Dietary calcium supplementation for preventing colorectal cancer and adenomatous polyps

Abstract

Background

Several dietary factors have been considered to be involved in the increasing incidence of colorectal cancer in industrialised countries. Experimental and epidemiological evidence has been suggestive but not conclusive for a protective role for high dietary calcium intake. Intervention studies with colorectal cancer as an endpoint are difficult to perform owing to the large number of patients and the long follow‐up required; studies using the appearance of colorectal adenomatous polyps as a surrogate endpoint are therefore considered in reviewing the existing evidence.

Objectives

This systematic review aims to assess the effect of supplementary dietary calcium on the incidence of colorectal cancer and the incidence or recurrence of adenomatous polyps.

Search methods

We searched the Cochrane Controlled Trials Register, the Cochrane Colorectal Cancer Group specialised register, MEDLINE, Cancerlit , and Embase, to December 2009. The reference lists of identified studies were inspected for further studies, and the review literature was scrutinized.

Selection criteria

Randomised controlled trials of the effects of dietary calcium on the development of colonic cancer and adenomatous polyps in humans are reviewed. Studies of healthy adults and studies of adults at higher risk of colon cancer due to family history, previous adenomatous polyps, or inflammatory bowel disease were considered; data from subjects with familial polyposis coli are excluded. The primary outcomes were the occurrence of colon cancer, and occurrence or recurrence of any new adenomas of the colon. Secondary outcomes were any adverse event that required discontinuation of calcium supplementation, and drop‐outs before the end of the study.

Data collection and analysis

Two reviewers independently extracted data, assessed trial quality and resolved discrepancies by consensus. The outcomes were reported as odds ratios (OR) with 95% confidence intervals (CI). The data were combined with the fixed effects model.

Main results

Two studies with 1346 subjects met the inclusion criteria. Both trials were well designed, double ‐ blind, placebo controlled trials, included participants with previous adenomas. The doses of supplementary elemental calcium used were 1200 mg daily for a mean duration of 4 years, and 2000 mg/day for three years.The rates of loss to follow ‐up were 14 % and 11%. 
 For the development of recurrent colorectal adenoma, a reduction was found (OR 0.74, CI 0.58,0.95) when the results from both trials were combined.

Authors' conclusions

Although the evidence from two RCTs suggests that calcium supplementation might contribute to a moderate degree to the prevention of colorectal adenomatous polyps, this does not constitute sufficient evidence to recommend the general use of calcium supplements to prevent colorectal cancer.

Plain language summary

Daily intake of 1 gr dietary calcium may have moderate protective effect on development of colorectal adenomatous polyps

There have been suggestions, based on observational studies and on laboratory markers, that dietary calcium may protect against colorectal cancer. This systematic review of the literature identified two well conducted randomised placebo‐controlled intervention studies involving 1346 subjects followed for 3‐4 years. The results suggest that there may be a moderate protective effect (OR 0.74; CI 0.58,0.95) for dietary supplementation of at least 1200mg elemental calcium per day on the development of colorectal adenomatous polyps. However, no trial has directly demonstrated an effect of calcium supplementation on the development of colorectal cancer itself.

