesketh, Institute for Cell & Molecular Biosciences, University of Newcastle, UK Professor Josef Köhrle, Charité Universitätsmedizin Berlin, Germany Dr Alan Kristal, Fred Hutchinson Cancer Research Center, Seattle, USA Professor Margaret Rayman, Faculty of Health and Medical Sciences, University of Surrey, UK Professor Lutz Schomburg, Charité Universitätsmedizin Berlin, Germany Dr Phil Taylor, Division of Cancer Epidemiology and Genetics, NCI, USA Piet van den Brandt, Professor of Epidemiology, Maastricht University, The Netherlands Professor David J. Waters, Purdue University, USA Professor Phil Whanger, Oregon State University, USA
I agree with the conflict of interest statement below:
I certify that I have no affiliations with or involvement in any organization or entity with a financial interest in the subject matter of my feedback.
Reply
21‐1‐2015
We wish to thank Dr. Brigelius‐Flohé and colleagues for their interest in our Cochrane review on selenium for preventing cancer.
Before addressing the specific points in their letter, we would like to clarify that our publication Vinceti et al. ‘Selenium for preventing cancer, Cochrane Database Syst Rev. 2014 Mar 30;3:CD005195’ was not a revised version of the previous Cochrane but rather an update, taking into account the additional three years of scientific literature on the topic, according to standard procedures of the Cochrane Collaboration.
With regard to the use of the term ‘healthy’ in RCTs, we used the term ‘healthy’ adult participants to mean that the (adult) individuals enrolled in the studies were free at the beginning of the trial from the disease representing the primary outcome, an incident cancer, as required when we deal with primary prevention trials. Being at high, low or intermediate risk of cancer, or affected by any other disease, or previously affected by another cancer, was not considered to be an exclusion criteria and did not preclude the term ‘healthy’ with respect to the trial outcome(s), which in all cases consisted of the incidence of a primary cancer. In our review, we specifically listed in detail the enrolment criteria for the trials, and before performing the meta‐analysis we excluded studies retrieved with our literature search that were not based on healthy adults (397 studies removed – see Figure 1 of our review). Being ‘totally’ healthy– i.e., apparently free from any disease and at a low risk for cancer or other chronic disease, was not a selection criteria for any of the selenium (Se) trials, including SELECT itself (for example, we used the term ‘apparently healthy men’ for the SELECT population in page 23).
As noted by Brigelius‐Flohé et al., the pooled analysis is obviously influenced by the largest trial, SELECT, and this is even more true when we limited the analysis, as recommended by the Cochrane review guidelines, to the trials at low risk of bias. SELECT has been of fundamental importance in selenium (Se) research for its large size, long follow‐up, and broad range of outcomes, all of which are important for defining the so far uncertain relation between Se and primary prevention of cancer and the adverse health effects of the metalloid. Results from SELECT, which are continuing to emerge in the literature (Kristal et al., JNCI; Martinez et al., Cancer Prev Res; Albanes et al., Cancer Prev Res 2014), in addition to other recent relevant trials (Karp et al., 2013), have been systematically confirmed by all the high‐quality, low‐bias trials so far carried out (some of which could unfortunately not be included in our review, having been published after our literature search deadline), with the exception of the excess high‐grade prostate cancer risk in the Se‐supplemented individuals with the highest baseline selenium status recently reported in SELECT (Kristal et al., JNCI 2014), an unexpected and concerning finding so far not investigated in the other trials with the partial exception of Marshall et al. (Cancer Prev Res 2011).
Assuming that the SELECT population was a group with ‘high Se status’ while NPC subjects had a low Se status, and suggesting that their different results were likely due to this, as claimed by Brigelius‐Flohé et al., is not well‐founded. Defining a low‐Se status and a high‐Se status is very subjective and debatable, but whatever approach is chosen, no such difference between these two trial populations emerges. We would argue that the more important distinctions between the two trials are that one had low risk of bias and high statistical power (SELECT), while the other one had high risk of bias and much lower power (NPC). The two trials also used different Se preparations. In fact, if we estimate Se intake though its relation with serum/plasma level computed with the rule of thumb proposed by Haldimann et al. (J Trace Elem Med Biol 1996) in the 30‐120 µg/l of plasma or serum Se, average baseline dietary exposure corresponding to their blood Se levels was around 90 µg/day for SELECT participants, and 76 for NPC subjects. If we compare these values to the Se recommended dietary intake (or comparable indexes defined as ‘recommended intake level’, ‘dietary reference value’, ‘average nutrient requirement’ etc.), both are well above these reference values for Se, whether using the 26‐34 µg/day recommended intake of the World Health Organization and Food Agriculture Organization (WHO‐FAO 2004), the 25‐35 µg/day range of the Japanese Ministry of Health Labour and Welfare (2005), the 55 µg/day of the US Institute of Medicine and Food Nutrition Board (2000), the 70 µg/day of European Food Safety Authority (EFSA 2014) or the 55 µg/day of the Italian Human Nutrition Society (SINU, Milan 2014). For a comprehensive review of this issue we refer to among other sources the Eurreca database at www.eurreca.org, Cavellaars et al., Eur J Clin Nutr 2010, Vinceti et al. 2013 Sci Total Environ, and to the EFSA journal, 2014. Thus, according to all of these standards, both the SELECT and NPC populations should be defined as having a ‘high Se status’. This would be further strengthened should we use the 110 µg/L serum Se cutpoint for Se toxicity (increased prevalence of depressive symptoms and higher levels of urinary 8‐oxo‐7,8‐dihydroguanine) recently suggested by two observational studies (Galan‐Chilet et al., Free Rad Biol Med 2014 and Conner et al., J Nutr 2014): according to such threshold values, all the RCTs included in our review including SELECT and NPC, with the exception of the Chinese ones, should be considered as carried out in populations with ‘very high’ Se status.
In our review, given the uncertainties and complexity of the issue, we consciously avoided labelling the populations in RCTs as low or high Se status, preferring instead to report baseline exposure levels and to use relative measures for their comparison (such as ‘lower status’, ‘the lowest exposure category’ instead of ‘low Se status’, and the converse for higher exposures). This was done particularly for the most influential studies in the review, the RCTs, to facilitate assessment of whether baseline Se exposure may influence the response to Se supplementation in terms of cancer risk and comparison of distributions of baseline Se exposure. We refer Brigelius‐Flohé et al. to our analysis in the review (pages 22/24), which found the following points, among others: the marginal difference in intake of around 15 µg/day between the SELECT and NPC populations, in contrast with usual differences of Se intake at the population and the individual level, which span hundreds of micrograms; the occurrence of adverse effects even in the trials with the lowest baseline exposure level, such as the increased incidence of skin cancer in NPC and of type 2 diabetes in all trials which so far investigated this outcome; and the considerable overlap of Se exposure levels between the various RCTs. Finally, in our review we had to state that ‘analyses stratified by baseline Se status are not available for SELECT: Such analyses would greatly help to elucidate this issue.’ Fortunately, such evaluation (though so far only for prostate cancer) has been subsequently published (Kristal et al., JNCI 2014, and specifically its table 4). As it happens, their finding based on quintiles of baseline Se exposure is consistent with our previous assessment. In fact, the abstract of that paper reported that ‘Se supplementation did not benefit men with low Se status but increased the risk of high‐grade PCa among men with high Se status.’
When programming the update of this Cochrane review, we decided not to further restrict the inclusion criteria for studies compared with the 2011 review, but rather to relax them somewhat. For example, meta‐analysis was carried out for site‐specific cancer types when only 2 randomised trials were available. We even discussed whether to include trials reported only as abstracts and not in extenso, but decided against this due to lack of consensus, even though this precluded consideration of at least two possible relevant RCTs, the Karp et al. trial for prevention of second primary tumours in patients with resected lung cancer (Karp et al., J Clin Oncol 2010) and a trial on the risk of cancer in BRCA1 carriers (Lubinski et al., Hered Cancer Clin Pract 2011). We agree with Dr. Brigelius‐Flohé et al. that a trial with only 4 weeks of supplementation would be very unsatisfactory, even in case of ‘mega‐dose’ Se administration, and such a dosage scheme would not have passed un‐remarked upon in our literature review, had we found such a study.
As far as Brigelius‐Flohé et al. comments about the excess diabetes incidence in SELECT among subjects allocated to Se administration, we are surprised to see this objection: reporting and commenting on the adverse effects of RCTs is mandatory according to the Cochrane Handbook for Systematic Reviews of Interventions (Higgins JPT and Green S, Chapter 4 ‘Adverse effects’) and more generally according to ethical and scientific issues. We also note that Brigelius‐Flohé et al. when commenting on the excess diabetes incidence rely entirely on statistical significance testing (‘..despite the confidence interval spanning 1’), an approach generally considered to be inappropriate for evaluating findings from epidemiologic studies (Sterne and Davey Smith, BMJ 2001; Rothman, Greenland and Lash, Modern Epidemiology 2008; Stang, Poole and Kuss, Eur J Epidemiol 2010), especially for adverse effects that the studies were not necessarily powered to detect. The excess diabetes risk was one of the concerning findings yielded by SELECT (Vinceti et al., Rev Environ Health 2009), mirroring the observation of an increased diabetes incidence detected in the previous NPC trial (Stranges et al., 2007). We also noted that such excess risk was found in all four RCTs that investigated this outcome, and this was also supported by some biological plausibility, though we did not carry out an in‐depth investigation of the diabetes & Se relation, for which we refer to recent literature (Steinbrenner 2013; Vinceti et al., J Trace Elem Med Biol 2015). Contrary to the claims of Brigelius‐Flohé et al., we did not mention the 2012 Rayman et al. trial published in PLoS One for the obvious reason that it did not include cancer nor diabetes among the outcomes under investigation.
The comment by Brigelius‐Flohé et al. stating that ‘Clearly SELECT was showing evidence of toxicity, which is unsurprising given the high baseline Se status and substantial level of supplementation’ is also unfounded. Being ourselves among the few investigators who have systematically reviewed the human health risks of chronic low‐dose Se overexposure, (Vinceti et al., Rev Environ Health 2001 and 2009; Vinceti et al., Sci Total Environ 2013; Vinceti et al., Toxicol Lett 2014), we must point out that the upper limit of ‘safe’ Se exposure was and is set at a higher level than that of the SELECT study groups allocated to Se administration, i.e. at 400 µg/day (US Institute of Medicine 2000; World Health Organization Food Agriculture Organization 2004, and the Office of Dietary Supplements of the National Institute of Health accessed at ods.od.nih.gov/factsheets/Selenium‐HealthProfessional/ on January 20, 2015).
Brigelius‐Flohè et al. challenge discussing the excess risk of diabetes and of non‐melanoma risk cancer ‘in the same breath’ since this would ‘ignore the likelihood of totally different mechanisms’. This misrepresents the review, which makes no claim that risk of non‐melanoma skin cancer and diabetes operate through the same mechanisms.
Brigelius‐Flohè et al. state that the participants in the Marshall et al. and Algotar et al. studies were at high risk for prostate cancer (as we mentioned in our review) and that prostate cancer was probably already initiated in them. The participants in these trials were biopsy‐negative for prostate cancer, and therefore the latter statement by Brigelius‐Flohè et al. is speculation not supported by the available evidence. Contrary to the claims of Brigelius‐Flohè et al., the Marshall et al. 2011 trial and the Algotar et al. 2013 trial were important, not only since they confirmed key results of SELECT trial, but also since they addressed the issue of influence of baseline Se status on the effect of Se supplementation on (prostate) cancer risk. We refer Brigelius‐Flohè et al. to pages 23 and 24 of our review where we analysed this issue in‐depth, and specifically to the following text: “Little evidence of a beneficial effect of Se supplementation was noted among participants with the lowest baseline Se exposure (plasma Se < 106 μg/L) in either the prostate cancer trial of Marshall et al. (Marshall 2011) or the prostate cancer trial of Algotar et al. (Algotar 2013), despite the fact that 45% of the participants in that study had baseline plasma Se levels < 123 μg/L‐the suggested threshold for beneficial effects of Se supplementation according to the NPCT (NPCT 2002)”. In addition, as previously mentioned, a 2014 report published after final submission of our review showed that SELECT subjects in the lowest baseline status categories did not benefit from Se supplementation with regard to (prostate) cancer risk, though they did not experience the increased risk of high‐grade prostate cancer induced by the Se supplementation observed in the highest exposure groups (Kristal et al. JNCI 2014).
We did not mention SNPs as a potential explanation of “the ... unexplained heterogeneity in the reaction of participants” since we were specifically reporting the comments of Ashton et al. Am J Clin Nutr 2009, who did not primarily focus on this possibility. However, as Brigelius‐Flohè et al. may note from several statements within our review, we agree about the potential importance of SNPs, and this is why we frequently mention the potential role of genetic factors in our review.
Page 25, column 1 (assessment of Se exposure): though we could not review in‐depth all studies concerning methods for assessing Se exposure and related issues, we wanted to mention the human studies finding an association between dietary and biomarker Se, those unable to find it, and the advantages and limitations of all these approaches. We refer Brigelius‐Flohè et al. to specific reviews or research papers on this important issue, which show that inadequate Se exposure classification made on the basis of dietary intake or of hair, blood, urine and toenail levels may have had a major role in the inconsistencies among various observational studies and between the observational and the experimental investigations. We stand behind the brief statement in our review concerning Se exposure assessment methods in the human body.
Contrary to the claims of Brigelius‐Flohé et al., the Blot and Hercberg trials indeed produced divergent results, and the statement about these two trials that ‘both saw beneficial effects’ is untrue. Though the effects of these trials administering (different) mixtures of vitamins and minerals and carried out in very different populations cannot be adequately summarized in few words, it can be easily appreciated that the Chinese trial found beneficial effects on decreased mortality, mainly due to reduced cancer rates (especially for stomach cancer) (Blot et al., JNCI 1993 and Am J Clin Nutr 1995) while the second trial found beneficial, null and adverse effects of supplementation overall as well as specifically for cancer (Hercberg et al., Arch Intern Med 2004 and Br J Nutr 2006). Among the adverse effects following supplementation, Hercberg et al. found an alteration of the lipid profile (Hercberg et al., Lipids 2005) and an increase in melanoma incidence (Hercberg et al., J Nutr 2007), later shown to decrease during the post‐intervention follow‐up, further supporting a causative role of the treatment (Ezzedine et al., Eur J Cancer 2010). However, since these two trials did not include an intervention arm receiving Se alone, they were excluded from our meta‐analysis as were all trials that administered Se together with other substances. They were included in a different Cochrane review (Bjelakovic et al. Cochrane Database Syst Rev 2012).
Page 30, column 1: contrary to the statements of Brigelius‐Flohé et al., the Karp et al. trial, published in extenso in J Clin Oncol 2014, was not a secondary prevention trial, but a primary prevention trial, as we indicated in our review. As literally abstracted from the Karp paper, study objectives were “to evaluate the efficacy of Se supplementation in reducing the incidence of lung second primary tumors in patients who had been treated for stage I non–small‐cell lung cancer; to evaluate the qualitative and quantitative toxicity of daily Se supplementation; and to compare the incidence of specific cancers, mortality from cancer, and overall survival of patients treated with Se supplementation versus placebo”. The study population was therefore comparable to that of the NPC trial in the sense that both included participants with a recent history of cancer: the first trial comprised 1561 individuals who had been treated for stage I non–small‐cell lung cancer with complete surgical resection, while the second RCT included 1312 individuals with a history of two or more basal cell carcinomas or one squamous cell carcinoma of the skin, with one of these occurring within the year prior to randomization. We note that the results of the low‐bias Karp et al. trial, which could not be meta‐analysed in our review having been published in extenso beyond the literature search deadline, were fully consistent with the conclusions of our review.
Brigelius‐Flohé et al. state that ‘A question that remains ignored by this review, by design, is whether Se in combination with other agents may be beneficial in cancer’. As they correctly recognize, this was not included among the objectives of our review. However, we agree with Brigelius‐Flohé et al. concerning the use of selenium compounds in cancer therapy warranting strong attention and in‐depth investigation, as stated in our section ‘Se as a potential cancer therapeutic agent’ in Vinceti et al., J Environ Sci Health C Environ Carcinog Ecotoxicol Rev 2013. However, caution must be used when addressing this issue, also due to the concerning results of a recent study in patients affected by nonmetastatic prostate cancer, where supplementation of ≥140 µg/day Se was found to be associated with excess mortality from prostate cancer (Kenfield et al., JNCI 2015).
We wish to thank Brigelius‐Flohé et al. for their search for typos and mistakes in our 193 page review. They claim that three errors were found; however, these were not errors. The acronym SU.VI.M.AX was sometimes used by the authors of that trial, and we used it in our review only when citing a reference titled with that form of the acronym (Arnaud et al., J Trace Elem Med Biol 2007), while we used the more common ‘SU.VI.MAX’ for the remaining papers. As far as the 78.96 ‘molecular weight’ of Se is concerned, we recognize that the adjective ‘atomic’ is more commonly used than ‘molecular’, but the latter may also be used in connection with ‘weight’ for Se, as it may be observed at the PubChem Open Chemistry database of the US National Institute of Health (http://pubchem.ncbi.nlm.nih.gov/compound/Se, accessed January 20, 2015) or the US Center for Disease Control and Prevention ‐ National Institute for Occupational Safety and Health website (http://www.cdc.gov/niosh/docs/81‐123/pdfs/0550.pdf, accessed January 20, 2015). Finally, aggressive marketing of Se supplements for breast cancer can be detected through a simple Google Internet search. Admittedly, this is also true for other cancers, including of course prostate cancer, and more generally for chronic disease or conditions claimed to be due to oxidative stress and alleged to be prevented by Se. However, such marketing approaches differed depending on the diseases, populations, sources of information, strategies, and periods involved, and were not analysed because they were outside the scope of our current review.
Contributors
Marco Vinceti, Gabriele Dennert, Catherine M Crespi, Marcel Zwahlen, Maree Brinkman, Maurice PA Zeegers, Markus Horneber, Roberto D’Amico, Cinzia Del Giovane
Further discussion on 'Selenium for preventing cancer'
Summary
We are pleased with your positive response to our concerns and the expressed willingness of the review authors to make changes as appropriate. In particular, we welcome the following proposed modifications.
A more accurate (and longer) abstract and plain language summary to take account of the concerns we specified in our letter and in the first of our “General criticisms”.
Modification of the review by ensuring that differences in baseline selenium exposure between trials are clarified and placed in the proper context.
More careful use of language in relation to statistical significance, as, for instance, in the two examples you cite in your letter. The preferred form you quote is much better than the misleading use of “lower” or “higher” for “non‐significant” effects, as occurred frequently in the review.
Removal of constraints on the use of section headings so that more appropriate headings can be used.
There is little point in revisiting all of our criticisms as they were clearly set out in our original letter and document, and most still stand. We would like to see the review amended as soon as possible to take account of those criticisms and specifically to correct the inaccuracies that we have noted. The review authors have replied with a number of points that we would like to challenge.
p2: Re the suggestion of an upper safe limit of organic selenium of 20 μg/d by Vinceti et al., the authors now justify the original inclusion of that statement on the basis of a study (ORDET) based on a semiquantitative FFQ at baseline and follow‐up for development of type 2 diabetes 16 years later. Based on that same study (p4), the authors refer to “Such a diabetogenic effect of selenium…”. A prospective study, especially one with a very weak study design such as ORDET, can only show an association—hardly a good basis for making such a statement in a Cochrane review. Furthermore, an upper safe limit of organic selenium of 20 μg/d would be just above that at which Keshan disease is seen—11 μg/d in a Chinese man, which translates to 14 μg/d in a man of Western body weight.[1]
p2: The authors say, “The recent availability of new data about endocrine (Lippman 2009; Stranges 2007) and dermatologic (Lippman 2009) toxicity of low doses of organic selenium adds new findings which support the recommendations by the WHO group.” The authors seem still not to have taken on board the fact that Lippman et al. 2009 doesnot show any endocrine toxicity of selenium. Furthermore, the dose given—200 μg/d—was not low.
p4: Diminution of the effect on type 2 diabetes over time. Proper interpretation of SELECT is that there was a null result during the trial (RR 1.07, P value 0.16) and a similarly null result with postintervention follow‐up time included (RR 1.04, P value 0.34). If trial‐only data versus post‐trial‐only data were compared, it is probably unlikely that there would be any difference statistically. However, we do understand the point the review authors make: Interpretation depends on how one thinks selenium acts. If we were talking about an effect that occurred immediately after starting a drug (e.g. platelet effect of aspirin, blood pressure reduction from antihypertensive) and stopped more or less immediately after cessation of the drug, then the review authors’ interpretation would have better credibility.
In contrast to the week or so that the effect of aspirin on platelets lasts, selenomethionine has a long half‐life of 252 d [363 d (turnover time) × 0.693 (from kinetic modelling)] (Swanson et al. AJCN 1991, 54:917‐26). In medicine, when calculating dosing intervals for drugs, it is typical to give doses every five to six half‐lives. When first‐order kinetics is applied, five half‐lives for total body selenium is 1260 days (3.45 years), and six half‐lives is 1512 days (4.14 years). Although it is true that the amount of the original dose still remaining is small after five (6.25%) or six (3.13%) half‐lives, excess residual selenium remains from the supplementation. So, on the basis of both observed effects with cancer and pharmacokinetic data, the events that occurred in the post‐trial period for SELECT participants (34 additional months) should still be considered a period of selenium exposure and therefore incompatible with the review authors’ hypothesis.
p6: We hotly dispute the assertion of the review authors (none of whom is a nutritionist) that “The assessment of selenium intake, despite the difficulties associated to its variability and the possible individual variability in absorption, in some cases might even yield better estimates of actual exposure compared with biomarkers”.
p7: Gender differences: The Schomburg references would have been preferable; Schomburg is the accepted authority in this area.
We very much hope that our original comments and those contained in this letter will help the review authors, guided by the editors, to revise the review, so that it sits more comfortably with the opinion of experienced investigators in the selenium‐cancer field.
Yours sincerely,
Professor Regina Brigelius‐Flohé, University of Potsdam, German Institute of Human Nutrition Professor GF Combs Jr, Grand Forks Human Nutrition Research Center, ARS/USDA, USA Dr Cindy D Davis, Office of Dietary Supplements, NIH, USA Dr Fiona R Green, Reader in Functional Genomics, University of Surrey, UK Professor John Hesketh, Institute for Cell & Molecular Biosciences, University of Newcastle, UK Professor Josef Köhrle, Charité Universitätsmedizin Berlin, Germany Dr Alan Kristal, Fred Hutchinson Cancer Research Center, Seattle, USA Professor Margaret P Rayman, Faculty of Health and Medical Sciences, University of Surrey, UK Professor Lutz Schomburg, Charité Universitätsmedizin Berlin, Germany Dr Phil Taylor, Division of Cancer Epidemiology and Genetics, NCI, USA Professor Piet van den Brandt, Department of Epidemiology, Maastricht University, The Netherlands Professor David J Waters, Purdue University, USA Professor Phil Whanger, Oregon State University, USA [1] National Academy of Sciences, Institute of Medicine’s Food and Nutrition Board, Dietary Reference Intakes for Vitamin C, Vitamin E, Selenium and Carotenoids. http://fnic.nal.usda.gov/nal_display/index.php?info_center=4&tax_level=4&tax_subject=256&topic_id=1342&level3_id=5141&level4_id=10591.
Reply
We would like to thank Drs Brigelius‐Flohé and colleagues for their continuing interest in our research activity on selenium.
We decided to shortly respond to some of their discussion points (citations from Dr Brigelius‐Flohé et al are provided in italics):
“more careful use of language in relation to statistical significance as, for instance, in the two examples you cite in your letter. The preferred form you quote is much better than the misleading use of “lower” or “higher” for “non‐significant” effects as occurred frequently in the review”
Dr Brigelius‐Flohé and colleagues do not acknowledge the limitations of their approach based on ‘statistical significance’ (please refer to the references provided in our previous reply). Their approach appears to have had major consequences for a number of considerations and statements in their two letters. It is of interest to note that even the SELECT “Data and Safety Monitoring Committee” expressed its concern “over a non‐significant increase in diabetes mellitus associated with selenium (P = 0.08 per interim data of August 1, 2008)” (cited from Lippman et al., JAMA 2009), which we consider a very correct approach given the decision‐making responsibility of such a Committee.
“The authors have replied with a number of points that we would like to challenge"
p2: "Re the suggestion of an upper safe limit of organic selenium of 20 μg/d by Vinceti et al., the authors now justify the original inclusion of that statement on the basis of a study (ORDET) based on a semi‐quantitative FFQ at baseline and follow‐up for development of type‐2 diabetes 16 years later. Based on that same study (p4), the authors refer to “Such a diabetogenic effect of selenium….”. A prospective study, especially one with a very weak study design such as ORDET, can only show an association—hardly a good basis for making such a statement in a Cochrane review. Furthermore, an upper safe limit of organic selenium of 20 μg/d would be just above that at which Keshan Disease is seen—11 mg/d in a Chinese man, which translates to 14 μg/d in a man of Western body weight.
As written in our original response, the suggestion of a safe upper limit of 20 µg/L was based on the ORDET study results already available and published as an abstract in Epidemiology in 2009. Stating that the ORDET study, one of the first and most methodologically sound European prospective studies, started in the 1980s by the Italian National Cancer Institute in Milan, was ‘weak’ is unacceptable. Its methodological value has been largely recognised in the scientific community and in the epidemiological literature.
Our review, however, never aimed at summarising the large epidemiological and laboratory literature addressing the issue of safe upper limit of Se exposure in humans, particularly the most recent studies.
p2: The authors say, “The recent availability of new data about endocrine (Stranges 2007; Lippman 2009) and dermatologic (Lippman 2009) toxicity of low doses of organic selenium adds new findings which support the recommendations by the WHO group.” The authors seem still not to have taken on board the fact that Lippman et al. 2009 shows no endocrine toxicity of selenium. Furthermore, the dose given—200 mg/d—was not low.
The relation between selenium and excess diabetes risk is an extremely important issue that clearly would require extensive review, but this was not the aim of our Cochrane review;therefore we would like to refer Dr Brigelius‐Flohé and colleagues to the most recent studies and reviews on the topic. It would also be useful to remind Dr Brigelius‐Flohé and colleagues that the SELECT trial found an excess risk of diabetes, which understandably caused concern for its “Data and safety monitoring Committee” (see above) and contributed to the anticipated ending of the trial. We took note that Dr Brigelius‐Flohé and colleagues do not consider the SELECT supplemental dose of 200 mg/Se/d to be a ‘low’ dose; actually, it was so high that it could be toxic.
p6: "We hotly dispute the assertion of the authors (none of whom is a nutritionist) that “The assessment of selenium intake, despite the difficulties associated to its variability and the possible individual variability in absorption, in some cases might even yield better estimates of actual exposure compared with biomarkers”.
Different exposure assessment methods have different advantages and disadvantages. What we stated in our review was, “A concern, which we cannot clarify to date, is that biomarkers do not adequately reflect intake of both organic and inorganic selenium species”. We still think there is currently no way of clarifying this.
We were very surprised in reading comments such as ‘None of the authors is a nutritionist’, not just because this is incorrect (one of the review authors, MB, is an accredited and practicing dietician and nutritionist), but also for the underlying and clearly ‘biased’ concept: that the right to conduct independent research should be determined by subjective value judgements by one’s peers.
Despite the detailed comments made by Dr Brigelius‐Flohé et al regarding key statements we have made and details of the studies we have identified in preparing the review, we remain convinced that the conclusions drawn from the original version of the review remain valid: We have not demonstrated a protective effect of selenium against cancer in men, women or children.
Contributors
Marco Vinceti, Maree Brinkman, Gabriele Dennert and Marcel Zwahlen on behalf of the review authors.
Selenium for preventing cancer, October 2018
Summary
Comment: John Endicott Consumer Peer Review of January 29, 2018 Cochrane Review Selenium for Preventing Cancer Authors: Vinceti M et al.
(1) “Suboptimal systematic reviews and meta‐analyses can be harmful given the major prestige and influence these types of studies have acquired.” John Ioannidis, The Mass Production of Redundant, Misleading, and Conflicted Systematic Reviews and Meta‐analyses, Milbank Quarterly, Vol. 94, No. 3, 2016 (emphasis added)
(2) “Indeed, research may find we would be better off to scrap peer review entirely. The readers … will continue to be the final and harshest judges.” Drummond Rennie, Guarding the Guardians: A Conference on Editorial Peer Review, JAMA, November 7, 1986
(3) “What a Load of Rubbish. Scientists have lost their taste for self‐policing and quality control. … The hallowed process of peer review is not all it is cracked up to be, either. … Peer review should be tightened—or perhaps dispensed with altogether, in favour of post‐publication evaluation in the form of appended comments.” How Science Goes Wrong, The Economist, October 19, 2013
I am posting this as an unsolicited consumer peer review Comment on the above extremely flawed Cochrane Review. I am a ‘consumer’ of one of the “selenium” supplements discussed in this Cochrane review, and am also a ‘reader’ of a number of relevant RCTs and interpretive standards which Cochrane peer reviewers have unaccountably overlooked. Accordingly, I am also a ‘harsh judge’ of the failure of Cochrane peer review in this instance. I do believe that “following the recommendations of [this 2018 Cochrane Review, ‘Selenium for Preventing Cancer’] could result in harm to patients or populations of interest”—within the meaning of Cochrane’s own publishing policy “Process in the event of serious errors in published Cochrane Reviews.” (emphasis added).
Specifically, I agree with Dr. Walter Willett of Harvard University who this month (March 13, 2018) in the Journal of the American Medical Association writes that “study findings can be buried in a poorly planned meta‐analysis”, and I also agree with Dr. Michael Bracken of Yale University, a long‐time Cochrane collaborator, when he says that he is “not going to defend” the meta‐analysis of cancer mortality effects appearing on page 160 of this 2018 Cochrane review. The flawed meta‐analysis on page 160 has, for the time being, “buried” an astounding finding of a 41% reduction in cancer mortality in study participants (confidence interval 0.42‐0.89, P value .008).
As Cochrane co‐founder Iain Chalmers pointed out only two years ago, “there is a vast potential gain from salvage operations [including] rescuing sunken trials from the bottom of the ocean … “ (1). Read on, and see if you do not agree that the cancer mortality RCT in question—styled NPCT 2002 in this Cochrane 2018 “selenium” review—must be unburied from the watery grave to which Cochrane authors Vinceti et al. have seen fit to consign it.
I. To begin at the beginning: Item 1 on the March 2017 “Consumer peer‐review form for a Cochrane intervention review” states “we do not expect you to comment on the title [but] please do so”, if you can “suggest an improvement.” Well, despite being warned off, I’m going to start by commenting on the review’s title, and will suggest an improvement.
Iain Chalmers, several years before he co‐founded the Cochrane Collaboration in 1992, wrote a letter to the British Medical Journal, “Proposal to outlaw the term ‘negative trial’” (2). In a similar vein, I would put it to you that Cochrane should have outlawed the term “selenium” from appearing anywhere in this review, insisting instead that the authors substitute—for each of their hundreds upon hundreds of uses of “selenium” (starting of course with the first word of the review’s title)—the proper term: “selenium compounds”. The term “selenium” means only one thing: elemental Se, which has never been employed as an intervention in any of the 388 studies referred to by Vinceti et al. in this 2018 version of their Cochrane review. (There was an earlier 2014 version of the review by the same authors).
Advice to Cochrane: Always insist that authors employ correct terminology for an intervention—otherwise, as in this case, the consequences to consumers can be dire. Harvard’s most noted living biologist E.O. Wilson insists on this point: “A great deal of the future of biology [and medicine, too] depends on the strengthening of taxonomy, for if you cannot tell one kind of plant [or therapeutic intervention] from another, you’re in trouble. Some kinds of research may be held up indefinitely [as has been the case with a replication trial here, see infra]. As the Chinese say, the beginning of wisdom is getting things by their right names.” (3).
Next point: the most important health outcome discussed in this 2018 review is cancer mortality. In “a recent popular survey in which people were asked how they would choose to die … 0% chose cancer.” (4).
Item 9 on the 2017 Cochrane consumer peer review form asks, “Are the most important outcomes to you listed in the ‘Summary of Findings’ table?” This question can only be answered “No”, since cancer mortality appears nowhere among the several outcomes listed in the Summary of Findings on pages 4‐5 of the review.
You must go all the way to page 160 of the review to locate the “poorly planned” (Walter Willett’s term, supra) meta‐analysis for cancer mortality, representing Vinceti et al.’s attempt to “bury” the NPCT results which Willett long ago described as “more important than anything else we know about in cancer prevention.” (5).
Dr. Michael Bracken, professor emeritus of epidemiology at Yale School of Public Health, has reacted as follows to Cochrane’s meta‐analysis (pooling results of NPCT 2002 and SELECT 2009, the only two studies which have reported total cancer mortality results for selenocompound intervention):
“I am not going to defend the [Cochrane] meta‐analysis. An [I2 value] of >75% is usually regarded as an indication that the meta‐analysis is inappropriate. Instead of calculating a summary risk estimate which is essentially meaningless in the presence of high heterogeneity, my preference would be to examine the two trials in detail to see why they give statistically different results (the confidence interval of each trial excludes the point estimate of the other).” (emphasis added) (6).
(The word “meaningless”, used by Professor Bracken, comes from section 9.5.1, “What is heterogeneity?”, in the Cochrane Handbook, q.v. Professor Bracken is listed in the acknowledgments for his contribution to this chapter of the Handbook).
Cochrane’s 2017 “peer review checklist” asks the question, “Have sources of heterogeneity been identified?” If the designated peer reviewer(s) did in fact answer “Yes” to this item on the checklist, there is only one brief, backhanded and wrongheaded reference in the entire 236‐page review to support such a “Yes” answer, as follows:
“The turning point of research on selenium and cancer was the SELECT trial (SELECT 2009) … . The intervention used in this trial was different from that used in NPCT (selenomethionine in SELECT, and selenised yeast in [NPCT]), although this is unlikely to have been responsible for observed differences (Waters 2013); in both cases, the intervention comprised organic selenium species (Block 2004).” Cochrane 2018 “Selenium … ” Review, at p. 31.
Block 2004, one of the dozens of speciation analyses of selenised yeast (SeY) published to date, refers to a number of selenium compounds other than selenomethionine (SeMet) which can be found in the nutritional supplement SeY (obtainable over the counter in any pharmacy). As analytic techniques have improved over the years, over 100 seleno compounds have now been detected in SeY. (7).
What is wanted—to support the Cochrane meta‐analysis combining the extremely dichotomous results of these two studies of extremely heterogeneous interventions—is a head‐to‐head trial of SeY and SeMet showing that they are bioequivalent with respect to the clinical outcome of total cancer mortality.
Waters 2013, in the version cited by Vinceti et al. is unlocatable (even in abstract form) on the Internet, but is clearly identical to the November 2012 report by Waters, Shen, et al. in the journal Nutrients, describing a head‐to‐head (biomarker) study of the effects of SeY and SeMet—in 49 elderly beagles. The Waters study proved unable to support “the possibility that SeMet and Se‐yeast are not equipotent in promoting … cancer risk reduction in the aging prostate [of beagles].” (8).
However, the 2018 Cochrane “selenium” review unaccountably overlooks a second head‐to‐head study comparing the effects of SeY and SeMet in humans. This 2014 study showed “reductions in [‘prostate cancer relevant’] biomarkers of oxidative stress following supplementation with SeY but not [with] SeMet in healthy men.” (emphases added). Comparative effects of two different forms of selenium on oxidative stress biomarkers in healthy men: a randomised clinical trial. (9). The 2018 Cochrane Review authors claim, on page 8, to have included clinicaltrials.gov in their literature search. Had they actually done so, the Richie trial would not have been hard to locate. The report of the Richie trial is among only 17 trials—all human, of course—which show up following a search for trials of “selenium” and “prostate cancer.” The 2018 Cochrane Review’s passing over (or suppression) of the Richie human head‐to‐head study showing cancer biomarker effectiveness of SeY, versus ineffectiveness of SeMet, calls to mind this observation by Ben Goldacre, in his 2012 book Bad Pharma:
“We proceed by testing things … in head‐to‐head trials and gathering together all of the evidence. This last step is crucial: if I withhold half the data from you, it’s very easy for me to convince you of something that is not true. … [S]o every time time we fail to publish [or cite] a piece of research, we expose real, living people to unnecessary, avoidable suffering [including death].” (emphasis in original) (10).
Citing a null beagle study, while at the same time failing to cite a positive human study, is a perfect example of Goldacre’s point, and is also a perfect example of John Ioannidis’ larger point, supra, which I will quote again:
“Suboptimal systematic reviews and meta‐analyses can be harmful given the major prestige and influence these types of studies have acquired.” (emphasis added).
And while I’m criticizing the external peer reviewer(s) for not catching up Vinceti et al. for their failure to credibly explain the “sources of heterogeneity” in their highly heterogeneous meta‐analysis, I feel I should call out as well the internal Cochrane peer reviewers, who appear to include: (1) an Information Specialist; (2) a Cochrane Review Group Advisor; (3) a Contact Editor; and (4) a Sign‐Off Editor.
Advice to Cochrane: At least one—and probably all—of the above internal Cochrane peer reviewers should, as a matter of the most basic auditing of a Cochrane review, have read in its entirety at least one cited study for the major outcome, total cancer mortality in this case. Here, that one study report would obviously have been the $150 million RCT, Effect of selenium and vitamin E on risk of prostate cancer and other cancers: the Selenium and Vitamin E Cancer Prevention Trial (SELECT). (11).
And in reading this SELECT report, it would be a good idea to have in mind a question such as: “I wonder if this study has anything to say about whether it considers its intervention, SeMet, to be bioequivalent to SeY, the intervention in NPCT?”
Had the Cochrane internal editors performed this uber‐simple spot check, they would have found that the 33 authors of the SELECT report explicitly disclaimed any notion that their study of SeMet could be viewed as refuting in any degree the results of the NPCT study of SeY:
“[T]he formulation (high‐selenium yeast) given in the NPCT trial may have been more active than the l‐selenomethionine given in SELECT … . [I]t is impossible to know now whether selenized yeast would have been more active than l‐selenomethionine … in SELECT.” (12).
I would put it to the Cochrane editorial team that these clear statements joined in by the 33 SELECT authors are, if read and absorbed, plain red flags that should have foreclosed Vinceti et al.’s misguided pooling of the results of these two RCTs in the Cochrane total cancer mortality meta‐analysis on page 160 of the 2018 review.
To sum up: SELECT is a “fair test” of SeMet. NPC is a “fair test” of SeY. And, as the SELECT authors make pellucidly clear to anyone paying attention, SELECT can in no way be viewed as a “fair test” of SeY—and so, SeY is very much still ‘in play’ as a potentially very effective chemopreventive agent, no reason at all not to perform a true replication trial, stat. But do not hold your breath waiting for anyone to step up to fund such a replication trial, at least not so long as the fatally flawed 2018 Cochrane “selenium” review is not withdrawn per Cochrane’s own “Process in the event of serious errors in published Cochrane Reviews.”
Of course, Cochrane can always first go back to Vinceti et al. and ask, “By any chance, can you cite some new evidence—better than your Waters 2013 elderly beagle study whose value (assuming it had any value in the first place) has been obliterated by the Richie 2014 human study—which could support your pooling of SELECT and NPCT cancer mortality results? If the answer is ‘Yes, we can’, by all means let’s look at this newly‐found evidence.” I strongly suspect there is none, but no harm in asking for some real “evidence”, as Iain Chalmers did in 1989 at the first JAMA/BMJ conference on peer review:
“The inaugural Peer Review Congress was held in a distinctly shabby hotel in Chicago, Illinois, in 1989. It was engaging and contentious: presenters studied the demography of reviewers at various journals, how often individuals conducted reviews, blinding, statistical reporting and much more. I was thrilled to see actual data.
“A distinguished editor in the audience took another view, excoriating presentation after presentation. Finally, Iain Chalmers (who later co‐founded the Cochrane Collaboration) stood and addressed him: ‘We have listened to your incessant criticisms of everyone who has gone to the trouble of obtaining data. What we have not heard from you is one single piece of evidence for your opinions.’ There was loud applause, and the future of these congresses was assured. They have taken place every four years since — in much better hotels.” Drummond Rennie, Let’s make peer review scientific, Nature, 05 July 2016, 31‐33.
[Final note, on “risk of bias”: The 1/29/18 Cochrane “Selenium” Review claims to follow GRADE guidelines in assessing an alleged “risk of bias” in the NPCT 2002 SeY trial. If you look further into this claim, which is repeated over and over again in the text of the review, you will see that it applies only to a risk of overstating the results of SeY supplementation for the outcome of prostate cancer incidence (resulting from a much higher rate of biopsy in the placebo arm of the study).
But, according the GRADE guidelines, “Summarizing study limitations must be outcome specific. Sources of bias may vary across outcomes. … For instance, RCTs of steroids for acute spinal cord injury measured both all‐cause mortality and, based on a detailed physical examination, motor function. Blinding of outcome assessors is irrelevant for mortality but crucial for motor function.”
Think about it: what effect would this alleged detection bias resulting in over ascertainment of prostate cancer incidence in the placebo group have on the outcome of total cancer mortality?
Isn’t it obvious—the purpose (and, indeed, the effect) of early CaP detection being to reduce cancer mortality—that the alleged detection bias actually favours SeY treatment for the mortality outcome? In other words, the alleged bias runs in two different directions for these two outcomes. As the Cochrane Risk of Bias Tool 2.0 states: “If the likely direction of the bias [for a particular outcome] can be predicted, it is helpful to predict this.” This flawed 2018 Cochrane review being such a shoddy product, it should come as no surprised to anyone that this extremely important point, along with many others, was overlooked entirely by the review authors, (and not picked up by the Cochrane peer reviewers, either, of course).
REFERENCES
1. Paul Glasziou and Iain Chalmers: Is 85% of health research really “wasted”?, thebmjopinion, January 14, 2016. 2. Chalmers I, Proposal to outlaw the term “negative trial”, Br Med J (Clin Res Ed) 1985;290:1002. 3. Wilson E.O. Biological Diversity: The Oldest Human Heritage. New York State Museum, 1999, at p. 7. 4. Machtay M and Glatstein E, "Just Another Statistic”, Oncologist. 1998;3(3):III‐IV. 5. Leibman B, “The Selenium Surprise”, Nutrition Action Jan/Feb 1997. 6. Bracken M, personal communication via email on January 31, 2016. 7. See, e.g.: Gilbert‐López B et al., Detection of over 100 selenium metabolites in selenized yeast, J Chromatogr B Analyt Technol Biomed Life Sci. 2017 Aug 15;1060:84‐90. Waters DJ, Shen S, Kengeri SS, Chiang EC, Combs GF Jr, Morris JS, Bostwick DG, Prostatic response to supranutritional selenium supplementation: comparison of the target tissue potency of selenomethionine vs. selenium‐yeast on markers of prostatic homeostasis, Nutrients. 2012 Nov 6;4(11):1650‐63. 8. Comparative effects of two different forms of selenium on oxidative stress biomarkers in healthy men: a randomised clinical trial, Cancer Prev Res (Phila). 2014 Aug;7(8):796‐804, Richie JP Jr, Das A, Calcagnotto AM, Sinha R, Neidig W, Liao J, Lengerich EJ, Berg A, Hartman TJ, Ciccarella A, Baker A, Kaag MG, Goodin S, DiPaola RS, El‐Bayoumy K. 9, Goldacre B, Bad Pharma: How Drug Companies Mislead Doctors and Harm Patients. Farrar, Straus & Giroux, 2013, at pp. 7‐8. 10. Effect of selenium and vitamin E on risk of prostate cancer and other cancers: the Selenium and Vitamin E Cancer Prevention Trial (SELECT),JAMA 2009 Jan 7;301(1):39‐51, Lippman SM1, Klein EA, Goodman PJ, Lucia MS, Thompson IM, Ford LG, et al.
Reply
We wish to thank Dr. Endicott for his comments. In response, we have amended the text of the review with reference to the following points:
a) Title. It was suggested that we change “selenium” to “selenium compounds” in the title and elsewhere. We acknowledge the relevance of the issue, and point out that the issue and importance of selenium speciation has been extensively considered throughout the current version of the review. We think that readers are unlikely to be misled into thinking that the review pertains exclusively to elemental selenium. Therefore we have not been asked by the Cochrane Editorial Methods Department to modify the original title of this Cochrane review, which has been unchanged since its original 2005 protocol.
b) Cancer mortality (and incidence). For the GRADE assessment and the related summary of findings, we were originally limited to seven outcomes, which have now been extended to eight. We have now added to the two summary of findings tables a) mortality from all cancer from experimental studies (RCTs) at low risk of bias, and b) mortality from all cancer from non‐experimental (observational) studies. Both estimates were already reported in the abstract.
c) NPC‐SELECT pooling. According to the methodology we have adopted for the entire assessment and all outcomes, i.e., to focus on randomised controlled trials and particularly on those at low risk of bias, we have not pooled NPC and SELECT for these additional outcomes reported in the GRADE assessment, due to their different risk of bias (possibly due to a breaking of blinding in the former, and for which the authors acknowledged a severe detection bias with reference to prostate biopsy rate). We have also extensively mentioned in the review how experimental human studies differed in term of selenium compounds administered, as well as how non‐experimental studies lacked exposure assessment of single selenium species. Both these issues may have been potential source of heterogeneity, as highlighted in the review.
d) The 2014 Richie et al. study and the Waters 2013 study. We now mentioned Richie et al. 2014 (as well as the comparable Rav‐Haren et al. Br J Nutr 2008 trial) which, though not eligible for our review, may be of help in showing how the proteomic and toxicological effects of the various selenium compounds administered to humans can be complex and inconsistent. These issues are definitively of interest to the relation between selenium and cancer risk. With reference to our Waters 2013 citation, we added more detail in the reference, to better allow readers to locate it.
October 30, 2018 ‐ Vinceti M, et al
Contributors
Marco Vinceti and Tommaso Filippini on behalf of the auhtor team.
What's new
| Date | Event | Description |
|---|---|---|
| 27 February 2020 | Amended | Minor edit to text made. |
History
Protocol first published: Issue 2, 2005 Review first published: Issue 5, 2011
| Date | Event | Description |
|---|---|---|
| 12 November 2018 | Amended | Minor edits made in relation to feedback submitted. |
| 12 November 2018 | Feedback has been incorporated | Feeback and author's response added. |
| 18 January 2018 | New citation required but conclusions have not changed | Addition of some recent references and minor edits. |
| 8 February 2017 | New search has been performed | New literature search conducted 8 February 2017. |
| 3 February 2015 | Feedback has been incorporated | Feedback and trial author's response added |
| 18 March 2014 | New citation required but conclusions have not changed | New trials added. Meta‐analysis of data from RCTs applied when at least 2 studies were available for each outcome |
| 15 February 2013 | New search has been performed | Search strategy updated |
| 14 August 2012 | Feedback has been incorporated | Additional feedback and trial author's response incorporated |
| 8 March 2012 | Feedback has been incorporated | Feedback submitted and trial author's reply added |
Acknowledgements
We thank the Cochrane Gynaecological, Neuro‐oncology and Orphan Cancer Group Editorial Team for their advice, in particular, Clare Jess, for her wisdom, kindness, and professionalism in overseeing the entire editorial process, and Joanne Platt, for designing the strategy and implementing the literature search. We also wish to thank Dr. Holger Schunemann, of McMaster University GRADE Center, for helping with the GRADE assessment.
