supplemental calcium in the chemoprevention of colorectal cancer: a systematic review and...

Clinical Therapeutics/Volume 32, Number 5, 2010

May 2010 789

This work was presented in part at the 6th Annual Meeting of Health Technology Assessment International, June 21–24, 2009, Singapore, Singapore.

Accepted for publication April 23, 2010.doi: 10.1016/j.clinthera.2010.04.0240149-2918/$ - see front matter

© 2010 Excerpta Medica Inc. All rights reserved.

ABSTRACTObjective: The aim of the review was to assess the

evidence for the effectiveness of calcium in reducing the recurrence of adenomas and the occurrence of colorectal cancer among populations at high, intermediate, and low risk of the disease.

Methods: A systematic review of randomized con-trolled trials (RCTs) was performed to compare calcium alone, and with other agents, versus placebo. Nine data-bases (Cochrane Library, MEDLINE, PreMEDLINE, CINAHL, EMBASE, Web of Science, Biological Abstracts, the National Research Register, and Current Controlled Trials) were searched for published and unpublished trials. Searches were not restricted by either language or date of publication. All searches were completed in January 2010. Database thesaurus and free text terms for calcium and adenomas or colorectal cancer were used to search for trial reports; additional terms were used to search for other agents of interest, such as NSAIDs and folic acid. Search terms consisted of a combination of terms for colorectal cancer (eg, colon or colorectal and neoplasm or cancer or adenoma) and terms for calcium and RCTs. The initial searches were conducted in June 2008, with update searches in January 2010 to identify more recent studies. The reference lists of relevant studies were also searched for additional papers not identified by the search of electronic databases. Studies had to satisfy the following criteria to be included: RCTs about calcium, with or without other chemopreventive agents, in adults with familial adenomatous polyposis (FAP), hereditary nonpolyposis colorectal cancer, or a history of colorectal adenomas, or with no increased baseline risk of colorectal cancer. Meta-analysis was performed. For discrete and numerical outcomes, relative risks (RRs) and risk differences were reported with 95% CIs. The random-effects model was used to account for clinical and methodologic variations between trials.

Results: The original and update searches of electronic databases produced 3835 citations, of which 6 studies (8 papers) met the inclusion criteria. Supplemental calcium had no effect on the number of adenomas in 1 small trial of patients with FAP. Meta-analysis of 3 trials in individu-als with a history of adenomas showed a statistically significant reduction in the RR for adenoma recurrence (RR = 0.80 [95% CI, 0.69–0.94], P = 0.006) for those receiving calcium 1200 to 2000 mg/d, but no effect was seen in advanced adenoma (RR = 0.77 [95% CI, 0.50–1.17], P = NS). Meta-analysis of 2 trials in populations with no increased baseline risk for colorectal cancer suggested that calcium, with or without vitamin D, had no effect on the RR for colorectal cancer (RR = 0.62 [95% CI, 0.11–3.40], P = NS).

Conclusion: Published reports indicated that supple-mental calcium was effective for the prevention of adeno-ma recurrence in populations with a history of adenomas, but no similar effect was apparent in populations at higher or lower risk. (Clinicaltrials.gov identifier: NCT00486512. (Clin Ther. 2010;32:789–803) © 2010 Excerpta Medica Inc.

Key words: adenoma, calcium, colorectal cancer, recurrence, risk.

INTRODUCTIONColorectal cancer is a malignant neoplasm arising from the lining of the large intestine and is the second most common fatal cancer in western Europe and the United

Supplemental Calcium in the Chemoprevention of Colorectal Cancer: A Systematic Review and Meta-Analysis

Christopher Carroll, PhD1; Katy Cooper, PhD1; Diana Papaioannou, MSc1; Daniel Hind, PhD2; Hazel Pilgrim, MSc1; and Paul Tappenden, MSc1

1Health Economics and Decision Science, School of Health and Related Research, University of Sheffield, Sheffield, United Kingdom; and 2Clinical Trials Research Unit, School of Health and Related Research, University of Sheffield, Sheffield, United Kingdom

790 Volume 32 Number 5

Clinical Therapeutics

or family history of colorectal polyps). Two previous reviews have examined the effect of supplemental calcium on adenoma recurrence in individuals with a history of adenomas,24,25 but a search of the literature identified no other systematic review or meta-analysis of random-ized controlled trials (RCTs) that addressed this question for all relevant populations in terms of colorectal cancer, as well as for the outcomes of advanced adenomas and colorectal cancer.

METHODSA literature search was performed to identify relevant research using database thesaurus and free text terms for calcium and adenomas or colorectal cancer. A vali-dated study design filter to identify RCTs was used.26

This search also included terms for other agents of interest, such as NSAIDs and folic acid, because this review was conducted, in part, as one element of a larger assessment of numerous potential chemopreventive agents for colorectal cancer. Nine databases were searched for published and unpublished trials: Cochrane Library, MEDLINE, PreMEDLINE, CINAHL, EMBASE, Web of Science, Biological Abstracts, the National Research Register, and Current Controlled Trials. Search terms consisted of a combination of terms for colorectal cancer (eg, colon or colorectal and neoplasm or canceror adenoma) and terms for calcium and RCTs. Searches were not restricted by either language or date of pub-lication. The initial searches for all relevant agents were conducted in June 2008, with update searches performed in January 2010 to identify any more recent calcium studies. The reference lists of relevant studies were also searched for additional papers not identified by the search of electronic databases.

Studies had to satisfy the following criteria to be included in the review: RCTs of calcium (with or without other chemopreventive agents) in adults with FAP, HNPCC, or a history of colorectal adenomas, or with no increased baseline risk of colorectal cancer. Relevant comparators were specified as either placebo or agents other than calcium. Relevant outcomes included the recurrence of adenomas or advanced adenomas, or the occurrence of colorectal cancer. Institutional review board approval and consent were not inclusion criteria for this review.

