Systematic Review and Meta-Analysis Medicine ®

OPEN

Effect of vitamin B supplementation

on cancer incidence, death due to cancer,
and total mortality
A PRISMA-compliant cumulative meta-analysis of randomized
controlled trials
Sui-Liang Zhang, MDa, Ting-Song Chen, MDb, Chen-Yun Ma, MDc, Yong-Bin Meng, MDd, Yu-Fei Zhang, MDa,
∗
Yi-Wei Chen, MDe, Yu-Hao Zhou, MDe,

Abstract
Background: Observational studies have suggested that vitamin B supplementation is associated with cancer risk, but this
association remains controversial. A pooled data-based meta-analysis was conducted to summarize the evidence from randomized
controlled trials (RCTs) investigating the effects of vitamin B supplementation on cancer incidence, death due to cancer, and total
mortality.
Methods: PubMed, EmBase, and the Cochrane Library databases were searched to identify trials to fit our analysis through August
2015. Relative risk (RR) was used to measure the effect of vitamin B supplementation on the risk of cancer incidence, death due to
cancer, and total mortality using a random-effect model. Cumulative meta-analysis, sensitivity analysis, subgroup analysis,
heterogeneity tests, and tests for publication bias were also conducted.
Results: Eighteen RCTs reporting the data on 74,498 individuals were included in the meta-analysis. Sixteen of these trials included
4103 cases of cancer; in 6 trials, 731 cancer-related deaths occurred; and in 15 trials, 7046 deaths occurred. Vitamin B
supplementation had little or no effect on the incidence of cancer (RR: 1.04; 95% confidence interval [CI]: 0.98–1.10; P = 0.216),
death due to cancer (RR, 1.05; 95% CI: 0.90–1.22; P = 0.521), and total mortality (RR, 1.00; 95% CI: 0.94–1.06; P = 0.952). Upon
performing a cumulative meta-analysis for cancer incidence, death due to cancer, and total mortality, the nonsignificance of the effect
of vitamin B persisted. With respect to specific types of cancer, vitamin B supplementation significantly reduced the risk of skin
melanoma (RR, 0.47; 95% CI: 0.23–0.94; P = 0.032).
Conclusion: Vitamin B supplementation does not have an effect on cancer incidence, death due to cancer, or total mortality. It is
associated with a lower risk of skin melanoma, but has no effect on other cancers.
Abbreviations: CI = confidence interval, ESRD = end-stage renal disease, IS = ischaemic stroke, MI = myocardial infarction,
RCT = randomized controlled trials, RR = relative risk.
Keywords: cancer, meta-analysis, mortality, vitamin B

Editor: Gokhan Cuce.

Names for PubMed indexing: S-LZ, T-SC, C-YM, Y-BM, Y-FZ, Y-WC, Y-HZ. Designed research: Y-HZ. Conducted research: S-LZ, Y-FZ, Y-HZ, Y-WC. Provided
essential reagents or provided essential materials: Y-HZ. Analyzed data or performed statistical analysis: Y-HZ and Y-WC. Wrote paper: Y-HZ and Y-WC. Had primary
responsibility for final content: Y-HZ. Revised the paper: T-SC, Y-HZ, C-YM.
All authors contributed to the planning, execution, and interpretation of the submitted manuscript and read and approved the final manuscript.
S-LZ, T-SC, C-YM contributed equally to this work.
This study was conducted under a grant from the talents training program of Shanghai seventh people’s hospital (QMX2015-01). The funders had no role in the study
design, data collection and analysis, decision to publish, or preparation of the manuscript.
The authors have no conflicts of interest to disclose.
a
Department of Invasive Technology, Shanghai Seventh People’s Hospital, b Department of Traditional Chinese and Western Medicine, Eastern Hepatobiliary Surgery
Hospital, Second Military Medical University, c Department of Medical Laboratory, Shanghai Seventh People’s Hospital, d Department of Traditional Chinese Medicine,
Changhai Hospital, Second Military Medical University, e Department of Rehabilitation Institute, Shanghai Seventh People’s Hospital, Shanghai, China.
∗
Correspondence: Yu-Hao Zhou, Department of Rehabilitation Institute, Shanghai Seventh People’s Hospital, Shanghai 200137, China (e-mail: zhou_ly@126.com).
Copyright © 2016 the Author(s). Published by Wolters Kluwer Health, Inc. All rights reserved.
This is an open access article distributed under the Creative Commons Attribution-ShareAlike License 4.0, which allows others to remix, tweak, and build upon the
work, even for commercial purposes, as long as the author is credited and the new creations are licensed under the identical terms.
Medicine (2016) 95:31(e3485)
Received: 28 September 2015 / Received in final form: 16 March 2016 / Accepted: 31 March 2016
http://dx.doi.org/10.1097/MD.0000000000003485

