Myocardial histopathologic phenogroups and differential prognostic implications in heart failure with preserved versus reduced ejection fraction
G Tersalvi, A U Fayyaz, S Dasari, S O Arab, K E Koepp, Y Wang, M C Bois, O F AbouEzzeddine, J J Maleszewski, M M RedfieldAbstract
Background
Myocardial remodeling in heart failure (HF) involves varying degrees of cardiomyocyte hypertrophy and interstitial fibrosis. We hypothesized that phenogroups defined by these two features (histogroups) would be associated with unique clinical profiles and outcomes in HF with preserved (HFpEF) or reduced ejection fraction (HFrEF).
Methods
We analyzed consecutive HF patients who underwent right ventricular endomyocardial biopsy (EMB) between 1999 and 2023 at our center. Patients with previous cardiac transplant, infiltrative, inflammatory or hypertrophic cardiomyopathy, primary valvular disease, constrictive pericarditis, isolated right HF, or inadequate samples were excluded. Based on index echocardiography, patients were classified as HFpEF (EF ≥50%) or HFrEF (EF <50%). Clinical, echocardiographic, and invasive data were collected. H&E-stained sections were retrospectively reviewed by an experienced cardiac pathologist blinded to EF and clinical data. Fibrosis was graded semiquantitatively (none, mild, moderate, marked), and cardiomyocyte diameter (CMD) was measured (x 5) at the nucleus, averaged, and rounded to 10 µm. Four histogroups were defined: “mild” (≤mild fibrosis and CMD<30 µm), “hypertrophic” (≤mild fibrosis and CMD≥30 µm), “fibrotic” (≥moderate fibrosis and CMD<30 µm), and “severe” (≥moderate fibrosis and CMD≥30 µm). The primary endpoint was all-cause mortality. Kaplan–Meier and Cox proportional hazards models were performed separately in HFpEF and HFrEF, with follow-up administratively censored at the group-specific median.
Results
In HFpEF (n=124), histogroup distribution was: “mild” n=37 (30%), “hypertrophic” n=52 (42%), “fibrotic” n=4 (3%), and “severe” n=31 (25%). The “severe” group more frequently comprised men and patients with diabetes; structural remodeling parameters worsened across groups, including higher LV mass, more advanced left atrial remodeling, and higher NT-proBNP levels (all p<0.05). In HFrEF (n=170), distribution was: “mild” n=29 (17%), “hypertrophic” n=66 (39%), “fibrotic” n=7 (4%), and “severe” n=68 (40%). In contrast with HFpEF, histogroups in HFrEF showed similar demographics, comorbidities, imaging, hemodynamics, and biomarkers (all p=NS). The “fibrotic” histogroup was excluded from survival analyses due to small sample size (HFpEF n=4; HFrEF n=7). In HFpEF, survival differed across histogroups (log-rank p=0.006, Figure 1), with the “severe” histogroup showing the worst outcomes after adjustment for age and sex (HR 2.7) and age, sex, and eGFR (HR 2.0). In HFrEF, histogroup was not associated with outcomes (Figure 2), without or with adjustment for age, sex or eGFR.
Conclusions
Integrated fibrosis and hypertrophy assessment identifies prognostically distinct phenogroups in HFpEF, whereas no independent association was observed in HFrEF. These findings suggest that structural myocardial heterogeneity carries differential prognostic relevance across HF phenotypes.Figure 1Figure 2