Drugs for preventing lung cancer in healthy people

Marcela Cortés‐Jofré; José‐Ramón Rueda; Claudia Asenjo‐Lobos; Eva Madrid; Xavier Bonfill Cosp

doi:10.1002/14651858.CD002141.pub3

. 2020 Mar 4;2020(3):CD002141. doi: 10.1002/14651858.CD002141.pub3

Drugs for preventing lung cancer in healthy people

Marcela Cortés‐Jofré ^1,^2,^✉, José‐Ramón Rueda ³, Claudia Asenjo‐Lobos ^2,⁴, Eva Madrid ⁵, Xavier Bonfill Cosp ⁶

Editor: Cochrane Lung Cancer Group

PMCID: PMC7059884 PMID: 32130738

Abstract

Background

This is the second update of this Cochrane Review. Some studies have suggested a protective effect of antioxidant nutrients and higher dietary levels of fruits and vegetables on lung cancer.

Objectives

To determine whether vitamins and minerals and other potential agents, alone or in combination, reduce lung cancer incidence and lung cancer mortality in healthy populations.

Search methods

We searched CENTRAL, MEDLINE and Embase from 1974 to May 2019 and screened references included in published studies and reviews.

Selection criteria

We included randomised controlled trials (RCTs) comparing vitamins or mineral supplements with placebo, administered to healthy people with the aim of preventing lung cancer.

Data collection and analysis

Four review authors independently selected the trials to be included in the review, assessed their methodological quality and extracted data. For dichotomous outcomes we calculated risk ratios (RRs) and 95% confidence intervals (CIs) and pooled results using the random‐effects model. We assessed the risk of bias using Cochrane's 'Risk of bias' assessment tool and certainty of evidence using the GRADE approach.

Main results

In this update, we identified three new trials for a total of 12 studies. Six analysed vitamin A, three vitamin C, three combined vitamin D3 + calcium, four vitamin E combined with other products, one selenium supplements and nine studied combinations of two or more products. Four studies included only men and five only women.

Vitamin A results in little to no difference in lung cancer incidence (RR 1.09, 95% CI 1.00 to 1.19; 5 RCTs, 212314 participants; high‐certainty evidence) and lung cancer mortality (RR 1.06, 95% CI 0.81 to 1.38; 3 RCTs, 190118 participants; high‐certainty evidence). But in smokers or asbestos workers vitamin A increases the risk of lung cancer incidence (RR 1.10, 95% CI 1.01 to 1.20; 3 RCTs, 43995 participants; high‐certainty evidence), lung cancer mortality (RR 1.18, 95% CI 1.01 to 1.38; 2 RCTs, 29426 participants; high‐certainty evidence) and all‐cause mortality (RR 1.09, 95% CI 1.05 to 1.13; 2 RCTs, 32883 participants; high‐certainty evidence). Vitamin A increases the risk of minor side effects, such as yellowing of the skin and minor gastrointestinal symptoms (high‐certainty evidence).

Vitamin C likely results in little to no difference in lung cancer incidence (RR 1.29, 95% CI 0.67 to 2.49; 2 RCTs, 14953 participants; moderate‐certainty evidence). In women, vitamin C increases the risk of lung cancer incidence (RR 1.84, 95% CI 1.14 to 2.95; 1 RCT, 7627 participants; high‐certainty evidence). In men, vitamin C results in little to no difference in mortality for lung cancer (RR 0.81, 95% CI 0.53 to 1.23; 1 RCT, 7326 participants; high‐certainty evidence).

Vitamin D + calcium may result in little to no difference in lung cancer incidence in postmenopausal women (RR 0.90, 95% CI 0.39 to 2.08; 3 RCTs, 37601 women; low‐certainty evidence).

Vitamin E results in little to no difference in lung cancer incidence (RR 1.01, 95% CI 0.90 to 1.14; 3 RCTs, 36841 participants; high‐certainty evidence) or to lung cancer mortality (RR 0.96, 95% CI 0.77 to 1.18; 2 RCTs, 29214 participants; high‐certainty evidence), but increases the risk of haemorrhagic strokes (hazard ratio (HR), 1.74, 95% CI 1.04 to 2.91; 1 RCT, 14641 participants; high‐certainty evidence).

Calcium results in little to no difference in lung cancer incidence in postmenopausal women (RR 0.65, 95% CI 0.13 to 3.18; 1 RCT, 733 participants) or in risk of renal calculi (RR 1.94, 95% CI 0.20 to 18.57; 1 RCT, 733 participants; low‐certainty evidence).

Selenium in men results in little to no difference in lung cancer incidence (RR 1.11, 95% CI 0.80 to 1.54; 1 RCT, 17448 participants; high‐certainty evidence) and lung cancer mortality (RR 1.09, 95% CI 0.72 to 1.66; 1 RCT, 17448 participants; high‐certainty evidence) and increases the risk for grade 1 to 2 dermatitis (RR 1.16, 95% CI 1.04 to 1.31; 1 RCT, 17448 participants; high‐certainty evidence) and for alopecia (RR 1.28, 95% CI 1.07 to 1.53; 1 RCT, 17448 participants; high‐certainty evidence).

The combination of vitamins A, C, E + selenium + zinc results in little to no difference in lung cancer incidence (RR 0.64, 95% CI 0.28 to 1.48; 1 RCT, 12741 participants; high‐certainty evidence).

Authors' conclusions

Well‐designed RCTs have shown no beneficial effect of supplements for the prevention of lung cancer and lung cancer mortality in healthy people. Vitamin A supplements increase lung cancer incidence and mortality in smokers or persons exposed to asbestos. Vitamin C increases lung cancer incidence in women. Vitamin E increases the risk of haemorrhagic strokes.

Plain language summary

Drugs for preventing lung cancer in healthy people

Review question

We reviewed the evidence assessing the relationship between vitamins or antioxidant intake and lung cancer prevention, mortality and adverse events for that cancer. This review updates our Cochrane Review on this topic published in 2012.

Background

Lung cancer is among the leading causes of cancer death throughout the world, and its prevention has become a public health priority. It has been suggested that vitamin supplements and some antioxidants may prevent lung cancer.

Study characteristics

This review includes 12 studies in which healthy adults were randomly assigned to receive vitamin supplements or placebo (a substance that has no physical effects) and were followed over time to evaluate their risk of developing lung cancer. The evidence is current to May 2019.

Key results

None of the treatments compared with placebo have shown a difference in the risk for lung cancer incidence or lung cancer mortality in healthy people. In smokers and people exposed to asbestos, vitamin A increases lung cancer incidence, lung cancer mortality and all‐cause mortality. Vitamin C increases lung cancer incidence in women. Vitamin E increases the risk of haemorrhagic strokes.

The certainty of the evidence is high for the following comparisons against placebo: vitamin A; vitamin E; selenium; and combinations of vitamins A, C, E, selenium and zinc.

Summary of findings

Background

This review is an update of a previously published Cochrane Review (Cortes‐Jofre 2012).

Description of the condition

Lung cancer is one of the most prevalent and lethal cancers in the world. It is classified into two main subtypes, small cell lung cancer and non‐small cell lung cancer, the latter representing approximately 85% of all cases (Rahal 2017).

Lung cancer is the leading cause of death from male cancer and the second‐leading cause of death in women around the world (Marshall 2013). In 2015, the global incidence was 2 million cases (1.3 for men and 0.6 for women), and overall mortality in the same year recorded 1.7 million deaths (1.2 in men and 0.5 in women) (Global 2017).

Geographic variation in incidence often reflects the national distribution of poverty and access to medical care, reflecting large historical and continuing differences in the prevalence of smoking (Siegel 2017).

Faced with this scenario, and considering the low average survival rate at five years after diagnosis for people diagnosed at a late stage (16.3%) (Marshall 2013), prevention emerges as an important strategy, focusing on known risk factors, such as smoking and exposure to environmental carcinogens: asbestos, arsenic, radon and polycyclic aromatic hydrocarbons, that predispose individuals to the development of lung cancer (Raaschou‐Nielsen 2013).

Advances in cell and molecular biology have increased understanding of the multiple events that lead to the development of lung cancer. The development of new technologies, such as genomic profiling and genome‐wide association studies has been helpful in the detection of new genetic variants, likely involved in lung cancer risk (Marshall 2013; Qu 2016; Sakoda 2011).

Description of the intervention

Cancer chemoprevention is the inhibition or reversal of carcinogenesis by intervention with pharmacologically active agents (Penny 2015) Several micronutrients have attracted the attention of the scientific community as potential cancer‐preventive agents and among them diet‐derived antioxidants have been studied intensively because of the protection they convey against oxidative stress (Benetou 2015).

Thus dietary supplements are commonly used to prevent chronic diseases, mainly cardiovascular disease and cancer and their use has increased over time, notably in the USA. A recent survey found that 77% of the USA adult population uses dietary supplements, the most predominant being multi vitamin/multi mineral supplements (CRN 2019). There are significant differences in dietary supplement intake within Europe with more prevalence in Northern countries than in Southern countries (Schwingshackl 2015), with for example, 2% of Greek men compared to 66% of Danish women taking dietary supplements.

How the intervention might work

Free radicals can lead to the development of cancer and cardiovascular disease by lipid peroxidation and DNA damage. Reactive oxygen species are ions or small molecules containing oxygen and an unpaired electron, and this free electron confers high reactivity to oxygen. Redox imbalance is induced by disequilibrium between the production and suppression of reactive oxygen species. Excess of oxidative damage can be controlled by exogenous antioxidants such as vitamins C and E, polyphenols, carotenes, flavonoids, omega‐3, and N‐acetylcysteine. These exogenous antioxidants deactivate excited oxygen molecules and organic free radicals, and in this way decrease the oxidative damage through distinct mechanisms of action (Hamishehkar 2016; Prevatto 2017).

It has been recognised that diet and nutrients play an important role in the development and progression of cancer, and many components of the diet are associated with the risk of cancer. With respect to lung cancer, it has been found that a high intake of vegetables, fruits, fish and soy may reduce the risk, while red meat and processed meat can increase it (Yang 2012). However, almost all trials of clinical intervention with isolated nutrients, such as supplements of vitamin A, vitamin E, vitamin C, folic acid, selenium and carotenoids, have not been able to demonstrate their protective effects against lung cancer (Wang 2015). Also high‐dose beta‐carotene supplementation appears to increase the risk of lung cancer among current smokers (Tanvetyanon 2008).

Since different foods are consumed in combinations, and interact with each other in a complex way, an integral diet analysis may better reflect dietary habits and provides a constructive tool for evaluating the overall effects of the total diet on human health (Cho 2006; Gnagnarella 2013).

No consistent evidence has been found that vitamin supplements affect cancer or mortality from all causes in healthy individuals without known nutritional deficiencies because the "multivitamin" preparations contain unbalanced formulations. The beta‐carotene (presumably all‐trans) used in studies has never been shown to be a safe supplement (Dror 2014; Fortmann 2014).

A systematic review that includes an uncontrolled trial in participants with early‐stage non‐small cell lung cancer suggested that beta‐carotene could modulate the expression of biomarkers such as cyclins; two trials did not show a significant effect on atypia and sputum bronchial cell metaplasia/dysplasia, also associated with lung cancer. In addition, an observational study found no correlation between serum retinol and oxo‐dGuo leukocytes, a marker of DNA damage. In conclusion, the beta‐carotene retinoid may be promising for use among a subset of people, and deserves further study (Fritz 2011).