1
Zhang et al. Medicine (2016) 95:31 Medicine

1. Introduction ongoing trials were identified from the metaRegister of

The potential role of vitamin B in relation to the risk of cancer, Controlled Trials. Finally, manual searches were performed
from the reference lists within the entire relevant nonrandom-
including breast and colorectal cancer, has been investigated in
several observational studies.[1–3] Although the mechanism of ized controlled trials in order to identify the additional eligible
action is unclear, vitamin B may affect the incidence of cancer studies.
According to a standardized approach, 2 authors carried out
because it is essential for the biosynthesis of nucleotides,
replication of DNA, supply of methyl-groups, and the growth literature search, data extraction, and quality assessment
and repair of cells.[4–7] However, observational studies often independently. The primary author solved any differences until
overestimate the magnitude of the effect and do not prove a consensus was achieved if there were any inconsistencies
causality, and the effect of vitamin B supplementation on the risk between 2 authors. In order to avoid less confounding variables
of cancer has not been confirmed by randomized controlled trials or biases, we restricted our study design to RCTs rather than
(RCTs).[8–25] Finally, previous studies have not investigated the observational studies. A study was eligible for inclusion if the
following criteria were met: it was an RCT; the trial evaluated the
potential interaction of supplementation with both vitamin B6
and B12 and its effect on cancer risk.[26,27] effects of vitamin B supplementation compared with those of
The reasons for the discordance between the findings of placebo or low-dose vitamin B; a follow-up period was of at least
1 year; and the trial reported at least 1 outcome as either cancer
RCTs[8–25] and earlier observational studies[1–3] could be as
follows: individual trials might have been underpowered to show incidence or death due to cancer.
clinical benefit, especially if event rates were lower than expected,
which always acquired broad confidence intervals; the duration 2.2. Data collection and quality assessment
of follow-up was shorter than that needed to show a clinical
A standard protocol was adopted independently by 2 authors to
benefit, or different types of supplements might provide a biased
extract the data from all included trials, and any differentials
view of the study question; observational studies are hypothesis-
between these 2 authors were resolved for an agreement though a
generating but cannot prove causality, and always overestimate
group discussion. The collected data include study characteristics
the magnitude of the effect; and most trials were designed with
(first author or study group’s name, publication year, study
vascular events as the primary endpoint, and their sample size did
design, type of blinding, intervention regimes, controls, and the
not allow adequate power to detect potential clinically relevant
duration of follow-up.), participants’ characteristics (number of
differences in cancer incidence.
patients, mean age, percentage of men, background fortification,
The effect of vitamin B supplementation on primary and
current diseases status, baseline total homocysteine level, and
secondary prevention of adverse cardiovascular outcomes has
baseline folate level), and outcomes variables (cancer incidence,
been studied in numerous RCTs.[8–25] With long-term follow-up
death due to cancer, total mortality, and specific-type cancer
and collection of cancer data in a majority of studies, insight into
incidence). Simultaneously, the quality of included trials was
the risk of cancer among participants with vitamin B supplemen-
evaluated using Jadad score[29] which ranged from 0 to 5, and
tation and those with placebo can be derived. In this study, a
based on the following items such as randomization, concealment
meta-analysis of RCTs was conducted to evaluate the effect of
of the treatment allocation, blinding, completeness of follow-up,
vitamin B on cancer incidence, death due to cancer, and total
and the use of intention-to-treat analysis. In our analysis, we
mortality in specific subpopulations, in an attempt to determine
considered a study with a score of 4 or 5 to be of high quality.
the effect of folic supplementation interaction with vitamin B6
and B12 on the risk of cancer-related outcomes. In addition,
cumulative meta-analyses were employed to determine the 2.3. Statistical analysis
evidence base for routine vitamin B supplementation in clinical The results of each RCT were assigned as dichotomous frequency
practice. data, and the event numbers were extracted from each trial to
calculate relative risks (RRs) and 95% confidence intervals (CIs)
2. Methods of each individual trial. The overall RRs with 95% CIs were
calculated for cancer incidence, death due to cancer, total
2.1. Data sources, search strategy, and selection criteria mortality, and specific-type cancer in participants who received
This review and pooled data based meta-analysis was conducted vitamin B supplementation. The comparison of pooled RR
and reported according to the Preferred Reporting Items for between vitamin B supplementation and placebo was performed
Systematic Reviews and Meta-Analysis (PRISMA) Statement, using fixed-effect and random-effects models respectively, and
issued in 2009.[28] Ethics approval was not necessary for this then the results from the random-effects model were presented
study, as only deidentified pooled data from individual studies here.[30,31] The heterogeneity of the treatment effects among
were analyzed. RCTs on vitamin B supplementation, written in included trials was evaluated using Q statistic; meanwhile, a P
the English language, were eligible for inclusion in our study, value for heterogeneity of less than 0.10 was considered to be
regardless of the publication status (published, in press, or in statistically significant.[32,33] In the cumulative meta-analysis,
progress), and the effects of vitamin B supplementation on cancer outcome data for cancer incidence, death due to cancer, and total
incidence, death due to cancer, total mortality, and any specific- mortality were shown sequentially in light of the year included in
type cancer were examined. We systematically searched the trials which first became available.
PubMed, EmBase, and Cochrane Central Register of Controlled Potential heterogeneity in estimates of the treatment effects was
Trials to identify all the trials related to vitamin B supplementa- explored using univariate meta-regression.[34] Subgroup analyses
tion through August 2015. The electronic databases were were also performed for cancer incidence. The estimates between 2
searched using the following keywords. subsets were compared by using interaction tests, which were
“vitamin B” AND “randomized controlled trials” AND based on Student t distribution rather than on normal distribution
“clinical trials” AND “human” AND “English.” Furthermore, as the number of included studies was small.[35] Sensitivity analyses