Background

Colon cancer is one of the leading cancers in men and women and it is associated with more industrialised societies; European Union average incidence between 1988‐1992 was 33.8/100000 for men and 23.7/100000 for women. The risk of colorectal cancer begins to increase after the age of 40 and rises sharply from the age of 50‐55 years. 
 Early detection and treatment may have some effect on reducing the mortality from colon cancer but effective primary prevention is the more important public health goal. Several dietary factors have been considered responsible for the changing incidence including calcium, fiber, sugar, fat, vegetables, and meat among others. The underlying mechanisms are not clear, but may in part be due to alterations in bile acids which have been found to be carcinogenic in animals (Nagengast 1995). Dietary calcium has been suggested as a chemoprotective agent against colorectal cancer; it is thought to bind fatty acids and bile acids in the colon thus inhibiting the fat‐induced hyperproliferation of colon epithelial cells (Newmark 1992). The mechanism may be mediated by the precipitation of colonic cytotoxic surfactants thus inhibiting luminal cytotoxicity (van der Meer 1997). Calcium has also been shown to reduce pathological cytokinetic crypt activity primarily in patients with a high familial risk of colorectal cancer (Rozen 1989). Although several animal studies have shown a protective effect of calcium in colon carcinogenesis (e.g. Pence 1993), the empirical evidence from cohort studies (e.g. Giovannucci 1998) and case control studies (e.g. Slattery 1997) is inconsistent on the question whether increased dietary calcium levels do indeed prevent the development of colon cancer in humans. Some epidemiological studies have suggested a protective effect for total calcium intake above 1500 to 1800mg daily (Newmark 1992), but for colorectal adenomas, which are potential precursors of most large bowel cancers, the evidence is conflicting (Bergsma‐Kadijk 1996, Martinez 1998, Hill 1978). The evidence from 24 case‐control and cohort studies and from over 20 epidemiological studies has been reviewed, and failed to find clear support for a significant protective effect of calcium on colon cancer (Bergsma‐Kadijk 1996; Martinez 1998). Calcium intake was associated with a decreased risk of colorectal cancer in a recent meta‐analysis of observational studies (Huncharek 2009). There are only a small number of completed randomised controlled trials, but the little evidence they provide is of better quality than that based on epidemiological studies, which could at best be used to strenghten the evidence from the randomised controlled trials if similar results were obtained. 
 Intervention studies with colorectal cancer as an endpoint are difficult to perform owing to the large number of patients and the long follow‐up required. Since almost all colorectal cancers develop from adenomas, and since 5% of all adenomas develop into cancer (Midgeley 1999), studies regarding the development of colorectal adenomatous polyps as a surrogate endpoint may reasonably be considered in reviewing the existing evidence. Colorectal adenoma could also be a reasonable target for primary prevention .The aim of this review is to assess critically the evidence from completed randomised controlled trials that calcium supplementation may be successful in reducing colon cancer risk. This review will be updated as results from new trials become available.

Objectives

This systematic review and meta‐analysis aims to assess the effect of supplementary dietary calcium on the incidence of colorectal cancer and the incidence or recurrence of adenomatous polyps in adults at different levels of risk, and the development of adverse effects. 
 The hypothesis to be tested is that higher levels of calcium intake have a beneficial effect in the prevention of colonic cancer and adenomatous polyps.

Methods

Criteria for considering studies for this review

Types of studies

Randomised controlled trials of the effects of dietary calcium on the development of colonic cancer and adenomatous polyps in humans are reviewed. The rationale for excluding other experimental designs is that randomised controlled trials provide the best current evidence upon which people can base their decisions; cohort studies are subject to recruitment bias, while case‐control and epidemiological studies of cancer risk are subject to survivor bias and they are also particularly prone to errors in dietary recall. 
 Trials reporting only physiological or laboratory parameters, such as epithelial cell proliferation rates, are not considered clinically relevant and are not reviewed here.

Types of participants

Studies of healthy adults, irrespective of gender or nationality and studies of adults at higher risk of colon cancer due to family history, previous adenomatous polyps, or inflammatory bowel disease are considered. Data from subjects with familial polyposis coli are excluded.

Types of interventions

Studies are included which used supplementation with calcium salts in doses above 1200 mg elemental calcium per day, with a duration of intervention longer than six months. 
 For the control group we accepted either placebo supplementation or no intervention. 
 Trials reporting combined interventions in which there was no arm testing for calcium supplementation alone are not included in this review.

Types of outcome measures

Primary 
 ‐occurrence of colon cancer 
 ‐occurrence of any new adenomas of the colon

Secondary 
 ‐any adverse event that required discontinuation of calcium supplementation 
 ‐drop‐outs (lost to follow up) before the end of the study.

Search methods for identification of studies

Relevant randomised controlled trials were identified by the search strategy described by Dickersin et al (Dickersin 1994). 
 The following search terms were used : calcium, prevention, cancer, neoplasms, colonic, colorectal, polyps, adenomas, adenomatous polyps, human, randomized controlled trials, random allocation, intervention studies. Searches for colo*, adenom*, polyp* were also performed. We searched the Cochrane Controlled Trials Register ( Cochrane 
 Library Issue 4,2009 ), the Cochrane Colorectal Cancer Group (CCCG) specialised register, the MEDLINE ( from 1966 to December 2009 ), Cancerlit ( from 1963 to December 2009 ), Embase ( from 1980 to December 2009 ). The Embase and the CCCG specialised register searches were done according to the search strategy of the CCCG. 
 The reference lists in reports of all identified studies were inspected for further studies. In addition the existing review literature (e. g. Bostick 1997) was scrutinized.

