Effect of Combined Folic Acid, Vitamin B₆, and Vitamin B₁₂ on Cancer Risk in Women:  A Randomized Trial FREE

Folate, vitamin B₆, and vitamin B₁₂ (water-soluble, essential B vitamins) are important for DNA synthesis and methylation, 2 critical processes for upholding DNA integrity and regulating gene expression, respectively,^1- 2 and are thus thought to play an important role in cancer prevention. Background fortification of the food supply with folic acid (a synthetic form of folate), a policy that began in the United States in 1998 to reduce risk of neural tube defects, has improved folate status in the general population.³

Approximately one-third of US adults currently take multivitamin supplements containing folic acid, vitamin B₆, and vitamin B₁₂.⁴ Observational studies, which were conducted mostly before folic acid fortification, in general support an inverse association between high intake or blood level of folate, vitamin B₆, and vitamin B₁₂ and risk of cancer, particularly colorectal neoplasia and breast cancer, and primarily among those individuals consuming alcohol, a known antagonist for these B vitamins.^5- 9 Data from randomized trials of folic acid alone or in combination with B vitamins and cancer risk are limited, not entirely consistent,^10- 14 and 1 trial¹³ has even raised concerns about deleterious effects.

Women have been underrepresented in previous trials with B vitamins. We conducted a detailed analysis of cancer end points in the Women's Antioxidant and Folic Acid Cardiovascular Study (WAFACS), a large randomized trial designed to test the effect of combined folic acid, vitamin B₆, and vitamin B₁₂ on prevention of cardiovascular disease (CVD) among high-risk women during the folic acid fortification era.^15- 18

The WAFACS was a randomized, double-blind, placebo-controlled trial that evaluated a combined folic acid, vitamin B₆ (pyridoxine hydrochloride), and vitamin B₁₂ (cyanocobalamin) treatment in the prevention of CVD events among women at high risk for CVD.^15- 18 This report describes the results of cancer end points. Details of the design have been reported previously.^15- 18 Briefly, between June 1995 and October 1996, 8171 US female health professionals were randomized into the vitamin C, vitamin E, and beta carotene groups of the Women's Antioxidant Cardiovascular Study, the parent trial of the WAFACS (Figure 1). Women were eligible for the trial if they were aged 40 years or older, were postmenopausal or had no intention of becoming pregnant, and had a reported history of CVD or at least 3 coronary risk factors. Women were excluded from participation if they had a previous history of cancer (excluding nonmelanoma skin cancer) within the past 10 years, any serious non-CVD illness, or were currently using warfarin or other anticoagulants.

All participants completed a baseline questionnaire inquiring about their medical history and lifestyle factors. At baseline, 98.8% also completed a comprehensive food-frequency questionnaire, inquiring about the average intake of foods and beverages during the past year and the dose and duration of use of vitamin and mineral supplements. Total intakes of folate, vitamin B₆, and vitamin B₁₂ included sources from both foods and supplements. Of the 8171 randomized participants, 5442 women who were additionally willing to forgo the use of individual folic acid, vitamin B₆, and vitamin B₁₂ supplements, or multivitamin with more than the recommended daily allowances of folic acid, vitamin B₆, and vitamin B₁₂, were further randomized in a retained factorial design to a daily combination of 2.5 mg of folic acid, 50 mg of vitamin B₆, and 1 mg of vitamin B₁₂ (active treatment group), or to a matching placebo in April 1998. The doses used in the WAFACS trial were higher than the adult recommended daily allowances for folate (400 μg/d), vitamin B₆ (1.5 mg/d), and vitamin B₁₂ (2.4 μg/d) in women.¹⁹

