De Schutter S; GENCOR Department University of Antwerp, Belgium.
&other centres in Belgiujm
Pugliese NR; Meems LMG; Moura-Ferreira S; et al

Journal of the American Heart Association. 15(14):e050524, 2026 Jul 21.

BACKGROUND: Heart failure with preserved ejection fraction (HFpEF) is a
heterogeneous disease characterized by exercise intolerance. Defining
pathophysiologically distinct subgroups allows more personalized therapy,
but efforts mainly relied on resting examinations. This study aimed to
define HFpEF phenotypes based on exercise limitations using combined
cardiopulmonary exercise testing with stress echocardiography.

METHODS: A total of 913 patients with HFpEF were recruited from 4
third-line centers and divided into derivation (n=623) and validation
cohorts (n=290). Unsupervised graph-based clustering of 61 cardiopulmonary
exercise testing with stress echocardiography variables was used to
identify HFpEF exercise phenotypes. Pathophysiological characteristics,
exercise capacity, and clinical outcomes were compared between phenotypes.

RESULTS: In the derivation cohort, cluster analysis identified 5 distinct
HFpEF exercise phenotypes characterized by specific exercise responses:
mild diastolic dysfunction (phenotype 1), impaired peripheral oxygen
extraction (phenotype 2), right ventricular-pulmonary artery uncoupling
(phenotype 3), reduced left ventricular systolic reserve (phenotype 4),
and chronotropic incompetence (phenotype 5). The composite outcome of
all-cause death and unplanned cardiovascular hospitalization differed
significantly across phenotypes, with phenotypes 2 (hazard ratio [HR],
1.76 [95% CI, 1.07-2.91]), 4 (HR, 2.15 [95% CI, 1.27-3.65]), and 5 (HR,
2.19 [95% CI, 1.33-3.61]) showing higher rates of the primary combined
outcome compared with phenotype 1. All phenotypes were replicated in the
validation cohort.

CONCLUSIONS: Deep phenotyping of the exercise response in patients with
HFpEF revealed 5 distinct phenogroups with marked differences in
pathophysiology, exercise performance, and clinical outcomes. This
subclassification may support more personalized therapeutic strategies and
improve risk stratification in HFpEF.