AHA Scientific Sessions 2026 Abstract · Nov 2, 2026
Baseline Lipoprotein(a) and Risk of Incident Heart Failure with Preserved Ejection Fraction
By Bawahab Y, Patel H, Vasile VC
Background: Lipoprotein(a) [Lp(a)] is an independent cardiovascular risk factor, but its relationship with incident heart failure with preserved ejection fraction (HFpEF) remains unclear. We examined whether baseline Lp(a) is independently and linearly associated with incident HFpEF in a large cohort. Methods: We conducted a retrospective cohort study of 59,985 patients without prevalent heart failure who underwent Lp(a) testing at Mayo Clinic (2006-2021). The index date was the first valid Lp(a) measurement. Incident HFpEF required a diastolic heart failure ICD code plus echocardiographic LVEF ≥50% within ±30 days of testing. Multivariable Cox proportional hazards models, stratified by sex and Lp(a) testing era, were adjusted for age, race, hypertension, type 2 diabetes, atrial fibrillation, chronic kidney disease, coronary artery disease, obesity, statins, PCSK9 inhibitors, and SGLT2 inhibitors. Proportional hazards assumptions were assessed using Schoenfeld residuals. Restricted cubic spline analysis assessed the functional form of the Lp(a)-HFpEF relationship. Results: Among 59,985 patients (mean age 55.3±12.9 years, 56.7% male, 87.9% White; median Lp(a) 14 mg/dL [IQR 6-42]; 21.8% with Lp(a) ≥50 mg/dL), 1,614 (2.69%) developed incident HFpEF over a median follow-up of 6.87 years (IQR 1.48-13.8). Incidence rates were 3.24 and 3.97 per 1,000 person-years for Lp(a) <50 versus ≥50 mg/dL, respectively. Lp(a) was associated with HFpEF in unadjusted analysis (per-standard deviation [SD] hazard ratio [HR] 1.13, p<0.001). After multivariable adjustment, the association was attenuated but persisted (perSD HR 1.06, 95% CI 1.01-1.11, p=0.010), with similar estimates after excluding PCSK9 inhibitor and SGLT2 inhibitor users (HR 1.06, p=0.013). Restricted cubic spline analysis supported an approximately linear dose-response across the Lp(a) distribution (p for non-linearity = 0.94). Conclusions: In a large cohort with nearly 7-year median follow-up, higher Lp(a) levels showed a linear, independent association with incident HFpEF after adjustment for traditional cardiovascular risk factors. These findings support Lp(a) as a novel continuous risk marker for HFpEF with potential implications for cardiovascular risk stratification.