We read the study conducted by Salinero-Fort and colleagues that was published in European Journal of Internal Medicine, enthusiastically and meticulously. The authors tried to examine the effect of body mass index (BMI) on the risk of all-cause mortality among type 2 diabetes mellitus patients in a 5-year follow-up [1]. The results were very interesting; however, some methodological issues should be noticed. First, continuous variable of BMI is categorized into five groups, which implies a loss of information.