Data collection and analysis

Two reviewers (MAW, AZ) independently reviewed the abstracts of potential studies to be included in the review. For possibly relevant articles or in cases of disagreement between the two reviewers the full article was obtained and inspected independently.Trials were categorised by patient baseline status (healthy versus adenomatous polyps or other risk factors for colonic cancer). The data from the studies were divided, as far as possible from published and unpublished information, into groups according to age at entry. 
 Two reviewers (YY, AZ) independently extracted the data from the included trials. Disagreements were documented and resolved by discussion and the authors of the study were contacted for clarification. Justification for excluding studies from the review were also documented.

Data extracted included: 
 ‐characteristics of trials ‐ publication status, year, country of study, setting, design, inclusion criteria, confounding variables, manner of recruitment, methods, analysis, results. 
 ‐characteristics of participants ‐ study population, number of participants in each group, age, gender, nationality, details on patients' basic bowel status, risk factors, average calcium intake in relation to the recommended daily allowance. 
 ‐characteristics of interventions ‐ calcium preparation used, dose, length of treatment and follow‐up. 
 ‐outcome measures ‐ number of subjects in whom a new adenomatous polyp occured in the intervention and control groups, number of subjects in whom colon cancer occured in both groups, number of subjects who developed adverse or medical events related to the intervention, drop‐outs before the end of study and reasons for dropping out.

Data synthesis 
 Dichotomous data was analysed by calculating the odds ratio for each trial with the uncertainity in each result being expressed as 95% confidence intervals. 
 Results are shown using the approach recommended in the Cochrane Handbook (Clarke 2000). All analyses were performed on the basis of intention to treat. We expected heterogeneity between study results because of different age categories, baseline risk levels, doses and preparations used and duration of studies. Subgroup analysis was planned to assess the impact of these possible sources of heterogeneity. In the absence of heterogeneity the fixed effects model is used appropriately. Associations for lower versus higher levels of calcium intake are reported. 
 The characteristics of eligible trials including those that were excluded and the reasons for exclusion are presented in the Tables. 
 The quality of the reports of the randomised trials was assessed using the method of Jadad et al (Jadad 1996). Concealment of randomisation was assessed according to Cochrane Collaboration standards. Subgroup analyses was planned for each quality item.Two reviewers independently assessed the methodological quality of each study. The studies were assessed for the quantification of concomitant use of other nutrients that have been implicated in colon carcinogenesis. 
 Additional information was obtained from the authors where the publications presented insufficient detail.

Results

Description of studies

See table of " Characteristics of included studies". Two studies, with 1346 subjects assigned to calcium supplements or placebo, met the prestated criteria for inclusion in this review.

The two studies were multicentre trials; one was conducted in ten European countries ‐ Belgium, Denmark, France, Germany, Ireland, Israel, Italy, Portugal, Spain, the UK (Bonithon‐Kopp 2000), and the other study involved six clinical centers in the USA (Baron 1999 a).

Participants

The entry criteria in the two trials were similar, including participants with previous adenomas and therefore considered at high risk for recurrence. Subjects included had a qualifying colonoscopy and polypectomy performed to ensure that the colon or rectum were free of polyps at baseline.

Intervention

Both studies used placebo in the control group. 1200 mg elemental calcium daily for a mean duration of 4 years was administered in one study (Baron 1999 a), and 2000 mg/day for three years in the second study (Bonithon‐Kopp 2000). The mean dietary intake was similar for the groups as stated in the studies.

Outcomes

The included studies reported the number of subjects with at least one recurrent adenoma, or colorectal cancer, and described the drop‐outs and the reasons for leaving the study before the end. Detailed data for the adverse event that required discontinuation of treatment could not be obtained in one study (Bonithon‐Kopp 2000). Information on adverse events that occured during the trials is presented in Table 1.

