Intracerebral Hemorrhage in Patients With CADASIL: Additive Impact of the <i>NOTCH3</i> R544C Variant and Hypertension?

Abstract

Intracerebral hemorrhage (ICH) is increasingly recognized in cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy, especially in Asian patients with the <i>NOTCH3</i> R544C variant. The associations between ICH, <i>NOTCH3</i> variants, and hypertension remain unclear. We enrolled patients from 2 independent cohorts with cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy in France (recruited since 2003) and Taiwan (recruited since 2019) and performed a cross-sectional retrospective analysis. Clinical history and evaluation were collected using standardized questionnaires and scales, while neuroimaging features were assessed with the CADA-MRIT inventory tool. Patients with and without a history of ICH were compared. Logistic regression and mediation analyses were conducted to identify factors associated with ICH. Of 552 patients with cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (440 from France and 112 from Taiwan), 34 (6.2%) had a history of ICH. Patients with ICH were older (62.9±11.4 versus 53.4±12.3 years), had a higher proportion of the <i>NOTCH3</i> R544C variant (79.4% versus 15.3%), and had hypertension (85.3% versus 24.9%). Analysis of magnetic resonance imaging data showed that they had more cerebral microbleeds, worse cerebral atrophy, and higher number of dilated perivascular spaces in basal ganglia. Hypertension (odds ratio, 7.90 [95% CI, 2.83-22.08]) and <i>NOTCH3</i> R544C variant (odds ratio, 9.91 [95% CI, 3.84-25.57]) were each independently associated with ICH, while no multiplicative interaction was detected between these 2 factors (<i>P</i>_interaction=0.81). Having both <i>NOTCH3</i> R544C variant and hypertension carried an additive effect on the risk of ICH (36.9% if both present, 8.8% if having <i>NOTCH3</i> R544C variant without hypertension, 5.4% if having hypertension without <i>NOTCH3</i> R544C variant, and 0.6% if both absent; <i>P</i>_trend<0.001). Finally, in the mediation analysis, 49.9% of the effects of the <i>NOTCH3</i> R544C variant on the occurrence of ICH would be explained by the effects of hypertension. Both the <i>NOTCH3</i> R544C variant and hypertension appear to be independent risk factors for ICH in cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy, with a potential additive effect. However, due to the study's cross-sectional design and population-specific factors, causality cannot be established. Prospective studies are thus needed to validate these findings and clarify underlying mechanisms.