Point-of-care lung ultrasound predicts in-hospital mortality in acute heart failure

Background: B-lines have been associated with adverse clinical outcomes in patients with heart failure (HF) when found at hospital discharge or during outpatient visits. Whether lung ultrasound (LUS) assessed B-lines may predict in-hospital mortality in patients with acute HF is still undetermined. Aim: To evaluate the association between B-lines on admission and in-hospital mortality among patients admitted with acute HF. Methods: Hand-held LUS was used to examine patients with acute HF. LUS was performed in eight chest zones with a pocket ultrasound device and analyzed offline. The association between B-lines and in-hospital mortality was assessed using Cox regression models. Results: We included 62 patients with median age 56years, 69.4% men, and median left ventricle ejection fraction 25%. The sum of B-lines ranged from 0 to 53 (median 6.5). An optimal receiver operating characteristic-determined cut-off of >= 19 B-lines demonstrated a sensitivity of 57% and a specificity of 86% (area under the curve 0.788) for in-hospital mortality. The incremental prognostic value of LUS when compared with lung crackles or peripheral edema by integrated discrimination improvement was 12.96% (95% CI: 7.0-18.8, P=0.02). Patients with >= 19 B-lines had a 4-fold higher risk of in-hospital mortality (HR 4.38; 95% CI: 1.37-13.95, P<0.01). Conclusion: In patients admitted with acute HF, point-of-care LUS measurements of pulmonary congestion (B-lines) are associated with in-hospital mortality.