1. Adverse events.

Study	Intervention	Side effects	Comments
Baron 1999a	calcium carbonate as 1200 mg elemental calcium daily	‐ digestive symptoms: only constipation reported ‐urinary stones in two subjects in the calcium group and one subject in the placebo group	Similar proportions in both groups had digestive symptoms or stopped treatment because of toxicity
Bonithon ‐ Kopp 2000	calcium gluconolactate and carbonate as 2000 mg elemental calcium daily	‐diarrhea severe and abdominal pain reported as major side effects	Major side effects more frecvent in the calcium group than in placebo group, were reported in six versus three subjects respectively. Poor compliance defined as taking less than 80% of the intervention was more frequent in the calcium group Any adverse event ‐14.8% in the calcium group vs. 6.7% placebo group , p = 0.043

Risk of bias in included studies

See table of " Characteristics of included studies". Both studies were well designed, double ‐ blind, placebo controlled trials.

Allocation

Both studies were randomised controlled trials with an adequate generation of allocation sequence. However, only one contained an adequate report of concealment of allocation (Bonithon‐Kopp 2000).

Blinding to intervention

Both trials were double ‐ blinded and the method of blinding was described and was adequate.

Follow ‐ up

In one study 25 subjects were excluded initially after the randomisation for lack of documentation of histologically confirmed adenoma, and total loss to follow ‐up in this trial was 14% (Bonithon‐Kopp 2000). The other study had a 11% loss to follow ‐ up (Baron 1999 a). Both studies described the reasons for drop ‐ out.

A maximum score of 5, according to the Jadad criteria of quality of studies for randomisation, blinding and follow ‐up, was obtained by both trials. However, the assessment for inclusion was based primarily on concealment of allocation.

Effects of interventions

For numerical details see "Analyses"

453 references were identified , the abstracts inspected , and 448 excluded for the following reasons: not randomised, observational studies, study subjects not responding to inclusion criteria, intervention of interest not used, no relevant outcomes reported, reports of biochemical endpoints, repeated reports of the same study, review articles. Five reports were considered potentially eligible for inclusion but, after inspection of the full papers three were excluded; one used a mixture for supplementation with no separate arm for calcium supplementation alone (Hofstad 1998), one was not randomised to the intervention (Hyman 1998), one included data from a similar report of the same study and we included only the paper with the most information (Baron 1999 b). 
 Two studies that included 1595 subjects met the inclusion criteria (Baron 1999 a; Bonithon‐Kopp 2000). In the Bonithon‐Kopp et al trial, there were three arms ‐ calcium, fiber, and placebo. We have analysed only the calcium and placebo arms in this review. Details of the results are presented in the summary analysis.

Primary outcomes:

a)number of subjects with a new diagnosis of colorectal cancer 
 There were too few events for any meaningful conclusion

b)number of subjects with any new adenoma 
 Information for the primary outcome of the number of subjects with at least one recurrent adenoma was found in both trials. When combined, a moderate, statistically significant reduction of recurrence of adenomas favouring the participants receiving calcium was obtained (OR 0.74; CI 0.58,0.95). Individually, only the larger trial (n=930) achieved a statistically significant effect (OR 0.73; CI 0.55,0.96) but not the smaller trial (n=416. OR 0.78; CI 0.45,1.33), but the combined results are strongly influenced by the larger trial (Baron 1999 a).

Secondary outcomes:

a)number of subjects experiencing any adverse event that required discontinuation of treatment 
 One trial (Baron 1999 a) did not find a difference between the treatment and control groups( OR 0.93; CI 0.42,2.05). The report of the other trial did not supply separate data for the 50 subjects who stopped the assigned treatment prematurely because of side effects; the proportions for those who stopped prematurely did not differ significantly between groups (p=0.27) (Bonithon‐Kopp 2000).

b)number of subjects that dropped‐out from the studies 
 No difference could be found for subjects that were lost to follow‐up in the two groups( OR 1.11; CI 0.80,1.55)

There was no evidence of heterogeneity between studies as assessed by inspection of the graphical presentations, therefore the fixed effect model was used for combining the studies. Planned subgroup analyses were not performed; the studies included high risk subjects only and reported that subgroup analyses conducted for age or baseline dietary intake of calcium within the trials did not change the conclusions derived from the overall result. Data for these subgroups were not available for the treatment and control groups separately.

