B vitamins are common ingredients of multivitamins, of course, but also of energy boosters, such as energy drinks. But while they are best known for their role in energy metabolism,^[1] they may play a role in cancer biology through partaking in one-carbon metabolism^[2] and thus in methylation reactions and DNA synthesis.

This hypothesis was substantiated in 2015 when a paper published by the New England Journal of Medicine caused a stir by reporting that nicotinamide (a form of vitamin B3 also known as niacinamide) could reduce the rate of new non-melanoma skin cancers.^[3]

B vitamins had gained an “anti-cancer” reputation.

Yet it was just one study showing that one form of vitamin B3 could reduce the rate of one type of skin cancer; it didn’t preclude the possibility that some B vitamins could worsen at least some types of cancers.

To look for other possible connections between B vitamin supplementation and cancer, Dr. Theodore Brasky at The Ohio State University, in collaboration with colleagues at the Fred Hutchinson Cancer Research Center and at University of Taipei, performed a large observational study.^[4] Since its publication in the Journal of Clinical Oncology, in 2017, this study has taken the supplement world by storm, for it linked the vitamins B₆ and B₁₂ each with a 30–40% increase in overall risk of lung cancer in men.

Let’s take a closer look at the study.

Study design

To look for possible connections between B vitamin supplementation and lung cancer, the researchers analyzed data from the 77,000 participants in the VITamins And Lifestyle (VITAL) prospective cohort study.^[5] The study itself was designed to look for possible associations between cancer risk and vitamin, mineral, and non-vitamin/non-mineral supplementation.

The researchers chose to focus on the vitamins B₆, B₁₂, and B₉, which play an important role in the one-carbon pathways and thus are most likely to affect carcinogenesis. The study participants, all residents of the State of Washington aged 50–76 at the beginning of the study, were classified into five groups based on their average daily dose of supplemental B vitamins over the previous 10 years. Statistical techniques were then used to adjust for confounding factors such as age, education, body size, and family history of lung cancer.

What were the results?

When the data were stratified by sex, B₆ and B₁₂ as individual supplements were each shown to increase lung cancer risk by 30–40% in men (but not in women).

The greatest risk was found among men with the highest average daily dose of B₆ (>20 mg/day was associated with an 82% greater risk) and B₁₂ (>55 mcg/day was associated with a 98% greater risk) over the ten years preceding the study.

When the data were stratified by smoking status, increased risk was associated with smoking. Smokers who had supplemented with high amounts of B₆ had nearly three times the risk of developing lung cancer, and those having supplemented with high amounts of B₁₂ had over three times the risk. The study found no association between supplementation and increased risk in either former smokers or recent smokers. As for never-smokers, the paper states they “were excluded from the smoking-stratified analysis because of the low number of participants with incident lung cancer in that group.”

The study showed that long-term supplementation with B₆ or B₁₂ increased lung-cancer risk in male current smokers, especially in those supplementing with high dosages of either vitamin.

What is the mechanism?

One-carbon chemical groups lack stability, so they need to be attached to larger molecules in a process called one-carbon metabolism. The vitamins B₆, B₉,^[6] and B₁₂ play an important part in one-carbon metabolism, which in turn plays a crucial part in methylation reactions and nucleotide synthesis.

The nucleus of each of your cells contains your complete DNA. In your DNA is encoded the genetic blueprint for every protein in your body. How then do cells maintain a unique identity? By each reading only certain parts of your DNA, so that only the appropriate genes are turned on at the appropriate time.

For that purpose, sections of your DNA can be “marked” with methyl groups that prevent the expression of nearby genes. This type of epigenetic imprinting is critical to keeping cells normal, healthy, and well behaved. When the process becomes dysfunctional, the wrong genes can be turned on at the wrong times, potentially leading to uncontrolled cell growth — to cancer.

So how would high amounts of B₆ or B₁₂ increase cancer risk? We might find some clues in a recent study on DNA methylation,^[7] which found that two years of supplementation with 400 mcg of B₉ and 500 mcg of B₁₂ changed DNA methylation. Thus, the increase in cancer risk seen in the Brasky study could be caused, in part, by changes in DNA methylation from long-term B vitamin supplementation.

Another curious finding from the Brasky study was that only men saw an increase in cancer risk from B₆ or B₁₂ supplementation. Women did not. We know androgens regulate some of the enzymes that participate in one-carbon metabolism,^[8] which might explain the difference.

Androgens and the vitamins B₆, B₉, and B₁₂ interact to play a role in DNA methylation. Since DNA methylation in part determines which genes are activated (or not) at any given time, this could explain the link between long-term B vitamin supplementation and cancer risk in men.

What does this study mean?

The Brasky study was not designed to show causation, but it did reveal a strong correlation between increased risk of lung cancer and long-term B₆/B₁₂ supplementation, especially in high doses and among smokers. There are several ways B vitamins may interact with cancer metabolism; more research is needed to determine the exact mechanisms at work. In the meantime, we are left with three takeaways:

Smoking, as you know, causes lung cancer. If you smoke, stop. If you are unable to stop, avoid supplementing with B vitamins for an extended period of time, especially if you are male. Long-term B vitamin supplementation seems to increase cancer risk in male smokers, possibly by potentiating carcinogenesis in precancerous cells in response to the carcinogens in cigarette smoke (which would explain why only current smokers, not former or recent smokers, seem affected).

