Burden and Outcomes of Sepsis Using Phoenix Criteria: Single-Center Cohort Study in a South African PICU, 2024

OBJECTIVES: To describe the burden and outcomes of pediatric sepsis using the Phoenix sepsis criteria in a South African PICU and to assess the Phoenix Sepsis Score (PSS) discriminatory performance for mortality and length of stay (LOS) in this setting. DESIGN: Single-center, retrospective cohort. SETTING: PICU at Red Cross War Memorial Children's Hospital, Cape Town, South Africa. PATIENTS: Emergency medical admissions during 2024, excluding preterm infants (< 37 wk postconceptual age), birth hospitalizations, surgical and palliative admissions, nosocomial sepsis (> 48 hr after admission), patients with incomplete data for sepsis classification, and PICU readmissions during the same hospitalization. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Of 1124 admissions screened, 332 children met inclusion criteria. The median age was 12 months (interquartile range [IQR], 3-45.8 mo). Based on the Phoenix criteria, 137 of 332 (41.3%) had sepsis and 101 of 332 (30.4%) had septic shock. Compared with nonsepsis cases, sepsis was associated with younger age, worse illness severity (median pediatric index of mortality, 0.07 vs. 0.02; p < 0.001), higher need for invasive ventilation (71.5% vs. 31.8%; p < 0.001), and higher mortality (19.7% vs. 4.1%; p < 0.001). Each one-point increase in PSS doubled the odds of mortality (adjusted odds ratio, 2.21 [95% CI, 1.8-2.8]; p < 0.001) and increased LOS by 1.3 days (95% CI, 0.86-1.77 d; p < 0.001). The PSS showed strong discriminatory capacity for mortality (with area under the receiver operating characteristic curve 0.894 [95% CI, 0.83-0.96]). The optimal mortality threshold was PSS greater than or equal to 4 (negative predictive value, 97.6%). CONCLUSIONS: In 2024, in our PICU setting, sepsis accounted for a high proportion of emergency medical admissions and deaths. The PSS demonstrated strong discriminatory performance for mortality and LOS, providing preliminary evidence supporting its feasibility as a risk-stratification tool in this South African PICU setting.