ReDS to detect pulmonary congestion in the emergency department: reproducibility and effect of pulmonary comorbidities

Abstract

Abstract Funding Acknowledgements Type of funding sources: Public hospital(s). Main funding source(s): This work was supported by the research fund of Bispebjerg University Hospital and Helsefonden. Background Remote Dielectric Sensing (ReDS) is a fast, non-invasive bedside method that estimates lung fluid content observer-independently. We previously described that ReDS has a moderate accuracy to detect acute heart failure in consecutive patients, and we hypothesise that unprecise ReDS estimates may arise from concomitant pulmonary diseases. Still, it is unknown how common comorbidities influence estimates of pulmonary congestion when using ReDS in the emergency department (ED). Purpose To examine the reproducibility of ReDS and the effect of pulmonary comorbidities in dyspnoeic patients in the ED. Method This prospective observational study included consecutive ED patients ≥50 years with acute dyspnoea. Upon admission, all patients were examined with a low-dose chest computed tomography (CT), an echocardiography and a ReDS examination. The default ReDS examination is performed with the ReDS device placed on the right hemithorax in sitting position. However, for reproducibility comparisons, we conducted three ReDS measurements on the right hemithorax after small adjustments, and in supine position (Picture 1: Scenario 1 and 2). Two blinded radiologists evaluated the CT scans for pulmonary congestion and other common pulmonary diseases. Results 97 patients were included and had study examinations conducted within a median of 4.8 hours from admission. Reproducibility comparisons revealed coefficients of variations (CV) of 9.6%, 8.2%, and 8.3% for three ReDS measurements on the right hemithorax (Picture 1: Scenario 1). Similarly, CV was 9.5% for sitting versus supine positions (Picture 1: Scenario 2). For scenario 2, patients with CT verified pulmonary congestion had CV of 5.9%, while those without pulmonary congestion had CV of 10.3%. In multivariate linear regression, we examined the relationship between pulmonary findings on CT and ReDS as a response variable (overall mean ReDS=31.6%, p<0.001). Lower ReDS estimates were observed in patients with pneumonia (-1.81, p=0.215, N=51) and emphysema (-5.44, p=0.001, N=26). Higher ReDS estimates were observed in fibrosis (5.58, p=0.032, N=8) and pulmonary congestion (5.79, p=0.002, N=17). Conclusion ReDS showed good reproducibility for patients with CT verified pulmonary congestion. However, the reproducibility was only moderate for consecutive patients with dyspnoea, who had a high prevalence of CT verified emphysema, lowering the ReDS estimates.Reproducibility comparisons