Abstract
BACKGROUND: Pulmonary hypertension (PH) is exhibits a profound pathophysiological association with right heart failure. The non-invasive Remote Dielectric Sensing (ReDS™) technology demonstrates exceptional sensitivity in identifying pathological states, ranging from mild cardiac compensation to severe pulmonary edema. This study aims to develop a PH-ReDS predictive model using ReDS™ technology, leveraging real-world clinical data to evaluate the risk of all-cause mortality or clinical deterioration in patients with PH. METHODS: We conducted a prospective, convenience-sampled observational pilot study involving adult patients with PH admitted to or followed as outpatients in the cardiology department of our institution. ReDS™ Pro was used to measure dielectric sensing values. Clinical data were extracted from the electronic medical record system of our hospital. Patients were followed up for 6 months post-discharge. The primary endpoints defined as all-cause mortality or PH-related clinical deterioration. RESULTS: A total of 202 patients were included, with 24.8% (n = 50) being male. Lung fluid was significantly correlated with left atrial diameter (r = 0.338, p < 0.001), left ventricular ejection fraction (r = -0.256, p < 0.001), and mean right atrial pressure (mRAP; r = 0.219, p = 0.007). Multivariate Cox regression analysis identified lung fluid level, renal insufficiency, and NT-proBNP as significant independent risk factors for PH deterioration. ROC analysis revealed that lung fluid and mRAP effectively discriminated between patients with and without short-term clinical deterioration related to PH, with optimal cut-off values of 30.5% for lung fluid and 6.5 mmHg for mRAP. The Combined PH-ReDS models incorporating lung fluid and mRAP significantly enhanced predictive accuracy, achieving the highest AUC value of 0.733 (95% CI: 0.551-0.916, p = 0.010). Patients with elevated lung fluid levels were found to have a significantly higher risk of short-term clinical deterioration related to PH, with a hazard ratio (HR) of 3.670 (95% CI: 1.274-10.571, p = 0.016). CONCLUSIONS: The PH-ReDS model demonstrated robust predictive efficacy for all-cause mortality or clinical deterioration in patients with PH. Notably, elevated lung fluid levels were significantly associated with an increased risk of short-term PH-related clinical deterioration, suggesting its potential as an early intervention indicator in clinical practice. TRIAL REGISTRATION: This trial was retrospectively registered with Clinical Trials. gov on 2025/04/24. The registration number was NCT06942871 .