Cerebral Amyloid Angiopathy and Risk of Isolated Nontraumatic Subdural Hemorrhage
Cyprien A Rivier,Hooman Kamel,Kevin N Sheth,Costantino Iadecola,Ajay Gupta,Mony J de Leon,Elizabeth Ross,Guido J Falcone,Santosh B Murthy
DOI: https://doi.org/10.1001/jamaneurol.2023.4918
2024-02-01
Abstract:Importance: Cerebral amyloid angiopathy (CAA) is a common cause of spontaneous intracerebral hemorrhage in older patients. Although other types of intracranial hemorrhage can occur in conjunction with CAA-related intracerebral hemorrhage, the association between CAA and other subtypes of intracranial hemorrhage, particularly in the absence of intracerebral hemorrhage, remains poorly understood. Objective: To determine whether CAA is an independent risk factor for isolated nontraumatic subdural hemorrhage (SDH). Design, setting, and participants: A population-based cohort study was performed using a 2-stage analysis of prospectively collected data in the UK Biobank cohort (discovery phase, 2006-2022) and the All of Us Research Program cohort (replication phase, 2018-2022). Participants included those who contributed at least 1 year of data while they were older than 50 years, in accordance with the diagnostic criteria for CAA. Participants with prevalent intracranial hemorrhage were excluded. Data were analyzed from October 2022 to October 2023. Exposure: A diagnosis of CAA, identified using the International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) diagnosis code. Main outcomes and measures: The outcome was an isolated nontraumatic SDH, identified using ICD-10-CM codes. Two identical analyses were performed separately in the 2 cohorts. First, the risk of SDH in patients with and without CAA was assessed using Cox proportional hazards models, adjusting for demographic characteristics, cardiovascular comorbidities, and antithrombotic medication use. Second, multivariable logistic regression was used to study the association between CAA and SDH. Results: The final analytical sample comprised 487 223 of the total 502 480 individuals in the UK Biobank cohort and 158 008 of the total 372 082 individuals in the All of Us cohort. Among the 487 223 participants in the discovery phase of the UK Biobank, the mean (SD) age was 56.5 (8.1) years, and 264 195 (54.2%) were female. There were 649 cases of incident SDH. Of the 126 participants diagnosed with CAA, 3 (2.4%) developed SDH. In adjusted Cox regression analyses, participants with CAA had an increased risk of having an SDH compared with those without CAA (hazard ratio [HR], 8.0; 95% CI, 2.6-24.8). Multivariable logistic regression analysis yielded higher odds of SDH among participants with CAA (odds ratio [OR], 7.6; 95% CI, 1.8-20.4). Among the 158 008 participants in the All of Us cohort, the mean (SD) age was 63.0 (9.5) years, and 89 639 (56.7%) were female. The findings were replicated in All of Us, in which 52 participants had CAA and 320 had an SDH. All of Us participants with CAA had an increased risk of having an SDH compared with those without CAA (HR, 4.9; 95% CI, 1.2-19.8). In adjusted multivariable logistic regression analysis, CAA was associated with higher odds of SDH (OR, 5.2; 95% CI, 0.8-17.6). Conclusions and relevance: In 2 large, heterogeneous cohorts, CAA was associated with increased risk of SDH. These findings suggest that CAA may be a novel risk factor for isolated nontraumatic SDH.