Is Body Mass Index a Useful Prognostic Factor for Critically Ill Patients

Abstract

Body mass index (BMI) is a measure of body fat calculated by dividing weight by the square of height. Since the index was devised by Adolphe Quetelet and is defined by dividing one’s body mass by the square of one’s height (kg/m), BMI has been commonly used as proxy measure of excess body fat. According to the literature, higher BMI was strongly associated with higher mortality in the general population.[1-4] However, critically ill patients’ relative mortality risks associated with overweight or obesity are still subject to debate although underweight has been established as a strong predictor of their deaths.[5-8] Several large-scale prospective studies have assessed the association between BMI and mortality in critically ill patients using (identical) cut off values. However, most of these studies were conducted on western populations.[9,10] The results of these studies cannot be generalized to Asian populations because of variations in body compositions and body fat distribution. Indeed, Asians have lower BMI but higher levels of body fat than Caucasians.[11] In the local research context, Lim SY et al claimed that BMI was not significantly associated with mortality in critically ill patients and that mortality risk in critically ill patients was more associated with failed extubation and severity of illness.[8] In their retrospective study, BMI values were classified into three categories: underweight 30.0 kg/m). Their findings showed that the Cox-proportional hazard ratios with exact partial likelihood to handle tied failures for hospital mortality comparing the BMI caretories with the reference category were 1.13 (0.88 to 1.44), 1.03 (0.84 to 1.26), 0.96 (0.76 to 1.22), and 0.68 (0.43 to 1.08) respectively, none of which were statistically significant. Nonetheless, a graded inverse association between BMI and mortality rate was evident. In other words, the lowest mortality rate was observed in the highest BMI group when surgical patients were excluded, and their findings were further explained by experimental evidence that adipocyte-secreted hormones such as