After randomization, the women were sent annual supplies of monthly calendar packs containing study medications and questionnaires on adherence, adverse effects, relevant disease end points, and risk factors. Study medications and disease ascertainment were continued in a blinded fashion until the scheduled end of the trial (July 31, 2005) for 7.3 years. Follow-up and validation of reported end points were completed in July 2006. Using the information at the end of the trial (July 31, 2005), mortality and morbidity follow-up data were calculated to be 92.6% and 90.5% complete, respectively. In terms of percentage of potential person-years of follow-up, a calculation that takes into account all information during follow-up, mortality and morbidity were 98.9% and 98.0% complete, respectively.¹⁸ The average adherence, defined as taking at least two-thirds of the study pills over the course of follow-up, was 83% for both the active treatment and placebo groups.¹⁸ In a random sample of 300 participants, there were no differences in median plasma folate (8.8 vs 8.9 ng/mL [to convert to nanomoles per liter, multiply by 2.266]; P = .94) and homocysteine (12.1 vs 12.5 μmol/L; P = .96) between the active treatment and placebo groups at the beginning of the trial, whereas median plasma folate was significantly higher (38.9 vs 15.4 ng/mL; P < .001) and homocysteine was significantly lower (9.8 vs 11.8 μmol/L; P < .001) in the active treatment group vs the placebo group at the end of the trial.¹⁸

The trial was approved by the institutional review board of the Brigham and Women's Hospital and was monitored by an external data and safety monitoring board. All patients provided written informed consent.

Women reported the occurrence of cancer through questionnaire, letter, or telephone call. Deaths were reported by family members, postal authorities, or through a search of the National Death Index (through December 2003). Following the report of a relevant medical event, written permission to obtain medical records was sought from the participant or next of kin in case of a death. Medical records were obtained from hospitals and treating physicians and reviewed by an end points committee of physicians blinded to randomized treatment assignment to confirm final diagnosis. Additional details of breast tumor characteristics were also extracted from medical records. Death from cancer cause was confirmed by examination of autopsy reports, death certificates, medical records, and information obtained from the next of kin or other family members. Death from any cause was confirmed by the end points committee or on the basis of a death certificate. Only confirmed invasive cancers were included in this study.

Statistical power for the effect of combined folic acid, vitamin B₆, and vitamin B₁₂ treatment on cancer risk reduction was calculated for 2-sided α = .05. Based on a total of 379 observed cancer cases at the end of the trial in the 5442 women who were randomized, we estimated that there would be 82.2% power to detect a 25% reduction and 61.1% power to detect a 20% reduction.

Primary analyses were conducted on an intention-to-treat basis. Baseline characteristics between the active treatment group and the placebo group were compared by using t tests or Wilcoxon signed rank tests for continuous variables and χ² tests for categorical variables. Kaplan-Meier survival curves were used to estimate event rates over time, and curves of the active treatment group and the placebo group were compared by the log-rank test. Only those events with large numbers of occurrences in each treatment group (total invasive cancer and breast cancer) were analyzed by Kaplan-Meier survival curves. The hazard ratios (HRs) and 95% confidence intervals (CIs) of cancer comparing women randomized to combined B vitamin treatment vs placebo were computed by Cox proportional hazards regression models with adjustments for age (in years) and randomized treatment assignments of vitamin E (vitamin E vs placebo), vitamin C (vitamin C vs placebo), and beta carotene (beta carotene vs placebo). The proportionality assumption was tested by including an interaction term of combined B vitamin treatment with the logarithm of time in the Cox proportional hazards regression models and was not violated for total invasive cancer (P = .26), breast cancer (P = .74), or any cancer death (P = .13). To examine the influence of preclinical cases, we conducted analyses excluding the first 2 years of follow-up. Those women with missing information on mortality and morbidity were treated as nonevents throughout follow-up until the end of trial in the intention-to-treat analysis.

Total invasive cancer and breast cancer, 2 events with sufficient numbers of occurrences in each treatment group, were further analyzed across subgroups of baseline characteristics. The cutpoints for age chosen were consistent with the analysis of cardiovascular events in the WAFACS trial.¹⁸ Prespecified subgroup analyses included those according to the baseline intakes of total folate, vitamin B₆, vitamin B₁₂, and alcohol, and the baseline multivitamin supplement use and history of cancer. Tests for multiplicative interaction between randomized combined B vitamin treatment assignment and categories of baseline characteristics in relation to cancer risk were performed by the Wald test for variables with 2 categories or log likelihood ratio tests comparing the models with or without interaction terms for variables with more than 2 categories. We also evaluated the effect of randomized combined B vitamin treatment on risk for invasive breast cancer according to tumor characteristics at diagnosis. To examine the influence of lack of adherence, in a sensitivity analysis, women were censored if and when they stopped taking at least two-thirds of their study medication. The 16 subgroup analyses by participant characteristics were conducted for each outcome, and 1 statistically significant test would be expected on the basis of chance alone.