2
Zhang et al. Medicine (2016) 95:31 www.md-journal.com

3. Results
The primary electronic search produced 13,334 articles in total.
After scanning titles and abstracts, 13,057 irrelevant or duplicate
articles were excluded during the initial review. The remaining
277 potentially eligible articles were retrieved after detailed
evaluations. Finally, 18 RCTs[8–25] were eligible for pooled
analysis (Fig. 1). Table 1 presents the general characteristics of
these included trials and baseline information of total 74,498
individuals. Of these, 3 trials[8,12,14] evaluated vitamin B
supplementation in patients with chronic renal disease or end-
stage renal disease, 7 trials[9–11,15,18–20] reported patients with
cardiovascular disease, 3 trials[13,21,23] evaluated patients with a
recent history of colorectal adenomas and no previous invasive
large intestine carcinoma, and the remaining 5 trials[16,17,22,24,25]
reported participants with cardiovascular risk factors. The
number of cases in each included trial ranged from 114 to
20,702 during the follow-up time of 2.0 to 7.3 years, the baseline
homocysteine level ranged from 9.6 to 31.7 mmol/L, the baseline
folate level ranged from 8.1 to 35.34 nmol/L, and the net change
in total homocysteine level ranged from 2.1 to 15.1 mmol/L.
In the intervention groups, the dose of folic acid ranged from 0.4
Figure 1. Flow diagram of the literature search and trials selection process. to 40 mg per day, that of vitamin B6 from 3.0 to 250 mg per day,
and that of vitamin B12 from 20 to 2000 mg per day. The
breakdown for the number of trials available for each outcome
was 16, 6, and 15 for cancer incidence,[9–11,13–25] death due to
by removing each individual trial from the meta-analysis were also cancer,[6,8,10,12,16,20] and total mortality,[8–11,13–21,23,24] respec-
conducted.[36] The publication bias for cancer incidence, death due tively. The quality of the trials was assessed using the Jadad
to cancer, and total mortality was statistically assessed using funnel score.[29] We considered a score ≥4 to indicate a high-quality
plots, Egger[37] and Begg tests,[38] and P values less than 0.05 was study. According to the Jaded scoring method, 6 tri-
considered to be statistically significant. STATA software (Version als[8,9,11,19,20,25] scored 5 points, another 6 trials[10,13–15,18,24]
10.0; Stata Corporation, College Station, TX, USA) was used to scored 4 points, 3[16,17,22] scored 3 points, 2[12,21] scored 2 points,
perform the statistical analyses. and the remaining 1[23] scored 1 point.

Table 1
Design and characteristic of trials included in our meta-analysis.
Baseline Baseline Dose of Dose of Dose of Net decrease Follow-
No. of Disease Background homocysteine folate level folic acid vitamin vitamin in homocysteine up
Source patients status fortification (umol/L) (nmol/L) (mg) B6 (mg) B12 (ug) (mmol/L) (y)
J Heinz[8] 650 ESRD No 29.0 14.1 2.5 10 25 8.6 2.1
VISP Trial Investigators[9] 3680 IS Yes 12.3 — 2.5 25 400 2.1 2.0
(HOPE) 2 Investigators[10] 5522 Vascular disease Parial 12.2 28.0 2.5 50 1000 3.3 5.0
or diabetes
NORVIT Trial Investigators[11] 3749 Had an acute MI No 13.1 10.95 0.8 40 400 2.3 3.3
within 7 days
M Righetti[12] 114 Hemodialysis No 31.7 22.32 5.0 250 500 15.1 2.4
Polyp Prevention Study Group[13] 1021 Colorectal adenomas Yes 9.8 23.70 1.0 — — — 7.0
Veterans Affairs 2056 Advanced chronic kidney Yes 22.4 35.34 40 100 2000 5.1 3.2
Site Investigators[14] disease or ESRD and high
homocysteine levels
WENBIT Study Group[15] 3096 Undergoing coronary No — 12.3 0.8 40 400 2.8 3.1
angiography
WAFACS Study Group[16] 5442 Health professionals No — 13.5 2.5 50 1000 — 7.3
BVAIT Research Group[17] 506 Initial tHcy >8.5 umol/L Yes 9.6 21.41 5.0 50 400 2.1 3.1
SEARCH Collaborative Group[18] 12,064 MI survivors No 13.5 16.76 2.0 — 1000 3.8 6.7
SU.FOL.OM3 Collaborative 2501 With a history of MI, No 12.8 15.29 0.56 3.0 20 2.7 4.7
Group[19] unstable angina, or IS
VITATOPS Study Group[20] 8164 Recent transient ischaemic Parial 14.3 — 2.0 25 500 3.8 3.4
attack or stroke
The ukCAP Trial Group [21]
939 Colorectal adenomas No — — 0.5 — — — 3.0
The VITRO Study Group[22] 701 hyperhomocysteinemic No 12.3 — 5.0 50 400 — 2.5
and health professionals
Wu K[23] 672 Colorectal adenomas Yes — — 1.0 — — — 5.3
CSPPT[24] 20,702 Hypertension No 12.5 8.1 0.8 — — — 4.5
B-PROOF[25] 2919 Elevated homocysteine Yes 14.4 18.9 0.4 — 500 4.4 2.0

= not available, ESRD = end-stage renal disease, IS = ischaemic stroke, MI = myocardial infarction.