This project was also supported by the National Institute for Health Research, via Cochrane Infrastructure funding to the Cochrane Gynaecological, Neuro‐oncology and Orphan Cancer Group. The views and opinions expressed therein are those of the authors and do not necessarily reflect those of the Systematic Reviews Programme, NIHR, NHS or the Department of Health.
Appendices
Appendix 1. Electronic search strategies
| Database | Date of most recent literature search | Search strategy | Comment |
| www.cancer.gov | 4 Feb 2011 | medication: selenium indication: prevention | |
| Cancerlit | Oct 2004 | 1 selen* OR organoselen* OR natriumselen* 2 random* OR placebo* OR clinical trial* OR controlled trial* OR controlled clinical trial* OR double blind* OR single blind* 3 epidemiologic stud* OR cohort OR case‐control stud* OR nested case‐control* OR case‐control design* OR prospectiv* 4 2 OR 3 5 1 AND 4 | Now included in MEDLINE database |
| Clinical Contents in Medicine (CCMed) | 4 Feb 2011 | selen* OR organoselen* OR natriumselen* | |
| CENTRAL | 2017, Issue 2 | #1 MeSH descriptor: [Selenium] this term only #2 MeSH descriptor: [Selenium Compounds] explode all trees #3 MeSH descriptor: [Organoselenium Compounds] explode all trees #4 selen* #5 #1 or #2 or #3 or #4 #6 MeSH descriptor: [Neoplasms] explode all trees #7 (neoplasm* or cancer* or tumor* or tumour* or carcino* or malignan* or adenocarcinoma* or sarcoma* or adenoma* or chondrosarcoma* or fibrosarcoma* or dermatofibrosarcoma* or neurofibrosarcoma* or hemangiosarcoma* or leiomyosarcoma* or liposarcoma* or myosarcoma* or rhabdomyosarcoma* or myxosarcoma* or osteosarcoma* or lymphoma*) #8 #6 or #7 #9 #5 and #8 |
|
| metaRegister of Controlled Trials (mRCT, www.controlled‐trials.com) | 4 Feb 2011 | selen AND cancer | Now included in the ISRCTN registry |
| Embase Ovid | 2017 week 6 | 1 selenium/ 2 selen*.mp. 3 selenium derivative/ 4 methylseleninic acid/ 5 methylselenium.mp. 6 exp organoselenium derivative/ 7 1 or 2 or 3 or 4 or 5 or 6 8 exp neoplasm/ 9 (neoplasm* or cancer* or tumor* or tumour* or carcino* or malignan* or adenocarcinoma* or sarcoma* or adenoma* or chondrosarcoma* or fibrosarcoma* or dermatofibrosarcoma* or neurofibrosarcoma* or hemangiosarcoma* or leiomyosarcoma* or liposarcoma* or myosarcoma* or rhabdomyosarcoma* or myxosarcoma* or osteosarcoma* or lymphoma*).mp. 10 8 or 9 11 7 and 10 12 exp clinical study/ 13 crossover procedure/ 14 double‐blind procedure/ 15 single‐blind procedure/ 16 cohort analysis/ 17 observational study/ 18 (random* or factorial* or crossover* or cross‐over* or cross over* or placebo* or (double adj blind*) or (singl* adj blind*) or assign* or allocat* or volunteer* or observ* or cohort* or prospectiv* or (case* and control*)).mp. 19 12 or 13 or 14 or 15 or 16 or 17 or 18 20 11 and 19 21 (exp animal/ or nonhuman/ or exp animal experiment/) not human/ 22 20 not 21 key: [mp=title, abstract, subject headings, heading word, drug trade name, original title, device manufacturer, drug manufacturer, device trade name, keyword] |
|
| German Cancer Study Register: www.studien.de |
Feb 2017 | selen | |
| MEDLINE (via Ovid) | Jan 2017, week 4 | 1 Selenium/ 2 exp Selenium Compounds/ 3 exp Organoselenium Compounds/ 4 selen*.mp. 5 1 or 2 or 3 or 4 6 exp Neoplasms/ 7 (neoplasm* or cancer* or tumor* or tumour* or carcino* or malignan* or adenocarcinoma* or sarcoma* or adenoma* or chondrosarcoma* or fibrosarcoma* or dermatofibrosarcoma* or neurofibrosarcoma* or hemangiosarcoma* or leiomyosarcoma* or liposarcoma* or myosarcoma* or rhabdomyosarcoma* or myxosarcoma* or osteosarcoma* or lymphoma*).mp. 8 6 or 7 9 5 and 8 10 randomized controlled trial.pt. 11 controlled clinical trial.pt. 12 randomized.ab. 13 placebo.ab. 14 drug therapy.fs. 15 randomly.ab. 16 trial.ab. 17 groups.ab. 18 exp case‐control studies/ 19 exp Cohort Studies/ 20 (cohort* or observ* or prospectiv* or (case* and control*)).mp. 21 10 or 11 or 12 or 13 or 14 or 15 or 16 or 17 or 18 or 19 or 20 22 9 and 21 23 exp animals/ not humans.sh. 24 22 not 23 key: mp=title, abstract, original title, name of substance word, subject heading word, keyword heading word, protocol supplementary concept, rare disease supplementary concept, unique identifier pt=publication type ab=abstract fs=floating subheading |
|
| SIGLE | Oct 2004 | ?selen? | database discontinued in 2005 |
| ISRCTN Registry (www.isrctn.com) |
Feb 2017 | selen AND cancer | |
| ClinicalTrials.gov Registry (www.clinicaltrials.gov) |
Feb 2017 | selen AND cancer |
Appendix 2. Newcastle‐Ottawa Scale for Cohort Studies
((*) means that a 'star' was assigned to the study for the corresponding item)
1) Selection
1.1) representativeness of the exposed cohort a) truly representative of the average _____________ (target population) in the community (*) b) somewhat representative of the average _____________ (target population) in the community (*) c) selected group of users, e.g., volunteers / nurses d) no description of the derivation of the cohort 1.2) selection of the non‐exposed cohort a) drawn from the same community as the exposed cohort (*) b) drawn from a different source c) no description 1.3) ascertainment of selenium exposure a) secure record (biochemical records) (*) b) structured interview (*) c) written self report or medical record only d) no description 1.4) demonstration that outcome of interest was not present at start of study a) no history of disease or exclusion of cases that occurred in the first 12 months (*) b) not stated
2) Comparability
2.1.) comparability of cohorts on the basis of the design or analysis a) study controls for AGE (*) b) study controls for SMOKING (*)
3) Outcome
3.1) assessment of outcome a) independent blind validation (> 1 person/record/time/process to extract information or reference to primary source such as X‐rays/hospital records) (*) b) record linkage (e.g., ICD codes in databases) (*) c) self report d) no description 3.2) Was follow‐up long enough for outcomes to occur? a) yes (> 3 years) b) no 3.3) adequacy of follow up of cohorts a) complete follow‐up of all subjects (*) OR b) subjects lost to follow‐up unlikely to introduce bias (< 5% lost to follow‐up or description provided of lost people) (*) c) follow‐up‐rate < 95% and no description of those lost d) no statement
Appendix 3. Additional Newcastle‐Ottawa Scale for Nested Case‐Control Studies
((*) means that a 'star' was assigned to the study for the corresponding item)
1) Selection
1.1) case definition a) independent validation (> 1 person/record/time/process to extract information or reference to primary source such as X‐rays/hospital records) (*) b) record linkage (e.g., ICD codes in databases) or self‐report with no reference to primary record c) no description 1.2) representativeness of cases: a) all eligible cases with outcome of interest over a defined period, cases in a defined catchment area/hospital etc. or an appropriate/random sample of those cases (*) b) not satisfying requirements in part (a) or not stated 1.3) selection of controls: a) community controls (same community and would be cases if had outcome) (*) b) hospital controls (within the same population e.g., city as cases) c) no description 1.4) definition of controls a) cases had no history of outcome controls had no history of outcome OR case had new (not necessarily first) occurrence of outcome controls with previous occurrence of outcome should not be excluded (*) b) no mention of history of outcome
2) Comparability
(validated in cohort assessment in question 2 ‐ number of stars was copied)
3) Exposure
3.1) ascertainment of selenium exposure: (validated in cohort assessment in question 1.3 ‐ number of stars was copied) 3.2) Same method of ascertainment for cases and controls a) yes (*) b) no 3.3) non‐response rate a) same rate for both groups (*) b) non‐respondents described c) rate different and no designation
Data and analyses
Comparison 1. Randomised controlled trials: highest versus lowest selenium exposure.
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
|---|---|---|---|---|
| 1 Any cancer risk | 5 | Risk Ratio (IV, Random, 95% CI) | Subtotals only | |
| 1.1 Studies with low RoB | 3 | 19475 | Risk Ratio (IV, Random, 95% CI) | 1.01 [0.93, 1.10] |
| 1.2 All studies | 5 | 21860 | Risk Ratio (IV, Random, 95% CI) | 0.99 [0.86, 1.14] |
| 2 Cancer mortality | 2 | Risk Ratio (IV, Random, 95% CI) | Subtotals only | |
| 2.1 Studies with low RoB | 1 | 17448 | Risk Ratio (IV, Random, 95% CI) | 1.02 [0.80, 1.30] |
| 2.2 All studies | 2 | 18698 | Risk Ratio (IV, Random, 95% CI) | 0.81 [0.49, 1.32] |
| 3 Head and neck cancer risk | 2 | Risk Ratio (IV, Random, 95% CI) | Subtotals only | |
| 3.1 Studies with low RoB | 1 | 1561 | Risk Ratio (IV, Random, 95% CI) | 1.00 [0.18, 5.45] |
| 3.2 All studies | 2 | 2811 | Risk Ratio (IV, Random, 95% CI) | 1.22 [0.52, 2.85] |
| 4 Oesophageal cancer risk | 2 | Risk Ratio (IV, Random, 95% CI) | Subtotals only | |
| 4.1 Studies with low RoB | 1 | 1561 | Risk Ratio (IV, Random, 95% CI) | 1.50 [0.06, 36.86] |
| 4.2 All studies | 2 | 2811 | Risk Ratio (IV, Random, 95% CI) | 0.53 [0.12, 2.28] |
| 5 Colorectal cancer risk | 3 | Risk Ratio (IV, Random, 95% CI) | Subtotals only | |
| 5.1 Studies with low RoB | 2 | 19009 | Risk Ratio (IV, Random, 95% CI) | 0.99 [0.69, 1.43] |
| 5.2 All studies | 3 | 20259 | Risk Ratio (IV, Random, 95% CI) | 0.74 [0.41, 1.33] |
| 6 Liver cancer risk | 4 | Risk Ratio (IV, Random, 95% CI) | Subtotals only | |
| 6.1 Studies with low RoB | 1 | 1561 | Risk Ratio (IV, Random, 95% CI) | 6.52 [0.37, 115.49] |
| 6.2 All studies | 4 | 6326 | Risk Ratio (IV, Random, 95% CI) | 0.52 [0.35, 0.79] |
| 7 Melanoma risk | 3 | Risk Ratio (IV, Random, 95% CI) | Subtotals only | |
| 7.1 Studies with low RoB | 2 | 2027 | Risk Ratio (IV, Random, 95% CI) | 1.35 [0.41, 4.52] |
| 7.2 All studies | 3 | 3277 | Risk Ratio (IV, Random, 95% CI) | 1.28 [0.63, 2.59] |
| 8 Non‐melanoma skin cancer risk | 4 | Risk Ratio (Random, 95% CI) | Subtotals only | |
| 8.1 Studies with low RoB | 2 | 2027 | Risk Ratio (Random, 95% CI) | 1.16 [0.30, 4.42] |
| 8.2 All studies | 4 | 3461 | Risk Ratio (Random, 95% CI) | 1.23 [0.73, 2.08] |
| 9 Lung cancer risk | 3 | Risk Ratio (IV, Random, 95% CI) | Subtotals only | |
| 9.1 Studies with low RoB | 2 | 19009 | Risk Ratio (IV, Random, 95% CI) | 1.16 [0.89, 1.50] |
| 9.2 All studies | 3 | 20259 | Risk Ratio (IV, Random, 95% CI) | 1.03 [0.78, 1.37] |
| 10 Breast cancer risk | 3 | Risk Ratio (IV, Random, 95% CI) | Subtotals only | |
| 10.1 Studies with low RoB | 1 | 802 | Risk Ratio (IV, Random, 95% CI) | 2.04 [0.44, 9.55] |
| 10.2 All studies | 3 | 2260 | Risk Ratio (IV, Random, 95% CI) | 1.44 [0.96, 2.17] |
| 11 Bladder cancer risk | 3 | Risk Ratio (IV, Random, 95% CI) | Subtotals only | |
| 11.1 Studies with low RoB | 2 | 19009 | Risk Ratio (IV, Random, 95% CI) | 1.07 [0.76, 1.52] |
| 11.2 All studies | 3 | 20259 | Risk Ratio (IV, Random, 95% CI) | 1.10 [0.79, 1.52] |
| 12 Prostate cancer risk | 5 | Risk Ratio (IV, Random, 95% CI) | Subtotals only | |
| 12.1 Studies with low RoB | 4 | 18942 | Risk Ratio (IV, Random, 95% CI) | 1.01 [0.90, 1.14] |
| 12.2 All studies | 5 | 19869 | Risk Ratio (IV, Random, 95% CI) | 0.91 [0.75, 1.12] |
| 13 Leukaemia and lymphoma risk | 2 | Risk Ratio (IV, Random, 95% CI) | Subtotals only | |
| 13.1 Studies with low RoB | 1 | 1561 | Risk Ratio (IV, Random, 95% CI) | 1.00 [0.25, 3.99] |
| 13.2 All studies | 2 | 2811 | Risk Ratio (IV, Random, 95% CI) | 1.21 [0.52, 2.80] |
Comparison 2. Observational studies: highest versus lowest selenium exposure.
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
|---|---|---|---|---|
| 1 Total cancer incidence and mortality | 14 | Odds Ratio (Random, 95% CI) | Subtotals only | |
| 1.1 Incidence | 7 | Odds Ratio (Random, 95% CI) | 0.72 [0.55, 0.93] | |
| 1.2 Mortality | 7 | Odds Ratio (Random, 95% CI) | 0.76 [0.59, 0.97] | |
| 2 Total cancer incidence and mortality (men) | 8 | Odds Ratio (Random, 95% CI) | Subtotals only | |
| 2.1 Incidence | 4 | Odds Ratio (Random, 95% CI) | 0.72 [0.46, 1.14] | |
| 2.2 Mortality | 4 | Odds Ratio (Random, 95% CI) | 0.65 [0.45, 0.94] | |
| 3 Total cancer incidence and mortality (women) | 6 | Odds Ratio (Random, 95% CI) | Subtotals only | |
| 3.1 Incidence | 2 | Odds Ratio (Random, 95% CI) | 0.90 [0.45, 1.77] | |
| 3.2 Mortality | 4 | Odds Ratio (Random, 95% CI) | 0.91 [0.80, 1.03] | |
| 4 Total cancer incidence and mortality (ascending order of selenium levels) | 13 | Odds Ratio (Random, 95% CI) | Subtotals only | |
| 4.1 Incidence | 7 | 1642 | Odds Ratio (Random, 95% CI) | 0.72 [0.55, 0.93] |
| 4.2 Mortality | 6 | 1230 | Odds Ratio (Random, 95% CI) | 0.63 [0.39, 1.01] |
| 5 Total cancer incidence and mortality (ascending order of differences in selenium levels) | 13 | Odds Ratio (Random, 95% CI) | Subtotals only | |
| 5.1 Incidence | 7 | 190 | Odds Ratio (Random, 95% CI) | 0.72 [0.55, 0.93] |
| 5.2 Mortality | 6 | 106 | Odds Ratio (Random, 95% CI) | 0.63 [0.39, 1.01] |
| 6 Stomach cancer risk | 5 | Odds Ratio (Random, 95% CI) | 0.66 [0.43, 1.01] | |
| 7 Stomach cancer risk (by sex) | 5 | Odds Ratio (Random, 95% CI) | 0.66 [0.42, 1.04] | |
| 7.1 All (male + female) | 2 | Odds Ratio (Random, 95% CI) | 0.75 [0.41, 1.36] | |
| 7.2 Male | 3 | Odds Ratio (Random, 95% CI) | 0.43 [0.14, 1.32] | |
| 7.3 Female | 2 | Odds Ratio (Random, 95% CI) | 0.73 [0.12, 4.35] | |
| 8 Colorectal cancer risk | 6 | Odds Ratio (Random, 95% CI) | 0.82 [0.72, 0.94] | |
| 9 Colorectal cancer risk (by sex) | 6 | Odds Ratio (Random, 95% CI) | 0.83 [0.72, 0.95] | |
| 9.1 All (male + female) | 1 | Odds Ratio (Random, 95% CI) | 0.80 [0.68, 0.94] | |
| 9.2 Male | 4 | Odds Ratio (Random, 95% CI) | 0.86 [0.65, 1.16] | |
| 9.3 Female | 4 | Odds Ratio (Random, 95% CI) | 0.96 [0.61, 1.50] | |
| 10 Colon cancer risk | 5 | Odds Ratio (Random, 95% CI) | 0.81 [0.69, 0.96] | |
| 11 Colon cancer risk (by sex) | 5 | Odds Ratio (Random, 95% CI) | 0.81 [0.69, 0.96] | |
| 11.1 All (male + female) | 2 | Odds Ratio (Random, 95% CI) | 0.84 [0.68, 1.03] | |
| 11.2 Male | 3 | Odds Ratio (Random, 95% CI) | 0.84 [0.56, 1.25] | |
| 11.3 Female | 2 | Odds Ratio (Random, 95% CI) | 0.68 [0.44, 1.04] | |
| 12 Lung cancer incidence and mortality | 13 | Odds Ratio (Random, 95% CI) | Subtotals only | |
| 12.1 Incidence | 11 | Odds Ratio (Random, 95% CI) | 0.82 [0.59, 1.14] | |
| 12.2 Mortality | 2 | Odds Ratio (Random, 95% CI) | 1.34 [0.93, 1.93] | |
| 13 Lung cancer risk (sex‐disaggregated data) | 13 | Odds Ratio (Random, 95% CI) | 0.89 [0.69, 1.14] | |
| 13.1 All (male + female) | 5 | Odds Ratio (Random, 95% CI) | 0.74 [0.43, 1.28] | |
| 13.2 Male | 7 | Odds Ratio (Random, 95% CI) | 0.98 [0.68, 1.39] | |
| 13.3 Female | 4 | Odds Ratio (Random, 95% CI) | 0.83 [0.43, 1.61] | |
| 14 Lung cancer risk (by exposure assessment) | 13 | Odds Ratio (Random, 95% CI) | 0.88 [0.65, 1.18] | |
| 14.1 Intake | 2 | Odds Ratio (Random, 95% CI) | 1.32 [0.95, 1.84] | |
| 14.2 Serum or plasma | 9 | Odds Ratio (Random, 95% CI) | 0.91 [0.70, 1.18] | |
| 14.3 Toenail | 2 | Odds Ratio (Random, 95% CI) | 1.05 [0.11, 10.36] | |
| 15 Lung cancer risk (ascending order of selenium levels) | 8 | 1938 | Odds Ratio (Random, 95% CI) | 0.97 [0.74, 1.27] |
| 16 Lung cancer risk (ascending order of differences in selenium levels) | 8 | 188 | Odds Ratio (Random, 95% CI) | 0.97 [0.74, 1.27] |
| 17 Breast cancer risk (women) | 8 | Odds Ratio (Random, 95% CI) | 1.09 [0.87, 1.37] | |
| 18 Bladder cancer risk | 5 | Odds Ratio (Random, 95% CI) | 0.67 [0.46, 0.97] | |
| 18.1 All (male + female) | 2 | Odds Ratio (Random, 95% CI) | 0.65 [0.46, 0.92] | |
| 18.2 Male | 3 | Odds Ratio (Random, 95% CI) | 0.82 [0.41, 1.62] | |
| 18.3 Female | 1 | Odds Ratio (Random, 95% CI) | 0.36 [0.14, 0.92] | |
| 19 Prostate cancer risk | 21 | Odds Ratio (Random, 95% CI) | 0.84 [0.75, 0.95] | |
| 20 Prostate cancer risk (by exposure assessment) | 21 | Odds Ratio (Random, 95% CI) | 0.84 [0.75, 0.95] | |
| 20.1 Intake and supplement | 4 | Odds Ratio (Random, 95% CI) | 0.99 [0.85, 1.15] | |
| 20.2 Serum or plasma | 13 | Odds Ratio (Random, 95% CI) | 0.86 [0.75, 0.99] | |
| 20.3 Toenail | 4 | Odds Ratio (Random, 95% CI) | 0.60 [0.44, 0.82] | |
| 21 Prostate cancer risk (ascending order of selenium levels) | 13 | 2816 | Odds Ratio (Random, 95% CI) | 0.86 [0.75, 0.99] |
| 22 Prostate cancer risk (ascending order of differences in selenium levels) | 13 | 345 | Odds Ratio (Random, 95% CI) | 0.86 [0.75, 0.99] |
Characteristics of studies
Characteristics of included studies [ordered by study ID]
Agalliu 2011.
| Methods | Nested case‐cohort study Country: Canada |
|
| Participants |
Name of parent cohort: Canadian Study of Diet, Lifestyle and Health (CSDLH)
Participants: 22,975 (alumni associations of the University of Western Ontario, 67% of 34,291) Recruitment: between 1995 and 1998 Outcome assessment: December 2003 Number of cases: • Prostate cancer: 661 Case definition: incidence Years of follow‐up: 4.3 to 7.7 mean Type of selenium marker: supplementation |
|
| Interventions | d.n.a. | |
| Outcomes | Statistical methods: Cox proportional hazard model Variables controlled in analysis: age at baseline, race, BMI, exercise activity, education | |
| Risk estimates [95% CI] |
Reference category: zero Results: Prostate cancer • Highest quartile: HR 0.76 (95% CI 0.43 to 1.33) |
|
| Selenium levels in exposure categories | Lowest quartile (median value): 15.7 µg Highest quartile (median value): 105.0 µg | |
| Notes | ||
Akbaraly 2005.
| Methods | Cohort/subcohort controlled cohort study Country: France |
|
| Participants |
Name of parent cohort: Etude du Vieillissement Antériel Study (EVA study)
Participants: 1389 (41% male, 59% female)
Inclusion criteria: 59 to 71 years of age; residents of Nantes; able to undergo examination at study centre Recruitment: 1991 to 1993 Outcome assessment: December 2001 Number of cases: • Any cancer: 45 (male/female: n.r.) Case definition: mortality Years of follow‐up: 9.0 Type of selenium marker: plasma |
|
| Interventions | d.n.a. | |
| Outcomes | Statistical methods: Cox proportional hazard model Variables controlled in analysis: gender, smoking, alcohol intake, medication use, obesity, diabetes mellitus, hypertension, CVD, age, education, dyslipidaemia, low cognitive function | |
| Risk estimates [95% CI] |
Reference category: highest quartile Results: Any cancer • Both genders: lowest quartile: RR 4.06 (95% CI 1.51 to 10.92) |
|
| Selenium levels in exposure categories | Lowest quartile: 14.2 to 75.0 µg/L Highest quartile: 96.3 to 155.6 µg/L | |
| Notes | ||
Algotar 2013.
| Methods | Randomised controlled trial Allocation: random Sequence generation: unclear Concealment: Study agent (2 doses) and matched placebo caplets were coated with titanium oxide to ensure identical appearance, weight, taste, and smell. Blinding: described only as double‐blinded Dropouts/withdrawals: Study dropout percentage was 34.1%, 41.9%, and 40.8% for placebo, 200 mg/d selenium group, and 400 mg/d selenium group, respectively (P = 0.173). Intention‐to‐treat‐analysis: yes Recruitment period: not specified Treatment duration: not specified Observation period/dermatological follow‐up: Participants were followed every 6 months for up to 5 years. Detection of cases: Tissue samples from participants' qualifying biopsies were requested by participants' physicians and were compiled in a biospecimen repository. Informed consent: An external Data and Safety Monitoring Committee (DSMC) was established before study initiation. This committee was responsible for reviewing protocol amendments, consent forms, accrual and retention rates, adverse events, and data analysis reports. |
|
| Participants | 699 male participants with a negative prostate biopsy Countries: United States, New Zealand Participants: 699 (randomised to selenium 200 µg/d: 234; to selenium 400 µg/d: 233; to placebo: 233) Condition: male patients at high risk for prostate cancer (prostate‐specific antigen (PSA) > 4 ng/mL and/or suspicious digital rectal examination and/or PSA velocity > 0.75 ng/mL/y), but with a negative prostate biopsy Demographics: mean age 65.2 ± SD 8 years (selenium 200 µg/d), 65.5 ± 7.7 years (selenium 400 µg/d), 65.5 ± 7.4 years (placebo) Recruitment and setting: urology offices at 20 sites in the United States and New Zealand |
|
| Interventions |
Intervention: • 200 µg/d selenium supplied as selenium yeast • 400 µg/d selenium supplied as selenium yeast Control: placebo Recruitment: not reported End of blinded treatment period: For participants in the United States, participation was complete at 5 years, whereas those in New Zealand received intervention for no longer than 3 years. |
|
| Outcomes |
Primary outcome measure: • Incidence of biopsy‐proven prostate cancer over the course of the study Other reported outcomes: • Secondary endpoint was rate of change of PSA over time (i.e. PSA velocity) based on biannual PSA measurements. |
|
| Risk estimates [95% CI] |
Primary outcomes: • Hazard ratios for risk of developing prostate cancer in the selenium 200‐µg/d or the selenium 400‐µg/d group were 0.94 (95% CI 0.52 to 1.7) and 0.90 (95% CI 0.48 to 1.70), respectively. Other reported outcomes: • PSA velocity in the selenium arms was not significantly different from that observed in the placebo group (P = 0.18 and P = 0.17, respectively). |
|
| Selenium levels in exposure categories | d.n.a. | |
| Notes | The DSMC recommended that the trial be stopped before all participants completed the full intervention duration. Adverse effects: No significant differences were seen in the incidences of cataract/glaucoma or in hair/nail changes in the 3 treatment groups. HR: adjusted for age at baseline, baseline PSA, baseline selenium concentrations |
|
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Random sequence generation (selection bias) | Low risk | Number‐based stratified randomisation |
| Allocation concealment (selection bias) | Low risk | Treatments and placebo tablets of identical appearance and taste |
| Blinding (performance bias and detection bias) All outcomes | Low risk | Identical appearance, weight, taste, and smell of tablets for treatments and placebo |
| Selective reporting (reporting bias) | Low risk | No problems found |
Allen 2008.
| Methods | Matched, nested case‐control study Countries: Denmark, Germany, Greece, Italy, the Netherlands, Spain, Sweden, the UK |
|
| Participants |
Participants: approximately 130,000 men
Inclusion criteria: male participants from the EPIC study Name of parent cohort: European Prospective Investigation into Cancer and Nutrition (EPIC) Recruitment: 1992 to 2000 Outcome assessment: at each country's study closure date (between June 1999 and January 2003) Number of cases: • Prostate cancer: 959 (male/female: 959/0) Case definition: incidence Years of follow‐up: median 2.6 (Greece) to 9.2 (Sweden) Type of selenium marker: plasma |
|
| Interventions | d.n.a. | |
| Outcomes |
Statistical methods: conditional logistical regression
Variables controlled in analysis: BMI, smoking, alcohol consumption, physical activity, marital status, education Variables controlled by matching: age, study centre, time of day of blood collection, time between blood collection and last meal, sex |
|
| Risk estimates [95% CI] |
Reference category: lowest quintile Results: Prostate cancer • Highest quintile: OR 0.96 (95% CI 0.70 to 1.31) |
|
| Selenium levels in exposure categories | Lowest quintile < 62.0 µg/L Highest quintile ≥ 84.1 µg/L |
|
| Notes | ||
Banim 2013.
| Methods | Nested case‐cohort study Country: UK |
|
| Participants |
Participants: 23,658 men and women Inclusion criteria: aged 40 to 74, resident in Norfolk county, registered at 35 general practices in rural, suburban, and inner city areas, no history of pancreatic cancer at enrolment or within 12 months of entering the study Name of parent cohort: European Prospective Investigation of Cancer‐Norfolk Study (EPIC‐Norfolk) Recruitment: 1993 to 1997 Case definition: incidence Type of selenium marker: intake Banim 2013: Outcome assessment: June 2010 Number of cases: • Pancreatic cancer: 86 (male/female: 38/48) Years of follow‐up: 17 Barrass 2013: Outcome assessment: December 2010 Number of cases: • Renal cell carcinoma: 65 (male/female: n.r.) Years of follow‐up: not reported (probably 17) |
|
| Interventions | d.n.a. | |
| Outcomes |
Statistical methods: Cox proportional hazard model Variables controlled in analysis: age, sex, smoking, diabetes, total energy intake, body mass index category, respective antioxidant supplement (only Banim 2013) |
|
| Risk estimates [95% CI] |
Reference category: lowest quartile, lowest quintile Results: Banim 2013: • Pancreatic cancer: highest quartile: HR 0.72 (95% CI 0.36 to 1.43) Barrass 2013: • Renal cell cancer: highest quintile: HR 0.40 (95% CI 0.17 to 0.98) |
|
| Selenium levels in exposure categories |
Banim 2013: • Lowest quartile < 43.6 µg/d • Highest quartile ≥ 72.0 µg/d Barrass 2013: • Lowest and highest quintiles not reported |
|
| Notes | ||
Bleys 2008.
| Methods | Cohort study Country: United States |
|
| Participants |
Name of parent cohort: Third National Health and Nutrition Examination Survey (NHANES III) Inclusion criteria: male and female adults, aged 20 to 90 years, participating in the NHANES III: "stratified, multistage probability cluster to provide data representing the noninstitutionalized US population" (Bleys 2008, p. 404) Recruitment: 1988 to 1994 Participants: 13,887 men and women Outcome assessment: 15 December 2000 Number of cases: • Cancer deaths: 457 (male/female: n.r.) Case definition: mortality Years of follow‐up: 6 to 12 Type of selenium marker: serum |
|
| Interventions | d.n.a. | |
| Outcomes |
Analysed cases: 457 (male/female: n.r.) Statistical methods: Cox proportional hazard regression Variables controlled in analysis: age, sex, race, education, annual family income, postmenopausal status (women), cigarette smoking, serum cotinine level, alcohol consumption |
|
| Risk estimates [95% CI] |
Reference category: lowest tertile Results: Cancer deaths • Both genders: highest tertile: HR 0.69 (95% CI 0.53 to 0.90) • Both genders: highest tertile: HR 0.68 (95% CI 0.48 to 0.97); cases at baseline excluded |
|
| Selenium levels in exposure categories | Lowest tertile < 117.31 µg/L Highest tertile ≥ 130.39 µg/L | |
| Notes | Updated results with longer follow‐up for the same population reported in Goyal 2013 | |
Brooks 2001.
| Methods | Matched, nested case‐control study Country: United States |
|
| Participants |
Name of parent cohort: Baltimore Longitudinal Study of Aging
Participants: 1555 men
Inclusion criteria: n.r. Recruitment: n.r. Outcome assessment: n.r. Number of cases: • Prostate cancer: 52 (male/female: 52/0) Case definition: incidence Years of follow‐up: n.r. Type of selenium marker: plasma |
|
| Interventions | d.n.a. | |
| Outcomes | Analysed cases: 52 of 133 (reason for non‐inclusion: plasma and/or histological confirmation of diagnosis not available) Statistical methods: logistical regression Variables controlled in analysis: years between blood donation and diagnosis/follow‐up, age, age by years before diagnosis interaction, BMI, smoking history, alcohol use Variables controlled by matching: age | |
| Risk estimates [95% CI] |
Reference category: lowest quartile Results: Prostate cancer • Highest quartile: OR 0.24 (95% CI 0.07 to 0.77) |
|
| Selenium levels in exposure categories | Lowest quartile: 82 to 107 µg/L Highest quartile: 133 to 182 µg/L | |
| Notes | ||
Clark 1985.
| Methods | Cohort/subcohort controlled cohort study Country: United States |
|
| Participants |
Participants: 177; no information on gender
Inclusion criteria: persons at high risk of non‐melanoma skin cancer Recruitment: n.r. Outcome assessment: n.r. Number of cases: • Skin (non‐melanoma): 19 (male/female: n.r.) Case definition: incidence Years of follow‐up: mean 3 Type of selenium marker: plasma |
|
| Interventions | d.n.a. | |
| Outcomes | Statistical methods: Cox proportional hazard model | |
| Risk estimates [95% CI] |
Reference category: lower half Results: Skin (non‐melanoma) • Sex n.r.: higher half: RR 0.77 (CI not reported) |
|
| Selenium levels in exposure categories | n.r. | |
| Notes | ||
Coates 1988.
| Methods | Matched, nested case‐control study Country: United States |
|
| Participants |
Participants: 6167; both genders
Inclusion criteria: employees of 2 Seattle companies Recruitment: 1972 to 1973 and 1976 Outcome assessment: not stated Number of cases: • Any cancer: 154 (male/female: n.r.) • Gastrointestinal cancer: 28 (male/female: n.r.) • Breast cancer: 20 (male/female: 0/20) • Prostate cancer: 13 (male/female: 13/0) • Haematological cancers: 12 (male/female: n.r.) • Cervical cancer: 12 (male/female: 0/12) • Lung cancer: 11 (male/female: n.r.) • Other: 58 (male/female: n.r.) Case definition: incidence Years of follow‐up: n.r. Type of selenium marker: serum and plasma |
|
| Interventions | d.n.a. | |
| Outcomes | Analysed cases: 154 (133 serum, 21 plasma) of 195 collected (reason for non‐inclusion: no sample available for analysis or no control available) Statistical methods: conditional logistic regression Variables controlled by matching: age, gender, race/ethnicity, year/month of sample collection, employer, plasma or serum sample | |
| Risk estimates [95% CI] |
Reference category: lowest Results: Any cancer • Both genders: highest quintile: OR 1.0 (95% CI 0.5 to 1.8) Gastrointestinal cancer • Both genders: highest tertile: OR 1.0 (CI not reported) Breast cancer Highest tertile: OR 3.4 (CI not reported) Prostate cancer Highest tertile: OR 0.3 (CI not reported) Haematological cancers Both genders: highest tertile: OR 0.6 (CI not reported) Cervical cancer Highest tertile: OR 1.1 (CI not reported) Lung cancer Both genders: highest tertile: OR 0.8 (CI not reported) Other cancers Both genders: highest tertile: OR 0.9 (CI not reported) |
|
| Selenium levels in exposure categories |
Serum:
• Lowest quintile: 98 to 142 µg/L
• Highest quintile: 181 to 240 µg/L
• Lowest tertile: 98 to 148 µg/L
• Highest tertile: 171 to 240 µg/L Plasma: • Lowest quintile: 115 to 129 µg/L • Highest quintile: 157 to 207 µg/L • Lowest tertile: 115 to 137 µg/L • Highest tertile: 151 to 207 µg/L |
|
| Notes | Primary publication:Coates 1988 Secondary publication: Coates 1987 | |
Combs 1993.
| Methods | Cohort/subcohort controlled cohort study Country: United States |
|
| Participants |
Participants: 1239 men and women
Inclusion criteria: participants from the NPCT with valid selenium measurement at baseline
Name of parent cohort: Nutritional Prevention of Cancer Trial (NPCT) Recruitment: see: Nutritional Prevention of Cancer Trial Outcome assessment: not stated Number of cases: • Squamous cell cancer: 204 (male/female: n.r.) Case definition: incidence Years of follow‐up: 2 Type of selenium marker: plasma |
|
| Interventions | d.n.a. | |
| Outcomes | Statistical methods: Cox proportional hazard model Variables controlled in analysis: age, gender, current smoking, alcohol drinking | |
| Risk estimates [95% CI] |
Reference category (unadjusted RR): lower half Results: Squamous cell cancer • Both genders: higher half: unadjusted RR 0.69 (95% CI 0.51 to 0.92) • Both genders: "interquartile contrast" (high vs low), adjusted RR 0.79 (95% CI 0.67 to 0.94) |
|
| Selenium levels in exposure categories | Lower half: ≤ 114.00 µg/L Higher half: ≥ 114.10 µg/L | |
| Notes | ||
Comstock 1997.
| Methods | Matched, nested case‐control study Country: United States |
|
| Participants |
Participants: 44,960 men and women (20,305 from CLUE I; 24,655 from CLUE II)
Inclusion criteria: residents of Washington County
Name of parent cohort: CLUE I and II Cohort Recruitment: 1974/75 or 1989 Outcome assessment: n.r. Number of cases: • Lung cancer: 258 (male/female: 157/101) Case definition: incidence Years of follow‐up: n.r. Type of selenium marker: serum/plasma |
|
| Interventions | d.n.a. | |
| Outcomes | Statistical methods: conditional logistical regression Variables controlled by matching: age, gender, race/ethnicity, year and month of sample collection, participant of Clue I or Clue II cohort | |
| Risk estimates [95% CI] |
Reference category: lowest quintile Results: Lung cancer • Both genders: highest quintile: OR 0.65 (CI n.r.) |
|
| Selenium levels in exposure categories | n.r. | |
| Notes | ||
Dong 2008.
| Methods | Cohort study Country: United States |
|
| Participants |
Participants: 339 (male/female: 275/64)
Inclusion criteria: participants from a surveillance programme for men and women with Barrett's oesophagus, no prior history of oesophageal cancer or diagnosis of cancer within first 3 months of baseline Name of parent cohort: Seattle Barrett's Esophagus Program Recruitment: 1983 to 2004, baseline assessment for this study: 1 February 1995 to 1 July 2004 Outcome assessment: n.r. Number of cases: oesophageal adenocarcinoma: 37 (male/female: 32/5) Case definition: incidence Years of follow‐up: mean: 5 Type of selenium marker: intake of selenium supplements (self‐administered food frequency questionnaire) |
|
| Interventions | d.n.a. | |
| Outcomes | Statistical methods: Cox proportional hazard regression Variables controlled in analysis: age, sex, fruit and vegetable consumption, per cent energy from fat, waist‐hip ratio, cigarette smoking, non‐steroidal anti‐inflammatory drug use | |
| Risk estimates [95% CI] |
Reference category: no supplemental selenium intake (lowest exposure category) Results: • Both genders: supplement intake ≥ 50 µg/d: HR 0.27 (95% CI 0.03 to 2.21) |
|
| Selenium levels in exposure categories | Lowest category: no supplemental selenium intake Middle category: supplemental selenium intake < 50 µg/d Highest category: supplemental intake ≥ 50 µg/d |
|
| Notes | ||
Dorgan 1998.
| Methods | Matched, nested case‐control study Country: United States |
|
| Participants |
Participants: 6426 women
Inclusion criteria: female volunteers with serum available at the Breast Cancer Serum Bank in Columbia (Missouri)/United States; no history of cancer at baseline; missing serum sample for analysis excluded Recruitment: 1987 to 1997 Outcome assessment: 1982 to 1983, 1989 Number of cases: • Breast cancer: 105 (male/female: 0/105) Case definition: incidence Years of follow‐up: median: 2.7 Type of selenium marker: serum |
|
| Interventions | d.n.a. | |
| Outcomes | Statistical methods: conditional logistical regression Variables controlled in analysis: serum cholesterol, packs of cigarettes/d, BMI Variables controlled by matching: age, year and month of sample collection, diagnosis of benign breast disease within 2 years before study enrolment, "sequence number of blood draw" for women who donate blood more than once | |
| Risk estimates [95% CI] |
Reference category: lowest quartile Results: Breast cancer • Highest quartile: OR 0.9 (95% CI 0.4 to 1.8) |
|
| Selenium levels in exposure categories | Lowest quartile: ≤ 112.9 µg/L Highest quartile: 131.9 to 156.3 µg/L | |
| Notes | ||
Dreno 2007.
| Methods | Multi‐centre, randomised, placebo‐controlled, parallel‐group trial Allocation: random Sequence generation: unclear Concealment: unclear Blinding: described only as double‐blinded Dropouts/withdrawals: During treatment phase, 38 in the selenium group and 37 in the placebo group withdrew from the study. This distribution was similar in both treatment groups. Intention‐to‐treat‐analysis: unclear Recruitment period: not specified Treatment duration: 3 years Observation period/dermatological follow‐up: Participants were followed for 2 years longer after treatment. Detection of cases: Participants were seen by a dermatologist before grafting; any participants presenting with a non‐malignant or malignant skin keratosis or viral warts that had been present for less than 3 months were not selected. Within 10 weeks following the graft, a second visit was performed by a dermatologist to check that no new cutaneous lesion had appeared. Informed consent: The protocol and the consent form had been approved by a National Ethics Committee before the start of the study. Written informed consent was mandatory. |
|
| Participants |
Participants: 184 (randomised to selenium 200 μg/d: 91; to placebo: 93) Condition: organ transplant recipient population Demographics: mean age 44.3 ± SD 13 years (selenium 200 μg/d), 44.4 ± 10.7 years (placebo) |
|
| Interventions |
Intervention: • 200 µg/d selenium supplied as selenium yeast Control: placebo |
|
| Outcomes |
Primary outcome measure: • Occurrence rates of warts and various keratoses Other reported outcomes: • Skin cancers |
|
| Risk estimates [95% CI] |
Primary outcome: Events in selenium group = 33 (36.3%), events in placebo group = 31 (33.3%); odds ratio 1.09, P = 0.72 Secondary outcome: Events in selenium group = 6 (6.6%), events in placebo group = 2 (2.2%); odds ratio 3.08, P = 0.15 |
|
| Selenium levels in exposure categories | ||
| Notes | ||
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Random sequence generation (selection bias) | Low risk | Multi‐centre randomised |
| Allocation concealment (selection bias) | Unclear risk | Not stated |
| Blinding (performance bias and detection bias) All outcomes | Unclear risk | Described only as double‐blinded |
| Selective reporting (reporting bias) | Low risk | No problems found |
Epplein 2009.
| Methods | Matched, nested case‐control study (Epplein 2009; Gill 2009) Country: United States |
|
| Participants |
Inclusion criteria: participants from the Multiethnic Cohort, aged 45 to 75 years (native Hawaiians: aged 42 years and older), blood sample provided before cancer diagnosis between 1997 and 2006 Name of parent cohort: Multiethnic Cohort Recruitment: 1993 to 1996 Case definition: incidence Type of selenium marker: serum Epplein 2009: Participants: 67,594 (male: 29,009/female: 38,585) men and women Outcome assessment: 2006 Number of cases: • Lung cancer: 207 (male/female: 136/71) Years of follow‐up: 0 to 10 Gill 2009: Participants: 29,009 men Outcome assessment: n.r. Number of cases: • Prostate cancer: 467 (male/female: 467/0) Years of follow‐up: n.r. |
|
| Interventions | d.n.a. | |
| Outcomes |
Statistical methods: conditional logistical regression Epplein 2009: Variables controlled in analysis: age, fasting hours, pack‐years, pack‐years squared, years of schooling, family history of lung cancer Variables controlled by matching: age, sex, race/ethnicity, date of sample collection, time of day of sample collection, fasting status, smoking Gill 2009: Analysed cases: 450 of 467 Variables controlled in analysis: age, fasting hours, BMI, family history of prostate cancer, education Variables controlled by matching: age, race/ethnicity, date of sample collection, geographic site (California, Hawaii), time of day of sample collection, fasting status |
|
| Risk estimates [95% CI] |
Epplein 2009:
Reference category: lowest tertile Results: Lung cancer Male: Highest tertile: OR 0.70 (95% CI 0.37 to 1.33) Female: Highest tertile: OR 0.98 (95% CI 0.42 to 2.29) Gill 2009: Reference category: lowest quartile Results: Prostate cancer Highest quartile: OR 0.82 (95% CI 0.59 to 1.14) |
|
| Selenium levels in exposure categories |
Epplein 2009:
Lowest tertile: median 0.12 µg/g of sodium
Highest tertile: median 0.15 µg/g of sodium Gill 2009: Lowest quartile: median 0.12 µg/g Highest quartile: median 0.16 µg/g |
|
| Notes | Primary publication:Epplein 2009 Other publications: Gill 2009 | |
Fex 1987.
| Methods | Matched, nested case‐control study Country: Sweden |
|
| Participants |
Participants: 7935 men
Inclusion criteria: 46 to 48 years of age; residents of Malmo/Sweden; no restriction regarding malignant disease at baseline (11 of 35 with diagnosis of cancer at baseline screening examination and/or died during first year of follow‐up)
Name of parent cohort: Malmo Preventive Programme Recruitment: 1975 to 1979 Outcome assessment: June 1981 Number of cases: • Any cancer: 35 (male/female: 35/0) Case definition: mortality Years of follow‐up: 3.5 to 8.0 Type of selenium marker: plasma |
|
| Interventions | d.n.a. | |
| Outcomes | Analysed cases: 35 of 61 (reason for non‐inclusion: no plasma sample available) Statistical methods: logistical regression, Mantel‐Haenszel Variables controlled by matching: age, month of sample collection | |
| Risk estimates [95% CI] |
Reference category: highest quintile Results: Any cancer Male: lowest quintiles: OR 3.8 (CI not reported) |
|
| Selenium levels in exposure categories | n.r. | |
| Notes | CI and number of cases not reported | |
Fujishima 2011.
| Methods | Prospective cohort study Country: northern part of Japan |
|
| Participants |
Participants: 1041 men and women
Inclusion criteria: adult haemodialysis patients
Name of parent cohort: "Kaleidoscopic Approaches to Patients with End‐stage RENal Disease Study" (the KAREN Study) Recruitment: June 2003 to March 2004 Number of cases: • Malignant disease‐related death: 17 Case definition: mortality Years of follow‐up: 5 Type of selenium marker: serum |
|
| Interventions | d.n.a. | |
| Outcomes | Statistical methods: Cox logistical regression Variables controlled by matching: age, male gender, BMI, hypertension, dyslipidaemia, diabetes mellitus, serum albumin levels, high‐sensitivity CRP levels, history of myocardial infarction, history of stroke, history of malignant disease, smoking status, regular drinking habit | |
| Risk estimates [95% CI] |
Reference category: lowest quartile Results: Malignant disease‐related death • Highest quartile: HR 2.98 (95% CI 0.62 to 14.35) |
|
| Selenium levels in exposure categories | Lowest quartile: 18.4 to 85.3 µg/L Highest quartile: 114.2 to 226.2 µg/L | |
| Notes | ||
Garland 1995.
