In my last email, I reviewed B vitamin supplementation and the history of the research. Initially the reductionist models and observational studies were promising. Unfortunately, the clinical trials showed something very different. B vitamin supplementation does not work, and when it does, there are trade-offs.
In this email, I will explain why reductionist models and observational studies are weak and flawed forms of evidence. In fact, newer reductionist models contradict the older ones, and predict that B vitamins could even accelerate aging. I will also explain how clinical trials are strong forms of evidence and able to avoid these flaws.
Why B vitamin reductionist models were wrong
Reductionist models try to simplify a complex system by looking at the individual parts. In medicine, it's easy to spot a reductionist model. When you hear a statement like, "Treatment X is an anti-inflammatory", or "Treatment X is an antioxidant". These are reductionist models trying to understand treatment X by looking at its effect on a single part of the body. This is a helpful process, but it has significant limitations. As Douglas Adams (Hitchhikers Guide to the Galaxy) quipped,
"If you try and take a cat apart to see how it works, the first thing you have on your hands is a nonworking cat."
When creating a reductionist model, the first challenge is to identify the crucial parts of a system. In medicine, these are the parts we suspect are responsible for health and disease. The second challenge is to avoid the omission of other crucial parts, or the omission of interconnections between parts. In other words, do not oversimplify the human body.
These challenges are so difficult to overcome that creating a working reductionist model is turning out to be a Sisyphean task. Our models fall apart as we uncover more and more parts, and more interconnections between parts. Older models are continually being replaced by newer and more complex ones. A seemingly simple treatment, with a simple pathway, turns out to be a highly nuanced treatment, with a very complicated pathway.
This is certainly the case with B vitamins. Our models from the 1970's (below) showed that B vitamins play a role in the methylation cycle. A pathway critical for production of DNA, protein, and neurotransmitters. The message from this model was, more B vitamins is good.
Since then, our knowledge of biochemical pathways has significantly expanded. It turns out we oversimplified B vitamins. The methylation cycle (circled in red in the diagram below) is connected to thousands of other reactions.
However, even this newer model is oversimplified; it omits crucial connections. Recently, it was discovered that DNA is connected to the methylation cycle. DNA gets methylated; and when this happens, genes get turned off (see diagram below). In the last decade, an entire field of biology has sprung into existence studying this phenomenon. It is called epigenetics. Genes appear to have a controller of their own, the epi-genome.
B vitamins methylate DNA
So, what happens to our DNA when we give people B vitamins? B vitamins increase methylation in the entire cell; therefore, it is logical to conclude they will increase methylation in DNA, and turn off genes. This is indeed what happens, and confirmed in genetic studies. Specifically, B vitamins turn off a group of tumor suppressor genes (DIRAS3, ARMC8, and NODAL). When tumor suppressor genes are turned off, it can lead to cancer. This can finally explain the increased cancer seen in the large clinical trials of B vitamins.
Unfortunately, further predictions from this new model get worse. DNA methylation is not only a risk for cancer, it is also directly correlated with biological age. The older you are, the sicker you are, the harder you live, the more your DNA is methylated. Like telomeres, DNA methylation is a new way of measuring your biological age - an epigenetic clock. Some gerontologists speculate that DNA methylation is so important, it is not only a marker of age, but the cause of aging itself. If this were true, B vitamin supplementation could actually accelerate aging. It would also explain the surprising increase in all-cause mortality seen in some some trials.
B vitamin supplementation could actually accelerate aging.
We are certainly witnessing a paradigm shift with respect to our understanding of B vitamins. In forty years we went from a reductionist model that predicts better health and less cancer because B vitamins provide essential molecules and repair DNA, to a model that predicts worse health and more cancer because B vitamins methylate DNA and turn off tumor suppressor genes. This should illustrate the point that one can never truly rely on reductionist models; they can change quickly, and are only as good as the part of the body they choose to highlight.
Why B vitamin observational studies were wrong
Like reductionist models, observational studies also predicted that more vitamins would be good. They were also flawed, but suffer from a different problem. They are unable to distinguish association from causation. Observational studies compare two groups of people; an exposed group and a non-exposed group. However, the exposed group is almost always different from the non-exposed group. It is these differences that account for the observed disease, not the exposure. This is called confounding.
Observational studies are unable to distinguish association from causation.
For example, folate (B9) is found in green leafy vegetables and expensive supplements. A group of people with high folate levels are different than a group of people with low folate levels. They are more likely to eat green leafy vegetables and consume expensive supplements. It also follows that they are more likely to take better care of themselves, get regular check-ups, eat healthier foods, avoid smoking, exercise more, have a higher income, etc. So when this group is observed to have better health outcomes, it is not necessarily because of the folate levels. It is because of the salubrious traits that lead to high folate levels. Ergo, the high folate levels don't CAUSE good health, they are ASSOCIATED with good health.
Clinical trials
Clinical trials are designed to solve both the problems encountered with reductionist models, and the problems encountered with observational studies.
Reductionist models, as explained above, can only test a treatment on a part of the body (like a chemical pathway or a cell in a test tube). Outcomes are measured by how the treatment affects that part, and that part alone. If the wrong part is chosen or the system is oversimplified, it will lead to erronious conclusions. Conversely, clinical trials test the entire body. The entire body gets the treatment, and the entire body is measured for real outcomes; like death, or a heart attack. Thus, clinical trials are methodologically "holistic". The irony here is that doctors who refer to themselves as "holistic", rely heavily on reductionist models, and almost never back their claims with clinical trials.
Doctors who refer to themselves as "holistic", rely on reductionist models, and almost never back their claims with clinical trials.
The problems of observational studies, on the other hand, are solved by addressing their main issue, confounding. The root cause of confounding is that the exposed group is different from the non-exposed group. Clinical trials correct this at the beginning of every trial during the process of randomization. Two equal groups are created and the only difference between the groups is that one will get the treatment (or exposure), and the other will not.
Conclusion
Clinical trials are our best tool in medicine to test treatments. They are the strongest kind of evidence. The one test to rule them all. This is why many in our field rank them along with germ theory, antibiotics, imaging, vaccines, anesthesia, and sanitation, as the most important advances in medical science.
In reality, most treatments you will encounter are based on weak kinds of evidence, like reductionist models and observational studies. Weak evidence far outnumbers strong evidence. It's cheap and easy to produce, and relatively malleable.
The take home message is: do not be enticed by weaker forms of evidence. Don't be misled by statements based on reductionist models, like: "This takes down inflammation", or "This is an antioxidant". These models are probably oversimplifications and already out of date. Likewise, don't be misled by statements based on observational studies, like: "NAD declines with aging, therefore it is the cause of aging", or "Testosterone declines with aging, therefore it is the cause of aging". These conflate association with causation; everything in your body declines with age.
Instead, demand the best test of a treatment - a clinical trial. The clinical trial is the ultimate challenge to any medical hypothesis. It's your body, you deserve the best available evidence. Why settle for less? After all, even vitamins can harm.
In my next email, I will connect the story of B vitamins with the story of antioxidants vitamins. Unsurprisingly, the antioxidants share a similar story, and a similar fate. Promising reductionist models and observational studies, disappointing clinical trial results, and cancer.