The PH score for early differentiation of necrotizing fasciitis: a retrospective diagnostic prediction study

Early differentiation of necrotizing fasciitis (NF) from non-necrotizing soft tissue infections (non-NSTI) remains a clinical challenge because delayed diagnosis is associated with substantial morbidity and mortality. Existing diagnostic scoring systems have shown inconsistent performance across different populations and often rely on multiple laboratory variables, limiting their routine clinical applicability. Therefore, this study aimed to develop and internally validate a simple diagnostic score based on routinely available laboratory parameters for the early identification of NF.

A single-center retrospective study enrolled 92 NF and 80 non-NSTI patients (2022-2026). Independent predictors were identified via multivariable logistic regression, and a PH score(Platelet-to-lymphocyte ratio-Hemoglobin Score)was developed and internally validated. Performance was compared with LRINEC, SIARI, and LARINF scores using ROC analysis, with bootstrap and 10-fold cross-validation for internal validation. Calibration and clinical utility were assessed.

PLR (OR = 1.02) and Hb (OR = 0.96) were independent predictors. The PH score showed an AUC of 0.853 (95% CI 0.798-0.909), outperforming LRINEC (0.654), SIARI (0.736), and LARINF (0.602). At cutoff ≥ 2, sensitivity 86.7%, specificity 73.4%, accuracy 80.5%. Bootstrap-corrected AUC was 0.847, 10-fold cross-validation AUC 0.852. Calibration was good, and decision curve analysis showed higher net benefit than existing scores.

The PH score, incorporating PLR and hemoglobin, is a simple and readily available diagnostic tool for the early differentiation of NF from non-NSTIs. The model showed favorable discrimination, calibration, and internal validation performance and provided greater net benefit than existing scoring systems. External validation in multicenter prospective cohorts is warranted.