| Methods | Matched, nested case‐control study Country: United States |
|
| Participants |
Participants: 62,641 women
Inclusion criteria: female registered nurses in 11 USA states; aged 30 to 55 years at baseline; completed questionnaire in 1976 and provided toenail sample in 1982; no history of cancer at baseline
Name of parent cohort: Nurses' Health Study (NHS) Recruitment: 1976 (toenail sample collection in 1982) Outcome assessment: 1 June 1986 Garland 1995: Number of cases: • Any cancer (without breast): 503 (male/female: 0/503) • Colon and rectal cancer: 89 (male/female: 0/89) • Melanoma: 63 (male/female: 0/63) • Ovarian cancer: 58 (male/female: 0/58) • Lung cancer: 47 (male/female: 0/47) • Other: 155 (male/female: 0/155) • Uterine cancer: 91 (male/female: 0/91) Hunter 1990: Number of cases: • Breast cancer: 434 (0/434) Case definition: incidence Years of follow‐up: 2.0 to 4.4 Type of selenium marker: toenail |
|
| Interventions | d.n.a. | |
| Outcomes | Statistical methods: logistical regression, conditional logistical regression Variables controlled in analysis: smoking status Variables controlled by matching: age, year and month of sample collection Hunter 1990 additionally controlled in analysis for age at first birth, age at menarche, alcohol use, history of benign breast disease, menopausal status, maternal breast cancer, breast cancer in sister(s), oral contraceptive use, parity, relative weight | |
| Risk estimates [95% CI] |
Reference category: lowest quintile, lowest tertile Results: Garland 1995: Any cancer (without breast) • Female: highest quintile: OR 1.44 (95% CI 0.97 to 2.13) Colon and rectal cancer • Female: highest tertile: OR 2.04 (95% CI 0.88 to 4.75) Melanoma • Female: highest tertile: OR 1.66 (95% CI 0.71 to 3.85) Ovarian cancer • Female: highest tertile: OR 1.22 (95% CI 0.44 to 3.38) Lung cancer • Female: highest tertile: OR 4.33 (95% CI 0.54 to 34.60) Other cancer • Female: highest tertile: OR 0.97 (95% CI 0.55 to 1.71) Uterine cancer • Female: highest tertile: OR 1.38 (95% CI 0.62 to 3.08) Hunter 1990: Breast cancer • Female: highest quintile: OR 1.10 (95% CI 0.70 to 1.72) |
|
| Selenium levels in exposure categories |
Garland 1995:
• Lowest quintile: ≤ 0.71 µg/g
• Highest quintile: ≥ 0.95 µg/g Hunter 1990: • Lowest quintile: ≤ 0.705 µg/g • Highest quintile: ≥ 0.906 µg/g |
|
| Notes | Primary publication:Garland 1995 Other publication:Hunter 1990 | |
Glattre 1989.
| Methods | Matched, nested case‐control study Country: Norway |
|
| Participants |
Participants: 100,000 men and women
Inclusion criteria: serum available at Janus serum bank (Norwegian serum bank, which is consolidated from several sources and is maintained by the Norwegian Cancer Society for research purposes) Recruitment: 1972 to 1985 Outcome assessment: end of 1985 Number of cases: • Thyroid cancer: 43 (male/female: 12/31) Case definition: incidence Years of follow‐up: 0.0 to 14.0 Type of selenium marker: serum |
|
| Interventions | d.n.a. | |
| Outcomes | Statistical methods: conditional logistical regression Variables controlled by matching: age, gender, year of sample collection, county of residence | |
| Risk estimates [95% CI] |
Reference category: highest tertile Results: Thyroid cancer • Both genders: lowest tertiles: OR 7.7 (95% CI 1.3 to 44.7) • Men: lowest tertiles: OR 6.5 (95% CI 0.2 to 201.9) • Women: lowest tertiles: OR 8.3 (95% CI 0.9 to 78.5) |
|
| Selenium levels in exposure categories | Lowest tertile: ≤ 98.7 µg/L Highest tertile: ≥ 130.3 µg/L | |
| Notes | ||
Goodman 2001.
| Methods | Matched, nested case‐control study Country: United States |
|
| Participants |
Participants: 18,314 (male/female: 12,025/6289)
Inclusion criteria: 4060 male asbestos workers: 45 to 74 years of age; 14,254 (male/female: 7965/6289) smokers > 20 pack‐years: 50 to 69 years of age; cohort of an RCT for lung cancer prevention in high‐risk populations
Name of parent cohort: Caret (Carotene and Retinol Efficacy Trial) Recruitment: 1988 to 1994 Outcome assessment: April 1999 Number of cases: • Lung cancer: 235 (male/female: n.r.) • Prostate cancer: 356 (male/female: 356/0) Case definition: incidence Years of follow‐up: 6.0 to 12.0 Type of selenium marker: serum |
|
| Interventions | d.n.a. | |
| Outcomes |
Analysed cases: 235 of 236 prostate cancer cases analysed (reason for non‐inclusion: no sample available for analysis or no control available); 356 of 385 lung cancer cases analysed (reason for non‐inclusion: missing selenium values for case‐control pairs) Statistical methods: conditional logistical regression Variables controlled by matching: age, smoking status at randomisation, year of randomisation, year of sample collection, treatment arm, exposure population |
|
| Risk estimates [95% CI] |
Reference category: lowest quartile Results: Lung cancer • Both genders: highest quartile: OR 1.20 (95% CI 0.77 to 1.88) • Male: highest quartile: OR 1.53 (95% CI 0.83 to 2.82) • Female: highest quartile: OR 0.76 (95% CI 0.29 to 2.01) Prostate cancer • Highest quartile: OR 1.02 (95% CI 0.65 to 1.60) |
|
| Selenium levels in exposure categories |
Lung cancer
• Lowest quartile: 63.9 to 105.5 µg/L
• Highest quartile: 129.4 to 172.3 µg/L Prostate cancer • Lowest quartile: 50.7 to 101.2 µg/L • Highest quartile: 126.0 to 219.6 µg/L |
|
| Notes | ||
Goyal 2013.
| Methods | Cohort study Country: United States |
|
| Participants |
Name of parent cohort: Third National Health and Nutrition Examination Survey (NHANES III) Inclusion criteria: male and female adults, aged 20 to 90 years, participating in the NHANES III: "stratified, multistage probability cluster to provide data representing the noninstitutionalized US population" (Bleys 2008, p. 404) Recruitment: 1988 to 1994 Participants: 13,887 men and women Outcome assessment: 31 December 2006 Number of cases: • Cancer deaths: 891 (male/female: n.r.) Case definition: mortality Years of follow‐up: 14.2 Type of selenium marker: serum |
|
| Interventions | d.n.a. | |
| Outcomes |
Analysed cases: 864 (male/female: n.r.) Statistical methods: Cox proportional hazard regression Variables controlled in analysis: age, sex, race‐ethnicity, level of education, annual family income, body mass index, smoking status, serum cotinine level, alcohol consumption, fruit and vegetable intake, physical activity, serum total cholesterol levels, hypertension status, diabetes status, history of heart attack, congestive heart failure, stroke or cancer, hormone use in women, supplement use, serum levels of other micronutrients in the study (analysis only for both genders) |
|
| Risk estimates [95% CI] |
Reference category: lowest tertile Results: Cancer deaths • Both genders: highest quintile: HR 0.84 (95% CI 0.61 to 1.17) • Male: highest quintile: HR 0.67 (95% CI 0.54 to 0.83) • Female: highest quintile: HR 0.91 (95% CI 0.71 to 1.16) |
|
| Selenium levels in exposure categories | Lowest quintile: ≤ 108.96 µg/L Highest quintile: ≥ 136.60 µg/L |
|
| Notes | Second report on the same cohort of Bleys 2008; results updated | |
Graff 2017.
| Methods | Matched, nested case‐control study Country: United States |
|
| Participants |
Name of the parent cohort: Health Professional Follow‐up Study Participants: 18,259 men Inclusion criteria: patients free from prostate cancer between 1993 and 1995 who returned EDTA‐preserved blood samples from HPFS cohort (35% of total cohort) Recruitment: 1986 Outcome assessment: 31 January 1998 Number of cases: • Prostate cancer: 166 (male/female: 166/0) Case definition: incidence Years of follow‐up: up to 5 Type of selenium marker: plasma |
|
| Interventions | d.n.a. | |
| Outcomes |
Analysesd cases: 154 Statistical methods: conditional logistical regression model Variables controlled in analysis: age at blood draw, smoking status at blood draw, every PSA test before blood draw, timing and season of blood draw, time between blood draw and index date Variables controlled by matching: year of birth, PSA test before blood draw, timing, season and year of blood draw |
|
| Risk estimates [95% CI] |
Reference category: lowest quartile Results: • Highest quartile: 1.57 (95% CI 0.92 to 2.69) |
|
| Selenium levels in exposure categories | Lowest quartile: 0.0894 ppm Highest quartile: 0.1308 ppm |
|
| Notes | Exposure category cutpoints provided by trial author | |
Grundmark 2011.
| Methods | Cohort study Country: Sweden |
|
| Participants |
Participants: 2322 males
Inclusion criteria: male residents in Uppsala county in January 1970, born from 1920 to 1924
Name of parent cohort: Uppsala Longitudinal Study of Adult Men (ULSAM) Recruitment: 1991 to 1995 Outcome assessment: 31 December 2003 Number of cases: • Prostate cancer: 208 Case definition: incidence Years of follow‐up: median: 26.5 Type of selenium marker: serum |
|
| Interventions | d.n.a. | |
| Outcomes | Statistical methods: proportional hazard model | |
| Risk estimates [95% CI] |
Reference category: lowest level Results: Prostate cancer • Highest level: RR 0.83 (95% CI 0.60 to 1.16) |
|
| Selenium levels in exposure categories | Lowest level: ≤ 70 µg/L Highest level: > 81 µg/L | |
| Notes | ||
Han 2013.
| Methods | Cohort study Country: United States |
|
| Participants |
Name of parent cohort: Vitamins and Lifestyles (VITAL) study Participants: 70,332 men and women Inclusion criteria: aged 50 to 76 years, participants recruited from subscribers to commercial mailing list, residents of western Washington state, no malignant disease at baseline, no (or missing) history of pancreatic cancer or neuroendocrine tumours Recruitment: 1 October 2000 to 31 December 2002 Outcome assessment: 31 December 2008 Number of cases: • Pancreatic cancer: 195 (male/female: n.r.); 184 adenocarcinoma pancreatic cancer and 11 neuroendocrine tumours Case definition: incidence Years of follow‐up: median: 7.1 Type of selenium marker: intake and supplement use (questionnaire: use of supplements over the past 10 years, mean supplemental intake/d calculated) |
|
| Interventions | d.n.a. | |
| Outcomes |
Analysed cases: Individuals with neuroendocrine tumours were excluded.
Daily intake: 162 out of 184 cases analysed (reason for exclusion: dietary questionnaire incomplete or implausible total energy intake)
Diet and 10‐year supplement use: 158 out of 184 cases analysed (reason for exclusion: dietary questionnaire incomplete or implausible total energy intake and missing supplement use) Statistical methods: Cox proportional hazard model Variables controlled in analysis: age, sex, ethnicity, education, body mass index, physical activity, cigarette smoking status, total alcohol consumption, family history of pancreatic cancer, history of diabetes, total energy intake |
|
| Risk estimates [95% CI] |
Reference category: lowest tertile Results: Adenocarcinoma pancreatic cancer • Daily intake: HR 0.44 (95% CI 0.23 to 0.85) • Diet and 10‐year supplement use: HR 0.69 (95% CI 0.39 to 1.20) |
|
| Selenium levels in exposure categories | Daily intake • Lowest tertile: 6.38 to 85.49 µg/d • Highest tertile: 127.50 to 641.60 µg/d Diet and 10‐year supplement use • Lowest tertile: 9.81 to 98.76 µg/d • Highest tertile: 145.66 to 646.60 µg/d | |
| Notes | ||
Hansen 2013.
| Methods | Cohort study Country: Denmark |
|
| Participants |
Participants: 54,208 men and women Inclusion criteria: aged 50 to 64, born in Denmark, no diagnosis of cancer registered in the Danish Cancer Registry, living in the Copenhagen, Frederiksberg Aarhus municipalities, Hinnerup or Hørning municipalities in Aarhus County, and nearly all in Copenhagen county Recruitment: 1993 to 1997 Outcome assessment: April 1995 to December 2009 Number of cases: 990 (male/female: n.r) Case definition: incidence Years of follow‐up: median: 13 Type of selenium marker: supplement use |
|
| Interventions | d.n.a. | |
| Outcomes |
Analysed cases: • Colon‐rectal cancer: 990 (male/female: n.r.) • Colon cancer: 642 (male/female: n.r.) • Rectal cancer: 348 (male/female: n.r.) Statistical methods: Cox proportional hazard model Variables controlled in analysis: alcohol consumption, smoking status (ever/never), physical activity at work and at leisure, non‐steroidal anti‐inflammatory drug use, body mass index, education level, intake of red and processed meat, dietary intake, supplemental intake of nutrients alternatively |
|
| Risk estimates [95% CI] |
Reference category: high use Results: • Colon‐rectal cancer: HR 1.25 (95% CI 1.05 to 1.48) • Colon cancer: HR 1.22 (95% CI 0.99 to 1.51) • Rectal cancer: HR 1.29 (95% CI 0.96 to 1.74) |
|
| Selenium levels in exposure categories | Supplement use: • Never use: 0 µg/d • High use: > 45.80 µg/d | |
| Notes | Data on dietary intake and Total intake + supplement use not reported according to inclusion criteria: only 2 categories ‐ high vs low use | |
Hartman 1998.
| Methods | Cohort/subcohort controlled cohort study Country: Finland |
|
| Participants |
Participants: 29,133 men
Inclusion criteria: 50 to 69 years of age; smokers; no history of cancer (other than non‐melanoma skin cancer) at baseline; no severe physical or psychiatric illness; intake of vitamin E/A/beta‐carotene supplements in excess of defined amounts
Name of parent cohort: Alpha‐Tocopherol, Beta‐Carotene Cancer Prevention (ATBC) Study Recruitment: 1985 to 1988 Outcome assessment: 30 April 1993 Number of cases: • Prostate cancer: 302 (male/female: 302/0) Case definition: incidence Years of follow‐up: 5.0 to 8.0 Type of selenium marker: intake |
|
| Interventions | d.n.a. | |
| Outcomes | Analysed cases: 302 of 317 (reason for non‐inclusion: no dietary information available) • Analysis stratified by randomisation status according to active interventions or placebo interventions in the RCT • Results reported separately for total selenium intake and non‐supplemental selenium intake Statistical methods: Cox regression analysis Variables controlled in analysis: age, living in urban area, beta‐carotene intervention, total energy, BPH | |
| Risk estimates [95% CI] |
Reference category: lowest quartile Results: Prostate cancer Total (nutritional and supplemental) selenium intake in participants without active alpha‐tocopherol intervention • Highest quartile: RR 1.27 (95% CI 0.70 to 2.20) Total (nutritional and supplemental) selenium intake in participants with alpha‐tocopherol intervention • Highest quartile: RR 0.84 (95% CI 0.43 to 1.67) Nutritional selenium intake in participants without active alpha‐tocopherol intervention • Highest quartile: RR 1.32 (95% CI 0.70 to 2.47) Nutritional selenium intake in participants with alpha‐tocopherol intervention • Highest quartile: RR 0.72 (95% CI 0.33 to 1.55) |
|
| Selenium levels in exposure categories | Total nutritional and supplemental selenium intake:
• Lowest quartile: ≤ 71.51 µg/d
• Highest quartile: ≥ 111.06 µg/d Nutritional selenium intake: • Lowest quartile: ≤ 70.10 µg/d • Highest quartile: ≥ 105.65 µg/d |
|
| Notes | ||
Hashemian 2015.
| Methods | Cohort study Country: Iran |
|
| Participants |
Name of parent cohort: Golestan Cohort Study Participants: 47,405 (male/female: 19,969/27,436) Inclusion criteria: aged 40 to 75, stable residents in Golestan region (Gonbad City and villages in Gonbad, Kalaleh, and Aq‐Qala counties); not having a current or previous diagnosis of upper gastrointestinal cancer Recruitment: 2004 to 2008 Outcome assessment: 2014 Number of cases: • Oesophageal squamous cell carcinoma: 201 (male/female: n.r.) Case definition: incidence Years of follow‐up: median: 7.2 Type of selenium marker: intake |
|
| Interventions | d.n.a. | |
| Outcomes |
Analysed cases: 201 (male/female: n.r.) Statistical methods: Cox proportional hazard model Variables controlled in analysis: age, sex, total energy, place of residence (urban or rural), smoking (never or ever), wealth score (low, medium, or high), ethnicity (non‐Turkmen or Turkmen), opiate use (never or ever), BMI, education (illiterate or formal education), marital status (single or married), physical activity score (continuous), fruit and vegetable intake |
|
| Risk estimates [95% CI] |
Reference category: lowest quartile Results: Oesophageal squamous cell carcinoma • Highest quartile: HR 0.67 (95% CI 0.53 to 1.30) |
|
| Selenium levels in exposure categories | Lowest quartile: < 116 µg/d Highest quartile: > 175 µg/d | |
| Notes | ||
Helzlsouer 2000.
| Methods | Matched, nested case‐control study Country: United States |
|
| Participants |
Participants: 10,456 men
Inclusion criteria: residents of Washington county; cases with second malignancy or missing pathological confirmation excluded
Name of parent cohort: CLUE II Cohort Recruitment: 1989 Outcome assessment: September 1996 Number of cases: • Prostate cancer: 117 (male/female: 117/0) Case definition: incidence Years of follow‐up: 6.8 to 7.8 Type of selenium marker: toenail |
|
| Interventions | d.n.a. | |
| Outcomes | Analysed cases: 117 of 145 (reason for non‐inclusion: no toenail clipping available) Statistical methods: conditional logistical regression Variables controlled in analysis: BMI at age 21, education, hours since last meal Variables controlled by matching: age, race/ethnicity, year and month of sample collection, size of toenail clipping | |
| Risk estimates [95% CI] |
Reference category: lowest quintile Results: Prostate cancer • Highest quintile: OR 0.38 (95% CI 0.17 to 0.85) |
|
| Selenium levels in exposure categories | Lowest quintile: ≤ 0.69 ppm Highest quintile: ≥ 0.92 ppm | |
| Notes | ||
Hotaling 2011.
| Methods | Cohort study Country: United States |
|
| Participants |
Participants: 77,050 men and women, aged 50 to 76 years, participants recruited from subscribers to commercial mailing list, residents of western Washington state, non‐whites excluded, no malignant disease at baseline Name of parent cohort: Vitamins and Lifestyle (VITAL) study Recruitment: 1 October 2000 to 31 December 2002 Outcome assessment: 31 December 2002 Number of cases: • Urothelial carcinoma: 330 Case definition: incidence Years of follow‐up: median: 6 Type of selenium marker: supplemental intake (questionnaire: use of supplements over the past 10 years, mean supplemental intake/day calculated) |
|
| Interventions | d.n.a. | |
| Outcomes | Statistical methods: Cox proportional hazard regression Variables controlled in analysis: age, gender, race (white, black, other), education, family history of bladder cancer, smoking (never; former, quit more than 10 years before start of VITAL; former, quit less than 10 years before start of VITAL; current), pack‐years (never‐smoker and tertiles), fruit and vegetable intake | |
| Risk estimates [95% CI] |
Reference category: non‐use Results: • Highest level: HR 0.97 (95% CI 0.72 to 1.31) |
|
| Selenium levels in exposure categories | Lowest level: non‐use Highest quartile: 20 µg/d | |
| Notes | ||
Hughes 2015.
| Methods | Matched, nested case‐control study Countries: Denmark, France, Germany, Greece, Italy, the Nederlands, Spain, United Kindom |
|
| Participants |
Name of parent cohort: European Prospective Investigation into Cancer and Nutrition (EPIC) Participants: 428,917 (male/female: 129,961/298,956) Inclusion criteria: aged 25 to 70, participants from the EPIC study Recruitment: 1992 to 2000 Outcome assessment: at each country’s study closure date (between June 2002 and 2003) Number of cases: • Colorectal cancer: 966 (male/female: 466/500) • Colon cancer: 598 (male/female: 272/326) • Rectal cancer: 368 (male/female: 194/174) Case definition: incidence Years of follow‐up: average: approximately 4 Type of selenium marker: serum |
|
| Interventions | d.n.a. | |
| Outcomes |
Statistical methods: conditional logistical regression Variables controlled in analysis: smoking status/duration/intensity, BMI, total physical activity, education level, total dietary energy consumption, intake of total calcium, fruits, vegetables, red and processed meats, and alcohol Variables controlled by matching: study centre of enrolment, sex, age at blood collection, time of blood collection and fasting status; among women, the following: menopausal status. Premenopausal women were matched on phase of menstrual cycle, and postmenopausal women were matched on current hormonal replacement therapy (HRT) use. |
|
| Risk estimates [95% CI] |
Reference category: lowest quintile Results: Colorectal cancer • Both genders: highest quintile: IRR 0.88 (95% CI 0.64 to 1.21) • Male: highest quintile: IRR 1.18 (95% CI 0.73 to 1.90) • Female: highest quintile: IRR 0.64 (95% CI 0.40 to 1.01) Colon cancer • Both genders: highest quintile: IRR 0.81 (95% CI 0.54 to 1.23) • Male: highest quintile: IRR 1.11 (95% CI 0.58 to 2.12) • Female: highest quintile: IRR 0.61 (95% CI 0.34 to 1.09) Rectal cancer • Both genders: highest quintile: IRR 1.09 (95% CI 0.63 to 1.89) • Male: highest quintile: IRR 1.32 (95% CI 0.55 to 3.19) • Female: highest quintile: IRR 0.76 (95% CI 0.32 to 1.80) |
|
| Selenium levels in exposure categories |
Both male and female • Lowest quintile: < 67.7 µg/L • Highest quintile: > 100.6 µg/L |
|
| Notes | Data for study population from Riboli 2002 | |
Hughes 2016.
| Methods | Matched, nested case‐control study Countries: Denmark, France, Germany, Greece, Italy, the Nederlands, Norway, Spain, Sweden, UK |
|
| Participants |
Name of parent cohort: European Prospective Investigation into Cancer and Nutrition (EPIC) Participants: 521,448 Inclusion criteria: aged 25 to 70 participants of the EPIC study Recruitment: 1992 to 2000 Outcome assessment: at each country’s study closure date (between December 2002 and December 2006) Number of cases: 261 (male/female: n.r.) Case definition: incidence Years of follow‐up: average: approximately 6 Type of selenium marker: serum |
|
| Interventions | d.n.a. | |
| Outcomes |
Analysed cases:
• Hepatocellular carcinoma (HCC): 106 (male/female: n.r)
• Gallbladder and biliary tract cancer (GBTC): 96 (male/female: n.r)
• Intrahepatic bile duct cancer (IHBC): 36 (male/female: n.r) Statistical methods: conditional logistical regression Variables controlled in analysis: BMI, waist circumference, baseline alcohol intake, physical activity (metabolic equivalent tasks), smoking status, education, alcohol intake pattern, self‐reported diabetes, total energy intake Variables controlled by matching: age at blood collection, sex, study centre, time of day, fasting status at blood collection. Additionally, women were matched by menopausal status and hormone replacement therapy use at the time of blood collection. |
|
| Risk estimates [95% CI] |
Reference category: lowest tertile Results • HCC: highest tertile: OR 0.41 (95% CI 0.23 to 0.72) • GBTCs: highest tertile: OR 0.74 (95% CI 0.47 to 1.18) |
|
| Selenium levels in exposure categories | Lowest tertile: ≤ 80.5 µg/L Highest tertile: ≥ 64.5 µg/L 20 µg/L increase | |
| Notes | Estimates for IHBC reported only for 20 µg/L increase: OR 0.42 (95% CI 0.15 to 1.20) | |
Kabuto 1994.
| Methods | Matched, nested case‐control study Country: Japan |
|
| Participants |
Participants: 20,000 men and women
Inclusion criteria: survivor of the atomic bomb in Hiroshima or Nagasaki; serum available for analysis
Name of parent cohort: Adult Health Study Hiroshima and Nagasaki Recruitment: 1960 (blood samples drawn in 1970 to 1972) Outcome assessment: 1983 Number of cases: • Stomach cancer: 201 (male/female: 113/88) • Lung cancer: 77 (male/female: 43/34) Case definition: incidence Years of follow‐up: 12.0 to 14.0 Type of selenium marker: serum |
|
| Interventions | d.n.a. | |
| Outcomes | Statistical methods: conditional logistical regression Variables controlled in analysis: radiation dose, smoking, age, gender Variables controlled by matching: age, gender, year/month of sample collection, city | |
| Risk estimates [95% CI] |
Reference category: highest quartile Results: Stomach cancer • Both genders: lowest quartile: OR 1.0 (95% CI 0.5 to 1.9) Lung cancer • Both genders: lowest quartile: OR 1.8 (95% CI 0.7 to 5.0) |
|
| Selenium levels in exposure categories | Lowest quartile ≤ 98.90 µg/L Highest quartile ≥ 128.10 µg/L | |
| Notes | ||
Karagas 1997.
| Methods | Matched, nested case‐control study Country: United States |
|
| Participants |
Participants: 1805 men and women
Inclusion criteria: at least 1 basal cell or squamous cell cancer before study entry; participants in an RCT for non‐melanoma skin cancer prevention with oral beta‐carotene supplementation
Name of parent cohort: Skin Cancer Prevention Study Recruitment: February 1983 to February 1986 Outcome assessment: 30 September 1989 Number of cases: • Squamous cell cancer: 131 (89% male/11% female) Case definition: incidence Years of follow‐up: 3.0 to 5.0 Type of selenium marker: plasma |
|
| Interventions | d.n.a. | |
| Outcomes | Statistical methods: conditional logistical regression Variables controlled in analysis: cigarette smoking Variables controlled by matching: age, gender, study centre of RCT, time in study (diagnosis date) | |
| Risk estimates [95% CI] |
Reference category: lowest quartile Results: Squamous cell cancer Both genders: highest quartile: OR 0.86 (95% CI 0.47 to 1.58) |
|
| Selenium levels in exposure categories | Lowest quartile: ≤ 0.12 ppm Highest quartile: ≥ 0.14 ppm | |
| Notes | ||
Karp 2013.
| Methods | Randomised controlled trial Phase III Chemoprevention Trial of Selenium Supplementation In Persons With Resected Stage I Non‐Small Cell Lung Cancer: ECOG 5597 Allocation: random, permuted blocks stratified by smoking status (current, former, or never), sex, and stage (IA vs IB with other therapy vs IB without other therapy) Sequence generation: permuted blocks within strata with dynamic balancing Concealment: central assignments at ECOG Coordinating Center Blinding: participant blinded, doctor blinded, outcome assessor/pathologist unclear, review/coding of medical records unblinded Dropouts/withdrawals: of 1561 randomised participants, no dropouts Intention‐to‐treat‐analysis: yes Recruitment period: 6 October 2000 to 5 November 2009 End of study period: 5 November 2009 Treatment duration: • Intervention was discontinued on 5 November 2009, following the Data Monitoring Committee recommendation that the study could eventually show significant evidence of benefit Observation period: After end of treatment phase, participants enter the follow‐up phase. Analyses till June 2011 reported (until January 2014 in Pillai 2014 with median follow‐up of 5.6 years) Detection of cases: visit at 3 months for adverse effects, annual visit for other endpoints Informed consent: yes |
|
| Participants | 1561 male and female participants with completely resected stage I non‐small‐cell lung cancer Countries: United States, Canada Participants: 1561 (randomised to selenium group: 1,040; to placebo group: 521) Condition: adult participants, 6 to 36 months from complete resection of histologically proven stage IA or IB non‐small‐cell lung cancer, with chest X‐ray or CT scan ≤ 8 weeks before registration without sign of new recurrent lung cancer, no recurrent cancers or any other prior cancer history within the past 5 years (except NMSC), normal hepatic function, ECOG performance status of 0 or 1, not taking selenium supplement regularly ≥ 70 μg/d, any therapy (chemo, radio, or biological therapy) completed at least 6 months before study registration and all related symptoms subsided Demographics: median age 66 in both intervention groups. Selenium and placebo participants were well balanced with respect to sex, age, smoking history, and stage at resection. Recruitment and setting: not reported |
|
| Interventions |
Intervention: 200 µg selenised yeast Control: placebo |
|
| Outcomes | Primary outcome: incidence of second primary lung tumours Secondary outcomes: incidence of any other second primary tumours, mortality, overall survival Other outcomes: qualitative and quantitative toxicity of selenium | |
| Risk estimates [95% CI] |
Karp 2013: Primary outcome: • Lung cancer: RR 1.23 (95% CI 0.80 to 1.80) Other outcomes: • Any cancer: RR 1.02 (95% CI 0.80 to 1.21) • Prostate cancer: RR 0.89 (95% CI 0.40 to 2.00) • Colorectal cancer: RR 0.50 (95% CI 0.13 to 1.91) • Melanoma: RR 1.25 (95% CI 0.24 to 6.43) • NMSC: RR 0.80 (95% CI 0.44 to 1.45) • Diabete mellitus: RR 1.19 (95% CI 0.61 to 2.35) Pillai 2014: Primary outcome: • Lung cancer: RR 1.29 (95% CI 0.87 to 1.93) |
|
| Selenium levels in exposure categories | d.n.a. | |
| Notes |
Karp 2013 Adverse effects: • Alopecia grade 1 to 2: RR 0.80 (95% CI 0.48 to 1.34) • Dermatitis grade 1 to 2: RR 1.59 (95% CI 0.75 to 3.37) • Nail changes grade 1 to 2: RR 0.72 (95% CI 0.46 to 1.12) • Fatigue grade 1 to 2: RR 1.02 (95% CI 0.68 to 1.53) • Nausea grade 1 to 2: RR 2.14 (95% CI 1.04 to 4.42) |
|
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Random sequence generation (selection bias) | Low risk | Random, permuted blocks stratified |
| Allocation concealment (selection bias) | Low risk | Central assignments |
| Blinding (performance bias and detection bias) All outcomes | Low risk | Participants blinded and doctor blinded, outcome assessor/pathologist unclear, review/coding of medical records unblinded |
| Selective reporting (reporting bias) | Low risk | No problems found |
Knekt 1990.
| Methods | Matched, nested case‐control study (Hakama 1990; Knekt 1988;Knekt 1990;Knekt 1996)
Cohort study (Knekt 1991) Country: Finland |
|
| Participants |
Inclusion criteria: no history of cancer at baseline
Name of parent cohort: Social Insurance Institution's Mobile Clinic Health Examination Survey Recruitment: 1968 to 1972 Knekt 1990: Participants: 39,268: 21,172 men and 18,096 women Outcome assessment: 31 December 1980 Number of cases: • Any cancer: 1096 (male/female: 597/499) • Stomach cancer: 95 (male/female: 58/37) • Colon and rectal cancer: 91 (male/female: 32/59) • Lung cancer: 198 (male/female: 189/9) • Prostate cancer: 51 (male/female: 51/0) • Urinary tract cancer: 47 (male/female: 34/13) • Pancreatic cancer: 45 (male/female: 22/23) • Breast cancer: 90 (male/female: 0/90) • Gynaecological cancer (without breast): 86 (male/female: 0/86) • Basal cell carcinoma (skin): 126 (male/female: 64/62) • Other: 267 (male/female: 147/120) Hakama 1990: Participants: number of participants n.r.; both genders Inclusion criteria: aged 15 years and older Outcome assessment: 1977 Number of cases: • Any cancer: 766 (male/female: n.r.) • Lung cancer: 151 (male/female: 151/0) • Breast cancer: 67 (male/female: 0/67) • Stomach cancer: 76 (male/female: n.r.) • Prostate cancer: 37 (male/female: 37/0) Knekt 1988: Participants: 36,265: 21,172 men and 15,093 women Outcome assessment: 31 December 1977 Number of cases: • Oesophageal and stomach cancer: 86 (male/female: 51/35) • Colon and rectal cancer: 57 (male/female: 21/36) Knekt 1991: Participants: 4538 men Inclusion criteria: aged 20 to 69 years, with dietary history taken Outcome assessment: 1986 Number of cases: • Lung cancer: 117 (male/female: 117/0) Knekt 1996: Participants: 1896 women Outcome assessment: 1980 Number of cases: • Ovarian cancer: 24 (male/female: 0/24) Case definition: incidence Years of follow‐up: 9 to 20 years Type of selenium marker: serum (Hakama 1990; Knekt 1988;Knekt 1990;Knekt 1996), intake (Knekt 1991: dietary history) |
|
| Interventions | d.n.a. | |
| Outcomes |
Knekt 1990
Statistical methods: conditional logistical regression
Variables controlled in analysis: smoking
Variables additionally controlled in analysis of highest 4 quintiles vs lowest quintile: occupation, BMI, parity, cholesterol, haematocrit
Variables controlled by matching: age, gender, municipality, time of baseline examination, duration of storage of sample Hakama 1990 Analysed cases: 766 of 864 (reason for non‐inclusion: no serum sample) Statistical methods: conditional logistical regression Variables controlled in analysis: smoking Variables additionally controlled in analysis of highest 4 quintiles vs lowest quintile: retinol level, alpha‐tocopherol level Variables controlled by matching: age, gender, municipality, time of baseline examination, duration of storage of sample Knekt 1988 Statistical methods: n.r. Variables controlled in analysis: smoking, serum cholesterol Variables controlled by matching: age, gender, municipality, time of baseline examination, duration of storage of sample Knekt 1991 Statistical methods: Cox proportional hazard model Variables controlled in analysis: age, smoking (data stratified according to smoking status) Knekt 1996 Statistical methods: conditional logistical regression Variables controlled by matching: age, gender, municipality, time of baseline examination, duration of storage of sample |
|
| Risk estimates [95% CI] |
Knekt 1990
Reference category: lowest quintile Results: Any cancer Male • Highest quintile: OR 0.41 (CI not reported) • Above 20th percentile: OR 0.67 (CI not reported); cases during first 2 years of follow‐up excluded: 476 cases: OR 0.65 (95% CI 0.48 to 0.89) Female • Highest quintile: OR 0.86 (CI not reported) • Above 20th percentile: OR 0.93 (CI not reported); cases during first 2 years of follow‐up excluded: 423 cases: OR 0.97 (95% CI 0.68 to 1.39) Stomach cancer Male • Highest quintile: OR 0.09 (CI not reported) • Above 20th percentile: OR 0.26 (CI not reported); cases during first 2 years of follow‐up excluded: 43 cases: OR 0.24 (95% CI 0.09 to 0.69) Female • Highest quintile: OR 0.27 (CI not reported) • Above 20th percentile: OR 0.59 (CI not reported); cases during first 2 years of follow‐up excluded: 30 cases: OR 0.48 (95% CI 0.14 to 1.66) Colon and rectal cancer Male • Highest quintile: OR 0.53 (CI not reported) • Above 20th percentile: OR 0.69 (CI not reported); cases during first 2 years of follow‐up excluded: 29 cases: OR 1.01 (95% CI 0.18 to 5.65) Female • Highest quintile: OR 0.80 (CI not reported) • Above 20th percentile: OR 1.26 (CI not reported); cases during first 2 years of follow‐up excluded: 48 cases: OR 1.10 (95% CI 0.42 to 2.92) Lung cancer Male • Highest quintile: OR 0.30 (CI not reported) • Above 20th percentile: OR 0.60 (CI not reported); cases during first 2 years of follow‐up excluded: 153 cases: OR 0.66 (95% CI 0.37 to 1.19) Female • Third highest quintile: OR 4.62 (CI not reported) (quintile 4 and 5 did not contain any cases) Prostate cancer • Highest quintile: OR 1.15 (CI not reported) • Above 20th percentile: OR 1.13 (CI not reported); cases during first 2 years of follow‐up excluded: 46 cases: OR 1.00 (95% CI 0.42 to 2.40) Urinary tract cancer Male • Highest quintile: OR 0.81 (CI not reported) • Above 20th percentile: OR 0.89 (CI not reported); cases during first 2 years of follow‐up excluded: 26 cases: OR 0.34 (95% CI 0.06 to 2.06) Female • Highest quintile: OR 4.12 (CI not reported) • Above 20th percentile: not reported; cases during first 2 years of follow‐up excluded: 9 cases: OR 2.51 (95% CI 0.13 to 47.9) Pancreatic cancer Male • Fourth quintile vs lowest: OR 0.58 (CI not reported) (highest quintile did not contain any cases) • Above 20th percentile: OR 0.11 (CI not reported); cases during first 2 years of follow‐up excluded: not reported Female • Highest quintile: OR 3.49 (CI not reported) • Above 20th percentile: not reported; cases during first 2 years of follow‐up excluded: 22 cases: OR 0.86 (95% CI 0.21 to 3.52) Breast cancer • Highest quintile: OR 0.64 (CI not reported) • Above 20th percentile: OR 0.52 (CI not reported); cases during first 2 years of follow‐up excluded: 74 cases: OR 0.57 (95% CI 0.18 to 1.81) Gynaecological cancer (without breast) • Highest quintile: OR 0.96 (CI not reported) • Above 20th percentile: OR 0.91 (CI not reported); cases during first 2 years of follow‐up excluded: 70 cases: OR 1.03 (95% CI 0.43 to 2.50) Basal cell carcinoma (skin) Male • Highest quintile: OR 0.54 (CI not reported) • Above 20th percentile: OR 0.65 (CI not reported); cases during first 2 years of follow‐up excluded: 54 cases: OR 0.86 (95% CI 0.35 to 2.12) Female • Highest quintile: OR 1.55 (CI not reported) • Above 20th percentile: OR 1.73 (CI not reported); cases during first 2 years of follow‐up excluded: 52 cases: OR 1.54 (95% CI 0.64 to 3.73) Other or unspecified cancer: Male • Highest quintile: OR 0.42 (CI not reported) • Above 20th percentile: OR 0.72 (CI not reported); cases during first 2 years of follow‐up excluded: 110 cases: OR 0.70 (95% CI 0.36 to 1.36) Female • Highest quintile: OR 0.71 (CI not reported) • Above 20th percentile: OR 0.87 (CI not reported); cases during first 2 years of follow‐up excluded: 111 cases: OR 0.92 (95% CI 0.44 to 1.92) Hakama 1990 Reference category: highest quintile Results: Any cancer Male • Lowest quintile: OR 2.40 (CI not reported) • Lowest quintile vs 4 highest quintiles: OR 1.60 (CI not reported) Female • Lowest quintile: OR 1.20 (CI not reported) • Lowest quintile vs 4 highest quintiles:0.90 (CI not reported) Lung cancer Male: • Lowest quintile vs 4 highest quintiles: OR 1.80 (CI not reported) Breast cancer • Lowest quintile vs 4 highest quintiles: OR 3.10 (CI not reported) Stomach cancer Male • Lowest quintile vs 4 highest quintiles: OR 6.70 (CI not reported) Female • Lowest quintile vs 4 highest quintiles: OR 2.00 (CI not reported) Prostate cancer • Lowest quintile vs 4 highest quintiles: OR 0.80 (CI not reported) Knekt 1988 Reference category: highest quintile Results: Oesophageal and stomach cancer Male • Lowest tertile: OR 2.20 (CI not reported) • Lowest quintile vs 4 highest quintiles: OR 3.3 (95% CI 1.3 to 9.1) Female • Lowest tertile: OR 1.50 (CI not reported) • Lowest quintile vs 4 highest quintiles: OR 2.4 (95% CI 0.7 to 8.3) Colon and rectal cancer Male • Lowest tertile: OR 0.90 (CI not reported) • Lowest quintile vs 4 highest quintiles: OR 1.7 (95% CI 0.4 to 7.7) Female • Lowest tertile: OR 0.60 (CI not reported) • Lowest quintile vs 4 highest quintiles: OR 0.8 (95% CI 0.2 to 2.4) Knekt 1991 Reference category: highest tertile Results: Lung cancer Male non‐smokers • Lowest tertile: OR 1.03 (CI not reported) Male smokers • Lowest tertile: OR 0.83 (CI not reported) Knekt 1996 Reference category: highest tertile Results: Ovarian cancer • Lowest tertile: OR 1.15 (95% CI 0.19 to 4.06) |
|
| Selenium levels in exposure categories |
Knekt 1990
• Lowest quintile: ≤ 48.90 µg/L • Highest quintile ≥ 78.00 µg/L Hakama 1990 • Quintiles: not specified Knekt 1988 Both genders • Lowest tertile: ≤ 56.90 µg/L • Highest tertile ≥ 70.10 µg/L • Lowest quintile: ≤ 50 µg/L • Highest 4 quintiles > 50 µg/L Knekt 1991 • Tertiles: n.r. Knekt 1996 • Lowest tertile: ≤ 56.90 µg/L • Highest tertile: ≥ 68.10 µg/L |
|
| Notes | Primary publication:Knekt 1990 Other publications: Hakama 1990, Knekt 1988, Knekt 1991, Knekt 1996 | |
Knekt 1998.
| Methods | Matched, nested case‐control study Country: Finland |
|
| Participants |
Participants: 9101 men and women
Inclusion criteria: 19 years or older; no history of cancer at baseline; serum sample available for analysis
Name of parent cohort: Social Insurance Institution's Mobile Clinic Health Examination Survey Recruitment: 1973 to 1976 Outcome assessment: end of 1991 Number of cases: • Lung cancer: 91 (male/female: approximately 95%/5%) Case definition: incidence Years of follow‐up: 16.0 to 19.0 Type of selenium marker: serum |
|
| Interventions | d.n.a. | |
| Outcomes | Analysed cases: 91 of 95 (male/female: 90/5) Statistical methods: conditional logistical regression Variables controlled in analysis: smoking, alpha‐tocopherol, serum cholesterol, copper, orosomucoid, BMI Variables controlled by matching: age, gender, municipality, season of sample collection, length of storage of sample | |
| Risk estimates [95% CI] |
Reference category: lowest tertile Results: Lung cancer • Analysis adjusted for smoking only: both genders: highest tertiles: OR 0.44 (95% CI 0.21 to 0.89) • Analysis adjusted for all variables (number of cases: 77): highest tertiles: OR 0.41 (95% CI 0.17 to 0.94) |
|
| Selenium levels in exposure categories | Lowest tertile: ≤ 45.49 µg/L Highest tertile: ≥ 60.60 µg/L | |
| Notes | ||
Kok 1987a.
| Methods | Matched, nested case‐control study Country: the Netherlands |
|
| Participants |
Participants: 10,532 men and women
Inclusion criteria: inhabitant of Zoetermeer; 5 years or older
Name of parent cohort: EPOZ Cohort (Epidemiologisch onderzoek naar risico‐indicatoren voor hart‐ en vaatziekten) Recruitment: 1975 to 1978 Outcome assessment: 31 December 1983 Number of cases: • Any cancer: 69 (male/female: 40/29) Case definition: mortality Years of follow‐up: 6.0 to 9.0 Type of selenium marker: serum |
|
| Interventions | d.n.a. | |
| Outcomes | Analysed cases: 69 of 114 (reason for non‐inclusion: serum or baseline data not available, deaths in first year of follow‐up excluded) Statistical methods: not specified Variables controlled in analysis: age, smoking, serum cholesterol, serum vitamins A and E, systolic and diastolic blood pressure, BMI, week of blood collection, years of education, gender (in group of both genders) Variables controlled by matching: age, gender, smoking status | |
| Risk estimates [95% CI] |
Reference category: highest 4 quintiles Results: Any cancer • Both genders: lowest quintile: OR 1.9 (90% CI 1.0 to 3.5) • Male: lowest quintile: OR 2.7 (90% CI 1.2 to 6.2) • Female: lowest quintile: OR 1.5 (90% CI 0.5 to 4.5) |
|
| Selenium levels in exposure categories |
Both genders
• Lowest quintile: ≤ 102.79 µg/L
• Highest 4 quintiles: ≥ 102.80 µg/L Males • Lowest quintile: ≤ 100.79 µg/L • Highest 4 quintiles: ≥ 100.80 µg/L Females • Lowest quintile: ≤ 107.29 µg/L • Highest 4 quintiles: ≥ 107.30 µg/L |
|
| Notes | Primary publication:Kok 1987b Other publication:Kok 1987a | |
Kornitzer 2004.
| Methods | Matched, nested case‐control study Country: Belgium |
|
| Participants |
Participants: 10,902 (male/female: 5,549/5,353)
Inclusion criteria: 25 to 74 years of age
Name of parent cohort: Belgian Interuniversity Study on Nutrition and Health Recruitment: 1980 to 1984 Outcome assessment: n.r. Number of cases: • Any cancer: 193 (male/female: 143/50) Case definition: mortality Years of follow‐up: 10 Type of selenium marker: serum |
|
| Interventions | d.n.a. | |
| Outcomes | Analysed cases: 143 male/50 female cases analysed from 252 male/91 female cases (reason for non‐inclusion: no selenium measurement available) Statistical methods: not specified Variables controlled in analysis: BMI, total energy, total fat, saturated fat, alcohol intake, fibre, retinol, vitamin C, smoking, beta‐carotene Variables controlled by matching: age, gender | |
| Risk estimates [95% CI] |
Reference category: highest tertile Results: Any cancer • Male: lowest tertile: OR 2.2 (95% CI 1.3 to 3.7) • Female: lowest tertile: OR 0.7 (95% CI 0.3 to 1.6) |
|
| Selenium levels in exposure categories | Lowest tertile ≤ 72.00 µg/L Highest tertile ≥ 85.00 µg/L | |
| Notes | ||
Kristal 2014.
| Methods | Matched, nested case‐control study Countries: United States, Canada, Puerto Rico |
|
| Participants |
Name of parent cohort: SELECT (Selenium and Vitamin E Cancer Prevention Trial), placebo arm Participants: 777 men from placebo arm of SELECT study Inclusion criteria: black men aged ≥ 50 years and all other men aged ≥ 55 years, without history of prostate cancer, serum PSA level ≤ 4 ng/L and non‐suspicious digital rectal examination Recruitment: July 2001 to May 2004 Outcome assessment: 31 July 2009 Number of cases: • Prostate cancer: 404 (male/female:404/0) Case definition: incidence Years of follow‐up: n.r. Type of selenium marker: toenail |
|
| Interventions | d.n.a. | |
| Outcomes |
Analysed cases: 404 (male/female: 404/0) Statistical methods: Cox proportional hazard model Variables controlled in analysis: age and race by matching, family history of prostate cancer, diabetes, body mass index, prostate‐specific antigen Variables controlled by matching: age and race |
|
| Risk estimates [95% CI] |
Reference category: lowest quintile Results: • Prostate cancer: highest quintile: HR 0.76 (95% CI 0.44 to 1.31) |
|
| Selenium levels in exposure categories | Lowest quintile: < 0.758 µg/g Highest quintile > 1.003 µg/g | |
| Notes | ||
Kromhout 1987.