All citations identified by the searching process were screened by 1 of 3 reviewers (C.C., K.C., or D.P.) to determine whether they met the inclusion criteria. For quality-control purposes, a double check for appropriate

States.1,2 Incidence of colorectal cancer increases with age, with the median age at diagnosis being between 70 and 74 years for both colon and rectal cancer pa-tients.3,4 Risk factors for the development of colorectal cancer include genetic syndromes, family or individual history of adenomas, and environmental factors related to diet and lifestyle.5,6 Approximately 5% of colorectal cancers are associated with the genetic syndromes fa-milial adenomatous polyposis (FAP) and hereditary nonpolyposis colorectal cancer (HNPCC, or Lynch syndrome),7,8 and up to 20% occur in individuals who have a family history of the disease but for whom no specific disease-causing mutations can be identified.9,10

The occurrence of colorectal cancer is otherwise sporadic, with the remaining 75% of patients having neither a clear family history nor any known predisposing condition.11

Colorectal cancer typically develops from adenoma-tous polyps arising from the lining of the intestine; indirect evidence suggests that adenomas may be present for ≥10 years before malignancy develops.12–14 The size and number of adenomas, as well as their histologic type and the presence of epithelial dysplasia, are thought to affect the risk of colorectal cancer development. Individuals in whom adenomatous polyps are identified undergo polypectomy (ie, removal of polyps) and are invited to attend routine endoscopic surveillance.15 The overall 5-year survival rate for colorectal cancer in the United Kingdom is ~50%, but it varies according to the stage of disease at diagnosis.16 The clinical effectiveness of several drug and micronutrient interventions for the prevention of colorectal cancer, its precursor (adenoma-tous polyps), or both in populations at differing risks for being diagnosed with colorectal cancer has been investigated and synthesized in a number of systematic reviews.17 These interventions include NSAIDs, folic acid, and antioxidants (eg, vitamin A, vitamin C, vitamin E, selenium, β-carotene).18–23 Aspirin and celecoxib have been found to have some chemopreventive effect in popu-lations with a history of colorectal adenomas,17,18 but the other agents appear to have no significant benefit and, in the case of folic acid, may even be harmful.21

Therefore, the purpose of the present work was to systematically review the existing evidence concerning the clinical effectiveness of calcium in reducing the recurrence of colorectal adenomas in populations with FAP/HNPCC or a history of adenomas (ie, with a high or intermediate risk), and in reducing the occurrence of colorectal cancer in the general population with no in-creased baseline risk of the disease (ie, no known personal

May 2010 791

C. Carroll et al.

identified29; 3 RCTs examined calcium in populations with a history of adenomas,31–34 and 2 RCTs examined calcium in general populations, with no baseline risk of colorectal cancer.36 A single ongoing study was identi-fied assessing aspirin, calcium, and vitamin D (as a com-bination treatment) versus placebo for prevention of adenomas in participants with a history of adenomas.

Individuals with FAPOne eligible controlled trial was identified comparing

calcium with placebo (Table I).29 This single-center study randomized 28 participants with FAP, who had previously undergone total abdominal colectomy with ileorectal anastomosis, and who were under surveillance for rectal polyps. Subjects in the intervention arm were given calcium carbonate 1500 mg/d. The duration of treatment and follow-up was 6 months. The report-ed trial was judged to be low quality because both the methods by which the allocation sequence was gener-ated and concealed, and the numbers randomized to and analyzed in each arm, were not reported. The pub-lished paper also did not report outcome event data. In the absence of all of these data, it was not possible to calculate the difference between the mean change in the number of rectal polyps for participants receiving cal-cium and placebo. The authors were contacted, but they were unable to provide the necessary data. The publica-tion reported that the effect of oral calcium on polyp number was not statistically significant.

Individuals with a History of AdenomasThree studies compared calcium with placebo (Table

I).31–34 All 3 were multicenter trials. The mean age of participants in 2 studies ranged from 59 to 61 years31,32; in the Hofstad et al study,33,34 the age range was 50 to 76 years. The Baron et al30,31 (Calcium Polyp Prevention Study) and Bonithon-Kopp et al32 trials compared calcium at 1200 and 2000 mg/d, respectively, with placebo. The Hofstad et al study administered a mixed intervention consisting of calcium 1600 mg/d, β-carotene 15 mg/d, vitamin C 150 mg/d, vitamin E 75 mg/d, and selenium 101 µg/d, compared with placebo. The dura-tion of treatment and the follow-up of participants ranged from 3 to 4 years. All participants underwent colonoscopy at baseline; subsequent colonoscopies were then undertaken to identify number and type of ade-nomas at follow-up. In the Baron et al and Bonithon-Kopp et al trials, all polyps identified at baseline were removed, but in the Hofstad et al trial, all polyps identi-

inclusion and exclusion was performed on 10% of the citations by a fourth reviewer (D.H.), and a κ statistic of 0.76 was recorded for interrater reliability. Full papers were retrieved for all citations that appeared to meet the inclusion criteria. The full paper was also retrieved to make a definitive judgment in cases where a decision could not be made about inclusion on the basis of title or abstract alone. The inclusion and exclusion criteria were then used to assess full papers. Queries concerning inclusion were resolved by discussion and consensus between ≥2 reviewers. Data from all included trials were extracted using a form designed specifically for this review and piloted on a sample paper. One reviewer (C.C.) extracted data from the final list of included studies into predesigned tables and appraised the quality of the included studies using a form based on standard criteria for RCTs.27 Both data extraction and quality assessment were then checked thoroughly by a second reviewer (K.C.). The aim of the quality assessment was to evaluate issues relating to the quality of the reported processes of allocation, randomization and blinding, the comparability of the treatment and control groups, and the appropriateness and quality of the analysis per-formed. The quality assessment process was undertaken to afford an idea of the respective quality of studies and to inform the internal validity of the review.

