Performance of Sepsis-3 criteria in burn patients: A multicenter real-world study

Sepsis remains a leading cause of morbidity and mortality among patients with burn injury, yet diagnosis is particularly challenging because burn-induced hypermetabolic and inflammatory responses frequently mimic infection-related organ dysfunction. Although multiple sepsis definitions are used in clinical practice, their diagnostic performance and prognostic relevance in real-world burn populations remain uncertain. We conducted a retrospective multicenter validation study using the TriNetX U.S. Collaborative Network. Adult patients hospitalized with burn injury were identified. Clinically documented sepsis, defined by ICD-10-CM codes A40, A41, or R65.2 occurring within 7 days after the index burn event, served as the reference standard. The diagnostic accuracy of Sepsis-3, American Burn Association (ABA), and Mann–Salinas (MN) criteria was assessed using sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV). Prognostic relevance was evaluated by comparing 28-day all-cause mortality across sepsis definitions using propensity score–matched cohorts and Cox proportional hazards models. Among 527,197 adult burn patients, 8810 (1.4%) had clinically documented sepsis. Sensitivity was high across definitions (ABA, 86.7%; Mann–Salinas, 86.6%; Sepsis-3, 81.4%), whereas specificity varied substantially (Mann–Salinas, 77.0%; ABA, 61.6%; Sepsis-3, 45.8%). PPVs were uniformly low (2.5%–6.0%), reflecting substantial overclassification in this low-prevalence population, while NPVs exceeded 99% for all criteria. No significant differences in 28-day mortality were observed between definitions, with all pairwise hazard ratios overlapping and crossing unity. In adult burn patients, commonly used sepsis definitions function primarily as sensitive screening tools rather than confirmatory diagnostic instruments. The Mann–Salinas criteria demonstrated the most balanced performance. Sepsis assessment in burn populations should integrate physiologic criteria, clinical judgment, and confirmatory infection testing.