Contrasting effects of preexisting hyperglycemia and higher body size on hospital mortality in critically ill patients: a prospective cohort study

Background: Obesity and diabetes mellitus are well-defined risk factors for cardiovascular mortality. The impact of antecedent hyperglycemia and body size on mortality in critical ill patients in intensive care units (ICUs) may vary across their range of values. Therefore, we prospectively analyzed the relationship between in-hospital mortality and preexisting hyperglycemia and body size in critically ill ICU patients to understand how mortality varied among normal, overweight, and obese patients and those with low, intermediate, and high glycated hemoglobin (HbA(1c)) levels. Methods: Medical history, weight, height, physiologic variables, and HbA1c were obtained during the first 24 h for patients who were consecutively admitted to the high complexity ICU of Hospital de Clinicas de Porto Alegre, Brazil, from April to August 2011. The relationships between mortality and obesity and antecedent hyperglycemia were prospectively analyzed by cubic spline analysis and a Cox proportional hazards model. Results: The study comprised 199 patients. The overall hospital mortality rate was 43.2% during a median 16 (8-28) days of follow-up. There was a progressive risk of in-hospital mortality with higher HbA1c levels, with the relationship becoming significant at HbA(1c) >9.3% compared with lower levels (hazard ratio 1.74; 95% confidence interval with Bonferroni correction 1.49-2.80). In contrast, mean body mass index (BMI) was higher in survivors than in nonsurvivors (27.2 kg/m(2) +/- 7.3 vs. 24.7 kg/m(2) +/- 5.0 P = 0.031, respectively). Cubic spline analysis showed that these relationships differed nonlinearly through the spectrum of BMI values. In a Cox proportional hazards model adjusted for Acute Physiology and Chronic Health Evaluation II score and HbA(1c), the risk of in-hospital mortality progressively decreased with increasing BMI (BMI <20 vs. 20-23.9 kg/m(2), P = 0.032; BMI < 20 vs. 24-34.9 kg/m(2), P = 0.010; BMI <20 vs. >= 35 kg/m(2), P = 0.032). Conclusions: Our findings suggest that significant hyperglycemia prior to ICU admission is a risk factor for in-hospital mortality. Conversely, increasing BMI may confer an advantageous effect against mortality in critical illness independently of previous glycemic control.