Comparison of Predictive Scoring Systems in Assessing Risk for Intensive Care Unit Admission and In-Hospital Mortality in Patients with Urinary Tract Infections

Abstract

Objective: We aimed to investigate the effectiveness of confusion, respiratory rate, blood pressure (CRB), CRB-65, and quick sequential organ failure assessment (qSOFA) in predicting intensive care unit (ICU) admission and in-hospital mortality of patients with urinary tract infections (UTI) compared with Systemic Inflammatory Response Syndrome (SIRS). Methods: Data of patients with UTI who visited the emergency department of a single centre between February 2018 and March 2020 were retrospectively analysed. Baseline characteristics were compared with the prevalence of ICU admission and in-hospital mortality. The effectiveness of CRB, CRB-65, qSOFA, and SIRS as indicators of ICU admission and in-hospital mortality were evaluated using the area under the receiver operating characteristic (AUROC) curve. Results: Overall, 1151 patients were included, of whom 132 (11.5%) were admitted to the ICU and 30 (2.6%) succumbed to in-hospital mortality. AUROC values of CRB, CRB-65, and qSOFA as predictors of ICU admission and in-hospital mortality were similar. CRB score >= 1 had a sensitivity and specificity of 71.3% and 73.5%, respectively, for ICU admission; 66.7% and 69.2%, respectively, for in-hospital mortality. CRB-65 score >= 2 had a sensitivity and specificity of 61.2% and 80.9%, respectively, for ICU admissions; 60% and 76.9%, respectively, for in-hospital mortality. A qSOFA score >= 1 had a sensitivity and specificity of 71.3% and 79.6%, respectively, for ICU admission; 66.7% and 74.8%, respectively, for in-hospital mortality. AUROC values of SIRS were 0.580 and 0.617 respectively for ICU admission and in-hospital mortality, which showed lower predictive performance than those of the other three scoring systems. Conclusion: In ICU admission, CRB, CRB-65, and qSOFA have better predictive performance than SIRS. CRB-65 and qSOFA have superior performance compared to CRB and SIRS in predicting mortality.