3
Zhang et al. Medicine (2016) 95:31 Medicine

cancer incidence
Study RR (95% CI
) P value I-square P value for heterogeneity

VISP Trial Investigator

s 0.98 (0.74, 1.30) 0.900 . .

NORVIT Trial Investigator

s 0.94 (0.75, 1.17) 0.579 0.0% 0.613

(HOPE) 2 Investigator
s 1.02 (0.91, 1.15) 0.729 0.0% 0.600

The VITRO Study Group 1.03 (0.92, 1.16) 0.628 0.0% 0.541

Polyp Prevention Study Group 1.08 (0.91, 1.27) 0.383 32.0% 0.208

Veterans Affairs Site Investigator

s 1.04 (0.90, 1.21) 0.551 26.9% 0.233

The ukCAP Trial Group 1.04 (0.93, 1.17) 0.494 12.4% 0.335

WENBIT Study Group 1.06 (0.95, 1.19) 0.305 16.3% 0.302

WAFACS Study Group 1.04 (0.95, 1.15) 0.394 10.4% 0.349

BVAIT Research Group 1.04 (0.96, 1.13) 0.355 0.0% 0.444

Wu K 1.04 (0.96, 1.13) 0.365 0.0% 0.536

SEARCH Collaborative Group 1.05 (0.98, 1.12) 0.155 0.0% 0.618

SU.FOL.OM3 Collaborative Group 1.06 (0.99, 1.13) 0.096 0.0% 0.622

VITATOPS Study Group 1.04 (0.98, 1.10) 0.220 0.0% 0.469

B−PROOF 1.04 (0.98, 1.10) 0.207 0.0% 0.544

CSPPT 1.04 (0.98, 1.10) 0.216 0.0% 0.615

A .3 .5 1 2
death due to cancer

Study RR (95% CI) P value I-square P value for heterogeneity

(HOPE) 2 Investigators 0.99 (0.75, 1.31) 0.953 . .

M Righetti 0.99 (0.75, 1.30) 0.916 0.0% 0.802

WENBIT plus NORVIT Trial Investigators 1.08 (0.88, 1.33) 0.436 0.0% 0.576

WAFACS Study Group 1.02 (0.85, 1.22) 0.825 0.0% 0.451

J Hein
z 1.02 (0.86, 1.22) 0.795 0.0% 0.613

VITATOPS Study Group 1.05 (0.90, 1.22) 0.521 0.0% 0.704

B .3 .5 1 2

total mortality

Study RR (95% CI) P value I-square P value for heterogeneity

VISP Trial Investigators 0.86 (0.66, 1.11) 0.249 . .

NORVIT Trial Investigators 0.96 (0.79, 1.15) 0.640 17.8% 0.270

(HOPE) 2 Investigators 0.98 (0.89, 1.08) 0.694 0.0% 0.516

M Righetti 0.98 (0.89, 1.08) 0.659 0.0% 0.700

Polyp Prevention Study Group 0.97 (0.87, 1.07) 0.506 3.7% 0.385

Veterans Affairs Site Investigators 0.99 (0.93, 1.06) 0.833 0.0% 0.451

The ukCAP Trial Group 0.97 (0.88, 1.07) 0.555 28.4% 0.212

WENBIT Study Group 0.98 (0.90, 1.07) 0.689 20.4% 0.268

WAFACS Study Group 0.99 (0.92, 1.06) 0.697 9.6% 0.356

Wu K 0.97 (0.89, 1.06) 0.511 23.2% 0.229

J Heinz 0.98 (0.91, 1.07) 0.699 19.6% 0.257

SEARCH Collaborative Group 1.00 (0.94, 1.06) 0.970 15.0% 0.297

SU.FOL.OM3 Collaborative Group 1.01 (0.93, 1.09) 0.801 38.0% 0.080

VITATOPS Study Group 1.00 (0.94, 1.07) 0.911 35.4% 0.092

CSPPT 1.00 (0.94, 1.06) 0.952 32.5% 0.108

.3 .5 1
1 2
C
Figure 2. Cumulative meta-analysis of the effect of vitamin B supplementation on the risk of cancer incidence (A), death due to cancer (B), and total mortality (C).

Data from 73,269 participants were used to evaluate the effect Data from 26,729 participants were used to evaluate the effect of
of vitamin B supplementation on cancer incidence and included vitamin B supplementation on death due to cancer and included
4103 cancer events. Vitamin B supplementation caused an 731 cases of cancer-related mortality. Vitamin B supplementation
increase of 4% in cancer incidence; however, this was not a increased the death rate due to cancer by 5%, but the change was
significant change (RR: 1.04; 95% CI: 0.98–1.10; P = 0.216; not significant (RR, 1.05; 95% CI: 0.90–1.22; P = 0.521; without
without evidence of heterogeneity; Fig. 2A). evidence of heterogeneity; Fig. 2B).

4
Zhang et al. Medicine (2016) 95:31 www.md-journal.com

Figure 3. Effect of vitamin B supplementation on specific-type cancer.