Meta-analysis of trials was performed using Review Manager, version 5.0 (Cochrane Collaboration, Copen-hagen, Denmark). For discrete and numerical outcomes, relative risks (RRs) and risk differences (RDs) were reported with 95% CIs. The random-effects model was used to account for clinical and methodologic variations between trials. Statistical heterogeneity was described using the I2 statistic. Only randomized participants for whom a valid outcome had been evaluated and reported were included in the analysis. Two types of comparison were presented: calcium alone versus placebo alone; and calcium with or without other chemopreventive agents versus placebo (with or without other interventions). The different population groups were not pooled in the analyses because their different levels of risk were likely to affect the impact of the intervention on outcomes.28

RESULTSThe original and updated searches of electronic databases produced 3835 citations, of which 8 articles (6 trials) satisfied the inclusion criteria (Figure 1).29–36 No ad-ditional references were found from other sources. One RCT examining calcium in individuals with FAP was



formed, in which the unpublished data from the Hofstad et al34 trial were excluded. The exclusion of this trial was justified for a number of reasons. First, the trial was smaller and of lower quality than the other 2 trials, and was therefore prone to a greater degree of bias.37

Second, the intervention was calcium plus antioxidants (vitamins C and E, β-carotene, and selenium) rather than calcium alone. These cointerventions may have also had an interaction effect with calcium, which could have further confounded the results.22,38

Meta-analysis of only the Baron et al30,31 and Bonithon-Kopp et al32 trials (with a total of 1186 par-ticipants in analysis) indicated a slightly lower 18% reduc-tion in the RR for being diagnosed with an adenoma of any type for those taking calcium rather than placebo (RR = 0.82 [95% CI, 0.69 to 0.98], P = 0.03) (Figure 3), and a 6% absolute risk reduction (RD = –0.06 [95% CI, –0.11 to –0.01], P = 0.03) (Table II). This significant effect was not apparent for the relative or absolute risks of being diagnosed with advanced adenoma (RR = 0.77 [95% CI, 0.50 to 1.17], P = NS; RD = –0.01 [95% CI, –0.05 to 0.03], P = NS) (Figure 4). Meta-analysis of the Baron et al and Bonithon-Kopp et al trials also found

fied were left in situ, and only new polyps were reported as outcome events at the follow-up colonoscopy.

The Baron et al31 and Bonithon-Kopp et al32 trials were both judged to be good quality. In both studies, alloca-tion concealment was adequate; a central, computer-generated randomization sequence was used, and meth-ods of blinding were adequate. The Hofstad et al34 trial was of lower quality: methods of allocation, randomiza-tion, and blinding were all unclear. Event data were also unpublished, but were provided by the author on request. In all 3 studies, between 10% and 20% of randomized participants were excluded from the analyses because of death, failure to have a colonoscopy, or loss to follow-up, but intent-to-treat analyses and power calculations were performed by all 3 studies and the required sample size was achieved.

Meta-analysis of all 3 trials (1279 participants in analysis) indicated a 20% reduction in the RR for being diagnosed with an adenoma of any type for those taking calcium rather than placebo (RR = 0.80 [95% CI, 0.69 to 0.94], P = 0.006) (Figure 2), and a 6% absolute risk reduction (RD = –0.06 [95% CI, –0.11 to –0.02], P = 0.009) (Table II). A sensitivity analysis was also per-

Unique citations retrieved by search of electronic databases and

reference tracking(n = 3835)

Citations satisfied inclusion criteria(n = 8)

Populations with a history of adenomas:

3 RCTs (5 articles)

Populations with no history of adenomas or colorectal cancer:

2 RCTs (2 articles)

Populations with FAP: 1 RCT (1 article)

Excluded studies not relevant for data extraction

(n = 3827)

Figure 1. Flow diagram of randomized controlled trials (RCTs) comparing supplemental calcium, with or without other agents, versus placebo for effectiveness in reducing the recurrence of adenomas and the occurrence of colorectal cancer among populations at high, intermediate, and low risk of the disease. FAP = familial adenomatous polyposis.

May 2010

793

C. C

arroll et al.

Table I. Characteristics of trials comparing supplemental calcium, with or without other agents, versus placebo for effectiveness in reducing the recurrence of adenomas and the occurrence of colorectal cancer among populations at high, intermediate, and low risk of the disease.

Study (Year)Study

DesignPopulation

and Age Intervention ControlTreatment Duration

Follow-Up Duration

FAP populations Thomas et al (1993)29

DB, RCT

28 FAP patients with previous colectomy and adenomas, aged 16–65 years (median, 38 years)

Calcium carbonate 1500 mg/d (number of patients not reported)

Placebo (number of patients not reported)

6 Months 6 Months

Populations with an increased risk of colorectal cancer (populations with a history of adenomas) Baron et al (1999)30,31

DB, RCT

History of adenomas, aged ≤80 years (mean, 61 years)

Calcium 1200 mg/d (n = 464) Placebo (n = 466)

4 Years 4 Years (from end of year 1 to end of year 4)

Bonithon-Kopp et al (2000)32

DB, RCT

History of adenomas, aged 35–75 years eligible (mean, 59 years)

Calcium 2000 mg/d (n = 204) Placebo (n = 212)

3 Years 3 Years

Hofstad et al (1998)34

DB, RCT

History of adenomas (polyps left in situ), aged 50–76 years

Calcium 1600 mg/d + β-carotene 15 mg/d + vitamin C 150 mg/d + vitamin E 75 mg/d + selenium 101 µg/d (n = 42 analyzed)

Placebo (n = 51 analyzed)

3 Years 3 Years

General populations and/or populations with an average risk of colorectal cancer Lappe et al (2007)35

RCT Postmenopausal women, aged ≥55 years

Calcium 1400–1500 mg/d (n = 445) or calcium 1400–1500 mg/d + vitamin D 1000–1100 IU/d (2 different doses of vitamin D in abstract and text) (n = 446)

Placebo (n = 288)

4 Years 4 Years

Wactawski-Wende et al (2006)36

DB, RCT

Postmenopausal women, aged 50–79 years (mean, 59 years)

Calcium 1000 mg/d + vitamin D 400 IU/d (n = 18,176)

Placebo (n = 18,106)

7 Years 7 Years

RCT = randomized controlled trial; FAP = familial adenomatous polyposis; DB = double-blind.

