Effects of Semaglutide in Obesity-Related Heart Failure with Preserved Ejection Fraction Across the Age Spectrum: Findings from the STEP-HFpEF Programme

The inclusion and exclusion criteria for both STEP‐HFpEF trials have been reported previously.5, 6 Briefly, the trial enrolled participants with symptomatic HF (left ventricular ejection fraction ≥45%) and obesity (body mass index [BMI] ≥30 kg/m²) with objective evidence of HFpEF based on elevated filling pressures or elevated natriuretic peptide levels and echocardiographic abnormalities, or HF hospitalization in the previous 12 months. Only patients with type 2 diabetes mellitus (T2DM) were included in the STEP‐HFpEF DM trial, in which semaglutide or placebo was added to background therapy for T2DM. Eligible participants were randomized 1:1 to receive once‐weekly semaglutide 2.4 mg or matching placebo in addition to standard care for 52 weeks.

The dual primary endpoints were: (1) change in Kansas City Cardiomyopathy Questionnaire clinical summary score (KCCQ‐CSS) and percentage change in body weight from baseline to 52 weeks. The confirmatory secondary endpoints were changes in 6‐min walk distance (6MWD) and C‐reactive protein (CRP), and a hierarchical composite endpoint (comprised of all‐cause death [from baseline to 57 weeks]). HF events (hospitalizations or urgent visits requiring intravenous therapy; an exploratory endpoint in both STEP‐HFpEF trials) were adjudicated by a blinded clinical events committee as previously described.5, 6 Change in N‐terminal pro‐B‐type natriuretic peptide (NT‐proBNP) concentrations from baseline to week 52 was also evaluated as an exploratory endpoint. Serious adverse events (SAEs), SAEs leading to permanent treatment discontinuation, and cardiac and gastrointestinal SAEs were also evaluated as safety endpoints. Finally, atrial fibrillation events and bone/joint injuries were also assessed as safety endpoints of interest for this analysis.

The study participants were stratified according to age into four categories: (i) <55 years, (ii) 55–64 years, (iii) 65–74 years, and (iv) ≥75 years. Baseline characteristics were compared across age strata using the Wilcoxon test and the Chi‐square test for continuous and categorical variables, respectively. The efficacy endpoints were examined according to the intention‐to‐treat principle. Analyses of continuous endpoints were performed using analysis of covariance models adjusted for the baseline value of the endpoint, treatment arm, trial, and BMI (<35 kg/m² or ≥35 kg/m²) as fixed factors using 1000 imputations; analyses also included an interaction term between treatment arm and age category. Estimates were combined using Rubin's rule. For outcomes of change in KCCQ‐CSS and 6MWD, missing observations at week 52 caused by cardiovascular death or previous HF events (if not observed) were single‐imputed using the lowest observed value across both treatment arms and visits. Missing values caused by other reasons were imputed from retrieved participants in the same randomized treatment arm. For other endpoints, missing observations at week 52 were imputed irrespective of death or prior HF events using the same imputation method. Interaction p‐values were derived from an F‐test of equality between the treatment differences across the age subgroups.

Analyses of the hierarchical composite endpoint (win ratio) were performed stratified by age, based on direct comparisons of each participant in semaglutide versus placebo arms, as reported previously.5, 6 A two‐sided p‐value of <0.05 was considered statistically significant and no adjustment for multiplicity was performed. Results are presented as estimated changes from baseline to 52 weeks for continuous endpoints and a win ratio for the hierarchical composite endpoint with 95% confidence intervals (CI) and two‐sided p‐values. NT‐proBNP and CRP were log‐transformed; hence, the treatment ratio with the corresponding 95% CI is reported. Statistical analyses were performed using SAS version 9.4 (SAS/STAT version 15.1).

The median age of participants was 69 years (interquartile range: 62–75). Participants in the older age subgroups (65–74 years and ≥75 years) had lower body weight but a higher prevalence of cardiovascular comorbidities than younger participants. Older participants had higher NT‐proBNP, while CRP was highest among younger participants. Older participants had lower 6MWD and had a larger proportion with New York Heart Association class III symptoms than younger participants. There were no differences in KCCQ‐CSS scores by age category at baseline. No differences were noted in the use of sodium‐glucose co‐transporter‐2 inhibitors and mineralocorticoid receptor antagonists by age category (Table 1).

The effect of semaglutide on the dual primary endpoint of KCCQ‐CSS and body weight was consistent across age strata without any significant heterogeneity in the treatment effect by age (Figure 1, Table 2). For body weight, a significant reduction ranging from −7.7% to −9.2% was observed across all age groups (p‐interaction = 0.41). Similarly, semaglutide use resulted in 5.6‐point to 8.4‐point improvement of KCCQ‐CSS across the age groups (p‐interaction = 0.80). Among confirmatory secondary endpoints, semaglutide led to an increase in 6MWD (p‐interaction = 0.96) and a favourable win ratio for the hierarchical composite endpoint (p‐interaction = 0.85) in each age category without significant heterogeneity in treatment effect. Semaglutide was also associated with consistent reductions in CRP and NT‐proBNP across different age groups (p‐interaction: CRP = 0.20, NT‐proBNP = 0.78). In terms of safety, the proportion of participants with SAEs was higher in the older age groups, with a higher frequency noted in the placebo group versus the semaglutide arm (Table 3). Frequency of investigator‐reported atrial fibrillation events and bone/joint injuries on follow‐up was low, with no difference between the semaglutide and placebo arms across age strata (Table 3).