The efficacy of chest radiographs in detecting parapneumonic effusions

Background and objective: Current guidelines recommend parapneumonic effusions (PPEs) with a thickness of >= 10 mm be sampled via thoracentesis. We hypothesized that anteroposterior (AP) CXRs are not as sensitive as posteroanterior (PA) and lateral radiographs in identifying PPEs and should not be routinely used in patients with suspected pneumonia. Methods: Sixty-one hospitalized patients with pneumonia and PPE were retrospectively studied, all of whom had a CXR and CT scan within 24 h of each other. The CXRs of these patients were independently read by three pulmonologists for an effusion in each hemithorax, which was correlated with measured pleural fluid thickness on chest CT. Results: Lateral, PA and AP radiographs were equivalent in identifying the presence of PPEs. All three views missed more than 10% of PPEs. The sensitivities of lateral, PA and AP CXRs were 85.7%, 82.1% and 78.4%, respectively (P = 0.749); the specificity was 87.5%, 81.3% and 76.4%, respectively (P = 0.198). The majority of effusions missed in each view were on films with lower lobe consolidation. Conclusions: All three CXR views missed effusions of a size significant enough to warrant thoracentesis. Consideration should be given to obtaining additional imaging at the time patients present with pneumonia, particularly in those with lower lobe consolidation.