Serial procalcitonin levels for predicting prognosis in community‐acquired pneumonia

Background and objective This study aimed to investigate the usefulness of addition of serial measurements of procalcitonin ( PCT ) to C‐reactive protein ( CRP ) values and pneumonia severity scores, such as CURB ‐65 (confusion, urea > 7 mmol/L, respiratory rate ≥ 30 breaths/min, low blood pressure (systolic < 90 mm Hg or diastolic ≤ 60 mm Hg) and age ≥ 65 years) and the Pneumonia Severity Index, and attempted to create and evaluate a new scoring system for predicting mortality risk using the biomarkers and pneumonia severity scores. Methods A total of 365 hospitalized community‐acquired pneumonia ( CAP ) patients in an observational cohort study in which PCT was measured serially from admission to 2–3 days after admission between December 2010 and December 2014 were reviewed retrospectively. PCT and CRP were measured on admission ( PCT D1 and CRP D1 ) and within 48–72 h after admission ( PCT D3 and CRP D3 ). Results Twenty‐one patients died (5.8%), and 52 patients (14.2%) did not respond to initial therapy. On multivariate analysis, CRP D1  ≥ 100 mg/L ( P  = 0.002), CURB ‐65 ≥ 3 ( P  < 0.001) and PCT D3 / D1  ≥ 1 ( P  < 0.001) were significant predictors of 30‐day mortality. Peak CRP ( P  = 0.02) and PCT D3 / D1  ≥ 1 ( P  = 0.03) were significant predictors of initial treatment failure. Using the new scoring system that defines CRP D1  ≥ 100 mg/L as 2 points, CURB ‐65 ≥ 3 as 1 point and PCT D3 / D1  ≥ 1 as 1 point, in CAP patients with both CRP D1  ≥ 100 mg/L and CURB ‐65 ≥ 3 on admission, the 30‐day mortality rate was 21.8%, and with PCT D3 / D1  ≥ 1, it increased to 50.0%. Conclusion It is useful to add serial measurements of PCT to CRP measurement and assessment of CURB ‐65 on admission of CAP patients to predict prognosis and initial treatment failure.