Unless you live under a rock it would have been impossible to miss this week's big story about vitamins preventing cancer in men. A paper by researchers at Brigham and Women's Hospital concluded: "Daily multivitamin supplementation modestly but significantly reduced the risk of total cancer."
Over an 11-year period, almost 15,000 male doctors, aged 50 or older were given either a placebo or a common multivitamin supplement, and the risk of cancer in two groups was compared. The researchers painted a fairly optimistic picture about the utility of vitamins in cancer prevention. But I have a problem with some of it.
First, the concept that cancer is a single disease is clearly false. What we call cancer is in fact, a conglomeration of at least 100 diseases, characterized by the site of origin and cell type. Is it even logical to ask whether a dozen or so vitamins could impact a set of 100+ different diseases in a meaningful way, and by some unknown mechanism?
A simpler version of this study would ask the same question, but confine the study to of one particular cancer, for example breast cancer. But breast cancer itself actually has 10 different types, each differing in the genetic makeup of the tumor, and the response (or lack of it) to a given therapy. So, even a study confined to only breast cancer would be still complicated by the heterogeneity of the disease. If different subtypes displayed different responses you couldn't even accurately make the statement that vitamin supplements (or anything, really) reduced the incidence of breast cancer--only that it reduced certain types of it. And studying all types of cancer together raises the complexity of the study enormously--perhaps to the point that the data is not interpretable.
In fact, the multivitamin study shows a similar pattern. Although there was an 8 percent reduction in overall cancer risk in the vitamin group compared to the placebo group, this did not hold true for prostate or colon cancer--there was no significant difference between the groups in these cases.
Furthermore, there was no significant difference in the cancer mortality rate between the groups, which is counterintuitive to what you might expect if the stated 8 percent reduction in incidence was real.
So, in order to believe that headlines, typically like CNN's "Multivitamins may prevent cancer in men," you need to accept that there is some mechanism by which a bunch of vitamins taken over 11 years has a small protective effect against some cancers, and this fails to translate into a lower mortality rate. I dunno. It just doesn't add up.
But the biggest problem I have is the wording of the conclusion: "[M]odestly but significantly." It may be unintentional, but this phrase automatically distorts the real findings of the study. And certainly makes it more newsworthy.
Scientists will understand that this term means modestly and statistically significant--in other words modestly with mathematically acceptable data. Everyone else, including the press, will no doubt read "significantly" as a lot. Which, I suspect is why this story has generating so much interest.
This hype will no doubt send men stampeding toward CVS in a frenzy that will make Pamplona's Running of the Bulls look like walking a poodle. This is the effect that bad headlines and exaggerated claims have on people that are perpetually confused by bad headlines and exaggerated claims.
If the take-home message from this study becomes "take vitamins and you won't get cancer," it will be just one more example of sloppy reporting, possibly encouraged by overoptimistic conclusions. We don't need any more of this. People are confused enough.
Do Vitamins Prevent Cancer? Not So Fast (Medical Progress Today)