Friday, July 31, 2009 11 Responses
If you follow obesity research with even a passing interest, there is a good chance that you have seen the now infamous editorial by Margaret Wente which was published in the Globe and Mail earlier this week. In her editorial, Ms Wente describes a recent study in Obesity which examined the links between body mass index and mortality in a group of 12,000 Canadians. Ms Wente makes several arguments - many of which are completely refuted by the available evidence - and I think that both her editorial and the study itself deserve some further analysis.
The study, which was led by Statistics Canada researcher Heather Orpana, reported that individuals who were overweight in 1994 were at significantly decreased risk of mortality over the 12-year follow-up when compared to individuals in the normal weight category. Now for several reasons that I will get to in a moment, that isn't actually all that shocking. But I can see why it would seem pretty earth-shattering to people unfamiliar with the field of obesity research, which apparently includes Ms Wente. You see, Ms Wente takes these findings to suggest that people should "Get fat, live longer". Now of course Orpana and colleagues did not examine whether gaining weight makes people live longer - they only looked at the associations between baseline weight and mortality. But that's not the biggest detail that Ms Wente overlooked. In fact, she says that
“being obese was associated with a 12-per-cent lower risk of dying”
But you see, the study didn't show that at all. In fact, the Globe published a letter to the editor from the study authors on Wednesday correcting Ms Wente, saying that
"None of the analyses reported in our paper...showed that obesity (a body mass index ranging between 30 and 35) was associated with lower mortality."
Let's get back to Ms Wente's argument that people should gain weight to live longer. To support her point, she cites an Annals of Internal Medicine review paper from 1993 which reported that individuals who gained a modest amount of weight as they aged had the lowest risk of mortality, while those who lost a large amount of weight had a dramatic increase in mortality. But the majority of studies in that review article failed to control for "unintentional" weight loss due to underlying cancer or other serious diseases (some did control for heart disease, however). This means that they lumped unintentional weight loss due to disease with "intentional" weight loss due to lifestyle changes. Unintentional weight loss is almost always bad, and may have been driving the relationship between weight loss and mortality in the Annals paper. When we look instead at more recent research focusing on "intentional" weight loss in overweight and obese adults, we see an overwhelming majority of studies which report that weight loss actually reduces the risk of mortality, just as we would expect. For example, a study in the American Journal of Epidemiology reports that "[in overweight women] intentional weight loss of any amount was associated with a 20% reduction in all-cause mortality". Why Ms Wente chose to avoid the wealth of current research in favour of a review article with such a major limitation I am not sure. If I had to guess, I would say it was because the majority of current research completely refutes her argument.
Now let's return to the current study in Obesity, because it is a genuinely interesting paper. As I mentioned earlier, the study assessed the association between body mass index and mortality in a sample of 12,000 Canadians. I've been chatting with a lot of people about this paper the past few days so I'd like to outline what I think we can learn from it.
The main point I would take from this paper is that body weight is associated with health risk, and as usual, it was a J shaped curve with increased risk observed at both very high and very low BMI's. So as usual, you want to avoid being significantly over or underweight. However, the reason that this study is so interesting is the nadir of the curve fell in the "overweight" range, rather than the "normal" range. But when seen in the context of other research, that finding isn't all that surprising. For example, Flegal and colleagues reported similar results in an American sample in 2005, as have some (but not all) others. In fact, I don't think anyone knows for certain what the "ideal" BMI is, and it is almost certainly different for people of different genders and ethnic backgrounds (for example Asian individuals often begin to experience metabolic complications at much lower BMI's than Caucasians). The current BMI guidelines may be a bit conservative for Caucasians, but public health guidelines should err on the conservative side, especially in a country with a wide range of ethnic backgrounds like Canada.
It is also important to keep in mind that where you store body fat is probably more important than how much body fat you have. For example, body fat stored in the legs has consistently been shown to protect against metabolic risk in longitudinal studies. In contrast, abdominal fat, and in particular visceral fat, is independently associated with increased risk of morbidity and mortality. People who have high BMI’s tend to have high amounts of visceral fat, which is probably responsible for much of the relationship between BMI and health risk. However, some people with high BMI’s store body fat mainly in their lower body (the “pear” body shape), and these individuals often have very little metabolic risk at all. Similarly, although less common, some individuals with a high BMI are merely very muscular (these people also have low health risk).
We also know that when people exercise, they preferentially reduce the visceral fat that is responsible for much of the obesity-related health risk. Even when there is little or no change in body weight, chronic exercise results in decreased visceral fat, and decreased health risk. This is likely one reason why research has consistently shown that it is better to be overweight and physically active, rather than lean and inactive.
This means that if you are exercising regularly, even if you are not losing weight, you are still dramatically reducing your risk of diabetes, heart disease, and several types of cancer. Too often people quit their exercise program because they feel they are not losing enough weight, or not losing weight fast enough, and they don't realize all of the benefits that come from exercise irrespective of changes in body weight. So I hope this study helps people realize that they should focus on what matters - living a healthy lifestyle - and stop focusing so much on their body weight.
So what's the take-home message from this lengthy post (aside from ignoring health advice from Margaret Wente)? Body weight affects your health, but not as much as diet and exercise. So focus on those healthy behaviors, and you'll be moving towards a longer, healthier life.
Hat tips to Alex Green and Alex Hutchinson for sending me the Globe article and/or related links. As always, to receive all of the latest obesity news and research via email, you can sign-up here.
Orpana, H., Berthelot, J., Kaplan, M., Feeny, D., McFarland, B., & Ross, N. (2009). BMI and Mortality: Results From a National Longitudinal Study of Canadian Adults Obesity DOI: 10.1038/oby.2009.191
Enjoyed this story? Share it with your friends by clicking the buttons below!
To get future posts delivered directly to your email inbox or to your RSS reader, be sure to subscribe to Obesity Panacea.