Discussion

The role of calcium in colon cancer prophylaxis has been suspected on the basis of animal experiments, such as that which showed fewer cancers following injection of carcinogens in rats that received calcium supplements in their feed (Adell‐Carceller 1997). In an attempt to examine the question in humans, dietary surveys have been used to identify a corelation between dietary calcium intake and recurrence of colorectal adenomas, and indeed it seemed that calcium might be protective (the rate ratio for the fifth quintile versus the first was 0.63 (95% confidence interval, 0.39‐1.02)) (Hyman 1998). The next step was to see if patients who took extra dietary supplements of their own accord fared better than those who did not, and once again for calcium supplementation there seemed to be a protective effect (odds ratio, 0.51; 95 percent confidence interval, 0.27‐0.96, Whelan 1999). 
 Adenomas rather than cancers were studied because they are so much more common, despite the limited applicability of findings from adnomas to cancer. 
 In order to relate calcium supplementation more closely with actual carcinogenesis, rather than extrapolating from the effect on the development of adenomas, attempts have been made to examine the effects of dietary calcium on human colonic or rectal epithelial proliferation, and some early studies suggested that calcium does indeed impede epithelial proliferation (O'Sullivan 1993). However, more recently, these results have not been confirmed, but it has also become clear that the proliferative index does not predict future colorectal neoplasia, although it may be weakly associated with the presence of current adenomas. The authors conclude that these results have important implications for the design of future intervention studies: "Although it may be attractive to include the measurement of intermediate markers in large controlled trials, until we have more confidence in their performance, we should rely on better proven and more reliable intermediates, such as adenomas" (Sandler 2000). Doubt on the importance of dietary calcium is cast by the results of a large case control study embedded in a major fecal occult blood screening programme ‐ no association of colon cancer was found with reported calcium intake (Little 1993). As indicated in the Background section, many epidemiological geographical studies have been reported, but on systematic review they present an inconsistent and inconclusive picture (Martinez 1998).The findings of several large cohort studies and one case‐control study performed worldwide, in Europe, USA, Hawai, China suggest that an increased intake of calcium may prevent the development of colorectal adenomas and cancer (Flood 2005; Kesse 2005; Larsson 2006; Park 2007; Shin 2006). Among the major potential sources of bias in epidemiological studies of diet and cancer are survivor bias, problems with dietary recall, and difficulties in identifying and controlling for countless confounding variables. We must therefore seek evidence from prospective controlled trials of calcium supplementation. The association between calcium and vitamin D intake and adenomatous polyps recurrence was evaluated and did not find significant results in a randomized controlled trial, where participants were randomized to either intensive counseling to adopt a low‐fat, high fiber, fruit and vegetable enriched eating plan, or to be given a standard brochure on healthy eating (Hartman 2005). 
 A recent large randomized controlled trial, the Women's Health Initiative (WHI) Study (Wactawski‐Wende 2006), including 36282 post‐menopausal women receiving calcium and vitamin D or placebo for an average of seven years, did not demonstrate any effect of the supplements on the incidence of colon cancer. However, we did not include this study in our review for a number of reasons. First, the WHI study used combined calcium and vitamin D supplements and the independent effect of vitamin D on prevention of colon cancer is not known. Second, the colon was not evaluated for adenomas or cancer at baseline ; the study did not use clear methods of diagnosing colon cancer, unlike the studies included in our review which used colonoscopy for evaluation. The studies included in our review also used higher doses of calcium. In future, a larger trial using the doses of calcium used in the studies we included, without vitamin D and for a longer follow‐up period may yield different results. 
