Prognostic values of B-lines combined with clinical congestion assessment at discharge in heart failure patients

Aims We aim to investigate the additive effect of B-lines on lung ultrasound (LUS) for predicting outcome in patients with heart failure (HF) when combined with conventional assessment of clinical congestion. Methods and results This study prospectively enrolled 117 hospitalized HF patients (61 +/- 16 years, 70.1% males) who underwent congestion assessment by the 'wet/dry' status, clinical congestion score (CCS), and B-lines on LUS. The primary endpoint was all-cause mortality or hospitalization for HF during the 180-day follow-up after discharge. The 'Wet', CCS >= 3, and B-lines >5, indicators of congestion positive (+), were observed in 83.8%, 76.1%, and 70.1% of the patients on admission, respectively; and the numbers significantly decreased to 41.9%, 41.9%, and 35.9% at discharge, respectively. The agreement between the 'wet/dry' status and B-lines (58.1%) or between CCS and B-lines (56.4%) was moderate at discharge, in terms of both positive and both negative. By incorporating the B-lines with assessment of clinical congestion, the patients at discharge were divided into three phenotypes as clinical congestion (+), clinical congestion (-) with B-lines (+), and clinical congestion (-) with B-lines (-). The Kaplan-Meier analysis showed a better survival in the both (-) group ('wet/dry' with B-lines: Chi-square 10.591, P = 0.005; CCS with B-lines: chi(2) 6.239, P = 0.031). When the 'wet' patients (n = 49) being taken as the reference, the 'dry' patients with B-lines (+) (n = 21) had an identical risk of the composite endpoint (hazard ratio [HR] adjusted for clinical covariates 1.021, 95% confidence interval [CI] 0.480-2.134, P = 0.974), while the 'dry' patients with B-lines (-) (n = 47) had a lower risk (HR 0.264, 95% CI 0.113-0.617, P = 0.002). When the CCS (+) patients (n = 49) being regarded as the reference, similar results were obtained in the patients with CCS (-) but B-lines (+) (n = 22) (HR 1.348, 95% CI 0.627-2.896, P = 0.444) as well as in those with both CCS (-) and B-lines (-) (n = 46) (HR 0.447, 95% CI 0.202-0.992, P = 0.048). Conclusions The combination of B-lines on LUS and conventional assessment helped to identify new phenotypes of congestion that aid in the risk stratification of discharged HF patients. Further investigation is warranted to determine whether this strategy could be adopted as a guide for decongestion therapy.