Data from 69,744 participants were used to evaluate the effect The effects of vitamin B supplementation on the risk of specific
of vitamin B supplementation on total mortality and included types of cancer were also evaluated. Overall, vitamin B
7046 death events. There were no significant differences in total supplementation was associated with a significantly reduced risk
mortality between participants receiving vitamin B and those of skin melanoma (RR, 0.47; 95% CI: 0.23–0.94; P = 0.032;
receiving placebo (RR, 1.00; 95% CI: 0.94–1.06; P = 0.952; with Fig. 3), whereas it had no significant effect on the risk of
moderate heterogeneity; Fig. 2C). A sensitivity analysis was gastrointestinal cancer (RR, 1.02; 95% CI: 0.87–1.19; P = 0.849),
conducted for total mortality. However, after sequential genitourinary cancer (RR, 1.09; 95% CI: 0.88–1.34; P = 0.445),
exclusion of each trial, the conclusion was not affected by the hematological cancer (RR, 1.08; 95% CI: 0.79–1.49; P = 0.625),
exclusion of any specific trial. respiratory and intrathoracic cancer (RR, 1.07; 95% CI:
When a cumulative meta-analysis for cancer incidence was 0.90–1.27; P = 0.470), breast cancer (RR, 0.82; 95% CI:
carried out, the original nonsignificant result for an effect of vitamin 0.63–1.07; P = 0.149), and other types of cancers (RR, 1.26;
B persisted; the effect was slight and borderline nonsignificant. 95% CI: 0.99–1.60; P = 0.056).
Similarly, the nonsignificant result persisted when cumulative meta- Heterogeneity testing for the analysis showed a P >0.10 for
analyses for death due to cancer and total mortality were conducted. cancer incidence and death due to cancer; no significant

5
Zhang et al. Medicine (2016) 95:31 Medicine

Figure 4. Subgroup analysis for cancer incidence.

heterogeneity was observed in the overall analysis, which Subgroup analyses were conducted for cancer incidence, death
suggests that most variation was attributable to chance alone. due to cancer, and total mortality to minimize heterogeneity and
However, moderate heterogeneity was observed in the magnitude explore the effect of vitamin B supplementation in any specific
of the effect on total mortality across the trials. Meta-regression subpopulations. Vitamin B supplementation might play an
analyses were performed[34] for cancer incidence that included important role in cancer incidence if the mean age of the
the mean age, baseline homocysteine level, baseline folate level, participants is <62 years (RR, 1.15; 95% CI: 0.99–1.34; P =
dose of folic acid, dose of vitamin B6, dose of vitamin B12, and 0.074; Fig. 4), and baseline homocysteine levels are <14 mmol/L
duration of follow-up. The results indicated that these variables (RR, 1.07; 95% CI: 0.99–1.14; P = 0.075; Fig. 4), although these
were not significant factors contributing to the association results were not statistically significant. When subgroup analyses
between vitamin B supplementation and cancer incidence (data based on other factors were carried out, no significant differences
not shown). were observed between vitamin B supplementation and placebo.

