Association between B Vitamins Supplementation and Risk of Cardiovascular Outcomes: A Cumulative Meta-Analysis of Randomized Controlled Trials

Background Observational studies suggest that B vitamin supplementation reduces cardiovascular risk in adults, but this association remains controversial. This study aimed to summarize the evidence from randomized controlled trials (RCTs) investigating B vitamin supplementation for the primary or secondary prevention of major adverse cardiovascular outcomes and to perform a cumulative meta-analysis to determine the evidence base. Methodology and Principal Findings In April 2013, we searched PubMed, Embase, and the Cochrane Library to identify relevant RCTs. We included RCTs investigating the effect of B vitamin supplementation on cardiovascular outcome. Relative risk (RR) was used to measure the effect using a random-effect model. Statistical heterogeneity scores were assessed using the Q statistic. We included data on 57,952 individuals from 24 RCTs: 12 primary prevention trials and 12 secondary prevention trials. In 23 of these trials, 10,917 major adverse cardiovascular events (MACE) occurred; in 20 trials, 7,203 deaths occurred; in 15 trials, 3,422 cardiac deaths occurred; in 19 trials, 3,623 myocardial infarctions (MI) occurred; and in 18 trials, 2,465 strokes occurred. B vitamin supplementation had little or no effect on the incidence of MACE (RR, 0.98; 95% confidence interval [CI]: 0.93–1.03; P = 0.37), total mortality (RR, 1.01; 95% CI: 0.97–1.05; P = 0.77), cardiac death (RR, 0.96; 95% CI: 0.90–1.02; P = 0.21), MI (RR, 0.99; 95% CI: 0.93–1.06; P = 0.82), or stroke (RR, 0.94; 95% CI: 0.85–1.03; P = 0.18). Conclusion/Significance B vitamin supplementation, when used for primary or secondary prevention, is not associated with a reduction in MACE, total mortality, cardiac death, MI, or stroke.

The effect of B vitamin supplementation on primary and secondary prevention of adverse cardiovascular outcomes has been studied in numerous RCTs . In this study, we performed a meta-analysis of these RCTs to evaluate the effect of Table 1. Design and characteristic of trials included in our meta-analysis*.  B vitamin supplementation on cardiovascular risk in specific subpopulations and attempt to determine the role of folic acid supplementation interaction with vitamin B6 and B12 in reducing cardiovascular risk. Furthermore, we used cumulative metaanalysis to determine the evidence base for routine B vitamin supplementation in clinical practice.

Data Sources, Search Strategy, and Selection Criteria
This review was conducted and reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis Statement, 2009 (Checklist S1) [44]. RCTs evaluating the effect of B vitamin supplementation on cardiovascular risk written in the English language were included in our study, regardless of the publication status (published, in press, or in progress) and the effect of B vitamin supplementation on MACE, total mortality, cardiac death, MI, and stroke were examined. Relevant trials were identified using the following procedure: (1) Electronic searches: we searched PubMed, Embase, and the Cochrane Central Register of Controlled Trials electronic databases for articles published through April 2013, using ''B vitamins'' AND ''randomized controlled trials'' AND ''clinical trials'' AND ''human'' AND ''English'' as the search terms.
(2) Other sources: we searched ongoing RCTs in the metaRegister of Controlled Trials, which lists trials that are registered as completed but not yet published. Furthermore, we reviewed bibliographies of publications for potentially relevant       trials. Medical subject headings, methods, patient population, interventions, and outcome variables of these studies were used to identify relevant trials. The literature search, data extraction, and quality assessment were undertaken by 2 investigators (CZ and PJM) independently with a standardized approach. Any inconsistencies between these investigators were identified by the primary investigator (YHZ) and resolved by consensus. We restricted our study to RCTs, which are less likely than observational studies to be subject to confounding variables or bias. A study was eligible for inclusion in our meta-analysis if the following criteria were met: (1) the study was a RCT; (2) the trial evaluated the effects of B vitamin supplementation compared with placebo; (3) the study's follow-up continued for at least 6 months; and (4) the trial reported at least 1 of the following outcomes: MACE, total mortality, cardiac death, MI, or stroke.