All P values were 2-sided, with a significance level of α = .05 (P≤.05). SAS version 9.1 (SAS Institute, Cary, North Carolina) was used for all analyses.

Of the 5442 women, 3492 (64.2%) had a prior CVD event and 1950 (35.8%) had 3 or more coronary risk factors. In addition, 418 (7.7%) had a prior cancer event at baseline (diagnosed at least 10 years before enrollment), 307 (5.6%) consumed 15 g/d or more of alcohol, and 1247 (22.9%) were current users of multivitamin supplements. The mean (SD) values at baseline were 62.8 (8.8) years for age and 30.6 (6.7) for body mass index, calculated as weight in kilograms divided by height in meters squared. There were no significant differences in baseline characteristics between the active treatment group and the placebo group (Table 1).

Three hundred seventy-nine women developed invasive cancer (187 in the active treatment group and 192 in the placebo group). Treatment with combined B vitamins had no significant effect on risk of total invasive cancer (101.1/10 000 person-years for the active treatment group vs 104.3/10 000 person-years for the placebo group; HR, 0.97; 95% CI, 0.79-1.18; P = .75) or individual cancer end points, including breast cancer (37.8/10 000 person-years vs 45.6/10 000 person-years, respectively; HR, 0.83; 95% CI, 0.60-1.14; P = .24) and colorectal cancer (9.7/10 000 person-years vs 12.0/10 000 person-years, respectively; HR, 0.81; 95% CI, 0.43-1.50; P = .50). There was no difference in any cancer death (24.6/10 000 person-years vs 30.1/10 000 person-years, respectively; HR, 0.82; 95% CI, 0.56-1.21; P = .32) or death from any cause (Table 2). Cumulative incidence curves indicated no variation over time for total invasive cancer, whereas a decrease in breast cancer risk appeared to emerge after approximately 3 years (Figure 2). After excluding the first 2 years of follow-up as prespecified, the HRs were 0.94 (95% CI, 0.73-1.20) for total invasive cancer, 0.73 (95% CI, 0.50-1.06) for breast cancer, and 0.92 (95% CI, 0.53-1.60) for any cancer death.

In analyses censoring women at the time they stopped taking at least two-thirds of their study pills, the null results remained for total invasive cancer (148 vs 140 cases; HR, 1.04; 95% CI, 0.83-1.32), breast cancer (59 vs 64 cases; HR, 0.91; 95% CI, 0.64-1.30), and any cancer death (29 vs 36 cases; HR, 0.80; 95% CI, 0.49-1.30).

Age significantly modified the effect of combined B vitamin treatment on risk of total invasive cancer and breast cancer (P for interaction = .02 and .05, respectively) (Table 3). A significantly reduced risk was observed for total invasive cancer (HR, 0.75; 95% CI, 0.57-0.99) and breast cancer (HR, 0.62; 95% CI, 0.40-0.98) among women aged 65 years or older at study entry, but no reductions in risk were observed among younger women (40-54 years or 55-64 years). There were no effect modifications by use of multivitamin supplements, intakes of total folate, vitamin B₆, and vitamin B₁₂, history of cancer, alcohol intake, or other risk factors at baseline. Finally, no significant effects were observed according to hormone receptor status or other characteristics of breast tumors, except that a borderline significant reduction was found for estrogen receptor–positive and progesterone receptor–negative tumors (Table 4).

In the WAFACS trial, up to 7.3 years of treatment with combined folic acid, vitamin B₆, and vitamin B₁₂ had no significant effect on overall risk of total invasive cancer, breast cancer, other individual cancers, or cancer death among women at high risk for CVD. There were no differences according to current use of multivitamin supplements, intakes of total folate, vitamin B₆, and vitamin B₁₂, or history of cancer at baseline. Lack of effect for total invasive cancer did not vary over time. We found a significant interaction by age group and a possible benefit among women aged 65 years or older at study entry.

