Lung Ultrasound Versus Chest Radiography for Acute Heart Failure: Impact of Heart Failure History and Pleural Effusion.

Background/

Objectives

This is the first prospective, same-day, multi-modality comparison of lung ultrasound (LUS) and chest radiography (CXR) for detecting acute heart failure (AHF) in non-critical patients with dyspnoea, examining the impact of chronic heart failure and pleural effusion, using low-dose chest CT (LDCT) as an objective comparator, and cardiologists-adjudicated AHF as reference standard.

Methods

An observational study of 240 consecutive non-critical patients ≥50 years admitted with dyspnoea was conducted. Unstable AHF cases were deemed ineligible. Each modality was evaluated at the population level with area under the curve (AUC), sensitivity, and specificity, and compared at the patient level using conditional odds ratio for the association to AHF adjudicated by blinded cardiologists. Congestion was defined by LUS as (a) ≥3 B-lines bilaterally, or (b) B-lines combined with pleural effusion, and (c) CXR, interpreted by two thoracic radiologists, using (d) LDCT as an objective comparator.

Results

Among 240 patients (66 with cardiologist-adjudicated AHF, 58 with chronic heart failure), LUS (b) demonstrated a diagnostic accuracy at population level of AUC = 0.82 (sensitivity = 80%, specificity = 84%), while CXR (c) achieved AUC = 0.80 (sensitivity = 68%, specificity = 91%), with CXR showing a modest but statistically significant difference over LUS at the patient level (OR = 1.51, p = 0.03). Incorporating pleural effusion into LUS increased its AUC from 0.67 to 0.82 (a vs. b, p < 0.001). The objective comparator, LDCT (d), achieved an AUC = 0.92 (sensitivity = 74%, specificity = 96%). In patients with chronic heart failure, LUS (b) and CXR (c) performed comparably (p = 0.87), whereas in those without chronic heart failure, CXR was superior (p = 0.04).

Conclusions

In non-critical, diagnostically challenging patients with dyspnoea, in whom critical AHF cases were not eligible, including pleural effusion improved LUS accuracy for AHF. Diagnostic performance differed by heart failure history, with CXR superior in new-onset heart failure, while LUS and CXR performed comparably in chronic heart failure.