Introduction: Antiplatelet therapy (APT) is used for prevention of vascular event, but carries higher risk of intracranial hemorrhage (ICH). Cerebral amyloid angiopathy (CAA) and microbleeds, a feature of Alzheimer's disease (AD), also increases the risk of ICH. This study aims to provide a large epidemiologic study that estimates the risk of ICH in patients with AD treated with APT. 
. 

Methods:                                        
We conducted a retrospective population-based cohort study, case-control-matched analysis. The AD cohort and non-AD cohort were selected from the National Health Insurance Research Database (NHIRD) from Taiwan. This study used NHIRD data from 2000-2013.The AD cohort was selected as patients diagnosed with AD or dementia (ICD-9 331.0, 290.0 to 290.3, 294.1, 331.2), who had received prescriptions of acetylcholinesterase inhibitors and memantine, between 2000 to 2012. AD patients with APT were then selected from AD cohort according to the use of APT. APT includes the use of aspirin, clopidogrel ,cilostazol, ticlopidine, aggrenox, and ticagrelor. The definition of APT was patients on continuous APT for at least 3 months. The cohort of AD without APT use were selected by 1:1 frequency match based on index year during 2000-2012. Non-AD groups were extracted from the remaining groups in NHIRD. The cohort of 
non-AD with APT use or without APT use were selected by 1:10 frequency match based on index year during 2000-2012 .To adjust for the potential baseline confounding factors due to imbalance in age, sex, or other covariates in four groups while preserving sample size, we used inverse probability of 
treatment weighting (IPTW) based on the propensity score.Primary outcome was the occurrence of ICH and subarachnoid hemorrhage (ICD-9-CM code 430-432). AD and non-AD groups with or without APT use were followed from index diagnosis of AD or index date of APT use until Dec. 31, 2013. Weighted Cox 
Proportional Hazards regression was used to estimate the risk of ICH in other 3 cohorts compared with non-AD without APT use (reference) group. Subgroup analyses were performed for the risk of ICH in categories defined by age and sex. 

Results: The incidence of ICH was 6.22, 5.30, 2.40 and 0.82 per 1,000 person-years in the AD with APT, AD without APT, non-AD with APT, and the non-AD without APT, with an CRR of 2.41 (95% CI=1.23–4.71), 2.02 (95% CI=1.10–3.72), 2.27 (95% CI=1.67–3.10) in AD with APT, AD without APT, non-AD with APT compared with non-AD without APT group . The risk of ICH was higher in AD patients with APT while age more than 65-year-old (CRR of 3.04 (95% CI=1.28–7.19))and the sex was male (CRR of 3.04 (95% CI=1.28–7.24)). 

Conclusions: Our retrospective, population-based cohort study suggested AD patients treated with APT are associated with higher risk of ICH compared with non-AD patients without APT. The ICH risk was higher in the elderly and male patients.

The risk of intracranial hemorrhage in Alzheimer’s disease patients treated with antiplatelet therapy