Two recent meta-analyses have reignited the public debate about whether it’s possible to be both fat and healthy. Neither publication changes the scientific picture much – and their findings are surprisingly similar, considering that the authors of the two papers drew opposite conclusions. The media coverage is what I found interesting, as it illustrates the deep cultural biases in the way we talk about these issues.
Let’s start with the facts. Cardiovascular fitness is well known to predict people’s risk of dying within a certain number of years of follow-up. “Risk of death” is the common shorthand for that lengthy concept, and I’ll use the term here (though it sounds weird to me because we’re all going to die eventually).
The first paper combined the results of ten studies that looked at whether obesity predicts some additional risk of death beyond the known risk associated with low fitness. All the studies used objective measures of weight, height, and fitness, rather than unreliable self-report. Participants were followed for 7.7 to 16 years afterward. The authors of the meta-analysis concluded that the risk of death depended only on fitness, not on fatness.
The second paper combined the results of eight studies of a related question, whether obesity predicts any additional risk of death beyond the known risks of poor metabolic health. That’s a fancy way of describing the set of risk factors that your doctor might monitor at check-ups: blood pressure, HDL (good) cholesterol, triglycerides, waist circumference, and fasting glucose. Those authors concluded that among people with good metabolic health, obesity was still associated with an increased risk of death, though only in studies with at least a ten-year follow-up.
(Dr. Arya Sharma points out that the definition of metabolic health in this paper was lenient. His work has shown no increased risk over 16 years for obese people who are metabolically healthy under a more stringent set of requirements.)
To me, the similarities between the two papers far exceed their differences. The first paper just missed finding a statistically significant effect on relative risk for obese people who are fit (average 1.13, 95% confidence interval 1.00-1.27). The second paper’s results barely achieved significance for obese people who are metabolically healthy (average 1.24, 95% confidence interval 1.02-1.55). As each paper reports on tens of thousands of participants, the problem is not a lack of statistical power but that the effect is weak, if it exists at all. At worst, the second paper suggests that the healthy obese have a 24% greater risk than healthy normal-weight people.
Compare that to the much larger risk increase for unfit or metabolically unhealthy people at any weight, reported in both papers. The relative risks ranged from 2.42 to 3.55, or roughly a threefold higher risk than healthy normal-weight people. So if obesity does present an independent risk, it’s about one-tenth as large as the risk associated with poor fitness or poor metabolic health.
What would a sensible society do with that information? Perhaps encourage everyone to concentrate on the risk factors that have a huge effect and are relatively controllable, by exercising and eating their veggies. Or worry about the many people who face a high risk of death because of unhealthy lifestyles but aren’t concerned because they’re thin.
What do we do instead? Ignore the first study and use the second to argue about whether it’s (sometimes) OK to be fat, with headlines like “Healthy Obesity is a Myth, Report Says” and “Overweight And Healthy: A Combo That Looks Too Good To Be True.” Sigh.