Data Collection and Quality Assessment
All data from included trials was extracted independently by 2 investigators (CZ and PJM) using a standardized protocol. Each data set was reviewed by a third investigator (YHZ), and any discrepancies between the 2 investigator's data were resolved by discussion. The data collected from each study included first author or study group name, publication year, study design, type of blinding, number of patients, percentage of men, mean age, background fortification, current disease status, baseline total homocysteine level, baseline folate level, intervention regimens, controls, and the duration of follow-up. The outcomes investigated included MACE, total mortality, cardiac death, MI, and stroke. Study quality was assessed using the Jadad score [45], which is based on the 5 following subscales: randomization (1 or 0), concealment of the treatment allocation (1 or 0), blinding (1 or 0), completeness of follow-up (1 or 0), and the use of intention-to-treat analysis (1 or 0). A score system ranging from 1 to 5 has been developed for quality assessment.

Statistical Analysis
We assigned the results of each RCT as dichotomous frequency data. Individual study relative risks (RR) and 95% confidence intervals (CI) were calculated from event numbers extracted from each trial before data pooling. The overall RR and 95% CI of MACE, total mortality, cardiac death, MI, and stroke were also calculated. Both fixed-effect and random-effect models were used to evaluate the pooled RR for B vitamin supplementation compared with placebo. Although both models yielded similar findings, results from the random-effect model, which assume that the true underlying effect varies among included trials, are presented here [46][47]. Heterogeneity of the treatment effect between studies was evaluated using the Q statistic, and P values ,0.10 were considered statistically significant [48][49]. In the cumulative meta-analysis, outcome data for MACE, total mortality, cardiac death, MI, and stroke from all available trials were included sequentially according to the year in which they first became available.
We explored potential sources of heterogeneity in estimates of the treatment effect on MACE with univariate meta-regression [50] (for number of participants, mean age, percentage of men, baseline homocysteine level, baseline folate level, dose of folic acid, dose of vitamin B6, dose of vitamin B12, net decrease in homocysteine level, and duration of follow-up). Subsequently, subgroup analyses were conducted for MACE on the basis of number of participants, mean age, percentage of men, number of trial centers, baseline homocysteine level, baseline folate level, intervention regimens, dose of folic acid, dose of vitamin B6, dose of vitamin B12, net decrease in homocysteine level, background fortification, control, disease prevention, renal status, duration of follow-up, and Jadad score [45]. Interaction tests [51] were performed to compare differences between estimates of the 2 subsets, which were based on Student t distribution rather than on normal distribution because the number of included studies was small. We also performed a sensitivity analysis by removing each individual trial from the meta-analysis [52]. Several methods were used to check for potential publication bias. Visual inspection of funnel plots for MACE, total mortality, cardiac death, MI, and stroke were conducted. The Egger (53) and Begg test (54) were used to statistically assess publication bias for MACE, total mortality, cardiac death, MI, and stroke. All reported P values were two-sided, and P values of ,0.05 were considered statistically significant for all included studies. Statistical analyses were performed using STATA software (version 10.0 StataCorp, Texas, USA).

Risk of Bias in Individual Trials
The quality of the trials was assessed using the Jadad score (45) and are summarized in Table 1. We considered a score $4 to indicate a high-quality study. Five trials [22,26,30,31,42]

