Blog | Friday, August 24, 2018

Multivitamins won't help you '____________'. Well, duh


Let's fill in the blank. Multivitamins won't help you fly. Or cliff dive.

They won't help you leap tall buildings in a single bound. They won't help you teleport, tesseract, read minds, or shoot laser beams from your eyeballs. Oh, and they won't help you measurably reduce your risk of dying from a heart attack, either. Surprise!

Or no surprise. In my 27 or so years of patient care, and with a patient population keenly interested in nutrient supplements, I cannot recall a single instance when any one of the above was the reason propounded by a patient of mine for taking a multivitamin/mineral supplement (MVM).

Yet the big medical news of this past week was yet another meta-analysis showing that multivitamins don't do what nobody I know thought they did in the first place: reduce the risk of heart attack, stroke, or cardiovascular death.

We've had just these findings before, with just such a plethora of breathless headlines and click-bait provocations attached to them, but here we go again. I concede, this is one of my pet peeves about the scientific literature-media-industrial complex. There seem to be no restrictions on how many times the same question can be posed, or on how surprised we can all pretend to be when the same answer ensues.

No, a multivitamin is not likely to reduce measurably your risk of cardiovascular death. Not this time, not last time, and, spoiler alert, not next time either. Get excited about the next batch of indistinguishable headlines as the spirit moves you.

So, leaving aside the media problem, let's address the more substantive matters of medical effects, and research motivations.

Consider what it would take to show that multivitamins do, in fact, reduce the risk of, say, cardiovascular mortality. One option would be to study perfectly healthy people, and compare those who do and don't take multis. Another would be to compare people with established cardiovascular disease, or at least significant risk factors, and do the same. Either way, for the information to be useful, the comparisons need to be between groups alike in all other relevant ways, but differing with regard to MVM use.

Do you see the problem? In a group of very healthy people who do, or don't take multis, but who eat well, don't smoke, exercise routinely, sleep enough, aren't stressed out, don't drink excessively, and have no overt cardiac risk factors, very, very few people are going to die of a heart attack. It simply doesn't matter in this group if multis are “good” for the heart, because you can't fix what isn't broken. In these healthy folks, cardiac deaths will be so rare with or without MVM use, that no differences will be apparent.

Then, there is the other end of the spectrum. People with significant risk factors for heart disease, or with established coronary disease already, are going to have heart attacks, and die from heart disease, at a much higher rate. Or, they would, but for treatment. And, of course, people known to have coronary disease participating in clinical trials are all receiving state-of-the-art treatment; any alternative would be unethical. So, in this group, multivitamins again cannot possibly fix what isn't broken. The study question in this case is not if multis reduce the risk of heart disease or death; rather, it is, do multis measurably reduce those risks beyond what the best drugs and procedures can do?

This is, simply, the statistical reality. Showing risk reduction by any given means is nearly impossible when risk is very low to begin with, because tiny or rare effects take massive sample size and duration to detect. Showing risk reduction when risk is high all around and being managed in all the ways we know how is also nearly impossible, because the effect is consigned to the realm of “residual benefit,” which is inevitably apt to be fairly small.

Despite all of these hurdles, the new paper in Circulation, a systematic review and meta-analysis of prior studies (both randomized and observational trials)- was, like its predecessors on this topic, actually rather far from entirely negative. In the author's own words, “… MVM supplement use was inversely related to the incidence of CHD when all studies were considered.” This finding simply didn't hold up when analyzed further, or when limited to randomized trials only.

The study also found significant benefit of MVM use when adjustment was not made for fruit and vegetable intake, or for exercise, suggesting that people who took MVMs were more likely to engage in other health-promoting behaviors. The benefit went away with adjustment for these, a statistical method to isolate the effects of just one variable.

The team also found, apparently to their surprise, that there was a net benefit of MVM use in studies conducted outside the U.S. The actual numbers are these: for studies outside the US, the relative risk of heart disease with MVM use was 0.74 (95% CI, 0.62–0.89). For studies conducted in the US, the relative risk was 0.91 (95% CI, 0.83–1.00). To explain: a relative risk below 1.0 means risk reduction; and the “95% CI” shows the range of risk values that are 95% probable based on that finding. So, for studies outside the US, there is 95% confidence in a risk reduction. For studies in the US, since the top end of the range is 1.00, there is 95% confidence that risk is either reduced, or unaltered.

That difference is rather modest, but why any difference at all? Perhaps some of the populations outside the US are more subject to nutrient deficiencies, and MVMs fixed those. If so, then this study actually suggests that MVM use to redress deficiencies may, indeed, reduce heart disease risk. As for the U.S., the findings leave room for benefit, and suggest that at worst MVMs do not harm the heart.

As for why studies like this one are conducted, funded, and published repeatedly, I can only speculate. The authors say it's because MVM use is popular and they apparently felt obligated to engage in a debunking crusade. I don't really buy that, because again, I don't recall ever meeting any patient who relied on a multivitamin to reduce risk of heart attack. I suspect the medical-industrial-establishment favors drugs and procedures over all less lucrative alternatives, and will reliably find reasons to study, and espouse, the benefits of our highly medicalized status quo.

Be that as it may, it's true you should not rely on a MVM to reduce your risk of heart attack or stroke. It might actually help a bit. This study doesn't rule that out by any means, and it appears very unlikely to hurt, but any effect is apt to be small. On the other hand, if you effectively use lifestyle as medicine, you can practically eliminate your risk of cardiovascular disease in the first place.

If, on the other hand, you take a multi for the many far more customary reasons as an insurance policy against gaps left by your diet, to have a bit more energy, perhaps to be a bit less subject to viral infections, this paper is just silent on the topic. It provides no information at all pertaining to any of the actual reasons for MVM use I've encountered among my own patients over the years. Those are all still potentially perfectly legitimate; those deserve to be studied.

Just one final comment, to be filed under ”have hammer, see nails.” The very word “supplement” begs the question: supplemental to what? The obvious answer is nutrients obtained from food. Supplements are most likely to do us good when they fill nutrient gaps left by our diet, and that, in turn, requires knowing what nutrients are being provided by diet. Historically, getting that information is tedious for all concerned, to it is rarely done. My associates and I have invented the requisite hammer (full disclosure, I am the founder and CEO of the company to which the link directs), a highly convenient, fast, fun method to assess dietary intake, including approximate levels of all important nutrients. By making such assessment easy, such a tool has clear implications for future research into nutrient supplements: do they make a difference based on what nutrient intake levels they are “supplementing” in the first place?

Let's conclude. If you want to reduce your risk of cardiovascular disease, eat well, be active, avoid toxins, and so on. Use lifestyle as medicine, in other words. Don't rely on a MVM for this purpose, obviously. But if you have other reasons for taking a MVM, this study, like all before it, leave room for potential benefit, while pretty reliably ruling out any meaningful harm. So, supplement as the spirit moves you, but don't bank on it to improve your cliff diving.

And let's all recall that no matter how effusive the media coverage, research rarely produces answers any better than the questions.