U-shaped association between triglyceride/HDL ratio and all-cause mortality in obese sepsis patients: A retrospective study based on eICU database.

Abstract

Objective: This study explored the link between the TG/HDL ratio and mortality in obese sepsis patients using the eICU database.

Design: Retrospective observational study.

Setting: Intensive Care Medicine.

Patients: According to Sepsis 3.0 criteria, sepsis is diagnosed with infection and a SOFA score ≥2. This study included adults (age ≥18) with ICU stays ≥48 h to ensure data stability. Patients with diabetes, acute pancreatitis, or those on lipid-lowering or antidiabetic treatments were excluded due to their impact on lipid and glucose metabolism, which could bias the analysis of the TG/HDL ratio and all-cause mortality.

Interventions: 938 obese sepsis patients were selected, and statistical methods (variable assessment, difference test, regression model, curve fitting, subgroup analysis) were used. The study was carried out using R 4.3.2 software.

Main variables of interest: TG/HDL.

Results: The TG/HDL ratio varied across quartiles and was significantly linked to mortality in obese sepsis patients. Regression and curve fitting showed a U-shaped relationship for TG/HDL < 8, with an inflection point at 3.59. The K-M curve confirmed this U-shaped pattern, and subgroup analysis revealed a significant interaction with hypertension.

Conclusions: A U-shaped relationship exists between TG/HDL and mortality in obese sepsis patients, with both low and high levels increasing death risk.