Journal of inflammation research
Association of SOFA-2 with in-Hospital Mortality in Refractory Septic Shock: A Multicohort Retrospective Study with a 12-Hour Landmark
By Hu E, Chen Q, Huang W and 2 others.
Abstract
Background
SOFA-2 is an updated organ dysfunction score designed to reflect contemporary intensive care practice. Its clinical relevance in refractory septic shock (RSS) remains unclear. We primarily evaluated the association between SOFA-2 and subsequent in-hospital mortality among patients who were alive and hospitalized at a 12-hour landmark after RSS identification. Comparison with conventional SOFA was secondary and exploratory.
Methods
This retrospective cohort study used MIMIC-IV as the primary cohort, with external consistency analyses in eICU-CRD and a local institutional cohort. RSS was operationalized using available electronic health record variables based on the SCCM/ESICM consensus framework, although some consensus elements were incompletely captured. Patients who died or were discharged before the 12-hour landmark were excluded from the primary analysis. SOFA-2 and conventional SOFA were assessed during the first 12 hours after RSS identification, with follow-up beginning at the landmark. Cause-specific Cox models were used for the primary analysis. Proportional-hazards diagnostics, time-varying coefficient models, time-dependent AUC, restricted cubic spline, subgroup, and sensitivity analyses were also performed.