| Methods | Cohort/subcohort controlled cohort study Country: the Netherlands |
|
| Participants |
Participants: 878 men
Inclusion criteria: 40 to 59 years of age; random sample of general male population at specific age in Zutphen
Name of parent cohort: Zutphen Study Recruitment: 1960 Outcome assessment: 1985 Number of cases: • Lung cancer: 63 (male/female: 63/0) Case definition: mortality Years of follow‐up: 25 Type of selenium marker: intake (interview) |
|
| Interventions | d.n.a. | |
| Outcomes | Statistical methods: Cox proportional hazard model Variables controlled in analysis: age, pack‐years of smoking | |
| Risk estimates [95% CI] |
Reference category: lowest quartile Results: Lung cancer • Male: highest quartile: RR 0.98 (95% CI 0.41 to 2.36) |
|
| Selenium levels in exposure categories | Lowest quartile: ≤ 55.00 µg/d Highest quartile: ≥ 72.10 µg/d | |
| Notes | ||
Li 2000.
| Methods | Randomised controlled trial Allocation: randomised, "based on their residence area" Sequence generation: unclear, not described Concealment: unclear, not described Blinding: of participants: adequate (placebo); of investigators and doctors: unclear, not described Dropouts/withdrawals: no significant difference between percentages of dropouts in intervention and control group (absolute numbers not reported) Intention‐to‐treat‐analysis: unclear Recruitment period: unclear, not described Observation period: 3 years, started in 1996 Study period: unclear, not described Detection of cases: unclear; the study followed the diagnostic menu published by the National Cancer Control and Prevention Center, follow‐up procedures not described Informed consent: unclear, not described |
|
| Participants |
Country: China Number of participants: 2065 (selenium group: 1112; placebo group: 953) Condition: HBs‐Ag carriers with negative AFP and normal ALT living in Qidong, Jiangsu province Demographics: men only; aged 20 to 65 years (screening group) Recruitment and setting: recruitment of 2065 HBs‐Ag carriers from 17 villages out of a screening group of 18,000 men |
|
| Interventions |
Intervention: 0.5 mg sodium selenite p.o. daily for 3 years Control: placebo |
|
| Outcomes |
Primary outcome measure: incidence of primary liver cancer Other: blood selenium levels, activity of glutathione peroxidase Results: person‐year incidence rate (number of cases/total number of persons) in intervention and control groups: • 1st year of follow‐up: selenium group 899.25/100,000 (10/1112); placebo group: 1888.77/100,000 (18/953) • 2nd year of follow‐up: selenium group 1708.60/100,000 (19/1112); placebo group: 4302.20/100,000 (41/953) • 3rd year of follow‐up: selenium group 3057.55/100,000 (34/1112); placebo group: 5981.11/100,000 (57/953) |
|
| Risk estimates [95% CI] | n.r. | |
| Selenium levels in exposure categories | d.n.a. | |
| Notes | Adverse effects were not mentioned. | |
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Random sequence generation (selection bias) | Unclear risk | Randomisation based only on residential area |
| Allocation concealment (selection bias) | Unclear risk | Not described |
| Blinding (performance bias and detection bias) All outcomes | Low risk | Blinding of participants and doctors |
| Selective reporting (reporting bias) | Low risk | No problems found |
Li 2004a.
| Methods | Matched, nested case‐control study Country: United States |
|
| Participants |
Participants: 14,916 men
Inclusion criteria: participants of Physicians' Health Study who provided blood sample (healthy male physicians); no history of cancer at baseline; several physical conditions excluded at baseline: chronic renal failure, unstable angina pectoris, liver disease, peptic ulcer, history of TIA/stroke/myocardial infarction/gout; no use of vitamin A or beta‐carotene supplements
Name of parent cohort: Physicians' Health Study Recruitment: 1982 Outcome assessment: 1995 Number of cases: • Prostate cancer: 586 (male/female: 586/0) Case definition: incidence Years of follow‐up: 13 Type of selenium marker: plasma |
|
| Interventions | d.n.a. | |
| Outcomes | Statistical methods: logistical regression Variables controlled in analysis: age at baseline, smoking status, duration of follow‐up Variables controlled by matching: age, smoking status | |
| Risk estimates [95% CI] |
Reference category: lowest quintile Results: Prostate cancer • Highest quintile: OR 0.78 (95% CI 0.54 to 1.13) |
|
| Selenium levels in exposure categories | Lowest quintile: 0.060 to 0.090 ppm Highest quintile: 0.121 to 0.190 ppm | |
| Notes | ||
Lubinski 2011.
| Methods | Randomised controlled trial Allocation: random Sequence generation: unclear Concealment: unclear Blinding: described only as double‐blinded Dropouts/withdrawals: no description Intention‐to‐treat‐analysis: unclear Recruitment period: not specified Treatment duration: unclear Observation period/dermatological follow‐up: • Median: 35 months (range 6 to 62 months) Detection of cases: not described Informed consent: not described |
|
| Participants |
Country: Poland Number of participants: 1135 (randomised to selenium group: 563, to placebo group: 572) Condition: adult women, BRCA1+ mutation carriers Demographics: not reported Recruitment and setting: not reported |
|
| Interventions |
Intervention: • 250 µg/d selenium supplied as sodium selenite Control: • Placebo |
|
| Outcomes |
Case definition: incidence • All cancer • Primary breast cancer • Ovarian cancer |
|
| Risk estimates [95% CI] | All cancer: HR 1.4 (95% CI 0.9 to 2.0), cases: selenium 60, placebo 45 Primary breast cancer: HR 1.3 (95% CI 0.7 to 2.5), cases: selenium 38, placebo 29 Ovarian cancer: HR 1.3 (95% CI 0.6 to 2.7), cases: selenium 17, placebo 10 |
|
| Selenium levels in exposure categories | d.n.a | |
| Notes | ||
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Random sequence generation (selection bias) | Unclear risk | Described only as randomised trial |
| Allocation concealment (selection bias) | Unclear risk | Not stated |
| Blinding (performance bias and detection bias) All outcomes | Unclear risk | Described only as double‐blinded |
| Selective reporting (reporting bias) | Low risk | No problems found |
Ma 2017.
| Methods | Cohort study Country: China |
|
| Participants | Name of parent cohorts: Shangai Men's Health Study (SMHS) and Shangai Women's Health Study (SWHS) Participants: 133,957 (male/female: 61,470/74,941) SMHS: 61,480 men SWHS: 74,941 women Inclusion criteria: SMHS: men aged 40 to 74; residents in Shangai with no history of cancer SWHS: women aged 40 to 70, residents in Shangai with no history of cancer Recruitment: SMHS: April 2002 to June 2006 SWHS: March 1997 to May 2000 Outcome assessment: 31 December 2012 Number of cases: 536 (male/female: 344/192) Case definition: incidence Years of follow‐up: SMHS: median: 9.3 SWHS: median: 15.2 Type of selenium marker: intake |
|
| Interventions | d.n.a. | |
| Outcomes |
Analysed cases:
• Hepatocellular carcinoma: 536 (male/female: 344/192) Statistical methods: Cox proportional hazard model Variables controlled in analysis: • Both genders: sex, age at recruitment, body mass index, total physical activity, total intake of energy, vegetable, fruit, red meat, egg, fish, and soy, vitamin E intake, income, education, smoking history, alcohol consumption, family history of liver cancer, history of viral hepatitis/chronic liver disease, history of diabetes, history of cholelithiasis and history of cholecystectomy • Men: age at recruitment, body mass index, total physical activity, total intake of energy, vegetable, fruit, red meat, egg, fish, and soy, vitamin E intake, income, education, smoking history, alcohol consumption, family history of liver cancer, history of viral hepatitis/chronic liver disease, history of diabetes, history of cholelithiasis and history of cholecystectomy • Women: age at recruitment, body mass index, total physical activity, total intake of energy, vegetable, fruit, red meat, egg, fish, and soy, vitamin E intake, income, education, smoking history, alcohol consumption, family history of liver cancer, history of viral hepatitis/chronic liver disease, history of diabetes, history of cholelithiasis, history of cholecystectomy, menopausal status, ever had oral contraceptive |
|
| Risk estimates [95% CI] |
Reference category: lowest quintile Results: Hepatocellular carcinoma • Both cohorts: highest quintile: HR 0.86 (95% CI 0.52 to 1.43) SMHS: highest quintile: HR 0.95 (95% CI 0.51 to 1.76) SWHS: highest quintile: HR 0.70 (95% CI 0.26 to 1.90) |
|
| Selenium levels in exposure categories | SMHS: • Lowest quintile: < 31.77 µg/d • Highest quintile: ≥ 54.52 µg/d SWHS: • Lowest quintile: < 36.24 µg/d • Highest quintile: ≥ 61.14 µg/d | |
| Notes | ||
Marshall 2011.
| Methods | Randomised controlled trial Allocation: random Sequence generation: unclear Concealment: unclear Blinding: described only as double‐blinded. The central pathologist was also blinded to study assignment. Dropouts/withdrawals: 13/227 in the selenium arm and 12/225 in the placebo arm were lost to follow‐up. Intention‐to‐treat‐analysis: yes Recruitment period: not specified Treatment duration: not specified Observation period/dermatological follow‐up: • Participants were followed for 3 years. They were seen in clinic at baseline and every 6 months thereafter. Detection of cases: Tissue blocks and corresponding pathology reports for all prostate procedures were to be submitted to the central study pathologist for review. Informed consent: All participants gave oral and written informed consent in accordance with institutional and federal guidelines. The protocol was approved by the Institutional Review Boards at participating institutions, and was monitored by the Data and Safety Monitoring Committee of SWOG. |
|
| Participants |
Country: United States Participants: 452 (randomised to selenium 200 µg/d: 227; to placebo group: 225) Condition: 40 years of age or older; digital rectal examination; biopsy‐confirmed diagnosis of HGPIN with no evidence of cancer; upper limit of prostate‐specific antigen (PSA) of 10 ng/mL (as measured locally); American Urological Association (AUA) symptom score < 20 (41), signifying no debilitating urinary problems; ambulatory and able to carry out work of a light or sedentary nature Demographics: Selenium and placebo participants were well balanced with respect to age, race, ethnicity, pre‐study PSA category, vitamin E supplements, and number of cores in the initial biopsy. They also were well balanced in body mass index, baseline blood selenium, performance status, and number of cores revealing HGPIN. |
|
| Interventions | Participants were randomised in fashion to placebo or 200 µg/d of selenium, with daily treatment scheduled for 3 years or until a prostate cancer diagnosis. Recruitment: not reported End of blinded treatment period: at 3 years |
|
| Outcomes |
Primary outcome measure: • progression of HGPIN to prostate cancer over a 3‐year period |
|
| Risk estimates [95% CI] |
Primary outcomes: • Adjusted OR 0.913 (95% CI 0.55 to 1.52, P = 0.727) for risk of prostate cancer as a function of treatment group (with placebo as referent group) |
|
| Selenium levels in exposure categories | d.n.a. | |
| Notes | OR estimate was given by the trial author. | |
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Random sequence generation (selection bias) | Low risk | Described as randomised |
| Allocation concealment (selection bias) | Low risk | Central randomisation with pathology review |
| Blinding (performance bias and detection bias) All outcomes | Low risk | Blinding of participants and personnel |
| Selective reporting (reporting bias) | Low risk | No problems found |
McNaughton 2005.
| Methods | Matched, nested case‐control study (McNaughton 2005b) Cohort study (Heinen 2007; van der Pols 2009) Country: Australia |
|
| Participants |
Name of parent cohort: Nambour Skin Cancer Study Recruitment: 1992 to 1996 Case definition: incidence McNaughton 2005b Participants: approximately 1000 men and women Inclusion criteria: randomly selected adults, aged 20 to 69 years; recruited for participation in a randomised controlled trial for skin cancer prevention with beta‐carotene supplements and sunscreen application in 1992; living in the Nambour community; free of SCC at baseline; blood sample and FFQ provided in 1996; participants with extreme energy intakes in FFQ excluded Outcome assessment: December 2001 Number of cases: • Basal cell carcinoma of the skin: 90 (male/female: 39/51) Years of follow‐up: 5.5 Type of selenium marker: serum and intake (FFQ) Heinen 2007 Participants: 1001 men and women Inclusion criteria: randomly selected adults, aged 20 to 69 years; recruited for participation in randomised controlled trial for skin cancer prevention with beta‐carotene supplements and sunscreen application in 1992; living in the Nambour community; blood sample and FFQ provided in 1996; participants with extreme energy intakes in FFQ and missing consumption frequencies for more than 10% of food items excluded Outcome assessment: 31 December 2004 Number of cases: • Basal cell carcinoma of the skin: 149 (male/female: 87/62) participants with 321 BCC tumours • Squamous cell carcinoma of the skin: 116 (male/female: 70/46) participants with 221 SCC tumours Case definition: incidence (tumour‐based incidence and person‐based incidence) Years of follow‐up: 8 Type of selenium marker: intake (FFQ) van der Pols 2009: Participants: 485 (male/female: 223/262) men and women Inclusion criteria: randomly selected adults, aged 20 to 69 years; recruited for participation in randomised controlled trial for skin cancer prevention with beta‐carotene supplements and sunscreen application in 1992; randomised to placebo in the intervention trial; living in the Nambour community; free of SCC at baseline; blood sample and FFQ provided in 1996; participants with extreme energy intakes in FFQ excluded Outcome assessment: 31 December 2004 Number of cases: • Basal cell carcinoma of the skin: 77 (male/female: 46/31) participants with 173 BCC tumours • Squamous cell carcinoma of the skin: 59 (male/female: 38/21) participants with 124 SCC tumours Years of follow‐up: 8 Type of selenium marker: serum |
|
| Interventions | d.n.a. | |
| Outcomes |
McNaughton 2005b:
Statistical methods: conditional logistical regression
Variables controlled in analysis: age, gender
Variables controlled by matching: age, gender Heinen 2007 Statistical methods: generalised linear models Variables controlled in analysis: age, sex, intervention arm in RCT, energy intake, skin colour, elastosis of the neck, smoking, use of dietary supplements, history of skin cancer van der Pols 2009 Statistical methods: generalised linear models Variables controlled in analysis: age, sex, pack‐years of smoking, alcohol intake, time spent outdoors on weekdays, history of skin cancer before 1996 |
|
| Risk estimates [95% CI] |
McNaughton 2005b Reference category: lowest quartile Results: Basal cell carcinoma (skin) • Both genders: highest quartile: OR 0.86 (95% CI 0.38 to 1.96) biochemical selenium level • Both genders: highest quartile: OR 1.13 (95% CI 0.47 to 2.74) selenium intake Heinen 2007 Reference category: lowest tertile Results: Basal cell carcinoma (skin) • Both genders: highest tertile: RR 0.95 (95% CI 0.59 to 1.50) Squamous cell carcinoma (skin) • Both genders: highest tertile: RR 1.3 (95% CI 0.77 to 2.3) van der Pols 2009 Reference category: lowest exposure category Results: Basal cell carcinoma (skin) • Both genders: highest exposure category: RR 0.58 (95% CI 0.32 to 1.07) Squamous cell carcinoma (skin) • Both genders: highest exposure category: RR 0.49 (95% CI 0.24 to 0.99) |
|
| Selenium levels in exposure categories |
McNaughton 2005b
n.r. Heinen 2007 • Lowest tertile ≤ 76.20 µg/d • Highest tertile ≥ 89.31 µg/d van der Pols 2009 • Lowest exposure category ≤ 78.96 µg/L • Highest exposure category ≥ 102.65 µg/L |
|
| Notes |
Primary publication:McNaughton 2005b
Other publications: Heinen 2007, van der Pols 2009 Tumour‐based incidence: number of newly developed histologically confirmed BCCs or SCCs divided by person‐years of follow‐up accumulated over follow‐up period Person‐based incidence: number of persons newly affected by BCC or SCC during the same person‐years of follow‐up time as calculated for the tumour‐based analysis |
|
Menkes 1986.
| Methods | Matched, nested case‐control study Country: United States |
|
| Participants |
Participants: 20,305 men and women
Inclusion criteria: female and male inhabitants of Washington county/Maryland; history of cancer at baseline excluded
Name of parent cohort: CLUE I Cohort Recruitment: September to November 1974 Menkes 1986b Outcome assessment: 1983 Number of cases: • Lung cancer: 99 (69% male/31% female) Helzlsour 1996 Inclusion criteria: women only; women who used hormones at baseline excluded Outcome assessment: 1989 Number of cases: • Ovarian cancer: 35 (male/female: 0/35) Breslow 1995 Outcome assessment: 1994 Number of cases: • Melanoma: 23 (male/female: n.r.) • Basal cell carcinoma (skin): 17 (male/female: n.r.) • Squamous cell cancer: 37 (male/female: n.r.) Zheng 1993 Outcome assessment: 1990 Number of cases: • Oral and pharyngeal: 28 (male/female: n.r.) Batieha 1993 Inclusion criteria: 15,161 women Outcome assessment: 31 May 1990 Number of cases: • Cervical cancer: 50 (male/female: 0/50) Helzlsour 1989 Inclusion criteria: 20,305 men and women Outcome assessment: 1986 Number of cases: • Bladder cancer: 35 (male/female: n.r.) Burney 1989 Outcome assessment: 1986 Number of cases: • Pancreatic cancer: 22 (male/female: 9/13) Ko 1994 Outcome assessment: 25 September 1991 Number of cases: • Colon cancer: 121 (male/female: 50/71) Case definition: incidence Years of follow‐up: 8.0 to 16.8 Type of selenium marker: serum |
|
| Interventions | d.n.a. | |
| Outcomes |
Menkes 1986b
Statistical methods: conditional logistical regression
Variables controlled by matching: age, gender, race/ethnicity, smoking status, year and month of sample collection Helzlsour 1986 Statistical methods: conditional logistical regression Variables controlled by matching: age, race/ethnicity, day and time of blood sample collection, hours since last meal, time since last menstrual period (postmenopausal: years, premenopausal: days) Breslow 1995 Statistical methods: conditional logistical regression Analysed cases: 17 of 98 basal cell carcinoma cases and 23 of 30 melanoma cases (and all squamous cell carcinoma cases) included in analysis Variables controlled by matching: age, gender, race/ethnicity Zheng 1993 Statistical methods: n.r. Variables controlled in analysis: smoking Variables controlled by matching: age, gender, race/ethnicity, year and month of sample collection, hours between previous meal and blood collection Batieha 1993 Statistical methods: conditional logistical regression Analysed cases: 50 of 60 (CIS and invasive cervical cancer) (reason for non‐inclusion: no matched control available) Variables controlled by matching: age, race/ethnicity, year and month of blood collection, hours since last meal, time since last menstrual period Helzlsour 1989 Statistical methods: n.r. Variables controlled in analysis: cigarette smoking, use of vitamin supplements Variables controlled by matching: age, gender, race/ethnicity, hours since last meal (all samples collected in same year) Burney 1989 Statistical methods: n.r. Variables controlled by matching: age, gender, race/ethnicity, hours since last meal Ko 1994 Analysed cases: 121 of 154 (reason for non‐inclusion: no serum sample available, tumour pathology or localisation unclear) Statistical methods: conditional logistical regression Variables controlled by matching: age, gender, race/ethnicity, year and month of sample collection, hours since last meal, women: time since last menstrual period, women: use of hormones/hormonal contraceptives |
|
| Risk estimates [95% CI] |
Menkes 1986b
Reference category: highest quintile Results: Lung cancer • Both genders: lowest quintile: OR 0.68 (CI not reported) Helzlsouer 1986 Reference category: lowest tertile Results: Ovarian cancer • Highest tertiles: OR 0.58 (95% CI 0.20 to 1.70) Breslow 1995 Reference category: lowest tertile Results: Melanoma • Both genders: highest tertile: OR 0.9 (95% CI 0.3 to 2.5) Basal cell carcinoma (skin) • Both genders: highest tertile: OR 0.8 (95% CI 0.1 to 4.5) Squamous cell cancer • Both genders: highest tertile: OR 0.6 (95% CI 0.2 to 1.5) Zheng 1993 Reference category: lowest tertile Results: Oral and pharyngeal cancer • Both genders: highest tertile: OR 5.43 (CI not reported) Batieha 1993 Reference category: highest tertile Results: Cervical cancer • Lowest tertile: OR 1.12 (95% CI 0.50 to 2.53) Helzlsour 1989 Reference category: highest tertile Results: Bladder cancer • Both genders: lowest tertile: OR 2.06 (95% CI 0.67 to 6.35) Burney 1989 Reference category: highest tertile Results: Pancreatic cancer • Both genders: lowest tertile: OR 4.5 (CI not reported) (unmatched analysis) • Both genders: lowest tertile vs higher 2 tertiles: OR 3.90 (95% CI 1.13 to 13.2) (matched analysis) • Male: 12.5 (95% CI 1.8 to 84.0) (unmatched analysis) • Female: 1.2 (95% CI 0.6 to 2.5) (unmatched analysis) Ko 1994 Reference category: highest quartile Results: Colon cancer • Both genders: lowest quartile: OR 0.82 (95% CI 0.35 to 1.92) |
|
| Selenium levels in exposure categories |
Menkes 1986b
• Quintiles: n.r. Helzlsouer 1986 Women • Lowest tertile: ≤ 105.0 µg/L • Highest tertile: ≥ 116.1 µg/L Breslow 1995 • Tertiles: n.r. Zheng 1993 • Tertiles: n.r. Batieha 1993 Women • Lowest tertile: ≤ 0.109 ppm • Highest tertile: ≥ 0.124 ppm Helzlsour 1989 Both genders • Lowest tertile: ≤ 109.0 µg/L • Highest tertile: ≥ 119.1 µg/L Burney 1989 • Lowest: 0.99 to 1.26 µmol/L; highest: 1.44 to 1.81 µmol/L Ko 1994 • Lowest quartile: ≤ 99.0 µg/L • Highest quartile: ≥ 118.1 µg/L |
|
| Notes | Primary publication:Menkes 1986b Other publications: Helzlsour 1996, Breslow 1995, Zheng 1993, Batieha 1993, Helzlsour 1989, Burney 1989, Ko 1994, Schober 1987 (cases included in Ko 1994), Menkes 1986a (case included in Menkes 1986b) | |
Michaud 2002.
| Methods | Matched, nested case‐control study Country: Finland |
|
| Participants |
Participants: 29,133 men
Inclusion criteria: 50 to 69 years of age; smokers; no history of cancer (other than non‐melanoma skin cancer) at baseline; no severe physical or psychiatric illness; intake of vitamin E/A/beta‐carotene supplements in excess of defined amounts
Name of parent cohort: Alpha‐Tocopherol, Beta‐Carotene Cancer Prevention (ATBC) Study Recruitment: 1985 to 1988 Outcome assessment: 30 April 1993 Number of cases: • Bladder cancer: 133 (male/female: 133/0) Case definition: incidence Years of follow‐up: 5 to 8 Type of selenium marker: toenail |
|
| Interventions | d.n.a. | |
| Outcomes | Statistical methods: conditional logistical regression Variables controlled in analysis: smoking dose and duration Variables controlled by matching: age, year/month of sample collection, intervention group status in RCT (only male smokers included in cohort) | |
| Risk estimates [95% CI] |
Reference category: lowest tertile/quartile Results: Bladder cancer • Male: highest tertile: OR 0.90 (95% CI 0.45 to 1.78) • Male: highest quartile: OR 0.87 (95% CI 0.30 to 2.52) |
|
| Selenium levels in exposure categories | n.r. | |
| Notes | ||
Michaud 2005.
| Methods | Matched, nested case‐control study Country: United States |
|
| Participants |
Participants: 101,950 (male/female: 33,737/68,213)
Inclusion criteria: cohort of HPFS (men) and NHS (women); no history of cancer at baseline
Name of parent cohort: Health Professional Follow‐Up Study (HPFS) and Nurses' Health Study (NHS) Recruitment: 1987 (HPFS), 1983 (NHS) Outcome assessment: 2000 Number of cases: • Bladder cancer: 337 (male/female: 221/116) Case definition: incidence Years of follow‐up: 13 to 17 Type of selenium marker: toenail |
|
| Interventions | d.n.a. | |
| Outcomes | Statistical methods: conditional logistical regression Variables controlled in analysis: pack‐years of smoking, heavy smoking at baseline Variables controlled by matching: age, gender, smoking status, month of sample collection | |
| Risk estimates [95% CI] |
Reference category: lowest quartile Results: Bladder cancer • Male: highest quartile: OR 1.17 (95% CI 0.66 to 2.07) • Female: highest quartile: OR 0.36 (95% CI 0.14 to 0.91) |
|
| Selenium levels in exposure categories |
Men
• Lowest quartile: ≤ 0.722 µg/g
• Highest quartile: ≥ 0.912 µg/g Women • Lowest quartile: ≤ 0.686 µg/g • Highest quartile: ≥ 0.840 µg/g |
|
| Notes | ||
Muka 2017.
| Methods | Cohort study Country: the Netherlands |
|
| Participants |
Name of parent cohort: The Rotterdam Study Participants: 5435 (male/female: n.r.) Inclusion criteria: aged ≥ 55 and living in the Ommoord district Recruitment: 1989 to 1993 Outcome assessment: December 2011 Number of cases: 211 (male/female: 128/83) Case definition: incidence Years of follow‐up: mean: 15.2 Type of selenium marker: intake |
|
| Interventions | d.n.a. | |
| Outcomes |
Analysed cases: 211 (male/female: 128/83) Statistical methods: Cox proportional hazard model Variables controlled in analysis: age, sex, alcohol intake, body mass index, smoking status, physical activity, Dutch healthy diet index, dietary processed meat intake, dietary unprocessed red meat intake, total energy intake, hormone replacement therapy, diabetes mellitus, education status, income status, total energy, adjusted sum of other minerals (excluding selenium), and family history of cancer |
|
| Risk estimates [95% CI] |
Reference category: lowest tertile Results: Lung cancer • Highest tertile: HR 1.39 (95% CI 0.97 to 1.99) |
|
| Selenium levels in exposure categories | n.r. | |
| Notes | Lung cancer: highest tertile: HR 1.44 (95% CI 0.98 to 2.11) after exclusion of lung cancer within the first 2 years of follow‐up | |
Nomura 1987.
| Methods | Unmatched, nested case‐control study Country: United States |
|
| Participants |
Participants: 6860 men
Inclusion criteria: born 1900 to 1919; Japanese ancestry; inhabitants of Oahu/Hawaii; participants in the Honolulu Heart Program (1965 to 1968)
Name of parent cohort: Honolulu Heart Program Recruitment: 1971 to 1975 Outcome assessment: n.r. Number of cases: • Any cancer: 280 (male/female: 280/0) • Stomach cancer: 66 (male/female: 66/0) • Rectal cancer: 32 (male/female: 32/0) • Lung cancer: 71 (male/female: 71/0) • Colon cancer: 82 (male/female: 82/0) • Bladder cancer: 29 (male/female: 29/0) Case definition: incidence Years of follow‐up: 11 Type of selenium marker: serum |
|
| Interventions | d.n.a. | |
| Outcomes | Statistical methods: proportional hazard regression/Cox regression Variables controlled in analysis: • Age at examination, cigarettes/d (any cancer, lung cancer, bladder cancer) • Age at examination (stomach, rectum, colon) | |
| Risk estimates [95% CI] |
Reference category: highest quintile Results: Stomach cancer • Male: lowest quintile: OR 0.9 (CI not reported) Rectal cancer • Male: lowest quintile: OR 1.6 (CI not reported) Lung cancer • Male: lowest quintile: OR 1.1 (CI not reported) Colon cancer • Male: lowest quintile: OR 1.8 (CI not reported) Bladder cancer • Male: lowest quintile: OR 3.1 (CI not reported) All five types of cancer • Male: lowest quintile: OR 1.3 (CI not reported) |
|
| Selenium levels in exposure categories | Lowest quintile: ≤ 103.0 µg/L Highest quintile: ≥ 133.1 µg/L | |
| Notes | N.B.: "Any cancer" in this study comprises all cancer cases for stomach, rectal, lung, colon, and bladder cancer. | |
Nomura 2000.
| Methods | Matched, nested case‐control study Country: United States |
|
| Participants |
Participants: 9345 men
Inclusion criteria: no cancer diagnosis at baseline, blood sample available for analysis, men from 2 cohorts: subcohort 1: participants of Nomura 1987; subcohort 2: brothers of participants in Nomura 1987 Recruitment: 1971 to 1977 Outcome assessment: 1995 Number of cases: • Prostate cancer: 249 (male/female: 249/0) Case definition: incidence Years of follow‐up: 19 to 25 Type of selenium marker: serum |
|
| Interventions | d.n.a. | |
| Outcomes | Analysed cases: random sample of 249 (out of 360) because of limited resources Statistical methods: generalised linear model Variables controlled in analysis: cigarette smoking history, age Variables controlled by matching: age, year/month of sample collection, recruitment in subcohort 1 or 2 | |
| Risk estimates [95% CI] |
Reference category: lowest quartile Results: Prostate cancer • Highest quartile: OR 0.5 (95% CI 0.3 to 0.9) |
|
| Selenium levels in exposure categories | Lowest quartile: ≤ 119.29 µg/L Highest quartile: ≥ 147.20 µg/L | |
| Notes | ||
NPCT 2002.
| Methods | Randomised controlled trial Nutritional Prevention of Cancer Trial (NPCT) Allocation: random, block/stratified by clinic Sequence generation: computer‐generated random numbers Concealment: central assignment (sealed pill bottles) Blinding: participant blinded, doctor blinded, outcome assessor/pathologist unclear, review/coding of medical records blinded Dropouts/withdrawals: “9 patients (5 in the selenium group and 4 in the placebo group) declined to provide additional illness information” (Clark 1996, p. 1959) ‐ 0 participants lost to vital follow‐up Intention‐to‐treat‐analysis: yes Recruitment period: 1983 to 1991 End of predefined study period: 31 December 1993 Blinded intervention continued until end of blinded period: 31 January 1996 Intervention duration: • 31 December 1993 (end of study period): mean = 4.5 years • 31 January 1996 (end of blinded period): mean = 7.9 years Observation period/dermatological follow‐up: • 31 December 1993 (end of study period): mean = 6.4 years • 31 January 1996 (end of blinded period): mean = 7.4 years Detection of cases: dermatological examination and interview every 6 months during follow‐up; incident BCC and SCC diagnosed by biopsy and confirmed by another dermatopathologist Informed consent: written informed consent forms, approval by institutional review board of participating institutions |
|
| Participants |
Country: United States Participants: 1312 (randomised to selenium group: 653; to placebo group: 659) Condition: male and female participants with history of 2 or more squamous cell or basal cell skin cancers Demographics: mean age 63.4 years (selenium)/63.0 years (placebo); 73.8% men (selenium), 75.6% men (placebo) Recruitment and setting: 7 dermatological clinics (3 academic units, 4 private practices) in the United States |
|
| Interventions |
Intervention: 200 µg selenium supplied as 500 mg selenium yeast tablets p.o. daily Control: placebo |
|
| Outcomes |
Primary outcome measure: incidence of basal and squamous cell carcinoma of the skin: • All analyses were based on 1250 participants with initial blood collection within 4 days after randomisation (621 in the selenium group and 629 in the placebo group) Other reported outcomes and secondary outcome measures: • Reported in Clark 1996: incidence of lung cancer, prostate cancer, colorectal cancer, any cancer, head and neck cancer, bladder cancer, oesophageal cancer, breast cancer, melanoma, haematological cancer • Reported in Duffield‐Lillico 2002: overall cancer mortality |
|
| Risk estimates [95% CI] |
Primary outcomes: At end of study period (31 December 1993) (Clark 1996) • BCC: RR 1.10 (95% CI 0.95 to 1.28); cases: selenium group: 377, placebo group: 350; incidence per person‐year under follow‐up: selenium group 0.16, placebo group 0.15 • SCC: RR 1.14 (95% CI 0.93 to 1.39); cases: selenium group 218, placebo group: 190; incidence per person‐year under follow‐up: selenium group 0.07, placebo group 0.06 At end of blinded period (31 January 1996) (Duffield‐Lillico 2003) • BCC: RR 1.17 (95% CI 1.02 to 1.35), HR 1.09 (95% CI 0.94 to 1.26); number of cases not reported; incidence per person‐year under follow‐up: selenium group: 0.16, placebo group 0.13 • SCC: RR 1.32 (95% CI 1.09 to 1.60), HR 1.25 (95% CI 1.03 to 1.51); number of cases not reported; incidence per person‐year under follow‐up: selenium group: 0.05, placebo group 0.07 • NMSC: RR 1.27 (95% CI 1.11 to 1.45), HR 1.17 (95% CI 1.02 to 1.34); number of cases not reported; incidence per person‐year under follow‐up: selenium group: 0.20, placebo group 0.16 Other reported outcomes and secondary outcomes: At end of study period (31 December 1993) (Clark 1996) • Lung cancer: RR 0.54 (95% CI 0.30 to 0.98), adjusted HR 0.56 (95% CI 0.31 to 1.01) cases selenium: 17, placebo: 31 • Prostate cancer: RR 0.37 (95% CI 0.18 to 0.71), adjusted HR 0.35 (95% CI 0.18 to 0.65) cases selenium: 13, placebo: 35 • Colorectal cancer: RR 0.42 (95% CI 0.18 to 0.95), adjusted HR 0.39 (95% CI 0.17 to 0.90) cases selenium: 8, placebo: 19 • Any cancer: RR 0.63 (95% CI 0.47 to 0.85), adjusted HR 0.61 (95% CI 0.46 to 0.82) cases selenium: 77, placebo: 119 • Head and neck cancer: RR 0.74 (95% CI 0.21 to 2.43), adjusted HR 0.77 (95% CI 0.27 to 2.24) cases selenium: 6, placebo: 8 • Bladder cancer: RR 1.32 (95% CI 0.40 to 4.61), adjusted HR 1.27 (95% CI 0.44 to 3.67) cases selenium: 8, placebo: 6 • Oesophageal cancer: RR 0.33 (95% CI 0.03 to 1.84), adjusted HR 0.30 (95% CI 0.06 to 1.49) cases selenium: 2, placebo: 6 • Breast cancer: RR 2.88 (95% CI 0.72 to 16.5), adjusted HR 2.95 (95% CI 0.80 to 10.9) cases selenium: 9, placebo:3 • Melanoma: RR 0.97 (95% CI 0.32 to 2.96), adjusted HR 0.92 (95% CI 0.34 to 2.45) cases selenium: 8, placebo: 8 • Haematological cancer: RR 1.58 (95% CI 0.46 to 6.14), adjusted HR 1.50 (95% CI 0.49 to 4.60) cases selenium: 8, placebo: 5 • Other specific carcinomas: RR 0.55 (95% CI 0.14 to 1.82), adjusted HR 0.54 (95% CI 0.18 to 1.62), cases selenium: 5, placebo: 9 • Total carcinoma: RR 0.55 (95% CI 0.40 to 0.77), adjusted HR 0.54 (95% CI 0.39 to 0.75), cases selenium: 59; placebo: 104 • Leukaemia/lymphoma: RR 1.58 (95% CI 0.46 to 6.14), adjusted HR 1.50 (95% CI 0.49 to 4.60), cases selenium: 8, placebo 5 • Other specific non‐carcinomas: RR 0.99 (95% CI 0.13 to 7.37), HR 0.99 (95% CI 0.20 to 4.94), cases selenium: 3, placebo: 3 • Total non‐carcinomas: RR 1.17 (95% CI 0.57 to 2.44), adjusted HR 1.16 (95% CI 0.60 to 2.27), cases selenium: 19; placebo: 16 At end of blinded period (31 January 1996) (Duffield‐Lillico 2002) • Lung cancer: RR 0.70 (95% CI 0.40 to 1.21), adjusted HR 0.74 (95% CI 0.44 to 1.24), cases selenium: 25, placebo: 35 • Prostate cancer: RR 0.51 (95% CI 0.29 to 0.87), adjusted HR 0.48 (95% CI 0.28 to 0.80), cases selenium: 22, placebo: 42 • Colorectal cancer: RR 0.46 (95% CI 0.19 to 1.08), adjusted HR 0.46 (95% CI 0.21 to 1.02), cases selenium: 9, placebo: 19 • Any cancer: RR 0.75 (95% CI 0.58 to 0.98), adjusted HR 0.75 (95% CI 0.58 to 0.97), cases selenium: 105, placebo: 137 • Head and neck cancer: RR 1.27 (95% CI 0.42 to 4.01), adjusted HR 1.27 (95% CI 0.47 to 3.42), cases selenium: 9, placebo: 7 • Bladder cancer: RR 1.24 (95% CI 0.44 to 3.61), adjusted HR 1.28 (95% CI 0.50 to 3.25), cases selenium: 10, placebo: 8 • Oesophageal cancer: RR 0.39 (95% CI 0.04 to 2.41), adjusted HR 0.40 (95% CI 0.08 to 2.07), cases selenium: 2, placebo: 5 • Breast cancer: RR 1.82 (95% CI 0.62 to 6.01), adjusted HR 1.89 (95% CI 0.69 to 5.14), cases selenium: 11, placebo: 6 • Melanoma: RR 1.21 (95% CI 0.46 to 3.30), adjusted HR 1.18 (95% CI 0.49 to 2.85), cases selenium: 11, placebo: 9 • Haematological cancer (lymphoma and leukaemia): RR 1.32 (95% CI 0.40 to 4.61), adjusted HR 1.25 (95% CI 0.43 to 3.61), cases selenium: 8, placebo: 6 • Cancer mortality, all sites: RR 0.59 (95% CI 0.39 to 0.89), adjusted HR 0.59 (95% CI 0.39 to 0.87), cases selenium: 40, placebo: 66 • Other carcinomas: RR 0.66 (95% CI 0.19 to 2.07), adjusted HR 0.67 (95% CI 0.24 to 1.88), cases selenium: 6, placebo:9 • Other non‐carcinomas: RR 0.59 (95% CI 0.09 to 3.04), adjusted HR 0.59 (95% CI 0.14 to 2.47), cases selenium: 3, placebo: 5 |
|
| Selenium levels in exposure categories | d.n.a. | |
| Notes |
Adverse effects: Clark 1996: 35 participants (21 in selenium and 14 in control group) complained of adverse effects, mostly involving gastrointestinal upset, and withdrew treatment. Post hoc introduced secondary outcomes: all‐cause mortality, total cancer mortality, total cancer incidence, and incidence of lung/prostate/colorectal cancers HR: adjusted for sex, age, smoking status, clinic site, plasma selenium concentration, clinical sun damage, sunscreen use at baseline, and number of BCCs/SCCs/NMSCs in the 12 months before randomisation |
|
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Random sequence generation (selection bias) | Low risk | Random, block/stratified by clinic, computer‐generated random numbers |
| Allocation concealment (selection bias) | Low risk | Central assignment (sealed pill bottles) |
| Blinding (performance bias and detection bias) All outcomes | Unclear risk | Occurrence of a detection bias, namely, a considerably higher rate of prostate biopsy in the placebo group |
| Selective reporting (reporting bias) | Low risk | No problems found |
O'Grady 2014.
| Methods | Cohort study Country: United States |
|
| Participants |
Name of parent cohort: National Institute of Health‐American Association of Retired Persons (NIH‐AARP) Diet and Health Study Participants: 482,807 (male/female: 287,944/194,863) Inclusion criteria: 50 to 71 years of age, AARP members, no previous diagnosis of cancer other than NMSC Recruitment: 1995 to 1996 Outcome assessment: December 2006 Number of cases: 592 (male/female: 257/335) Case definition: incidence Years of follow‐up: mean: 9.1 Type of selenium marker: intake |
|
| Interventions | d.n.a. | |
| Outcomes |
Analysed cases:
• Total thyroid cancer: 592 (male/female: 257/335)
• Papillary thyroid cancer subtype: 406 (male/female: 164/242)
• Follicular thyroid cancer subtype: 113 (male/female: 57/56) Statistical methods: Cox proportional hazard model Variables controlled in analysis: entry age, sex, calories, smoking status, race, education, BMI, physical activity, vitamin C, vitamin E, beta‐carotene, and folate |
|
| Risk estimates [95% CI] |
Reference category: lowest quintile Results: Total thyroid cancer • Both genders: highest quintile: HR 1.35 (95% CI 0.99 to 1.84) • Male: highest quintile: HR 1.23 (95% CI 0.71 to 2.12) • Female: highest quintile: HR 1.14 (95% CI 0.65 to 2.02) Papillar subtype • Both genders: highest quintile: HR 1.35 (95% CI 0.92 to 1.98) • Male: highest quintile: HR 1.32 (95% CI 0.65 to 2.69) • Female: highest quintile: HR 1.29 (95% CI 0.68 to 2.46) Follicular subtype • Both genders: highest quintile: HR 1.41 (95% CI 0.71 to 2.79) • Male: highest quintile: HR 1.32 (95% CI 0.43 to 4.03) • Female: highest quintile: HR 0.88 (95% CI 0.20 to 3.87) |
|
| Selenium levels in exposure categories | Lowest quintile: median 47 µg/d Highest quintile: median 150.1 µg/d | |
| Notes | ||
Outzen 2016.
| Methods | Matched, nested case‐control study Country: Denmark |
|
| Participants |
Name of parent cohort: Danish Prospective Diet, Cancer and Health Study Participants: 27,179 men Inclusion criteria: aged 50 to 64, born in Denmark, residents in the Copenhagen and Aarhus areas, no previous history of cancer Recruitment: December 1993 to May 1997 Outcome assessment: 31 December 2007 Number of cases: 911 (male/female: 911/0) Case definition: incidence Years of follow‐up: 8 Type of selenium marker: plasma |
|
| Interventions | d.n.a. | |
| Outcomes |
Analysed cases: Prostate cancer • 784 (male/female: 784/0) Statistical methods: conditional logistical regression Variables controlled in analysis: body mass index, education, smoking status, duration and frequency, and participation in sport Variables controlled by matching: age at blood collection, time of day of blood collection, and fasting status |
|
| Risk estimates [95% CI] |
Reference category: lowest quartile Results: Prostate cancer • Highest quartile: OR 0.95 (95% CI 0.70 to 1.29) |
|
| Selenium levels in exposure categories | Lowest quartile: ≤ 71.4 μg/d Highest quartile: > 88.9 μg/d | |
| Notes | ||
Overvad 1991.
| Methods | Cohort/subcohort controlled cohort study Country: Channel Islands (UK) |
|
| Participants |
Participants: 5162 women
Inclusion criteria: ≥ 35 years of age; ostensibly healthy inhabitants of Guernsey
Name of parent cohort: Channel Island Cohort Recruitment: 1967 to 1976 Outcome assessment: end of 1985 Number of cases: • Breast cancer: 46 (male/female: 0/46) Case definition: incidence Years of follow‐up: mean: 11 years for cases Type of selenium marker: plasma |
|
| Interventions | d.n.a. | |
| Outcomes | Analysed cases: 46 of 88 (reason for non‐inclusion: no plasma available) Statistical methods: logistical regression Variables controlled in analysis: age, age at menarche, age at first baby, parity, BMI | |
| Risk estimates [95% CI] |
Reference category: highest quartile Results: Breast cancer • Lowest quartile: RR 0.80 (95% CI 0.29 to 2.19) |
|
| Selenium levels in exposure categories | Lowest quartile: ≤ 84.90 µg/L Highest quartile: ≥ 116.00 µg/L | |
| Notes | ||
Pantavos 2015.
| Methods | Cohort study Country: the Netherlands |
|
| Participants |
Name of parent cohort: The Rotterdam Study Participants: 4877 women Inclusion criteria: aged ≥ 55 and living in the Ommoord district. no history of previous breast cancer Recruitment: July 1989 to September 1993 Outcome assessment: December 2010 Number of cases: 199 (male/female: 0/199) Case definition: incidence Years of follow‐up: median: 17 years Type of selenium marker: intake |
|
| Interventions | d.n.a. | |
| Outcomes |
Analysed cases: 199 (male/female: 0/199) Statistical methods: Cox proportional hazard model Variables controlled in analysis: age, body mass index, education level, family history of breast cancer, smoking status, alcohol consumption, use of multi‐vitamin supplement |
|
| Risk estimates [95% CI] |
Reference category: lowest tertile Results: Breast cancer • Highest tertile: HR 1.34 (95% CI 0.94 to 1.91) |
|
| Selenium levels in exposure categories | Lowest tertile: median 23.58 µg/d Highest tertile: median 37.46 µg/d | |
| Notes | ||
Park 2015.
| Methods | Cohort study Country: United States (Hawaii and California) |
|
| Participants |
Name of parent cohort: The Multiethnic Cohort Participants: 75,216 men Inclusion criteria: aged 45 to 75, African Americans, Native Hawaiians, Japanese American, Latinos, and white men, without a previous diagnosis of prostate cancer Recruitment: 1993 to 1996 Outcome assessment: 31 December 2010 Number of cases: • Prostate cancer: 7115 Case definition: incidence Years of follow‐up: mean: 13.9 Type of selenium marker: intake |
|
| Interventions | d.n.a. | |
| Outcomes |
Analysed cases: • Prostate cancer: 7115 Statistical methods: Cox proportional hazard model Variables controlled in analysis: age at entry, race/ethnicity, family history of prostate cancer, body mass index, height, smoking status, education level, history of diabetes, physical activity, daily intakes of alcohol, calcium, legume, and lycopene |
|
| Risk estimates [95% CI] |
Reference category: lowest quintile Results: Prostate cancer • Highest quintile: RR 1.01 (95% CI 0.84 to 1.20) |
|
| Selenium levels in exposure categories | Lowest quintile: < 44.0 µg/1000 kcal/d Highest quintile ≥ 60.1 µg/1000 kcal/d | |
| Notes | ||
Peleg 1985.
| Methods | Matched, nested case‐control study Country: United States |
|
| Participants |
Participants: 2530 men and women
Inclusion criteria: 15 years of age and older; residents of Evans County; cases within first 2 years of follow‐up excluded
Name of parent cohort: Evans County Study Recruitment: 1967 to 1969 Outcome assessment: January 1981 Number of cases: • Any cancer: 130 (male/female: 78/52) Case definition: incidence Years of follow‐up: 11 to 14 Type of selenium marker: serum |
|
| Interventions | d.n.a. | |
| Outcomes | Statistical methods: n.r. Variables controlled by matching: age, gender, race/ethnicity, year/month of sample collection | |
| Risk estimates [95% CI] |
Reference category: highest quartile Results: Any cancer • Both genders: lowest quartile: OR 1.0 (CI not reported) |
|
| Selenium levels in exposure categories | Lowest quartile: ≤ 103 µg/L Highest quartile: ≥ 127 µg/L | |
| Notes | ||
Peters 2007.