Effects of B Vitamins on Major Cardiovascular Outcomes
Data from 56,925 individuals was used to assess the effect of B vitamin supplementation on MACE and included 10,917 MACE. Overall, B vitamin supplementation reduced the risk of MACE by 2%, but this was not statistically significant (RR, 0.98; 95% CI: 0.93-1.03; P = 0.37, Figure 2 and Figure S1). Heterogeneity was observed in the magnitude of the effect across the trials (P = 0.07). However, after sequential exclusion of each trial from all pooled analysis, the conclusion was not affected by the exclusion of any specific trial.
Data from 55,482 individuals was used to assess the effect of B vitamin supplementation on total mortality and included 7,203 deaths. No significant difference in the number of deaths was observed between participants receiving B vitamins compared with those receiving placebo (RR, 1.01; 95% CI: 0.97-1.05; P = 0.77, without evidence of heterogeneity; Figure 3 and Figure S2).
Data from 47,926 individuals was used to assess the effect of B vitamin supplementation on cardiac death and included 3,422 cardiac deaths. B vitamin supplementation caused a 4% reduction in cardiac death; however, this was not a significant change (RR, 0.96; 95% CI: 0.90-1.02; P = 0.21, without evidence of heterogeneity; Figure 4 and Figure S3).
Data from 54,976 individuals was used to assess the effect of B vitamin supplementation on MI, and included 3,623 MIs. There were no significant differences between participants receiving B vitamins compared with placebo for MI (RR, 0.99; 95% CI: 0.93-1.06; P = 0.82, without evidence of heterogeneity, Figure 5 and Figure S4).
Data from 54,102 individuals and 2,465 stroke events was used to assess the effect of B vitamin supplementation on stroke. B vitamin supplementation reduced incident stroke by 6%, but this was not a significant reduction (RR, 0.94; 95% CI: 0.85-1.03; P = 0.18; with no statistical heterogeneity, Figure 6 and Figure  S5). Sensitivity analysis was conducted for the incidence of stroke. However, after sequential exclusion of each trial from all pooled analysis, the conclusion was not affected by the exclusion of any specific trial.

Cumulative Meta-Analysis
On cumulative meta-analysis for MACE (Figure 2), the original nonsignificant result for a B vitamin effect persisted, and the effect was slight and borderline nonsignificant. Similarly, the nonsignificant result persisted when cumulative meta-analyses for total mortality (Figure 3), cardiac death (Figure 4), MI (Figure 5), and stroke ( Figure 6) were conducted.

Meta-Regression and Subgroup Analyses
Heterogeneity testing for the analysis showed a P,0.10 for MACE. We concluded that heterogeneity was statistically significant in the overall analysis. We, therefore, conducted a meta-regression analysis [50] for MACE that included the number of participants, mean age, percentage of men, baseline homocysteine level, baseline folate level, dose of folic acid, dose of vitamin B6, dose of vitamin B12, net decrease in homocysteine level, and duration of follow-up. Overall, we detected a net decrease in homocysteine (P = 0.043) that contributed to the association between B vitamin supplementation and MACE ( Figure S6). However, the number of participants (P = 0.435), mean age (P = 0.763), percentage of men (P = 0.866), baseline homocysteine level (P = 0.094), baseline folate level (P = 0.797), dose of folic acid (P = 0.840), dose of vitamin B6 (P = 0.933), dose of vitamin B12 (P = 0.614), and duration of follow-up (P = 0.186) were not significant factors contributing to the association between B vitamin supplementation and MACE ( Figure S6).
Subgroup analyses were conducted for MACE to minimize heterogeneity among the included trials and to evaluate the effect of B vitamin supplementation in specific subpopulations. B vitamin supplementation significantly reduced the risk of MACE if the percentage of men in the study was ,65% (RR, 0.94; 95% CI: 0.89-1.00; P = 0.034; Figure 7) and the study was a single-center trial (RR, 0.78; 95% CI: 0.66-0.93; P = 0.005; Figure 7).
Furthermore, we noted that B vitamin supplementation was associated with a nonsignificant reduction in MACE when the baseline homocysteine level was.14 mmol/L (RR, 0.93; 95% CI: 0.86-1.00; P = 0.051; Figure 7) or dietary grain fortification had been taken to boost the folate level (RR, 0.96; 95% CI: 0.91-1.00; P = 0.061; Figure 7). No other significant differences were identified in the predefined factors between those who took B vitamin supplements and those who took placebo (Figure 7). Furthermore, there was no significant difference in the effect of B vitamins on MACE between the 2 subgroups by factors that could affect the treatment effect except center of trials (P = 0.008; Figure 7).

