Is obesity good for you?
John Menzies, Edinburgh, 17th April 2014.
Some of the All Blacks team - Go on, tell them they’re fat!
We know instinctively that being overweight is unhealthy. There is plenty of evidence to support this idea - obesity increases the risk of developing diabetes, heart disease and certain cancers and, put bluntly, it increases the chances of dying . [article]
But over the years several very large studies have looked at the chances of dying due to any cause (termed “all-cause mortality”), not just diseases associated with obesity, and compared death rates in different ranges of body mass index (BMI). BMI is a simple calculation - body mass in kilograms divided by the square of height in metres. It is an imperfect measure of fat stores (and the setting of the BMI boundary between “normal “and “overweight” is an interesting story in itself (read more), but BMI typically gives a good impression of how much excess weight a population (not a specific individual!) carries.
We might expect that populations with BMI values in the normal range may be at less risk of death than those in the overweight or obese range. In many studies this is proven to be the case. [article] However, a recent meta-analysis of 97 separate sets of data containing information for almost 3 million individuals and 270,000 deaths suggested that being overweight was protective in all-cause mortality compared to the normal range. [article]
The study expressed the risk as a hazard ratio - the risk at any given time compared to a reference group. For example, a group with a hazard ratio of 2 indicates that members of that group are at twice the risk of those in the reference group. To give a more concrete example, the hazard ratio for smokers to develop lung cancer is ~5 compared to never-smokers.
In this study those of normal body weight served as the reference group. As one might expect, those in the very obese group (BMI >35) had an increased hazard ratio of 1.29 - an increased risk of death. But those in the overweight (BMI 25-30) and obese (BMI 30-35) groups had hazard ratios of 0.94 and 0.95 respectively. In other words, compared to the normal weight group they seemed to have a reduced risk of death from any cause.
This was rather a surprising finding - after all we have learned about diseases associated with obesity, surely it could not be the case that being overweight was protective against death? Perhaps overweight individuals are more likely to survive illness because of extra energy stores, or recieve more attention in hospitals? In any case, a number of scientists suggested explanations for the result.
A major criticism was the use of BMI - this measurement does not discriminate between weight from fat and lean weight from muscle, bone and so on. The average BMI of the All Blacks team that played at the last Rugby World Cup was 29.8. One would have to be brave to accuse that group of being overweight.
In addition, perhaps that the “normal” group may have contained the ill and frail, or those losing weight because of undiagnosed disease or in an effort to reverse a disease process ? [article] Perhaps the “normal” group featured smokers or ex-smokers, or undernourished individuals ? [article] Importantly, it was pointed out that this apparent protective effect disappeared after the age of 65. [article] Given that our population is aging and that, presumably, death becomes more likely as one ages, these data are not so encouraging.
The authors replied to these criticisms and defended their method and interpretation. [article] But what’s the moral of the story? Don’t necessarily trust BMI as an indicator of health in an individual, and look deeper if a scientific study suggests something that sounds too good to be true.