STOP-DHF (a non-invasive hemodynamic-guided strategy to prevent heart failure decompensation) trial: determinants of left ventricular filling pressure elevation
A Bohm, J Lucka, A Segev, J Jankova, V Hlodakova, J Stevkova, M Kollarova, N Jajcay, L Jurasko Jancovic, O Holly, E Pogran, M Spilak, F Skorec, S Karolcik, B BezakAbstract
Background
Elevated left ventricular filling pressures (LVFP) typically precede clinical heart failure (HF) decompensation by days to weeks. Non-invasive remote hemodynamic monitoring enables early detection of rising LVFP, creating an opportunity for timely intervention before overt clinical deterioration occurs.
Purpose
To identify clinical determinants associated with LVFP elevation detected during remote monitoring in the STOP-DHF trial.
Methods
STOP-DHF was a prospective, multicenter, single-arm pragmatic trial enrolling 330 patients with chronic HF (NYHA class II–III), irrespective of left ventricular ejection fraction (LVEF). Non-invasive LVFP was assessed using a CE-certified photoplethysmography-based system. The primary composite event rate of HF hospitalization and all-cause mortality was 4.2% (95% CI 2.3–7.0%), significantly lower than the predefined performance goal of 9% (p < 0.001). Following predefined hemodynamic alerts, structured telephone assessments complemented by medical chart review were performed to evaluate clinical status and identify potential factors associated with LVFP elevation. Precipitating determinants were categorized, and their relative frequencies were analyzed.
Results
Over a mean follow-up of 6 months, 606 structured medical assessments triggered by LVFP alerts were conducted in 235 patients (mean age 66.6 ± 12.3 years; 36% women; 42% LVEF ≤40%; NYHA II 65%, NYHA III 35%). Thirteen alerts were excluded due to measurement errors, leaving 593 assessments for analysis. An identifiable precipitating factor was present in 300 (51%) cases. Infection was the most frequent determinant (91 cases, 30%), predominantly respiratory (78; 86%). Other contributors included medication nonadherence or recent dose reduction (52; 17%), arrhythmias or conduction disturbances (42; 14%), drug-related factors including NSAID or cardiotoxic medication exposure (39; 13%), and hypertensive episodes (36; 12%). Less common determinants were dietary nonadherence (12; 4%), anemia (8; 3%), myocardial ischemia (3; 1%), and other causes (17; 6%). The distribution of determinants did not differ significantly according to LVEF phenotype (global p = 0.25).
Conclusion
In this telemonitoring HF cohort, a precipitating clinical factor was identified in approximately half of LVFP alert-triggered assessments, most commonly infection, medication-related factors, and arrhythmias. The similar distribution across HF phenotypes highlights the importance of universal preventive strategies. Non-invasive remote monitoring may enable early recognition of these triggers and facilitate targeted ambulatory interventions to prevent clinical deterioration.