The Usefulness of Lung Ultrasound for the Aetiological Diagnosis of Community-Acquired Pneumonia in Children

Abstract

The aetiology of community-acquired pneumonia (CAP) is not easy to establish. As lung ultrasound (LUS) has already proved to be an excellent diagnostic tool for CAP, we analysed its usefulness for discriminating between the aetiologically different types of CAP in children. We included 147 children hospitalized because of CAP. LUS was performed in all patients at admission, and follow-up LUS was performed in most patients. LUS-detected consolidations in viral CAP were significantly smaller, with a median diameter of 15 mm, compared to 20 mm in atypical bacterial CAP (p = 0.05) and 30 mm in bacterial CAP (p < 0.001). Multiple consolidations were detected in 65.4% of patients with viral CAP and in 17.3% of patients with bacterial CAP (p < 0.001). Bilateral consolidations were also more common in viral CAP than in bacterial CAP (51.9% vs. 8.0%, p < 0.001). At follow-up, a regression of consolidations was observed in 96.6% of patients with bacterial CAP and in 33.3% of patients with viral CAP (p < 0.001). We found LUS to be especially suitable for differentiating bacterial CAP from CAP due to other aetiologies. However, LUS must be interpreted in light of clinical and laboratory findings.

Figure 1

Stratification of patients with community-acquired pneumonia according to the aetiology. CAP: community-acquired pneumonia; PCR: polymerase chain reaction-based assay from nasopharyngeal swab; ↑ PCT: increased serum procalcitonin concentration (>0.25 ng/ml); ↑ WBC: increased white blood cell count (>15 × 10⁹/L); CXR: chest X-ray. Thirty-six patients were excluded from the study because we could not determine the aetiology. Two patients were excluded from the study because of lack of the serologic confirmation of Mycoplasma pneumoniae infection and three patients because of the transfer to the intensive care unit. Thirteen patients in whom viruses were detected in the nasopharynx were classified as having bacterial CAP (co-infection or superinfection).

Figure 2

A receiver operating characteristics (ROC) curve analysis of the lung-ultrasound-detected consolidation size, discriminating between the bacterial and viral pneumonia. The optimal cut-off size was 21 mm, with a sensitivity of 80% and a specificity of 75% to diagnose bacterial CAP. The area under the ROC curve (AUC) was 0.85 (p < 0.001; 95% CI 0.79–0.92).