Nonlinear relationship between BMI and 30-day in-hospital mortality in ICU sepsis patients: a multicenter retrospective cohort study

While obesity paradox has been observed in various critical conditions, its role in sepsis remains unclear. This study aimed to investigate the relationship between body mass index (BMI) and 30-day in-hospital mortality among intensive care unit (ICU) patients with sepsis. This retrospective study analyzed data from 17,510 adult sepsis patients (aged ≥ 18 years) with ICU stays between 24 h and 30 days, complete mortality data, and available BMI measurements, using the eICU Collaborative Research Database (eICU-CRD v2.0) collected in the United States between 2014 and 2015. BMI was the exposure variable, and 30-day in-hospital mortality was the outcome. Covariates included demographics, clinical parameters, comorbidities, and treatments. Univariate and multivariate logistic regression models examined the association between BMI and mortality, while generalized additive models with smooth curve fitting explored potential nonlinear relationships. The median BMI among the 17,510 patients was 28.91 kg/m². The overall 30-day in-hospital mortality was 16.7% (95% CI: 16.2%–17.3%). A non-linear relationship between BMI and mortality was identified with an inflection point at 21.5 kg/m². Below this point, each unit increase in BMI was associated with a 9% decrease in mortality risk (OR: 0.91, 95% CI: 0.87–0.94, p < 0.001). Above this point, the association became insignificant (OR: 0.99, 95% CI: 0.99-1.00, p = 0.054). In ICU patients with sepsis, BMI is nonlinearly associated with 30-day in-hospital mortality, showing a significant negative correlation when BMI is ≤ 21.5 kg/m². These findings provide valuable insights for clinical risk assessment.