794 V

olume 32 N

umber 5

Clinical T

herapeutics

Calcium + Antioxidants Placebo

Weight, %

Random Risk Ratio

(95% CI)Random Risk Ratio

(95% CI)Study or Subgroup Adenomas Total Adenomas Total

Baron et al (1999)30,31 127 409 159 423 68.3 0.83 (0.68–1.00)Bonithon-Kopp et al (2000)32 28 176 36 178 12.2 0.79 (0.50–1.23)Hofstad et al (1998)34 21 42 35 51 19.5 0.73 (0.51–1.04)

Total 627 652 100.0 0.80 (0.69–0.94)Total events 176 230Heterogeneity: t2 = 0.00, x2 = 0.39, df = 2 (P = 0.82), I2 = 0%Test for overall effect: z = 2.77 (P = 0.006)

Figure 2. Results of meta-analysis comparing supplemental calcium, with or without antioxidants, versus placebo for effectiveness in reducing the recurrence of any adenoma among populations with a history of adenomas.

20.5 1 5Favors placeboFavors calcium

0.2

May 2010

795

C. C

arroll et al.

Table II. Summary of meta-analyses comparing supplemental calcium, with or without other agents, versus placebo for effectiveness in reducing the recurrence of adenomas and the occurrence of colorectal cancer among populations at intermediate and low risk of the disease.

Variable Intervention ControlFollow-Up,

y

Event Rates: Intervention vs

ControlRR

(95% CI)I2, %

RD (95% CI)

I2, %

History of adenomasAny adenoma30–34

Baron et al (1999)30,31


Hofstad et al (1998)34

Calcium 1200–2000 mg/d with or without antioxidants

Placebo alone

3–4 *176/627 vs 230/652

0.80†

(0.69 to 0.94)0 –0.06†

(–0.11 to –0.02)0

Baron et al (1999)30,31


Calcium 1200–2000 mg/d alone

Placebo alone

3–4 155/585 vs 195/601

0.82‡

(0.69 to 0.98)0 –0.06‡

(–0.11 to –0.01)0

Advanced adenoma Baron et al (1999)30,31



Placebo alone

3–4 35/585 vs 47/601

0.77(0.50 to 1.17)

0 –0.01(–0.05 to 0.03)

0

Colorectal cancer Baron et al (1999)30,31



Placebo alone

3–4 1/668 vs 4/678

0.34(0.05 to 2.14)

0 0.00(–0.01 to 0.00)

0

No increased baseline risk of colorectal cancerColorectal cancer Lappe et al (2007)35


Calcium 1000–1500 mg/d + vitamin D 400–1100 IU/d

Placebo alone

4 or 7 169/18,622 vs 156/18,394

1.08(0.87 to 1.34)

0 0.00 (0.00 to 0.00)

0

Lappe et al (2007)35


Calcium 1000–1500 mg/d ± vitamin D 400–1100 IU/d

Placebo alone

4 or 7 169/19,067 vs 156/18,394

0.62(0.11 to 3.40)

68 0.00(0.00 to 0.01)

39

RR = relative risk; RD = absolute risk difference; I2 = measure of heterogeneity.*Includes unpublished data provided by authors of Hofstad et al33 study report. † P < 0.01. ‡ P < 0.05.

796 V

olume 32 N

umber 5

Clinical T

herapeutics

Calcium PlaceboWeight,

%

Random Risk Ratio



Baron et al (1999)30,31 127 409 159 423 84.8 0.83 (0.68–1.00)Bonithon-Kopp et al (2000)32 28 176 36 178 15.2 0.79 (0.50–1.23)

Total (95% CI) 585 601 100.0 0.82 (0.69–0.98)Total events 155 195Heterogeneity: t2 = 0.00, x2 = 0.04, df = 1 (P = 0.84), I2 = 0%Test for overall effect: z = 2.23 (P = 0.03)

Figure 3. Results of meta-analysis comparing supplemental calcium alone versus placebo for effectiveness in reducing the recurrence of any adenoma among populations with a history of adenomas.


0.2

May 2010 797

C. Carroll et al.

no significant reduction in the RR or absolute risk of being diagnosed with colorectal cancer among those taking calcium (RR = 0.34 [95% CI, 0.05 to 2.14], P = NS; RD = 0.00 [95% CI, –0.01 to 0.00], P = NS). These results for colorectal cancer incidence should be interpreted with caution because of the relatively short follow-up duration and the low event rate (1/668 in the calcium arms and 4/678 in the placebo arms).

Data were also reported for the Grau et al39 trial for participants with a subsequent colonoscopy for up to 10 years after the end of the 4-year treatment period. The follow-up reported a statistically significant 37% re-duction in the RR for adenoma among the 347 patients who had a further colonoscopy (37% of the 930 origi-nally randomized patients) in the first 5 years after the trial had ended (ie, years 4–8) (RR = 0.63 [95% CI, 0.46–0.87], P = 0.005). However, among the 424 patients who had a further colonoscopy (46% of the 930 originally randomized patients) in the period 6 to 10 years after the trial (ie, years 9–13 after the end of the trial), the risk reduction for adenomas was no longer significant (RR = 1.09 [95% CI, 0.85–1.39], P = NS). In both in-tervals, there was also no significant effect on the numbers of advanced adenomas. These data were omit-ted from the meta-analysis because posttrial calcium intake was not controlled for, and because the rates of attrition render these long-term follow-up data less robust for the purposes of the present analysis.

Individuals at Average Risk of Being Diagnosed With Colorectal Cancer

Two studies of participants with no known baseline risk of colorectal cancer were identified which compared calcium, with or without vitamin D, with a control group (Table I).35,36 These were multicenter trials, but the number of centers involved was not reported. The number of participants randomized was 1179 in the Lappe et al35 study and 36,282 in the Wactawski-Wende et al36 trial (the Women’s Health Initiative study). The study population consisted of postmenopausal women with a mean age of 59 years in the Wactawski-Wende et al study, and of participants aged ≥55 years in the study by Lappe et al. The study by Wactawski-Wende et al assessed calcium at 1000 mg/d plus vitamin D at 400 IU/d in comparison with placebo. Lappe et al com-pared 3 groups: calcium citrate at 1400 mg/d or calcium carbonate at 1500 mg/d alone; calcium at the same dose plus vitamin D at 1000 to 1100 IU/d; and placebo. The duration of treatment and follow-up was 4 years

in the study by Lappe et al and 7 years in the study by Wactawski-Wende et al.

