A common mistake in medical epidemiological studies is to mistake coincidence for causality. This is a very easy trap to fall into if the predicted cause-and-effect result is one that we already believe, or would like to believe, or that previous data leads us to expect. Again, the results of epidemiological studies should be used to formulate a hypothesis to be tested in the future, and not to prove a hypothesis.
With this in mind, let us evaluate a group of world-wide nutritional studies that was published by Lancet, the top English medical journal,the third week in February. They studied BMI's throughout the world. (for those of you in the non-medical world, this is (weight)/(height)(height), i.e. height squared. We have all been told that the greater the BMI (i.e. the heavier you are for a given height), the greater is the risk of heart disease and death. A recent anomaly reported elsewhere was a study that showed that those who were slightly overweight (BMI's between 26 and 30) had the best chance of surviving a heart attack. This result fit no theories, so it has not been further addressed, especially since this was the first evidence that maybe obesity wasn't the killer that it was claimed to be.
The Lancet study, which studied countries through the New World, the Old World, and the Far East came up with three results that held up statistically:
(1) Over the past 30 years the BMI is steadily increasing in all countries, almost monotonically.
(2) Over the past 30 years, the world-wide rate of heart attacks and cardiovascular deaths is decreasing.
(3) In the Far East, and only in the Far East, the greater the BMI, the LESS was your risk of heart attack or cardiovascular death.
This last result, indicating that overweight is cardioprotective, has many cardiologists and statisticians scrambling to find fault with it. However, except for saying they don't believe the result, there has been no significant rebuttal of the result. It is up to we doctors and researchers to try to understand the implications of this result, if we assume it represents a true state of affairs, but since the relationship was observed for over 30 years of studies, it is difficult to claim that it is not significant, and perhaps we should re-evaluate our definition of the "normal" range of BMI's.