On the other hand, combined healthy foods could also reduce the risk of lung cancer, through antioxidants, polyphenols, fibre and minerals that they contain; in addition, their interrelations could synergistically improve their individual protective effects as a whole. Recent studies have shown that the dietary pattern could affect the structure and metabolome of the human intestinal microbiome and may contribute to the health or pathogenesis of disorders such as coronary vascular disease and inflammatory bowel disease (Albenberg 2014). However, the general effects of different dietary patterns on human health are not clear and still require more research (Sun 2016).

Why it is important to do this review

Most people believe that even if vitamins are not effective, at least they are safe. But in spite of a variety of research studies into the cellular biologic function of vitamins and interesting messages about their roles on health, the long‐term health consequences of vitamin consumption are unknown (Hamishehkar 2016).

Given the continuing cancer burden, the relatively low impact of proven cancer treatment strategies in reducing lung cancer mortality, and the possibility that food‐based or other components may have chemopreventive properties, it is essential to evaluate the use of these agents. Our aim was to review the evidence for the effectiveness of chemoprevention in lung cancer in healthy people.

Objectives

To determine whether vitamins and minerals and other potential agents, alone or in combination, reduce lung cancer incidence and lung cancer mortality in healthy populations.

Methods

Criteria for considering studies for this review

Types of studies

We included only randomised controlled trials (RCTs) comparing any eligible intervention with placebo.

Types of participants

Healthy men and women of all ages, independent of their smoking status or other risk factors for lung cancer.  Smokers and those exposed to asbestos are considered as people at high risk; people not known to have been exposed to such risk factors are considered as people at low risk.

Types of interventions

Dietary supplementation with specific vitamins, minerals (selenium, zinc or others) and other potential agents, natural or synthetic, such as retinoids, isothiocyanates, flavonoids, monoterpenes, or pharmaceuticals such as N‐acetyl cysteine, alone or in combination, at any doses. Administration could be in capsule or tablet form, to be consumed orally.

In this review we assessed the following comparisons.

Vitamin A (beta‐carotene or retinol) versus placebo
Vitamin C (ascorbic acid) versus placebo
Vitamin D plus calcium versus placebo
Vitamin D plus calcium versus calcium alone
Vitamin E (alpha‐tocopherol) versus placebo
Calcium versus placebo
Selenium versus placebo
Vitamin A plus vitamin E versus placebo
Vitamin C plus vitamin E versus placebo
Vitamin E plus selenium versus placebo
Vitamins A and E plus selenium versus placebo
Vitamins A, C, E plus selenium plus zinc versus placebo

Types of outcome measures

We considered the following primary outcomes in this review.

Lung cancer incidence
Lung cancer mortality
Adverse events

Since the role of the drugs included in this review could also have an impact on other cancers or diseases, we also considered the following secondary outcomes.

Total cancer incidence
Total cancer mortality
Total mortality

Search methods for identification of studies

Electronic searches

For this update, Cochrane Lung Cancer Information Specialists redesigned our search strategies for the following three main databases.

CENTRAL (Appendix 1)
MEDLINE (Appendix 2)
Embase (Appendix 3)

The search string for MEDLINE was developed according to the Cochrane Highly Sensitive Search Strategy, sensitivity maximising version (2008 version) as referenced in Chapter 6.4.11.1 and detailed in box 6.4.b of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). We searched the three databases from 1974 to 2 May 2019.

We also searched the following clinical trials registries for possible unpublished or ongoing trials:

ClinicalTrials.gov.
WHO International Clinical Trials Registry Platform (ICTRP).

Searching other resources

We also searched and screened published meta‐analyses and recent reviews addressing the topic of our review for RCTs from 1983 to May 2019.

Data collection and analysis

Selection of studies

In this update, four review authors (MC‐J, JRR, CA and EM) independently evaluated the titles and abstracts obtained from the electronic search. They examined the full‐text of the provisionally included studies to determine if the study met the inclusion criteria. We resolved disagreements by discussion and consensus.

All of the included studies presented their results in several articles, and in some cases postintervention follow‐up data are also available. For all studies we used the most recently published data for each relevant outcome variable.

Data extraction and management

As in the first published version of this review, in this update we used a standardised form designed for the purposes of this review to collect data. The extracted data included details of the methods of randomisation, comparisons of interest, the number and type of people originally randomised in each arm of the study, the losses to follow‐up and the results of interest of each arm of the study.

In this update of the review, three review authors (MC‐J, CA and JRR) extracted data from the new included studies, and the most relevant information about the study is presented in the Characteristics of included studies section. The same three review authors also extracted and analysed the most recent postintervention follow‐up data of the trials already included in the first published versions of this review.

Assessment of risk of bias in included studies

Three review authors (MC‐J, CA and JRR) independently assessed the risk of bias for each study for the following domains:

sequence generation,
allocation concealment,
blinding of participants and personnel,
outcome assessment,
incomplete outcome data,
selective reporting.

We judged each potential source of bias as high, low or unclear and provided a quote from the study report together with a justification for our judgement in the 'Risk of bias' table. We judged the risk of bias for each domain according to the criteria defined in chapter 8 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). We resolved any disagreement by discussion and consensus.

Measures of treatment effect

For each study, we calculated risk ratios (RRs) and their 95% confidence intervals (CIs) for dichotomous outcomes. Where appropriate, we pooled results of comparable groups of trials, using the random‐effects model.

Unit of analysis issues

The unit of analysis was the participant. We did not anticipate the possibility of cross‐over trials or cluster trials for this topic.

Dealing with missing data

In case of missing or incomplete data, we tried to obtain it by writing to the interested authors. Authors who did not respond were sent a second letter and all but one provided the required additional information and data on their studies.

Assessment of heterogeneity

We tested heterogeneity between trials with the I² statistic. We interpreted the I² value according to the following thresholds (Higgins 2011): 0% to 40% heterogeneity might not be important; 30% to 60% may represent moderate heterogeneity; 50% to 90% may represent substantial heterogeneity; 75% to 100%: considerable heterogeneity. We investigated substantial heterogeneity (I² > 50%) by prespecified subgroup analysis.

Assessment of reporting biases

We planned to generate funnel plots and to perform Egger's linear regression tests to investigate reporting biases for considered outcomes when the number of trials included in a single meta‐analysis is sufficient (at least 10 trials). We followed recommendations provided in Chapter 10 of the Cochrane Handbook for Systematic Reviews of Interventions (Sterne 2011).

Data synthesis

When sufficient clinically similar studies were available, we grouped their results into meta‐analyses and performed meta‐analyses based on an intention‐to‐treat (ITT) analysis when they were available.

We performed meta‐analyses according to the recommendations contained in Chapter 9 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). For the meta‐analyses, a review author (JR) entered data into Review Manager 5 (Review Manager 2014), and a second review author (MC‐J) reviewed the data to verify its accuracy.

'Summary of findings' tables

We followed the GRADE approach by creating 'Summary of findings' tables, as suggested in Chapters 11 and 12 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). For each outcome we rated the certainty of the evidence according to one of the following categories: 'high', moderate', 'low', or 'very low', after assessing and taking into account: risk of bias of the studies, inconsistency of results, indirectness, imprecision and publication bias.

For this update, we added the following seven 'Summary of findings' tables.

Vitamin A (beta‐carotene or retinol) compared to placebo (Table 1).
Vitamin C compared to placebo (Table 2).
Vitamin D plus calcium compared to placebo (Table 3).
Vitamin E (alpha‐tocopherol) compared to placebo (Table 4).
Calcium compared to placebo (Table 5).
Selenium compared to placebo (Table 6).
Vitamins A, C, E + selenium + zinc compared to placebo (Table 7).

Summary of findings for the main comparison. Vitamin A compared to placebo for preventing lung cancer in healthy people.

Vitamin A compared to placebo for preventing lung cancer in healthy people
Patient or population: healthy people and healthy male physicians  Setting: outpatients  Intervention: vitamin A  Comparison: placebo
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	№ of participants  (studies)	Certainty of the evidence  (GRADE)	Comments
	Risk with placebo	Risk with vitamin A
Incidence: lung cancer	Study population		RR 1.09  (1.00 to 1.19)	212314  (5 RCTs)	⊕⊕⊕⊕  High	Subgroup analysis shows increase in risk of lung cancer incidence in smokers and asbestos workers taking vitamin A
	9 per 1000	9 per 1000 (9 to 10)
Mortality: lung cancer	Study population		RR 1.06 (0.81 to 1.38)	190118 (3 RCTs)	⊕⊕⊕⊕  High	Subgroup analysis shows increase in risk of lung cancer mortality in smokers and asbestos workers taking vitamin A
	3 per 1000	4 per 1000 (3 to 5)
Adverse events: yellowing of the skin	Study population		RR 1.14 (1.07 to 1.21)	22071 (1 RCT)	⊕⊕⊕⊕  High
	139 per 1000	159 per 1000  (149 to 168)
Adverse events: minor gastrointestinal symptoms	Study population		RR 2.22 (1.80 to 2.74)	22071 (1 RCT)	⊕⊕⊕⊕  High
	11 per 1000	25 per 1000  (20 to 31)
Incidence: all cancers	Study population		RR 1.02 (0.97 to 1.07)	44267 (3 RCTs)	⊕⊕⊕⊕  High
	132 per 1000	135 per 1000  (128 to 142)
Mortality: all cancers	Study population		RR 1.02 (0.88 to 1.77)	22071 (1 RCT)	⊕⊕⊕⊕  High
	34 per 1000	35 per 1000  (30 to 61)
Mortality: all causes ‐ smokers and asbestos workers	Study population		RR 1.09 (1.05 to 1.13)	32883 (2 RCTs)	⊕⊕⊕⊕  High
	226 per 1000	246 per 1000  (237 to 255)
*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; RCT: randomised controlled trial; RR: risk 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.

Vitamin C compared to placebo for preventing lung cancer in healthy people
Patient or population: healthy male physicians, period: 1997 to 2007 and female health professionals, period: 1998 to 2005  Setting: outpatients  Intervention: vitamin C  Comparison: placebo
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	№ of participants  (studies)	Certainty of the evidence  (GRADE)	Comments
	Risk with placebo	Risk with vitamin C
Incidence: lung cancer	Study population		RR 1.29  (0.67 to 2.49)	14953  (2 RCTs)	⊕⊕⊕⊝  Moderate^a
	11 per 1000	14 per 1000  (7 to 26)
Incidence: lung cancer ‐ men	Study population		RR 0.94  (0.64 to 1.38)	7326  (1 RCT)	⊕⊕⊕⊕  High
	15 per 1000	14 per 1000  (9 to 20)
Incidence: lung cancer ‐ women	Study population		RR 1.84  (1.14 to 2.95)	7627  (1 RCT)	⊕⊕⊕⊕  High
	7 per 1000	13 per 1000  (8 to 20)
Mortality: lung cancer ‐ men	Study population		RR 0.81  (0.53 to 1.23)	7326  (1 RCT)	⊕⊕⊕⊕  High
	13 per 1000	11 per 1000  (7 to 16)
Adverse events	Study population		Not available	1 RCT	⊕⊕⊕⊕  High	Numeric data not provided, but one study reported that there were no significant effects of either agent on minor bleeding or gastrointestinal tract symptoms, fatigue, drowsiness, skin discolouration, rashes, or migraine
	Not available	Not available
Incidence: all cancers	Study population		RR 1.03  (0.94 to 1.13)	14953 (2 RCTs)	⊕⊕⊕⊕  High
	170 per 1000	175 per 1000  (159 to 192)
Mortality: all cancers	Study population		RR 1.11  (0.93 to 1.34)	14953 (2 RCTs)	⊕⊕⊕⊕  High
	45 per 1000	49 per 1000  (41 to 60)
*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; RCT: randomised controlled trial; RR: risk 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.