The Wactawski-Wende et al36 study was found to be of good quality: it used adequate methods of allocation concealment, randomization, and blinding, and excluded <5% of randomized participants. A power calculation was performed, and the required sample size was achieved. The trial reported by Lappe et al35 was of lower quality: the generation of the randomization sequence was adequate, but allocation concealment and methods of blinding were unclear; between 5% and 20% of ran-domized participants were excluded and no power calculation was performed. Intent-to-treat analyses were performed in both studies.

No adenoma data were reported by these trials, but colorectal cancer was measured as an outcome. Neither study found (individually or combined in a meta-analysis) a statistically significant effect of calcium with or without vitamin D on either the relative or absolute risk of being diagnosed with colorectal cancer (Table II). Meta-analysis of the 2 studies (a total of 37,016 participants) compar-ing the effect of calcium plus vitamin D with the effect of placebo found no differences in the relative or absolute risk of being diagnosed with colorectal cancer (RR = 1.08 [95% CI, 0.87 to 1.34], P = NS; RD = 0.00 [95% CI, 0.00 to 0.00], P = NS) (Table II; Figure 5). The same result was found for a comparison of the effects of calcium with or without vitamin D with the effects of placebo, including a total of 37,461 participants in the analysis (RR = 0.62 [95% CI, 0.11 to 3.40], P = NS; RD = 0.00 [95% CI, –0.01 to 0.00], P = NS), and the degree of heterogeneity in this analysis was high (58%) (Table II; Figure 6). Only one trial (Lappe et al35: 733 par-ticipants in the analysis) compared calcium alone with placebo, and also reported no significant effect.

Adverse EventsFour of the 6 trials reported no adverse events (AEs)

or complications at doses of 1000 to 1500 mg/d, but 2 trials with higher doses observed AEs. The Bonithon-Kopp et al32 trial, which used a dose of 2000 mg/d, reported a statistically significant difference between the calcium and placebo groups in terms of all AEs (26/176 vs 12/178 [P = 0.043]), but there was no dif-ference in terms of the major AEs of severe diarrhea or abdominal pain (6/176 vs 3/178 [P = NS]; P value generated by authors). The Hofstad et al34 trial, which used a dose of 1600 mg/d, also reported a number of cases of diarrhea and abdominal pain in the calcium

798 V

olume 32 N

umber 5

Clinical T

herapeutics

Calcium + Vitamin D PlaceboWeight,

%

Random Risk Ratio



Lappe et al (2007)35 1 446 2 288 0.8 0.32 (0.03–3.54)Wactawski-Wende et al (2006)36

168 18,176 154 18,106 99.2 1.09 (0.87–1.35)

Total (95% CI) 18,622 18,394 100.0 1.08 (0.87–1.34)Total events 169 156Heterogeneity: t2 = 0.00, x2 = 0.98, df = 1 (P = 0.32), I2 = 0%Test for overall effect: z = 1.66 (P = 0.51)

Figure 5. Results of meta-analysis comparing supplemental calcium plus vitamin D versus placebo for reducing the incidence of colorectal cancer among populations with no increased baseline risk of colorectal cancer.

100.1 1 100Favors placeboFavors calcium +

vitamin D

0.01

Calcium PlaceboWeight,

%

Random Risk Ratio



Baron et al (1999)30,31 30 409 43 423 89.4 0.72 (0.46–1.13)Bonithon-Kopp et al (2000)32 5 176 4 178 10.6 1.26 (0.35–4.63)

Total (95% CI) 585 601 100.0 0.77 (0.50–1.17)Total events 35 47Heterogeneity: t2 = 0.00, x2 = 0.064, df = 1 (P = 0.42), I2 = 0%Test for overall effect: z = 1.24 (P = 0.21)

Figure 4. Results of meta-analysis comparing supplemental calcium alone versus placebo for effectiveness in reducing the recurrence of ad-vanced adenoma among populations with a history of adenomas.


0.2

May 2010

799

C. C

arroll et al.

Calcium ± Vitamin D PlaceboWeight,

%

Random Risk Ratio



Lappe et al (2007)35 1 891 2 288 29.6 0.16 (0.01–1.78)Wactawski-Wende et al (2006)36

168 18,176 154 18,106 70.4 1.09 (0.87–1.35)

Total (95% CI) 19,067 18,394 100.0 0.62 (0.11–3.40)Total events 169 156Heterogeneity: t2 = 1.06, x2 = 2.41, df = 1 (P = 0.12), I2 = 58%Test for overall effect: z = 0.55 (P = 0.58)

Figure 6. Results of meta-analysis comparing supplemental calcium, with or without vitamin D, versus placebo for reducing the incidence of colorectal cancer among populations with no increased baseline risk of colorectal cancer.

100.1 1 100Favors placeboFavors calcium ±

vitamin D

0.01



(Baron et al30 and Bonithon-Kopp et al) showed a slightly reduced (18%), statistically significant reduction in the risk of adenoma recurrence. The more conserva-tive, smaller effect size reported by the present review for this population may therefore be the most robust of the results in all of the published reviews.

It should also be noted that the effect found during the first 4-year follow-up period in the Grau et al39 trial did not continue consistently over the following 10 years; benefits appeared to increase during the first 5 years but disappear during the following 5 years. However, the intake of calcium and other agents was not controlled, after the treatment period had ended; attrition was substantial. Nevertheless, it is possible that some protec-tive effect might have persisted if those taking supple-mental calcium had continued to do so. All 3 trials in samples with a history of adenomas reported the same direction of effect, although the relative reduction in risk was statistically significant only in the Baron et al30

and Hofstad et al34 trials. On the other hand, the num-bers of patients with an adenoma were very different in all 3 trials. The Baron et al trial had twice as many participants but 4 times the number of patients with adenomas, in both arms, than the Bonithon-Kopp et al32 trial (Figure 2). The Hofstad et al trial had a quarter of the participants, but a similar number of patients with adenomas, in both arms, as the Bonithon-Kopp et al trial. The higher incidence of adenomas in the former trial may be due to the fact that polyps were left in situ, which may have also affected the risk of new adenomas.34,42,43

The substantial differences between the Baron et al and Bonithon-Kopp et al trials may also have been the result of differences between the populations of the 2 trials, principally in terms of the type, number, and location of the adenomas present but removed at baseline (ie, larger numbers of proximal and/or advanced adenomas in the Baron et al trial, increasing the risk of recurrence).44,45

Different doses and follow-up (only 1 year difference between the trials) are unlikely to explain the discrepancies in the numbers of events in both arms across the trials.