 Colorectal cancer itself is sufficiently rare that it is extremely difficult to conduct a randomised controlled trial large enough and long enough to isolate the protective effect of any given single nutrient or nutritional supplement. That is why we have to rely on the indirect evidence supplied by studying the development of the far more common colorectal adenomatous polyps, some of which may be pre‐cancerous. There are major difficulties in the conduct of such trials, such as concommitent use of other supplements that might also affect carcinogenesis, fibre, anti‐oxidants and others. The effects of other nutrients have been reviewed elsewhere ‐ e.g. fiber (Asano 2002). The trials must also take into account regular dietary calcium intake as well as the supplement, and the different baseline risk of colorectal cancer in different people ‐ those with inflammatory bowel disease, family history and familial polyposis and previous history of colon neoplasia. To date only two groups have succeeded in completing such trials, one in North America and one in Europe. 
 These two trials were conducted well and produced results that suggest a clinically relevant protective effect of dietary calcium supplementation on the development of colorectal adenomatous polyps. Only in the larger, American, trial did the effect reach statistical significance. It is difficult to assess which factors contributed to the small difference in the results of the studies. Baron et al (Baron 1999 a; Baron 1999 b) used a lower dose of calcium (1200mg/day versus 2000mg/day in the smaller European trial), but for a longer period (4 years versus 3 years). It was larger, achieving an adequate sample size, unlike Bonithon‐Kopp et al. (Bonithon‐Kopp 2000) whose smaller sample size reduced the power of the study with a lack of statistical significance. There were more events in the control group in Baron et al, which could be attributed to the longer follow‐up. The adequacy of concealment was only documented in Bonithon‐Kopp et al. Otherwise, the methodological differences between the two trials were also so slight that no meaningful sub‐analysis was possible in order to try and explain the differences in the results. 
 Within‐trial subgroup analyses for age, sex, and compliance showed no significant differences in the results for both trials. Despite the relatively large numbers of subjects who successfuly completed the trials, there was not enough data to perform sub‐group analyses for other risk factors. Recent studies based on subsets of subjects from the original trials included in the review found associations between colorectal adenoma characteristics and the risk of recurrence. The number of adenomas and their proximal location at baseline were the main predictors of recurrence; patients with three or more adenomas with at least one of them located on the proximal colon had a much higher risk of overall recurrence, proximal recurrence and advanced adenoma recurrence ( Bonithon‐Kopp 2004). Calcium supplementation may have a more pronounced antineoplastic effect on advanced colorectal lesions than on other types of polyps (Wallace 2004). Multiple adenomas and the presence of large or advanced adenomas have been identified as risk factors for adenoma recurrence( Martinez 2001). Results also suggest that the protective effect of calcium could be greater when serum levels of 25‐ hydroxyvitamin D are in the higher range( Grau 2003). 
 Only subjects at high risk due to previous adenomas were included in the trials, so that we may conclude that the intervention could be applied to this category of population, if confirmed as effective by further studies. Perhaps a higher dose of calcium for a longer period is needed, but these doses may be less tolerable and the risk‐benefit ratio must be weighed. Studies on healthy adults not considered at high risk would be difficult to conduct considering the problems of recruitment of sufficient subjects, the necessity of a qualifying and subsequent follow‐up colonoscopies, as well as the long‐term adherence to treatment requested from a healthy adult. 
 In general, where we find weak clinical effects it is interesting to look for reasons to explain why only some individuals respond to the therapeutic intervention and not others. The genetic propensity to colon cancer has been studied in relation to calcium intake and it seems that people with the BB BsmI genotype for the vitamin D receptor may be at reduced risk of colorectal adenoma. However, they seem to be less at risk when exposed to lower rather than higher levels of calcium and vitamin D intake. So the direction of the effect is opposite to that we expect from previous observations, and more work is needed to identify and explain any heterogeneity of response (Kim 2001). 
 Calcium supplementation is attractive as a potential intervention to reduce the risk of colorectal adenomas and colorectal cancer; it is relatively cheap and readily 
 available. Although it is likely to be safe, this, and its efficacy would need to be more clearly demonstrated in further controlled studies before any attempt at widespread introduction into clinical practice. Given the small number of studies eligible for inclusion in this review, and the conflicting evidence derived from them, no better conclusions can be drawn at this stage than those resulting from the vast amount of epidemiological data.