The effect of B vitamins on non-smokers is still uncertain. In this study, sample sizes for never-smokers were too small to evaluate associations accurately.

Although observational studies cannot show causation, the associations between B vitamins and cancer risk found in this study raise an important point, which is that high-dose, long-term consumption of any supplement can potentially interact with your biochemistry in unexpected ways. Exceeding the recommended, tested doses of even the most healthful micronutrients may not be innocuous.

Q&A with Dr. Theodore Brasky, PhD, lead author of the study

When this study was published, its finding that B vitamin supplements increased cancer risk in men generated a lot of press. But isn’t there some nuance to that finding, especially with regard to smoking habits? What ultimate take-home message can be extracted from the data?

The nuance is sort of centered around the general idea that once you start chopping up data, you lose precision. In epidemiology, our best estimates come from data reflecting the largest sample sizes. Our most cited finding was that long-term, high-dose supplementation of vitamin B6 and long-term, high-dose supplementation of vitamin B12 were each associated with about a doubling of lung-cancer risk in men. This is an entirely true representation of our results. However, when we drilled down further — and thus lost some precision — we found that this twofold increase in risk was an average across different groups of men, some with no increase in risk (men who had never smoked or had stopped smoking at the time the study began), and some with a threefold to fourfold increase in risk (men who smoked at the time the study began).

Here the scientist is left with two possibilities. Is the real finding (a) based on the larger sample size with more precise data? — men who use these supplements have twice the risk of lung cancer as do men who don’t use these supplements; or (b) based on the subgroups within men with less precise results? — men who currently smoke and who use these supplements have three to four times the risk of lung cancer as do men who currently smoke and don’t use these supplements. To me, the take-home message is the latter.

Supplementation dose, frequency, and duration are all important from a biological standpoint. How were those factors taken into account in the design of the study questionnaire? What were the pros and cons of the different ways of using those factors (and others) to identify meaningful associations with lung-cancer risk?

Put simply, we had a number of options. We could analyze separately a given supplement’s frequency of use (i.e., days per week), duration of use (i.e., number of years in the past 10 [our questionnaire only asked about the past 10 years of use]), and most common dose used, or we could combine those data.

Analyzing separately any single aspect removes the influence of the other two, which is, in my view, not ideal. Combining the data gives two additional options. We could determine a cumulative dose over the past 10 years or an average daily dose over the past 10 years. We chose the latter because it’s easier to understand and because it allowed us to compare risks with what might be expected for intakes at the level of a multivitamin taken daily for the same amount of time.

However, the disadvantage of this option — which, I contend, remains better than the alternatives — is that the 10-year, daily-dose calculation equates short-term, high-dose intakes with long-term, lower-dose intakes. The highest category of intake for supplemental B₁₂, for instance, was >55 mcg/day. This is >55 mcg taken daily, on average, over 10 years. For some people, it may actually have been about that amount daily for 10 years, but for most it was shorter-term use at higher doses that averaged out to this level.

Therefore, >55 mcg is not meant to be interpreted as the actual dose that might confer risk. Indeed, most B₁₂ supplements are sold at much, much higher doses. A standard pill from a bottle at the grocery store might contain between 500 and 2,000 mcg, with instructions that it should be taken daily. This is why the comparison to what might be consumed from a multivitamin (100% RDA) comes in handy.

Although the question “Does B vitamin supplementation increase cancer risk?” is straightforward, extracting a solid answer from a given study population is another matter. Epidemiologists like yourself are experts at identifying risk associations within large study populations. At the other end of the spectrum, basic scientists like myself tend to use defined experimental models to identify important cellular/molecular controls that drive disease processes. Could you comment on how epidemiological studies and basic science (i.e., bench research) fit in the big picture of biomedical science? Do you feel they complement each other?

A better scientist than I could probably comment on this with real nuance. I can only give my interpretation, which is, sadly, not based off any firsthand experience with bench science. I once pipetted something, but my assay didn’t run properly. C’est la vie. Looking broadly across disciplines, I can say that epidemiologists and “basic” scientists have a complementary relationship born out of necessity.

Epidemiologists cite rodent studies because in these experiments a lot of the variables can be controlled. The animals are very similar genetically, they’re all fed the same diet (unless it’s a nutrition study), handled the same way, etc. Moreover, we can perform some trials in animals that are considered unethical in humans — exposing rodents to tobacco smoke, for instance. We often see the results of these studies as hypothesis generating because, after all, the animal is a model for the human. People do not, in fact, have fur or tails, and we are much more genetically diverse than rodents purpose bred for disease models. In some instances, animal models are better approximations than others. Mice have estrous cycles rather than menstrual cycles, so some similarities for reproductive cancers are muddied by physiology here. Similarly, a mouse’s prostate gland is structured differently from a man’s; again, models. The idea is the same for work involving cells in petri dishes, although the contrast is starker. On the other hand, from what I’m told by my colleagues in these fields, epidemiologic research, which is predominantly done in an observational manner, is seen as hypothesis generating.

That we all work together towards the same goal is what’s important. Although we definitely give each other grief, epidemiologists appreciate basic scientists for their explanation of biologic mechanisms, and (I’m assuming) basic scientists appreciate epidemiologists for their findings in need of biologic explanation.