Vitamin D plus calcium compared to placebo for preventing lung cancer in healthy people
Patient or population: healthy postmenopausal women  Setting: outpatients  Intervention: vitamin D plus calcium  Comparison: placebo
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	№ of participants  (studies)	Certainty of the evidence  (GRADE)	Comments
	Risk with placebo	Risk with vitamin D plus calcium
Incidence: lung cancer	Study population		RR 0.90 (0.39 to 2.08)	37061 (3 RCTs)	⊕⊕⊝⊝  Low ^{a, b}
	7 per 1000	6 per 1000 (3 to 15)
Mortality: lung cancer	Study population		Not available	Not available		The studies did not evaluate this outcome.
	Not available	Not available
Adverse events: renal calculi	Study population		RR 1.49 (0.70 to 3.17)	2931 (2 RCTs)	⊕⊕⊕⊝  Moderate ^a
	8 per 1000	12 per 1000 (6 to 25)
Adverse events: serum calcium above normal	Study population		RR 2.98 (0.60 to 14.74)	2197 (1 RCT)	⊕⊕⊝⊝  Low ^{a, b}
	2 per 1000	5 per 1000 (1 to 27)
Incidence: all cancers	Study population		RR 0.73 (0.48 to 1.11)	37061 (3 RCTs)	⊕⊕⊝⊝  Low ^{a, c}
	76 per 1000	55 per 1000 (36 to 84)
Mortality: all cancers	Study population		RR 0.91 (0.78 to 1.05)	34670 (1 RCT)	⊕⊕⊕⊝  Moderate ^a
	20 per 1000	18 per 1000 (16 to 21)
Mortality: all causes	Study population		Not available	Not available		The studies did not evaluate this outcome.
	Not available	Not available
*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; RCT: randomised controlled trial; RR: risk 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.

Vitamin E compared to placebo for preventing lung cancer in healthy people
Patient or population: healthy people  Setting: outpatients  Intervention: vitamin E  Comparison: placebo
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	№ of participants  (studies)	Certainty of the evidence  (GRADE)	Comments
	Risk with placebo	Risk with vitamin E
Incidence: lung cancer	Study population		RR 1.01  (0.90 to 1.14)	36841  (3 RCTs)	⊕⊕⊕⊕  High
	30 per 1000	30 per 1000  (27 to 34)
Mortality: lung cancer	Study population		RR 0.96  (0.77 to 1.18)	29214  (2 RCTs)	⊕⊕⊕⊕  High
	12 per 1000	12 per 1000  (9 to 14)
Adverse events: haemorrhagic stroke	Study population		HR 1.74  (1.04 to 2.91)	14641 (1 RCT)	⊕⊕⊕⊕  High	Reported no significant effects on minor bleeding or gastrointestinal tract symptoms.
	3 per 1000	5 per 1000  (3 to 9)
Incidence: all cancers	Study population		RR 0.99  (0.94 to 1.04)	36832 (3 RCTs)	⊕⊕⊕⊕  High
	142 per 1000	140 per 1000  (133 to 147)
Mortality: all cancers	Study population		RR 1.11  (0.99 to 1.24)	54517 (2 RCTs)	⊕⊕⊕⊕  High
	19 per 1000	21 per 1000  (19 to 24)
Mortality: all causes	Study population		RR 1.03  (0.99 to 1.07)	69090 (3 RCTs)	⊕⊕⊕⊕  High
	117 per 1000	120 per 1000  (116 to 125)
*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; HR: hazard ratio; RCT: randomised controlled trial; RR: risk 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.

Calcium compared to placebo for preventing lung cancer in healthy people
Patient or population: white healthy postmenopausal women over 55 years  Setting: outpatients  Intervention: calcium  Comparison: placebo
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	№ of participants  (studies)	Certainty of the evidence  (GRADE)	Comments
	Risk with placebo	Risk with calcium
Incidence: lung cancer ‐ women	Study population		RR 0.65  (0.13 to 3.18)	733  (1 RCT)	⊕⊕⊝⊝  Low ^{a, b}
	10 per 1000	7 per 1000  (1 to 33)
Mortality: lung cancer	Study population		Not available	Not available		The study did not evaluate this outcome.
	Not available	Not available
Adverse events: renal calculi	Study population		RR 1.94  (0.20 to 18.57)	733 (1 RCT)	⊕⊕⊝⊝  Low ^{a, b}
	3 per 1000	7 per 1000  (1 to 64)
Incidence: all cancers	Study population		RR 0.55 (0.29 to 1.03)	733 (1 RCT)	⊕⊕⊝⊝  Low ^{a, b}
	69 per 1000	38 per 1000  (20 to 72)
Mortality: all cancers	Study population		Not available	Not available		The study did not evaluate this outcome.
	Not available	Not available
Mortality: all causes	Study population		Not available	Not available		The study did not evaluate this outcome.
	Not available	Not available
*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; RCT: randomised controlled trial; RR: risk 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.

Selenium compared to placebo for preventing lung cancer in healthy people
Patient or population: healthy men  Setting: outpatients  Intervention: selenium  Comparison: placebo
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	№ of participants  (studies)	Certainty of the evidence  (GRADE)	Comments
	Risk with placebo	Risk with selenium
Incidence: lung cancer	Study population		RR 1.11  (0.80 to 1.54)	17448  (1 RCT)	⊕⊕⊕⊕  High
	8 per 1000	9 per 1000  (6 to 12)
Mortality: lung cancer	Study population		RR1.09  (0.72 to 1.66)	17448  (1 RCT)	⊕⊕⊕⊕  High
	5 per 1000	5 per 1000  (3 to 8)
Adverse events: alopecia	Study population		RR 1.28  (1.07 to 1.53)	17448  (1 RCT)	⊕⊕⊕⊕  High
	24 per 1000	30 per 1000  (25 to 36)
adverse events: grade 1 to 2 dermatitis	Study population		RR 1.16  (1.04 to 1.31)	17448  (1 RCT)	⊕⊕⊕⊕  High
	59 per 1000	69 per 1000 (62 to 78)
Incidence: all cancers	Study population		RR 1.01  (0.92 to 1.11)	17448  (1 RCT)	⊕⊕⊕⊕  High
	95 per 1000	96 per 1000  (87 to 105)
Mortality: all cancers	Study population		RR 1.02  (0.80 to 1.30)	17448  (1 RCT)	⊕⊕⊕⊕  High
	14 per 1000	15 per 1000  (11 to 19)
Mortality: all causes	Study population		RR 0.98  (0.86 to 1.13)	17448  (1 RCT)	⊕⊕⊕⊕  High
	44 per 1000	43 per 1000  (38 to 50)
*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; RCT: randomised controlled trial; RR: risk 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.

Vitamins A, C, E + selenium + zinc compared to placebo for preventing lung cancer in healthy people
Patient or population: healthy women, aged 35 to 60, and men, aged 45 to 60, living in France  Setting: outpatients  Intervention: vitamins A, C, E + selenium + zinc  Comparison: placebo
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	№ of participants  (studies)	Certainty of the evidence  (GRADE)	Comments
	Risk with placebo	Risk with vitamins A, C, E + selenium + zinc
Incidence: lung cancer	Study population		RR 0.64  (0.28 to 1.48)	12741  (1 RCT)	⊕⊕⊕⊕  High
	2 per 1000	1 per 1000  (1 to 3)
Mortality: lung cancer	Study population		Not available	Not available		The study did not evaluate this outcome.
	Not available	Not available
Adverse events	Study population		Not available	Not available		No adverse data published
	Not available	Not available
Incidence: all cancers	Study population		RR 0.96  (0.83 to 1.10)	12741  (1 RCT)	⊕⊕⊕⊕  High
	80 per 1000	77 per 1000  (67 to 88)
Mortality: all cancers	Study population		Not available	Not available		The study did not evaluate this outcome.
	Not available	Not available
Mortality: all causes	Study population		RR 0.88  (0.70 to 1.11)	12741  (1 RCT)	⊕⊕⊕⊕  High
	28 per 1000	25 per 1000  (20 to 31)
*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; RCT: randomised controlled trial; RR: risk 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.

Study	Vitamin A (beta‐carotene or retinol)	Vitamin C (ascorbic acid)	Vitamin D + calcium	Vitamin E (alpha‐tocopherol)	Selenium	Other combinations of two or more products
ATBC 1994	20 mg daily			50 mg daily		Alpha‐tocopherol (50 mg) + beta‐carotene (20 mg), daily
Brunner 2011 (WHI)			400 IU of vitamin D3 and 1000 mg of calcium
Gaziano 2009 (PHS II)		500 mg daily		400 IU every other day		Vitamin E 400 IU every other day + vitamin C 500 mg daily
Hennekens 1996 (PHS)	50 mg every other day
Hercberg 2010 (SU.VI.MAX)						Combination of antioxidants (120 mg of ascorbic acid, 30 mg of vitamin E, 6 mg of beta‐carotene, 100 μg of selenium (as selenium‐enriched yeast), and 20 mg of zinc (as gluconate) in a single daily capsule
Kamangar 2006 (LINXIAN)	Retinol (as palmitate 5000 IU) and zinc (as zinc oxide 22.5 mg) daily	120 mg and molybdenum (yeast complex 30 ug) daily				Beta‐carotene (15 mg), vitamin E (alpha‐tocopherol 30 mg) and selenium (as selenium yeast 50 ug) daily Riboflavin B2 (3.2 mg) and niacin B3 (40 mg) daily
Lappe 2007			1100 IU of vitamin D3 plus 1400 to 1500 mg of calcium 1400 to 1500 mg of calcium
Lappe 2017			2000 IU of vitamin D3 and 1500 mg of calcium
Lee 2005 (WHS)	50 mg every other day			600 IU every other day
Lin 2009 (WACS)	50 mg every other day	500 mg daily		600 IU every other day
Lippman 2009 (SELECT)				400 IU daily	200 μg daily	Selenium 200 μg + vitamin E 400 IU daily
Omenn 1996 (CARET)	30 mg + 25000 IU retinol daily

Study	Length of treatment (years)	Follow‐up (years)
ATBC 1994	5 to 8	16
Brunner 2011 (WHI)	7	7
Gaziano 2009 (PHS II)	6	8
Hennekens 1996 (PHS)	12	12
Hercberg 2010 (SU.VI.MAX)	8	12
Kamangar 2006 (LINXIAN)	5	15
Lappe 2007	4	4
Lappe 2017	4	4
Lee 2005 (WHS)	2	6
Lin 2009 (WACS)	9	9
Lippman 2009 (SELECT)	7	7
Omenn 1996 (CARET)	4	12

Date	Event	Description
2 May 2019	New search has been performed	Search strategies fully redesigned. Full search ran and 10,428 references screened (Figure 1). Three new studies included (Brunner 2011; Lappe 2007 ; Lappe 2017), for a total of 12 included studies
18 May 2018	New citation required but conclusions have not changed	Background, methods and results sections updated to be compliant with the current MECIR standards. A new primary outcome included: adverse events. 3 new studies integrated in the analysis, 7 'Summary of findings' tables added, GRADE approach applied. Conclusions not changed. Two new review authors.

