A prediction model for admission to the intensive care unit in patients with perianal necrotizing fasciitis

Abstract

Introduction: Patients with perianal necrotizing fasciitis (PNF) frequently require admission to the intensive care unit (ICU). The study aimed to develop a novel scoring system to predict ICU admission in PNF patients.

Methodology: This cohort study retrospectively recruited patients in the Nanjing Hospital of Chinese Medicine. The outcome was the admission to the ICU. Random forest was used to select variables for the development of a new scoring system, whose performance was assessed using the area under the curve (AUC) with a 95% confidence interval (CI). Integrated discrimination improvement (IDI) and net reclassification improvement (NRI) were applied to assess the performance improvement of Sequential Organ Failure Assessment (SOFA) score and Fournier’s Gangrene Severity Index (FGSI) compared to the new scoring system.

Results: Totally 106 eligible individuals with PNF were enrolled. SOFA, age, course of disease, and extent of disease were selected to develop the new scoring system, which was named “modified SOFA” (mSOFA). The AUC of the mSOFA was 0.974 (95% CI: 0.931-1.000). SOFA (NRI: -0.72, p = 0.010; IDI = -0.05, p = 0.002) and FGSI (NRI: -1.50, p < 0.001; IDI = -0.55, p < 0.001) demonstrated a decreased predictive performance for the ICU admission compared to mSOFA.

Conclusions: The mSOFA scoring system had a better predictive performance for the ICU admission than SOFA and FGSI, indicating that mSOFA may be a reliable tool for the prediction of ICU admission in PNF patients.