Mortality prediction using serum biomarkers and various clinical risk scales in community-acquired pneumonia

Abstract

We evaluated the predictive value of serum biomarkers and various clinical risk scales for the 28-day mortality of community-acquired pneumonia (CAP). Serum biomarkers including procalcitonin (PCT) and C-reactive protein (CRP) were evaluated in the emergency department. Scores for the pneumonia severity index (PSI); CURB65 (confusion, urea, respiration, blood pressure; age >65 years); Infectious Disease Society of America (IDSA) and American Thoracic Society (ATS) guidelines for severe CAP; Acute Physiology, Chronic Health Evaluation (APACHE) II; Sequential Organ Failure Assessment (SOFA); and quick SOFA (qSOFA) were calculated. Receiver-operating characteristic curves for 28-day mortality were calculated for each predictor using cut-off values, and we applied logistic regression models and area under the curve (AUC) analysis to compare the performance of predictors. Of the 125 enrolled patients, 13 died within 28 days. The AUCs of the PCT and CRP were 0.83 and 0.77, respectively. Using a PCT level >5.6 μg/L as the cut-off, the sensitivity and specificity for mortality were 76.9% and 90.2%, respectively. The three pneumonia severity scales showed an AUC of 0.86 (PSI), 0.87 (IDSA/ATS) and 0.77 (CURB65). The AUCs of the APACHE II, SOFA and qSOFA scores were 0.85, 0.83 and 0.81, respectively. The models combining CRP and/or PCT with PSI or the IDSA/ATS guidelines demonstrated superior performance to those of either PSI or the IDAS/ATS guidelines alone. In conclusion, serum PCT is a reliable single predictor for short-term mortality. Inclusion of CRP and/or PCT could significantly improve the performance of the PSI and IDAS/ATS guidelines.