No.	Query
#41	#6 AND #30 AND #39 AND #40
#40	'crossover procedure'/exp OR 'double‐blind procedure'/exp OR 'randomized controlled trial' /exp OR 'single‐blind procedure'/exp OR random* OR factorial* OR crossover* OR cross NEXT/1 over* OR placebo* OR doubl* NEAR/1 blind* OR singl* NEAR/1 blind* OR assign* OR allocat* OR volunteer*
#39	#31 OR #32 OR #33 OR #34 OR #35 OR #36 OR #37 OR #38
#38	risk:ti,ab AND lower:ti,ab
#37	risk:ti,ab AND decreas:ti,ab
#36	risk:ti,ab AND reduc:ti,ab
#35	risk:ti,ab AND modif:ti,ab
#34	'pre‐exposure prophylaxis'/exp OR 'prep':ti,ab OR *'prophyla'**:ti,ab
#33	'chemoprophylaxis'/exp OR *'chemoprevent'**:ti,ab
#32	'primary prevention'/exp OR *'prevent'**:ti,ab
#31	'prevention and control'/exp
#30	#7 OR #8 OR #9 OR #10 OR #11 OR #12 OR #13 OR #14 OR #15 OR #16 OR #17 OR #18 OR #19 OR #20 OR #21 OR #22 OR #23 OR #24 OR #25 OR #26 OR #27 OR #28 OR #29
#29	*'multivitamin'**:ti,ab
#28	'antineoplastic agent'/exp OR *'anticarcinogen':ti,ab OR 'anticancer':ti,ab OR 'antitumo'**:ti,ab
#27	'protective agent'/exp OR *'protect'**:ti,ab
#26	'zinc'/exp OR 'zinc':ti,ab
#25	'selenium'/exp OR 'selenium':ti,ab
#24	'vitamin e':ti,ab
#23	'ascorbic acid'/exp OR 'ascorbic acid':ti,ab OR 'vitamin c':ti,ab
#22	'terpene'/exp OR *'monoterpene'**:ti,ab
#21	'flavonoid'/exp OR *'flavonoid'**:ti,ab
#20	'isothiocyanic acid derivative'/exp OR *'isothiocyanate'**:ti,ab
#19	'acetylcysteine'/exp OR 'n acetyl cysteine':ti,ab
#18	'retinoid'/exp OR *'retinoid'**:ti,ab
#17	'retinol'/exp OR 'vitamin a':ti,ab OR 'retinol':ti,ab
#16	'alpha tocopherol'/exp OR *'tocopherol'**:ti,ab
#15	'beta carotene'/exp OR *'beta carotene'**:ti,ab
#14	'medicinal plant'/exp OR 'herb'/exp OR *'medicinal plant':ti,ab OR 'herb'*:ti,ab
#13	'mineral'/exp OR 'minerals':ti,ab
#12	'diet supplementation'/exp OR *'supplement'**:ti,ab
#11	'diet'/exp OR *'diet'**:ti,ab
#10	'glutathione'/exp OR 'glutathione':ti,ab
#9	'vitamin'/exp OR *'vitamin'**:ti,ab
#8	'antioxidant'/exp OR *'antioxid'**:ti,ab
#7	'carotenoid'/exp OR *'carotenoid'**:ti,ab
#6	#1 OR #2 OR #3 OR #4 OR #5
#5	'pleura tumor'/exp OR 'mpm':ti,ab OR *'pleural cancer':ti,ab OR 'pleural malignan':ti,ab OR 'pleural mesothelioma':ti,ab OR 'pleural neoplasm':ti,ab OR 'pleural tumo'**:ti,ab
#4	'small cell lung cancer'/exp OR *'oat cell carcinoma':ti,ab OR 'oat cell lung cancer':ti,ab OR 'sclc':ti,ab OR 'small cell lung cancer':ti,ab OR 'small cell lung carcinoma'*:ti,ab
#3	'non small cell lung cancer'/exp OR *'nonsmall cell lung cancer'*:ti,ab OR 'non small cell lung cancer ':ti,ab OR *'nonsmall cell lung carcinoma':ti,ab OR 'non small cell lung carcinoma':ti,ab OR 'nsclc'*:ti,ab
#2	*'bronchopulmonary carcino':ti,ab OR 'cancer' NEAR/3 'lung'* OR 'lung adenocarcimoma* ':ti,ab OR'lung cancer':ti,ab OR 'lung carcinoma':ti,ab OR 'lung malignan':ti,ab OR 'lung neoplasm ':ti,ab OR *'lung tumo':ti,ab OR 'pulmonary adenocarcinoma':ti,ab OR 'pulmonary cancer ':ti,ab OR *'pulmonary carcino':ti,ab OR 'pulmonary malignan':ti,ab OR 'pulmonary neoplasm ':ti,ab OR *'pulmonary tumo'**:ti,ab
#1	'lung tumor'/exp

