What question did this study set out to answer?

To evaluate the predictive ability of various sepsis scoring systems for in-hospital mortality in PICU patients with suspected infection.

April 6, 2026Open Access

Comparison of Phoenix Sepsis Score and traditional scoring systems in predicting mortality in PICU patients with suspected infection

Key Points

To evaluate the predictive ability of various sepsis scoring systems for in-hospital mortality in PICU patients with suspected infection.
Prospective single-center cohort study
Children with suspected infection admitted to PICU
Assessing discrimination using area under the receiver operating characteristic curve (AUC)
Calculating calibration with Hosmer–Lemeshow test
Among 519 eligible children, 32 in-hospital deaths (6.2%) occurred
AUC for different scores: PSS 0.86, Phoenix-8 0.89, SIRS 0.73, pSOFA 0.87, PELOD-2 0.90, PRISM III 0.89
PSS showed comparable performance to pSOFA but was lower than Phoenix-8, PELOD-2, and PRISM III for mortality prediction
No significant miscalibration was found among the scoring systems

Abstract

Sepsis is a major contributor to mortality in pediatric intensive care units (PICU). Timely and accurate risk stratification in children with sepsis is essential to guide clinical management. The objective of this study was to evaluate and compare the ability of the Phoenix Sepsis Score (PSS), Phoenix-8 score, systemic inflammatory response syndrome (SIRS) score, Pediatric Sequential Organ Failure Assessment (pSOFA), Pediatric Logistic Organ Dysfunction-2 (PELOD-2) score, and Pediatric Risk of Mortality III (PRISM III) score to predict in-hospital mortality among PICU patients with suspected infection in low- and middle-income countries. We conducted a prospective, single-center cohort study of children admitted to the PICU with suspected infection. The primary outcome was in-hospital mortality. We assessed scoring system discrimination by calculating the area under the receiver operating characteristic curve (AUC) and evaluated calibration using the Hosmer–Lemeshow test. Among 519 eligible children, 32 in-hospital deaths occurred (6.2%). The median time to death was 4 days (IQR, 1–9). For mortality prediction, the AUCs were 0.86 (95% CI, 0.77–0.95) for PSS, 0.89 (95% CI，0.82–0.95) for Phoenix-8, 0.73 (95% CI，0.63 –0.83) for SIRS, 0.87 (95% CI，0.79–0.94) for pSOFA, 0.90 (95% CI，0.85–0.96) for PELOD-2, and 0.89 (95% CI，0.82–0.97) for PRISM III. PSS exceeded SIRS and was comparable to pSOFA for in-hospital mortality prediction, while Phoenix-8, PELOD-2, and PRISM III showed higher AUCs. For the combined outcome of early death (within 72 h of PICU admission) or extracorporeal membrane oxygenation (ECMO) use, PSS showed similar performance to pSOFA, Phoenix-8, and PELOD-2. No significant miscalibration was observed across scoring systems. PSS demonstrated good discriminatory ability in identifying high-risk patients for mortality in the context of low- and middle-income countries. It is suitable for the early identification of adverse outcomes in PICU patients with infections.

AI에게 질문

Bookmark

View Full Paper