No previous review has evaluated advanced adenoma as an outcome, or performed analyses for the outcome of colorectal cancer in populations with a history of adenomas or with no increased baseline risk of colorectal cancer. The meta-analyses described in the present article found that the significant positive effect identified for any adenoma was not mirrored for advanced adenomas, either for the original treatment period or for additional follow-up data.39 This is important because the presence

group (5 and 4, respectively), but found no significant differences between groups (placebo group had 7 and 4, respectively).

DISCUSSIONSupplemental calcium does not appear to benefit FAP populations, based on the limited evidence identified by this review, although the only randomized clinical trial included very few participants, so its findings must be considered cautiously. Only the NSAIDs sulindac, celecoxib, and tiracoxib have been reported to have any chemopreventive effect in individuals with FAP or HNPCC.17 Meta-analysis of 3 trials that enrolled indi-viduals with a history of adenomas found a significant reduction in the risk of adenoma recurrence for those taking supplemental calcium with or without antioxi-dants compared with those taking placebo (RR = 0.80 [95% CI, 0.69–0.94], P = 0.006). This is similar to the RR reported by 2 previous reviews.24,25 Shaukat et al25

performed the same 3-trial analysis with nearly identical data (36 rather than 35 events in the placebo group of the Hofstad et al33,34 trial) and generated the same risk (RR = 0.80 [95% CI, 0.68 to 0.93], P = 0.004). However, the possible exclusion of the Hofstad et al trial in a sensitivity analysis is justified for the reasons noted previously, and, as demonstrated, the risk reduction associated with calcium was reduced when only the 2 higher quality trials of calcium alone versus placebo alone were analyzed (RR = 0.82 [95% CI, 0.69 to 0.98], P = 0.03). Despite including only the Baron et al30,31

and Bonithon-Kopp et al32 trials in their analysis, and excluding the Hofstad et al trial because of the use of antioxidants as a cointervention, Weingarten et al24

reported the largest protective effect for calcium (odds ratio [OR] = 0.74 [95% CI, 0.58 to 0.95], P = 0.02). This OR translated to a risk reduction of 20% (RR = 0.80 [95% CI, 0.67 to 0.96], P = 0.02). This result was a consequence of using numbers of randomized partici-pants, rather than the numbers for whom an outcome measurement had been recorded, as the denominator in the analysis. The former approach assumes that none of those lost to follow-up experienced the target out-come.40,41 The relative percentage of events in the calcium arm of the analysis in the Weingarten et al review was therefore smaller than that reported in the original intent-to-treat analysis, and so may have exaggerated the effect of calcium. The sensitivity analysis performed by the current review, including only the largest, best-quality trials evaluating supplemental calcium alone

May 2010 801

C. Carroll et al.

tion, P = 0.018). Supplemental calcium may, therefore, have an effect in some populations with no increased baseline risk.

Supplemental calcium only appeared to have a sta-tistically significant, positive chemopreventive effect in reducing numbers of adenomas in populations with a history of adenomas. Nevertheless, despite this limita-tion, calcium still compares favorably with other po-tential chemopreventive agents. Aspirin has also been found to have a significant chemopreventive effect in populations with a history of adenomas, but has con-siderable AEs.17 Other NSAIDs, such as celecoxib, have also been found to be effective, but induce serious AEs, which has led to the termination of a trial evaluat-ing their effectiveness in colorectal cancer prevention.17

Antioxidants and folic acid have been found to have no proven beneficial effect, and may even be potentially harmful.21,23 By contrast, calcium has few reported AEs, although, as noted, these may increase if higher doses are tried to produce greater effects. An assessment of the cost-effectiveness of calcium chemoprevention in addition to screening, compared with aspirin chemo-prevention in addition to screening, and compared with screening alone, suggested that calcium has ~50% to ~60% probability of being the most economically attractive option in the at-risk population, depend-ing on the age at chemoprevention inception and dura-tion of therapy.17 The uncertainty about this finding is primarily the result of the uncertainty about the ef-fectiveness of calcium compared with aspirin chemo-prevention, and the uncertainty about the AE profile of the agents. Given that the present review found only 3 trials, at most, for those populations at different levels of risk, further trials are needed to confirm the potential chemopreventive effect of calcium demonstrated by meta-analysis of these trials, especially at higher doses of the agent. Such trials are also justified by the appar-ent cost-effectiveness of and general lack of AEs with calcium as a chemopreventive agent. There is, therefore, good reason to continue assessing the chemopreventive effects of supplemental calcium, as well as its AE profile, at different doses, especially in populations with a his-tory of adenomas, but also in general populations, such as women who do not take supplemental estrogen.

CONCLUSIONSPublished reports indicated that supplemental calcium was effective for the prevention of adenoma recurrence in individuals with a history of adenomas, but no effect

of advanced adenomas has been identified as a principal risk factor for adenoma recurrence.44 Furthermore, calcium appeared to have no beneficial impact on the incidence of colorectal cancer, either for populations with a history of adenomas or for populations with no increased baseline risk. The number of events was very small for the former population, but far more meaning-ful for the general population studies.

