Treatment With B Vitamins and Incidence of Cancer in Patients With Previous Stroke or Transient Ischemic Attack
Results of a Randomized Placebo-Controlled Trial
Background and Purpose—To determine the effect of B vitamin treatment on the incidence of cancer among patients with stroke or transient ischemic attack.
Methods—A total of 8164 patients with recent stroke or transient ischemic attack were randomly allocated to double-blind treatment with 1 tablet daily of placebo or B vitamins (2 mg folic acid, 25 mg vitamin B6, 500 μg vitamin B12) and followed for a median of 3.4 years for any cancer as an adverse event.
Results—There was no significant difference in the incidence of any cancer among participants assigned B vitamins compared with placebo (4.04% versus 4.59%; risk ratio, 0.86; 95% CI, 0.70–1.07) and no difference in cancer mortality (2.35% versus 2.09%; risk ratio, 1.09; 0.81–1.46). Among 1899 patients with diabetes, the incidence of cancer was higher among participants assigned B vitamins compared with placebo (5.35% versus 3.28%; adjusted risk ratio, 2.21; 1.31–3.73), whereas among 6168 patients without diabetes, the incidence of cancer was lower among participants assigned B vitamins compared with placebo (3.66% versus 5.03%; adjusted risk ratio, 0.67; 0.51–0.87; P for interaction=0.0001).
Conclusions—Daily administration of folic acid, vitamin B6, and vitamin B12 to 8164 patients with recent stroke or transient ischemic attack for a median of 3.4 years had no significant effect, compared with placebo, on cancer incidence or mortality. However, a post hoc subgroup analysis raises the hypothesis that folic acid treatment may increase the incidence of cancer among diabetics and reduce the incidence of cancer among nondiabetics with a history of stroke or transient ischemic attack.
Folate is a B vitamin that is essential for the biosynthesis of nucleotides, replication of DNA, supply of methyl groups, and the growth and repair of cells.1 Observational studies have reported folate deficiency to be associated with an increased risk of breast2 and colorectal3 cancer and higher intakes of dietary folate with a lower risk of colorectal cancer.4 Functional polymorphisms involving the methylenetetrahydrofolate reductase gene involved in folate metabolism are also associated with an increased risk of cancer.5,6 However, the introduction of food fortification with folic acid (the synthetic form of folate used in vitamin supplements and fortified foods) in the United States and Canada in 1996 coincided with a deceleration, rather than an acceleration, in the decline in colorectal cancer incidence in the United States and Canada, raising the hypothesis that folic acid treatment may increase the rate of transformation of adenomas into cancers and small cancers into larger cancers.7,8 This hypothesis is supported by experimental studies showing that folic acid treatment may compromise immunity against cancer9,10 and accelerate the growth of established cancer cells.11,12
Three randomized trials of folic acid treatment in participants with a history of colorectal adenoma reported conflicting results on the recurrence of adenomas and cancer incidence,13–15 whereas 2 trials of treatment with folic acid plus vitamin B12 in patients with ischemic heart disease reported increased cancer outcomes and all-cause mortality.16
A meta-analysis of 8 randomized placebo-controlled trials (RCTs) of folic acid treatment involving 35 603 individuals at risk of cardiovascular disease did not find any significant effect of folic acid on 3010 incident cancers during a median follow-up of 5 years (rate ratio [RR], 1.05; 95% CI, 0.98–1.13) nor any effect on cancer mortality (RR, 1.00; 0.85–1.18).17 Furthermore, there was no heterogeneity in the effects on cancer by dose of folic acid or by duration of folic acid treatment. However, because the 95% CIs of the RR are consistent with folic acid increasing the risk of cancer by up to 13%, and lowering the risk by only up to 2%, it is possible that the meta-analysis is underpowered to reliably exclude a modest, but clinically significant, increase in the risk of cancer.
We aimed to explore the effect of folic acid on any cancer as an adverse event in a population of 8164 patients with previous stroke enrolled in the VITAmins TO Prevent Stroke (VITATOPS) trial.18,19 The VITATOPS trial is registered with ClinicalTrials.gov, number NCT00097669, and Current Controlled Trials, number ISRCTN74743444.
The methods and primary results of the VITATOPS trial have been published.18,19 Briefly, 8164 patients with recent stroke or transient ischemic attack (TIA; <7 months) were randomly assigned to take B vitamins (2 mg folic acid, 25 mg vitamin B6, 500 μg vitamin B12) or placebo for a median duration of 3·4 years (interquartile range, 2.0–5.5).19 The primary outcome was the composite of stroke, myocardial infarction, or death from vascular causes.19
Data on adverse events, which included all incident cancers and cancer mortality, were reported by investigators on an adverse event form at each 6-month follow-up visit and adjudicated by a blinded Outcome and Adverse Events Adjudication Committee.17 Incident cancers were the first occurrence after randomization of any new cancers excluding, where possible, nonfatal nonmelanoma skin cancers. Cancers were subclassified into colorectal, other gastrointestinal, prostate, other genitourinary, lung, breast, melanoma, hematologic, and other. Linkage to population-wide cancer or cause-of-death registries was not performed.
