Breakthrough Findings from the Latest Pediatric Sepsis Research

The recent publication in the Journal of the American Medical Association (JAMA) has shed new light on the persistent challenge of pediatric sepsis in United States hospitals. Titled 'National Estimates of Pediatric Sepsis in US Hospitals Using Pediatric Sepsis Event Surveillance Criteria,' this study reveals that nearly 1 in 5 pediatric hospital deaths—specifically 17.8%—involve sepsis. Conducted by a team of researchers from leading institutions, the analysis draws from nearly 4 million hospitalizations between 2016 and 2023, projecting over 18,000 cases and approximately 1,877 deaths annually in 2022 alone.

This research underscores the critical role of academic medical centers in advancing surveillance methods for life-threatening conditions like sepsis. By adapting the Phoenix criteria into a scalable electronic health record (EHR)-based Pediatric Sepsis Event (PSE) definition, the study provides a robust framework for tracking incidence and outcomes, paving the way for improved quality care across hospitals nationwide.

Defining Pediatric Sepsis: From Phoenix Criteria to PSE Surveillance

Pediatric sepsis, or sepsis in children beyond the neonatal period (ages 30 days to 17 years), is a life-threatening condition characterized by organ dysfunction resulting from a dysregulated immune response to infection. Unlike adult sepsis, which uses the Sepsis-3 criteria, pediatric cases have historically lacked standardized definitions, leading to inconsistent reporting.

The Phoenix Sepsis Score, developed through an international consensus and validated in 2024, marks a significant advancement. It assigns points across four organ systems—respiratory, cardiovascular, coagulation, and neurological—totaling a score of 2 or more points in the presence of suspected infection to diagnose sepsis. Septic shock adds cardiovascular dysfunction. Universities like the University of Zurich and Harvard Medical School played key roles in its creation.

The PSE definition adapts this for EHRs, using routinely captured data such as blood cultures, antimicrobial administration, ventilation status, platelet counts, and Glasgow Coma Scale (GCS) scores. Validation against physician-reviewed cases showed 69.9% sensitivity and 93.1% specificity, outperforming traditional ICD-10 codes. This tool enables real-time surveillance, essential for academic health systems training future clinicians.

Core Statistics: Incidence, Mortality, and Burden

Across 3.9 million hospitalizations in datasets from Epic Cosmos and HCA Healthcare, sepsis affected 1.3% of cases, equating to 51,542 identified events. In-hospital mortality stood at 10.1% overall—4.0% for sepsis without shock and 14.0% for septic shock. Sepsis contributed to 17.8% of all pediatric hospital deaths, highlighting its disproportionate impact.

Seventy-two percent of cases were community-onset, with 61.6% involving septic shock. These figures were consistent across secondary datasets from PINC AI, Phoenix Children's Hospital, and PECARN, affirming the PSE's reliability.

Demographics and Risk Factors in Pediatric Sepsis

The average age of affected children was 6.6 years, with 44.3% female and a slight male predominance in incidence (1.4% vs. 1.2%). Comorbidities were prevalent: 36.5% had neurologic conditions, 38.3% cardiovascular issues, and 43.2% gastrointestinal disorders—far higher than in general pediatric admissions.

• Higher incidence in infants due to immature immune systems.
• Hospital-onset cases showed elevated mortality, often linked to invasive procedures.
• Common infection sources: respiratory (48.3%), with blood cultures positive in 16.1% (top pathogens: Staphylococcus aureus 19.4%, Streptococcus spp. 10.6%).

These patterns emphasize the need for targeted research at universities like Children's Hospital of Philadelphia (CHOP), affiliated with the University of Pennsylvania Perelman School of Medicine.

Temporal Trends: Stability Amid Fluctuations

From 2016 to 2022, national estimates showed no significant change in cases (annual 0.2%) or deaths (0.3%), despite a temporary incidence peak at 1.6% in 2020, possibly tied to COVID-19. Mortality dipped pre-2020, rose during the pandemic, then stabilized. Incidence rose in older children (6-17 years), signaling evolving epidemiology.

This stability contrasts with adult sepsis declines, urging pediatric-specific interventions developed through collaborative academic efforts.

Global Context: US vs. Worldwide Pediatric Sepsis

Globally, sepsis claims over 3 million child lives yearly, with higher mortality in low-resource settings (up to 28.5% vs. US 10.1%). US incidence (1.3%) aligns with high-resource benchmarks, but hospital mortality remains a concern. The PSE framework offers potential for international adaptation, as noted by editorial authors from the University of Zurich and University of Queensland.

Academic Contributions: Universities Driving the Research

Lead author Chanu Rhee, MD, MPH, from Harvard Medical School and Harvard Pilgrim Health Care Institute, spearheaded this work alongside collaborators from the University of Pennsylvania, University of Colorado, Northwestern University Feinberg School of Medicine, University of Utah, and Nemours Children’s Health (affiliated with Thomas Jefferson University). These institutions validated the PSE through manual reviews at diverse sites.

CHOP's Pediatric Sepsis Program exemplifies university-led innovation in early recognition and treatment.Learn more about CHOP's efforts.

Implications for Hospital Surveillance and Quality Improvement

The PSE enables benchmarking, policy-making, and interventions like the CDC's Sepsis Core Elements. Hospitals can implement open-source toolkits for prospective tracking, reducing hospital-onset cases through better hygiene and timely antimicrobials. Editorial experts stress standardizing EHR data for neurology to boost accuracy.

Prevention Strategies: Lessons from University Research

Prevention starts with vaccines, hand hygiene, and prompt infection treatment. Quality improvement programs at institutions like Primary Children's Hospital (University of Utah) have sustained sepsis alerts and bundles, lowering mortality. Machine learning models from CU Anschutz predict sepsis early, bridging diagnosis gaps.

• Family education on fever signs.
• Antibiotic stewardship to curb resistance.
• Screening tools in emergency departments.

Future Outlook: Research Gaps and Innovations

Challenges include neonatal exclusion, viral sepsis undercounting, and international validation. Future studies may integrate AI for real-time alerts and expand to community settings. Academic careers in pediatric critical care offer opportunities to shape these advances.Read the accompanying JAMA editorial.

Photo by CDC on Unsplash