Prognostic value of serial score measurements of the national early warning score, the quick sequential organ failure assessment and the systemic inflammatory response syndrome to predict clinical outcome in early sepsis

BACKGROUND AND IMPORTANCE: Sepsis is a common and potentially lethal syndrome, and early recognition is critical to prevent deterioration. Yet, currently available scores to facilitate recognition of sepsis lack prognostic accuracy. OBJECTIVE: To identify the optimal time-point to determine NEWS, qSOFA and SIRS for the prediction of clinical deterioration in early sepsis and to determine whether the change in these scores over time improves their prognostic accuracy. DESIGN: Post hoc analysis of prospectively collected data. SETTINGS AND PARTICIPANTS: This study was performed in the emergency department (ED) of a tertiary-care teaching hospital. Adult medical patients with (potential) sepsis were included. OUTCOME MEASURES AND ANALYSIS: The primary outcome was clinical deterioration within 72 h after admission, defined as organ failure development, the composite outcome of ICU-admission and death. Secondary outcomes were the composite of ICU-admission/death and a rise in SOFA at least 2. Scores were calculated at the ED with 30-min intervals. ROC analyses were constructed to compare the prognostic accuracy of the scores. RESULTS: In total, 1750 patients were included, of which 360 (20.6%) deteriorated and 79 (4.5%) went to the ICU or died within 72 h. The NEWS at triage (AUC, 0.62; 95% CI, 0.59-0.65) had a higher accuracy than qSOFA (AUC, 0.60; 95% CI, 0.56-0.63) and SIRS (AUC, 0.59; 95% CI, 0.56-0.63) for predicting deterioration. The AUC of the NEWS at 1 h (0.65; 95% CI, 0.63-0.69) and 150 min after triage (0.64; 95% CI, 0.61-0.68) was higher than the AUC of the NEWS at triage. The qSOFA had the highest AUC at 90 min after triage (0.62; 95% CI, 0.58-0.65), whereas the SIRS had the highest AUC at 60 min after triage (0.60; 95% CI, 0.56-0.63); both are not significantly different from triage. The NEWS had a better accuracy to predict ICU-admission/death