This subgroup analysis, aimed at examining the effect of once-daily B vitamin treatments on the overall incidence of the adverse events, cancer, and cancer mortality, was planned before the study began. Hence, we collected cancer outcomes throughout the study period. The study protocol does not make specific mention of our plan to analyze cancer outcomes.18,19
Baseline characteristics and laboratory data were tabulated according to the assigned treatment groups and expressed as proportions for categorical variables and means (SD) for continuous variables with a normal distribution.
The primary analysis compared the incidence of the any cancer event between the placebo and B vitamin groups during follow-up by intention to treat. The event rates were calculated as the number of events that occurred during the follow-up period divided by the total number of patients randomized. The ratios of the event rates (RR:treatment/placebo) and their 95% CIs were calculated to describe the treatment effect. We used Kaplan-Meier methods to construct cumulative time-to-event curves for the 2 groups and the main comparison was based on a log-rank test.
A Cox proportional hazard model analysis was used to control for any potential imbalance in baseline characteristics and follow-up between the 2 groups.
Subgroup analyses compared the effect of B vitamins with placebo on cancer incidence according to patient age at randomization, sex, ethnic group, folate fortification, diabetes, antiplatelet use at randomization or any time, smoking at randomization or at any time, alcohol use at randomization, red cell folate concentrations, serum folate, total plasma homocysteine concentrations, and year of follow-up.
Two-sided significance tests were used throughout and a 2-sided probability value <0.05 was considered significant.
The composite of stroke, myocardial infarction, or death from vascular causes occurred among 15% of patients assigned to B vitamins and 17% assigned to placebo (RR, 0.91; 95% CI, 0.82–1.00; P=0.05; absolute risk reduction, 1.56%; 95% CI, −0.01 to 0.16%).19
Table 1 and Figure 1 show that any cancer occurred in 165 (4.04%) participants assigned B vitamins and 187 (4.59%) assigned placebo (RR, 0.86; 95% CI, 0.70–1.07). There was no significant difference in the RR for any subtype of cancer (Table 1).
A Cox proportional hazard model analysis revealed similar hazard ratios to the RR both before and after adjusting for any potential imbalance in the baseline characteristics and follow-up duration between the 2 groups.
Cancer Subtypes and Cancer Death
Table 1 shows that, compared with placebo, B vitamins did not significantly reduce any cancer death (2.35% B vitamins versus 2.09% placebo; RR, 1.09; 95% CI, 0.81–1.46).
Table 2 and Figure 2 show that results for incident cancer in the whole trial population were consistent among all patient subgroups, except for diabetes. Among 1899 patients with diabetes, any cancer occurred in 51 of 954 (5.35%) participants assigned B vitamins and 31 of 945 (3.28%) assigned placebo (adjusted RR, 2.21; 95% CI, 1.31–3.73). For patients without diabetes, the adjusted RR is 0.67 (95% CI, 0.51–0.87). The interaction between diabetes and treatment was significant (P=0.0001).
The principal result of this study is that daily administration of B vitamins to patients with recent stroke or TIA for a median of 3.4 years had no significant effect, compared with placebo, on overall cancer incidence or mortality. However, a post hoc subgroup analysis raises the hypothesis that folic acid treatment may increase the incidence of cancer among diabetics with a history of stroke or TIA and decrease the incidence of cancer among nondiabetics with a history of stroke or TIA.
The strengths of the study, supporting the validity of the results, are that systematic bias in treatment allocation was minimized by the randomization process and observer bias in the evaluation of cancer outcomes was minimized by the blinding of assessors, clinicians, and patients to the treatment allocation.
Potential limitations of the study are that it was a secondary analysis of an adverse event, it was underpowered to reliably investigate the effect of B vitamins on cancer outcomes, adherence to study medication was incomplete, which may have biased the results toward the null hypothesis, and ascertainment of cancer outcomes may have been incomplete. However, the results were consistent after excluding centers with incomplete follow-up.