| Methods | Matched, nested case‐control study Country: United States |
|
| Participants |
Participants: 26,975 white non‐Hispanic men
Inclusion criteria: 55 to 74 years of age; excluded: no baseline questionnaire/informed consent/blood sample, no further contact after screening
Name of parent cohort: Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial Recruitment: September 1993 to June 2001 Outcome assessment: 1 October 2001 Number of cases: • Prostate cancer: 724 (male/female: 724/0) Case definition: incidence Years of follow‐up: 0.3 to 8.0 Type of selenium marker: serum |
|
| Interventions | d.n.a. | |
| Outcomes | Analysed cases: 724 of 803 (reason for non‐inclusion: no selenium measurement available) Statistical methods: n.r. Variables controlled in analysis: age, time since initial screening, year of blood collection, study centre Variables controlled by matching: age, month of sample collection, time since initial screening | |
| Risk estimates [95% CI] |
Reference category: lowest quartile Results: Prostate cancer • Highest quartile: OR 0.84 (95% CI 0.62 to 1.14) |
|
| Selenium levels in exposure categories | Lowest quartile: 50.5 to 126.7 µg/L Highest quartile: 158.0 to 253.0 µg/L | |
| Notes | ||
Peters 2008.
| Methods | Cohort study Country: United States |
|
| Participants |
Inclusion criteria: aged 50 to 76 years, participants recruited from subscribers to commercial mailing list, residents of western Washington state, non‐whites excluded, no malignant disease at baseline Name of parent cohort: Vitamins and Lifestyle (VITAL) study Recruitment: 1 October 2000 to 31 December 2002 Type of selenium marker: supplemental intake (questionnaire: use of supplements over past 10 years, mean supplemental intake/day calculated) Case definition: incidence Peters 2008 Participants: 35,242 men Outcome assessment: 31 December 2004 Number of cases: • Prostate cancer: 818 (male/female: 818/0) Years of follow‐up: 2 to 4 Asgari 2009 Participants: 69,671 men and women Outcome assessment: 31 December 2006 Number of cases: • Melanoma: 461 (male/female: n.r.) Years of follow‐up: 4 to 5 years |
|
| Interventions | d.n.a. | |
| Outcomes |
Peters 2008 Analysed cases: 818 of 830 (reason for non‐inclusion: not reported) Statistical methods: Cox proportional hazard regression analysis Variables controlled in analysis: age, family history of prostate cancer, BPH, income, multi‐vitamin use Asgari 2009 Analysed cases: 1 case not analysed (reason for non‐inclusion: not reported) Statistical methods: Cox proportional hazard regression Variables controlled in analysis: age, sex, education, family history of melanoma, personal history of non‐melanoma skin cancer, mole removal, freckles, sunburns, hair colour, reaction to sunlight exposure |
|
| Risk estimates [95% CI] |
Reference category: no supplemental selenium intake (lowest exposure category) Peters 2008 Results: Prostate cancer • Highest exposure category: RR 0.90 (95% CI 0.62 to 1.30) Asgari 2009 Results: Melanoma • Highest exposure category HR 0.98 (95% CI 0.69 to 1.41) |
|
| Selenium levels in exposure categories | Stratification according to supplemental selenium intake Peters 2008 • Lowest category: no supplemental intake • Highest category ≥ 51 µg/d Asgari 2009 • Lowest exposure category: no supplemental intake • Highest exposure category ≥ 50 µg/d |
|
| Notes | ||
Ratnasinghe 2000.
| Methods | Matched, nested case‐control study Country: China |
|
| Participants |
Participants: 9143 men
Inclusion criteria: 35 years or older; tin miners employed by the Yunnan Tin Corporation; 10 or more years of underground mining/smelting; no history of cancer at baseline Recruitment: 1992 to 1997 Outcome assessment: 1997 Number of cases: • Lung cancer: 108 (male/female: 108/0) Case definition: incidence Years of follow‐up: ≈ 3 Type of selenium marker: serum |
|
| Interventions | d.n.a. | |
| Outcomes | Analysed cases: plasma available for 108 of a total of 339 identified cases Statistical methods: logistical regression, conditional logistical regression, Wilcoxon rank sum test Variables controlled in analysis: radon exposure, smoking Variables controlled by matching: age, year and month of sample collection | |
| Risk estimates [95% CI] |
Reference category: lowest tertile Results: Lung cancer • Highest tertile: OR 1.2 (95% CI 0.6 to 2.4) |
|
| Selenium levels in exposure categories | Lowest tertile: 20 to 39 µg/L Highest tertile: 55 to 121 µg/L | |
| Notes | ||
Reid 2008.
| Methods | Randomised controlled trial Substudy of the Nutritional Prevention of Cancer Trial (NPCT 2002) Allocation: random Sequence generation: computer‐generated random numbers Concealment: central assignment (sealed pill bottles) Blinding: participant blinded, doctor blinded, outcome assessor/pathologist unclear, review/coding of medical records blinded Dropouts/withdrawals: 2 participants declined to provide additional illness information, no participant lost to vital follow ‐up Intention‐to‐treat‐analysis: yes Recruitment period: 1989‐1992 Treatment duration: • Blinded intervention continued until the end of the blinded period; 1 February 1996. Observation period/dermatological follow‐up: 1 February 1996 Detection of cases: dermatological examination and interview every 6 months during follow‐up; incident BCC and SCC diagnosed by biopsy and confirmed by another dermatopathologist Informed consent: written informed consent forms, approval by institutional review boards of participating institutions |
|
| Participants | 423 male and female participants with prior non‐melanoma skin cancer Country: United States Participants: 423 (randomised to selenium group: 210, to placebo group: 213) Condition: male and female with history of 2 or more squamous cell or basal cell skin cancers Demographics: mean age 63.8 years (selenium)/63.8 years (placebo); 66.2% men (selenium). 68.2% men (placebo) Recruitment and setting: dermatological clinic in Macon, Georgia |
|
| Interventions |
Intervention: • 400 µg selenium supplied as selenium yeast tablets p.o. daily Control: • Placebo • 400 µg/d of selenium yeast or identical‐appearing low selenium yeast placebo Recruitment: 12 September 1989 to 3 April 1992 End of blinded treatment period: 2 February 1996 |
|
| Outcomes |
Primary outcome measure: incidence of basal and squamous cell carcinoma of the skin • All analyses were based on n = 423 participants with initial blood collection within 4 days after randomisation Other reported outcomes: • Total internal cancer incidence |
|
| Risk estimates [95% CI] |
Primary outcomes: • BCC: RR 0.90 (95% CI 0.65 to 1.24); cases: selenium group: 76, placebo group: 83; adjusted HR: 0.95 (95% CI 0.69 to 1.29) • SCC: RR 1.05 (95% CI 0.71 to 1.56); cases: selenium group: 56, placebo group: 53; adjusted HR: 1.05 (95% CI 0.72 to 1.53) • NMSC: RR 0.88 (95% CI 0.66 to 1.16); cases: selenium group: 98, placebo group: 108; adjusted HR: 0.91 (95% CI 0.69 to 1.20) • NMSC in women: RR 0.40 (95% CI 0.20 to 0.80) Other reported outcomes: • Total internal cancer incidence: RR 1.10 (95% CI 0.57 to 2.17); cases: selenium group: 21, placebo group: 19 |
|
| Selenium levels in exposure categories | d.n.a. | |
| Notes | Information on study design, which was not reported in Reid 2008, was taken from information available on the Nutritional Prevention of Cancer Trial. Adverse effects: not reported HR: adjusted for: age (continuous), smoking status (never, former, current), gender |
|
Ringstad 1988.
| Methods | Matched, nested case‐control study Country: Norway |
|
| Participants |
Participants: 9364 men and women
Inclusion criteria: 20 to 54 years of age (men), 20 to 49 years of age (women); inhabitants of Tromso; blood sample provided in 1979; no history of cancer at baseline
Name of parent cohort: Tromso Heart Study II Recruitment: 1979 to 1980 Outcome assessment: 1985 Number of cases: • Any cancer: 60 (male/female: 26/34) Case definition: incidence Years of follow‐up: 5 to 7 Type of selenium marker: serum |
|
| Interventions | d.n.a. | |
| Outcomes | Analysed cases: 60 of 72 (reason for non‐inclusion: no sample available) Statistical methods: n.r. Variables controlled by matching: age, gender, smoking status, month of sample collection, place of residence (district of Tromso) | |
| Risk estimates [95% CI] |
Reference category: highest 3 quartiles Results: Any cancer • Both genders: lowest quartile: OR 1.4 (95% CI 0.6 to 3.5) |
|
| Selenium levels in exposure categories | Lowest quartile: ≤ 114.49 µg/L Highest 3 quartiles: 114.50 to 114.51 µg/L | |
| Notes | ||
Sakoda 2005.
| Methods | Matched, nested case‐control study Country: China |
|
| Participants |
Participants: 41,563 men and women
Inclusion criteria: inhabitants of Haiman city of Chinese origin; written consent; toenail clipping available Recruitment: January 1993 to December 1993 Outcome assessment: 30 September 2000 Number of cases: • Primary liver cancer: 166 (male/female: 154/12) Case definition: mortality Years of follow‐up: 6.8 to 7.8 Type of selenium marker: toenail |
|
| Interventions | d.n.a. | |
| Outcomes | Analysed cases: 166 of 455 observed cases (only cases with questionnaire, blood sample, and toenail specimen analysed after 2000 owing to different methods of selenium analysis) Statistical methods: not specified Variables controlled in analysis: • Both genders: age, gender, HBsAg status, alcohol intake, history of acute hepatitis, occupation • Men: age, HBs‐Ag status, alcohol intake, history of acute hepatitis, family history of HCC, occupation • Women: HBs‐Ag status, age, history of acute hepatitis Variables controlled by matching: age, gender, township of residence | |
| Risk estimates [95% CI] |
Reference category: lowest quartile Results: Primary liver cancer • Both genders: highest quartile: OR 0.50 (95% CI 0.28 to 0.90) • Male: highest quartile: OR 0.57 (95% CI 0.31 to 1.05) • Female: highest 3 quartiles: OR 0.18 (95% CI 0.03 to 1.13) |
|
| Selenium levels in exposure categories |
Both genders and men
• Lowest quartile: 0 to 1.70 ppm
• Highest quartile: ≥ 4.43 ppm Women • Lowest quartile: 0.00 to 1.70 ppm • Highest 3 quartiles: ≥ 1.71 ppm |
|
| Notes | ||
Salonen 1984.
| Methods | Matched, nested case‐control study Country: Finland |
|
| Participants |
Participants: 8113 men and women
Inclusion criteria: 31 to 59 years of age; random sample of inhabitants of 2 Finnish provinces; initially free of cancer
Name of parent cohort: North Karelia Project Recruitment: February to April 1972 Outcome assessment: 31 December 1978 Number of cases: • Any cancer: 128 (male/female: n.r.) Case definition: incidence Years of follow‐up: 8.5 Type of selenium marker: serum |
|
| Interventions | d.n.a. | |
| Outcomes | Statistical methods: logistical regression/paired‐sample OR Variables controlled in analysis: tobacco consumption, serum cholesterol, beer consumption, dietary saturated fats, years of education, study area Variables controlled by matching: age, gender, smoking (tobacco use/d), total serum cholesterol | |
| Risk estimates [95% CI] |
Reference category: above 30th percentile Results: Any cancer • Both genders: ≤ 30th percentile: OR 3.1 (95% CI 1.5 to 7.7) • Both genders: ≤ 0 percentile: OR 3.0 (95% CI 1.2 to 21.9) |
|
| Selenium levels in exposure categories | 1st to 10th percentile ≤ 34.00 µg/L Above 30th percentile ≥ 45.00 µg/L | |
| Notes | ||
Salonen 1985.
| Methods | Matched, nested case‐control study Country: Finland |
|
| Participants |
Participants: 12,155 men and women
Inclusion criteria: 30 to 64 years of age; random sample of residents of 2 Finnish provinces; initially free of cancer
Name of parent cohort: North Karelia Project Recruitment: January to March 1977 Outcome assessment: 31 December 1980 Number of cases: • Any cancer: 51 (male/female: 30/21) Case definition: mortality Years of follow‐up: 3.7 Type of selenium marker: serum |
|
| Interventions | d.n.a. | |
| Outcomes | Analysed cases: 51 out of 56 (reason for non‐inclusion: no serum sample available) Statistical methods: logistical regression Variables controlled by matching: age, gender, smoking (tobacco use/d) | |
| Risk estimates [95% CI] |
Reference category: highest 2 tertiles Results: Any cancer • Both genders: lowest tertile: OR 5.8 (95% CI 1.2 to 29.0) |
|
| Selenium levels in exposure categories | Lowest tertile: ≤ 47.00 µg/L Highest 2 tertiles ≥ 47.10 µg/L | |
| Notes | ||
SELECT 2009.
| Methods | Randomised controlled trial SELECT (Selenium and Vitamin E Cancer Prevention Trial) Allocation: random, block/stratified by clinic Sequence generation: computer‐generated random numbers Concealment: central assignment (pill bottles) Blinding: participant blinded, doctor blinded, outcome assessor/pathologist blinded, review/coding of medical records blinded Dropouts/withdrawals: of 35,533 randomised participants, 645 were excluded from analysis because they had prior prostate cancer, did not give informed consent, or participated at 2 study sites that were excluded owing to management and regulatory issues Intention‐to‐treat‐analysis: yes Recruitment period: 22 August 2001 to 24 June 2004 End of study period: 1 August 2009 Blinded intervention was discontinued on 23 October 2008 following the recommendation of the Data Safety and Monitoring Committee after the second formal interim analysis in September 2008. Detection of cases: Participants had clinic visits once every 6 months and reported prostate cancers to the study staff. Study staff obtained medical records to verify the diagnosis. Tissue and the corresponding pathology report were sent to the central pathology laboratory for confirmation. Informed consent: yes |
|
| Participants |
Countries: United States, Canada, Puerto Rico Number of participants: 34,888 men, randomised to 4 groups: placebo (8696), vitamin E (8737), selenium (8752), selenium + vitamin E (8703) Condition: healthy men, aged 50 years or older (African American) or 55 years or older (all other), no prior diagnosis of prostate cancer, 4 ng/mL or less of PSA in serum, a digital rectal examination not suspicious for cancer, no current use of anticoagulant therapy other than 175 mg/d or less of acetylsalicylic acid, or 81 mg/d or less of acetylsalicylic acid with clopidogrel bisulphate, no history of haemorrhagic stroke, normal blood pressure Demographics: median age: 62.3 to 62.6 years in all 4 intervention groups, 79% white in all 4 intervention groups Recruitment and setting: 427 participating sites |
|
| Interventions | Group 1: placebo + placebo Group 2: 400 IU/d all rac‐alpha‐tocopheryl acetate + placebo Group 3: 200 µg/d L‐selenomethionine + placebo Group 4: 400 IU/d all rac‐alpha‐tocopheryl acetate + 200 µg/d L‐selenomethionine |
|
| Outcomes |
Primary outcome: incidence of prostate cancer as determined by routine clinical management Secondary outcomes: incidence of any cancer/lung cancer/colorectal cancer, diabetes mellitus, cardiovascular events, death from any cause |
|
| Risk estimates [95% CI] | Results are presented for the comparison of selenium alone (group 3) vs placebo (group 1) Primary outcome: • Prostate cancer: HR 1.04 (95% CI 0.90 to 1.18) (99% CI 0.87 to 1.24), cases: selenium 432 (5‐year rate: 4.56%), placebo 416 (5‐year rate 4.43%) Secondary outcomes: • Any cancer: HR 1.01 (95% CI 0.89 to 1.15) • Lung cancer: HR 1.12 (99% CI 0.73 to 1.72) • Colorectal cancer: HR 1.05 (99% CI 0.66 to 1.67) • Other primary cancer (excluding prostate cancer, basal cell and squamous cell skin cancer): HR 0.95 (99% CI 0.77 to 1.17) • Diabetes mellitus: HR 1.07 (99% CI 0.94 to 1.22) • Cardiovascular events: HR 1.02 (99% CI 0.92 to 1.13) • Deaths: HR 0.99 (99% CI 0.82 to 1.19) • Deaths from cancer: HR 1.02 (99% CI 0.74 to 1.41) |
|
| Selenium levels in exposure categories | d.n.a. | |
| Notes | Adverse effects: • Alopecia: RR 1.28 (99% CI 1.01 to 1.62) • Dermatitis grade 1 to 2: RR 1.17 (99% CI 1.00 to 1.35) • Dermatitis grade 3 to 4: RR 1.74 (99% CI 0.56 to 5.44) • Halitosis: RR 1.17 (99% CI 0.99 to 1.38) • Nail changes: RR 1.04 (99% CI 0.94 to 1.16) • Fatigue grade 1 to 2: RR 1.09 (99% CI 0.95 to 1.26) • Fatigue grade 3 to 4: RR 0.87 (99% CI 0.40 to 1.88) • Nausea grade 1 to 2: RR 1.19 (99% CI 0.94 to 1.52) • Nausea grade 3: RR 0.99 (99% CI 0.30 to 3.34) |
|
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Random sequence generation (selection bias) | Low risk | Random, block/stratified by clinic, computer‐generated random numbers |
| Allocation concealment (selection bias) | Low risk | Central assignment |
| Blinding (performance bias and detection bias) All outcomes | Low risk | Participants, doctors, outcomes |
| Selective reporting (reporting bias) | Low risk | No problems found |
Steevens 2010.
| Methods | Cohort/subcohort controlled cohort study Country: the Netherlands |
|
| Participants |
Name of parent cohort: Netherlands Cohort Study (NLCS) Recruitment: 1986 van den Brandt 1993b Participants: 120,852 (male/female: 58,279/ 62,573); aged 55 to 69 years; returned baseline questionnaire; no history of cancer at baseline Outcome assessment: 31 December 2002 Number of cases: • Oesophageal squamous cell carcinoma (ESCC): 64 (male/female: 40/24) • Oesophageal adenocarcinoma (EAC): 112 (male/female: 93/19) • Gastric cardia adenocarcinoma (GCA): 114 (male/female: 97/17) Case definition: incidence Years of follow‐up: 16.3 Type of selenium marker: toenail |
|
| Interventions | d.n.a. | |
| Outcomes |
Analysed cases: • Oesophageal squamous cell carcinoma (ESCC): 64 of 71 • Oesophageal adenocarcinoma (EAC): 112 of 129 • Gastric cardia adenocarcinoma (GCA): 114 of 127 Statistical methods: Cox proportional hazard model Variables controlled in analysis: age, sex, cigarette smoking (current yes/no, number of cigarettes smoked daily, and number of smoking years), alcohol consumption (g/d), and BMI (kg/m²) |
|
| Risk estimates [95% CI] |
Reference category: lowest quartile Results: Esophageal squamous cell carcinoma (ESCC) • Both genders: highest quartile: RR 0.37 (95% CI 0.16 to 0.86) • Men: highest quartile: RR 0.81 (95% CI 0.64 to 1.4) • Women: highest quartile: RR 0.79 (95% CI 0.63 to 0.99) Oesophageal adenocarcinoma (EAC) • Both genders: highest quartile: RR 0.76 (95% CI 0.41 to 1.40) • Men: highest quartile: RR 1.07 (95% CI 0.99 to 1.15) • Women: highest quartile: RR 0.72 (95% CI 0.61 to 0.84) Gastric cardia adenocarcinoma (GCA) • Both genders: highest quartile: RR 0.52 (95% CI 0.27 to 1.02) • Men: highest quartile: RR 0.94 (95% CI 0.84 to 1.06) • Women: highest quartile: RR 0.73 (95% CI 0.56 to 0.95) |
|
| Selenium levels in exposure categories | Lowest quartile: ≤ 0.498 µg/g Highest quartile: ≥ 0.613 µg/g | |
| Notes | ||
Steinbrecher 2010.
| Methods | Nested case‐control study Country: Germany |
|
| Participants |
Participants: 11,928 men (from the total cohort of 25,540 men and women) Name of parent cohort: EPIC‐Heidelberg cohort Recruitment: 1994 to 1998 Outcome assessment: 2/2007 Number of cases: • Prostate cancer: 248 Case definition: incidence Years of follow‐up: mean: 3 Type of selenium marker: serum |
|
| Interventions | d.n.a. | |
| Outcomes |
Statistical methods: conditional logistical regression
Variables controlled in analysis: family history of prostate cancer, participation in PSA testing, smoking status, and vigorous physical activity Variables controlled in matching: age group and time of recruitment |
|
| Risk estimates [95% CI] |
Reference category: lowest quartile Results: Prostate cancer • Highest quartile: OR 1.10 (95% CI 0.58 to 2.09) |
|
| Selenium levels in exposure categories | Lowest quartile: ≤ 78.9 µg/L Highest quartile: ≥ 95.0 µg/L | |
| Notes | ||
Suadicani 2012.
| Methods | Cohort study Country: Denmark |
|
| Participants |
Participants: 3333 males; male participants were derived from 14 workplaces in Copenhagen: the Air Force, Army, Navy, Emergency Management Agency, Postal Service, Customs Service, a railroad company, a national bank, a telephone company, 3 municipal service centres (for electricity and engineering and a fire brigade), a pharmaceutical company, and a building contractor company Name of parent cohort: Copenhagen male study Recruitment: from 1970 to 1971/1985 to 1986 Outcome assessment: 1985 to 1986/2001 Number of cases: • Deaths for lung cancer: 167 Case definition: death for lung cancer Years of follow‐up: 16 Type of selenium marker: serum |
|
| Interventions | d.n.a. | |
| Outcomes | Statistical methods: Cox logistical regression Variables controlled in analysis: age, pack‐years of smoking, spirits intake, and dietary markers | |
| Risk estimates [95% CI] |
Reference category: lowest exposure category: 0.4 to 1.0 µmol/L Results: Deaths from lung cancer • Highest exposure category: HR 1.43 (95% CI 0.96 to 2.14) |
|
| Selenium levels in exposure categories | Lowest category: 31.58 to 78.96 µg/L Highest category: 120.65 to 236.88 µg/L |
|
| Notes | ||
Sun 2016.
| Methods | Cohort study Country: China |
|
| Participants |
Name of parent cohorts: Shangai Men's Health Study (SMHS) and Shangai Women's Health Study (SWHS) Participants: 133,957 (male/female: 61,470/74,941) • SMHS: 61,480 men • SWHS: 74,941 women Inclusion criteria: • SMHS: men aged 40 to 74, residents in Shangai with no history of cancer • SWHS: women aged 40 to 70, residents in Shangai with no history of cancer Recruitment: • SMHS: April 2002 to June 2006 • SWHS: March 1997 to May 2000 Outcome assessment: 31 December 2012 Number of cases: 2603 (male/female: 1798/805) Case definition: mortality Years of follow‐up: • SMHS: median: 8.37 • SWHS: median: 13.90 Type of selenium marker: intake |
|
| Interventions | d.n.a | |
| Outcomes |
Analysed cases:
Cancer mortality: 2603 (male/female: 1798/805) Statistical methods: Cox proportional hazard model Variables controlled in analysis: age, birth cohort, education, income, marital status, occupation, body mass index, physical activity, total energy intake, dietary fat intake, supplement use, smoking status, drinking status, status with regard to history of hypertension, diabetes, coronary hearth disease, or stroke, family history of cancer and menopausal status (women only) |
|
| Risk estimates [95% CI] |
Reference category: lowest quintile Results: Cancer mortality • SMHS: highest quintile: HR 0.97 (95% CI 0.81 to 1.13) • SWHS: highest quintile: HR 0.90 (95% CI 0.77 to 1.05) |
|
| Selenium levels in exposure categories | SMHS: • Lowest quintile: < 19.36 µg/1000 kcal/d • Highest quintile: ≥ 31.92 µg/1000 kcal/d SWHS: • Lowest quintile: < 19.05 µg/1000 kcal/d • Highest quintile: ≥ 33.36 µg/1000 kcal/d | |
| Notes | ||
Thomson 2008.
| Methods | Cohort study Country: United States |
|
| Participants |
Participants: 133,614 women
Inclusion criteria: postmenopausal participants (aged 50 to 79 years) of the WHI clinical trial and observational study Name of parent cohort: Women's Health Initiative (WHI) Recruitment: n.r. Outcome assessment: December 2004 Number of cases: • Ovarian cancer: 451 Case definition: incidence Years of follow‐up: mean: 7 Type of selenium marker: supplemental selenium intake |
|
| Interventions | d.n.a. | |
| Outcomes | Statistical methods: Cox logistical regression Variables controlled in analysis: participation in observational or intervention study, age, log calories, number of relatives with breast/ovarian cancer, dietary modification randomisation arm, hysterectomy, minority race, pack‐years of smoking, physical activity, NSAID use, parity, infertility, duration of oral contraceptive use, number of lifetime ovulatory cycles, partial oophorectomy, age at menopause, hormone therapy at study entry | |
| Risk estimates [95% CI] |
Reference category: no intake of supplemental selenium (lowest exposure category) Results: Ovarian cancer • Highest exposure category: HR 1.00 (95% CI 0.73 to 1.37) |
|
| Selenium levels in exposure categories | Lowest exposure category: no supplemental selenium intake Highest exposure category: > 20 µg/d supplemental selenium intake |
|
| Notes | ||
van den Brandt 1993.
| Methods | Cohort/subcohort controlled cohort study Country: the Netherlands |
|
| Participants |
Name of parent cohort: Netherlands Cohort Study (NLCS) Recruitment: 1986 Case definition: incidence van den Brandt 1993b Participants: 120,852: 58,279 men and 62,573 women; aged 55 to 69 years; returned baseline questionnaire; no history of cancer at baseline Outcome assessment: n.r. Number of cases: • Stomach cancer: 104 (male/female: 84/20) • Colon cancer: 234 (male/female: 121/113) • Rectal cancer: 113 (male/female: 77/36) van den Brandt 1993a Participants: 120,852: 58,279 men and 62,573 women; age 55 to 69 years; returned baseline questionnaire; no history of cancer at baseline Outcome assessment: n.r. Number of cases: • Lung cancer: 370 (male/female: 335/35) van den Brandt 1994 Participants: 62,573 postmenopausal women Outcome assessment: 1989 Number of cases: • Breast cancer (postmenopausal): 355 (male/female: 0/355) • Breast cancer (postmenopausal), multi‐variate analysis: 270 (male/female: 0/270) Zeegers 2002 Participants: 120,852: 58,279 men and 62,573 women Outcome assessment: December 1992 Number of cases: • Bladder cancer: 431 (male/female: 372/59) van den Brandt 2003 Participants: 58,279 men Outcome assessment: n.r. (probably December 1992) Number of cases: • Prostate cancer: 540 (male/female: 540/0) Years of follow‐up: • 3.3 (Brandt 1993a; Brandt 1993b; Brandt 1994) • 6.3 (Zeegers 2002; Brandt 2003) Type of selenium marker: toenail |
|
| Interventions | d.n.a. | |
| Outcomes |
van den Brandt 1993b
Analysed cases: 234 of 351 colon cancer cases/104 of 176 stomach cancer cases/113 of 185 rectal cancer cases analysed (reasons for non‐inclusion: history of cancer at baseline not available, no pathological confirmation or CIS, no toenail clipping available)
Statistical methods: Mantel‐Haenszel
Variables controlled in analysis: age, gender van den Brandt 1993a Analysed cases: 370 of 617 (reasons for non‐inclusion: history of cancer at baseline not available, no toenail clipping, no pathological confirmation, problems with selenium measurement) Statistical methods: Mantel‐Haenszel Variables controlled in analysis: age, gender van den Brandt 1994 Analysed cases: 355 of 553 (reasons for non‐inclusion: history of cancer at baseline not available, CIS, no toenail sample or problems with selenium detection) Statistical methods: multi‐variate case‐cohort analysis Variables controlled in analysis: age, history of benign breast disease, maternal breast cancer, breast cancer in sister(s), age at menarche, age at menopause, oral contraceptive use, parity, age at first birth, body mass index, education, current cigarette smoking, alcohol intake, energy intake Zeegers 2002 Analysed cases: 431 of 619 (reason for non‐inclusion: no toenails available) Statistical methods: exponentially distributed failure time regression models Variables controlled in analysis: age, gender, number of cigarettes/d, years of cigarette smoking van den Brandt 2003 Analysed cases: 540 of 704 (reason for non‐inclusion: no toenail samples or selenium detection not possible) Statistical methods: exponentially distributed failure time regression models Variables controlled in analysis: age, family history of prostate cancer, number of cigarettes/d, years of cigarette smoking, level of education |
|
| Risk estimates [95% CI] |
Reference category: lowest quartile/quintile Results: van den Brandt 1993b Stomach cancer • Both genders: highest quintile: RR 0.61 (95% CI 0.33 to 1.11); highest quintile: RR 0.64 (95% CI 0.33 to 1.27) (max. adj.) • Men: highest quintile: RR 0.40 (95% CI 0.17 to 0.96) (max. adj.) • Women: highest quartile: RR 1.68 (95% CI 0.43 to 6.54) (max. adj.) Colon cancer • Both genders: highest quintile: RR 0.77 (95% CI 0.49 to 1.19); highest quintile: RR 0.80 (95% CI 0.50 to 1.29) (max. adj.) • Men: highest quintile: RR 0.82 (95% CI 0.43 to 1.58) (max. adj.) • Women: highest quintile: RR 0.77 (95% CI 0.41 to 1.45) (max. adj.) Rectal cancer • Both genders: highest quintile: RR 1.01 (95% CI 0.55 to 1.84); highest quintile: RR 1.05 (95% CI 0.54 to 2.03) (max. adj.) • Men: highest quintile: RR 0.91 (95% CI 0.41 to 2.00) (max. adj.) • Women: highest quartile: RR 1.58 (95% CI 0.59 to 4.22) (max. adj.) van den Brandt 1993a Lung cancer • Both genders: highest quintile: RR 0.40 (95% CI 0.27 to 0.59) • Men: highest quintile: RR 0.50 (95% CI 0.30 to 0.82) • Women: highest quartile: RR 0.40 (95% CI 0.13 to 1.24) van den Brandt 1994 Breast cancer • Multi‐variate analysis: highest quintile: RR 0.84 (95% CI 0.55 to 1.27) • Age‐stratified analysis: highest quintile: RR 0.93 (95% CI 0.65 to 1.33) Zeegers 2002 Bladder cancer • Both genders: highest quintile: RR 0.67 (95% CI 0.46 to 0.97) van den Brandt 2003 Prostate cancer • Highest quintile: RR 0.69 (95% CI 0.48 to 0.99) |
|
| Selenium levels in exposure categories |
van den Brandt 1993b
• Lowest quintile: ≤ 0.483 µg/g
• Highest quintile: ≥ 0.631 µg/g
• Lowest quartile: ≤ 0.497 µg/g
• Highest quartile: ≥ 0.613 µg/g van den Brandt 1993a Both genders and men • Lowest quintile: ≤ 0.483 µg/g • Highest quintile: ≥ 0.631 µg/g Women • Lowest quartile ≤ 0.497 µg/g • Highest quartile ≥ 0.613 µg/g van den Brandt 1994 Women • Lowest quintile: ≤ 0.499 µg/g • Highest quintile: ≥ 0.646 µg/g Zeegers 2002 • Lowest quintile: ≤ 0.483 µg/g • Highest quintile: ≥ 0.631 µg/g van den Brandt 2003 Men • Lowest quintile: ≤ 0.467 µg/g • Highest quintile: ≥ 0.617 µg/g |
|
| Notes | Primary publication:van den Brandt 1993b Other publications: Zeegers 2002, van den Brandt 1993a, van den Brandt 1994, van den Brandt 2003 | |
van Noord 1987.
| Methods | Matched, nested case‐control study Country: the Netherlands |
|
| Participants |
Participants: 8760 women
Inclusion criteria: 42 to 52 years of age; premenopausal; inhabitants of Utrecht
Name of parent cohort: DOM (Diagnostic onderzoek mammacarcinoom) Study Recruitment: n.r. Outcome assessment: 1 February 1986 Number of cases: • Breast cancer (premenopausal): 27 (male/female: 0/27) Case definition: incidence Years of follow‐up: 0.6 to 3.5, mean: 2.1 Type of selenium marker: toenail |
|
| Interventions | d.n.a. | |
| Outcomes |
Analysed cases: 7 detected during initial mammography screening in this study and not included in the analysis of incident cases Statistical methods: n.r. Variables controlled by matching: age, date of birth, premenopausal status |
|
| Risk estimates [95% CI] |
Reference category: lowest quartile Results: Breast cancer (premenopausal) • Highest quartile: OR 1.1 (95% CI 0.5 to 2.9) |
|
| Selenium levels in exposure categories | n.r. | |
| Notes | ||
Virtamo 1987.
| Methods | Cohort/subcohort controlled cohort study Country: Finland |
|
| Participants |
Participants: 1110 men
Inclusion criteria: 55 to 74 years of age; inhabitants of Finnish rural areas; participants of prior study on CHD; serum sample available: cases within first year of follow‐up excluded
Name of parent cohort: Men in rural East and West Finland Recruitment: 1974 Outcome assessment: 31 December 1983 Number of cases: • Any cancer: 109 (male/female: 109/0) Case definition: incidence Years of follow‐up: 10 Type of selenium marker: serum |
|
| Interventions | d.n.a. | |
| Outcomes | Statistical methods: conditional logistical regression Variables controlled in analysis: age, area of residence, smoking, serum cholesterol, alcohol intake | |
| Risk estimates [95% CI] |
Reference category: highest tertile Results: Any cancer • Lowest tertile OR 1.14 (95% CI 0.66 to 1.98) |
|
| Selenium levels in exposure categories | Lowest tertile: 15 to 46 µg/L Highest tertile: 60 to 136 µg/L | |
| Notes | ||
Walter 2011.
| Methods | Cohort study Country: United States |
|
| Participants |
Inclusion criteria: aged 50 to 76 years, recruited from subscribers to commercial mailing list, residents of western Washington state, non‐whites excluded, no malignant disease at baseline Name of parent cohort: Vitamins and Lifestyle (VITAL) study Number of participants: 66,227 men and women (male/female: n.r.) Recruitment: 1 October 2000 to 31 December 2002 Outcome assessment: 31/12/2008 Number of cases: • Haematological malignancies: 588 Case definition: incidence Years of follow‐up: mean: 6.5 years Type of selenium marker: supplemental intake (questionnaire: use of supplements over past 10 years, mean supplemental intake/d calculated) |
|
| Interventions | d.n.a. | |
| Outcomes | Statistical methods: Cox proportional hazard regression Variables controlled in analysis: sex, race/ethnicity (white, Hispanic, other), education (high school graduate or less, some college, college or advanced degree), smoking (pack‐years), self‐rated health (excellent, very good, good, fair, poor), vegetable servings per day (excluding potato servings); fruit servings per day; history of coronary artery disease (defined as history of heart attack, coronary bypass surgery, angioplasty, and/or angina; yes, no), history of rheumatoid arthritis (yes, no), history of fatigue or lack of energy over the year before baseline (yes, no), and number of first‐degree relatives with a history of leukaemia or lymphoma (none, 1, 2) | |
| Risk estimates [95% CI] |
Reference category: none Results: Highest level: RR 0.95 (95% CI 0.75 to 1.20) |
|
| Selenium levels in exposure categories | Lowest level: none Highest level: 20.1 to 400.0 µg/d |
|
| Notes | ||
Wei 2004.
| Methods | Frequency‐matched cohort‐controlled study Country: China |
|
| Participants |
Participants:Mark 2000: 29,584 men and women; Wei 2004: 1103 people who were originally selected as disease‐free controls in Mark 2000
Inclusion criteria: 40 to 69 years of age; healthy inhabitants of 4 Linxian communities; participants of a randomised controlled trial
Name of parent cohort: General Population Trial Linxian Recruitment: 1985 Outcome assessment: May 1991 (Mark 2000); n.r. (Wei 2004) Number of cases: Wei 2004 • Oesophageal cancer: 75 (male/female: 49/26) mortality • Stomach, cardia cancer: 36 (male/female: 22/14) mortality • Stomach, non‐cardia cancer: 24 (male/female: 20/4) mortality • Other: 32 (male/female: 22/10) mortality Mark 2000 • Oesophageal cancer: 590 (male/female: 286/304) incidence • Oesophageal cancer: 332 (male/female: n.r.) mortality • Stomach, cardia cancer: 402 (male/female: 239/163) incidence • Stomach, cardia cancer: 232 (male/female: n.r.) mortality • Stomach, non‐cardia cancer: 87 (male/female: 66/21) incidence • Stomach, non‐cardia cancer: 68 (male/female: n.r.) mortality Case definition: mortality, incidence Years of follow‐up: unclear/approximately 9 (Wei 2004), 6 (Mark 2000) Type of selenium marker: serum |
|
| Interventions | d.n.a. | |
| Outcomes |
Statistical methods: Cox‐proportional hazard model Variables controlled in analysis:Wei 2004: age, cholesterol, smoking, alcohol intake, BMI; Mark 2000: age Variables controlled by matching: age category, gender |
|
| Risk estimates [95% CI] |
Wei 2004
Reference category: lowest quartile Results: Oesophageal cancer • Both genders: highest quartile: RR 0.35 (95% CI 0.16 to 0.81) Stomach, cardia cancer • Both genders: highest quartile: RR 0.31 (95% CI 0.11 to 0.87) Stomach, non‐cardia cancer • Both genders: highest quartile: RR 1.64 (95% CI 0.49 to 5.48) Other cancers • Both genders: highest quartile: RR 1.95 (95% CI 0.66 to 5.81) Mark 2000 Reference category: lowest quartile Results: Oesophageal cancer • Both genders/incidence: highest quartile: RR 0.56 (95% CI 0.44 to 0.71) • Both genders/mortality: highest quartile: RR 0.62 (95% CI 0.44 to 0.89) Stomach, cardia cancer • Both genders/incidence: highest quartile: RR 0.47 (95% CI 0.33 to 0.65) • Both genders/mortality: highest quartile: RR 0.59 (95% CI 0.39 to 0.90) Stomach, non‐cardia cancer • Both genders/incidence: highest quartile: OR 1.07 (95% CI 0.55 to 2.08) • Both genders/mortality: highest quartile: OR 1.03 (95% CI 0.85 to 2.02) |
|
| Selenium levels in exposure categories |
Wei 2004
• Lowest quartile: 0.0 to 60.0 µg/L
• Highest quartile ≥ 84.5 µg/L Mark 2000 • Lowest quartile: 0.00 to 59.70 µg/L • Highest quartile ≥ 82.20 µg/L |
|
| Notes |
Primary publication: Wei 2004
Other publication: Mark 2000 Remark: Wei 2004 measured serum selenium in a subcohort derived from 29,584 male and female participants of the Linxian Population Trial. The earlier publication of this study, Mark 2000, reported 332 fatal cases and 590 incident cases. The later publication, Wei 2004, reported deaths from oesophageal cancer among disease‐free controls in Mark 2000 and analysed 75 fatal cases. |
|
Willett 1983.
| Methods | Matched, nested case‐control study Country: United States |
|
| Participants |
Participants: 10,940 men and women
Inclusion criteria: 30 to 69 years of age; serum sample available (only 4480 samples of cohort were available because of freezer breakdown); participants of an RCT on hypertension; institutionalised and bedfast people excluded
Name of parent cohort: Hypertension Detection Follow‐Up Programme (HDFP) Recruitment: 1973 to 1974 Outcome assessment: n.r. Number of cases: • Any cancer: 111 (male/female: 60/51) Case definition: incidence Years of follow‐up: 5 Type of selenium marker: serum |
|
| Interventions | d.n.a. | |
| Outcomes | Statistical methods: logistical regression of unmatched data Variables controlled by matching: age, gender, race/ethnicity, smoking status, year/month of sample collection, initial blood pressure, use of antihypertensive medication, randomisation group • In women: parity, menopausal status | |
| Risk estimates [95% CI] |
Reference category: highest quintile, highest 3 quintiles Results: Any cancer • Both genders: lowest quintile vs highest quintile: OR 2.0 (CI not reported) • Both genders: lowest quintile vs highest 3 quintiles: OR 1.9 (95% CI 1.1 to 3.3) |
|
| Selenium levels in exposure categories | Lowest quintile: ≤ 114 µg/L Highest quintile: ≥ 154 µg/L | |
| Notes | ||
Yoshizawa 1998.
| Methods | Matched, nested case‐control study Country: United States |
|
| Participants |
Participants: 33,737 men
Inclusion criteria: 40 to 75 years of age; physicians from all 50 US states; provision of toenails in 1987 and completed baseline questionnaire in 1986; exclusion of histologically confirmed prostate cancer at baseline, and cases within first 2 years of follow‐up
Name of parent cohort: Health Professionals Follow‐Up Study (HPFS) Recruitment: 1986 to 1987 Outcome assessment: 1994 Number of cases: • Prostate cancer: 181 (male/female: 181/0) Case definition: incidence Years of follow‐up: 8 to 9 Type of selenium marker: toenail |
|
| Interventions | d.n.a. | |
| Outcomes | Statistical methods: logistical regression, conditional logistical regression Variables controlled in analysis: quintiles of lycopene, saturated fat, calcium, family history of prostate cancer, BMI, vasectomy Variables controlled by matching: age, smoking status, year/month of sample collection | |
| Risk estimates [95% CI] |
Reference category: lowest quintile Results: Prostate cancer (advanced) • Highest quintile: OR 0.39 (95% CI 0.18 to 0.84) |
|
| Selenium levels in exposure categories | Lowest quintile: 0.530 to 0.730 µg/g Highest quintile: 0.941 to 7.090 µg/g | |
| Notes | ||
Yu 1991.
| Methods | Randomised controlled trial Allocation: random Sequence generation: unclear, not described Concealment: unclear, not described Blinding: described as double‐blind; blinding of participants: adequate, placebo tablets; blinding of investigators and doctors: unclear Dropouts/withdrawals: unclear, not described Intention‐to‐treat‐analysis: unclear, not described Recruitment period: unclear, not described Observation period: 2 years Study period: 2 years Detection of cases: unclear, use of "national standards" for the diagnosis of liver cancer Informed consent: unclear, not described |
|
| Participants |
Country: China Number of participants: 2474 Condition: first‐degree relatives within 3 generations of families with 2 or more cases of liver cancer during the period 1972 to 1985 Demographics: gender distribution not reported; age: 15 to 75 years Recruitment and setting: Participants were residents in Qidong province. |
|
| Interventions |
Intervention: 200 µg selenium as selenised yeast p.o. daily, intervention period unclear Control: placebo |
|
| Outcomes |
Primary outcome measure: incidence of primary liver cancer within 2 years after start of intervention Results: • 13 cases in 1030 placebo participants • 10 cases in 1444 selenium participants |
|
| Risk estimates [95% CI] | n.r. | |
| Selenium levels in exposure categories | d.n.a. | |
| Notes | Data were extracted from Yu 1991. We identified 2 later publications (Li 1992, Yu 1993), which we assumed to report on the same trial as Yu 1991. However, the total number of participants differed from the initial report (N = 3849 in the later publications, with 1485 receiving placebo and 2364 receiving selenium). The total number of cases was not reported in either Li 1992 or Yu 1993. Reported results were as follows: Li 1992 Person‐year incidence rate in intervention and control groups: • Within 1 year of follow‐up: selenium group 175.36/100,000; placebo group: 414.65/100,000 • Within 2 years of follow‐up: selenium group 219.37/100,000; placebo group: 553.15/100,000 Yu 1993 Cumulated incidence • After 1 year: selenium group 1.75/1000; placebo group: 4.15/1000 • After 2 years: selenium group 2.19/1000; placebo group: 5.53/1000 • We could not make contact with study investigators to clarify these discrepancies. As we could not clarify the actual number of liver cancer cases in the later publications, we decided to use the data of Yu 1991 for this review. • Adverse effects were not mentioned in Yu 1991 or Li 1992. Yu 1993 stated that no cases of selenosis were observed in the trial. |
|
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Random sequence generation (selection bias) | Unclear risk | Sequence generation not described |
| Allocation concealment (selection bias) | Unclear risk | Not described |
| Blinding (performance bias and detection bias) All outcomes | Low risk | Participants blinded, doctors stated only as double‐blind |
| Selective reporting (reporting bias) | Unclear risk | Recruitment period unclear; dropout unclear |
Yu 1997.
| Methods | Randomised controlled trial Allocation: random Sequence generation: unclear, not described Concealment: unclear, not described Blinding: of participants: adequate (placebo); of investigators and doctors: unclear, not described Dropouts/withdrawals: unclear, not described Recruitment period: unclear, not described Intention‐to‐treat‐analysis: unclear, not described Observation period: 1987 to 1994 Intervention period: 1987 to 1990 Detection of cases: unclear, monthly blood sample during follow‐up for liver enzymes (SGPT, ZnTT), use of "national standards" for the diagnosis of liver cancer Informed consent: unclear, not described |
|
| Participants |
Country: China Number of participants: 226 (selenium group: 113; placebo group 113) Condition: HBs‐antigen carriers with normal liver function Demographics: 95 men, 131 women; age: 21 to 63 years Recruitment and setting: recruitment “through screening in a village in the city Qidong” (Li 1992) |
|
| Interventions |
Intervention: 200 µg selenium as selenised yeast p.o. daily for 4 years Control: placebo |
|
| Outcomes |
Primary outcome measure: incidence of primary liver cancer (defined as increase in SGPT and ZnTT) Results: At end of intervention period • 0 cases in the selenium group • 7 cases in the placebo group for a total of 445 person‐years of observation (person‐time incidence rate: 1573.03/100,000) |
|
| Risk estimates [95% CI] | n.r. | |
| Selenium levels in exposure categories | d.n.a. | |
| Notes |
Adverse effects: "No side effects have been found in these trials" (Yu 1997, p. 124) Further data reported in: Li 1992 (Chinese, translated); Yu 1991 In Yu 1991, a different incidence was reported for the selenium group (5 cases). We could not clarify this discrepancy with later papers Li 1992 and Yu 1997. |
|
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Random sequence generation (selection bias) | Unclear risk | Sequence generation not described |
| Allocation concealment (selection bias) | Unclear risk | Not described |
| Blinding (performance bias and detection bias) All outcomes | Low risk | Participants blinded, doctors stated only as double‐blind |
| Selective reporting (reporting bias) | Unclear risk | Recruitment period unclear; dropout unclear |
Yu 1999.
| Methods | Matched, nested case‐control study Country: China (Taiwan) |
|
| Participants |
Participants: 4841 men
Inclusion criteria: 30 to 65 years of age; HBs‐Ag‐positive or/and HCV‐positive; recruited at 2 centres: Government Employee Central Clinics and Liver Unit of Chang‐Gung Memorial Hospital Recruitment: August 1988 to June 1992 Outcome assessment: 31 December 1996 Number of cases: • Primary liver cancer: 69 (male/female: 69/0) Case definition: incidence Years of follow‐up: 4.5 to 8.3 Type of selenium marker: plasma |
|
| Interventions | d.n.a. | |
| Outcomes | Analysed cases: 69 of 73 (reason for non‐inclusion: no sample available) Statistical methods: conditional logistical regression Variables controlled in analysis: age, cigarette smoking, alcohol intake, plasma levels of retinol/alpha‐tocopherol/alpha‐carotene/beta‐carotene/lycopene Variables controlled by matching: age, year and season of sample collection, recruitment clinic | |
| Risk estimates [95% CI] |
Reference category: lowest quintile Results: Primary liver cancer • Highest quintile: OR 0.62 (95% CI 0.21 to 1.86) |
|
| Selenium levels in exposure categories | Lowest quintile ≤ 124.90 µg/L Highest quintile ≥ 162.40 µg/L | |
| Notes | ||
µ: micro. AFP: alpha‐fetoprotein. ALT: alanine aminotransferase. ATBC: alpha‐tocopherol, beta‐carotene cancer prevention study. AU: arbitrary unit. AUA: American Urological Association. BCC: basal cell carcinoma. BMI: body mass index. BPH: benign prostate hyperplasia. CARET: Carotene and Retinol Efficacy Trial. CHD: coronary heart disease. CI: confidence interval. CIS: carcinoma in situ. CSDLH: Canadian Study of Diet, Lifestyle and Health. CT: computed tomography. CVD: cardiovascular disease. dL: deciliter. d.n.a.: does not apply. DOM: Diagnostic onderzoek mammacarcinoom. DSMC: Data and Safety Monitoring Committee. ECOG: Eastern Cooperative Oncology Group.EPIC: European Prospective Investigation of Cancer.EVA: Etude du Vieillissement Antériel. EPOZ: Epidemiologisch onderzoek naar risico‐indicatoren voor hart‐ en vaatziekten. FFQ: food‐frequency questionnaire. g: gram. GBTC: gallbladder and biliary tract cancer. HBs‐Ag: hepatitis B surface antigen. HCC: hepatocellular carcinoma. HCV: hepatitis C virus. HGPIN: high‐grade prostatic intraepithelial neoplasia. HPFP: Hypertension Detection Follow‐up Programme. HPFS: Health Professionals Follow‐up Study. HR: hazard ratio. HRT: hormone replacement therapy. IHBC: intrahepatic bile duct cancer. IRR: incident rate ratio. IU: international unit. L: litre. m: milli. max. adj.: maximally adjusted. MHC: Mobile Health Clinic. n: nano. NHS: Nurses‘ Health Study. NLCS: Netherlands Cohort Study. NMSC: non‐melanoma skin cancer. NPCT: Nutritional Prevention of Cancer Trial. n.r.: not reported. NSAID: non‐steroidal anti‐inflammatory drug. OR: odds ratio. p.: page. p.o.: per os. ppm: parts per million. PSA: prostate‐specific antigen. RCT: randomised controlled trial. RR: risk ratio. SCC: squamous cell carcinoma. SD: standard deviation. SGPT: alanine aminotransferase. TIA: transient ischaemic attack. UK: United Kingdom. USA: United States of America. VITAL: Vitamins and Lifestyle study. WHI: Women's Health Initiative. ZnTT: zinc turbidity test.
