Why 68 Vitamin Studies Are Better than None

Not all scientific studies are created equal--a fact that most reporters don't make note of in their stories. The recent study from Europe that concludes that taking vitamins may be harmful to your health is a perfect case in point.

It's not really a new study--in the way most of us think of studies. The investigators, lead by Dr. Goran Bjelakovic of Denmark, did not recruit thousands or hundreds or even ten participants and then give them vitamin supplements and see how they did. Instead, the researchers pooled together the data from 68 previous studies on vitamins to see what they could learn.

The idea of combining many smaller studies into one big study, called a meta-analysis, was championed in the 20th century by a British scientist named Richard Peto. Indeed, he was knighted by the Queen for his use of meta-analysis in cancer research, so technically, we should refer to him as Sir Richard Peto.

But the tricky thing about meta-analysis is that all those studies you want to combine were conducted by different investigators using very different methods. In other words, you cannot compare apples to apples. You're basically comparing apples to oranges and strawberries and bananas and maybe an eggplant or two--and that requires some pretty fancy statistics.

It can also lead to its own form of bias since meta-analyzers have to rank the studies they're looking at by how scientifically rigorous the studies are and that can lead to some very subjective decisions.

So if you look at all 68 studies that were pooled together by Bjelakovic and company, you find that taking vitamin supplements have no effect on lifespan--a fact the researchers duly noted and something you might expect in countries that have a plentiful supply of food.

But when they looked only at what they considered to be the most scientifically rigorous studies, they found a slight increase in premature death in participants who took vitamin A, beta-carotene (a precursor of vitamin A), and vitamin E.

There's a certain amount of biological sense in finding a problem with vitamin A, beta-carotene and vitamin E since they are all fat-soluble vitamins and so can build up to high levels in the body's tissue. But as statisticians are fond of saying, association does not mean causation. Just because researchers found a link doesn't mean that they have proven cause and effect.

After all, if you're taking vitamins because you're a smoker, you shouldn't be surprised if you develop lung cancer. And a surprising number of people take vitamins because they think the supplements will make up for whatever bad habits they have.

At any rate, the finding of potential premature death is what got trumpeted by the press--although without the long explanation.

The real conclusion to take from this meta-analysis is that there are no shortcuts to good health. If you eat right, exercise regularly and are an active participant in your community, you are more likely to be healthy--but even then there are no guarantees.