Analysis of the heterogeneous treatment effect of vasoactive drug dosage and time on hospital mortality across different sepsis phenotypes: a retrospective cohort study

• DOI: 10.1186/s40001-025-02660-x
• Publication Date: 2025-05-24T01:20:16Z
• Authors: Jiacheng Shen; Kun Fang; Jianhong Xie; Dongsheng Sun; Li Li
• Keywords: Vasoactive

The heterogeneity of sepsis poses challenges for the individualized treatment vasoactive drugs. This study used data from ICUs in MIMIC-IV (2008-2019) and eICU (2014-2015) databases, identified by sepsis-3 criteria, stratified into phenotypes consensus K-means. norepinephrine equivalence (NEE) formula balance different drugs, with NEE captured hourly up to 72 h record both time use dosage. logistic regression model, including phenotype-dosage-time interactions, examined heterogeneous effects on hospital mortality. To address confounding, three models were fitted: Model 1 unadjusted, 2 adjusted age sex, 3 additionally included 7 clinical variables via machine learning directed acyclic graph. Nonlinear dosage was further analyzed based restricted cubic splines. P values interaction Bonferroni-adjusted. A total 54,673 patients phenotype identification, 8,803 evaluate effect Four identified: A, B, C D. Phenotype D most severe subgroup, followed C, while B mild subgroups. In 3, each 0.05 μg/kg/min increase linked higher mortality (OR 1.328, 95% CI 1.314-1.342; p < 0.001). Longer also significantly increased risk 1.006, 1.005-1.007; addition, these associations varied (P RCS consistently showed than other at dosages 0.1-0.5 µg/kg/min, this gap increasing over time, showing a clear dosage-time dependence. displayed lower overall but steepest relative as dosage: 1.309; OR time: 1.005) 3. reached highest when exceeded 0.5 which Finally, U-shaped curve minimum around 20% 0.03-0.05 µg/kg/min. Sepsis differ their drug use, indicating need phenotype-specific strategies improve outcomes.