Characteristics of excluded studies [ordered by study ID]
| Study | Reason for exclusion |
|---|---|
| Albanes 2014 | Same results of SELECT 2009, stratified according to tocopherol status |
| Bates 2011 | Results not reported according to inclusion criteria: HR estimated per SD increase of selenium level reported |
| Bostick 1993 | Cohort study: Iowa Women's Health Study cohort Selenium exposure not assessed according to eligibility: only intake of selenium supplements yes/no in questionnaire assessed |
| Brock 1991 | Case‐control study with precancerous condition (carcinoma in situ of the cervix) |
| Chen 1988 | Case‐control study |
| Chen 2003 | Case‐control study |
| Connelly‐Frost 2009 | Case‐control study |
| Costello 2001 | APPOSE (Australian Prostate Cancer Prevention Trial Using Selenium): Publication describes study design; trial was not started. |
| Criqui 1991 | Population‐based, prospective case‐control study: Lipid Research Clinic Prevalence and Follow‐Up study Results not reported according to inclusion criteria: differences in mean selenium levels reported |
| Cui 2007 | Nested case‐control study Selenium exposure not assessed according to eligibility: selenium measurement conducted in tissue of benign breast disease |
| Davies 2002 | Nested case‐control study: EPIC Norfolk study cohort Results not reported according to inclusion criteria: RR estimate per unit increase in selenium level reported |
| Epplein 2014 | Results not reported according to inclusion criteria: selenium not reported as independent variable ‐ only selenoprotein P |
| Fleshner 2003 | Randomised Study of Vitamin E, Selenium, and Soy Protein Isolate in Patients with High‐Grade Prostatic Intraepithelial Neoplasia: Multi‐component Intervention |
| Geybels 2013 | Same population as van den Brandt 1993, restricted only to advanced prostate cancer cases |
| Geybels 2014 | Same population as Geybels 2013, stratified according to genetic variation in SePP1 and GPX1 |
| Hagmar 1992 | Historical cohort study |
| Harris 2012 | Cancer was not a study endpoint. |
| Hartman 2002 | Nested case‐control study: ATBC cohort Results not reported according to inclusion criteria: differences in mean selenium levels reported; OR reported as graph and could not be calculated from reported data |
| Huzarski 2006 | Interventional study without control group with 1489 female participants with BRCA1 mutation who received a selenium‐containing nutritional supplement |
| Joniau 2007 | Intervention study without control group with male participants with high‐grade intraepithelial neoplasia of the prostate who received a selenium‐containing nutritional supplement |
| Karunasinghe 2012 | Results not reported according to inclusion criteria: differences in mean selenium levels reported |
| Kellen 2008 | Case‐control study |
| Kilander 2001 | Cohort study in Uppsala/Sweden Results not reported according to inclusion criteria: RR estimate per unit increase in selenium level reported |
| Knekt 1988a | Nested case‐control study: Mobile Health Clinic cohort Results not reported according to inclusion criteria: differences in mean selenium levels reported |
| Knekt 1988b | Nested case‐control study: Mobile Health Clinic cohort Results not reported according to inclusion criteria: differences in mean selenium levels reported |
| Knekt 1991 | Nested case‐control study: Mobile Health Clinic cohort Results not reported according to inclusion criteria: differences in mean selenium levels reported |
| Kok 1987b | Nested case‐control study: Zoetermeer cohort Results not reported according to inclusion criteria: differences in mean selenium levels reported |
| Kune 2006 | Case‐control study |
| Kuroda 1988 | Case‐control study |
| Lane 2017 | Some study participants had cancer at baseline. |
| Lawson 2007 | Cohort study on multi‐vitamin use and risk of prostate cancer |
| Le Marchand 2006 | Case‐control study |
| Li 2004b | RCT for gastric cancer prevention with multi‐component intervention (200 mg synthetic allitridum and 100 µg selenium per day) |
| Limburg 2005 | Randomised controlled trial: Primary endpoint in this 2‐by‐2 factorial design trial with selenomethionine 200 µg daily and/or celecoxib 200 mg twice daily was the per‐participant change (regression, stable, progression) in pre‐existing oesophageal dysplasia ‐ cancer incidence and mortality were not endpoints in this study. |
| Linxian Pilot 2000 | Randomised controlled trial of selenium supplements and celecoxib in participants with oesophageal squamous dysplasia in Linxian, China Endpoint was "regression of disease"; cancer was not an endpoint in this investigation. |
| Loeb 2015 | Selenium exposure not assessed according to eligibility: only intake of selenium supplements yes/no on questionnaire assessed |
| Martinez 2014 | Same participants as SELECT 2009, stratified according to NKX3.1 genetic variant |
| Neuhouser 2009 | Cohort study (Women's Health Initiative) on multi‐vitamin use and risks of cancer and cardiovascular disease No data reported for selenium and cancer risk |
| Persson 2000 | Selenium exposure not assessed according to eligibility |
| Ray 2006 | Cohort study (Women's Health and Aging Studies I and II) on selenium and carotenoid serum levels and mortality No data reported for selenium and cancer mortality |
| Rayman 2001 | PRECISE trial (Prevention of Cancer by Intervention with Selenium): Trial has been stopped. |
| Rendon | Randomised controlled trial: Vitamin E, Selenium, and Soy Protein in Preventing Cancer in Patients With High‐Grade Prostate Neoplasia: Multi‐component Intervention |
| Steevens 2010b | Cancer was not a study endpoint. |
| Thompson 2009 | Cohort study: Iowa Women's Health Study cohort Selenium exposure was not assessed according to eligibility; only intake of selenium supplements yes/no on questionnaire was assessed. |
| Tsugane 1996 | Case‐control and cross‐sectional studies |
| Ujiie 2002 | Part of this study is a prospective cohort study in Miyagi/Japan. Results were not reported according to inclusion criteria; differences in mean selenium levels were reported. |
| van Noord 1992 | Nested case‐control study: DOM cohort Results were not reported according to inclusion criteria; differences in mean selenium levels were reported. |
| van Noord 1993 | Nested case‐control study: DOM II cohort Results were not reported according to inclusion criteria; RR estimate per unit increase in selenium levels were reported. |
| van't Veer 1996 | Case‐control study |
| Wallace 2009 | Case‐control study |
| Watters 2009 | Cohort study on smoking and prostate cancer risk. Selenium was not reported as an independent variable. |
| Wright 2004 | Cohort study: ATBC cohort Exposure to antioxidants was assessed via a self‐developed index. |
| You 2005 | Randomised controlled trial to test retardation of progression of precancerous gastric lesions among 3400 adults in Shandong, China. Intervention: vitamin C, vitamin E, selenium, garlic preparation Multi‐component intervention |
| Yuan 2006 | Nested case‐control study: Shanghai cohort study No data reported on selenium and cancer risk |
| Zeegers 2009 | Cohort study on factors influencing recurrence or progression of bladder cancer: West Midlands Bladder Cancer Prognosis Programme |
µ: micro. APPOSE: Australian Prostate Cancer Prevention Trial Using Selenium. ATBC: alpha‐tocopherol, beta‐carotene cancer prevention study. BRCA: breast cancer. DOM: Diagnostic Onderzoek Mammacarcinoom. EPIC: European Prospective Investigation of Cancer. m: milli. g: gram. OR: odds ratio. PRECISE: Prevention of Cancer by Intervention with Selenium. RCT: randomised controlled trial. SELECT: Selenium and Vitamin E Cancer Prevention Trial.
Characteristics of ongoing studies [ordered by study ID]
Argos 2013.
| Trial name or title | Bangladesh Vitamin E and Selenium Trial (BEST) |
| Methods | Double‐blind, placebo‐controlled, 2‐by‐2 factorial, randomised controlled trial |
| Participants | 7000 adults having manifest arsenical skin lesions in Bangladesh Inclusion • Manifest arsenical skin lesions • Aged 25 to 65 years • Signed informed consent Exclusion • Currently pregnant • Not a permanent resident of study area • Unwillingness to discontinue current vitamin use • History of cancer • Too ill to participate • Unwillingness to provide biological samples (blood and urine) |
| Interventions | 6‐year supplementation, divided into 4 study arms: • Vitamin E (alpha‐tocopherol, 100 mg daily) • Selenium (L‐selenomethionine, 200 μg daily) • Vitamin E and selenium • Placebo |
| Outcomes |
Primary endpoints • Prevention of non‐melanoma skin cancer Secondary endpoints: • All‐cause and cancer mortality • Diabetes mellitus • Oxidative stress biomarkers |
| Starting date | April 2006 |
| Contact information | Dr. Habibul Ahsan Center for Cancer Epidemiology and Prevention, The University of Chicago 5841 South Maryland Avenue, MC 2007 Chicago, IL 60637 |
| Notes |
BEST: Bangladesh Vitamin E and Selenium Trial.
Differences between protocol and review
In the previous Cochrane review, review authors adapted the risk of bias assessment for RCTs, which was introduced by Cochrane after publication of our protocol; we used the Jadad score and the Delphi list to assess the quality of RCTs, but because the results of these checklist assessments were of no relevance for this review, we have omitted them.
With respect to the protocol, in this second updated review (as well as in the previous update), we decided to perform meta‐analysis of RCTs when at least two studies were available, and to emphasise the analysis conducted for all RCTs and for RCTs at low risk of bias, to highlight the most reliable and recent evidence on the selenium and cancer relation, which comes from well‐designed experimental studies. As in the previous version of the review, we included in our analysis both primary and secondary outcomes of RCTs, as well as adverse effects reported in these studies. Furthermore, we updated the methods section to clarify that the main ‘primary’ analysis included analyses examining low risk of bias trials only, and ‘sensitivity analyses’ consisted of analyses that included all trials, regardless of risk of bias.
In this update, we included a 'Summary of findings' table for RCTs with low risk of bias, and one for observational studies.
Contributions of authors
MV co‐ordinated the current update, commented on the protocol and the review, screened search results, appraised study bias, and updated the draft in collaboration with the other review authors.
TF and CDG extracted data from the added papers, appraised study bias, conducted data analyses, commented on the review, wrote part of the draft, and provided a methodological perspective.
CDG commented on the review, appraised study bias, prepared the 'Summary of findings' (GRADE) tables, wrote part of the draft, and provided a methodological perspective.
GD is the primary author of the first version of the review and was involved in all steps of the present update, including commenting on the protocol and the manuscript, extracting data from papers, and providing a methodological perspective.
MZw commented on the protocol and the review and provided a methodological perspective.
MB commented on the protocol and provided feedback at various stages of the review.
MZe commented on the protocol and the review and provided feedback on different portions of these documents.
MH commented on the protocol, extracted data from papers, and commented on the review text at various stages of the review.
RDA commented on the protocol and provided feedback at various stages of the review.
CMC commented on the protocol and on the review, wrote part of the draft, and provided a methodological perspective.
All review authors have reviewed and approved the final draft of this update.
Sources of support
Internal sources
-
Department of Biomedical, Metabolic and Neural Sciences, University of Modena and Reggio Emilia Medical School, Modena, Italy.
This work was funded in part by the Department of Biomedical, Metabolic and Neural Sciences, University of Modena and Reggio Emilia, Modena. The funding source had no role in designing, conducting, or writing this systematic review. The contents of this systematic review are solely the responsibility of the review authors and do not necessarily represent the official views of this Department.
External sources
-
Dr. Ernst und Anita Bauer Foundation, Germany.
This work was partially funded by the Dr. Ernst and Anita Bauer Foundation. The funding source had no role in designing, conducting, or writing this systematic review. The contents of this systematic review are solely the responsibility of the review authors and do not necessarily represent the official views of the Dr. Ernst and Anita Bauer Foundation.
-
EU (European Union) Project: Concerted action for complementary and alternative medicine in the cancer field (EU CAM‐Cancer) (Contract no.: QLG4‐CT‐2002‐00786), Other.
This work was funded in part by the EU CAM‐Cancer Project. The funding source had no role in designing, conducting, or writing this systematic review. The contents of this systematic review are solely the responsibility of the review authors and do not necessarily represent the official views of the EU CAM‐Cancer Project.
-
Deutsche Krebshilfe (German Cancer Aid), Germany.
This work was funded in part by the German Cancer Aid. The funding source had no role in designing, conducting, or writing this systematic review. The contents of this systematic review are solely the responsibility of the review authors and do not necessarily represent the official views of the German Cancer Aid.
-
NCCAM, USA.
This work was funded in part by Grant Number R24 AT001293 from the National Center for Complementary and Alternative Medicine (NCCAM). The funding source had no role in designing, conducting, or writing this systematic review. The contents of this systematic review are solely the responsibility of the review authors and do not necessarily represent the official views of the NCCAM or the National Institutes of Health.
-
NIH, USA.
This work was partially funded by Grant Number CA16042 from the National Institutes of Health, National Cancer Institute (NCI). The funding source had no role in designing, conducting, or writing this systematic review. The contents of this systematic review are solely the responsibility of the review authors and do not necessarily represent the official views of the NCI or the National Institutes of Health.
-
Italian League against Cancer‐LILT (Reggio Emilia), Italy.
This work was funded in part by the Italian League against Cancer (LILT), Reggio Emilia section. The funding source had no role in designing, conducting, or writing this systematic review. The contents of this systematic review are solely the responsibility of the review authors and do not necessarily represent the official views of LILT‐Reggio Emilia.
-
Fondazione Pietro Manodori (Reggio Emilia), Italy, Other.
This work was funded in part by the Fondazione Pietro Manodori of Reggio Emilia. The funding source had no role in designing, conducting, or writing this systematic review. The contents of this systematic review are solely the responsibility of the review authors and do not necessarily represent the official views of the Fondazione Pietro Manodori.
Fondazione di Vignola, Italy.
Declarations of interest
MV: none known.
TF: none known.
CDG: none known.
GD: none known.
MZw: none known.
MB: none known.
MZe: Maurice Zeegers is the first investigator and the coauthor of included observational and experimental studies.
MH: none known.
RDA: none known.
CMC: none known.
Edited (no change to conclusions)
References
References to studies included in this review
Agalliu 2011 {published data only}
- Agalliu I, Kirsh VA, Kreiger N, Soskolne CL, Rohan TE. Oxidative balance score and risk of prostate cancer: results from a case‐cohort study. Cancer Epidemiology 2011;35(4):353‐61. [DOI] [PubMed] [Google Scholar]
Akbaraly 2005 {published data only}
- Akbaraly NT, Arnaud J, Hininger‐Favier I, Gourlet V, Roussel AM, Berr C. Selenium and mortality in the elderly: results from the EVA study. Clinical Chemistry 2005;51(11):2117‐23. [DOI] [PubMed] [Google Scholar]
Algotar 2013 {published data only}
- Algotar AM, Stratton MS, Ahmann FR, Ranger‐Moore J, Nagle RB, Thompson PA, et al. Phase III clinical trial investigating the effect of selenium supplementation in men at high‐risk for prostate cancer. The Prostate 2013;73(3):328‐35. [DOI] [PMC free article] [PubMed] [Google Scholar]
Allen 2008 {published data only}
- Allen NE. Reply to B Bekaert and MP Rayman. American Journal of Clinical Nutrition 2009;89(4):1277. [DOI] [PubMed] [Google Scholar]
- Allen NE, Appleby PN, Roddam AW, Tjønneland A, Johnsen NF, Overvad K, et al. Plasma selenium concentration and prostate cancer risk: results from the European Prospective Investigation into Cancer and Nutrition (EPIC). American Journal of Clinical Nutrition 2008;88(6):1567‐75. [DOI] [PubMed] [Google Scholar]
Banim 2013 {published data only}
- Banim PJR, Luben R, McTaggart A, Welch A, Wareham N, Khaw KT, et al. Dietary antioxidants and the aetiology of pancreatic cancer: a cohort study using data from food diaries and biomarkers. Gut 2013;62(10):1489‐96. [DOI] [PubMed] [Google Scholar]
- Barrass B, Rochester M, Luben R, Khaw KT, Hart A. Dietary selenium and lower risk of developing renal cancer ‐ a prospective cohort study using food diaries in epic Norfolk. Journal of Urology 2013;189(4):E248. [Google Scholar]
Bleys 2008 {published data only}
- Bleys J, Navas‐Acien A, Guallar E. Serum selenium levels and all‐cause, cancer, and cardiovascular mortality among US adults. Archives of Internal Medicine 2008;168(4):404‐10. [DOI] [PubMed] [Google Scholar]
Brooks 2001 {published data only}
- Brooks JD, Metter EJ, Chan DW, Sokoll LJ, Landis P, Nelson WG, et al. Plasma selenium level before diagnosis and the risk of prostate cancer development. Journal of Urology 2001;166(6):2034‐8. [PubMed] [Google Scholar]
Clark 1985 {published data only}
- Clark L, Graham G, Bray J. Nonmelanoma skin cancer and plasma selenium: a prospective cohort study. American Journal of Epidemiology 1985;122(3):528. [Google Scholar]
Coates 1988 {published data only}
- Coates RJ. Cancer risk in relation to serum levels of selenium, retinol, and copper. Dissertation Abstract International (Sci) 1987; Vol. 47, issue 12:4836.
- Coates RJ, Weiss NS, Daling JR, Morris JS, Labbe RF. Serum levels of selenium and retinol and the subsequent risk of cancer. American Journal of Epidemiology 1988;128(3):515‐23. [DOI] [PubMed] [Google Scholar]
Combs 1993 {published data only}
- Combs GF Jr, Clark LC, Turnbull BW, Graham GF, Smith CL, Sanders B Jr, et al. Low plasma selenium (Se) predicts the 24 month incidence of squamous cell carcinoma of the skin in a cancer prevention trial. FASEB 1993; Vol. 7, issue 3:A278.
Comstock 1997 {published data only}
- Comstock GW, Alberg AJ, Huang HY, Wu K, Burke AE, Hoffman SC, et al. The risk of developing lung cancer associated with antioxidants in the blood: ascorbic acid, carotenoids, alpha‐tocopherol, selenium, and total peroxyl radical absorbing capacity. Cancer Epidemiology, Biomarkers & Prevention 1997;6(11):907‐16. [PubMed] [Google Scholar]
Dong 2008 {published data only}
- Dong LM, Kristal AR, Peters U, Schenk JM, Sanchez CA, Rabinovitch PS, et al. Dietary supplement use and risk of neoplastic progression in esophageal adenocarcinoma: a prospective study. Nutrition and Cancer 2008;60(1):39‐48. [DOI] [PMC free article] [PubMed] [Google Scholar]
Dorgan 1998 {published data only}
- Dorgan JF, Sowell A, Swanson CA, Potischman N, Miller R, Schussler N, et al. Relationships of serum carotenoids, retinol, alpha‐tocopherol, and selenium with breast cancer risk: results from a prospective study in Columbia, Missouri (United States). Cancer Causes and Control 1998;9(1):89‐97. [DOI] [PubMed] [Google Scholar]
Dreno 2007 {published data only}
- Dreno B, Euvrard S, Frances C, Moyse D, Nandeuil A. Effect of selenium intake on the prevention of cutaneous epithelial lesions in organ transplant recipients. European Journal of Dermatology 2007;17(2):140‐5. [DOI] [PubMed] [Google Scholar]
Epplein 2009 {published data only}
- Epplein M, Franke AA, Cooney RV, Morris JS, Wilkens LR, Goodman MT, et al. Association of plasma micronutrient levels and urinary isoprostane with risk of lung cancer: the multiethnic cohort study. Cancer Epidemiology, Biomarkers & Prevention 2009;18(7):1962‐70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gill JK, Franke AA, Steven MJ, Cooney RV, Wilkens LR, Marchand L, et al. Association of selenium, tocopherols, carotenoids, retinol, and 15‐isoprostane F(2t) in serum or urine with prostate cancer risk: the multiethnic cohort. Cancer Causes Control 2009;20(7):1161‐71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kolonel LN, 5P01CA033619‐20. Epidemiologic studies of diet and cancer in Hawaii. http://cancercontrol.cancer.gov/grants/abstract.asp?ApplID=6805844 (accessed 6 April 2011).
Fex 1987 {published data only}
- Fex G, Pettersson B, Akesson B. Low plasma selenium as a risk factor for cancer death in middle‐aged men. Nutrition and Cancer 1987;10(4):221‐9. [DOI] [PubMed] [Google Scholar]
Fujishima 2011 {published data only}
- Fujishima Y, Ohsawa M, Itai K, Kato K, Tanno K, Turin TC, et al. Serum selenium levels are inversely associated with death risk among hemodialysis patients. Nephrology, Dialysis, Transplantation 2011;26(10):3331‐8. [DOI] [PubMed] [Google Scholar]
Garland 1995 {published data only}
- Garland M, Morris JS, Stampfer MJ, Colditz GA, Spate VL, Baskett CK, et al. Prospective study of toenail selenium levels and cancer among women. Journal of the National Cancer Institute 1995;87(7):497‐505. [DOI] [PubMed] [Google Scholar]
- Hunter DJ, Morris JS, Stampfer MJ, Colditz GA, Speizer FE, Willett WC. A prospective study of selenium status and breast cancer risk. JAMA 1990;264(9):1128‐31. [PubMed] [Google Scholar]
Glattre 1989 {published data only}
- Glattre E, Thomassen Y, Thoresen SO, Haldorsen T, Lund‐Larsen PG, Theodorsen L, et al. Prediagnostic serum selenium in a case‐control study of thyroid cancer. International Journal of Epidemiology 1989;18(1):45‐9. [DOI] [PubMed] [Google Scholar]
Goodman 2001 {published data only}
- Goodman GE, Schaffer S, Bankson DD, Hughes MP, Omenn GS, The Carotene and Retinol Efficacy Trial (CARET) Co‐Investigators. Predictors of serum selenium in cigarette smokers and the lack of association with lung and prostate cancer risk. Cancer Epidemiology, Biomarkers & Prevention 2001;10(10):1069‐76. [PubMed] [Google Scholar]
- Omenn GS, Goodman G, Thornquist M, Grizzle J, Rosenstock L, Barnhart S, et al. The beta‐carotene and retinol efficacy trial (CARET) for chemoprevention of lung cancer in high risk populations: smokers and asbestos‐exposed workers. Cancer Research 1994;54(7 Suppl):2038s‐43s. [PUBMED: 8137335] [PubMed] [Google Scholar]
Goyal 2013 {published data only}
- Goyal A, Terry MB, Siegel AB. Serum antioxidant nutrients, vitamin A, and mortality in US adults. Cancer Epidemiology, Biomarkers & Prevention 2013;22(12):2202‐11. [PUBMED: 23897583] [DOI] [PMC free article] [PubMed] [Google Scholar]
Graff 2017 {published data only}
- Graff RE, Judson G, Ahearn TU, Fiorentino M, Loda M, Giovannucci EL, et al. Circulating antioxidant levels and risk of prostate cancer by TMPRSS2:ERG. The Prostate 2017;77(6):647‐53. [PUBMED: 28102015] [DOI] [PMC free article] [PubMed] [Google Scholar]
Grundmark 2011 {published data only}
- Grundmark B, Zethelius B, Garmo H, Holmberg L. Serum levels of selenium and smoking habits at age 50 influence long term prostate cancer risk; a 34 year ULSAM follow‐up. BMC Cancer 2011;11:431. [EPI_EU_Grundmark 2011] [DOI] [PMC free article] [PubMed] [Google Scholar]
Han 2013 {published data only}
- Han X, Li J, Brasky TM, Xun P, Stevens J, White E, et al. Antioxidant intake and pancreatic cancer risk: the Vitamins and Lifestyle (VITAL) study. Cancer 2013;119(7):1314‐20. [PUBMED: 23280534] [DOI] [PMC free article] [PubMed] [Google Scholar]
Hansen 2013 {published data only}
- Hansen RD, Albieri V, Tjonneland A, Overvad K, Andersen KK, Raaschou‐Nielsen O. Effects of smoking and antioxidant micronutrients on risk of colorectal cancer. Clinical Gastroenterology and Hepatology 2013;11(4):406‐15.e3. [PUBMED: 23142208] [DOI] [PubMed] [Google Scholar]
- Tjonneland A, Olsen A, Boll K, Stripp C, Christensen J, Engholm G, et al. Study design, exposure variables, and socioeconomic determinants of participation in diet, cancer and health: a population‐based prospective cohort study of 57,053 men and women in Denmark. Scandinavian Journal of Public Health 2007;35(4):432‐41. [PUBMED: 17786808] [DOI] [PubMed] [Google Scholar]
Hartman 1998 {published data only}
- Hartman TJ, Albanes D, Pietinen P, Hartman AM, Rautalahti M, Tangrea JA, et al. The association between baseline vitamin E, selenium, and prostate cancer in the alpha‐tocopherol, beta‐carotene cancer prevention study. Cancer Epidemiology, Biomarkers & Prevention 1998;7(4):335‐40. [PubMed] [Google Scholar]
Hashemian 2015 {published data only}
- Hashemian M, Poustchi H, Abnet CC, Boffetta P, Dawsey SM, Brennan PJ, et al. Dietary intake of minerals and risk of esophageal squamous cell carcinoma: results from the Golestan Cohort Study. American Journal of Clinical Nutrition 2015;102(1):102‐8. [PUBMED: 26016858] [DOI] [PMC free article] [PubMed] [Google Scholar]
- Pourshams A, Khademi H, Malekshah AF, Islami F, Nouraei M, Sadjadi AR, et al. Cohort Profile: The Golestan Cohort Study ‐ a prospective study of oesophageal cancer in Northern Iran. International Journal of Epidemiology 2010;39(1):52‐9. [PUBMED: 19332502] [DOI] [PMC free article] [PubMed] [Google Scholar]
Helzlsouer 2000 {published data only}
- Helzlsouer KJ, Huang HY, Alberg AJ, Hoffman S, Burke A, Norkus EP, et al. Association between alpha‐tocopherol, gamma‐tocopherol, selenium, and subsequent prostate cancer. Journal of the National Cancer Institute 2000;92(24):2018‐23. [DOI] [PubMed] [Google Scholar]
Hotaling 2011 {published data only}
- Hotaling JM, Wright JL, Pocobelli G, Bhatti P, Porter MP, White E. Long‐term use of supplemental vitamins and minerals does not reduce the risk of urothelial cell carcinoma of the bladder in the VITamins And Lifestyle study. The Journal of Urology 2011;185(4):1210‐5. [DOI] [PMC free article] [PubMed] [Google Scholar]
Hughes 2015 {published data only}
- Hughes DJ, Fedirko V, Jenab M, Schomburg L, Meplan C, Freisling H, et al. Selenium status is associated with colorectal cancer risk in the European prospective investigation of cancer and nutrition cohort. International Journal of Cancer 2015;136(5):1149‐61. [PUBMED: 25042282] [DOI] [PubMed] [Google Scholar]
Hughes 2016 {published data only}
- Hughes DJ, Duarte‐Salles T, Hybsier S, Trichopoulou A, Stepien M, Aleksandrova K, et al. Prediagnostic selenium status and hepatobiliary cancer risk in the European Prospective Investigation into Cancer and Nutrition cohort. American Journal of Clinical Nutrition 2016;104(2):406‐14. [PUBMED: 27357089] [DOI] [PMC free article] [PubMed] [Google Scholar]
Kabuto 1994 {published data only}
- Kabuto M, Imai H, Yonezawa C, Neriishi K, Akiba S, Kato H, et al. Prediagnostic serum selenium and zinc levels and subsequent risk of lung and stomach cancer in Japan. Cancer Epidemiology, Biomarkers & Prevention 1994;3(6):465‐9. [PubMed] [Google Scholar]
Karagas 1997 {published data only}
- Karagas MR, Greenberg ER, Nierenberg D, Stukel TA, Morris JS, Stevens MM, et al. Risk of squamous cell carcinoma of the skin in relation to plasma selenium, alpha‐tocopherol, beta‐carotene, and retinol: a nested case‐control study. Cancer Epidemiology, Biomarkers & Prevention 1997;6(1):25‐9. [PubMed] [Google Scholar]
Karp 2013 {published data only}
- Karp DD, Lee SJ, Keller SM, Wright GS, Aisner S, Belinsky SA, et al. Randomized, double‐blind, placebo‐controlled, phase III chemoprevention trial of selenium supplementation in patients with resected stage I non‐small‐cell lung cancer: ECOG 5597. Journal of Clinical Oncology 2013;31(33):4179‐87. [PUBMED: 24002495] [DOI] [PMC free article] [PubMed] [Google Scholar]
- Pillai RN, Lee SJ, Karp DD, Aisner SC, Ruckdeschel JC, Ramalingam SS, Khuri FR. Second primary lung cancers: analysis of E5597 selenium chemoprevention study. Journal of Clinical Oncology 2014;32(15):7518. [Google Scholar]
Knekt 1990 {published data only}
- Hakama M, Aaran RK, Alfthan G, Aromaa A, Hakulinen T, Knekt P, et al. Linkage of serum sample bank and cancer registry in epidemiological studies. Progress in Clinical and Biological Research, v. 346. New York: Wiley‐Liss, 1990:169‐78. [PubMed]
- Knekt P, Aromaa A, Alfthan G, Maatela J, Hakama M, Hakulinen T, et al. Re: Prospective study of serum micronutrients and ovarian cancer. Journal of the National Cancer Institute 1996;88(19):1408. [DOI] [PubMed] [Google Scholar]
- Knekt P, Aromaa A, Maatela J, Alfthan G, Aaran RK, Hakama M, et al. Serum selenium and subsequent risk of cancer among Finnish men and women. Journal of the National Cancer Institute 1990;82(10):864‐8. [DOI] [PubMed] [Google Scholar]
- Knekt P, Aromaa A, Maatela J, Alfthan G, Aaran RK, Teppo L, et al. Serum vitamin E, serum selenium and the risk of gastrointestinal cancer. International Journal of Cancer 1988;42(6):846‐50. [DOI] [PubMed] [Google Scholar]
- Knekt P, Jarvinen R, Seppanen R, Rissanen A, Aromaa A, Heinonen OP, et al. Dietary antioxidants and the risk of lung cancer. American Journal of Epidemiology 1991;134(5):471‐9. [DOI] [PubMed] [Google Scholar]
Knekt 1998 {published data only}
- Knekt P, Marniemi J, Teppo L, Heliovaara M, Aromaa A. Is low selenium status a risk factor for lung cancer?. American Journal of Epidemiology 1998;148(10):975‐82. [DOI] [PubMed] [Google Scholar]
Kok 1987a {published data only}
- Kok FJ, Bruijn AM, Hofman A, Valkenburg HA. Selenium status and chronic disease mortality: Dutch epidemiological findings. International Journal of Epidemiology 1987a;16(2):329‐32. [DOI] [PubMed] [Google Scholar]
- Kok FJ, Bruijn AM, Hofman A, Vermeeren R, Valkenburg HA. Is serum selenium a risk factor for cancer in men only?. American Journal of Epidemiology 1987b;125(1):12‐6. [DOI] [PubMed] [Google Scholar]
Kornitzer 2004 {published data only}
- Kornitzer M, Valente F, Bacquer D, Neve J, Backer G. Serum selenium and cancer mortality: a nested case‐control study within an age‐ and sex‐stratified sample of the Belgian adult population. European Journal of Clinical Nutrition 2004;58(1):98‐104. [DOI] [PubMed] [Google Scholar]
Kristal 2014 {published data only}
- Kristal AR, Darke AK, Morris JS, Tangen CM, Goodman PJ, Thompson IM, et al. Baseline selenium status and effects of selenium and vitamin E supplementation on prostate cancer risk. Journal of the National Cancer Institute 2014;106(3):djt456. [PUBMED: 24563519] [DOI] [PMC free article] [PubMed] [Google Scholar]
Kromhout 1987 {published data only}
- Kromhout D. Essential micronutrients in relation to carcinogenesis. American Journal of Clinical Nutrition 1987;45(5 Suppl):1361‐7. [DOI] [PubMed] [Google Scholar]
Li 2000 {published data only}
- Li W, Zhu Y, Yan X, Zhang Q, Li X, Ni Z, et al. The prevention of primary liver cancer by selenium in high risk populations. Zhonghua Yu Fang Yi Xue Za Zhi [Chinese Journal of Preventive Medicine] 2000;34(6):336‐8. [Li 2000] [PubMed] [Google Scholar]
Li 2004a {published data only}
- Li H, Kantoff PW, Giovannucci E, Leitzmann MF, Gaziano JM, Stampfer MJ, et al. Manganese superoxide dismutase polymorphism, prediagnostic antioxidant status, and risk of clinical significant prostate cancer. Cancer Research 2005a;65(6):2498‐504. [DOI] [PubMed] [Google Scholar]
- Li H, Stampfer MJ, Giovannucci EL, Morris JS, Willett WC, Gaziano JM, et al. A prospective study of plasma selenium levels and prostate cancer risk. Journal of the National Cancer Institute 2004;96(9):696‐703. [DOI] [PubMed] [Google Scholar]
- Li H, Stampfer MJ, Giovannucci EL, Morris JS, Willett WC, Gaziano JM, et al. Plasma selenium levels associated with subsequent risk of prostate cancer. American Journal of Urology Review 2005b;3(1):28‐34. [Google Scholar]
Lubinski 2011 {published data only}
- Lubinski J, Jaworska K, Durda K, Jakubowska A, Huzarski T, Byrski T, et al. Selenium and the risk of cancer in BRCA1 carriers. Hereditary Cancer in Clinical Practice 2011;9(Suppl 2):A5. [DOI] [PMC free article] [PubMed] [Google Scholar]
Ma 2017 {published data only}
- Ma X, Yang Y, Li HL, Zheng W, Gao J, Zhang W, et al. Dietary trace element intake and liver cancer risk: results from two population‐based cohorts in China. International Journal of Cancer 2017;140(5):1050‐9. [PUBMED: 27859272] [DOI] [PMC free article] [PubMed] [Google Scholar]
Marshall 2011 {published data only}
- Marshall JR, Tangen CM, Sakr WA, Wood DP Jr, Berry DL, Klein EA, et al. Phase III trial of selenium to prevent prostate cancer in men with high‐grade prostatic intraepithelial neoplasia: SWOG S9917. Cancer Prevention Research 2011;4:1761‐9. [DOI] [PMC free article] [PubMed] [Google Scholar]
McNaughton 2005 {published data only}
- Heinen MM, Hughes MC, Ibiebele TI, Marks GC, Green AC, Pols JC. Intake of antioxidant nutrients and the risk of skin cancer. European Journal of Cancer 2007;43(18):2707‐16. [DOI] [PubMed] [Google Scholar]
- McNaughton SA, Marks GC, Gaffney P, Williams G, Green AC. Antioxidants and basal cell carcinoma of the skin: a nested case‐control study. Cancer Causes and Control 2005;16(5):609‐18. [DOI] [PubMed] [Google Scholar]
- Pols JC, Heinen MM, Hughes MC, Ibiebele TI, Marks GC, Green AC. Serum antioxidants and skin cancer risk: an 8‐year community‐based follow‐up study. Cancer Epidemiology, Biomarkers & Prevention 2009;18(4):1167‐73. [DOI] [PubMed] [Google Scholar]
Menkes 1986 {published data only}
- Batieha AM, Armenian HK, Norkus EP, Morris JS, Spate VE, Comstock GW. Serum micronutrients and the subsequent risk of cervical cancer in a population‐based nested case‐control study. Cancer Epidemiology, Biomarkers & Prevention 1993;2(4):335‐9. [PubMed] [Google Scholar]
- Breslow RA, Alberg AJ, Helzlsouer KJ, Bush TL, Norkus EP, Morris JS, et al. Serological precursors of cancer: malignant melanoma, basal and squamous cell skin cancer, and prediagnostic levels of retinol, beta‐carotene, lycopene, alpha‐tocopherol, and selenium. Cancer Epidemiology, Biomarkers & Prevention 1995;4(8):837‐42. [PubMed] [Google Scholar]
- Burney PG, Comstock GW, Morris JS. Serologic precursors of cancer: serum micronutrients and the subsequent risk of pancreatic cancer. American Journal of Clinical Nutrition 1989;49(5):895‐900. [DOI] [PubMed] [Google Scholar]
- Helzlsouer KJ, Alberg AJ, Norkus EP, Morris JS, Hoffman SC, Comstock GW. Prospective study of serum micronutrients and ovarian cancer. Journal of the National Cancer Institute 1996;88(1):32‐7. [DOI] [PubMed] [Google Scholar]
- Helzlsouer KJ, Comstock GW, Morris JS. Selenium, lycopene, alpha‐tocopherol, beta‐carotene, retinol, and subsequent bladder cancer. Cancer Research 1989;49(21):6144‐8. [PubMed] [Google Scholar]
- Ko W. The Associations of Serologic Precursors and the Anatomic‐Site Specific Incidence of Colon Cancer. Baltimore, MD: John Hopkins University, 1994. [Google Scholar]
- Menkes MJ. Vitamins A, E, selenium and risk of lung cancer. Dissertation Abstract International (Sci) 1986a; Vol. 46, issue 11:3807.
- Menkes MS, Comstock GW, Vuilleumier JP, Helsing KJ, Rider AA, Brookmeyer R. Serum beta‐carotene, vitamins A and E, selenium, and the risk of lung cancer. New England Journal of Medicine 1986b;315(20):1250‐4. [DOI] [PubMed] [Google Scholar]
- Schober SE. Vitamin A, vitamin E, selenium, and colon cancer risk. Dissertation Abstract International (Sci) 1986; Vol. 46, issue 11:3808.
- Schober SE, Comstock GW, Helsing KJ, Salkeld RM, Morris JS, Rider AA, Brookmeyer R. Serologic precursors of cancer. I. Prediagnostic serum nutrients and colon cancer risk. American Journal of Epidemiology 1987;126(6):1033‐41. [DOI] [PubMed] [Google Scholar]
- Zheng W, Blot WJ, Diamond EL, Norkus EP, Spate V, Morris JS, et al. Serum micronutrients and the subsequent risk of oral and pharyngeal cancer. Cancer Research 1993;53(4):795‐8. [PubMed] [Google Scholar]
Michaud 2002 {published data only}
- Michaud DS, Hartman TJ, Taylor PR, Pietinen P, Alfthan G, Virtamo J, et al. No association between toenail selenium levels and bladder cancer risk. Cancer Epidemiology, Biomarkers & Prevention 2002;11(11):1505‐6. [PubMed] [Google Scholar]
Michaud 2005 {published data only}
- Michaud DS, Vivo I, Morris JS, Giovannucci E. Toenail selenium concentrations and bladder cancer risk in women and men. British Journal of Cancer 2005;93(7):804‐6. [DOI] [PMC free article] [PubMed] [Google Scholar]
Muka 2017 {published data only}
- Muka T, Kraja B, Ruiter R, Lahousse L, Keyser CE, Hofman A, et al. Dietary mineral intake and lung cancer risk: the Rotterdam Study. European Journal of Nutrition 2017;56(4):1637‐46. [PUBMED: 27073037] [DOI] [PMC free article] [PubMed] [Google Scholar]
Nomura 1987 {published data only}
- Nomura A, Heilbrun LK, Morris JS, Stemmermann GN. Serum selenium and the risk of cancer, by specific sites: case‐control analysis of prospective data. Journal of the National Cancer Institute 1987;79(1):103‐8. [PubMed] [Google Scholar]
Nomura 2000 {published data only}
- Nomura AMY, Lee J, Stemmermann GN, Combs GF Jr. Serum selenium and subsequent risk of prostate cancer. Cancer Epidemiology, Biomarkers & Prevention 2000;9(9):883‐7. [PubMed] [Google Scholar]
NPCT 2002 {published data only}
- Clark L, Krongrad A, Dalkin B, Witherington R, Herlong H, Carpenter D. Decreased incidence of prostate cancer with selenium supplementation: 1983‐96 results of a double‐blind cancer prevention trial. European Journal of Cancer Prevention 1997; Vol. 6, issue 5:497‐8. [DOI] [PubMed]
- Clark LC, Combs GF Jr, Turnbull BW, Slate EH, Chalker DK, Chow J, et al. Effects of selenium supplementation for cancer prevention in patients with carcinoma of the skin: a randomized controlled trial. JAMA 1996;276(24):1957‐63. [PubMed] [Google Scholar]
- Clark LC, Dalkin B, Krongrad A, Combs GF, Turnbull BW, Slate EH, et al. Decreased incidence of prostate cancer with selenium supplementation: results of a double‐blind cancer prevention trial. British Journal of Urology 1998;81(5):730‐4. [DOI] [PubMed] [Google Scholar]
- Clark LC, UARIZ‐CA49764 NCI‐P89‐0003. Double‐blind, randomized trial of selenium‐enriched brewer's yeast vs brewer's yeast placebo for the prevention of skin cancer in patients with a history of squamous or basal cell skin cancer (Summary Last Modified 05/91). http://www.cancer.gov. National Cancer Institute, (accessed 1 March 2004).
- Combs GF, Clark LC, Turnbull BW. Reduction of cancer mortality and incidence by selenium supplementation. Medizinische Klinik 1997;92(Suppl 3):42‐5. [DOI] [PubMed] [Google Scholar]
- Combs GFJr, Clark LC, Turnbull BW. Reduction of cancer risk with an oral supplement of selenium. Biomedical and Environmental Sciences 1997;10(2‐3):227‐34. [PubMed] [Google Scholar]
- Duffield‐Lillico AJ, Dalkin BL, Reid ME, Turnbull BW, Slate EH, Jacobs ET, et al. Nutritional Prevention of Cancer Study Group. Selenium supplementation, baseline plasma selenium status and incidence of prostate cancer: an analysis of the complete treatment period of the Nutritional Prevention of Cancer Trial. BJU International 2003b;91(7):608‐12. [DOI] [PubMed] [Google Scholar]
- Duffield‐Lillico AJ, Reid ME, Turnbull BW, Combs GF Jr, Slate EH, Fischbach LA, et al. Baseline characteristics and the effect of selenium supplementation on cancer incidence in a randomized clinical trial: a summary report of the Nutritional Prevention of Cancer Trial. Cancer Epidemiology, Biomarkers & Prevention 2002;11(7):630‐9. [PubMed] [Google Scholar]
- Duffield‐Lillico AJ, Slate EH, Reid ME, Turnbull BW, Wilkins PA, Combs GF Jr, et al. Nutritional Prevention of Cancer Study Group. Selenium supplementation and secondary prevention of nonmelanoma skin cancer in a randomized trial. Journal of the National Cancer Institute 2003a;95(19):1477‐81. [DOI] [PubMed] [Google Scholar]
- Marshall JR, 5R01CA049764‐15. Nutritional Prevention of Cancer. http://crisp.cit.nih.gov (accessed 1 March 2004).
