bias and confounding
There are major concerns related to confounding and bias in this study. More intensive medication in the obese subjects is obviously mainly due to twice as high prevalence rates of hypertension and diabetes compared to subjects with normal BMI. Do the authors suggest that diabetes and hypertension improve life expectancy? Statin use was apparently not different across BMI categories. On the other hand, antidabetic drug therapy is not reported. However, intensification of blood glucose lowering therapy has been associated with increased cardiovascular risk and mortality. Finally, drug prescription may relate to varying rates of drug adherence across BMI groups. Socioeconomic factors may be important in relation to survival and BMI but are not considered in this analysis. The present data are not convincing but rather misleading.
Conflict of Interest: