Serial Procalcitonin as a Predictor of Bacteremia and Need for Intensive Care Unit Care in Adults With Pneumonia, Including Those With Highest Severity: A Prospective Cohort Study.

Abstract

BackgroundProcalcitonin (PCT) is a prohormone that rises in bacterial pneumonia and has promise in reducing antibiotic use. Despite these attributes, there are inconclusive data on its use for clinical prognostication. We hypothesize that serial PCT measurements can predict mortality, intensive care unit (ICU) admission, and bacteremia.MethodsA prospective cohort study of inpatients diagnosed with pneumonia was performed at a large tertiary care center in Boston, Massachusetts. Procalcitonin was measured on days 1 through 4. The primary endpoint was a composite adverse outcome defined as all-cause mortality, ICU admission, and bacteremia. Regression models were calculated with area under the receiver operating characteristic curve (AUC) as a measure of discrimination.ResultsOf 505 patients, 317 patients had a final diagnosis of community-acquired pneumonia (CAP) or healthcare-associated pneumonia (HCAP). Procalcitonin was significantly higher for CAP and HCAP patients meeting the composite primary endpoint, bacteremia, and ICU admission, but not mortality. Incorporation of serial PCT levels into a statistical model including the Pneumonia Severity Index (PSI) improved the prognostic performance of the PSI with respect to the primary composite endpoint (AUC from 0.61 to 0.66), bacteremia (AUC from 0.67 to 0.85), and need for ICU-level care (AUC from 0.58 to 0.64). For patients in the highest risk class PSI >130, PCT was capable of further risk stratification for prediction of adverse outcomes.ConclusionSerial PCT measurement in patients with pneumonia shows promise for predicting adverse clinical outcomes, including in those at highest mortality risk.