Vitamin B6 and Cancer Risk: A Field Synopsis and Meta-analysis

Simone Mocellin; Marta Briarava; Pierluigi Pilati


J Natl Cancer Inst. 2016;109(3) 

In This Article


We systematically reviewed and meta-analyzed the evidence regarding the relationship between vitamin B6 intake and blood levels with the risk of cancer. To this aim, we included both observational (prospective and retrospective) and intervention studies, gathering together the largest collection of data ever reported on this subject (almost 100 000 cancer cases).

The results of observational data support a strong association between both vitamin B6 dietary intake and PLP blood levels and the risk of cancer from any site, in terms of both high vs low category comparison and dose-response relationship. However, the high degree of between-study heterogeneity does not allow us to draw definitive conclusions on the role of vitamin B6 in the development of malignant disease in general. When we considered cancers by site, the most consistent findings were those regarding gastrointestinal tumors (both as single tumors—especially colorectal carcinoma—and as a whole), which are based on a large number of studies (n = 45) and consequently a high number of cases (n = 34 428). For the gastrointestinal tract, the inverse association between cancer risk and both vitamin intake and blood levels was statistically significant across all subgroup analyses (by study design and ethnicity), heterogeneity remaining an issue for intake (moderate level of evidence) but not for blood levels results (high level of evidence, although some limitations should be considered, as discussed below).

Nevertheless, other available data do not appear to be in line with these findings. For example, the evidence from vitamin B6 total intake (dietary and supplements) is overall less convincing, although the association with gastrointestinal cancer risk does remain statistically significant. This observation raises the question of whether vitamin B6 can exert a direct preventive effect against cancer development or if it rather represents an indicator of the presence of other protective micronutrients in a healthy diet. The strong and homogeneous association between cancer risk and vitamin B6 blood levels might tip the balance in favor of a direct role of the vitamin against carcinogenesis. On the other side, another piece of evidence appears to oppose this hypothesis: In fact, the meta-analysis of RCTs testing the cancer-preventive role of vitamin B6 found no proof of such an effect. However, this evidence was graded as low level because of the high risk of bias, which is mainly due to the fact that trials did not test vitamin B6 alone (the co-administration of other vitamins can be regarded as a confounding factor) and because their focus was not cancer (which means that participants had not followed up with the aim of detecting the occurrence of cancer); furthermore, evidence from RCTs regarded the risk of malignancies arising from any site of the body (no meta-analysis could be conducted for single tumor types): In particular, no data were available for gastrointestinal carcinomas, for which the evidence from observational studies is strongest.

Regarding tumors from other sites, the evidence for an association is weaker simply because of lack of evidence (eg, prostate carcinoma); because favorable dietary intake data are not confirmed by PLP levels (eg, breast carcinoma) or vice versa (eg, lung carcinoma); or because positive data existing for intake are coupled with a complete lack of information on PLP levels (eg, upper airway tumors), which calls for further investigation in this field.

Overall, our findings provide two main points: 1) vitamin B6 should be further considered as a risk reduction strategy for gastrointestinal cancers; in particular, dedicated RCTs appear justified, possibly with the enrollment of subjects selected on the basis of their personal risk as determined by already established risk factors, as well as according to their blood levels of PLP; 2) PLP blood levels should be considered a new cancer risk factor and their value as biomarker of cancer predisposition validated in large screening programs, possibly along with other biomarkers such as germline polymorphisms involved in carcinogenesis. As a corollary, the current daily recommended dose of vitamin B6 (1.5 mg) might be revised upwards; alternatively, a diet based on foods rich in this vitamin might be endorsed.

Limitations of the present study must also be acknowledged. The most important is represented by the confounding factors potentially influencing the results; in particular, both vitamin B6 intake and plasma PLP levels might be surrogate markers for the assumption of other dietary nutrients with anticancer effects. This is why some investigators prefer to rely on findings obtained from food studies rather than nutrient studies. In the case of vitamin B6, food containing this vitamin, such as fruit and vegetables, does contain other potential anticancer nutrients; however, vitamin B6 is also present in other food (such as meat) that is associated with an increased cancer risk, which might lead to a balance between opposite effects of nutrients present in vitamin B6–containing food. Overall, the true effect of vitamin B6 (like any other nutrient) remains very difficult to examine separately from that of other nutrients. Analogously, PLP plasma levels have been associated with inflammation, which in turn is associated with cancer risk; therefore, PLP might not be primarily involved in the carcinogenesis process but might rather represent a surrogate of the inflammatory status of a given person: again, discerning between the two conditions is an unmet challenge. These considerations call for the conduction of large studies where all these issues are properly addressed in order to enable investigators to provide clinicians with more robust and possibly definitive evidence on this subject.

Another limitation of the available evidence is that we found no data on some tumor types (such as malignant melanoma and soft tissue sarcomas) and very scarce data on others (such as childhood malignancies), which deserve attention by both basic and clinical researchers. As mentioned above, for other cancers (eg, breast and lung carcinomas) the evidence is not fully consistent. These observations, coupled with the heterogeneity of results that often remain unexplained (especially as regards intake data), suggest that other unveiled factors might act as confounders and should be carefully investigated in order to further dissect and ultimately fully exploit the cancer-protective potential of vitamin B6. For example, single studies demonstrated different vitamin effects on cancer risk based on primary tumor subsets (eg, colon vs rectal cancer;[21,22] hormone receptor–positive vs –negative breast cancer[23]) or in some participants subgroups (eg, women in pre- vs postmenopause;[24,25] men vs women[26]). Moreover, available data did not enable us to address the issue of the relationship between cancer risk and the interaction of vitamin B6 with other micronutrients (eg, vitamin B12, folate, methionine) or germline polymorphisms involved in one carbon metabolism, which has been advocated in some studies.[27–30]

Finally, we acknowledge that other meta-analyses have been already published on this topic over the past few years. They all focused on single tumor sites (colorectal, breast, and kidney cancer, respectively) and observational studies; ours is the first synopsis in this field (including intervention studies). In one article dedicated to colorectal cancer and published 10 years ago,[31] the authors restricted the search to prospective studies in order to avoid the limits of retrospective design (eg, recall bias); in our work—based on a larger number of studies and participants—we demonstrated that the association is statistically significant in both retrospective and prospective studies; moreover, we extended this observation to other gastrointestinal cancers, for which the evidence can be currently categorized as high. Other investigators could not demonstrate a statistically significant relationship between vitamin intake and breast cancer risk,[25] which is inconsistent with the present study. Yet, some authors did not consider the effect of PLP blood levels (which is instead the source of strong evidence in favor of the protective activity of vitamin B6),[32] and others based their conclusion regarding the relationship between PLP levels and kidney cancer on a single series,[33] a finding we could confirm based on multiple studies.

In conclusion, our work provides both clinicians and researchers with the first comprehensive and quantitative overview of the available evidence that links vitamin B6 to cancer risk. Although observational studies suggest an inverse association between higher intake and blood concentrations of vitamin B6, randomized trials do not support these findings. Further studies are needed to clarify the association between vitamin B6 and cancer risk, particularly gastrointestinal cancers.