A meta-analysis of 8 previous RCTs of folic acid treatment found that the incidence of cancer was similar among diabetics assigned B vitamins compared with placebo (8.6% versus 8.4%; RR, 1.05; 0.86–1.29) and also similar among nondiabetics assigned B vitamins compared with placebo (8.7% versus 8.2%; adjusted RR, 1.06; 0.95–1.18; P for interaction=1.0).17
The most likely reason for the discrepancy between the results of the 8 previous RCTs and the VITATOPS trial is random error (ie, chance). The meta-analysis of 8 RCTs was based on many more cancer events (n=3009) than the VITATOPS trial (n=352) and is therefore likely to be more reliable. However, although our VITATOPS trial results are those of a subgroup analysis, and therefore only hypothesis-generating, the interaction between B vitamin therapy and cancer outcomes among diabetics versus nondiabetics was highly significant (P=0.0001, or 1 in 10 000 chance of being a false-positive) and consistent across outcomes for any cancer and cancer mortality. Because the subgroups may not have been balanced between the 2 treatment groups, we conducted a multivariate analysis, by means of a Cox multiple regression model, to control for potential confounding effects and confirmed the significant interaction between diabetes and treatment on cancer incidence. Although diabetes itself was not a significant, independent predictor of cancer incidence in the Cox model (P=0.0596), it was a significant effect modifier (P=0.0001).
If our results are valid, other possible explanations for the discrepancy between the results of the 8 previous RCTs and the VITATOPS trial could be differences in the types of patients studied and their folate status. Patients enrolled in VITATOPS all had a previous TIA or stroke, most were not white, and many were residing in regions of low folate status (eg, Asia), unlike the previous 8 RCTs.
There are several possible mechanisms by which folate treatment could increase the risk of cancer in diabetics. Folate is a water-soluble B vitamin that plays a critical role in DNA synthesis, methylation, and repair.1,20 Imbalance in these 3 functions may contribute to carcinogenesis. Unmetabolized folic acid may compromise the body's immune defense against cancer9,10 and augment the growth of established cancer cells.11,12 Type 2 diabetes is associated with an increased risk for several cancers (liver, pancreas, endometrium, colon and rectum, breast, bladder) and a reduced risk of only prostate cancer.20 Although the association between diabetes and cancer may confounded by shared risk factors such as aging, obesity, diet, and physical inactivity, it could also be causal due to hyperinsulinemia, hyperglycemia, inflammation, or failure to metabolize folic acid.21
Our overall results are consistent with the meta-analysis of 8 placebo-controlled trials of folic acid treatment involving 35 603 individuals and support its conclusion that folic acid treatment for a median of 5 years does not significantly affect the rate of incident cancers or cancer mortality.17 However, we are not aware of any reports of a potential increase in cancer events among diabetic people with a history of stroke or TIA who are exposed to B vitamins. This post hoc hypothesis requires validation in independent data sets.
Although awaiting the results of these studies, our data extend the generalizability of the results of the B vitamin treatment trialists collaboration17 to patients with a history of stroke or TIA and suggest no adverse or favorable effect of B vitamins on cancer outcomes.
Sources of Funding
The National Health and Medical Research Council of Australia, the Medical Research Council of the United Kingdom, the Biomedical Research Council of Singapore, the National Medical Research Council of Singapore, the National Heart Foundation of Australia, the Royal Perth Hospital Medical Research Foundation, and the Health Department of Western Australia. Blackmores Ltd, Australia supplied the B vitamin and matching placebo tablets.
Dr Hankey was a principal investigator for the VITATOPS trial grants funded by the National Health and Medical Research Council of Australia. Dr Lees was a principal investigator for the VITATOPS trial grants funded by the Medical Research Council of the United Kingdom. Dr Chen was a principal investigator for the VITATOPS trial grants funded by the Biomedical Research Council of Singapore and the National Medical Research Council of Singapore.
Bruce Ovbiagele, MD, MSc, was the Guest Editor for this paper.
All authors were members of the International Steering Committee of the VITATOPS trial.
- Received October 17, 2011.
- Revision received January 17, 2012.
- Accepted February 9, 2012.
- © 2012 American Heart Association, Inc.
- Smith AD,
- Kim YI,
- Refsum H
- Larsson SC,
- Giovannucci E,
- Wolk A
- Giovannucci E
- Li D,
- Ahmed M,
- Li Y,
- Jiao L,
- Chou T-H,
- Wolff RA,
- et al
- Ulrich CM,
- Neuhouser M,
- Liu AY,
- Boynton A,
- Gregory JFR III.,
- Shane B,
- et al
- Mason JB,
- Dickstein A,
- Jacques PF,
- Haggarty P,
- Selhub J,
- Dallal G,
- et al
- Kim YI
- Troen AM,
- Mitchell B,
- Sorensen B,
- Wener MH,
- Johnston A,
- Wood B,
- et al
- Wu K,
- Plarz EA,
- Willett WC,
- Fuchs CS,
- Selhub J,
- Rosner BA,
- et al
- Clarke R,
- Halsey J,
- Lewington S,
- Lonn E,
- Armitage J,
- Manson JE,
- et al
- Hankey GJ,
- Algra A,
- Chen C,
- Wong MC,
- Cheung R,
- Wong L,
- et al
- Giovannucci E,
- Harlan DM,
- Archer MC,
- Bergenstal RM,
- Gapstur SM,
- Habel LA,
- et al