ABSTRACT NUMBER: ESOC2026A1463 STATIN THERAPY AND STROKE RISK IN PATIENTS WITH HYPERCHOLESTEROLAEMIA: A POPULATION-BASED LONGITUDINAL STUDY USING ELECTRONIC HEALTH RECORDSIN SOUTH LONDON

Abstract Background and aims We aim to evaluate the impact of statin prescription on stroke risk in patients with hypercholesterolaemia and to assess disparities in statin prescribing. Methods We analysed electronic health records from patients with hypercholesterolaemia, registered in 41 general practices in south London between 2005 and 2021. The causespecific hazard ratio of statin prescription on stroke, adjusted for patients’ sociodemographic characteristics and stroke risk factors (smoking ever, hypertension, and diabetes), was estimated using a time-varying exposure Cox proportional hazards model stratified by history of heart diseases. The association between statin prescription and patients’ sociodemographic characteristics was evaluated using a logistic regression. Results Of the 849,968 registered patients, 166,124 (19.5%) had records of hypercholesterolaemia. Among them, 33.5% were prescribed statins, 2.6% had a record of stroke, and 50.6% were female, 31.7%, 16.2% and 8.9% had records of hypertension, diabetes and history of heart diseases respectively. In a Cox model stratified by history of heart diseases, statin prescription was associated with a reduced hazard of stroke (cause-specific hazard ratio: 0.74; 95% confidence interval (CI): 0.68–0.80, p<0.001), with follow-up administratively censored at 79 years (n=161,527; 97.2%). Statins were less likely prescribed to female patients and patients of Black ethnicity (odds-ratio: 0.70, 95% CI: 0.68-0.72, p<0.001 and odds-ratio: 0.82, 95% CI: 0.79-0.85, p<0.001, respectively, see attached figure). Conclusions Statin therapy prescription is associated with reduced stroke risk in patients with hypercholesterolaemia, yet it was under-prescribed to women and patients of Black ethnicity, highlighting avoidable disparities in preventive care. Conflict of interest Marc Delord and Abdel Douiri: nothing to disclose Table 1 - belongs to Results