Eike Nagel, Philip Wenzel, Tim Seidler, Karl-Philipp Rommel, Florian André, Sebastian Kelle, Andreas Rolf, Mariuca Vasa-Nicotera, Eva Herrmann, Felicitas Escher, Jedrzej Hoffmann, Frank Rademakers, Philipp Lurz, Valentina O Puntmann, Decipher HFpEF Study Group
In invasively characterized HFpEF, CMR tissue mapping captures active relaxation impairment and passive stiffness conditional on filling pressure, robust to clinical covariate adjustment. The exploratory 2-phenotype partition is hypothesis-generating; prospective external validation, including treatment-response data, is required before clinical stratification. (Validation of CMR Against Invasive Heamodynamics in Patients With HFpEF [DECIPHER HFpEF]; NCT03251183).
BACKGROUND: Diastolic dysfunction (DD) in heart failure with preserved ejection fraction (HFpEF) reflects active relaxation (τ) and passive stiffness (β), conflated by end-diastolic pressure (EDP) and potentially requiring different therapies. Whether cardiac magnetic resonance (CMR) tissue mapping captures these beyond filling pressure is unknown.
OBJECTIVES: This study aims to determine whether CMR tissue mapping (native T1 corrected for field strength and scanner [T1c], native T2 corrected for field strength and scanner [T2c]) captures τ and β as markers of mechanistic severity in HFpEF beyond EDP.
METHODS: In this prospective multicenter study, patients with suspected HFpEF underwent CMR with sequence-corrected tissue mapping and invasive pressure-volume loop analysis. HFpEF was defined as EDP ≥16 mm Hg (n = 52). Age- and sex-matched control patients (n = 32) and healthy volunteers (n = 18) provided cross-group context. Eighteen catheterized symptomatic patients with normal resting EDP underwent handgrip provocation for a continuum analysis. Partial correlations adjusted T1c-β and T2c-τ associations for EDP.
RESULTS: T1c and T2c differed between HFpEF and matched control patients (Cliff's δ 0.74 and 0.80; both P < 0.001). Within HFpEF, T1c correlated with β (partial ρ = 0.47; P = 0.003) and T2c with τ (ρ = 0.35; P = 0.018), both conditional on EDP. Exploratory partitioning around medoids clustering on τ and β identified 2 phenotypes (mild DD [n = 24], severe DD [n = 28]; adjusted Rand index: 0.78) that separated T1c and T2c (both P < 0.001), whereas an EDP median split did not, consistent with tissue mapping reflecting mechanistic severity beyond filling pressure.
CONCLUSIONS: In invasively characterized HFpEF, CMR tissue mapping captures active relaxation impairment and passive stiffness conditional on filling pressure, robust to clinical covariate adjustment. The exploratory 2-phenotype partition is hypothesis-generating; prospective external validation, including treatment-response data, is required before clinical stratification. (Validation of CMR Against Invasive Heamodynamics in Patients With HFpEF [DECIPHER HFpEF]; NCT03251183).