The Etiology of Community-Acquired Pneumonia Correlates with Serum Inflammatory Markers in Children

Abstract

Community-acquired pneumonia (CAP) severely affects pediatric hospitalizations. This study assessed the contribution of CAP to hospitalizations, its etiology in relationship with age, and the inflammatory markers. Between 2013 and 2018, 1064 CAP patients were hospitalized and diagnosed with bacterial/possibly bacterial pneumonia (BP), viral/possibly viral pneumonia (VP) and atypical pneumonia (AP). The etiology was confirmed using blood/pleural fluid culture/polymerase chain reaction (PCR), rapid antigen test/PCR in nasopharyngeal swabs, or serological studies. CAP accounted for 9.9% of hospitalizations and 14.8% of patient days. BP was diagnosed in 825 (77.5%), VP in 190 (17.9%), and AP in 49 (4.6%) cases; the confirmed etiology (n = 209; 20%) included mostly influenza (39%; n = 82), respiratory syncytial virus (RSV, 35%; n = 72), and Mycoplasma pneumoniae (19%; n = 39). VP frequency decreased with age (41% in < 3 mo to 9% in ≥ 60 mo), in contrast to AP (13% in ≥ 60 mo). Among the analyzed parameters, the best differentiating potential was shown by: C-reactive protein (CRP, AUCBP-VP = 0.675; 95% CI: 0.634–0.715), procalcitonin (AUCBP-AP = 0.73; 95% CI: 0.67–0.794), and CRP/procalcitonin (AUCAP-VP = 0.752; 95% CI: 0.67–0.83); a good positive predictive value (88.8%, 98.3%, and 91.6%, respectively) but a low negative predictive value (29.5%, 13.1%, and 40.7%, respectively) was observed. CAP influences hospital patient days more than the crude number of patients would suggest. On a clinical basis, BP is mainly recognized, although viral pneumonia is confirmed most often. RSV and influenza are responsible for a huge percentage of hospitalized cases, as well as M. pneumoniae in children aged ≥ 5 years. Serum inflammatory markers may help differentiate etiological factors.