Assessing pulmonary congestion in patients hospitalized with decompensated chronic heart failure according to lung ultrasound and remote dielectric sensing (ReDS)

Aim. To conduct a comparative assessment of parameters and dynamics of pulmonary congestion according to lung ultrasound and remote dielectric sensing (ReDS) in patients hospitalized with decompensated chronic heart failure (CHF) Materials and methods. The pilot single-center study included patients hospitalized with decompensated CHF. Lung ultrasound and ReDS were simultaneously performed within 24 hours from the moment of hospitalization and at discharge. Eight-zone lung ultrasound was performed with the calculation of the sum of B-lines. Pulmonary congestion was confirmed with the sum of B-lines >= 5. ReDS was performed according to the manufacturer's protocol. Congestion was confirmed at the value of more than 35%. To determine ReDS interoperator variability, each patient was examined by two operators who were blind to each other's findings with a 20-30-minute interval. Results. Thirty-five patients were included in the study: 40% (n = 14) men, the average age was 71 (65.5; 78.5) years, the median NT-proBNP was 1,379 (470; 4,277) pg / l. Hydrothorax at admission was observed in 31,4% (n = 11) of patients. The incidence of pulmonary congestion according to lung ultrasound was 57.1% (n = 20): 31.4% (n = 11) of patients had mild congestion, 22.9% (n = 8) - moderate, and 2.9% (n = 1) - severe congestion. ReDS data revealed pulmonary congestion in 62.9% (n = 22) of cases, of which 37,1% (n = 13) of cases were characterized by mild, 22.9% (n = 8) - by moderate, and 2.9% (n = 1) - by severe congestion. A moderate correlation was found between ReDS (%) and lung ultrasound (sum of B-lines) findings at admission (Spearman's rank correlation coefficient = 0.402; p = 0.017). No correlation between the two methods was found at discharge (p = 0.613). The frequency of agreement between lung ultrasound and ReDS on signs of congestion at admission was 77.1% (p = 0.004) with an average Cohen's Kappa coefficient (kappa = 0.53). The average interoperator variability in ReDS was 9.9%. Conclusion. A moderate correlation was revealed between ReDS (%) and lung ultrasound (sum of B-lines) in detecting pulmonary congestion (Spearman's rank correlation coefficient = 0.402; p = 0.017). No correlation between the two methods was found at discharge (p = 0.613).