6
Zhang et al. Medicine (2016) 95:31 www.md-journal.com

gastric cancer.[50] In addition, Zhang et al[53] suggested that folic

acid supplementation might affect subsequent lung cancer risk in
men. Tio et al[52] indicated that high blood folate level was
associated with an increased risk of prostate cancer. However, the
hypothesized effect of vitamin B supplementation comes from
meta-analyses of observational studies, which may overestimate its
effect on the incidence of specific types of cancer. Three important
systematic reviews and meta-analyses of RCTs have evaluated the
impact of folic acid supplementation on cancer incidence and have
found no evidence to support a significant effect.[26,27,54] Clarke
et al[26] performed a meta-analysis of 8 RCTs involving 37,485
individuals and found that vitamin B supplementation had no
significant effect on cancer incidence (RR, 1.05; 95% CI:
0.98–1.13), cancer mortality (RR, 1.00; 95% CI: 0.85–1.18),
and total mortality (RR, 1.02; 95% CI: 0.97–1.08) during the
whole scheduled treatment period or in the subsequent years.
Vollset et al[27] suggested that folic acid supplementation was
Figure 5. Funnel plot for cancer incidence. associated with higher plasma concentrations of folic acid, but had
no significant effect on cancer incidence. For specific types of
cancers, there was no significant difference between vitamin B and
placebo for cancer at any specific sites. Qin et al[54] indicated that
Furthermore, there was no significant difference in the effect of while folic acid supplementation had no significant effect on total
vitamin B supplementation between the 2 subgroups with respect cancer incidence, and the incidence of colorectal, prostate, lung,
to cancer incidence. Finally, similar nonsignificant results were breast, and hematological malignancy cancers, it significantly
detected for death due to cancer and total mortality (data not reduced the risk of melanoma. In the present study, all pooled RR
shown). estimate points for cancer incidence were >1 (evidence accumulat-
A review of funnel plots did not rule out the potential for ed up to 2006) with a potential trend toward moving rightward in
publication bias for cancer incidence. The Egger[37] and Begg the cumulative meta-analysis of vitamin B supplementation. A
tests,[38] however, showed no evidence of publication bias for potentially harmful effect of vitamin B on total cancer incidence was
cancer incidence (P value for Egger: 0.183; P value for Begg: detected, but this trend was nonsignificant and requires validation.
0.893; Fig. 5). For death due to cancer and total mortality, the nonsignificant
effects persisted and remained.
There was no significant difference between vitamin B
4. Discussion
supplementation and placebo in terms of the effect on the
Previous observational studies[39–46] have suggested that vitamin relative risks of cancer incidence, death due to cancer, and total
B supplementation has a marked effect on cancer incidence. mortality. Cumulative findings of out meta-analysis indicated,
However, observational studies may overestimate the effect of with evidence accumulated up to 2006, that the pooled RR
vitamin B supplementation. So far, the effect of vitamin B estimate points for cancer incidence were >1. A study conducted
supplementation on the risk of cancer has not been confirmed by by Ulrich and Potter[55] indicated that folic acid may have
any RCT. We therefore conducted a meta-analysis of RCTs to influenced growth in cancers that were silent at baseline or during
quantitatively assess the effect of vitamin B supplementation on trials, leading to excess subsequent clinical surfacing and
the risk of cancer-related outcomes, and the findings of our study, diagnosis during extended follow-up. Furthermore, evidence
which has a large sample size, are potentially more robust than suggests that aggressive supplementation may enhance the
those of any individual trial. In this study, RCTs were included to growth of established, microscopic lesions.[56] Data on the
explore all possible correlations between vitamin B supplemen- incidence of specific types of cancer were available in our study;
tation and the outcomes of cancer incidence, death due to cancer, however, no significant difference was detected between vitamin
and total mortality. This comprehensive, large-scale, quantitative B supplementation and placebo, except in the case of skin
study included 74,498 individuals from 18 trials with a broad melanoma. These results may be attributed to chance, because a
range of baseline characteristics. The findings of our study small number of trials[10,16,20] were included.
indicated that vitamin B supplementation has no significant effect Subgroup analyses were conducted for cancer incidence, death
on cancer incidence, death due to cancer, and total mortality. due to cancer, and total mortality. No significant effect on cancer
Considering specific cancer types, vitamin B significantly reduced incidence was observed in subpopulations with a mean age less
the risk of skin melanoma, but did not have any significant effect than 62 years and those with a baseline homocysteine level >14 m
on other types of cancers. In a cumulative meta-analysis, the mol/L; the effect seemed to be slight, but nonsignificant. In the
effect of vitamin B on cancer incidence, death due to cancer, and current study, mean age and baseline homocysteine levels in
total mortality persisted and remained nonsignificant. participants were available for whole populations; individual
Several meta-analyses have evaluated the impact of vitamin B data were not available, which prevented us from performing
supplementation on the risk of cancer-related outcomes.[26,27,47–54] more comprehensive analyses. Furthermore, participants with
For meta-analysis based on observational studies, folic acid different backgrounds and intervention regimens might contrib-
supplementation was associated with a lower risk of oral and ute to the biased treatment effect. Finally, nearly all included trials
pharyngeal,[47] breast,[48] bladder,[49] esophageal and pancreatic included participants from the Western countries except 1 trial,
cancer.[50] Furthermore, dietary folic acid supplementation was not which specifically included Chinese people.[24] The findings of
associated with the risk of colorectal,[51] prostate,[52] lung,[53] and CSPPT were consistent with those of trials conducted in Western

