Prognostic accuracy of qSOFA in predicting 28-day mortality among infected patients in an emergency department: a prospective validation study

Abstract

BackgroundFew prospective studies have evaluated the quick Sequential (Sepsis-Related) Organ Failure Assessment (qSOFA) criteria in emergency department (ED)settings. The aim of this study was to determine the prognostic accuracy of qSOFA compared with systemic inflammatory response syndrome (SIRS) in predicting the 28-day mortality of infected patients admitted to an ED.MethodsA prospective observational cohort study of all adult (≥18 years) infected patients admitted to the ED of Slagelse Hospital, Denmark, was conducted from 1 October 2017 to 31 March 2018. Patients were enrolled consecutively and data related to SIRS and qSOFA criteria were obtained from electronic triage record. Information regarding mortality was obtained from the Danish Civil Registration System. The original cut-off values of ≥2 was used to determine the prognostic accuracy of SIRS and qSOFA criteria for predicting 28-day mortality and was assessed by analyses of sensitivity, specificity, positive predictive value, negative predictive value, likelihood ratios and area under the receiver operating characteristic curve (AUROC) with 95% confidence intervals (CI).ResultsA total of 2112 patients were included in this study. A total of 175 (8.3%) patients met at least two qSOFA criteria, while 1012 (47.9%) met at least two SIRS criteria on admission. A qSOFA criteria of at least two for predicting 28-day mortality had a sensitivity of 19.5% (95% CI 13.6% to 26.5%) and a specificity of 92.6% (95% CI 91.4% to 93.7%). A SIRS criteria of at least two for predicting 28-day mortality had a sensitivity of 52.8% (95% CI 44.8% to 60.8%) and a specificity of 52.5% (95% CI 50.2% to 54.7%). The AUROC values for qSOFA and SIRS were 0.63 (95% CI 0.59 to 0.67) and 0.52 (95% CI 0.48 to 0.57), respectively.ConclusionBoth SIRS and qSOFA had poor sensitivity for 28-day mortality. qSOFA improved the specificity at the expense of the sensitivity resulting in slightly higher prognostic accuracy overall.