Publication Bias
Review of funnel plots did not rule out the potential for publication bias for MACE, total mortality, cardiac death, MI, and stroke ( Figure S7). However, the Egger [53] and Begg tests [54] showed no evidence of publication bias for MACE (P value for Egger, 0.342; P value for Begg, 0.267), total mortality (P value for Egger, 0.312; P value for Begg, 0.183), cardiac death (P value for Egger, 0.631; P value for Begg, 0.621), MI (P value for Egger, 0.816; P value for Begg, 0.944), and stroke (P value for Egger, 0.992; P value for Begg, 1.000).

Discussion
Several observational studies [1][2][3][4][5] have suggested that B vitamin supplementation may improve major cardiovascular outcomes. However, this effect has not been confirmed to date. In the present study, we included RCTs and explored all possible correlations between B vitamin supplementation and the outcomes of MACE, total mortality, cardiac death, MI, and stroke. Furthermore, we conducted a cumulative meta-analysis to explore the value of routine B vitamin supplementation in clinical practice. This large quantitative study included 57,952 individuals from 24 trials with a broad range of baseline characteristics. Our results suggest that B vitamin supplementation has no significant effect on MACE, total mortality, cardiac death, MI, or stroke. Furthermore, in a cumulative meta-analysis, this B vitamin effect persisted and remained nonsignificant.
We reviewed previous meta-analyses and found that the hypothesized protective effect of B vitamins comes from observational studies [9], which, we suggest, may overestimate the effect on major cardiovascular outcomes. Several systematic reviews and meta-analyses of RCTs [12][13][14][15][16][17] have evaluated the impact of folic acid supplementation on major cardiovascular outcomes and have found no evidence to support a significant association. In a metaanalysis, Wang et al. [19] indicated that B vitamins significantly reduced the risk of stroke, but subgroup analysis showed that this effect was based on trials that included only participants without a previous stroke history. Lee et al. [55] suggested that folic acid supplementation administered as primary prevention significantly reduced the risk of stroke at.3 years of follow-up and that the homocysteine-lowering effect was.20%. However, this metaanalysis included trials with a male: female ratio of .2. Gao et al. [56] found that B vitamin supplementation given to reduce homocysteine levels influenced the risk of stroke events, especially in subgroups with a follow-up of.3 years (RR, 0.92; 95% CI: 0.84-1.00), without a history of grain fortification (RR, 0.91; 95% CI: 0.83-1.00), or without chronic kidney disease (RR, 0.93; 95% CI: 0.86-1.01), but these differences were not significant. In the present study, all pooled RR estimate points for stroke were ,1 (evidence accumulated up to 2006) with a potential trend toward moving leftward in the cumulative meta-analysis of B vitamin supplementation. We suggest a potential protective effect of B vitamins on stroke events. However, this trend is not obvious and requires validation.
There was no significant difference between B vitamin supplementation and placebo for the RR of major cardiovascular outcomes. However, several trials included in our study reported inconsistent results. Schnyder et al. [21] indicated that homocysteine-lowering therapy with B complex vitamins reduced the risk of MACE in patients after percutaneous coronary intervention. However, Lange et al. [24] indicated that B vitamin complex supplementation after coronary stenting increases the risk of instent restenosis and the need for target-vessel revascularization. The possible reasons for these inconsistencies are as follows: (1) a history of B vitamin supplementation may mask treatment effect; (2) patients in our meta-analysis had variable disease status, which also affects major cardiovascular outcomes; (3) elevated homocysteine levels impair vascular function, but the effect of homocysteine on cardiovascular outcomes may not be directly related to coronary disease pathogenesis, which is largely attributable to plaque formation and rupture [57]; and (4) high doses of B vitamin may adversely affect vascular remodeling and myocardial repair and may increase complications and death in patients with cardiovascular disease [58]. Therefore, although B vitamin supplementation may have direct effects on major cardiovascular outcomes, these effects may be counterbalanced by other effects.