- Reid ME, Duffield‐Lillico AJ, Garland L, Turnbull BW, Clark LC, Marshall JR. Selenium supplementation and lung cancer incidence: an update of the Nutritional Prevention of Cancer Trial. Cancer Epidemiology, Biomarkers & Prevention 2002;11(11):1285‐91. [PubMed] [Google Scholar]
- Reid ME, Duffield‐Lillico AJ, Slate E, Natarajan N, Turnbull B, Jacobs E, et al. The nutritional prevention of cancer: 400 mcg per day selenium treatment. Nutrition and Cancer 2008;60(2):155‐63. [DOI] [PubMed] [Google Scholar]
- Stranges S, Marshall JR, Natarajan R, Donahue RP, Trevisan M, Combs GF, et al. Effects of long‐term selenium supplementation on the incidence of type 2 diabetes: a randomized trial. Annals of Internal Medicine 2007;147(4):217‐23. [DOI] [PubMed] [Google Scholar]
O'Grady 2014 {published data only}
- O'Grady TJ, Kitahara CM, DiRienzo AG, Gates MA. The association between selenium and other micronutrients and thyroid cancer incidence in the NIH‐AARP Diet and Health Study. PloS One 2014;9(10):e110886. [PUBMED: 25329812] [DOI] [PMC free article] [PubMed] [Google Scholar]
Outzen 2016 {published data only}
- Outzen M, Tjonneland A, Larsen EH, Friis S, Larsen SB, Christensen J, et al. Selenium status and risk of prostate cancer in a Danish population. British Journal of Nutrition 2016;115(9):1669‐77. [PUBMED: 26971676] [DOI] [PubMed] [Google Scholar]
- Tjonneland A, Olsen A, Boll K, Stripp C, Christensen J, Engholm G, et al. Study design, exposure variables, and socioeconomic determinants of participation in Diet, Cancer and Health: a population‐based prospective cohort study of 57,053 men and women in Denmark. Scandinavian Journal of Public Health 2007;35(4):432‐41. [PUBMED: 17786808] [DOI] [PubMed] [Google Scholar]
Overvad 1991 {published data only}
- Overvad K, Wang DY, Olsen J, Allen DS, Thorling EB, Bulbrook RD, Hayward JL. Selenium in human mammary carcinogenesis: a case‐cohort study. European Journal of Cancer 1991;27(7):900‐2. [DOI] [PubMed] [Google Scholar]
Pantavos 2015 {published data only}
- Pantavos A, Ruiter R, Feskens EF, Keyser CE, Hofman A, Stricker BH, et al. Total dietary antioxidant capacity, individual antioxidant intake and breast cancer risk: the Rotterdam Study. International Journal of Cancer 2015;136(9):2178‐86. [PUBMED: 25284450] [DOI] [PubMed] [Google Scholar]
Park 2015 {published data only}
- Park SY, Haiman CA, Cheng I, Park SL, Wilkens LR, Kolonel LN, et al. Racial/ethnic differences in lifestyle‐related factors and prostate cancer risk: the Multiethnic Cohort Study. Cancer Causes & Control 2015;26(10):1507‐15. [PUBMED: 26243447] [DOI] [PMC free article] [PubMed] [Google Scholar]
Peleg 1985 {published data only}
- Peleg I, Morris S, Hames CG. Is serum selenium a risk factor for cancer?. Medical Oncology and Tumor Pharmacotherapy 1985;2(3):157‐63. [DOI] [PubMed] [Google Scholar]
Peters 2007 {published data only}
- Peters U, Foster CB, Chatterjee N, Schatzkin A, Reding D, Andriole GL, et al. Serum selenium and risk of prostate cancer ‐ a nested case‐control study. American Journal of Clinical Nutrition 2007;85(1):209‐17. [DOI] [PMC free article] [PubMed] [Google Scholar]
Peters 2008 {published data only}
- Asgari MM, Maruti SS, Kushi LH, White E. Antioxidant supplementation and risk of incident melanomas: results of a large prospective cohort study. Archives of Dermatology 2009;145(8):879‐82. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Peters U, Littman AJ, Kristal AR, Patterson RE, Potter JD, White E. Vitamin E and selenium supplementation and risk of prostate cancer in the vitamins and lifestyle (VITAL) study cohort. Cancer Causes Control 2008;19(1):75‐87. [DOI] [PubMed] [Google Scholar]
Ratnasinghe 2000 {published data only}
- Ratnasinghe D, Tangrea JA, Forman MR, Hartman T, Gunter EW, Qiao YL, et al. Serum tocopherols, selenium and lung cancer risk among tin miners in China. Cancer Causes and Control 2000;11(2):129‐35. [DOI] [PubMed] [Google Scholar]
Reid 2008 {published data only}
- Reid ME, Duffield‐Lillico AJ, Slate E, Natarajan N, Turnbull B, Jacobs E, et al. The nutritional prevention of cancer: 400 mcg per day selenium treatment. Nutrition and Cancer 2008;60(2):155‐63. [DOI] [PubMed] [Google Scholar]
Ringstad 1988 {published data only}
- Ringstad J, Jacobsen BK, Tretli S, Thomassen Y. Serum selenium concentration associated with risk of cancer. Journal of Clinical Pathology 1988;41(4):454‐7. [DOI] [PMC free article] [PubMed] [Google Scholar]
Sakoda 2005 {published data only}
- Sakoda LC, Graubard BI, Evans AA, London WT, Lin WY, Shen FM, et al. Toenail selenium and risk of hepatocellular carcinoma mortality in Haimen City, China. International Journal of Cancer 2005;115(4):618‐24. [DOI] [PubMed] [Google Scholar]
Salonen 1984 {published data only}
- Salonen JT, Alfthan G, Huttunen JK, Puska P. Association between serum selenium and the risk of cancer. American Journal of Epidemiology 1984;120(3):342‐9. [DOI] [PubMed] [Google Scholar]
Salonen 1985 {published data only}
- Salonen JT, Salonen R, Lappetelainen R, Maenpaa PH, Alfthan G, Puska P. Risk of cancer in relation to serum concentrations of selenium and vitamins A and E: matched case‐control analysis of prospective data. British Medical Journal (Clinical Research Edition) 1985;290(6466):417‐20. [DOI] [PMC free article] [PubMed] [Google Scholar]
SELECT 2009 {published data only}
- Cook ED. Selenium and Vitamin E Cancer Prevention Trial ‐ this one's for us. Journal of the National Medical Association 2002;94(9):856‐8. [PMC free article] [PubMed] [Google Scholar]
- Cook ED, Moody‐Thomas S, Anderson KB, Campbell R, Hamilton SJ, Harrington JM, et al. Minority recruitment to the Selenium and Vitamin E Cancer Prevention Trial (SELECT). Clinical Trials 2005;2(5):436‐42. [DOI] [PubMed] [Google Scholar]
- DeFrancesco L. Prostate cancer prevention trial launched. Nature Medicine 2001;7(10):1076. [DOI] [PubMed] [Google Scholar]
- Dunn BK, Ryan A, Ford LG. Selenium and Vitamin E Cancer Prevention Trial: a nutrient approach to prostate cancer prevention. Recent Results in Cancer Research 2009;181:183‐93. [DOI] [PubMed] [Google Scholar]
- FDA. Largest‐ever prostate cancer prevention trial. FDA Consumer 2001;35(5):8. [PubMed] [Google Scholar]
- Ford LG, Minasian LM, McCaskill‐Stevens W, Pisano ED, Sullivan D, Smith RA. Prevention and early detection clinical trials: opportunities for primary care providers and their patients. CA: A Cancer Journal for Clinicians 2003;53(2):82‐101. [DOI] [PubMed] [Google Scholar]
- Hoque A, Albanes D, Lippman SM, Spitz MR, Taylor PR, Klein EA, et al. Molecular epidemiologic studies within the Selenium and Vitamin E Cancer Prevention Trial (SELECT). Cancer Causes and Control 2001;12(7):627‐33. [DOI] [PubMed] [Google Scholar]
- Kardinal C, Brooks J. Ochsner Cancer Institute studies vitamin E and selenium as prostate cancer prevention agents. Ochsner Journal 2003;5(2):51. [PMC free article] [PubMed] [Google Scholar]
- Klein EA. Clinical models for testing chemopreventative agents in prostate cancer and overview of SELECT: the Selenium and Vitamin E Cancer Prevention Trial. Recent Results in Cancer Research 2003;163:212‐25. [DOI] [PubMed] [Google Scholar]
- Klein EA. Selenium and vitamin E cancer prevention trial. Annals of the New York Academy of Sciences 2004;1031:234‐41. [DOI] [PubMed] [Google Scholar]
- Klein EA, Atkins MB, Walther P, Klotz L, SWOG‐S000. Phase III randomized study of selenium and vitamin E for the prevention of prostate cancer (SELECT trial). http://clinicaltrials.gov/ (accessed 12 January 2004).
- Klein EA, Lippman SM, Thompson IM, Goodman PJ, Albanes D, Taylor PR, et al. The selenium and vitamin E cancer prevention trial. World Journal of Urology 2003;21(1):21‐7. [DOI] [PubMed] [Google Scholar]
- Klein EA, Thompson IM Jr, Tangen CM, Crowley JJ, Lucia MS, Goodman PJ, et al. Vitamin E and the risk of prostate cancer: the Selenium and Vitamin E Cancer Prevention Trial (SELECT). JAMA 2011;306(14):1549‐56. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Klein EA, Thompson IM, Lippman SM, Goodman PJ, Albanes D, Taylor PR, et al. SELECT: the Selenium and Vitamin E Cancer Prevention Trial. Urologic Oncology 2003;21(1):59‐65. [DOI] [PubMed] [Google Scholar]
- Klein EA, Thompson IM, Lippman SM, Goodman PJ, Albanes D, Taylor PR, et al. SELECT: the Selenium and Vitamin E Cancer Prevention Trial: rationale and design. Prostate Cancer and Prostatic Diseases 2000;3(3):145‐51. [DOI] [PubMed] [Google Scholar]
- Klein EA, Thompson IM, Lippman SM, Goodman PJ, Albanes D, Taylor PR, et al. SELECT: the next prostate cancer prevention trial. Selenium and Vitamin E Cancer Prevention Trial. Journal of Urology 2001;166(4):1311‐5. [DOI] [PubMed] [Google Scholar]
- Lippman SM, Goodman PJ, Klein EA, Parnes HL, Thompson IM Jr, Kristal AR, et al. Designing the Selenium and Vitamin E Cancer Prevention Trial (SELECT). Journal of the National Cancer Institute 2005;97(2):94‐102. [DOI] [PubMed] [Google Scholar]
- Lippman SM, Klein EA, Goodman PJ, Lucia MS, Thompson IM, Ford LG, et al. Effect of selenium and vitamin E on risk of prostate cancer and other cancers: the Selenium and Vitamin E Cancer Prevention Trial (SELECT). JAMA 2009;301(1):39‐51. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lotan Y, Goodman PJ, Youssef RF, Svatek RS, Shariat SF, Tangen CM, et al. Evaluation of vitamin E and selenium supplementation for the prevention of bladder cancer in SWOG coordinated SELECT. Journal of Urology 2012;187(6):205‐10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Miller M. Enrollment begins for largest‐ever prostate cancer prevention trial. Journal of the National Cancer Institute 2001;93(15):1132. [DOI] [PubMed] [Google Scholar]
- Pak RW, Lanteri VJ, Scheuch JR, Sawczuk IS. Review of vitamin E and selenium in the prevention of prostate cancer: implications of the selenium and vitamin E chemoprevention trial. Integrative Cancer Therapies 2002;1(4):338‐44. [DOI] [PubMed] [Google Scholar]
- South West Oncology Group. Selenium and Vitamin E Cancer Prevention Trial (SELECT). http://www.controlled‐trials.com/mrct/trial/SELENIUM/1059/42795.html (accessed 16 April 2004).
- South West Oncology Group. Selenium and Vitamin E in preventing prostate cancer. http://www.controlled‐trials.com/mrct/trial/SELENIUM/1059/32859.html. Southwest Oncology Group, (accessed 16 April 2004).
- Tangen CM, Goodman PJ, Crowley JJ, Thompson IM. Statistical design issues and other practical considerations for conducting phase III prostate cancer prevention trials. Journal of Urology 2004;171(2 Pt 2):S64‐7. [DOI] [PubMed] [Google Scholar]
- Tom J. SELECT (opportunity for prostate cancer prevention). Hawaii Medical Journal 2002;61(6):126‐9. [PubMed] [Google Scholar]
Steevens 2010 {published data only}
- Steevens J, Brandt PA, Goldbohm RA, Schouten LJ. Selenium status and the risk of esophageal and gastric cancer subtypes: the Netherlands cohort study. Gastroenterology. 138. W.B. Saunders (Independence Square West, Philadelphia PA 19106‐3399, United States), 2010; Vol. 138, issue 5:1704‐13. [DOI] [PubMed]
Steinbrecher 2010 {published data only}
- Steinbrecher A, Meplan C, Hesketh J, Schomburg L, Endermann T, Jansen E, et al. Effects of selenium status and polymorphisms in selenoprotein genes on prostate cancer risk in a prospective study of European men. Cancer Epidemiology, Biomarkers & Prevention. 19. American Association for Cancer Research Inc. (615 Chestnut Street, 17th Floor, Philadelphia PA 19106‐3483, United States), 2010; Vol. 19, issue 11:2958‐68. [DOI] [PubMed]
Suadicani 2012 {published data only}
- Suadicani P, Hein HO, Gyntelberg F. Serum selenium level and risk of lung cancer mortality: a 16‐year follow‐up of the Copenhagen Male Study. European Respiratory Journal 2012;39(6):1443‐8. [DOI] [PubMed] [Google Scholar]
Sun 2016 {published data only}
- Sun JW, Shu XO, Li HL, Zhang W, Gao J, Zhao LG, et al. Dietary selenium intake and mortality in two population‐based cohort studies of 133 957 Chinese men and women. Public Health Nutrition 2016;19(16):2991‐8. [PUBMED: 27197889] [DOI] [PMC free article] [PubMed] [Google Scholar]
Thomson 2008 {published data only}
- Thomson CA, Neuhouser ML, Shikany JM, Caan BJ, Monk BJ, Mossavar‐Rahmani Y, et al. The role of antioxidants and vitamin A in ovarian cancer: results from the Women's Health Initiative. Nutrition and Cancer 2008;60(6):710‐9. [DOI] [PubMed] [Google Scholar]
van den Brandt 1993 {published data only}
- Zeegers MPA, Goldbohm RA, Bode P, Brandt PA. Prediagnostic toenail selenium and risk of bladder cancer. Cancer Epidemiology, Biomarkers & Prevention 2002;11(11):1292‐7. [PubMed] [Google Scholar]
- Brandt PA, Goldbohm RA, van't Veer P, Bode P, Dorant E, Hermus RJ, et al. A prospective cohort study on selenium status and the risk of lung cancer. Cancer Research 1993a;53(20):4860‐5. [PubMed] [Google Scholar]
- Brandt PA, Goldbohm RA, van't Veer P, Bode P, Dorant E, Hermus RJ, et al. A prospective cohort study on toenail selenium levels and risk of gastrointestinal cancer. Journal of the National Cancer Institute 1993b;85(3):224‐9. [DOI] [PubMed] [Google Scholar]
- Brandt PA, Goldbohm RA, van't Veer P, Bode P, Dorant E, Hermus RJ, et al. Toenail selenium levels and the risk of breast cancer. American Journal of Epidemiology 1994;140(1):20‐6. [DOI] [PubMed] [Google Scholar]
- Brandt PA, Zeegers MPA, Bode P, Goldbohm RA. Toenail selenium levels and the subsequent risk of prostate cancer: a prospective cohort study. Cancer Epidemiology, Biomarkers & Prevention 2003;12(9):866‐71. [PubMed] [Google Scholar]
van Noord 1987 {published data only}
- Noord PAH, Collette HJA, Maas MJ, Waard F. Selenium levels in nails of premenopausal breast cancer patients assessed prediagnostically in a cohort‐nested case‐referent study among women screened in the DOM project. International Journal of Epidemiology 1987;16(2):318‐22. [DOI] [PubMed] [Google Scholar]
Virtamo 1987 {published data only}
- Virtamo J, Valkeila E, Alfthan G, Punsar S, Huttunen JK, Karvonen MJ. Serum selenium and risk of cancer. A prospective follow‐up of nine years. Cancer 1987;60(2):145‐8. [DOI] [PubMed] [Google Scholar]
Walter 2011 {published data only}
- Walter RB, Brasky TM, Milano F, White E. Vitamin, mineral, and specialty supplements and risk of hematologic malignancies in the prospective VITamins And Lifestyle (VITAL) study. Cancer Epidemiology, Biomarkers & Prevention 2011;20(10):2298‐308. [DOI] [PMC free article] [PubMed] [Google Scholar]
Wei 2004 {published data only}
- Mark SD, Qiao YL, Dawsey SM, Wu YP, Katki H, Gunter EW, et al. Prospective study of serum selenium levels and incident esophageal and gastric cancers. Journal of the National Cancer Institute 2000;92(21):1753‐63. [DOI] [PubMed] [Google Scholar]
- Wei WQ, Abnet CC, Qiao YL, Dawsey SM, Dong ZW, Sun XD, et al. Prospective study of serum selenium concentrations and esophageal and gastric cardia cancer, heart disease, stroke, and total death. American Journal of Clinical Nutrition 2004;79(1):80‐5. [DOI] [PubMed] [Google Scholar]
Willett 1983 {published data only}
- Willett WC, Polk BF, Morris JS, Stampfer MJ, Pressel S, Rosner B, et al. Prediagnostic serum selenium and risk of cancer. Lancet 1983;2(8342):130‐4. [DOI] [PubMed] [Google Scholar]
Yoshizawa 1998 {published data only}
- Yoshizawa K, Willett WC, Morris SJ, Stampfer MJ, Spiegelman D, Rimm EB, et al. Study of prediagnostic selenium level in toenails and the risk of advanced prostate cancer. Journal of the National Cancer Institute 1998;90(16):1219‐24. [DOI] [PubMed] [Google Scholar]
Yu 1991 {published data only}
- Li WG. [Preliminary observations on effect of selenium yeast on high risk populations with primary liver cancer]. Zhonghua Yu Fang Yi Xue Za Zhi [Chinese Journal of Preventive Medicine] 1992;26(5):268‐71. [PubMed] [Google Scholar]
- Yu S, Li W, Zhu Y. Chemoprevention of liver cancer. CCPC‐93: Second International Cancer Chemo Prevention Conference; 1993 April 28‐30; Berlin (Meeting Abstracts)..
- Yu SY, Zhu YJ, Li WG, Huang QS, Huang CZ, Zhang QN, et al. A preliminary report on the intervention trials of primary liver cancer in high‐risk populations with nutritional supplementation of selenium in China. Biological Trace Element Research 1991;29(3):289‐94. [DOI] [PubMed] [Google Scholar]
Yu 1997 {published data only}
- Li WG. [Preliminary observations on effect of selenium yeast on high risk populations with primary liver cancer]. Zhonghua Yu Fang Yi Xue Za Zhi [Chinese Journal of Preventive Medicine] 1992;26(5):268‐71. [PubMed] [Google Scholar]
- Yu SY, Zhu YJ, Li WG. Protective role of selenium against hepatitis B virus and primary liver cancer in Qidong. Biological Trace Element Research 1997;56(1):117‐24. [MEDLINE: ] [DOI] [PubMed] [Google Scholar]
- Yu SY, Zhu YJ, Li WG, Huang QS, Huang CZ, Zhang QN, et al. A preliminary report on the intervention trials of primary liver cancer in high‐risk populations with nutritional supplementation of selenium in China. Biological Trace Element Research 1991;29(3):289‐94. [DOI] [PubMed] [Google Scholar]
Yu 1999 {published data only}
- Yu MW, Horng IS, Hsu KH, Chiang YC, Liaw YF, Chen CJ. Plasma selenium levels and risk of hepatocellular carcinoma among men with chronic hepatitis virus infection. American Journal of Epidemiology 1999;150(4):367‐74. [DOI] [PubMed] [Google Scholar]
References to studies excluded from this review
Albanes 2014 {published data only}
- Albanes D, Till C, Klein EA, Goodman PJ, Mondul AM, Weinstein SJ, et al. Plasma tocopherols and risk of prostate cancer in the Selenium and Vitamin E Cancer Prevention Trial (SELECT). Cancer Prevention Research (Philadelphia, Pa.) 2014;7(9):886‐95. [PUBMED: 24961880] [DOI] [PMC free article] [PubMed] [Google Scholar]
Bates 2011 {published data only}
- Bates CJ, Hamer M, Mishra GD. Redox‐modulatory vitamins and minerals that prospectively predict mortality in older British people: the National Diet and Nutrition Survey of people aged 65 years and over. British Journal of Nutrition 2011;105(1):123‐32. [EPI_EU_Bates 2011] [DOI] [PMC free article] [PubMed] [Google Scholar]
Bostick 1993 {published data only}
- Bostick RM, Potter JD, McKenzie DR, Sellers TA, Kushi LH, Steinmetz KA, et al. Reduced risk of colon cancer with high intake of vitamin E: the Iowa Women's Health Study. Cancer Research 1993;53(18):4230‐7. [PubMed] [Google Scholar]
Brock 1991 {published data only}
- Brock KE, Gridley G, Morris JS, Willett WC. Serum selenium level in relation to in situ cervical cancer in Australia. Journal of the National Cancer Institute 1991;83(4):292‐3. [DOI] [PubMed] [Google Scholar]
Chen 1988 {published data only}
- Chen Q. [Protective effects of selenium, zinc and copper on lung cancer]. Zhonghua Yu Fang Yi Xue Za Zhi [Chinese Journal of Preventive Medicine] 1988;22(4):221‐4. [PubMed] [Google Scholar]
Chen 2003 {published data only}
- Chen K, Qiu JL, Sui LM, Yu WP, Wang JY, Zhang LJ. Nutrient intake and gastric cancer in residents of Zhoushan Islands, China. Digestive and Liver Disease 2003;35(12):912‐3. [DOI] [PubMed] [Google Scholar]
Connelly‐Frost 2009 {published data only}
- Connelly‐Frost A, Poole C, Satia JA, Kupper LL, Millikan RC, Sandler RS. Selenium, folate, and colon cancer. Nutrition and Cancer 2009;61(2):165‐78. [DOI] [PMC free article] [PubMed] [Google Scholar]
Costello 2001 {published data only}
- Costello AJ. A randomized, controlled chemoprevention trial of selenium in familial prostate cancer: rationale, recruitment, and design issues. Urology 2001;57(4 Suppl 1):182‐4. [DOI] [PubMed] [Google Scholar]
Criqui 1991 {published data only}
- Criqui MH, Bangdiwala S, Goodman DS, Blaner WS, Morris JS, Kritchevsky S, et al. Selenium, retinol, retinol‐binding protein, and uric acid. Associations with cancer mortality in a population‐based prospective case‐control study. Annals of Epidemiology 1991;1(5):385‐93. [DOI] [PubMed] [Google Scholar]
Cui 2007 {published data only}
- Cui Y, Vogt S, Olson N, Glass AG, Rohan TE. Levels of zinc, selenium, calcium, and iron in benign breast tissue and risk of subsequent breast cancer. Cancer Epidemiology, Biomarkers & Prevention 2007;16(8):1682‐5. [DOI] [PubMed] [Google Scholar]
Davies 2002 {published data only}
- Davies TW, Treasure FP, Welch AA, Day NE. Diet and basal cell skin cancer: results from the EPIC‐Norfolk cohort. British Journal of Dermatology 2002;146(6):1017‐22. [DOI] [PubMed] [Google Scholar]
Epplein 2014 {published data only}
- Epplein M, Burk RF, Cai Q, Hargreaves MK, Blot WJ. A prospective study of plasma Selenoprotein P and lung cancer risk among low‐income adults. Cancer Epidemiology, Biomarkers & Prevention 2014;23(7):1238‐44. [DOI] [PMC free article] [PubMed] [Google Scholar]
Fleshner 2003 {published data only}
- Fleshner N, CAN‐CNIC‐PRP1. Phase II randomized study of vitamin E, selenium, and soy protein isolate in patients with high‐grade prostatic intraepithelial neoplasia. http://www.cancer.gov (accessed 1 April 2004).
Geybels 2013 {published data only}
- Geybels MS, Verhage BA, Schooten FJ, Goldbohm RA, Brandt PA. Advanced prostate cancer risk in relation to toenail selenium levels. Journal of the National Cancer Institute 2013;105(18):1394‐401. [PUBMED: 23878355] [DOI] [PubMed] [Google Scholar]
Geybels 2014 {published data only}
- Geybels MS, Brandt PA, Schouten LJ, Schooten FJ, Breda SG, Rayman MP, et al. Selenoprotein gene variants, toenail selenium levels, and risk for advanced prostate cancer. Journal of the National Cancer Institute 2014;106(3):dju003. [PUBMED: 24563517] [DOI] [PubMed] [Google Scholar]
Hagmar 1992 {published data only}
- Hagmar L, Linden K, Nilsson A, Norrving B, Akesson B, Schutz A, et al. Cancer incidence and mortality among Swedish Baltic Sea fishermen. Scandinavian Journal of Work, Environment and Health 1992;18(4):217‐24. [DOI] [PubMed] [Google Scholar]
Harris 2012 {published data only}
- Harris Holly R, Brgkvist L, Wolk A. Selenium intake and breast cancer mortality in a cohort of Swedish women. Breast Cancer Research and Treatment 2012;134(3):1269‐77. [DOI] [PubMed] [Google Scholar]
Hartman 2002 {published data only}
- Hartman TJ, Taylor PR, Alfthan G, Fagerstrom R, Virtamo J, Mark SD, et al. Toenail selenium concentration and lung cancer in male smokers (Finland). Cancer Causes & Control 2002;13(10):923‐8. [DOI] [PubMed] [Google Scholar]
Huzarski 2006 {published data only}
- Huzarski T, Byrski T, Gronwald J, Kowalska E, Zajaczek S, Gorski B, et al. A lowering of breast and ovarian cancer risk in women with a BRCA1 mutation by selenium supplementation of diet. Hereditary Cancer in Clinical Practice 2006;4(1):58. [DOI] [PMC free article] [PubMed] [Google Scholar]
Joniau 2007 {published data only}
- Joniau S, Goeman L, Roskams T, Lerut E, Oyen R, Van PH. Effect of nutritional supplement challenge in patients with isolated high‐grade prostatic intraepithelial neoplasia. Urology 2007;69(6):1102‐6. [DOI] [PubMed] [Google Scholar]
Karunasinghe 2012 {published data only}
- Karunasinghe N, Han DY, Goudie M, Zhu S, Bishop K, Wang A, et al. Prostate disease risk factors among a New Zealand cohort. Journal of Nutrigenetics and Nutrigenomics 2012;5(6):339‐51. [PUBMED: 23363810] [DOI] [PubMed] [Google Scholar]
Kellen 2008 {published data only}
- Kellen E, Zeegers MP, Bruckers L, Buntinx F. The investigation of a geographical cluster of bladder cancer. Acta Clinica Belgica 2008;63(5):313‐20. [DOI] [PubMed] [Google Scholar]
Kilander 2001 {published data only}
- Kilander L, Berglund L, Boberg M, Vessby B, Lithell H. Education, lifestyle factors and mortality from cardiovascular disease and cancer. A 25‐year follow‐up of Swedish 50‐year‐old men. International Journal of Epidemiology 2001;30(5):1119‐26. [DOI] [PubMed] [Google Scholar]
Knekt 1988a {published data only}
- Knekt P, Reunanen A, Aromaa A, Heliovaara M, Hakulinen T, Hakama M. Serum cholesterol and risk of cancer in a cohort of 39,000 men and women. Journal of Clinical Epidemiology 1988;41(6):519‐30. [DOI] [PubMed] [Google Scholar]
Knekt 1988b {published data only}
- Knekt P, Aromaa A, Maatela J, Aaran RK, Nikkari T, Hakama M, et al. Serum vitamin E and risk of cancer among Finnish men during a 10‐year follow‐up. American Journal of Epidemiology 1988;127(1):28‐41. [DOI] [PubMed] [Google Scholar]
Knekt 1991 {published data only}
- Knekt P, Aromaa A, Maatela J, Alfthan G, Aaran RK, Nikkari T, et al. Serum micronutrients and risk of cancers of low incidence in Finland. American Journal of Epidemiology 1991;134(4):356‐61. [DOI] [PubMed] [Google Scholar]
Kok 1987b {published data only}
- Kok FJ, Duijn CM, Hofman A, Vermeeren R, Bruijn AM, Valkenburg HA. Micronutrients and the risk of lung cancer (letter). New England Journal of Medicine 1987c;316:1416. [DOI] [PubMed] [Google Scholar]
Kune 2006 {published data only}
- Kune G, Watson L. Colorectal cancer protective effects and the dietary micronutrients folate, methionine, vitamins B6, B12, C, E, selenium, and lycopene. Nutrition and Cancer 2006;56(1):11‐21. [DOI] [PubMed] [Google Scholar]
Kuroda 1988 {published data only}
- Kuroda M, Imura T, Morikawa K, Hasegawa T. Decreased serum levels of selenium and glutathione peroxidase activity associated with aging, malignancy and chronic hemodialysis. Trace Elements in Medicine 1988;5(3):97‐103. [Google Scholar]
Lane 2017 {published data only}
- Lane JA, Oliver SE, Appleby PN, Lentjes MA, Emmett P, Kuh D, et al. Prostate cancer risk related to foods, food groups, macronutrients and micronutrients derived from the UK Dietary Cohort Consortium food diaries. European Journal of Clinical Nutrition 2017;71(2):274‐83. [PUBMED: 27677361] [DOI] [PMC free article] [PubMed] [Google Scholar]
Lawson 2007 {published data only}
- Lawson KA, Wright ME, Subar A, Mouw T, Hollenbeck A, Schatzkin A, et al. Multivitamin use and risk of prostate cancer in the National Institutes of Health‐AARP Diet and Health Study. Journal of the National Cancer Institute 2007;99(10):754‐64. [DOI] [PubMed] [Google Scholar]
Le Marchand 2006 {published data only}
- Marchand L, Saltzman BS, Hankin JH, Wilkens LR, Franke AA, Morris SJ, et al. Sun exposure, diet, and melanoma in Hawaii Caucasians. American Journal of Epidemiology 2006;164(3):232‐45. [DOI] [PubMed] [Google Scholar]
Li 2004b {published data only}
- Li H, Li HQ, Wang Y, Xu HX, Fan WT, Wang ML, et al. An intervention study to prevent gastric cancer by micro‐selenium and large dose of allitridum. Chinese Medical Journal 2004;117(8):1155‐60. [PubMed] [Google Scholar]
Limburg 2005 {published data only}
- Limburg PJ, Wei W, Ahnen DJ, Qiao Y, Hawk ET, Wang G, et al. Randomized, placebo‐controlled, esophageal squamous cell cancer chemoprevention trial of selenomethionine and celecoxib. Gastroenterology 2005;129(3):863‐73. [DOI] [PubMed] [Google Scholar]
Linxian Pilot 2000 {published data only}
- NCI‐OH95‐C‐N026NCI‐P00‐0157. Pilot randomized chemoprevention study of selenium and celecoxib, alone or in combination, in patients with esophageal squamous dysplasia who are residing in Linxian, People's Republic of China. http://www.cancer.gov/search/ViewClinicalTrials.aspx?cdrid=67930&version=HealthProfessional&protocolsearchid=5845119 (accessed 25 February 2009).
Loeb 2015 {published data only}
- Loeb S, Peskoe SB, Joshu CE, Huang WY, Hayes RB, Carter HB, et al. Do environmental factors modify the genetic risk of prostate cancer?. Cancer Epidemiology, Biomarkers & Prevention 2015;24(1):213‐20. [PUBMED: 25342390] [DOI] [PMC free article] [PubMed] [Google Scholar]
Martinez 2014 {published data only}
- Martinez EE, Darke AK, Tangen CM, Goodman PJ, Fowke JH, Klein EA, et al. A functional variant in NKX3.1 associated with prostate cancer risk in the Selenium and Vitamin E Cancer Prevention Trial (SELECT). Cancer Prevention Research 2014;7(9):950‐7. [PUBMED: 24894197] [DOI] [PMC free article] [PubMed] [Google Scholar]
Neuhouser 2009 {published data only}
- Neuhouser ML, Wassertheil‐Smoller S, Tomson C, Aragaki A, Anderson GL, Manson JE, et al. Multivitamin use and risk of cancer and cardiovascular disease in the women's health initiative cohorts. Archives of Internal Medicine 2009;169(3):294‐304. [DOI] [PMC free article] [PubMed] [Google Scholar]
Persson 2000 {published data only}
- Persson‐Moschos ME, Stavenow L, Akesson B, Lindgarde F. Selenoprotein P in plasma in relation to cancer morbidity in middle‐aged Swedish men. Nutrition and Cancer 2000;36(1):19‐26. [DOI] [PubMed] [Google Scholar]
Ray 2006 {published data only}
- Ray AL, Semba RD, Walston J, Ferrucci L, Cappola AR, Ricks MO, et al. Low serum selenium and total carotenoids predict mortality among older women living in the community: the women's health and aging studies. Journal of Nutrition 2006;136(1):172‐6. [DOI] [PubMed] [Google Scholar]
Rayman 2001 {published data only}
- Larsen EH. Prevention of cancer by intervention with Selenium (pilot) ‐ PRECISE‐PILOT. http://www.food.dtu.dk/Default.aspx?ID=20782 (accessed 6 April 2011).
- MRC Clinical Trials Unit, ISRCTN64336220. A randomised double‐blind placebo‐controlled cancer prevention trial with an estimated duration of 5 years with 52,000 subjects recruited from the general populations of the UK, Denmark, Sweden, Finland and the United States. http://www.controlled‐trials.com/ISRCTN64336220 (accessed 6 April 2011).
- Rayman M, ISRCTN25193534. UK prevention of cancer by intervention with selenium. http://www.controlled‐trials.com/ISRCTN25193534/ (accessed 6 April 2011).
- Rayman M, NCT00022165. Selenium in the prevention of cancer. http://clinicaltrials.gov/ct2/show/NCT00022165. England, W12 ONN, United Kingdom: Hammersmith Hospital London, (accessed 6 April 2011).
Rendon {published data only}
- Rendon RA, Fleshner N. Vitamin E, selenium and soy protein in preventing cancer in patients with high‐grade prostate neoplasia. http://clinicaltrials.gov/ct2/show/NCT00064194. www.clinicaltrials.gov, (accessed 6 April 2011).
Steevens 2010b {published data only}
- Steevens J, Schouten LJ, Driessen AL, Huysentruyt CJ, Keulemans YC, Goldbohm RA, et al. Toenail selenium status and the risk of Barrett's esophagus: the Netherlands Cohort Study. Cancer Causes Control 2010;21(12):2259‐68. [DOI] [PMC free article] [PubMed] [Google Scholar]
Thompson 2009 {published data only}
- Thompson CA, Habermann TM, Wang AH, Vierkant RA, Folsom AR, Ross JA, et al. Antioxidant intake from fruits, vegetables and other sources and risk of non‐Hodgkin lymphoma: the Iowa Women's Health Study. International Journal of Cancer 2009;126(4):992‐1003. [DOI: 10.1002/ijc.24830] [DOI] [PMC free article] [PubMed] [Google Scholar]
Tsugane 1996 {published data only}
- Tsugane S, Hamada GS, Karita K, Tsubono Y, Laurenti R. Cancer patterns and lifestyle among Japanese immigrants and their descendants in the city of Sao Paulo, Brazil. Gann Monographs on Cancer Research 1996;44:43‐50. [Google Scholar]
Ujiie 2002 {published data only}
- Ujiie S, Kikuchi H. The relation between serum selenium value and cancer in Miyagi, Japan: 5‐year follow up study. Tohoku Journal of Experimental Medicine 2002;196(3):99‐109. [DOI] [PubMed] [Google Scholar]
van't Veer 1996 {published data only}
- van't Veer P, Strain JJ, Fernandez‐Crehuet J, Martin BC, Thamm M, Kardinaal AF, et al. Tissue antioxidants and postmenopausal breast cancer: the European Community Multicentre Study on Antioxidants, Myocardial Infarction, and Cancer of the Breast (EURAMIC). Cancer Epidemiology, Biomarkers & Prevention 1996;5(6):441‐7. [PubMed] [Google Scholar]
van Noord 1992 {published data only}
- Noord PAH, Tweel I, Kaaks R, Waard F. Selenium levels and subsequent colorectal cancers: the efficiency gain of a sequential test to a cohort‐nested study with a 1:4 matching ratio. In: Noord PAH editor(s). Selenium and Human Cancer Risk: Nail Keratin as a Tool in Metabolic Epidemiology. Amsterdam: Thesis Publishers, 1992:139‐53. [Google Scholar]
van Noord 1993 {published data only}
- Noord PAH, Maas MJ, Tweel I, Collette C. Selenium and the risk of postmenopausal breast cancer in the DOM cohort. Breast Cancer Research and Treatment 1993;25(1):11‐9. [DOI] [PubMed] [Google Scholar]
Wallace 2009 {published data only}
- Wallace K, Kelsey KT, Schned A, Morris JS, Andrew AS, Karagas MR. Selenium and risk of bladder cancer: a population‐based case‐control study. Cancer Prevention Research (Phila Pa) 2009;2(1):70‐3. [DOI] [PMC free article] [PubMed] [Google Scholar]
Watters 2009 {published data only}
- Watters JL, Park Y, Hollenbeck A, Schatzkin A, Albanes D. Cigarette smoking and prostate cancer in a prospective US cohort study. Cancer Epidemiology, Biomarkers & Prevention 2009;18(9):2427‐35. [DOI] [PMC free article] [PubMed] [Google Scholar]
Wright 2004 {published data only}
- Wright ME, Mayne ST, Stolzenberg‐Solomon RZ, Li Z, Pietinen P, Taylor PR, et al. Development of a comprehensive dietary antioxidant index and application to lung cancer risk in a cohort of male smokers. American Journal of Epidemiology 2004;160(1):68‐76. [DOI] [PubMed] [Google Scholar]
You 2005 {published data only}
- You WC, Li JY, Zhang L, Jin ML, Chang YS, Ma JL, et al. Etiology and prevention of gastric cancer: a population study in a high risk area of China. Chinese Journal of Digestive Diseases 2005;6(4):149‐54. [DOI] [PubMed] [Google Scholar]
Yuan 2006 {published data only}
- Yuan JM, Gao YT, Ong CN, Ross RK, Yu MC. Prediagnostic level of serum retinol in relation to reduced risk of hepatocellular carcinoma. JNCI Cancer Spectrum 2006;98(7):482‐90. [DOI] [PubMed] [Google Scholar]
Zeegers 2009 {published data only}
- Zeegers MP, Bryan RT, Langford C, Billingham L, Murray P, Deshmukh NS, et al. The West Midlands Bladder Cancer Prognosis Programme: rationale and design. BJU International 2010;705(6):784‐8. [DOI: 10.1111/j.1464-410X.2009.08849.x] [DOI] [PubMed] [Google Scholar]
References to ongoing studies
Argos 2013 {published data only}
- Argos M, Rahman M, Parvez F, Dignam J, Islam T, Quasem I, et al. Baseline comorbidities in a skin cancer prevention trial in Bangladesh. European Journal of Clinical Investigation 2013;43(6):579‐88. [PUBMED: 23590571] [DOI] [PMC free article] [PubMed] [Google Scholar]
Additional references
Allen 2016
- Allen NE, Travis RC, Appleby PN, Albanes D, Barnett MJ, Black A, et al. Selenium and prostate cancer: analysis of individual participant data from fifteen prospective studies. Journal of the National Cancer Institute 2016;108(11):djw153. [PUBMED: 27385803] [DOI] [PMC free article] [PubMed] [Google Scholar]
Allingstrup 2015
- Allingstrup M, Afshari A. Selenium supplementation for critically ill adults. Cochrane Database of Systematic Reviews 2015, Issue 7. [DOI: 10.1002/14651858.CD003703.pub3] [DOI] [PMC free article] [PubMed] [Google Scholar]
Arnaud 2007
- Arnaud J, Arnault N, Roussel AM, Bertrais S, Ruffieux D, Galan P, et al. Relationships between selenium, lipids, iron status and hormonal therapy in women of the SU.VI.M.AX cohort. Journal of Trace Elements in Medicine and Biology 2007;21(Suppl 1):66‐9. [DOI] [PubMed] [Google Scholar]
Ashton 2009
- Ashton K, Hooper L, Harvey LJ, Hurst R, Casgrain A, Fairweather‐Tait SJ. Methods of assessment of selenium status in humans: a systematic review. American Journal of Clinical Nutrition 2009;89(6):2025S‐39S. [DOI] [PubMed] [Google Scholar]
Barany 2002
- Barany E, Bergdahl IA, Bratteby LE, Lundh T, Samuelson G, Schutz A, et al. Trace elements in blood and serum of Swedish adolescents: relation to gender, age, residential area, and socioeconomic status. Environmental Research 2002;89(1):72‐84. [DOI] [PubMed] [Google Scholar]
Beane Freeman 2015
- Beane Freeman LE, Karagas MR, Baris D, Schwenn M, Johnson AT, Colt JS, et al. Is the inverse association between selenium and bladder cancer due to confounding by smoking?. American Journal of Epidemiology 2015;181(7):488‐95. [PUBMED: 25776013] [DOI] [PMC free article] [PubMed] [Google Scholar]
Behne 1996
- Behne D, Gessner H, Kyriakopoulos A. Information on the selenium status of several body compartments of rats from the selenium concentrations in blood fractions, hair and nails. Journal of Trace Elements in Medicine and Biology 1996;103:174‐9. [DOI] [PubMed] [Google Scholar]
Behne 2010
- Behne D, Alber D, Kyriakopoulos S. Long‐term selenium supplementation of humans: selenium status and relationships between selenium concentrations in skeletal muscle and indicator materials. Journal of Trace Elements in Medicine and Biology 2010;24(2):99‐105. [DOI] [PubMed] [Google Scholar]
Bhattacharjee 2017
- Bhattacharjee A, Basu A, Biswas J, Sen T, Bhattacharya S. Chemoprotective and chemosensitizing properties of selenium nanoparticle (Nano‐Se) during adjuvant therapy with cyclophosphamide in tumor‐bearing mice. Molecular and Cellular Biochemistry 2017;424(1‐2):13‐33. [PUBMED: 27696310] [DOI] [PubMed] [Google Scholar]
Bjelakovic 2012
- Bjelakovic G, Nikolova D, Gluud LL, Simonetti RG, Gluud C. Antioxidant supplements for prevention of mortality in healthy participants and patients with various diseases. Cochrane Database of Systematic Reviews 2012, Issue 3. [DOI: 10.1002/14651858.CD007176] [DOI] [PubMed] [Google Scholar]
Bjelakovic 2014
- Bjelakovic G, Gluud LL, Nikolova D, Whitfield K, Krstic G, Wetterslev J, et al. Vitamin D supplementation for prevention of cancer in adults. Cochrane Database of Systematic Reviews 2014, Issue 6. [DOI: 10.1002/14651858.CD007469.pub2] [DOI] [PMC free article] [PubMed] [Google Scholar]
Block 2004
- Block E, Glass RS, Jacobsen NE, Johnson S, Kahakachchi C, Kaminski R, et al. Identification and synthesis of a novel selenium‐sulfur amino acid found in selenized yeast: rapid indirect detection NMR methods for characterizing low‐level organoselenium compounds in complex matrices. Journal of Agricultural and Food Chemistry 2004;52(12):3761‐71. [DOI] [PubMed] [Google Scholar]
Bodnar 2012
- Bodnar M, Konieczka P, Namiesnik J. The properties, functions, and use of selenium compounds in living organisms. Journal of Environmental Science and Health, Part C, Environmental Carcinogenesis & Ecotoxicology Reviews 2012;30(3):225‐52. [DOI] [PubMed] [Google Scholar]
Brigelius‐Flohe 2017
- Brigelius‐Flohe R, Flohe L. Selenium and redox signalling. Archives of Biochemistry and Biophysics 2017;617:48‐59. [PUBMED: 27495740] [DOI] [PubMed] [Google Scholar]
Brinkman 2006
- Brinkman M, Reulen RC, Kellen E, Buntinx F, Zeegers MP. Are men with low selenium levels at increased risk of prostate cancer?. European Journal of Cancer 2006;42(15):2463‐71. [DOI] [PubMed] [Google Scholar]
Brodin 2015
- Brodin O, Eksborg S, Wallenberg M, Asker‐Hagelberg C, Larsen EH, Mohlkert D, et al. Pharmacokinetics and toxicity of sodium selenite in the treatment of patients with carcinoma in a phase I clinical trial: the SECAR study. Nutrients 2015;7(6):4978‐94. [PUBMED: 26102212] [DOI] [PMC free article] [PubMed] [Google Scholar]
Bruhn 2009
- Bruhn RL, Stamer WD, Herrygers LA, Levine JM, Noecker RJ. Relationship between glaucoma and selenium levels in plasma and aqueous humour. British Journal of Ophthalmology 2009;93(9):1155‐8. [DOI] [PubMed] [Google Scholar]
Burri 2008
- Burri J, Haldimann M, Dudler V. Selenium status of the Swiss population: assessment and change over a decade. Journal of Trace Elements in Medicine and Biology 2008;22(2):112‐9. [DOI] [PubMed] [Google Scholar]
Chawla 2016
- Chawla R, Loomba R, Chaudhary RJ, Singh S, Dhillon KS. Impact of high selenium exposure on organ function and biochemical profile of the rural population living in seleniferous soils in Punjab, India [2015]. Proceedings of the 4th International Conference on Selenium in the Environment and Human Health. 2015:93‐4.