Authors' conclusions

Implications for practice.

Although the evidence from two RCTs suggests that calcium supplementation might contribute to the prevention of colorectal adenomatous polyps, this does not constitute sufficient evidence to recommend the general use of calcium supplements to prevent colorectal cancer in healthy adults or in adults with a previous history of adenomatous polyps. There is however no evidence that such supplements would be detrimental in doses of up to 2 grams elemental calcium per day.

Implications for research.

Large multicentre trials could have the power to answer the outstanding questions ‐ whether calcium supplementation has a protective effect in patients at particularly high risk, or in totally healthy subjects. Further work is needed on the factors involved in individual susceptibility to colorectal cancer in order to identify any subgroups that might benefit from calcium supplementation. 
 Further trials with an adequate sample size and duration need to be conducted to confirm the effects of this intervention, including assessments of dose‐response relationships and the risk‐ benefit balance.

What's new

Date	Event	Description
5 January 2010	New search has been performed	Number of studies found on last searches was added . No new studies were found for inclusion

History

Protocol first published: Issue 3, 2002
 Review first published: Issue 1, 2004

Date	Event	Description
23 July 2008	Amended	Converted to new review format.
7 September 2007	New citation required and conclusions have changed	Substantive amendment

Notes

This is the third update of the review. The searches for new trials were performed December 2009, but didn't identify any new trials to be included.

Data and analyses

Comparison 1. Calcium versus placebo.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Number of subjects diagnosed with colorectal cancer	2	1346	Odds Ratio (M‐H, Fixed, 95% CI)	0.34 [0.05, 2.15]
2 Number of subjects with at least one new adenoma	2	1346	Odds Ratio (M‐H, Fixed, 95% CI)	0.74 [0.58, 0.95]
3 Number of subjects with at least one adverse event requiring discontinuation of treatment	1	930	Odds Ratio (M‐H, Fixed, 95% CI)	0.93 [0.42, 2.05]
4 Number of subjects that dropped ‐ out	2	1346	Odds Ratio (M‐H, Fixed, 95% CI)	1.11 [0.80, 1.55]

1.1. Analysis.

Comparison 1 Calcium versus placebo, Outcome 1 Number of subjects diagnosed with colorectal cancer.

1.2. Analysis.

Comparison 1 Calcium versus placebo, Outcome 2 Number of subjects with at least one new adenoma.

1.3. Analysis.

Comparison 1 Calcium versus placebo, Outcome 3 Number of subjects with at least one adverse event requiring discontinuation of treatment.

1.4. Analysis.

Comparison 1 Calcium versus placebo, Outcome 4 Number of subjects that dropped ‐ out.

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Baron 1999 a.

Methods	multicentre Randomisation : computerised generation of allocation sequence, blocked according to study center ( 2 )* Allocation concealment : not mentioned Double‐ blind : yes ( 2 ) Intention to treat : not mentioned 89% included in the main analysis from the main study period. Every six months questionnaires on adherence to intervention, medication,nutritional supplements Two periods included in the study : from qualifying and including the first f / u colonoscopy (at 1 year) and the main study period after the first and including the second f / u colonoscopy ( at 4 years) f / u described ( 1 ) 89% completed study Design : parallel Jadad score = 5
Participants	N =930 670 ‐ men 260 ‐women Mean age (SD) : 61 (9) yrs with recent history of colorectal adenomas confirmed histologically removed within 3 months before recruitment , free of polyps on complete colonoscopy Inclusion criteria ‐ < 80 yrs ‐in good health Exclusion criteria ‐FAP, CRC ‐malabsorbtion ‐conditions worsened by calcium Baseline characteristics and dietary pattern similar in both groups, no significant differences. Less than 3% taking calcium supplements at the start of the trial discontinued it
Interventions	Treatment group: calcium carbonate 3g (1200 mg elemental calcium ) daily N = 464 Control group: identical placebo N = 466 Mean duration : 4 yrs Diet assessed at enrollment and at the end of the study by a validated food questionnaire Mean daily dietary calcium intake 865+/‐ 423 mg in Tx group 889+/‐ 451 in C group Compliance assessed as percentage of tablets taken, similar in both groups. > 80% of subjects taking study agents > 90% of the time at 4 yrs
Outcomes	‐the proportion of subjects in whom at least one adenoma was detected after the first f / u and including the second f / u colonoscopy ‐the average number of adenomas in both groups All polyps removed were histologically diagnosed
Notes	Recruitment 1988 ‐ 1992 Six clinical centers in USA 913 subjects underwent at least one f / u colonoscopy. 832 ( 89%) included in the main analysis 409‐ Tx group 423‐ C group and had two f / u colonoscopies with polyps removed. Interim colonoscopies only when necessary Drop‐outs: 55 (11, 8%)‐ Tx group: 22 died ,11lost interest, 8 ill or moved, 2 unknown /other 43 (9,2%)‐ C group: 25 died, 14 lost interest, 10 ill /moved, 6 other/unknown
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Bonithon‐Kopp 2000.