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Incidence lung cancer	5	212314	Risk Ratio (M‐H, Random, 95% CI)	1.09 [1.00, 1.19]
1.1 High‐risk people (smokers and asbestos workers)	3	43995	Risk Ratio (M‐H, Random, 95% CI)	1.10 [1.01, 1.20]
1.2 Low‐risk people (non‐smokers or mixed population)	3	168319	Risk Ratio (M‐H, Random, 95% CI)	0.99 [0.69, 1.42]
2 Mortality lung cancer	3	190118	Risk Ratio (M‐H, Random, 95% CI)	1.06 [0.81, 1.38]
2.1 High‐risk people (smokers)	2	29426	Risk Ratio (M‐H, Random, 95% CI)	1.18 [1.01, 1.38]
2.2 Low‐risk people (non‐smokers or mixed population)	2	160692	Risk Ratio (M‐H, Random, 95% CI)	0.71 [0.35, 1.44]
3 Adverse events	1		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
3.1 Yellowing of the skin	1	22071	Risk Ratio (M‐H, Random, 95% CI)	1.14 [1.07, 1.21]
3.2 Minor gastrointestinal symptoms	1	22071	Risk Ratio (M‐H, Random, 95% CI)	2.22 [1.80, 2.74]
4 Incidence all cancers	3	44267	Risk Ratio (M‐H, Random, 95% CI)	1.02 [0.97, 1.07]
4.1 High‐risk people (smokers and asbestos workers)	1	14569	Risk Ratio (M‐H, Random, 95% CI)	1.05 [0.98, 1.12]
4.2 Low‐risk people (non‐smokers or mixed population)	1	7627	Risk Ratio (M‐H, Random, 95% CI)	1.00 [0.86, 1.16]
4.3 Global PHS study population	1	22071	Risk Ratio (M‐H, Random, 95% CI)	0.98 [0.92, 1.06]
5 Mortality all cancers	1	22071	Risk Ratio (M‐H, Random, 95% CI)	1.02 [0.88, 1.17]
6 Mortality all causes	2	32883	Risk Ratio (M‐H, Random, 95% CI)	1.09 [1.05, 1.13]
6.1 High‐risk people (smokers and asbestos workers)	2	32883	Risk Ratio (M‐H, Random, 95% CI)	1.09 [1.05, 1.13]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Incidence lung cancer	2	14953	Risk Ratio (M‐H, Random, 95% CI)	1.29 [0.67, 2.49]
1.1 Males	1	7326	Risk Ratio (M‐H, Random, 95% CI)	0.94 [0.64, 1.38]
1.2 Females	1	7627	Risk Ratio (M‐H, Random, 95% CI)	1.84 [1.14, 2.95]
2 Mortality lung cancer	1	7326	Risk Ratio (M‐H, Random, 95% CI)	0.81 [0.53, 1.23]
2.1 Males	1	7326	Risk Ratio (M‐H, Random, 95% CI)	0.81 [0.53, 1.23]
3 Incidence all cancers	2	14953	Risk Ratio (M‐H, Random, 95% CI)	1.03 [0.94, 1.13]
3.1 Males	1	7326	Risk Ratio (M‐H, Random, 95% CI)	1.00 [0.92, 1.08]
3.2 Females	1	7627	Risk Ratio (M‐H, Random, 95% CI)	1.11 [0.95, 1.29]
4 Mortality all cancers	2	14953	Risk Ratio (M‐H, Random, 95% CI)	1.11 [0.93, 1.34]
4.1 Males	1	7326	Risk Ratio (M‐H, Random, 95% CI)	1.05 [0.89, 1.23]
4.2 Females	1	7627	Risk Ratio (M‐H, Random, 95% CI)	1.28 [0.95, 1.72]
5 Mortality all causes	1	7326	Risk Ratio (M‐H, Random, 95% CI)	1.06 [0.97, 1.15]
5.1 Males	1	7326	Risk Ratio (M‐H, Random, 95% CI)	1.06 [0.97, 1.15]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Incidence lung cancer	3	37601	Risk Ratio (M‐H, Random, 95% CI)	0.90 [0.39, 2.08]
1.1 Females	3	37601	Risk Ratio (M‐H, Random, 95% CI)	0.90 [0.39, 2.08]
2 Adverse events	2		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
2.1 Renal calculi	2	2931	Risk Ratio (M‐H, Random, 95% CI)	1.49 [0.70, 3.17]
2.2 Serum calcium value above normal	1	2197	Risk Ratio (M‐H, Random, 95% CI)	2.98 [0.60, 14.74]
3 Incidence all cancers	3	37601	Risk Ratio (M‐H, Random, 95% CI)	0.73 [0.48, 1.11]
3.1 Females	3	37601	Risk Ratio (M‐H, Random, 95% CI)	0.73 [0.48, 1.11]
4 Mortality all cancers	1	34670	Risk Ratio (M‐H, Random, 95% CI)	0.91 [0.78, 1.05]
4.1 Females	1	34670	Risk Ratio (M‐H, Random, 95% CI)	0.91 [0.78, 1.05]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Incidence lung cancer	1	891	Risk Ratio (M‐H, Random, 95% CI)	0.33 [0.03, 3.19]
1.1 Females	1	891	Risk Ratio (M‐H, Random, 95% CI)	0.33 [0.03, 3.19]
2 Renal calculi	1	891	Risk Ratio (M‐H, Random, 95% CI)	0.33 [0.03, 3.19]
2.1 Females	1	891	Risk Ratio (M‐H, Random, 95% CI)	0.33 [0.03, 3.19]
3 Incidence all cancers	1	891	Risk Ratio (M‐H, Random, 95% CI)	0.76 [0.38, 1.55]
3.1 Females	1	891	Risk Ratio (M‐H, Random, 95% CI)	0.76 [0.38, 1.55]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Incidence lung cancer	3	36841	Risk Ratio (M‐H, Random, 95% CI)	1.01 [0.90, 1.14]
1.1 High‐risk people (smokers and asbestos workers)	1	14573	Risk Ratio (M‐H, Random, 95% CI)	1.01 [0.89, 1.15]
1.2 Males 50 years or older	1	14641	Risk Ratio (M‐H, Random, 95% CI)	0.87 [0.59, 1.29]
1.3 Women	1	7627	Risk Ratio (M‐H, Random, 95% CI)	1.26 [0.80, 1.98]
2 Mortality lung cancer	2	29214	Risk Ratio (M‐H, Random, 95% CI)	0.96 [0.77, 1.18]
2.1 High‐risk people (male smokers)	1	14573	Risk Ratio (M‐H, Random, 95% CI)	0.93 [0.73, 1.19]
2.2 Males 50 years or older	1	14641	Risk Ratio (M‐H, Random, 95% CI)	1.02 [0.67, 1.56]
3 Incidence all cancers	3	36832	Risk Ratio (M‐H, Random, 95% CI)	0.99 [0.94, 1.04]
3.1 High‐risk people (smokers and asbestos workers)	1	14564	Risk Ratio (M‐H, Random, 95% CI)	0.98 [0.91, 1.05]
3.2 Males 50 years or older	1	14641	Risk Ratio (M‐H, Random, 95% CI)	1.03 [0.95, 1.12]
3.3 Women	1	7627	Risk Ratio (M‐H, Random, 95% CI)	0.94 [0.81, 1.09]
4 Mortality all cancers	2	54517	Risk Ratio (M‐H, Random, 95% CI)	1.11 [0.99, 1.24]
4.1 Males	1	14641	Risk Ratio (M‐H, Random, 95% CI)	1.09 [0.92, 1.29]
4.2 Females	1	39876	Risk Ratio (M‐H, Random, 95% CI)	1.12 [0.95, 1.32]
5 Mortality all causes	3	69090	Risk Ratio (M‐H, Random, 95% CI)	1.03 [0.99, 1.07]
5.1 High‐risk people (smokers and asbestos workers)	1	14573	Risk Ratio (M‐H, Random, 95% CI)	1.03 [0.98, 1.07]
5.2 Males 50 years or older	1	14641	Risk Ratio (M‐H, Random, 95% CI)	1.03 [0.94, 1.12]
5.3 Females	1	39876	Risk Ratio (M‐H, Random, 95% CI)	1.03 [0.93, 1.15]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Incidence lung cancer	1	733	Risk Ratio (M‐H, Random, 95% CI)	0.65 [0.13, 3.18]
1.1 Females	1	733	Risk Ratio (M‐H, Random, 95% CI)	0.65 [0.13, 3.18]
2 Adverse events: renal calculi	1	733	Risk Ratio (M‐H, Random, 95% CI)	1.94 [0.20, 18.57]
3 Incidence all cancers	1	733	Risk Ratio (M‐H, Random, 95% CI)	0.55 [0.29, 1.03]
3.1 Females	1	733	Risk Ratio (M‐H, Random, 95% CI)	0.55 [0.29, 1.03]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Incidence lung cancer	1	17448	Risk Ratio (M‐H, Random, 95% CI)	1.11 [0.80, 1.54]
1.1 Males	1	17448	Risk Ratio (M‐H, Random, 95% CI)	1.11 [0.80, 1.54]
2 Mortality lung cancer	1	17448	Risk Ratio (M‐H, Random, 95% CI)	1.09 [0.72, 1.66]
2.1 Males	1	17448	Risk Ratio (M‐H, Random, 95% CI)	1.09 [0.72, 1.66]
3 Adverse events	1		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
3.1 Alopecia	1	17448	Risk Ratio (M‐H, Random, 95% CI)	1.28 [1.07, 1.53]
3.2 Dermatitis 1 to 2	1	17448	Risk Ratio (M‐H, Random, 95% CI)	1.16 [1.04, 1.31]
4 Incidence all cancers	1	17448	Risk Ratio (M‐H, Random, 95% CI)	1.01 [0.92, 1.11]
4.1 Males	1	17448	Risk Ratio (M‐H, Random, 95% CI)	1.01 [0.92, 1.11]
5 Mortality all cancers	1	17448	Risk Ratio (M‐H, Random, 95% CI)	1.02 [0.80, 1.30]
5.1 Males	1	17448	Risk Ratio (M‐H, Random, 95% CI)	1.02 [0.80, 1.30]
6 Mortality all causes	1	17448	Risk Ratio (M‐H, Random, 95% CI)	0.98 [0.86, 1.13]
6.1 Males	1	17448	Risk Ratio (M‐H, Random, 95% CI)	0.98 [0.86, 1.13]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Incidence lung cancer	1	14565	Risk Ratio (M‐H, Random, 95% CI)	1.10 [0.97, 1.24]
1.1 High‐risk people (smokers and asbestos workers)	1	14565	Risk Ratio (M‐H, Random, 95% CI)	1.10 [0.97, 1.24]
2 Incidence all cancers	1	14565	Risk Ratio (M‐H, Random, 95% CI)	1.04 [0.97, 1.11]
2.1 High‐risk people (smokers and asbestos workers)	1	14565	Risk Ratio (M‐H, Random, 95% CI)	1.04 [0.97, 1.11]
3 Mortality all causes	1	14565	Risk Ratio (M‐H, Random, 95% CI)	1.06 [1.02, 1.11]
3.1 High‐risk people (smokers and asbestos workers)	1	14565	Risk Ratio (M‐H, Random, 95% CI)	1.06 [1.02, 1.11]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Incidence lung cancer	1	7328	Risk Ratio (M‐H, Random, 95% CI)	0.83 [0.50, 1.39]
1.1 Males	1	7328	Risk Ratio (M‐H, Random, 95% CI)	0.83 [0.50, 1.39]
2 Incidence all cancers	1	7309	Risk Ratio (M‐H, Random, 95% CI)	1.03 [0.91, 1.16]
2.1 Males	1	7309	Risk Ratio (M‐H, Random, 95% CI)	1.03 [0.91, 1.16]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Incidence lung cancer	1	17399	Risk Ratio (M‐H, Random, 95% CI)	1.16 [0.84, 1.61]
1.1 Males	1	17399	Risk Ratio (M‐H, Random, 95% CI)	1.16 [0.84, 1.61]
2 Mortality lung cancer	1	17399	Risk Ratio (M‐H, Random, 95% CI)	0.95 [0.61, 1.47]
2.1 Males	1	17399	Risk Ratio (M‐H, Random, 95% CI)	0.95 [0.61, 1.47]
3 Adverse events: halitosis	1	17399	Risk Ratio (M‐H, Random, 95% CI)	1.24 [1.10, 1.41]
4 Incidence all cancers	1	17399	Risk Ratio (M‐H, Random, 95% CI)	1.03 [0.94, 1.12]
4.1 Males	1	17399	Risk Ratio (M‐H, Random, 95% CI)	1.03 [0.94, 1.12]
5 Mortality all cancers	1	17399	Risk Ratio (M‐H, Random, 95% CI)	0.94 [0.73, 1.20]
5.1 Males	1	17399	Risk Ratio (M‐H, Random, 95% CI)	0.94 [0.73, 1.20]
6 Mortality all causes	1	17399	Risk Ratio (M‐H, Random, 95% CI)	0.94 [0.82, 1.08]
6.1 Males	1	17399	Risk Ratio (M‐H, Random, 95% CI)	0.94 [0.82, 1.08]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Incidence lung cancer	1	12741	Risk Ratio (M‐H, Random, 95% CI)	0.64 [0.28, 1.48]
1.1 Males (65% exposed to tobacco or asbestos)	1	5028	Risk Ratio (M‐H, Random, 95% CI)	0.60 [0.22, 1.64]
1.2 Females (45% exposed to tobacco or asbestos)	1	7713	Risk Ratio (M‐H, Random, 95% CI)	0.75 [0.17, 3.37]
2 Incidence all cancers	1	12741	Risk Ratio (M‐H, Random, 95% CI)	0.96 [0.83, 1.10]
2.1 Males	1	5028	Risk Ratio (M‐H, Random, 95% CI)	0.89 [0.74, 1.06]
2.2 Females	1	7713	Risk Ratio (M‐H, Random, 95% CI)	1.02 [0.87, 1.20]
3 Mortality all causes	1	12741	Risk Ratio (M‐H, Random, 95% CI)	0.88 [0.70, 1.11]
3.1 Males	1	5028	Risk Ratio (M‐H, Random, 95% CI)	0.79 [0.60, 1.05]
3.2 Females	1	7713	Risk Ratio (M‐H, Random, 95% CI)	1.01 [0.72, 1.40]

Methods	Alpha‐Tocopherol Beta‐Carotene Cancer prevention (ATBC) study Randomised, double‐blind, placebo‐controlled trial  Objective: to evaluate the effectiveness of alpha‐tocopherol or beta‐carotene supplements for the prevention of lung cancer  Southwestern Finland, period 1985‐1993
Participants	29133 male smokers  50‐69 years of age  Excluded: those with previous cancer or other serious illness, users of vitamin E, vitamin A, or beta‐carotene supplements in excess of predefined doses, or treatment with anticoagulants. Smoking status definition criteria: 5+ cigarettes/day. Median of 20 cigarettes smoked daily and duration of smoking prior to study entry 36 years.
Interventions	Four intervention groups: a) alpha‐tocopherol (50 mg/day), n: 7286. b) beta‐carotene (20 mg/day), n: 7282. c) alpha‐tocopherol (50 mg/day) + beta‐carotene (20 mg/day), n: 7278. d) placebo, n: 7287. Comparisons: (1) alpha‐tocopherol (a+c) versus no alpha‐tocopherol (b+d), n: 14564 versus 14569) (2) beta‐carotene (b+c) versus no beta‐carotene (a+d), n: 14560 versus 14573 (3) alpha‐tocopherol (a+c) versus placebo (d), n: 14564 versus 7287 (4) beta‐carotene (b+c) versus placebo (d), n: 14560 versus 7287 (5) alpha‐tocopherol + beta‐carotene (c) versus placebo (d), n: 7278 versus 7287 All doses were administered daily. Duration of treatment: five to eight years, median 6.1 years
Outcomes	Lung cancer incidence Lung cancer mortality Total cancer mortality All‐cause mortality Follow‐up: trial period 29133 participants (8 years: median, 6.1), total of 169751 participant‐years. Post‐trial period another 3 years with 25283 participants, and another 3 years with 22838 participants; six years for cancer incidence and mortality and eight years for total mortality.
Notes	Comparisons 1 and 2: intention‐to‐treat analysis Comparisons 3, 4 and 5: not intention‐to‐treat analysis Information on cancer incidence and mortality was mainly taken from the cancer registry. Trial registration identifier: NCT00342992 Funding: supported by Public Health Service of Finland contracts N01‐CN‐45165 and N01‐RC‐45035 from the US National Cancer Institute, National Institutes of Health, Department of Health and Human Services
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Mentioned as "randomly assigned" but sequence generation process is not explained in a detailed way, but probably done centrally, given that: (quote) "A coded reserve supply of capsule packs was maintained centrally in the event of lost capsules requiring replacement."
Allocation concealment (selection bias)	Low risk	Randomisation was probably done centrally. Quote: "A coded reserve supply of capsule packs was maintained centrally in the event of lost capsules requiring replacement."
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Stated as "double‐blind". Quote: "A coded reserve supply of capsule packs was maintained centrally in the event of lost capsules requiring replacement. All formulation were coloured with quinoline yellow”.
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Quote: "Participants and all study staff involved in the ascertainment of end points and the assignment of final diagnoses remained blinded to the participants' treatment assignment throughout the trial".
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Causes for withdrawal from the study well reported. Quote: "The chest film at study exit was available for all but 494 surviving men, yielding a 98% success rate that was equal across the supplementation groups". Quote: "The dropout rate varied only slightly across the randomised groups".
Selective reporting (reporting bias)	Low risk	Authors present results on all outcome measures that were prespecified as relevant.
Other bias	Low risk	The study appears to be free of other sources of bias.