Chen 2000
- Chen X, Mikhail SS, Ding YW, Yang G, Bondoc F, Yang CS. Effects of vitamin E and selenium supplementation on esophageal adenocarcinogenesis in a surgical model with rats. Carcinogenesis 2000;21(8):1531‐6. [PubMed] [Google Scholar]
Chen 2013
- Chen YC, Prabhu KS, Mastro AM. Is selenium a potential treatment for cancer metastasis?. Nutrients 2013;5(4):1149‐68. [PUBMED: 23567478] [DOI] [PMC free article] [PubMed] [Google Scholar]
Combs 2012
- Combs GF Jr, Jackson MI, Watts JC, Johnson LK, Zeng H, Idso J, et al. Differential responses to selenomethionine supplementation by sex and genotype in healthy adults. British Journal of Nutrition 2012;107(10):1514‐25. [DOI] [PubMed] [Google Scholar]
Connor Gorber 2009
- Connor Gorber S, Schofield‐Hurwitz S, Hardt J, Levasseur G, Tremblay M. The accuracy of self‐reported smoking: a systematic review of the relationship between self‐reported and cotinine‐assessed smoking status. Nicotine & Tobacco Research 2009;11(1):12‐24. [PUBMED: 19246437] [DOI] [PubMed] [Google Scholar]
Cortes‐Jofre 2012
- Cortes‐Jofre M, Rueda JR, Corsini‐Munoz G, Fonseca‐Cortes C, Caraballoso M, Bonfill Cosp X. Drugs for preventing lung cancer in healthy people. Cochrane Database of Systematic Reviews 2012, Issue 10. [DOI: 10.1002/14651858.CD002141.pub2] [DOI] [PubMed] [Google Scholar]
Cui 2017a
- Cui K, Li X, Du Y, Tang X, Arai S, Geng Y, et al. Chemoprevention of prostate cancer in men with high‐grade prostatic intraepithelial neoplasia (HGPIN): a systematic review and adjusted indirect treatment comparison. Oncotarget 2017; Vol. 8, issue 22:36674‐84. [PUBMED: 28415774] [DOI] [PMC free article] [PubMed]
Cui 2017b
- Cui Z, Liu D, Liu C, Liu G. Serum selenium levels and prostate cancer risk: a MOOSE‐compliant meta‐analysis. Medicine 2017;96(5):e5944. [PUBMED: 28151881] [DOI] [PMC free article] [PubMed] [Google Scholar]
Dalton 2008
- Dalton SO, Schuz J, Engholm G, Johansen C, Kjaer SK, Steding‐Jessen M, et al. Social inequality in incidence of and survival from cancer in a population‐based study in Denmark, 1994‐2003: summary of findings. European Journal of Cancer 2008;44(14):2074‐85. [DOI] [PubMed] [Google Scholar]
Egger 1998
- Egger M, Schneider M, Davey Smith G. Spurious precision? Meta‐analysis of observational studies. British Medical Journal 1998;316(7125):140‐4. [DOI] [PMC free article] [PubMed] [Google Scholar]
Fairweather‐Tait 2011
- Fairweather‐Tait SJ, Bao Y, Broadley MR, Collings R, Ford D, Hesketh JE, et al. Selenium in human health and disease. Antioxidants & Redox Signaling 2011;14(7):1337‐83. [DOI] [PubMed] [Google Scholar]
Fernandes 2015
- Fernandes AP, Gandin V. Selenium compounds as therapeutic agents in cancer. Biochimica et Biophysica Acta 2015;1850(8):1642‐60. [PUBMED: 25459512] [DOI] [PubMed] [Google Scholar]
Fortmann 2013
- Fortmann SP, Burda BU, Senger CA, Lin JS, Whitlock EP. Vitamin and mineral supplements in the primary prevention of cardiovascular disease and cancer: an updated systematic evidence review for the U.S. Preventive Services Task Force. Annals of Internal Medicine 2013;159(12):824‐34. [DOI] [PubMed] [Google Scholar]
Galan‐Chilet 2017
- Galan‐Chilet I, Grau‐Perez M, Marco G, Guallar E, Martin‐Escudero JC, Dominguez‐Lucas A, et al. A gene‐environment interaction analysis of plasma selenium with prevalent and incident diabetes: the Hortega study. Redox Biology 2017;12:798‐805. [PUBMED: 28437656] [DOI] [PMC free article] [PubMed] [Google Scholar]
Gammelgaard 2011
- Gammelgaard B, Jackson MI, Gabel‐Jensen C. Surveying selenium speciation from soil to cell ‐ forms and transformations. Analytical and Bioanalytical Chemistry 2011;399(5):1743‐63. [DOI] [PubMed] [Google Scholar]
Gerritsen 2015
- Gerritsen M, Berndt N, Lechner L, Vries H, Mudde A, Bolman C. Self‐reporting of smoking cessation in cardiac patients: how reliable is it and is reliability associated with patient characteristics?. Journal of Addiction Medicine 2015;9(4):308‐16. [PUBMED: 26083956] [DOI] [PubMed] [Google Scholar]
Gerstenberger 2015
- Gerstenberger JP, Bauer SR, Blarigan EL, Sosa E, Song X, Witte JS, et al. Selenoprotein and antioxidant genes and the risk of high‐grade prostate cancer and prostate cancer recurrence. The Prostate 2015;75(1):60‐9. [PUBMED: 25284284] [DOI] [PMC free article] [PubMed] [Google Scholar]
Gong 2016
- Gong HY, He JG, Li BS. Meta‐analysis of the association between selenium and gastric cancer risk. Oncotarget 2016;7(13):15600‐5. [PUBMED: 26862854] [DOI] [PMC free article] [PubMed] [Google Scholar]
Goossens 2016
- Goossens ME, Zeegers MP, Poppel H, Joniau S, Ackaert K, Ameye F, et al. Phase III randomised chemoprevention study with selenium on the recurrence of non‐invasive urothelial carcinoma. The SELEnium and BLAdder cancer Trial. European Journal of Cancer 2016;69:9‐18. [PUBMED: 27814472] [DOI] [PubMed] [Google Scholar]
Gorlova 2006
- Gorlova OY, Zhang Y, Schabath MB, Lei L, Zhang Q, Amos CI, et al. Never smokers and lung cancer risk: a case‐control study of epidemiological factors. International Journal of Cancer 2006;118(7):1798‐804. [DOI] [PubMed] [Google Scholar]
GRADE Working Group 2004
- GRADE Working Group 2004. Grading quality of evidence and strength of recommendations. BMJ 2004;328(7454):1490–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
GRADEpro GDT [Computer program]
- Hamilton (ON): McMaster University (developed by Evidence Prime), 2015. Available from gradepro.org. GRADEproGDT [Computer program]. Version accessed 10 May 2017. Hamilton (ON): McMaster University (developed by Evidence Prime), 2015. Available from gradepro.org, 2015.
Greenhalgh 2005
- Greenhalgh T, Peacock R. Effectiveness and efficiency of search methods in systematic reviews of complex evidence: audit of primary sources. BMJ 2005;331(7524):1064‐5. [PUBMED: 16230312] [DOI] [PMC free article] [PubMed] [Google Scholar]
Guallar 2013
- Guallar E, Stranges S, Mulrow C, Appel LJ, Miller ER III. Enough is enough: stop wasting money on vitamin and mineral supplements. Annals of Internal Medicine 2013;159(2):850‐1. [DOI] [PubMed] [Google Scholar]
Gundacker 2006
- Gundacker C, Komarnicki G, Zodl B, Forster C, Schuster E, Wittmann K. Whole blood mercury and selenium concentrations in a selected Austrian population: does gender matter?. Science of the Total Environment 2006;372(1):76‐86. [DOI] [PubMed] [Google Scholar]
Guyatt 2011
- Guyatt GH, Oxman AD, Vist G, Kunz R, Brozek J, Alonso‐Coello P, et al. GRADE guidelines: 4. Rating the quality of evidence ‐ study limitations (risk of bias). Journal of Clinical Epidemiology 2011;64(4):407‐15. [PUBMED: 21247734] [DOI] [PubMed] [Google Scholar]
Haldimann 1996
- Haldimann M, Venner TY, Zimmerli B. Determination of selenium in the serum of healthy Swiss adults and correlation to dietary intake. Journal of Trace Elements in Medicine and Biology 1996;10(1):31‐45. [DOI] [PubMed] [Google Scholar]
Hatfield 2014
- Hatfield DL, Tsuji PA, Carlson BA, Gladyshev VN. Selenium and selenocysteine: roles in cancer, health, and development. Trends in Biochemical Sciences 2014;39(3):112‐20. [DOI] [PMC free article] [PubMed] [Google Scholar]
Hazane‐Puch 2013
- Hazane‐Puch F, Champelovier P, Arnaud J, Garrel C, Ballester B, Faure P, et al. Long‐term selenium supplementation in HaCaT cells: importance of chemical form for antagonist (protective versus toxic) activities. Biological Trace Elements Research 2013;154(2):288‐98. [DOI] [PubMed] [Google Scholar]
Hercberg 2004
- Hercberg S, Galan P, Preziosi P, Bertrais S, Mennen L, Malvy D, et al. The SU.VI.MAX study: a randomized, placebo‐controlled trial of the health effects of antioxidant vitamins and minerals. Archives of Internal Medicine 2004;164(21):2335‐42. [DOI] [PubMed] [Google Scholar]
Higgins 2003
- Higgins JPT, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta‐analyses. British Medical Journal 2003;327(7414):557‐60. [DOI] [PMC free article] [PubMed] [Google Scholar]
Higgins 2009
- Higgins JPT, Thompson SG, Spiegelhalter DJ. A re‐evaluation of random‐effects meta‐analysis. Journal of the Royal Statistical Society Series A (Statistics in Society) 2009;172(1):137‐59. [DOI] [PMC free article] [PubMed] [Google Scholar]
Higgins 2011a
- Higgins JPT, Green S, editor(s). Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0 (updated March 2011). The Cochrane Collaboration, 2011. Available from www.cochrane‐handbook.org.
Higgins 2011b
- Higgins JP, Altman DG, Gøtzsche PC, Jüni P, Moher D, Oxman AD, et al. Cochrane Bias Methods Group, Cochrane Statistical Methods Group. The Cochrane Collaboration's tool for assessing risk of bias in randomised trials. British Medical Journal 2011;343:d5928. [DOI] [PMC free article] [PubMed] [Google Scholar]
Hunter 1990
- Hunter DJ, Morris JS, Chute CG, Kushner E, Colditz GA, Stampfer MJ, et al. Predictors of selenium concentration in human toenails. American Journal of Epidemiology 1990;132(1):114‐22. [DOI] [PubMed] [Google Scholar]
Hurst 2013a
- Hurst R, Collings R, Harvey LJ, King M, Hooper L, Bouwman J, et al. EURRECA‐estimating selenium requirements for deriving dietary reference values. Critical Reviews in Food Science and Nutrition 2013;53(10):1077‐96. [DOI] [PubMed] [Google Scholar]
Hurst 2013b
- Hurst R, Siyame EW, Young SD, Chilimba AD, Joy EJ, Black CR, et al. Soil‐type influences human selenium status and underlies widespread selenium deficiency risks in Malawi. Scientific Reports 2013;3:1425. [PUBMED: 23478344] [DOI] [PMC free article] [PubMed] [Google Scholar]
IARC 2014
- IARC. In: Stewart BW, Wild CP editor(s). World Cancer Report 2014. Lyon, France: IARC Press, International Agency for Research on Cancer, 2014. [Google Scholar]
Institute of Medicine 2009
- Institute of Medicine. Dietary reference intakes: elements. http://www.iom.edu/Object.File/Master/54/395/DRIs.Elements.pdf (accessed 18 May 2009).
Jablonska 2013
- Jablonska E, Gromadzinska J, Klos A, Bertrandt J, Skibniewska K, Darago A, et al. Selenium, zinc and copper in the Polish diet. Journal of Food Composition and Analysis 2013;31(2):259–65. [Google Scholar]
Jablonska 2015a
- Jablonska E, Vinceti M. Selenium and human health: witnessing a Copernican revolution?. Journal of Environmental Science and Health. Part C, Environmental Carcinogenesis & Ecotoxicology Reviews 2015;33(3):328‐68. [PUBMED: 26074278] [DOI] [PubMed] [Google Scholar]
Jablonska 2015b
- Jablonska E, Gromadzinska J, Peplonska B, Fendler W, Reszka E, Krol MB, et al. Lipid peroxidation and glutathione peroxidase activity relationship in breast cancer depends on functional polymorphism of GPX1. BMC Cancer 2015;15:657. [PUBMED: 26446998] [DOI] [PMC free article] [PubMed] [Google Scholar]
Jablonska 2016
- Jablonska E, Raimondi S, Gromadzinska J, Reszka E, Wieczorek E, Krol MB, et al. DNA damage and oxidative stress response to selenium yeast in the non‐smoking individuals: a short‐term supplementation trial with respect to GPX1 and SEPP1 polymorphism. European Journal of Nutrition 2016;55(8):2469‐84. [PUBMED: 26658762] [DOI] [PMC free article] [PubMed] [Google Scholar]
Jaffé 1992
- Jaffé W. Selenio, un elemento esencial y toxico. Datos de Latinoamerica. Archivos Latinoamericanos de Nutrición 1992;42(2):90‐3. [PubMed] [Google Scholar]
Jerome‐Morais 2011
- Jerome‐Morais A, Diamond AM, Wright ME. Dietary supplements and human health: for better or for worse?. Molecular Nutrition & Food Research 2011;55(1):122‐35. [DOI] [PubMed] [Google Scholar]
Jossa 1991
- Jossa F, Trevisan M, Krogh V, Farinaro E, Giumetti D, Fusco G, et al. Serum selenium and coronary heart disease risk factors in southern Italian men. Atherosclerosis 1991;87(2‐3):129‐34. [DOI] [PubMed] [Google Scholar]
Kafai 2003
- Kafai MR, Ganji V. Sex, age, geographical location, smoking, and alcohol consumption influence serum selenium concentrations in the USA: third National Health and Nutrition Examination Survey, 1988‐1994. Journal of Trace Elements in Medicine and Biology 2003;17(1):13‐8. [DOI] [PubMed] [Google Scholar]
Kandas 2009
- Kandas NO, Randolph C, Bosland MC. Differential effects of selenium on benign and malignant prostate epithelial cells: stimulation of LNCaP cell growth by noncytotoxic, low selenite concentrations. Nutrition and Cancer 2009;61(2):251‐64. [DOI] [PMC free article] [PubMed] [Google Scholar]
Kant 2007
- Kant AK, Graubard BI. Ethnicity is an independent correlate of biomarkers of micronutrient intake and status in American adults. Journal of Nutrition 2007;137(11):2456‐63. [DOI] [PubMed] [Google Scholar]
Karamali 2015
- Karamali M, Nourgostar S, Zamani A, Vahedpoor Z, Asemi Z. The favourable effects of long‐term selenium supplementation on regression of cervical tissues and metabolic profiles of patients with cervical intraepithelial neoplasia: a randomised, double‐blind, placebo‐controlled trial. British Journal of Nutrition 2015;114(12):2039‐45. [PUBMED: 26439877] [DOI] [PubMed] [Google Scholar]
Karita 2003
- Karita K, Sasaki S, Ishihara J, Tsugane S. Validity of a self‐administered food frequency questionnaire used in the 5‐year follow‐up survey of the JPHC Study to assess selenium intake: comparison with dietary records and blood levels. Journal of Epidemiology 2003;13(1 Suppl):S92‐7. [DOI] [PMC free article] [PubMed] [Google Scholar]
Kasaikina 2013
- Kasaikina MV, Turanov AA, Avanesov A, Schweizer U, Seeher S, Bronson RT, et al. Contrasting roles of dietary selenium and selenoproteins in chemically induced hepatocarcinogenesis. Carcinogenesis 2013;34(5):1089‐95. [PUBMED: 23389288] [DOI] [PMC free article] [PubMed] [Google Scholar]
Kim 2001
- Kim YY, Mahan DC. Comparative effects of high dietary levels of organic and inorganic selenium on selenium toxicity of growing‐finishing pigs. Journal of Animal Science 2001;79(4):942‐8. [DOI] [PubMed] [Google Scholar]
Kryscio 2017
- Kryscio RJ, Abner EL, Caban‐Holt A, Lovell M, Goodman P, Darke AK, et al. Association of antioxidant supplement use and dementia in the prevention of Alzheimer's disease by vitamin E and selenium trial (preadvise). JAMA Neurology 2017;74(5):567‐73. [PUBMED: 28319243] [DOI] [PMC free article] [PubMed] [Google Scholar]
Kushi 2012
- Kushi LH, Doyle C, McCullough M, Rock CL, Demark‐Wahnefried W, Bandera EV, et al. American Cancer Society Guidelines on nutrition and physical activity for cancer prevention: reducing the risk of cancer with healthy food choices and physical activity. CA: a Cancer Journal for Clinicians 2012;62(1):30‐67. [PUBMED: 22237782] [DOI] [PubMed] [Google Scholar]
Labunskyy 2014
- Labunskyy VM, Hatfield DL, Gladyshev VN. Selenoproteins: molecular pathways and physiological roles. Physiological Reviews 2014;94(3):739‐77. [PUBMED: 24987004] [DOI] [PMC free article] [PubMed] [Google Scholar]
Lance 2017
- Lance P, Alberts DS, Thompson PA, Fales L, Wang F, San Jose J, et al. Colorectal adenomas in participants of the select randomized trial of selenium and vitamin E for prostate cancer prevention. Cancer Prevention Research 2017;10(1):45‐54. [PUBMED: 27777235] [DOI] [PMC free article] [PubMed] [Google Scholar]
Langendam 2013
- Langendam MW, Akl EA, Dahm P, Glasziou P, Guyatt G, Schunemann HJ. Assessing and presenting summaries of evidence in Cochrane Reviews. Systematic Reviews 2013;23(2):81. [DOI] [PMC free article] [PubMed] [Google Scholar]
Lawlor 2004
- Lawlor DA, Davey SG, Kundu D, Bruckdorfer KR, Ebrahim S. Those confounded vitamins: what can we learn from the differences between observational versus randomised trial evidence?. Lancet 2004;363(9422):1724‐7. [DOI] [PubMed] [Google Scholar]
Li 2012
- Li S, Xiao T, Zheng B. Medical geology of arsenic, selenium and thallium in China. Science of the Total Environment 2012;421‐422:31‐40. [PUBMED: 21440288] [DOI] [PubMed] [Google Scholar]
Longnecker 1996
- Longnecker MP, Stram DO, Taylor PR, Levander OA, Howe M, Veillon C, et al. Use of selenium concentration in whole blood, serum, toenails, or urine as a surrogate measure of selenium intake. Epidemiology 1996;7(4):384‐90. [DOI] [PubMed] [Google Scholar]
Lu 2016
- Lu J, Zhang J, Jiang C, Deng Y, Ozten N, Bosland MC. Cancer chemoprevention research with selenium in the post‐SELECT era: promises and challenges. Nutrition and Cancer 2016;68(1):1‐17. [PUBMED: 26595411] [DOI] [PMC free article] [PubMed] [Google Scholar]
Mandrioli 2017
- Mandrioli J, Michalke B, Solovyev N, Grill P, Violi F, Lunetta C, et al. Elevated levels of selenium species in cerebrospinal fluid of amyotrophic lateral sclerosis patients with disease‐associated gene mutations. Neuro‐degenerative Diseases 2017;17(4‐5):171‐80. [PUBMED: 28478440] [DOI] [PubMed] [Google Scholar]
Marschall 2017
- Marschall TA, Kroepfl N, Jensen KB, Bornhorst J, Meermann B, Kuehnelt D, et al. Tracing cytotoxic effects of small organic Se species in human liver cells back to total cellular Se and Se metabolites. Metallomics : Integrated Biometal Science 2017;9(3):268‐77. [PUBMED: 28184394] [DOI] [PubMed] [Google Scholar]
McNaughton 2005b
- McNaughton SA, Marks GC, Green AC. Role of dietary factors in the development of basal cell cancer and squamous cell cancer of the skin. Cancer Epidemiology, Biomarkers & Prevention 2005;14(7):1596‐607. [DOI] [PubMed] [Google Scholar]
Meader 2014
- Meader N, King K, Llewellyn A, Norman G, Brown J, Rodgers M, et al. A checklist designed to aid consistency and reproducibility of GRADE assessments: development and pilot validation. Systematic Reviews 2014;3:82. [DOI] [PMC free article] [PubMed] [Google Scholar]
Meplan 2012
- Meplan C, Rohrmann S, Steinbrecher A, Schomburg L, Jansen E, Linseisen J, et al. Polymorphisms in thioredoxin reductase and selenoprotein K genes and selenium status modulate risk of prostate cancer. PloS One 2012;7(11):e48709. [PUBMED: 23133653] [DOI] [PMC free article] [PubMed] [Google Scholar]
Meplan 2014
- Meplan C, Hesketh J. Selenium and cancer: a story that should not be forgotten‐insights from genomics. Cancer Treatment and Research 2014;159:145‐66. [PUBMED: 24114479] [DOI] [PubMed] [Google Scholar]
Meplan 2015
- Meplan C. Selenium and chronic diseases: a nutritional genomics perspective. Nutrients 2015;7(5):3621‐51. [PUBMED: 25988760] [DOI] [PMC free article] [PubMed] [Google Scholar]
Michalke 2017
- Michalke B, Solovyev N, Vinceti M. Se‐speciation investigations at neural barrier. Se2017: The 11th International Symposium on Selenium in Biology and Medicine and the 5th International Conference on Selenium in the Environment and Human Health; August 13‐17, 2017; Stockholm (Meeting Abstract). 2017.
Misra 2015
- Misra S, Boylan M, Selvam A, Spallholz JE, Bjornstedt M. Redox‐active selenium compounds ‐ from toxicity and cell death to cancer treatment. Nutrients 2015;7(5):3536‐56. [PUBMED: 25984742] [DOI] [PMC free article] [PubMed] [Google Scholar]
Morales 2013
- Morales NA, Romano MA, Michael Cummings K, Marshall JR, Hyland AJ, Hutson A, et al. Accuracy of self‐reported tobacco use in newly diagnosed cancer patients. Cancer Causes and Control 2013;24(6):1223‐30. [PUBMED: 23553611] [DOI] [PMC free article] [PubMed] [Google Scholar]
Morris 2013
- Morris JS, Crane SB. Selenium toxicity from a misformulated dietary supplement, adverse health effects, and the temporal response in the nail biologic monitor. Nutrients 2013;5(4):1024‐57. [DOI] [PMC free article] [PubMed] [Google Scholar]
Moyad 2012
- Moyad MA. Heart healthy=prostate healthy: SELECT, the symbolic end of preventing prostate cancer via heart unhealthy and over anti‐oxidation mechanisms?. Asian Journal of Andrology 2012;14(2):243‐4. [DOI] [PMC free article] [PubMed] [Google Scholar]
Moyer 2014
- Moyer VA. Vitamin, mineral, and multivitamin supplements for the primary prevention of cardiovascular disease and cancer: U.S. Preventive Services Task Force recommendation statement. Annals of Internal Medicine 2014;160(8):558‐64. [PUBMED: 24566474] [DOI] [PubMed] [Google Scholar]
Muecke 2014
- Muecke R, Micke O, Schomburg L, Glatzel M, Reichl B, Kisters K, et al. Multicenter, phase III trial comparing selenium supplementation with observation in gynecologic radiation oncology: follow‐up analysis of the survival data 6 years after cessation of randomization. Integrative Cancer Therapies 2014;13(6):463‐7. [PUBMED: 25015649] [DOI] [PubMed] [Google Scholar]
National Toxicology Program 2011
- National Toxicology Program. Selenium sulfide. Report on carcinogens: carcinogen profiles/U.S. Dept. of Health and Human Services, Public Health Service, National Toxicology Program 2011;12:376‐7. [PubMed] [Google Scholar]
Nishino 2001
- Nishino Y, Tsubono Y, Tsuji I, Komatsu S, Kanemura S, Nakatsuka H, et al. Passive smoking at home and cancer risk: a population‐based prospective study in Japanese nonsmoking women. Cancer Causes and Control 2001;12(9):797‐802. [DOI] [PubMed] [Google Scholar]
Niskar 2003
- Niskar AS, Paschal DC, Kieszak SM, Flegal KM, Bowman B, Gunter EW, et al. Serum selenium levels in the US population: Third National Health and Nutrition Examination Survey, 1988‐1994. Biological Trace Element Research 2003;91(1):1‐10. [DOI] [PubMed] [Google Scholar]
Novoselov 2005
- Novoselov SV, Calvisi DF, Labunskyy VM, Factor VM, Carlson BA, Fomenko DE, et al. Selenoprotein deficiency and high levels of selenium compounds can effectively inhibit hepatocarcinogenesis in transgenic mice. Oncogene 2005;24(54):8003‐11. [DOI] [PubMed] [Google Scholar]
Office of Dietary Supplements 2009
- Office of Dietary Supplements, NIH Clinical Centers, National Institutes of Health. Dietary Supplement Fact Sheet: Selenium. http://ods.od.nih.gov/factsheets/selenium.asp (accessed 18 May 2009).
Ovaskainen 1993
- Ovaskainen ML, Virtamo J, Alfthan G, Haukka J, Pietinen P, Taylor PR, et al. Toenail selenium as an indicator of selenium intake among middle‐aged men in an area with low soil selenium. American Journal of Clinical Nutrition 1993;57(5):662‐5. [DOI] [PubMed] [Google Scholar]
Panter 1996
- Panter KE, Hartley WJ, James LF, Mayland HF, Stegelmeier BL, Kechele PO. Comparative toxicity of selenium from seleno‐DL‐methionine, sodium selenate, and Astragalus bisulcatus in pigs. Fundamental and Applied Toxicology 1996;32(2):217‐23. [PubMed] [Google Scholar]
Pearce 2004
- Pearce N. The globalization of epidemiology: introductory remarks. International Journal of Epidemiology 2004;33(5):1127‐31. [DOI] [PubMed] [Google Scholar]
Penney 2010
- Penney KL, Schumacher FR, Li H, Kraft P, Morris JS, Kurth T, et al. A large prospective study of SEP15 genetic variation, interaction with plasma selenium levels, and prostate cancer risk and survival. Cancer Prevention Research (Philadelphia, Pa.) 2010;3(5):604‐10. [PUBMED: 20424130] [DOI] [PMC free article] [PubMed] [Google Scholar]
Penney 2013
- Penney KL, Li H, Mucci LA, Loda M, Sesso HD, Stampfer MJ, et al. Selenoprotein P genetic variants and mrna expression, circulating selenium, and prostate cancer risk and survival. The Prostate 2013;73(7):700‐5. [PUBMED: 23129481] [DOI] [PMC free article] [PubMed] [Google Scholar]
Pestitschek 2013
- Pestitschek M, Sonneck‐Koenne C, Zakavi SR, Li S, Knoll P, Mirzaei S. Selenium intake and selenium blood levels: a novel food frequency questionnaire. Wiener Klinische Wochenschrift 2013;125(5‐6):160‐4. [DOI] [PubMed] [Google Scholar]
Pildal 2007
- Pildal J, Hrobjartsson A, Jorgensen KJ, Hilden J, Altman DG, Gotzsche PC. Impact of allocation concealment on conclusions drawn from meta‐analyses of randomized trials. International Journal of Epidemiology 2007;36(4):847‐57. [DOI] [PubMed] [Google Scholar]
Posadzki 2013
- Posadzki P, Lee MS, Onakpoya I, Lee HW, Ko BS, Ernst E. Dietary supplements and prostate cancer: a systematic review of double‐blind, placebo‐controlled randomised clinical trials. Maturitas 2013;75(2):125‐30. [PUBMED: 23567264] [DOI] [PubMed] [Google Scholar]
Ramamoorthy 2015
- Ramamoorthy V, Rubens M, Saxena A, Shehadeh N. Selenium and vitamin E for prostate cancer ‐ justifications for the SELECT study. Asian Pacific Journal of Cancer Prevention 2015;16(7):2619‐27. [PUBMED: 25854337] [DOI] [PubMed] [Google Scholar]
Rapiti 2009
- Rapiti E, Fioretta G, Schaffar R, Neyroud‐Caspar I, Verkooijen HM, Schmidlin F, et al. Impact of socioeconomic status on prostate cancer diagnosis, treatment, and prognosis. Cancer 2009;115(23):5556‐65. [DOI] [PubMed] [Google Scholar]
Ravn‐Haren 2008
- Ravn‐Haren G, Bugel S, Krath BN, Hoac T, Stagsted J, Jorgensen K, et al. A short‐term intervention trial with selenate, selenium‐enriched yeast and selenium‐enriched milk: effects on oxidative defence regulation. The British journal of nutrition 2008;99(4):883‐92. [PUBMED: 17888202] [DOI] [PubMed] [Google Scholar]
Rayman 2004
- Rayman MP. The use of high‐selenium yeast to raise selenium status: how does it measure up?. British Journal of Nutrition 2004;92(4):557‐73. [DOI] [PubMed] [Google Scholar]
Rayman 2008
- Rayman MP. Food‐chain selenium and human health: emphasis on intake. British Journal of Nutrition 2008;100(2):254‐68. [DOI] [PubMed] [Google Scholar]
Rayman 2009
Rayman 2012
- Rayman MP. Selenium and human health. Lancet 2012;379(9822):1256‐68. [DOI] [PubMed] [Google Scholar]
Rees 2013
- Rees K, Hartley L, Day C, Flowers N, Clarke A, Stranges S. Selenium supplementation for the primary prevention of cardiovascular disease. Cochrane Database of Systematic Reviews 2013;1:CD009671. [DOI] [PMC free article] [PubMed] [Google Scholar]
Review Manager 2014 [Computer program]
- Copenhagen: Nordic Cochrane Centre, The Cochrane Collaboration, 2014. Review Manager 5 (RevMan 5) [Computer program]. Version 5.3. Copenhagen: Nordic Cochrane Centre, The Cochrane Collaboration, 2014, 2014.
Riboli 2002
- Riboli E, Hunt KJ, Slimani N, Ferrari P, Norat T, Fahey M, et al. European prospective investigation into cancer and nutrition (epic): study populations and data collection. Public Health Nutrition 2002;5(6B):1113‐24. [PUBMED: 12639222] [DOI] [PubMed] [Google Scholar]
Richie 2014
- Richie JP Jr, Das A, Calcagnotto AM, Sinha R, Neidig W, Liao J, et al. Comparative effects of two different forms of selenium on oxidative stress biomarkers in healthy men: a randomized clinical trial. Cancer prevention research (Philadelphia, Pa.) 2014;7(8):796‐804. [PUBMED: 24938534] [DOI] [PMC free article] [PubMed] [Google Scholar]
Rocourt 2013
- Rocourt CR, Cheng WH. Selenium supranutrition: are the potential benefits of chemoprevention outweighed by the promotion of diabetes and insulin resistance?. Nutrients 2013;5(4):1349‐65. [DOI] [PMC free article] [PubMed] [Google Scholar]
Rodriguez 1995
- Rodriguez Rodriguez EM, Sanz Alaejos MT, Diaz Romero C. Urinary selenium status of healthy people. European Journal of Clinical Chemistry and Clinical Biochemistry 1995;33(3):127‐33. [DOI] [PubMed] [Google Scholar]
Rose 2014
- Rose AH, Bertino P, Hoffmann FW, Gaudino G, Carbone M, Hoffmann PR. Increasing dietary selenium elevates reducing capacity and ERK activation associated with accelerated progression of select mesothelioma tumors. American Journal of Pathology 2014;184(4):1041‐9. [DOI] [PMC free article] [PubMed] [Google Scholar]
Sacco 2013
- Sacco JE, Dodd KW, Kirkpatrick SI, Tarasuk V. Voluntary food fortification in the United States: potential for excessive intakes. European Journal of Clinical Nutrition 2013;67(6):592‐7. [DOI] [PubMed] [Google Scholar]
Sandsveden 2017
- Sandsveden M, Manjer J. Selenium and breast cancer risk: a prospective nested case‐control study on serum selenium levels, smoking habits and overweight. International Journal of Cancer 2017;141(8):1741‐50. [PUBMED: 28681438] [DOI] [PubMed] [Google Scholar]
Satia 2006
- Satia JA, King IB, Morris JS, Stratton K, White E. Toenail and plasma levels as biomarkers of selenium exposure. Annals of Epidemiology 2006;16(1):53‐8. [DOI] [PubMed] [Google Scholar]
Schrauzer 1977
- Schrauzer GN, White DA, Schneider CJ. Cancer mortality correlation studies ‐ III: statistical associations with dietary selenium intakes. Biological Chemistry 1977;7(1):23‐31. [DOI] [PubMed] [Google Scholar]
Schwingshackl 2017
- Schwingshackl L, Schwedhelm C, Hoffmann G, Lampousi AM, Knuppel S, Iqbal K, et al. Food groups and risk of all‐cause mortality: a systematic review and meta‐analysis of prospective studies. American Journal of Clinical Nutrition 2017;105(6):1462‐73. [PUBMED: 28446499] [DOI] [PubMed] [Google Scholar]
Shamberger 1969
- Shamberger RJ, Frost DV. Possible protective effect of selenium against human cancer. Canadian Medical Association Journal 1969;100(14):682. [PMC free article] [PubMed] [Google Scholar]
Shigemi 2017
- Shigemi Z, Manabe K, Hara N, Baba Y, Hosokawa K, Kagawa H, et al. Methylseleninic acid and sodium selenite induce severe ER stress and subsequent apoptosis through UPR activation in PEL cells. Chemico‐biological Interactions 2017;266:28‐37. [PUBMED: 28161410] [DOI] [PubMed] [Google Scholar]
Slattery 2012
- Slattery ML, Lundgreen A, Welbourn B, Corcoran C, Wolff RK. Genetic variation in selenoprotein genes, lifestyle, and risk of colon and rectal cancer. PloS One 2012;7(5):e37312. [PUBMED: 22615972] [DOI] [PMC free article] [PubMed] [Google Scholar]
Slavik 2008
- Slavik P, Illek J, Brix M, Hlavicova J, Rajmon R, Jilek F. Influence of organic versus inorganic dietary selenium supplementation on the concentration of selenium in colostrum, milk and blood of beef cows. Acta Veterinaria Scandinavica 2008;50:43. [DOI] [PMC free article] [PubMed] [Google Scholar]
Smith 2000
- Smith AM, Chang MP, Medeiros LC. Generational differences in selenium status of women. Biological Trace Element Research 2000;75(1‐3):157‐65. [DOI] [PubMed] [Google Scholar]
Solovyev 2013
- Solovyev N, Berthele A, Michalke B. Selenium speciation in paired serum and cerebrospinal fluid samples. Analytical and Bioanalytical Chemistry 2013;405(6):1875‐84. [DOI] [PubMed] [Google Scholar]
Steen 2008
- Steen A, Strom T, Bernhoft A. Organic selenium supplementation increased selenium concentrations in ewe and newborn lamb blood and in slaughter lamb meat compared to inorganic selenium supplementation. Acta Veterinaria Scandinavica 2008;50:7. [DOI] [PMC free article] [PubMed] [Google Scholar]
Steinbrenner 2013
- Steinbrenner H, Speckmann B, Sies H. Toward understanding success and failures in the use of selenium for cancer prevention. Antioxidants & Redox Signaling 2013;19(2):181‐91. [DOI] [PMC free article] [PubMed] [Google Scholar]
Stranges 2007
- Stranges S, Marshall JR, Natarajan R, Donahue RP, Trevisan M, Combs GF, et al. Effects of long‐term selenium supplementation on the incidence of type 2 diabetes: a randomized trial. Annals of Internal Medicine 2007;147(4):217‐23. [DOI] [PubMed] [Google Scholar]
Stranges 2010
- Stranges S, Sieri S, Vinceti M, Grioni S, Guallar E, Laclaustra M, et al. A prospective study of dietary selenium intake and risk of type 2 diabetes. BMC Public Health 2010;10:564. [DOI] [PMC free article] [PubMed] [Google Scholar]
Stratton 2010
- Stratton MS, Algotar AM, Ranger‐Moore J, Stratton SP, Slate EH, Hsu CH, et al. Oral selenium supplementation has no effect on prostate‐specific antigen velocity in men undergoing active surveillance for localized prostate cancer. Cancer Prevention Research (Philadelphia, Pa.) 2010;3(8):1035‐43. [PUBMED: 20647337] [DOI] [PMC free article] [PubMed] [Google Scholar]
Su 2005
- Su YP, Tang JM, Tang Y, Gao HY. Histological and ultrastructural changes induced by selenium in early experimental gastric carcinogenesis. World Journal of Gastroenterology 2005;11(29):4457‐60. [DOI] [PMC free article] [PubMed] [Google Scholar]
Su 2016
- Su LQ, Jin YL, Unverzagt FW, Cheng YB, Hake AM, Ran L, et al. Nail selenium level and diabetes in older people in rural China. Biomedical and Environmental Sciences 2016;29(11):818‐24. [PUBMED: 27998388] [DOI] [PubMed] [Google Scholar]
Takata 2011
- Takata Y, Kristal AR, King IB, Song X, Diamond AM, Foster CB, et al. Serum selenium, genetic variation in selenoenzymes, and risk of colorectal cancer: primary analysis from the Women's Health Initiative Observational Study and Meta‐analysis. Cancer Epidemiology, Biomarkers & Prevention 2011;20(9):1822‐30. [PUBMED: 21765007] [DOI] [PMC free article] [PubMed] [Google Scholar]
Thompson 2016
- Thompson PA, Ashbeck EL, Roe DJ, Fales L, Buckmeier J, Wang F, et al. Selenium supplementation for prevention of colorectal adenomas and risk of associated type 2 diabetes. Journal of the National Cancer Institute 2016;108(12):djw152. [PUBMED: 27530657] [DOI] [PMC free article] [PubMed] [Google Scholar]
Tiwary 2006
- Tiwary AK, Stegelmeier BL, Panter KE, James LF, Hall JO. Comparative toxicosis of sodium selenite and selenomethionine in lambs. Journal of Veterinary Diagnostic Investigation 2006;18(1):61‐70. [DOI] [PubMed] [Google Scholar]
Tsubota‐Utsugi 2012
- Tsubota‐Utsugi M, Imai E, Nakade M, Tsuboyama‐Kasaoka N, Morita A, Tokudome S. Dietary Reference Intakes for Japanese: the summary report from the Scientific Committee of “Dietary Reference intakes for Japanese 2010”. National Institute of Health and Nutrition (http://www.mhlw.go.jp/bunya/kenkou/sessyu‐kijun.html [Report in Japanese] 2012).
Tsuji 2015
- Tsuji PA, Carlson BA, Anderson CB, Seifried HE, Hatfield DL, Howard MT. Dietary selenium levels affect selenoprotein expression and support the interferon‐gamma and IL‐6 immune response pathways in mice. Nutrients 2015;7(8):6529‐49. [PUBMED: 26258789] [DOI] [PMC free article] [PubMed] [Google Scholar]
Vinceti 1998
- Vinceti M, Rothman KJ, Bergomi M, Borciani N, Serra L, Vivoli G. Excess melanoma incidence in a cohort exposed to high levels of environmental selenium. Cancer Epidemiology, Biomarkers & Prevention 1998;7(10):853‐6. [PubMed] [Google Scholar]
Vinceti 2000
- Vinceti M, Rovesti S, Bergomi M, Vivoli G. The epidemiology of selenium and human cancer. Tumori 2000;86(2):105‐18. [DOI] [PubMed] [Google Scholar]
Vinceti 2001
- Vinceti M, Wei ET, Malagoli C, Bergomi M, Vivoli G. Adverse health effects of selenium in humans. Reviews on Environmental Health 2001;16(4):233‐51. [DOI] [PubMed] [Google Scholar]
Vinceti 2009
- Vinceti M, Maraldi T, Bergomi M, Malagoli C. Risk of chronic low‐dose selenium overexposure in humans: insights from epidemiology and biochemistry. Reviews on Environmental Health 2009;24(3):231‐48. [DOI] [PubMed] [Google Scholar]
Vinceti 2012
- Vinceti M, Crespi CM, Malagoli C, Bottecchi I, Ferrari A, Sieri S, et al. A case‐control study of the risk of cutaneous melanoma associated with three selenium exposure indicators. Tumori 2012;98(3):287‐95. [DOI] [PMC free article] [PubMed] [Google Scholar]
Vinceti 2013a
- Vinceti M, Crespi CM, Bonvicini F, Malagoli C, Ferrante M, Marmiroli S, et al. The need for a reassessment of the safe upper limit of selenium in drinking water. Science of the Total Environment 2013;443:633‐42. [DOI] [PubMed] [Google Scholar]
Vinceti 2013b
- Vinceti M, Crespi CM, Malagoli C, Giovane C, Krogh V. Friend or foe? The current epidemiologic evidence on selenium and human cancer risk. Journal of Environmental Science and Health, Part C, Environmental Carcinogenesis & Ecotoxicology Reviews 2013;31(4):305‐41. [DOI] [PMC free article] [PubMed] [Google Scholar]
Vinceti 2013c
- Vinceti M, Solovyev N, Mandrioli J, Crespi CM, Bonvicini F, Arcolin E, et al. Cerebrospinal fluid of newly diagnosed amyotrophic lateral sclerosis patients exhibits abnormal levels of selenium species including elevated selenite. Neurotoxicology 2013;38:25‐32. [DOI] [PMC free article] [PubMed] [Google Scholar]
Vinceti 2014a
Vinceti 2015
- Vinceti M, Grioni S, Alber D, Consonni D, Malagoli C, Agnoli C, et al. Toenail selenium and risk of type 2 diabetes: the ORDET cohort study. Journal of Trace Elements in Medicine and Biology 2015;29:145‐50. [PUBMED: 25169979] [DOI] [PubMed] [Google Scholar]
Vinceti 2016a
- Vinceti M, Rothman KJ. More results but no clear conclusion on selenium and cancer. American Journal of Clinical Nutrition 2016;104(2):245‐6. [PUBMED: 27413135] [DOI] [PubMed] [Google Scholar]
Vinceti 2016b
- Vinceti M, Ballotari P, Steinmaus C, Malagoli C, Luberto F, Malavolti M, et al. Long‐term mortality patterns in a residential cohort exposed to inorganic selenium in drinking water. Environmental Research 2016;150:348‐56. [PUBMED: 27344266] [DOI] [PubMed] [Google Scholar]
Vinceti 2017a
- Vinceti M, Filippini T, Cilloni S, Bargellini A, Vergoni AV, Tsatsakis A, et al. Health risk assessment of environmental selenium: emerging evidence and challenges (Review). Molecular Medicine Reports 2017;15:3323‐35. [PUBMED: 28339083] [DOI] [PMC free article] [PubMed] [Google Scholar]
Vinceti 2017b
- Vinceti M, Filippini T, Cilloni S, Crespi CM. The epidemiology of selenium and human cancer. Advances in Cancer Research 2017;136:1‐48. [DOI] [PubMed] [Google Scholar]
Vinceti 2017c
- Vinceti M, Filippini T, Cilloni S, Platt J, Jess C. Implications of different methods for literature searching and assessment in systematic reviews and meta‐analyses: a case study. Global Evidence Summit: Using evidence ‐ Improving lives; September 13‐16, 2017; Cape Town (Meeting Abstract). 2017.
Vinceti 2017d
- Vinceti M, Chiari A, Eichmüller M, Rothman KJ, Filippini T, Malagoli C, et al. A selenium species in cerebrospinal fluid predicts conversion to Alzheimer'sdementia in persons with mild cognitive impairment. Alzheimers Res Ther 2017;9(1):100. [DOI] [PMC free article] [PubMed] [Google Scholar]
Visser 2017
- Visser ME, Durao S, Sinclair D, Irlam JH, Siegfried N. Micronutrient supplementation in adults with HIV infection. Cochrane Database of Systematic Reviews 2017;5:CD003650. [PUBMED: 28518221] [DOI] [PMC free article] [PubMed] [Google Scholar]
Waters 2004
- Waters DJ, Chiang EC, Cooley DM, Morris JS. Making sense of sex and supplements: differences in the anticarcinogenic effects of selenium in men and women. Mutation Research 2004;551(1‐2):91‐107. [DOI] [PubMed] [Google Scholar]
Waters 2013
- Waters DJ, Shen S, Kengeri SS, Chiang EC, Combs GF, Morris JS, Bostwick DG. Toward interpreting the null results of SELECT: Direct comparison of the prostatic tissue potency of Supranutritional Selenomethionine vs. Selenium‐yeast on markers of prostatic homeostasis and cancer risk reduction. Proceedings of the 10th International Symposium on Selenium in Biology and Medicine. Berlin: Deutsche Forschungsgemeinschaft, 2013:62.
Weekley 2013
- Weekley CM, Aitken JB, Finney L, Vogt S, Witting PK, Harris HH. Selenium metabolism in cancer cells: the combined application of XAS and XFM techniques to the problem of selenium speciation in biological systems. Nutrients 2013;5(5):1734‐56. [DOI] [PMC free article] [PubMed] [Google Scholar]
Wells 2004
- Wells GA, Shea B, O´Connell D, Peterson J, Welch V, Losos M, et al. The Newcastle‐Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta‐analyses. http://www.ohri.ca/programs/clinical_epidemiology/oxford.htm. Ottawa, (accessed 1 April 2004).
WHO 2004
- Joint FAO/WHO Expert Consultation on Human Vitamin, Mineral Requirements. Vitamin and mineral requirements in human nutrition: report of a joint FAO/WHO expert consultation; Bangkok, Thailand; 21‐30 September 1998. http://whqlibdoc.who.int/publications/2004/9241546123.pdf (accessed 9 March 2011). World Health Organization (WHO), 2004.
WHO 2017
- WHO. Cancer fact sheet. http://www.who.int/mediacentre/factsheets/fs297/en/ (accessed 6 November 2017). e.
Wichman 2016
- Wichman J, Winther KH, Bonnema SJ, Hegedus L. Selenium supplementation significantly reduces thyroid autoantibody levels in patients with chronic autoimmune thyroiditis: a systematic review and meta‐analysis. Thyroid 2016;26(12):1681‐92. [PUBMED: 27702392] [DOI] [PubMed] [Google Scholar]
Wood 2008
- Wood L, Egger M, Gluud LL, Schulz KF, Juni P, Altman DG, et al. Empirical evidence of bias in treatment effect estimates in controlled trials with different interventions and outcomes: meta‐epidemiological study. British Medical Journal 2008;336(7644):601‐5. [DOI] [PMC free article] [PubMed] [Google Scholar]
Zeng 2005
- Zeng H, Uthus EO, Combs GF Jr. Mechanistic aspects of the interaction between selenium and arsenic. Journal of Inorganic Biochemistry 2005;99(6):1269‐74. [DOI] [PubMed] [Google Scholar]
Zhou 2013
- Zhou J, Huang K, Lei XG. Selenium and diabetes ‐ evidence from animal studies. Free Radical Biology & Medicine 2013;65:1548‐56. [PUBMED: 23867154] [DOI] [PMC free article] [PubMed] [Google Scholar]
Zwolak 2012
- Zwolak I, Zaporowska H. Selenium interactions and toxicity: a review. Cell Biology and Toxicology 2012;28(1):31‐46. [DOI] [PubMed] [Google Scholar]
References to other published versions of this review
Dennert 2005
- Dennert G, Zwahlen M, Brinkman M, Vinceti M, Zeegers MPA, Horneber M. Selenium for preventing cancer. Cochrane Database of Systematic Reviews 2005, Issue 2. [DOI: 10.1002/14651858.CD005195] [DOI] [PMC free article] [PubMed] [Google Scholar]
Dennert 2011
- Dennert G, Zwahlen M, Brinkman M, Vinceti M, Zeegers MP, Horneber M. Selenium for preventing cancer. Cochrane Database of Systematic Reviews 2011, Issue 5. [DOI: 10.1002/14651858.CD005195] [DOI] [PMC free article] [PubMed] [Google Scholar]
Vinceti 2014
- Vinceti M, Dennert G, Crespi CM, Zwahlen M, Brinkman M, Zeegers MP, et al. Selenium for preventing cancer. Cochrane Database of Systematic Reviews 2014, Issue 3. [DOI: 10.1002/14651858.CD005195.pub3; PUBMED: 24683040] [DOI] [PMC free article] [PubMed] [Google Scholar]