Lung Ultrasound for Heart Failure Diagnosis in Primary Care

Background: Lung ultrasound (LUS) is useful for diagnosing pulmonary congestion, but its value in primary care remains unclear. We investigated whether LUS improved diagnostic accuracy in outpatients with heart failure (HF) suspicion. Methods and results: LUS was performed on 2 anterior (A), 2 lateral (L), and 2 posterior (P) areas per hemithorax. An area was positive when >= 3 B-lines were observed. Two diagnostic criteria were used: for LUS-C1, 2 positive areas of 4 (A-L) on each hemithorax; and for LUS-C2, 2 positive areas of 6 (A-L-P) on each hemithorax. A cardiologist blinded to LUS validated HF diagnosis. 162 patients were included (age 75.6 +/- 9.4 years, 70.4% women). Both LUS criteria, alone and combined with other HF diagnostic criteria, were accurate for identifying HF. LUS-C2 outperformed LUS-C1, showing remarkable specificity (0.99) and positive predictive value (0.92). LUS-C2, together with Framingham criteria, N-terminal pro-B-type natriuretic peptide, and electrocardiogram, added diagnostic value (area under the receiver operating characteristic curves 0.90 with LUS-C2 vs 0.84 without; P =.006). In the absence of N-terminal pro-B-type natriuretic peptide, LUS-C2 significantly reclassified one-third of patients above Framingham criteria and electrocardiogram (net reclassification improvement 0.65, 95% confidence interval 0.04-1.1). Conclusions: LUS was accurate enough to rule-in HF in a primary care setting. The accuracy of diagnostic workup for HF in primary care is enhanced by incorporating LUS, irrespective NT-proBNP availability.