However, the duration of follow-up in these general-population studies was only up to 7 years, which may not be long enough to detect a reliable effect on cancer occurrence; at least 10 years may be required.12,13 Also, participants in these trials were not tested for the pres-ence of adenomas at baseline, which may confound the results. Further caution must be exercised with all of these results because there have been few such studies overall: these analyses only included 2 or 3 trials in total, and the mean age in the majority of trials was 59 to 61 years, which is somewhat younger than the median age of individuals at the time of diagnosis of colorectal cancer.3,4

Some antioxidants, as noted previously, have been found to have no beneficial effect, and may possibly even be harmful. The effects of vitamin D are less clear: Lappe et al35 reported results for both calcium alone and calcium plus vitamin D, finding no difference between these trial arms. The potential for doses of vitamin D to enhance the protective effect of supplemental calcium, suggested by previous research,43 also cannot be dem-onstrated by the limited trial evidence for low-risk populations that was identified by the present review. It must also be noted that the meta-analysis of the Lappe et al and Wactawski-Wende et al36 trials generated substantial statistical heterogeneity, most likely because of the small number of events in the Lappe et al study, which may have skewed the results. A further weakness of the analysis was that it was a marginal analysis (ie, the intervention may have interacted with cointerven-tions, such as antioxidants).46

Despite these possible confounding factors affecting the reliability of the results of the meta-analysis, more recent analysis of the Wactawski-Wende et al36 study data suggests that estrogen may modify calcium and vitamin D supplementation.45,47 The analysis by Ding et al45 demonstrated that, for women taking calcium and vitamin D without estrogen, the data suggested a possible benefit (hazard ratio = 0.71 [95% CI, 0.46 to 1.01]), although it failed to reach the 5% level of prob-ability for statistical significance (for estrogen interac-



11. University of York Centre for Reviews and Dissemination. The management of colorectal cancers. Effect Health Care. 2004;8:1–12.

12. Leslie A, Carey FA, Pratt NR, Steele RJ. The colorectal adenoma-carcinoma sequence. Br J Surg. 2002;89:845– 860.

13. Stryker SJ, Wolff BG, Culp CE, et al. Natural history of untreated colonic polyps. Gastroenterology. 1987;93:1009– 1013.

14. Flossmann E, Rothwell PM, for the British Doctors Aspirin Trial and the UK-TIA Aspirin Trial. Effect of aspirin on long-term risk of colorectal cancer: Consistent evidence from randomised and observational studies. Lancet. 2007; 369:1603–1613.

15. Atkin WS, Saunders BP, for the British Society for Gastro-enterology and the Association of Coloproctology for Great Britain and Ireland. Surveillance guidelines after removal of colorectal adenomatous polyps. Gut. 2002;51 (Suppl 5):V6–V9.

16. Cancer Research UK. Bowel cancer: Survival statistics. http://info.cancerresearchuk.org/cancerstats/types/bowel/ survival/index.htm. Accessed April 23, 2010.

17. Cooper K, Pilgrim H, Carroll C, et al. Chemoprevention of colorectal cancer: Systematic review and economic evalua-tion. Health Technol Assess. In press.

18. Dubé C, Rostom A, Lewin G, et al. The use of aspirin for primary prevention of colorectal cancer: A systematic review prepared for the US Preventive Services Task Force. Ann Intern Med. 2007;146:365–375.

19. Rostom A, Dubé C, Lewin G, et al. Nonsteroidal anti- inflammatory drugs and cyclooxygenase-2 inhibitors for primary prevention of colorectal cancer: A systematic re-view prepared for the US Preventive Services Task Force. Ann Intern Med. 2007;146:376–389.

20. Asano TK, McLeod RS. Non steroidal anti-inflammatory drugs (NSAID) and aspirin for preventing colorectal ade-nomas and carcinomas. Cochrane Database Syst Rev. 2004;(2): CD004079.

21. Carroll C, Cooper K, Papaioannou D, et al. Meta-analysis: Folic acid in the prevention of colorectal adenomas and the chemoprevention of colorectal cancer. Aliment Pharmacol Ther. 2010;31:708–718.

22. Bjelakovic G, Nikolova D, Simonetti RG, Gluud C. Anti- oxidant supplements for preventing gastrointestinal can-cers. Cochrane Database Syst Rev. 2008;(3):CD004183.

23. Bjelakovic G, Nagorni A, Nikolova D, et al. Meta-analysis: Antioxidant supplements for primary and secondary pre-vention of colorectal adenoma. Aliment Pharmacol Ther. 2006;24:281–291.

24. Weingarten MA, Zalmanovici A, Yaphe J. Dietary calcium supplementation for preventing colorectal cancer and ad-enomatous polyps. Cochrane Database Syst Rev. 2008;(1): CD003548.

was apparent in those at higher or lower risk. The validity of this review was robust, given that the findings and direction of effect in all but 1 of the 5 trials were consis-tent, and 3 of the 5 principal trials were of good quality, evaluating the effect of supplemental calcium in sizable samples.

ACKNOWLEDGMENTSThis study was funded, in part, by the UK National Coordinating Centre for Health Technology Assessment. The authors have indicated that they have no conflicts of interest to declare, and that there was no industry support for or involvement in this study.

The authors wish to thank Bjorn Hofstad for provid-ing unpublished data.

REFERENCES1. Bond JH. Colorectal cancer update. Prevention, screen-

ing, treatment, and surveillance for high-risk groups. Med Clin North Am. 2000;84:1163–1182.

2. Ferlay J, Autier P, Boniol M, et al. Estimates of the cancer incidence and mortality in Europe in 2006. Ann Oncol. 2007;18:581–592.

3. Office for National Statistics. Cancer statistics registra-tion: Registrations of cancer diagnosed in 2006, England. Series MB1, No. 37. http://www.statistics.gov.uk/StatBase/Product.asp?vlnk=8843. Accessed March 30, 2010.

4. Welsh Cancer Intelligence and Surveillance Unit. Cancer incidence in Wales, 2003–2007. http://www.wales.nhs.uk/ sites3/docmetadata.cfm?orgid=242&id=110389. Accessed March 30, 2010.

5. Giovannucci E, Stampfer MJ, Colditz GA, et al. Folate, methionine, and alcohol intake and risk of colorectal ad-enoma. J Natl Cancer Inst. 1993;85:875–884.

6. National Institute for Health and Clinical Excellence. Im-proving outcomes in colorectal cancers: Manual update. http://guidance.nice.org.uk/CSGCC/Guidance/pdf/ English. Accessed March 30, 2010.

