Diagnostic performance and 3-month prognostic impact of multimodal assessment of residual congestion at discharge in acute heart failure

Background

Residual congestion at discharge is common in patients hospitalized for heart failure (HF) and is strongly associated with early post-discharge adverse outcomes. Despite its clinical relevance, detection of residual congestion remains an unmet need, as several clinical, device-based, and biomarker tools are available, yet it is unclear which approach best identifies residual congestion and provides prognostic value.

Purpose

The aim of this study was: (1) to evaluate the diagnostic performance of bedside ultrasound, including lung ultrasound (LUS) and venous excess ultrasound (VExUS), remote dielectric sensing (ReDS), bioimpedance, and NT-proBNP for the detection of residual congestion at hospital discharge compared with clinical assessment; and (2) to assess the association between residual congestion detected by each method and 3-month adverse outcomes in patients hospitalized for HF.

Methods

This was a prospective, single-centre study including consecutive patients hospitalized for HF, either with de novo HF or decompensated chronic HF. At discharge, congestion was assessed using ultrasound (LUS and VExUS), clinical scores [Composite Congestive Score (CCS) and EVEREST score], ReDS, bioimpedance, and NT-proBNP. The results of ultrasound, ReDS, and bioimpedance assessments were blinded to the treating physicians. Residual congestion at discharge was defined for each modality. The primary endpoint was a composite of all-cause mortality, HF hospitalization, and urgent HF visits requiring intravenous diuretics at 3 months (cEP3).

Results

A total of 233 patients were included (mean age 70.6 years, 64.4% men, 61.4% with de novo HF, mean left ventricular ejection fraction 42.6%, and mean discharge glomerular filtration rate 55.9 mL/min). At 3 months, cEP3 occurred in 24.9% of patients (n=58).

Residual congestion at discharge was detected in 46.4% of patients by ultrasound (LUS and VExUS), compared with 9.0% by CCS, 20.6% by EVEREST, 31.7% by ReDS, 78.8% by bioimpedance, and 47.7% by NT-proBNP. Concordance among methods was low (Cohen’s kappa <0.6). Ultrasound showed the highest sensitivity (79%) and negative predictive value (90%) for predicting cEP3 (Table 1). The highest discriminative ability was also observed with ultrasound (AUC 0.72, 95% CI 0.66–0.78) (Table 2). Residual congestion detected by ultrasound was associated with the strongest prognostic impact (OR 6.99, 95% CI 3.54–14.72; p<0.0001) (Table 3).

Conclusion

Residual congestion at discharge is common in patients hospitalized for HF and remains a major determinant of early post-discharge adverse outcomes. Ultrasound-based assessment (LUS and VExUS) outperforms clinical, other device-based, and biomarker approaches for detecting residual congestion and provides superior prognostic stratification at 3 months. These findings support the integration of ultrasound-based congestion assessment at discharge into routine clinical practice.