Methods	Women's Health Initiative (WHI) calcium plus vitamin D Randomized clinical trial with blinding. Objective: to determine whether calcium plus vitamin D supplementation would prevent hip fracture. The original study examined effects on incidence and mortality for all invasive cancers and lung cancer was specifically reported. Study period between 1994 and 1999, mean follow‐up of 7 years
Participants	17343 women in the supplement group and 17327 in the placebo group. Postmenopausal women 50 to 79 years of age Exclusion criteria: a predicted survival of less than 3 years, a history of renal calculi or hypercalcaemia, current use of oral corticosteroids, and current daily use of at least 600 IU of supplemental vitamin D (single supplement and multivitamin combined) or calcitriol.
Interventions	Study group: 1000 mg of elemental calcium as calcium carbonate combined with 400 IU of vitamin D3. Each pill contained half the dose and 2 pills were taken daily. Control group: received identical‐appearing tablets, each taken twice daily
Outcomes	Lung cancer incidence (lung cancer incidence was reported separately) Incidence all cancer Mortality all cancers
Notes	Intention‐to‐treat analysis
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote:"Randomization was done using a permuted‐block algorithm with participants stratified according to clinical center and age".
Allocation concealment (selection bias)	Unclear risk	Comment: not stated
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Quote: "Blinding of the study was achieved by identical‐looking pills and bottles". Comment: however, participant's blinding not affecting the outcomes of this study because the medical records were obtained for any self‐reported cancers and were reviewed and verified by both local and central physician adjudicators who were blinded to randomisation status.
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Quote: "local and central physician adjudicators were blinded to randomization status".
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	Quote: "about one‐quarter of the participants had stopped taking pills by the end of the study". Comment: not stated how many participants were lost to follow‐up in each group.
Selective reporting (reporting bias)	Low risk	The study protocol is available and published reports include all prespecified outcomes
Other bias	Low risk	The study appears to be free of other sources of bias.

Methods	Physicians' Health Study II (PHS II) Randomised in blocks of 16 and stratified by age, double‐blind, placebo‐controlled, 2 x 2 x 2 x 2 factorial trial  Objective: to evaluate whether long‐term vitamin E or C supplementation decreases risk of prostate and total cancer events among men. The original study examined the effect on lung cancer incidence and incidence for all cancers. Lung cancer incidence was specifically reported.  Period: began in 1997 and continued until its scheduled completion on 31 August 2007. Analysis was in terms of number of events per participant‐years of follow‐up for each study agent, and was conducted on intention‐to‐treat basis.
Participants	14641 male physicians (7641 from PHS I and 7000 new physicians). Mean age 64.3 years; 56.4% never smokers, 40% former smokers, 3.6% current smokers. Inclusion criteria: 50 years and older; no history of cancer (except non‐melanoma skin cancer), myocardial infarction, stroke or transient cerebral ischemias; no current liver or renal disease, peptic ulcer, gout, and were required to indicate their willingness to avoid the use of non‐study vitamin supplements. Smoking status definition criteria: none explicit. Categories: never, former, current.
Interventions	Vitamin E (400 IU) every other day Vitamin C (500 mg) daily Multivitamin, daily Placebo daily Duration of treatment: six years
Outcomes	Incidence of prostate and total cancer (including lung cancer incidence) adverse events Mean follow‐up 8 years, median 7.6
Notes	PHS I participants who enrolled in PHS II (approximately 7500) continued on their original randomised beta‐carotene treatment assignment and also were randomised to vitamin C, vitamin E, and a multivitamin, or their placebos. New physician participants in PHS II (approximately 7500), identified from a roster of all potentially eligible USA male physicians provided by the American Medical Association, randomised to beta‐carotene, vitamin E, vitamin C, and a multivitamin, or their placebos. Included 1307 men with a history of prior cancer at randomisation. For analyses of the secondary endpoints of total mortality, any cancer mortality, and site‐specific cancer deaths, we included all participants Trial registration, ClinicalTrials.gov Identifier: NCT00270647 Funding, sponsors and collaborators: Brigham and Women's Hospital and National Cancer Institute
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Randomised according to a two‐by‐two factorial design, with use of a computer‐generated list of random numbers" Quote: "will be stratified according to age (55‐59, 60‐64, 65‐69, 70‐74, and 751 years) in blocks of sixteen."
Allocation concealment (selection bias)	Low risk	Central provision of active drugs and placebo. Quote: "The participants were sent monthly calendar packs."
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Quote: "Blinding of participants and physicians."
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Quote: "reported diagnoses were confirmed after examination of all available information by a committee of physicians..., all blinded to treatment assignment."
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Very small percentage of losses in follow‐up (0.01% of participant‐years of follow‐up).
Selective reporting (reporting bias)	Low risk	The study protocol is available and published reports include all prespecified outcomes.
Other bias	Low risk	The study appears to be free of other sources of bias.

PERMALINK

Drugs for preventing lung cancer in healthy people

Marcela Cortés‐Jofré

José‐Ramón Rueda

Claudia Asenjo‐Lobos

Eva Madrid

Xavier Bonfill Cosp

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

Plain language summary

Summary of findings

Background

Description of the condition

Description of the intervention

How the intervention might work

Why it is important to do this review

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Types of participants

Types of interventions

Types of outcome measures

Search methods for identification of studies

Electronic searches

Searching other resources

Data collection and analysis

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

Measures of treatment effect

Unit of analysis issues

Dealing with missing data

Assessment of heterogeneity

Assessment of reporting biases

Data synthesis

'Summary of findings' tables

Summary of findings for the main comparison. Vitamin A compared to placebo for preventing lung cancer in healthy people.

Summary of findings 2. Vitamin C compared to placebo for preventing lung cancer in healthy people.

Summary of findings 3. Vitamin D plus calcium compared to placebo for preventing lung cancer in healthy people.

Summary of findings 4. Vitamin E compared to placebo for preventing lung cancer in healthy people.

Summary of findings 5. Calcium compared to placebo for preventing lung cancer in healthy people.

Summary of findings 6. Selenium compared to placebo for preventing lung cancer in healthy people.

Summary of findings 7. Vitamins A, C, E + selenium + zinc compared to placebo for preventing lung cancer in healthy people.

Subgroup analysis and investigation of heterogeneity

Sensitivity analysis

Results

Description of studies

Results of the search

1.

Included studies

1. Active intervention compared to placebo.

2. Length of treatment and follow‐up period.

Excluded studies

Risk of bias in included studies

2.

3.

Allocation

Blinding

Incomplete outcome data

Selective reporting

Other potential sources of bias

Effects of interventions

Vitamin A (beta‐carotene or retinol) versus placebo

Primary outcomes

Lung cancer incidence

1.1. Analysis.

Lung cancer mortality

1.2. Analysis.

Adverse events

1.3. Analysis.

Secondary outcomes

Total cancer incidence

1.4. Analysis.

Methods	Physicians' Health Study (PHS) Randomised controlled trial. Double‐blind.  Objective: to evaluate the effectiveness of beta‐carotene supplements for preventing cancer.  Harvard Medical School and Brigham and Women's Hospital, Boston, USA  Period: 1982‐1995  Intention‐to‐treat analysis
Participants	22071 healthy male physicians (11112 smokers and 10919 non‐smokers), selected from the American Medical Association; aged 40‐84; no history of cancer (except non‐melanoma skin cancer), myocardial infarction, stroke or transient cerebral ischemias; no current liver or renal disease, peptic ulcer, gout, no contraindications to aspirin, or use of aspirin, other platelet active drugs, nonsteroidal anti‐inflammatory agents or vitamin A supplements; no side effects to aspirin. High compliance, measured in a run‐in phase. Smoking status definition criteria: none explicit. Categories: never, former, current
Interventions	Intervention: beta‐carotene (50 mg on alternate days), n = 11036 Placebo, n = 11035 Duration of treatment: average 12 years (range, 11.6 to 14.2)
Outcomes	Incidence of cancer (except non‐melanoma skin cancer); histopathologically confirmed; only the first diagnosed cancer during follow‐up was counted Lung cancer incidence Lung cancer mortality adverse events Total mortality and total cancer mortality Lung cancer incidence and mortality was specifically reported.  Follow‐up: average 12 years (from randomisation)
Notes	The study tested two hypotheses: 1) aspirin (325 mg alternate days) reduces cardiovascular mortality; 2) beta‐carotene reduces incidence of cancer. Only data on objective 2 were included in this review. The aspirin component was terminated early, in 1988, due to a statistically extreme reduction in incidence of first myocardial infarction. 5% of participants did not give consent to confirm their potential events and were not included in the analysis. Trial Registration, ClinicalTrials.gov Identifier: NCT00005252 Funding, sponsors and collaborators: National Heart, Lung, and Blood Institute (NHLBI).
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Randomised according to a two‐by‐two factorial design, with use of a computer‐generated list of random numbers" Quote: "The Physicians' Health Study is a randomised, double‐blind, placebo‐controlled trial with a two‐by‐two factorial design."
Allocation concealment (selection bias)	Low risk	Central provision of active drugs and placebo. Quote: "The participants were sent monthly calendar packs."
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Quote: "Blinding of participants and physicians."
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Quote: "reported diagnoses were confirmed after examination of all available information by a committee of physicians..., all blinded to treatment assignment."
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Quote: "Data on all 22,071 participants were analysed according to their treatment assignment, 0.3% were lost to follow‐up.The data were analysed according to intention to treat."
Selective reporting (reporting bias)	Low risk	The study protocol is available and published reports include all prespecified outcomes (mortality of cardiovascular disease and lung cancer incidence).
Other bias	Low risk	The study appears to be free of other sources of bias.