7. Bishop DT, Hall NR. The genetics of colorectal cancer. Eur J Cancer. 1994;30A:1946–1956.

8. Aaltonen LA, Salovaara R, Kristo P, et al. Incidence of hereditary nonpolyposis colorectal cancer and the feasi-bility of molecular screening for the disease. N Engl J Med. 1998;338:1481–1487.

9. Hill MJ, Morson BC, Bussey HJ. Aetiology of adenoma—carcinoma sequence in large bowel. Lancet. 1978;1:245– 247.

10. Ramsoekh D, Van Leerdam ME, Wagner A, Kuipers EJ. Review article: Detection and management of hereditary non-polyposis colorectal cancer (Lynch syndrome). Aliment Pharmacol Ther. 2007;26(Suppl 2):101–111.

May 2010 803

C. Carroll et al.

of published randomised controlled trials. BMJ. 1999;319:670–674.

41. Montori V, Guyatt G. Intention- to-treat principle. CMAJ. 2001;165: 1339–1341.

42. Martínez ME, Sampliner R, Mar-shall JR, et al. Adenoma character-istics as risk factors for recurrence of advanced adenomas. Gastroenterol-ogy. 2001;120:1077–1083.

43. Bonithon-Kopp C, Piard F, Fenger C, et al, for the European Cancer Pre-vention Organisation Study Group. Colorectal adenoma characteristics as predictors of recurrence. Dis Colon Rectum. 2004;47:323–333.

44. Grau MV, Baron JA, Sandler RS, et al. Vitamin D, calcium supplemen-tation, and colorectal adenomas: Results of a randomized trial. J Natl Cancer Inst. 2003;95:1765–1771.

45. Ding EL, Mehta S, Fawzi WW, Giovannucci EL. Interaction of es-trogen therapy with calcium and vitamin D supplementation on colorectal cancer risk: Reanalysis of Women’s Health Initiative random-ized trial. Int J Cancer. 2008;122: 1690–1694.

46. McAlister FA, Straus SE, Sackett DL, Altman DG. Analysis and reporting of factorial trials: A systematic re-view. JAMA. 2003;289:2545–2553.

47. Stajner I. Comment on interaction of hormone replacement therapy with calcium and vitamin D supple-mentation on colorectal cancer risk. Int J Cancer. 2009;124:1736; author reply, 1737–1738.

Cancer Prev. 1993;2:189]. Eur J Can-cer Prev. 1992;1:415–422.

34. Hofstad B, Almendingen K, Vatn M, et al. Growth and recurrence of colorectal polyps: A double-blind 3-year intervention with calcium and antioxidants. Digestion. 1998;59:148– 156.

35. Lappe JM, Travers-Gustafson D, Davies KM, et al. Vitamin D and calcium supplementation reduces cancer risk: Results of a randomized trial [published correction appears in Am J Clin Nutr. 2008;87:794]. Am J Clin Nutr. 2007;85:1586–1591.

36. Wactawski-Wende J, Kotchen JM, Anderson GL, et al, for the Wom-en’s Health Initiative Investigators. Calcium plus vitamin D supplemen-tation and the risk of colorectal can-cer [published correction appears in N Engl J Med. 2006;354:1102]. N Engl J Med. 2006;354:684–696.

37. Tierney JF, Stewart LA. Investigating patient exclusion bias in meta-analysis. Int J Epidemiol. 2005;34:79–87.

38. Bjelakovic G, Nikolova D, Simonetti RG, Gluud C. Antioxidant supple-ments for prevention of gastrointestinal cancers: A systematic review and meta-analysis. Lancet. 2004;364: 1219–1228.

39. Grau MV, Baron JA, Sandler RS, et al. Prolonged effect of calcium supple-mentation on risk of colorectal ade-nomas in a randomized trial. J Natl Cancer Inst. 2007;99:129–136.

40. Hollis S, Campbell F. What is meant by intention to treat analysis? Survey

25. Shaukat A, Scouras N, Schünemann HJ. Role of supplemental calcium in the recurrence of colorectal ade-nomas: A meta-analysis of random-ized controlled trials. Am J Gastroen-terol. 2005;100:390–394.

26. Scottish Intercollegiate Guidelines Network. Search filters: Randomised controlled trials. http://www.sign.ac.uk/methodology/filters.html# random. Accessed March 30, 2010.

27. National Health Service Centre for Reviews and Dissemination. System- atic reviews: CRD’s guidance for un-dertaking reviews in healthcare. http://www.york.ac.uk/inst/crd/SysRev/!SSL!/WebHelp/TITLEPAGE.htm. Accessed March 30, 2010.

28. Higgins J, Green S. Cochrane Hand-book for Systematic Reviews of Interven-tions, version 5.0.2. http://www. cochrane-handbook.org. Accessed April 23, 2010.

29. Thomas MG, Thomson JP, William-son RC. Oral calcium inhibits rectal epithelial proliferation in familial ad-enomatous polyposis. Br J Surg. 1993; 80:499–501.

30. Baron JA, Beach M, Mandel JS, et al, for the Calcium Polyp Prevention Study Group. Calcium supplements for the prevention of colorectal adenomas. N Engl J Med. 1999;340: 101–107.

31. Baron JA, Beach M, Mandel JS, et al, for the Polyp Prevention Study Group. Calcium supplements and colorectal adenomas. Ann N Y Acad Sci. 1999;889:138–145.

32. Bonithon-Kopp C, Kronborg O, Giacosa A, et al, for the European Cancer Prevention Organisation Study Group. Calcium and fibre supplementation in prevention of colorectal adenoma recurrence: A randomised intervention trial. Lancet. 2000;356:1300–1306.

33. Hofstad B, Vatn M, Hoff G, et al. Growth of colorectal polyps: Design of a prospective, randomized, placebo- controlled intervention study in pa-tients with colorectal polyps [published correction appears in Eur J

Address correspondence to: Christopher Carroll, PhD, Health Economics and Decision Science, School of Health and Related Research, University of Sheffield, Regent Court, Regent Street, Sheffield, United Kingdom, S1 4DA. E-mail: [email protected]

supplemental calcium in the chemoprevention of colorectal cancer: a systematic review and...

Documents