7
Zhang et al. Medicine (2016) 95:31 Medicine

countries. Furthermore, alimentation habits might play an [13] The Polyp Prevention Study GroupFolic acid for the prevention of
colorectal adenomas a randomized clinical trial. JAMA 2007;297:2351–9.
important role in the risk of cancer;[24] however, data about
[14] The Veterans Affairs Site InvestigatorsEffect of homocysteine lowering
alimentation status were not available to us. Therefore, we just on mortality and vascular disease in advanced chronic kidney disease and
performed a relative comprehensive review to evaluate the effect end-stage renal disease: a randomized controlled trial. JAMA 2007;298:
of vitamin B on the risk of cancer, death due to cancer, and total 1163–70.
mortality. [15] The WENBIT Study GroupMortality and cardiovascular events in
patients treated with homocysteine-lowering B vitamins after coronary
The present meta-analysis has certain limitations. First, angiography: a randomized controlled trial. JAMA 2008;300:795–804.
different types and doses of vitamin B supplements were [16] The WAFACS Study GroupEffect of combined folic acid, vitamin B6,
included, which could bias the results. Second, the background and vitamin B12 on cancer risk in women: a randomized trial. JAMA
among participants taking vitamin B might have impaired our 2008;300:2012–21.
[17] The BVAIT Research GroupHigh-dose B vitamin supplementation and
ability to identify the treatment effect. Third, the different results
progression of subclinical atherosclerosis: a randomized controlled trial.
of cancer surveillance and reporting may lead to various Stroke 2009;40:730–6.
incidences of cancer among trials. Fourth, patients who have [18] Study of the Effectiveness of Additional Reductions in Cholesterol and
had background therapy for previous diseases were not available Homocysteine (SEARCH) Collaborative GroupEffects of homocysteine-
in stratified analyses. Fifth, several included trials with low Jadad lowering with folic acid plus vitamin B12 vs placebo on mortality and
major morbidity in myocardial infarction survivors: a randomized trial.
score, which hampered the quality of our work. Finally, more JAMA 2010;303:2486–94.
detailed relevant analysis could be restricted by conducting [19] The SU.FOL.OM3 Collaborative GroupEffects of B vitamins and omega
analysis using pooled data instead of individual data. 3 fatty acids on cardiovascular diseases: a randomised placebo controlled
In conclusion, vitamin B supplementation has no significant Trial. BMJ 2010;341:c6273.
[20] The VITAmins TO Prevent Stroke (VITATOPS) Trial Study Group-
effect on cancer incidence, death due to cancer, and total
Treatment with B vitamins and incidence of cancer in patients with
mortality. Subgroup analyses suggested that vitamin B might previous stroke or transient ischemic attack results of a randomized
have a detrimental effect on cancer incidence when the mean age placebo-controlled trial. Stroke 2012;43:1572–7.
of the participants was less than 62 years and baseline [21] The ukCAP Trial GroupAspirin and folic acid for the prevention of
homocysteine levels were >14 mmol/L. In addition, vitamin B recurrent colorectal adenomas. Gastroenterology 2008;134:29–38.
[22] The Vitamins and Thrombosis (VITRO) Study GroupHomocysteine
supplementation significantly reduced the risk of skin melanoma. lowering by B vitamins and the secondary prevention of deep vein
Future trials should focus on specific younger subpopulations and thrombosis and pulmonary embolism: a randomized, placebo-con-
participants with baseline homocysteine level >14 mmol/L. We trolled, double-blind trial. Blood 2007;109:139–44.
suggest that ongoing trials should be improved in the following [23] Wu K, Platz EA, Willett WC, et al. A randomized trial on folic acid
supplementation and risk of recurrent colorectal adenoma. Am J Clin
ways: total cancer incidence, and death due to cancer or any
Nutr 2009;90:1623–31.
specific type of cancer should be recorded and reported [24] CSPPT InvestigatorsEfficacy of folic acid therapy in primary prevention
normatively, and it should be evaluated in future trials, and of stroke among adults with hypertension in China: the CSPPT
the role of intervention duration and dosage of supplementation randomized clinical trial. JAMA 2015;313:1325–35.
should be taken into consideration before evaluating clinical [25] B-PROOF Study GroupEffect of daily vitamin B-12 and folic acid
supplementation on fracture incidence in elderly individuals with an
outcomes. elevated plasma homocysteine concentration: B-PROOF, a randomized
controlled trial. Am J Clin Nutr 2014;100:1578–86.
References [26] Clarke R, Halsey J, Lewington S, et al. Effects of lowering homocysteine
levels with B vitamins on cardiovascular disease, cancer, and cause-
[1] Larsson SC, Giovannucci E, Wolk A. Folate and risk of breast cancer: a specific mortality: meta-analysis of 8 randomized trials involving 37 485
meta-analysis. J Natl Cancer Inst 2007;99:64–76. individuals. Arch Intern Med 2010;170:1622–31.
[2] Giovannucci E. Epidemiologic studies of folate and colorectal neoplasia: [27] Vollset SE, Clarke R, Lewington S, et al. Effects of folic acid
a review. J Nutr 2002;132(suppl):2350S–5S. supplementation on overall and site-specific cancer incidence during
[3] Sanjoaquin MA, Allen NA, Couto E, et al. Folate intake and colorectal the randomised trials: meta-analyses of data on 50,000 individuals.
cancer risk: a meta-analytical approach. Int J Cancer 2005;113:825–38. Lancet 2013;381:1029–36.
[4] Smith AD, Kim YI, Refsum H. Is folic acid good for everyone? Am J Clin [28] Moher D, Liberati A, Tetzlaff J, et al. PRISMA GroupPreferred reporting
Nutr 2008;87:517–33. items for systematic reviews and meta-analyses: the PRISMA statement.
[5] Liu L, Wylie RC, Andrews LG, et al. Aging, cancer and nutrition: the PLos Med 2009;6:e1000097.
DNA methylation connection. Mech Ageing Dev 2003;124:989–98. [29] Jadad AR, Moore RA, Carroll D. Assessing the quality of reports of
[6] Ebbing M, B⊘naa KH, Nygard O, et al. Cancer incidence and mortality randomized clinical trials: is blinding necessary? Control Clin Trials
after treatment with folic acid and vitamin B12. JAMA 2009;302: 1996;17:1–2.
2119–26. [30] DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin
[7] Zhou Y-H, Tang J-Y, Wu M-J, et al. Effect of folic acid supplementation Trials 1986;7:177–88.
on cardiovascular outcomes: a systematic review and meta-analysis. [31] Ades AE, Lu G, Higgins JP. The interpretation of random-effects
PLoS One 2011;6:e25142. metaanalysis in decision models. Med Decis Making 2005;25:646–54.
[8] Heinz J, Kropf S, Domröse U, et al. B vitamins and the risk of total [32] Deeks JJ, Higgins JPT, Altman DG. Higgins J, Green S. Analyzing data
mortality and cardiovascular disease in end-stage renal disease: results of and undertaking meta-analyses. Cochrane Handbook for Systematic
a randomized controlled trial. Circulation 2010;121:1432–8. Reviews of Interventions 5. 0. 1. Oxford, UK:The Cochrane Collabora-
[9] The VISP Trial InvestigatorsLowering homocysteine in patients with tion; 2008;chap 9.
ischemic stroke to prevent recurrent stroke, myocardial infarction, and [33] Higgins JPT, Thompson SG, Deeks JJ, et al. Measuring inconsistency in
death: the vitamin intervention for stroke prevention (VISP) randomized meta-analyses. BMJ 2003;327:557–60.
controlled trial. JAMA 2004;291:565–75. [34] Thompson SG, Higgins JP. How should meta-regression analyses be
[10] The Heart Outcomes Prevention Evaluation (HOPE) 2 Investigator- undertaken and interpreted? Stat Med 2002;21:1559–73.
sHomocysteine lowering with folic acid and B vitamins in vascular [35] Altman DG, Bland JM. Interaction revisited: the difference between two
disease. N Engl J Med 2006;354:1567–77. estimates. BMJ 2003;326:219.
[11] The NORVIT Trial InvestigatorsHomocysteine lowering and cardiovas- [36] Tobias A. Assessing the influence of a single study in meta-analysis. Stata
cular events after acute myocardial infarction. N Engl J Med 2006;354: Tech Bull 1999;47:15–7.
1578–88. [37] Egger M, Davey Smith G, Schneider M, et al. Bias in meta-analysis
[12] Righetti M, Serbelloni P, Milani S, et al. Homocysteine-lowering vitamin detected by a simple, graphical test. BMJ 1997;315:629–34.
B treatment decreases cardiovascular events in hemodialysis patients. [38] Begg CB, Mazumdar M. Operating characteristics of a rank correlation
Blood Purif 2006;24:379–86. test for publication bias. Biometrics 1994;50:1088–101.

