Background: Much debate exists around the relative diagnostic accuracy of physical exam (PE) versus bedside ultrasound (US) in patients with dyspnea. We compared the accuracy of structured comprehensive PE and lung US in ED patients with dyspnea.
Methods: Using a prospective, observational crossover study design, we enrolled hemodynamically stable adults who presented with dyspnea to a large academic ED. ED physicians performed both a structured PE (inspection, palpation, percussion, auscultation) and lung US (presence of A or B lines, sliding, effusion, and consolidation in 6 lung fields bilaterally) on each patient, the order of which was randomized. After each evaluation, the physician rated degree of certainty (low, moderate, high) that any of 6 diagnoses (pulmonary edema, COPD, pneumonia, pleural effusion, asthma and pneumothorax) were present. Patients also rated their confidence in PE and US, before and after evaluation. A masked adjudicator determined final diagnosis based on all clinical data. We compared the accuracies (true positive plus true negative/all) of PE and US using the adjudicated diagnosis as the criterion standard. Patients’ relative confidence level with each evaluation was also assessed.
Results: We enrolled 67 patients. Mean (SD) age was 63 (18), 38 (57%) were male. Patients were randomized to PE (n=37) or US (n=30) first. 20 patients had more than 1 diagnosis. Final diagnoses included pulmonary edema (n=22), COPD (n=18), pneumonia (n=11), pleural effusion (n=7), and asthma (n=3). 34 patients had none of the six conditions. Accuracies of PE and US for the various diagnoses were pulmonary edema (72 [59-82] vs 72 [59-82]%), COPD (79 [67-88] vs 84 [72-91]%), pneumonia (79 [67-88] vs 88 [77-94]%), pleural effusion (97 [89-99] vs 87 [76-93] %), and asthma (91 [81-96] vs 94 [85-98] %); P>0.05 for all. Diagnostic accuracies did not differ based on order of evaluation. Prior to evaluation only 1 patient had more confidence in PE than US, 36 patients had more confidence in US, and 30 had equal confidence. After evaluation 3 patients had more confidence in PE, 33 had more confidence in US, and 31 considered them equal.
Conclusion: PE and US have similar accuracies at diagnosing cardiopulmonary conditions in ED patients with dyspnea. Patients were more confident in US than PE.