Blood eosinophil numbers as predictor of severity of community-acquired pneumonia in a prospective multicentre validation cohort

<bold>Background:</bold> A retrospective multicentre study suggested that blood eosinopenia (≤50/µL) predicts severity of community-acquired pneumonia (CAP) including mortality. <bold>Objective:</bold> This study aimed to validate these findings in a second independent prospective multicentre cohort. <bold>Methods:</bold> We analysed patient cases (≥18-years-old) with CAP using data from PROGRESS, a prospective observational study on hospitalised CAP. Outcome parameters reflecting the severity of the disease were compared between blood eosinopenia (≤50/µL) and non-eosinopenia (>50/µL). <bold>Results:</bold> Overall, 1,763 (860 eosinopenic and 903 non-eosinopenic) patient cases were included. Eosinopenia (versus non-eosinopenia) correlated with increased in-hospital mortality (2.8% versus 1.2%; RR 2.29; 95% CI 1.13-4.65; p=0.025), risk of mechanical ventilation (14.7% versus 7.1%; RR 2.07; 95% CI 1.55-2.75; p<0.0001), and length of stay (8 versus 7 days; p<0.0001), while time to in-hospital death was not different (17 versus 12 days; p=0.644). <bold>Conclusion:</bold> Our study using data from a prospective multicentre CAP cohort confirmed that blood eosinophil numbers predict severity of disease in hospitalised patients with CAP.