Obesity and Overweight Are Independently Associated with Greater Survival in Critically Ill Diabetic Patients: A Retrospective Cohort Study

Abstract

Background. The relationship between obesity and the outcomes of critically ill diabetic patients is not completely clear. We aimed to assess the effects of obesity and overweight on the outcomes among diabetic patients in the intensive care unit (ICU). Methods. Critically ill diabetic patients in the ICU were classified into three groups according to their body mass index. The primary outcomes were 30-day and 90-day mortality. ICU and hospital length of stay (LOS) and incidence and duration of mechanical ventilation were also assessed. Cox regression models were developed to evaluate the relationship between obesity and overweight and mortality. Results. A total of 6108 eligible patients were included. The 30-day and 90-day mortality in the normal weight group were approximately 1.8 times and 1.5 times higher than in the obesity group and overweight group, respectively ( $P<0.001$ , respectively). Meanwhile, the ICU (median (IQ): 2.9 (1.7, 5.3) vs. 2.7 (1.6, 4.8) vs. 2.8 (1.8, 5.0)) and hospital (median (IQ): 8.3 (5.4, 14.0) vs. 7.9 (5.1, 13.0) vs. 8.3 (5.3, 13.6)) LOS in the obesity group and overweight group were not longer than in the normal weight group. Compared with normal weight patients, obese patients had significantly higher incidence of mechanical ventilation (58.8% vs. 64.7%, $P<0.001$ ) but no longer ventilation duration (median (IQ): 19.3 (7.0, 73.1) vs. 19.0 (6.0, 93.7), $P=1$ ). Multivariate Cox regression showed that obese and overweight patients had lower 30-day (HR (95% CI): 0.62 (0.51, 0.75); 0.76 (0.62, 0.92), respectively) and 90-day (HR (95% CI): 0.60 (0.51, 0.70); 0.79 (0.67, 0.93), respectively) mortality risks than normal weight patients. Conclusions. Obesity and overweight were independently associated with greater survival in critically ill diabetic patients, without increasing the ICU and hospital LOS. Large multicenter prospective studies are needed to confirm our findings and the underlying mechanisms warrant further investigation.