Methods	Trial registration: "Primary Prevention Trial of the Health Effects of Antioxidant Vitamins and Minerals." Randomised, placebo‐controlled trial Objective: assess the efficacy of nutritional doses of supplementation with a combination of antioxidant vitamins and minerals in reducing the incidence of cancer and Ischaemic cardiovascular disease in the general population. France.
Participants	Inclusion criteria: women in the range of 35‐60 and men in the age range of 45‐60 from all over France. At randomisation: 6364 in placebo and 6377 in supplemented group 60.5% women; mean age 49 years. Women: 54.1% non‐smokers, 28.9% former smokers, 17% current smokers; men: 32.3% non‐smokers, 46.2% former smokers, 21.5% current smokers Smoking status definition criteria: none explicit
Interventions	Intervention: combination of antioxidants (120 mg of ascorbic acid, 30 mg of vitamin E, 6 mg of beta‐carotene, 100 μg of selenium (as selenium‐enriched yeast), and 20 mg of zinc (as gluconate)) in a single daily capsule Placebo in a single daily capsule Duration of treatment: 8 years
Outcomes	Primary outcomes: major fatal and nonfatal ischaemic cardiovascular events and cancer of any kind, except for basal cell carcinoma of the skin Secondary outcome: all‐cause mortality Follow‐up period: 12.5 years
Notes	Trial registration, ClinicalTrials.gov Identifier: NCT00272428 The SU.VI.MAX project received public and private support form the following companies or subsidiaries, all located in France: Institut National de la Santé Et de la Recherche Médicale, Fruit d'Or Recherche, Lipton, Cereal, Candia, Kellogg's, CERIN, LU/Danone, Sodexho, L'Oréal, Estée Lauder, Peugeot, Jet Service, RP Scherer, France Telecom, Becton Dickinson, Fould Springer, Boehringer Diagnostic, Seppic Givaudan Lavirotte, Le Grand Canal, Air Liquide, Carboxyque, Klocke, Trophy Radio, Jouan, and Perkin Elmer.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Random treatment allocation was performed by block sequence generation stratified by sex, age group, smoking habits, and residential area.
Allocation concealment (selection bias)	Low risk	Capsules were prepared in 52 weekly packages of 7 capsules and delivered each year in a box labelled with the participant's number and a 10‐digit lot number.
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Randomisation was concealed from subjects and all investigators except for the few who were in charge of capsule  labelling.
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Randomisation was concealed from subjects and all investigators except for the few who were in charge of capsule labelling.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	At the end of the supplementation period, 5501 participants in the intervention group and 5553 in the placebo group remained. Losses in the postintervention period explained.
Selective reporting (reporting bias)	Low risk	Authors present results on all outcome measures that were prespecified as relevant.
Other bias	Low risk	The study appears to be free of other sources of bias.

Methods	Linxian General Population Nutrition Intervention Trial Randomised, double‐blind, placebo‐controlled trial  Objective: to evaluate the effect of supplementation with four different combinations of vitamins and minerals in the prevention of lung cancer mortality in healthy adults from Linxian, China
Participants	29584 adults Median age 52 years; 55% female; 30% smoked tobacco; 23% reported alcohol use the past year, and 32% had a family history of oesophageal or stomach cancer. Inclusion criteria: 40 to 69‐year‐old adults, with no history of malignancy. Smoking status definition criteria: ever smoking cigarettes for 6 or more months.
Interventions	Five groups: retinol (as palmitate 5000 IU) and zinc (as zinc oxide 22.5 mg) daily riboflavin (3.2 mg) and niacin (40 mg) daily vitamin C (ascorbic acid 120 mg) and molybdenum (as molybdenum yeast complex 30 ug) daily beta‐carotene (15 mg), vitamin E (a‐tocopherol 30 mg) and selenium (as selenium yeast 50 ug) daily placebo daily Doses for those daily supplements ranged from 1 to 2 times USA recommended daily allowances. Duration of treatment: 5.25 years
Outcomes	Lung cancer mortality Follow‐up: 15 years. Intervention: 5.2 years and postintervention follow‐up: 10 years
Notes	Quote: "The people of Linxian are deficient in many micronutrients, which may limit the generalization of these results. Nevertheless, the results of this study are similar to other chemoprevention studies, which did not find benefit from vitamins in reducing lung cancer incidence or mortality". In Linxian area oesophageal and gastric cardia cancer mortality were among the highest in the world. Trial registration, ClinicalTrials.gov Identifier: NCT00342654 Funding, sponsors and collaborators: National Cancer Institute
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Computer‐generated random numbers."
Allocation concealment (selection bias)	Low risk	Central allocation
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Stated as "double‐blind or "double masked." Coded pill bottles kept in the central study management centre and available only to the study data manager.
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Stated as "double‐blind or "double masked."
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Quote: "case ascertainment was considered complete and loss to follow‐up minimal (n = 276, or < 1%)."
Selective reporting (reporting bias)	Low risk	Authors present results on all outcome measures that were prespecified as relevant.
Other bias	Low risk	The study appears to be free of other sources of bias.

Methods	Double‐blind, randomised placebo‐controlled trial Objective: to determine the efficacy of calcium alone and calcium plus vitamin D in reducing cancer incident risk of all types. Study period was four years. Participants from nine county farming areas of eastern Nebraska, USA
Participants	White women, healthy postmenopausal women over 55 years. Women needed to be at least four years past last menses, in generally good health, living independently in the community, and weighing less than 300 pounds. Exclusion criteria: medical diagnosis of any chronic kidney disease, Paget's or other metabolic bone disease, and history of cancer except for superficial basal or squamous cell carcinoma of the skin and other malignancies treated curatively more than 10 years prior to entry into study.
Interventions	Study group: 1400 to 1500 mg supplemental calcium/d alone (Ca‐only), supplemental calcium plus 1100 IU vitamin D3/d (Ca D) Control group: placebo
Outcomes	This study presented data related to a secondary endpoint: cancer incidence. Lung cancer incidence was specifically reported and used in this analysis. Adverse events were reported.
Notes	Study registered at ClinicalTrials.gov as NCT00352170 Intention‐to‐treat analysis
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: use of a computer‐generated permuted blocks (n = 5) randomisation scheme
Allocation concealment (selection bias)	Unclear risk	Not stated
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Stated as "double‐blind".
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	In our review relevant outcomes are all objective (cancer incidence, mortality and adverse events)
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	Quote: "Of 1180 women enrolled, 1024 (86.8%) completed the 4 year of study. Most of the losses (n = 92) occurred within the first year." Comment: not stated how many participants were lost to follow‐up in each group
Selective reporting (reporting bias)	Low risk	The protocol of the study is available and results on all outcomes published.
Other bias	Low risk	Comment: only women were included, which limits the study's external validity

Methods	Creighton cancer and vitamin D study Double‐blind, placebo‐controlled, population‐based randomised clinical trial Objective: to determine if dietary supplementation with vitamin D3 and calcium reduces the risk of cancer among healthy older women Duration: 4 years (24 June 2009 to 26 August 2015 ‐ the final date of follow‐up) Location: 31 rural counties of Nebraska
Participants	Postmenopausal women 55 years and older Mean age: 65.2 years (standard deviation 7.0)
Interventions	Intervention group: (vitamin D3 + calcium group) received 2000 IU/d of vitamin D3 and 1500 mg/d of calcium; the placebo group received identical placebos
Outcomes	Primary outcome: incidence of all‐types of cancer (excluding non‐melanoma skin cancers). Secondary outcomes were planned for common specific types of cancer including: cancers of the breast, lung and colon lymphoma leukaemia myeloma Lung cancer incidence was specifically reported and used in this review. adverse events were reported.
Notes	ClinicalTrials.gov: NCT01052051
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote:"randomized by a statistician (P.W.) to 1 of 2 groups by computerized block randomization, using a block size of 8."
Allocation concealment (selection bias)	Unclear risk	Not stated
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Stated as double‐blind. Quotes: "identical placebo to the intervention". "Only statistician and a research assistant who had no contact with participants were unblinded to group assignment."
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Quote: "Only statistician and a research assistant who had no contact with participants were unblinded to group assignment."
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Withdrawal and lost to follow‐up were low and similar in both groups.
Selective reporting (reporting bias)	Low risk	All outcome were prespecified in protocol
Other bias	Low risk	Quote: "The cohort included older women, primarily non‐Hispanic white, and no men." Comment: this limits the study's external validity.

Methods	Women's Health Study (WHS) Randomised, double‐blind, placebo‐controlled trial, using a 2 x 2 x 2 factorial design  Objective: to test the balance of benefits and risks of aspirin, vitamin E, and beta‐carotene in the primary prevention of cancer and cardiovascular disease. Brigham and Women's Hospital and Harvard Medical School, Boston, USA. Period: 1993‐1998. Intention‐to‐treat analysis
Participants	39876 female health professionals; aged 45 or older. No history of cancer (except non‐melanoma skin cancer), coronary heart disease or cerebrovascular disease. 13% (2635) of women assigned to beta‐carotene and 13% (2635) to placebo group were cigarette smokers at baseline. Smoking status definition criteria: none explicit. Categories: current, past or never
Interventions	Eight treatment groups: all three active agents three groups of two active agents and one placebo three groups of one active agent and two placebo, or all three placebos Comparisons: beta‐carotene (50 mg alternate days) groups (n = 19,939) or versus placebo groups (n = 19,937) vitamin E (600 IU every other day) versus placebo groups Duration of treatment: 2.1 years
Outcomes	Primary outcome: major cardiovascular event total invasive cancer (apart from non‐melanoma skin cancer) Secondary outcomes: individual cardiovascular events cardiovascular death main site‐specific cancers in women: breast, lung, and colon cancers Lung cancer incidence was specifically reported and used in this review.
Notes	The beta‐carotene component was terminated early because of harmful results of an interim analysis in the CARET study for beta‐carotene. Trial registration, ClinicalTrials.gov Identifier: NCT00000479 Sponsors and collaborators: National Heart, Lung, and Blood Institute (NHLBI)
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Insufficient information in publications to permit judgement, but probably done using a "computer‐generated list of random numbers" since some of the members of the team were the same as the Physicians Health Study. Quote: "Is a randomised study, with a two‐by‐two factorial design."
Allocation concealment (selection bias)	Low risk	Quote: "Study agents provided in convenient monthly calendar packs."
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Quote: "Participants and investigators will be blinded to treatment groups."
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Quote: "Reported diagnoses will be confirmed by an End points Committee of physicians (all of whom will be blinded to participant's treatment assignments)."
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Quote: "Only three losses were reported among the 39,876 participants."
Selective reporting (reporting bias)	Low risk	The study protocol is available and published reports include all prespecified outcomes.
Other bias	Low risk	The study appears to be free of other sources of bias.

Methods	Women's Antioxidant and Cardiovascular Study (WACS) Randomised, double‐blind, placebo‐controlled, 2 × 2 × 2 factorial trial  Objective: to evaluate the individual and combined effects of three antioxidant supplements, ascorbic acid, vitamin E, and beta‐carotene, in the prevention of cardiovascular diseases Location: USA
Participants	7627 USA female health professionals. Median age 60.4 years; 77% postmenopausal, 77% overweight or obese, 27% reported having taken multivitamins at baseline. Inclusion criteria: at least 40 years‐old; were postmenopausal or not intending to become pregnant; and had known cardiovascular disease or at least three of the following cardiac risk factors: hypertension, high cholesterol level, diabetes, parental history of myocardial infarction, or obesity (i.e. body mass index ≥ 30 kg/m²). Exclusion criteria: self‐reported history of cancer (except non‐melanoma skin cancer) within the past 10 years, had active liver disease or cirrhosis, had chronic kidney failure, were current users of anticoagulants, or were unwilling to avoid out of study use of vitamins A, C, and E and beta‐carotene at intakes exceeding the recommended daily allowance during the trial. Smoking status definition criteria: none explicit. Categories: never, past, current
Interventions	WACS was designed as a 3‐group trial: vitamin C (500 mg of ascorbic acid) versus placebo daily vitamin E (600 IU of α –tocopherol) versus placebo every other day beta‐carotene (50 mg of Lurotin) versus placebo every other day Duration of treatment: average 9.2 years
Outcomes	Primary outcome: total cancer incidence (excluding non‐melanoma skin cancers) and site‐specific types of cancer Lung cancer incidence was specifically reported and used in this review.
Notes	The trial was conducted as a companion to the Women's Health Study (WHS). Quote: “This study had very limited statistical power to investigate any effect of dietary antioxidants on the risk of specific cancers.” Trial registration, ClinicalTrials.gov Identifier: NCT00000541 Funding, sponsors and collaborators: National Heart, Lung, and Blood Institute (NHLBI)
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Subjects were randomised in a 2 x 2 x 2 factorial design to 500 mg of vitamin C or placebo daily, 600 mg of vitamin E or placebo on alternate days and/or 50 mg of beta‐carotene or placebo on alternate days. There was a three‐month run‐in phase in which eligible participants received placebo caplets. Subjects were randomised only if they reported good compliance, willingness to continue in the trial, had no history of cancer, active liver disease, or use of coumadin, and expressed continued willingness to forego the use of beta‐carotene and vitamin A, C, or E supplements. In 1998, participants were further randomised to the B‐vitamin intervention (folic acid, vitamin B6, vitamin B12).
Allocation concealment (selection bias)	Low risk	Central allocation
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Stated as "double‐blind"
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Stated as "double‐blind”
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	No detailed information published on losses to follow‐up
Selective reporting (reporting bias)	Low risk	Authors present results on all outcome measures that were prespecified as relevant.
Other bias	Low risk	The study appears to be free of other sources of bias.

