Wenbo Fu, Yumei Li, Kee Chong Ng, Sheng Fu
Pediatric sepsis remains a significant global health burden with considerable mortality and morbidity, particularly in resource-limited settings. This umbrella review synthesizes current evidence from systematic reviews and meta-analyses on the definitions, epidemiology, diagnostic approaches, prognostic models, and emerging prediction technologies of pediatric sepsis worldwide. Through comprehensive evaluation of 18 high-quality systematic reviews published between 2019 and 2024, we identified critical gaps in early recognition and management protocols while highlighting promising technological innovations that demonstrate robust prediction potential. Our analysis revealed substantial variation in sepsis outcomes across different healthcare settings, with approximately 85% of global pediatric sepsis deaths occurring in low- and middle-income countries (LMICs). Quantitative synthesis of prediction models demonstrated pooled sensitivity of 0.81 (95% confidence interval [CI]: 0.76-0.85) and specificity of 0.79 (95% CI: 0.74-0.84), with area under the receiver operating characteristic curve values ranging from 0.78 to 0.94 depending on resource setting and model complexity. The integration of machine learning algorithms with traditional clinical markers showed particular promise for developing context-appropriate prediction tools that function effectively across resource settings. Etiological analysis revealed distinct pathogen profiles by age and geography: Gram-negative organisms were predominant in neonates (68%) and LMICs (72%), while Gram-positive pathogens accounted for 58% of cases in high-income country settings. Antimicrobial resistance exceeded 50% for common pathogens in LMICs, critically complicating empirical treatment. This review provides a comprehensive foundation for future research directions and implementation strategies to reduce the global burden of pediatric sepsis through precision medicine approaches and health system strengthening.