There is a concern that increasing folate status resulting from mandatory folic acid fortification and supplementation may increase cancer risk in some people because folate may play a dual role in carcinogenesis—prevent tumor initiation when it is administered early in carcinogenesis and in individuals with suboptimal folate status, but promote tumor development when it is administered later in carcinogenesis (ie, once premalignant lesions are established) and in individuals with high folate intake.^20- 23 In the WAFACS trial, consistent with the implementation of US folic acid fortification policy, plasma folate levels were increased in the placebo group during 7.3 years of follow-up (8.9 vs 15.4 ng/mL). With the long duration of treatment and follow-up, the WAFACS trial showed no effect on cancer risk even among those women taking multivitamin supplements or with higher intake of total folate, vitamin B₆, and vitamin B₁₂, or with a history of cancer at baseline. Two other large randomized trials assessing folic acid, vitamin B₆, and vitamin B₁₂ in relation to CVD risk have reported cancer outcomes.^11- 12 In the Heart Outcomes Protection Evaluation (HOPE)-2 trial,¹² which included 5522 patients aged 55 years or older who had vascular disease or diabetes, an average of 5 years of treatment with the same daily combination of 2.5 mg of folic acid, 50 mg of vitamin B₆, and 1 mg of vitamin B₁₂ had no significant effect on risk of total cancer (relative risk, 1.06; 95% CI, 0.91-1.23), cancers of breast, colon, lung, or prostate, or cancer death. In the HOPE-2 trial, 71.1% of participants were from countries with folic acid fortification. The NORVIT trial,¹¹ which included 3749 men and women aged 30 to 85 years who recently had acute myocardial infarction in Norway, a country that has no mandatory folic acid fortification in foods, also reported no effect on cancer outcomes. Compared with the placebo group, there was no difference in cancer risk in those patients receiving the combined folic acid (0.8 mg), vitamin B₆ (40 mg), and vitamin B₁₂ (0.4 mg) treatment (40 vs 40 cases; relative risk, 1.02; 95% CI, 0.65-1.58) or combined folic acid (0.8 mg) and vitamin B₁₂ (0.4 mg) treatment (39 vs 40 cases) over an average of 3.3 years of treatment and follow-up.¹¹ Also, in this trial,¹¹ there was a suggestive beneficial effect on cancer risk in those patients treated with vitamin B₆ (40 mg) alone compared with those assigned to placebo (25 vs 40 cases). Taken together, data from randomized trials of combined B vitamins provide reassurance that folic acid supplementation up to a dose of 2.5 mg/d when combined with vitamin B₆ and vitamin B₁₂ does not appear to increase cancer occurrences and deaths among individuals at high risk for CVD during the folic acid fortification era. Because cancer has a long latency period, we cannot exclude the possibility that there might be beneficial or harmful effects of combined B vitamins on cancer that were not detectable within 7.3 years of treatment.

The results from randomized trials of folic acid alone and colorectal adenomas and of combined B vitamins and colorectal cancer have been mixed. In a trial¹⁰ that involved 60 men and women and was conducted before folic acid fortification, 24 months of treatment with 1 mg/d of folic acid nonsignificantly reduced the recurrence of colon adenomas. In another trial,¹³ which involved 1021 men and women and was conducted during folic acid fortification, 3 years of treatment with 1 mg/d of folic acid had no benefit on the recurrence of colorectal adenomas at 3 years of follow-up and at another 3 or 5 years of follow-up. Instead, there was an increased risk for advanced lesions at 3 years (relative risk, 1.32; 95% CI, 0.90-1.92) and at the second follow-up (relative risk, 1.67; 95% CI, 1.00-2.80). Those results did not differ significantly by sex, age, alcohol use, body mass index, or baseline plasma folate.¹³ In the third trial,¹⁴ which involved 945 men and women in England and was conducted before folic acid fortification, treatment with 0.5 mg/d of folic acid for 3 years had no effect on the recurrence of colorectal adenomas (relative risk, 1.07; 95% CI, 0.85-1.34) or advanced adenomas (relative risk, 0.98; 95% CI, 0.68-1.40). There was no increase in risk of colon cancer among women treated with combined B vitamins in the WAFACS trial, but a slightly increased risk was observed in the HOPE-2 trial.¹² However, the number of colon cancer cases was limited in these randomized trials. The evidence for a beneficial effect of folate, vitamin B₆, and vitamin B₁₂ on colorectal cancer is strong in observational studies.⁵ Emerging data also suggest that vitamin B₆ may be more important than folate for colon cancer prevention in postmenopausal women.⁸ Thus, the effect of folic acid or combined B vitamins on colon cancer risk remains uncertain.

