Lp(a) above 125 nmol/L predicts ischaemic stroke in men but not women, and equally across age and ethnicity, UK Biobank analysis of 353,309 finds (Am J Prev Cardiol 2025)

Original title: Demographics of lipoprotein(a) and stroke: The UK biobank

Am J Prev Cardiol · 29 June 2025 · 8

Kao AS, Kermanshahchi J, Khosrowjerdi S, Razavi AC, Levene J, Reyes M, Garg P, Longstreth WT, Rannikmae K, Sudlow C, Tsai M, Tsimikas S et al.

This UK Biobank analysis included 353,309 participants with an Lp(a) measurement and no prior atherosclerotic cardiovascular disease (55.7% women, 94% White, average age 56), examining whether Lp(a) above 125 nmol/L (present in 11.1%, n = 39,067) predicts ischaemic and haemorrhagic stroke, and whether this varies by age, sex or race/ethnicity. Over a median 13.8-year follow-up, there were 5,002 (1.4%) ischaemic strokes and 1,462 (0.4%) haemorrhagic strokes. Lp(a) above 125 nmol/L was associated with higher ischaemic stroke risk (HR 1.12, 95% CI 1.02-1.22, P = 0.019) but not haemorrhagic stroke (HR 0.95, 95% CI 0.79-1.13, P = 0.545). The ischaemic stroke association did not vary by age or race/ethnicity, but did vary by sex (P for interaction = 0.041), present in men (HR 1.20, 95% CI 1.07-1.36) but not women. This large biobank study confirms Lp(a) as an independent ischaemic stroke risk factor with a notable sex-specific pattern that current guidelines do not yet reflect.

Original abstract

Introduction: Lipoprotein(a) [Lp(a)] is associated with ischemic stroke, but the strength of association based on demographic differences remains unclear. We aimed to investigate the association between Lp(a)>125 nmol/L and stroke by age, sex, and racial/ethnic subgroups.

Methods: Using data from the UK Biobank, we included 353,309 participants with an Lp(a) measurement without a history of atherosclerotic cardiovascular disease (ASCVD). Stroke was defined as ischemic stroke or hemorrhagic stroke (with subtypes subarachnoid hemorrhage and intracerebral hemorrhage). Cox proportional hazards models were used to evaluate the association between elevated Lp(a) and stroke (ischemic or hemorrhagic), adjusted for ASCVD risk factors, race/ethnicity, lipid lowering therapy, antiplatelet and anticoagulation medications. Outcomes were defined using ICD-10 codes.

Results: The study population consisted of 55.7 % women and 94 % White individuals with average age of 56 years. The median Lp(a) level was 20.9 [IQR 9.5, 61.4] nmol/L and prevalence of Lp(a) >125 nmol/L was 11.1 % (n = 39,067). Over a median follow-up of 13.8 [13.1, 14.5] years, there were 5002 (1.4 %) ischemic strokes and 1462 (0.4 %) hemorrhagic strokes. Lp(a) > 125 nmol/L was associated with increased risk for ischemic stroke (HR 1.12, 95 % CI 1.02-1.22, P = 0.019), but not hemorrhagic stroke (HR 0.95, 95 % CI 0.79-1.13, P = 0.545). The association between Lp(a)>125 nmol/L and ischemic stroke did not vary by age (p-interaction=0.691) or race/ethnicity (p-interaction 0.526) but did vary by sex (p-interaction=0.041) with an association among men (HR 1.20, 95 % CI 1.07-1.36) but not among women.

Conclusion: Lp(a) is independently associated with ischemic stroke, with variation by sex but not age or race/ethnicity. Lp(a) was not significantly associated with hemorrhagic stroke.

Summary written by lp-a.org from the published abstract; figures as published. Page updated 17 August 2026. Methods.