Methods	Selenium and Vitamin E Cancer Prevention Trial (SELECT) Phase III randomised, placebo‐controlled trial  Objective: to determine whether selenium, vitamin E, or both could prevent prostate cancer and other diseases with little or no toxicity in relatively healthy men
Participants	35533 men from 427 participating sites in the USA, Canada, and Puerto Rico. Median age 62.4 years; 78% White, 20% African Americans, 3% Hispanics, 1% Asians and 17% had a family history of prostate cancer. Inclusion criteria: age 50 years or older for African American men and 55 years or older for all other men, no prior prostate cancer diagnosis, 4 ng/mL or less of PSA in serum, and a digital rectal examination (DRE) 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, and normal blood pressure. Smoking status definition criteria: none explicit. Categories: never, current, former, ever (unknown status), unknown.
Interventions	SELECT was designed as a 4‐group trial with 5 prespecified comparisons: selenium (200 μg) versus placebo daily vitamin E (400 IU) versus placebo daily selenium (200 μg) and vitamin E (400 IU) versus placebo daily selenium (200 μg) versus selenium(200 μg) and vitamin E(400 IU) daily vitamin E (400 IU) versus selenium and vitamin E(400 IU) daily Duration of treatment: seven years
Outcomes	Primary outcome: prostate cancer incidence Prespecified secondary outcomes: lung, colorectal, and overall primary cancer incidence and survival Lung cancer incidence was specifically reported and used in this review. adverse events were reported.
Notes	Quote: "On September 15, 2008, the independent data and safety monitoring committee met, reviewed data as of August 1, 2008, for the second formal interim analysis, and recommended the discontinuation of study supplements because the alternative hypothesis of no evidence of benefit from either study agent was convincingly demonstrated (P.0001) and there was no possibility of a benefit to the planned degree with additional follow‐up. Study sites were notified to discontinue supplements on October 23, 2008, and the data presented in this article are current as of this date." Trial registration, ClinicalTrials.gov Identifier: NCT00006392 Funding, sponsors and collaborators: Southwest Oncology Group, National Cancer Institute, National Center for Complementary and Alternative Medicine (NCCAM), Eastern Cooperative Oncology Group Cancer and Leukemia Group B, NCIC Clinical Trials Group
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Participants were randomised in a randomised block scheme, in which the block was the study site. This ensured a balance of the 4 intervention groups within each study site."
Allocation concealment (selection bias)	Low risk	Central randomisation
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Doble‐blind and main outcome measure assessments are not likely to be biassed by lack of blinding. Quote: "To ensure the quality Supplement Quality Control and Quality Assurance of the blind was maintained, capsules received in each subsequent lot were compared with the previous lot and with matching capsules in the current shipment for their characteristics of weight, shape and size, colour and external marking, odour, and comparability of contents of opened capsules."
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Doble‐blind and main outcome measure assessments are not likely to be biased by lack of blinding.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	The authors' detailed information published on losses to follow‐up. Quote: "All analyses were performed by using an intention‐to‐treat analysis in which men were classified according to the group to which they were randomised."
Selective reporting (reporting bias)	Low risk	Authors presented results on all outcome measures that were prespecified as relevant.
Other bias	Low risk	Monitoring policy for stopping the trial and interim analyses previously defined.

Methods	Carotene and Retinol Efficacy Trial (CARET) Randomised, double‐blind, placebo‐controlled trial  Objective: to evaluate the effectiveness of retinol and beta‐carotene supplements for the prevention of lung cancer.  Seattle, Portland, San Francisco, Baltimore, Conneticut, Irvine (USA).  Period: pilot study: 1985‐1988; efficacy cohort: 1989 and 1991 recruitment of additional study centres, follow‐up until 1995 and 2000.  Intention‐to‐treat analysis
Participants	4060 male workers exposed to asbestos, aged 45‐69 (45‐74 in pilot study), 15 years. Previous exposure to asbestos, and asbestos‐related disease or workers in high‐risk trades for 5 years Current non‐smokers or smoking in previous 15 years Smoking status definition criteria: none explicit 14,254 male and female smokers recruited from healthcare organisations Aged 50‐69 years Heavy current or ex‐smokers (previous 6 years) of 20 or more pack‐years cigarettes Inclusion criteria: asbestos‐exposed men who were: current smokers or quit within 15 years prior to enrolment had first exposure to asbestos on the job at least 15 years prior to enrolment had chest X‐ray positive for changes compatible with asbestos exposure according to International Labour Organization criteria; or had been employed in a protocol‐defined high‐risk trade for at least 5 years, at least 10 years prior to enrolment heavy smokers, men and women: cigarette smoking history of 20+ pack‐years, either current smokers or had quit within previous 6 years. Exclusion criteria: premenopausal women history of cirrhosis or hepatitis within 12 months prior to enrolment taking > 5500 IU daily vitamin A supplement taking any beta‐carotene supplement history of cancer within 5 years prior to enrolment serum glutamic‐oxaloacetic transaminase (SGOT) > than 2.5x upper limit of normal, or alkaline phosphatase > 1.5x upper limit of normal taking less than 50% of study vitamins during the enrolment period between the first and second visit
Interventions	Intervention: 30 mg/day beta‐carotene + 25000 IU/day retinol Comparison: two placebos, one each/day. Duration of treatment: planned for eight years but stopped ahead of schedule after interim analysis
Outcomes	Lung cancer incidence Incidence of other cancers Cancer mortality During the postintervention phase, primary endpoints were incidences of lung cancer, all‐cause mortality, and mortality from cardiovascular disease. Duration of follow‐up: stopped ahead of schedule after interim analysis
Notes	Information on cancer incidence and mortality was obtained from clinical records.  The CARET intervention was stopped 21 months early because of clear evidence of no benefit and substantial evidence of possible harms. Because the CARET Steering Committee decided to end active intervention on 11 January 1996, all participants were asked to stop taking the intervention agents and to return to their study centre, where a final blood sample was collected from each participant and written informed consent was obtained for postintervention follow‐up. A total of 1174 participants who were enrolled in CARET did not contribute participant‐years of follow‐up to this postintervention analysis; of these, 1092 (93%) died during the intervention phase and 82 (7%) were lost to follow‐up. In the ongoing postintervention follow‐up in CARET, 93% of the living participants are being followed actively through mailed questionnaires; the remainder (including those considered lost to follow‐up during the intervention phase) are being followed passively through searches of local cancer registries and the National Death Index. Trial registration identifier: NCT00712647 Funding, sponsors and collaborators: Fred Hutchinson Cancer Research Center and national Cancer Institute
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Randomization is based on a permuted blocks algorithm with random block size and equal allocation to the two arms, stratified by study centre and exposure population. The unit of randomizations is the household to guard against household members taking the wrong vitamin type."
Allocation concealment (selection bias)	Low risk	Central allocation
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Stated as "double‐blind or "double masked"
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Stated as "double‐blind or "double masked"
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Quote: "As of December 15, 1995, ascertainment of vital status was more than 98 percent complete."
Selective reporting (reporting bias)	Low risk	Authors present results on all outcome measures that were prespecified as relevant.
Other bias	Low risk	Monitoring policy for stopping the trial and interim analyses previously defined

Study	Reason for exclusion
Albanes 1986	Feasibility study; pilot study
Alpha‐Tocopherol Study Group 1994	A postintervention follow‐up
Arnold 1992	Randomised placebo‐controlled trial of chemoprevention in healthy smokers, but outcome measure was change in sputum atypia (intermediate endpoint)
Ayoub 1999	Randomised controlled trial of chemoprevention in healthy smokers, but outcome measure was abnormalities in the expression of Retinoic acid receptor beta (intermediate endpoint)
Cullen 2005	Observational post‐trial follow‐up
De Klerk 1998	Compares retinol and B‐carotene, but no placebo or intervention group
Ebbing 2009	Not healthy people. Quote "A total of 6837 patients with ischaemic heart disease were treated with B vitamins or placebo between 1998 and 2005, and were followed up through December 31, 2007."
Goodman 1993	Feasibility study; pilot trial
Holick 2002	Study post‐trial follow‐up that ended on 30 April 1993, with the passive case analysis for this study that continued until the date of death or until December 1998. Follow‐up data up to 14 years (median, 11 years)
Kelly 2009	Wrong outcomes. The primary outcome measure was treatment "failure", defined as histologic progression.
Kurie 2000	Randomised, double‐blind, placebo‐controlled trial of chemoprevention in healthy smokers, but the outcome measure was change in bronchial epithelium (intermediate endpoint)
Lee 1994	Randomised placebo‐controlled trial of chemoprevention of lung cancer in healthy smokers, but the outcome measure was change in bronchial metaplasia index (intermediate endpoint)
Lee 1998	Comparative study with placebo‐controlled group in healthy smokers, but was not randomised and the outcome was oxidative DNA and protein (globin) damage (intermediate endpoint)
Shiels 2011	Observational post‐trial follow‐up
Tao 2017	Postintervention follow‐up of Womens Health Initiative calcium and Vitamin D supplementation double‐blinded, randomised, placebo‐controlled trial in 36,382 postmenopausal women aged 50–79 years, recruited at 40 USA centres. Postintervention follow‐up continued among 29,862 (86%) of the surviving participants.
Van Poppel 1997	Randomised controlled trial of chemoprevention in healthy smokers, but the outcome was sputum cytology (intermediate endpoint)
Virtamo 2014	Postintervention follow‐up of the Alpha‐Tocopherol, Beta‐Carotene Cancer Prevention (ATBC) study. The trial period continued through 30 April 1993, and the study cohort was followed through national registries thereafter.
Wang 2014	Post‐trial follow‐up of the Physicians' Health Study II. The vitamin E and vitamin C treatment ended in 2007, and observational follow‐up continued through June 2011.
Willett 1984	Randomised placebo‐controlled trial of chemoprevention in healthy female workers, but the outcome was plasma retinol level (intermediate endpoint)
Woodson 1999	Observational post‐trial analysis
Xuan 1991	Feasibility study; a study to determine the feasibility of conducting a large‐scale, lung cancer chemoprevention trial
Yu 1990	Feasibility study; pilot study