Cumulative meta-analyses suggest that a nonsignificant response persisted for MACE, total mortality, cardiac death, MI, and stroke. However, the proposed nonsignificant protective effect of B vitamin supplementation for MI has been refuted by the NORVIT Trial Investigators [26], who specifically included individuals with MI, which may have contributed to higher recurrence and suggests a nonsignificant harmful effect for B vitamin supplementation. However, the proposed nonsignificant harmful effect of B vitamin supplementation for stroke was refuted by the HOPE 2 Trial [27]. This trial suggested that B vitamin supplementation reduced the risk of stroke by 25% (95% CI: 0.59-0.97), but this may be due to a considerably lower number of stroke events than coronary events. In addition, the CIs of the estimated risk reduction were wide. Furthermore, the results were not adjusted for the multiplicity of potential outcomes.
In our study, subgroup analysis suggested that B vitamin supplementation reduced the risk of MACE in specific subgroups. The main findings were as follows. First, although no significant difference was observed in MACE for the B vitamin supplementation group compared with the placebo group, in populations with a baseline total homocysteine level .14 mmol/L and those with a net decrease in homocysteine level of .20% there seemed to be a slight, but nonsignificant, benefit. In this study, baseline homocysteine levels were only available for whole populations, not individuals. Second, folate fortification of the grain supply may influence the risk of MACE, because it reduces the prevalence of low folate levels and high total homocysteine levels, so the mean difference in total homocysteine levels between groups narrows. Fortification reduces the number of participants with a high total homocysteine level, which is the population most likely to benefit from supplementation [59]. In this study, we found that B vitamin supplementation reduced the risk of MACE if participants had a background of grain fortification. Those with background fortification received higher doses of vitamin B12, which may be more effective at preventing MACE. Third, the protective effect of B vitamins was more evident if the percentage of men in the study was ,65%. Among men, there is a higher rate of smoking and drinking, both of which increase the risk of MACE. Therefore, the effect of B vitamins on the risk of MACE in studies with a higher percentage of male subjects may be reduced or balanced by a higher rate of smoking and drinking. Fourth, there was a significant difference between single-center and multi-center trials in the risk for MACE. This difference could be due to chance because only 6 trials were single-center, and they had a low sample size and low rate of cardiovascular disease occurrence. Fifth, no significant differences were observed in MACE between participants with a low folate level and those with a normal folate level. Folate level data were not available for several trials, so this conclusion was similar to those of a previous meta-analysis [60]. Sixth, the B vitamin dose showed no significant effect because all doses investigated were in the mega-dose range, and there may have been differences in the control groups in different studies. Seventh, B vitamin supplementation seems to offer a benefit for primary prevention of MACE but not for secondary prevention. Participants with pre-existing cardiovascular disease have higher recurrence and mortality rates. Finally, B vitamins seem to benefit participants with kidney disease but not those without kidney disease; patients with kidney disease may have higher homocysteine concentrations and a higher rate of cardiovascular disease.
The limitations of our study are as follows: (1) different types and doses of supplements might result in bias; (2) the use of background B vitamin supplementation might have impaired our ability to identify a treatment effect; (3) differences in diagnosis and reporting may have contributed to the differences in major cardiovascular outcomes in some trials; (4) stratified analyses based on background therapy in patients with previous disease were unavailable; (5) data on hemoglobin levels were not available, so we could not evaluate the potential confounding role of hemoglobin when evaluating the effect of vitamins B on the risk of major cardiovascular outcomes; (6) several trials with low Jadad score were included in our study, which could bias the results; and (7) since inherent assumptions are made for any meta-analysis, the analysis used pooled data, and individual patient data were unavailable; this restricted us from performing a more detailed analysis and obtaining more comprehensive results.
In conclusion, we found that B vitamin supplementation has little or no effect on major cardiovascular outcomes across various patient populations. We believe the use of B vitamin supplementation as a structured intervention in everyday clinical practice is not justified. A meta-analysis of individual patient data might be more appropriate to determine how the duration and dose of supplementation influence treatment outcomes.