Association of Statin Use and Lipid Levels with Cerebral Microbleeds and Intracranial Hemorrhage in Patients with Atrial Fibrillation: a Prospective Cohort Study

Background: An increased risk of intracranial hemorrhage (ICH) associated with statins has been reported; but data on the relationship between statin-use and cerebral microbleeds (CMBs) in patients with atrial fibrillation (AF), a population at high bleeding and cardiovascular risk are lacking. Aims: To explore the association between statin-use and blood lipid-levels with the prevalence and progression of CMBs in patients with atrial fibrillation with particular focus on anticoagulated patients. Methods: Data of Swiss-AF, a prospective cohort of patients with established AF were analyzed. Statin-use was assessed during baseline and throughout follow-up. Lipid values were measured at baseline. CMBs were assessed using magnetic resonance imagining (MRI) at baseline and at 2-years follow-up. Imaging data were centrally assessed by blinded investigators. Associations of statin-use and low-density lipoprotein (LDL) levels with CMB prevalence at baseline or CMB progression (at least one additional or new CMB on follow-up MRI at 2-years compared to baseline) were assessed using logistic regression models; the association with ICH was assessed using flexible parametric survival models. Models were adjusted for hypertension, smoking, body mass index, diabetes, stroke/transient ischemic attack, coronary heart disease, antiplatelet use, anticoagulant use and education. Results: Of the 1693 patients with CMB data at baseline MRI (mean±SD age 72.5±8.4y, 27.6% women, 90.1% on oral anticoagulants), 802 patients (47.4%) were statin users. The multivariable adjusted odds ratio (adjOR) for CMBs prevalence at baseline for statin users was 1.10 (95% CI, 0.83–1.45). AdjOR for 1 unit increase in LDL-levels was 0.95 (95% CI, 0.82–1.10). At 2-years, 1188 patients had follow-up MRI. CMBs progression was observed in 44 (8.0%) statin users and 47 (7.4%) non-statin users. Of these patients 64 (70.3%) developed a single new CMB, 14 (15.4%) developed 2 CMBs and 13 developed more than 3 CMBs. The multivariable adjOR for statin users was 1.09 (95% CI, 0.66–1.80). There was no association between LDL-levels and CMBs progression (adjOR 1.02, 95% CI, 0.79–1.32). At follow-up 14 (1.2%) statin users had ICH vs 16 (1.3%) non-users. The age and sex adjusted Hazard Ratio (adjHR) was 0.75 (95% CI, 0.36–1.55). Results remained robust in sensitivity analyses excluding participants without anticoagulants. Conclusions: In this prospective cohort of patients with AF, a population at increased hemorrhagic risk due to anticoagulation, the use of statins was not associated with an increased risk of CMBs.