Adverse event prediction in immunocompetent adult patients hospitalized with community-acquired pneumonia

Abstract

Background: Community-acquired pneumonia (CAP) causes significant morbidity and mortality in adults. Aim: To compare the accuracy of four validated rules for predicting adverse outcomes in patients hospitalized with CAP. Patients and methods: We compared the pneumonia severity index (PSI), British Thoracic Society score (CURB-65), SMART-COP and severe CAP score (SCAP) in 659 immunocompetent adult patients aged 18 to 101 years, 52% male, hospitalized with CAP. Major adverse outcomes were: admission to ICU, need for mechanical ventilation (MV), in-hospital complications and 30-day mortality. Mean hospital length of stay (LOS) was also evaluated. The predictive indexes were compared based on sensitivity, specificity, and area under the curve of the receiver operating characteristic curve. Results: Of the studied patients, 77% had comorbidities, 23% were admitted to the intensive care unit and 12% needed mechanical ventilation. The rate of all adverse outcomes and hospital LOS increased directly with increasing PSI, CURB-65, SMART-COP and SCAP scores. The sensitivity, specificity and area under the curve of the prognostic indexes to predict adverse events were: Admission to ICU (PSI: 0.48, 0.84 and 0.73; SMART-COP: 0.97, 0.23 and 0.75; SCAP: 0.57, 0.81 and 0.76); use of MV (PSI: 0.44, 0.84 and 0.75; SMART-COP: 0.96, 0.35 and 0.84; SCAP: 0.53, 0.87 and 0.78); 30-days mortality (PSI: 0.45, 0.97 and 0.83; SMART-COP: 0.94, 0.29 and 0.77; SCAP: 0.53, 0.95 and 0.81). CURB-65 had a lower discriminatory power compared to the other indices. Conclusions: PSI score and SCAP were more accurate and specific and SMART-COP was more sensitive to predict the risk of death. SMART-COP was more sensitive and SCAP was more specific in predicting the use of mechanical ventilation.