Association of SGLT2-inhibitor and GLP1-receptor agonist use with clinical outcomes in aortic stenosis: a large-scale propensity-matched study

Background

Aortic stenosis (AS) currently lacks pharmacological treatment to mitigate disease progression and heart failure (HF) development. Adverse myocardial remodelling in AS, characterised by left-ventricular hypertrophy and diastolic dysfunction, parallels that in HFpEF. Since recent trials have demonstrated that SGLT2 inhibitors (SGLT2i) and GLP-1 receptor agonists (GLP1RA) have clinical benefit in HFpEF patients, we hypothesized these agents may also improve clinical outcomes in patients with AS.

Objectives

To investigate the association between SGLT2i and GLP1RA-use and long-term clinical outcomes in AS patients, stratified by pre- and post-aortic valve replacement (AVR) status.

Methods

Using a global federated research network, we identified AS patients aged >18 without prior HF from 2015 to 2026. Exposure was defined as SGLT2i or GLP1RA prescription within 6 months of AS diagnosis. Patients were further stratified by AVR status: pre-AVR cohort indexed at AS diagnosis and post-AVR cohort indexed at the time of surgical/transcatheter intervention.

1:1 propensity-score matching was performed across 18 baseline characteristics, including demographics, co-morbidities and concomitant medications. Primary outcomes included all-cause mortality, HF hospitalisation, time-to-AVR (surgical/transcatheter) and composite of death or HF - assessed via Kaplan-Meier and Cox models.

Results

27,576 SGLT2i-users and 22,914 GLP1RA-users were equally matched with non-users (mean age 73 years, 39% female). Before matching, SGLT2i-users (31.1+6.8 vs 28.9+6.5 kg/m2) and GLP1RA-users (34.3+7.4 vs 28.7+6.3 kg/m2) exhibited higher BMI than non-users. Similarly, there was a greater prevalence of type 2 diabetes amongst the SGLT2i-users (73.6% vs 17.6%) and GLP1RA-users (76.8% vs 18.5%) compared to non-users.

In the pre-AVR cohort, both SGLT2i (HR 0.73, 95% CI 0.70-0.76, p<0.001) and GLP1RA use (HR 0.80, 95% CI 0.76-0.84, p<0.001) were associated with reduced 10-year mortality. Reflecting this, SGLT2i-use (HR 1.33, 95% CI 1.25-1.43, p<0.001) and GLP1RA-use (HR 2.16, 95% CI 1.98-2.35, p<0.001) were both associated with increased likelihood of undergoing AVR.

In the post-AVR cohort, SGLT2i was associated with a reduced 1-year risk of the composite outcome, death or HF hospitalisation (HR 0.83, 95% CI 0.72-0.95, p=0.008), and mortality alone (HR 0.86, 95% CI 0.80-0.93, p<0.001). Conversely, GLP1RA-use was associated with reduced 1-year mortality (HR 0.48, 95% CI 0.41-0.55, p<0.001) but no significant benefit for composite endpoint (HR 0.92, 95% CI 0.78-1.08, p=0.29).

Conclusions

In aortic stenosis patients, SGLT2i and GLP1RA use was associated with improved outcomes prior to and after aortic valve replacement. SGLT2i use was associated with reduced incidence of HF following AVR. These findings provide a strong rationale for randomised controlled trials to evaluate these agents as potential disease-modifying therapies in valvular heart disease.Pre-AVR and Post-AVR Kaplan Meier Graphs