Methods	multicentre Randomisation : stratified by center balanced every six patients 2calcium, 2fibre, 1placebo calcium, 1placebo fibre ( 2 ) Allocation concealment : independent randomisation center Double‐blind: yes ( 2 ) Intention to treat: yes f / u described (1) f / u complete colonoscopy at 3 yrs 86% Design: parallel Jadad score= 5
Participants	N =640 Mean age (SD) : 58.8 ( 8.8 ) yrs Tx group, 59.3 ( 8.4 ) yrs C group, with history of colorectal adenomas. Polyps removed at qualifying colonoscopy. Inclusion criteria ‐at least two adenomas or one > 5mm on complete colonoscopy confirmed by pathologist ‐ age 35‐75 ‐ no debilitating disease Exclusion criteria ‐ FAP, IBD, CRC, colonic resection ‐contraindication to calcium / fibre ‐current calcium treatment that could not be stopped Baseline characteristics and mean dietary intake similar in both groups, no significant differences
Interventions	Treatment group: ‐ calcium gluconolactate and carbonate PO twice daily (2000 mg elemental calcium daily) N =204 ‐ fibre ‐isphagula 3.5g per day N =224 Control group: placebo sucrose with taste, appearance and excipient as the intervention N =212 Diet assessed by a standard validated questionnaire at baseline and at 3 yrs Mean calcium intake: 944 (364) mg ‐ calcium group 1023 (404) mg C group, p =0.28 Duration‐ 3 yrs Compliance assessed by unused sachets returned, standard interview, fecal calcium. 69% in Tx and 82% in C groups, p =0.044 took > 80% of the sachets prescribed
Outcomes	‐ at least one new adenoma at 3 yr examination All removed polyps were pathologically diagnosed
Notes	21 centres from ten countries Recruitment 1991‐ 1994 25 subjects excluded initially after the randomisation because not histologically confirmed adenomas (665 subjects initially) Randomisation after complete qualifiyng colonoscopy Interim colonoscopies only if necessary, with polyps removed counted as endpoints, only if removed >12 months after the qualifying colonoscopy ( 1 subject not counted only and this had a hyperplastic polyp). The proportions of final colonoscopies similar between groups p =0.41 Calcium group: 176 subjects completed study, 8 died, 2 severe illness, 5 lost to f/u 13 refused to continue Control group: 178 completed study, 9 died, 1 severe illness, 5 lost to f/u, 19 refused to continue
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Low risk	A ‐ Adequate

*( ) number of points obtained in the Jadad scale; FAP familial adenomatous polyposis; IBD inflammatory bowel disease 
 CRC colorectal cancer; SD standard deviation; yr year; Tx treatment; C control; f /u follow ‐up

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Baron 1999 b	Included participants and data from a similar report of the same study
Hofstad 1998	Randomised, parallel Participants: N= 116 polyp bearing subjects Intervention: mixture of calcium and antioxidants vs. placebo Not separate arm for calcium supplementation
Hyman 1998	Not randomised for calcium supplementation; assessment based on a questionnaire .

Contributions of authors

The original idea for this review was Professor Berry's, of The Hebrew University, Jerusalem, but he withdrew as a co‐author for lack of time. 
 Dr Zalmanovici did most of the work, selecting the studies to be included, and assessing their quality. She also entered all the data into Revman and wrote the Background, Methods and Results sections. 
 Professor Weingarten supervised the project, was the contact for CCCG and for the authors of the trials included in the review. He was also responsible for the literature search, and was an independent assessor in selecting the trials to be included. He edited the review and wrote the Discussion. 
 Dr John Yaphe assessed the selected reviews and criticized and improved the review

Declarations of interest

None known

New search for studies and content updated (no change to conclusions)