Being overweight or obese is associated with decreased mortality in critically ill patients: a retrospective analysis of a large regional Italian multicenter cohort

J Crit Care. 2012 Dec;27(6):714-21. doi: 10.1016/j.jcrc.2012.08.013. Epub 2012 Oct 24.

Abstract

Purpose: To describe the epidemiology of obesity in a large cohort of intensive care unit (ICU) patients and study its impact on outcomes.

Methods: All 3902 patients admitted to one of 24 ICUs in the Piedmont region of Italy from April 3 to September 29, 2006, were included in this retrospective analysis of data from a prospective, multicenter study.

Results: Mean body mass index (BMI) was 26.0 ± 5.4 kg/m(2): 32.8% of patients had a normal BMI, 2.6% were underweight, 45.1% overweight, 16.5% obese, and 2.9% morbidly obese. ICU mortality was significantly (P < .05) lower in overweight (18.8%) and obese (17.5%) patients than in those of normal BMI (22%). In multivariate logistic regression analysis, being overweight (OR = 0.73; 95%CI: 0.58-0.91, P = .007) or obese (OR = 0.62; 95%CI: 50.45-0.85, P = .003) was associated with a reduced risk of ICU death. Being morbidly obese was independently associated with an increased risk of death in elective surgery patients whereas being underweight was independently associated with an increased risk of death in patients admitted for short-term monitoring and after elective surgery.

Conclusions: In this cohort, overweight and obese patients had a reduced risk of ICU death. Being underweight or morbidly obese was associated with an increased risk of death in some subgroups of patients.

MeSH terms

  • Adult
  • Age Factors
  • Aged
  • Aged, 80 and over
  • Body Mass Index
  • Comorbidity
  • Critical Illness / mortality*
  • Female
  • Humans
  • Intensive Care Units / statistics & numerical data*
  • Italy / epidemiology
  • Male
  • Middle Aged
  • Multicenter Studies as Topic
  • Obesity / epidemiology
  • Obesity / mortality
  • Overweight / epidemiology
  • Overweight / mortality*
  • Prognosis
  • Retrospective Studies
  • Sepsis / epidemiology
  • Sepsis / mortality*
  • Sex Factors
  • Thinness / epidemiology
  • Thinness / mortality