Lack of an overall effect for breast cancer in the WAFACS trial is generally consistent with observational studies that suggest no overall association between intake or blood level of folate, vitamin B₆, and vitamin B₁₂, and breast cancer risk.⁶ Data from several large cohort and case-control studies suggest that adequate folate may reduce breast cancer risk associated with alcohol intake⁶ or decrease risk of developing estrogen receptor–negative breast cancer.²⁴ However, we observed no clear pattern of an effect according to alcohol intake and a borderline significant reduction was found for estrogen receptor–positive, progesterone receptor–negative breast cancer. The number of women consuming 15 g/d or more of alcohol or developing estrogen receptor–positive, progesterone receptor–negative or estrogen receptor–negative, progesterone receptor–negative breast cancers was limited in the WAFACS trial.

In 1 randomized trial²⁵ of folic acid supplementation in pregnancy followed up for 36 years, women who received 0.2 or 5 mg of folic acid supplements during pregnancy had increased deaths from total cancer (n = 112) and breast cancer (n = 31); those women who received the higher dose of 5 mg had the highest risk. However, that study had short (several months in pregnancy) and remote (36 years ago) treatment and small number of breast cancer events. The possibility that the findings may be a result of chance and confounding cannot be excluded. The results from experimental studies have been mixed. Folate deficiency suppressed the N-methyl-N-nitrosourea–induced mammary tumors,^26- 28 while folate supplementation enhanced the initiation or early promotion of the N-methyl-N-nitrosourea–induced mammary tumors in rats in 1 study,²⁶ but had no effect in 2 other studies.^27- 28

In the WAFACS trial, a significant benefit of combined folic acid, vitamin B₆, and vitamin B₁₂ treatment was observed among women aged 65 years or older. If the finding is real and substantiated, the results may have public health significance because the incidence rates of cancer are high in elderly persons.²⁹ The finding is biologically plausible because elderly individuals have increased requirements for these B vitamins.¹⁹ In addition, 2 large prospective investigations have shown that an inverse association between plasma vitamin B₆ and breast cancer risk is primarily present in postmenopausal women.^7,9 In the HOPE-2¹² and NORVIT¹¹ trials, there was no presentation by age for cancer end points, which were secondary outcomes. Because many subgroups were evaluated, we cannot exclude the possibility that the results from the subgroup analyses in the WAFACS trial are due to chance.

The strengths of the WAFACS trial include a randomized, double-blind, and placebo-controlled design, which minimizes confounding and bias that potentially affect the results from observational studies. To our knowledge, the WAFACS trial also had the longest duration of treatment with combined B vitamins of any trial to date, high follow-up rates, and relatively high adherence to treatment. Women have been underrepresented in other B vitamin trials, and the WAFACS trial is the largest trial of women. However, we cannot distinguish the effect attributable to any single B vitamin supplement compared with their combination. In addition, statistical power was insufficient to examine site-specific cancers. The WAFACS participants were female health professionals who tended to be health conscious, have well-balanced diets, and greater access to health care and screening, which may have led to lower occurrences of cancer. However, more than two-thirds of participants also were overweight or obese, and thus were at elevated risk for CVD and cancer. Therefore, the findings may not be directly generalizable to the entire US population. However, it seems unlikely that the exposure-disease relationships observed among women in the WAFACS differ from women in general.

In conclusion, treatment with combined folic acid, vitamin B₆, and vitamin B₁₂ provided neither beneficial nor harmful effects on overall risk of total cancer, breast cancer, or deaths from cancer among women at high risk for CVD.