The utility of serial procalcitonin measurements in addition to pneumonia severity scores in hospitalised community-acquired pneumonia: A multicentre, prospective study

•Serial measurements of PCT were useful for predicting prognosis in CAP.•Serial measurements of PCT were also useful for predicting treatment failure.•Increased PCT from admission to days 3-4 was a risk factor for poor prognosis.•PCT measurements improved the accuracy of PSI for predicting prognosis. The usefulness of serial procalcitonin (PCT) measurements for predicting the prognosis and treatment efficacy for hospitalised community-acquired pneumonia (CAP) patients was investigated. This prospective, multicentre, cohort study enrolled consecutive CAP patients who were hospitalised at 10 hospitals in western Japan from September 2013 to September 2016. PCT and C-reactive protein (CRP) were measured on admission (PCT D1 and CRP D1), within 48–72 h after admission (PCT D3 and CRP D3), and within 144–192 h after admission. CURB-65 and the Pneumonia Severity Index (PSI) were assessed on admission. The primary outcome was 30-day mortality; secondary outcomes were early and late treatment failure rates. A total of 710 patients were included. The 30-day mortality rate was 3.1%. On multivariate analysis, only PCT D3/D1 ratio >1 [odds ratio (95% confidence interval): 4.33 (1.46–12.82),P = 0.008] and PSI [odds ratio (95% confidence interval): 2.32 (1.07–5.03), P = 0.03] were significant prognostic factors. Regarding treatment efficacy, PCT D3/D1 >1 was a significant predictor of early treatment failure on multivariate analysis. PCT D3/D1 with the PSI significantly improved the prognostic accuracy over that of the PSI alone. PCT should be measured consecutively, not only on admission, to predict the prognosis and treatment efficacy in CAP.