Refining Atrial Shunt Therapy for Heart Failure With Preserved Ejection Fraction: Lessons From Pivotal Randomized Trials and Emerging Directions.

Heart failure (HF) with preserved ejection fraction (HFpEF) is a challenging and heterogeneous condition with limited effective treatments. Elevated left atrial pressure is a key driver of HF symptoms and complications, making it a compelling therapeutic target. Interatrial shunts have been proposed as a novel strategy to reduce left atrial pressure and improve outcomes in patients with HF. Two major randomized controlled trials, REDUCE LAP-HF II (Reduce Elevated Left Atrial Pressure in Patients With Heart Failure) and RELIEVE-HF (Reducing Lung Congestion Symptoms in Advanced Heart Failure), investigated the efficacy of atrial shunts and yielded seemingly contradictory results. REDUCE LAP-HF II, which focused on patients with HF and ejection fraction (EF) >40%, identified a responder subgroup-patients without latent pulmonary vascular disease or cardiac rhythm management devices-who experienced significant clinical benefits and favorable cardiac remodeling. Nonresponders, however, demonstrated worse outcomes. RELIEVE-HF, which included patients across a broad range of EFs, reported worse outcomes in patients with HF and EF >40% treated with shunts. Notably, these patients shared key characteristics with REDUCE LAP-HF II nonresponders, including more advanced echocardiographic abnormalities, elevated resting pulmonary vascular resistance, and increased markers of HF severity. Here, we compare the design and outcomes of both trials to explore the possibility of underlying concordance in their findings. By highlighting the importance of detailed phenotyping, including invasive exercise hemodynamics, we demonstrate the potential benefit of interatrial shunts for carefully selected patients with HFpEF and provide supportive evidence for ongoing trials like RESPONDER-HF using refined patient selection to optimize outcomes for atrial shunts in HFpEF and HF with mildly reduced EF.