177 Predictive Value of SEPSIS-3 Criteria in the Emergency Department

Abstract

In 2016 the Sepsis-3 Task Force proposed the quick Sequential Organ Failure Assessment (qSOFA) score to predict adverse outcomes in patients with a suspected infection in non-intensive care unit (ICU) settings. We aimed to compare the qSOFA to previous used scoring systems including the Systemic Inflammatory Response Syndrome (SIRS) criteria, Modified Early Warning System (MEWS) and National Early Warning Score (NEWS) for the prediction of adverse outcomes in patients admitted to the emergency department (ED) with a suspected or proven systemic infection. It was a prospective Dutch multicenter observational study performed at 4 large EDs. All patients ≥ 18 years presenting a suspected or proven systemic infection were included between November 2016 and February 2018. At ED arrival the vital signs and laboratory tests were recorded. Follow-up time was 30 days. The qSOFA, SIRS, MEWS and NEWS were calculated and receiver operating characteristics analysis was performed to determine the predictive value for adverse outcomes (ie, ICU admission, 30-day mortality and the composite outcome of ICU admission or 30-day mortality). A total of 1228 patients were included in the study, aged 63 (± 18.0) years and 592 (48.2%) were woman. Of the included patients, 59 (4.9%) were admitted to the ICU and 55 (4.5%) died within 30 days of admission. From those, 15 were previously admitted to the ICU (composite outcome of 8.1% of the patients). For ICU admission, the qSOFA had the lowest area under the curve (AUC) compared to other scorings systems (AUC 0.69 versus AUC > 0.73, P <0.001). The most accurate predictor of mortality at 30-days was NEWS with AUC = 0.75 (p<0.001). The AUC of qSOFA was qSOFA AUC = 0.69 (<0.001). The composite outcome was better predicted by NEWS with a AUC = 0.77 (P<.001) and again poorly by the other scoring systems (qSOFA = 0.69, P<0.001). The pairwise comparisons for composite outcome and 30-day mortality between qSOFA versus NEWS were statistically significant (P <0.008). Compared with the other scoring systems, qSOFA had the lowest sensitivity (18.2%) in predicting 30-day mortality in patients with (suspected) infection at the ED (sensitivity > 61%). Optimal cutoff values for predicting 30-day mortality were found for qSOFA ≥1 (sensitivity 69%, specificity of 66%), SIRS ≥2 (sensitivity 76%, specificity of 44%), MEWS ≥2 (sensitivity 86%, specificity 41%) and NEWS ≥5 (sensitivity 78%, specificity 66%). Due to its low sensitivity, qSOFA is not a proper bedside tool to predict adverse outcomes in the ED. For patients presenting to the ED with suspected or proven infection NEWS provides the most accurate tool for predicting adverse outcomes with an optimal cutoff value of NEWS ≥5.