Data file anonymized.

Background
Although the paradoxical association between obesity and improved survival has been reported in a variety of clinical settings, its applicability to intensive care unit (ICU) outcomes in older critically ill patients remains unclear. We sought to examine the association between obesity and 30-day mortality and other adverse outcomes in this population.
Methods
We analyzed data of older patients (≥ 60 years) in the eICU Collaborative Research Database. Body mass index (BMI) was stratified according to the World Health Organization obesity classification. Logistic regression model was used to estimate adjusted odds ratios (ORs), and cubic spline curve was used to explore the nonlinear association between BMI and 30-day ICU outcomes. Stratified analysis and sensitivity analysis were also performed.
Results
Compared with class I obesity, under- and normal-weight were associated with higher all-cause, cardiovascular and noncardiovascular mortality, and class III obesity was associated with greater all-cause and cardiovascular mortality (OR, 1.18 [95% CI, 1.06–1.32], 1.28 [1.08–1.51]). Obesity classes II and III were associated with higher composite all-cause mortality, mechanical ventilation, or vasoactive drug usage risks (OR, 1.12 [95% CI, 1.04–1.20], 1.33 [1.24–1.43]). Mechanical ventilation was strongly positively associated with BMI. A significant BMI-by-sex interaction was observed for cardiovascular mortality, such that the association between severe obesity and mortality was more pronounced among older men than older women.
Conclusions
The obesity paradox does not appear to apply to short-term ICU outcomes in older patients with critical illness, mainly due to increased all-cause and cardiovascular mortality in severely obese patients, particularly in men.