Background: Retrospective evidence suggests that eosinopenia is associated with worse outcomes
in patients with community-acquired pneumonia (CAP). However, data about the relationship
between blood eosinophil levels and well-established risk stratification scores for
CAP are missing, and a suitable eosinophil count threshold for distinguishing high-
from low-risk patients has not been determined yet. This study therefore investigated
blood eosinophil count at hospital admission as a biomarker for risk stratification
of CAP.
Methods: Adult patients were recruited in a prospective observational multi-centre study on
hospitalised CAP (PROGRESS). The correlation between blood eosinophil numbers at hospital
admission as continuous variable and risk stratification scores was analysed. A continuous
analysis of eosinophil numbers versus mortality and risk of mechanical ventilation
was performed for threshold determination of eosinopenia as a biomarker for risk stratification.
Based on this analysis, patients were allocated to an eosinopenia (≤ 50/µL) and non-eosinopenia
(> 50/µL) group. Their baseline characteristics and outcomes were compared.
Findings: Overall, 1763 (60.0% males, median age 63.0 years) patients were included. Eosinophil
counts correlated inversely with CRB-65 (p<0·0001), CURB-65 (p<0·0001), PSI (p<0·0001),
Infectious Disease Society of America/American Thoracic Society minor criteria (p<0·0001),
SOFA (p<0·0001), and Quick SOFA (p=0·0155). An eosinophil count threshold of 50/µL
was proven suitable for risk stratification: Eosinopenia (versus non-eosinopenia)
was associated with increased in-hospital mortality (2·8% versus 1·2%; RR 2·29; p=0·0251),
need for mechanical ventilation (14·7% versus 7·1%; RR 2·07; p<0·0001), and length
of stay (8 versus 7 days; p=0·0048). After compensating for multiple confounders including
glucocorticoid treatment in the multivariate analysis, eosinopenia≤50/µL (versus non-eosinopenia)
correlated with an increased (p=0·0003) need for mechanical ventilation.
Interpretation: In a prospective multicentre study, blood eosinophil count at hospital admission
correlates with well-established risk stratification scores for CAP. Eosinopenia≤50/µL
seems to be a promising biomarker for risk stratification.