8
Zhang et al. Medicine (2016) 95:31 www.md-journal.com

[39] Su LJ, Arab L. Nutritional status of folate and colon cancer risk: evidence [48] Chen P, Li C, Li X, et al. Higher dietary folate intake reduces the breast
from NHANES I epidemiologic follow-up study. Ann Epidemiol cancer risk: a systematic review and meta-analysis. Br J Cancer
2001;11:65–72. 2014;110:2327–38.
[40] Wei EK, Giovannucci E, Wu K, et al. Comparison of risk factors for [49] He H, Shui B. Folate intake and risk of bladder cancer: a meta-analysis of
colon and rectal cancer. Int J Cancer 2004;108:433–42. epidemiological studies. Int J Food Sci Nutr 2014;65:286–92.
[41] Ferraroni M, La Vecchia C, D’Avanzo B, et al. Selected micronutrient [50] Tio M, Andrici J, Cox MR, et al. Folate intake and the risk of upper
intake and the risk of colorectal cancer. Br J Cancer 1994;70:1150–5. gastrointestinal cancers: a systematic review and meta-analysis. J
[42] Lajous M, Romieu I, Sabia S, et al. Folate, vitamin B12 and Gastroenterol Hepatol 2014;29:250–8.
postmenopausal breast cancer in a prospective study of French women. [51] Heine-Bröring RC, Winkels RM, Renkema JM, et al. Dietary supplement
Cancer Causes Control 2006;17:1209–13. use and colorectal cancer risk: a systematic review and meta-analyses of
[43] Levi F, Pasche C, Lucchini F, et al. Dietary intake of selected prospective cohort studies. Int J Cancer 2015;136:2388–401.
micronutrients and breast-cancer risk. Int J Cancer 2001;91:260–3. [52] Tio M, Andrici J, Cox MR, et al. Folate intake and the risk of prostate
[44] Shrubsole MJ, Jin F, Dai Q, et al. Dietary folate intake and breast cancer cancer: a systematic review and meta-analysis. Prostate Cancer Prostatic
risk: results from the Shanghai Breast Cancer Study. Cancer Res Dis 2014;17:213–9.
2001;61:7136–41. [53] Zhang YF, Zhou L, Zhang HW, et al. Association between folate intake
[45] Adzersen KH, Jess P, Freivogel KW, et al. Raw and cooked vegetables, and the risk of lung cancer: a dose-response meta-analysis of prospective
fruits, selected micronutrients, and breast cancer risk: a case-control studies. PLoS One 2014;9:e93465.
study in Germany. Nutr Cancer 2003;46:131–7. [54] Qin X, Cui Y, Shen L, et al. Folic acid supplementation and cancer risk: a
[46] Lajous M, Lazcano-Ponce E, Hernandez-Avila M, et al. Folate, vitamin B meta-analysis of randomized controlled trials. Int J Cancer 2013;133:
(6), and vitamin B(12) intake and the risk of breast cancer among 1033–41.
Mexican women. Cancer Epidemiol Biomarkers Prev 2006;15:443–8. [55] Ulrich CM, Potter JD. Folate and cancer: timing is everything. JAMA
[47] Galeone C, Edefonti V, Parpinel M, et al. Folate intake and the risk of 2007;297:2408–9.
oral cavity and pharyngeal cancer: a pooled analysis within the [56] im Y-I, Salomon RN, Graeme-Cook F, et al. Dietary folate protects
International Head and Neck Cancer Epidemiology Consortium. Int J against the development of macroscopic colonic neoplasia in a dose
Cancer 2015;136:904–14. responsive manner in rats. Gut 1996;39:732–40.