“The Quiet Signature: A Hidden Inflammation Index Predicting Sudden Cardiometabolic Collapse in ‘Stable’ Type 2 Diabetes”

¹Ioanna Kyriazis ²Nikos Papadopoulos

Abstract

`Background: Pediatric sepsis can progress rapidly, and clinical signs may appear late. We evaluated whether a compact bedside algorithm combining biomarkers and early warning scores can predict pediatric sepsis-related deterioration earlier than standard clinical assessment.
Methods: We conducted a prospective diagnostic accuracy study in a tertiary children’s hospital including 612 patients (median age 4.1 years; IQR 1.2–9.3) presenting with suspected infection to the emergency department or pediatric wards. The index test (“SEPSIS-EDGE”) integrated procalcitonin, lactate, absolute neutrophil count, and a modified pediatric early warning score (PEWS). The reference standard was sepsis-related clinical deterioration within 24 hours (PICU transfer, vasoactive initiation, mechanical ventilation, or death), adjudicated by an independent panel blinded to the index score. We calculated sensitivity, specificity, AUC, and compared SEPSIS-EDGE against PEWS alone.
Results: Sepsis-related deterioration occurred in 68/612 children (11.1%). SEPSIS-EDGE ≥7 demonstrated sensitivity 0.88 (95% CI 0.78–0.94), specificity 0.73 (95% CI 0.69–0.77), PPV 0.28, and NPV 0.98. Discrimination was excellent (AUC 0.89; 95% CI 0.85–0.93), outperforming PEWS alone (AUC 0.76; 95% CI 0.70–0.82; p<0.001). Median lead time to deterioration signal was 5.6 hours earlier (IQR 3.2–8.4) compared with clinician-triggered escalation documentation (p<0.001). In multivariable analysis, SEPSIS-EDGE ≥7 independently predicted deterioration (adjusted OR 9.4; 95% CI 4.8–18.2; p<0.001).
Conclusion: A practical biomarker-enhanced bedside algorithm identified high-risk pediatric infection cases hours before overt clinical decline, offering a scalable approach for earlier escalation and targeted intervention in resource-intense settings.