Back

Corticosteroid Treatment and Hemorrhage Risk in Cerebral Amyloid Angiopathy with Cortical Superficial Siderosis: A Matched Cohort Analysis

Arndt, P.; Khadhraoui, E.; Mueller, S. J.; Neumann, K.; Mattern, H.; Loeding, S.-M.; Tas, M.; Meuth, S. G.; Perosa, V.; Charidimou, A.; Schreiber, S.

2025-09-25 neurology
10.1101/2025.09.24.25336535 medRxiv
Show abstract

Cortical superficial siderosis (cSS) is the strongest marker of future intracranial hemorrhage (ICH) in cerebral amyloid angiopathy (CAA), a condition without disease-modifying therapy. Emerging evidence suggests that cSS may reflect meningovascular inflammation potentially amenable to immunosuppression. We conducted a single-center matched cohort study comparing high-dose corticosteroid therapy versus no immunosuppression in patients with probable CAA and cSS. Treated patients were matched 1:2 to untreated patients by demographics, cSS multifocality, and baseline intracerebral hemorrhage. The primary outcome was incident ICH; secondary outcomes were any cerebrovascular event and mortality. Kaplan-Meier analysis assessed event-free survival. cSS progression was evaluated in patients with [≥]6 months MRI follow-up, and inflammation in a subgroup with post-contrast vessel-wall-imaging (VWI). Thirty-six patients were included (12 treated, 24 untreated; mean age 77.4 years; 91% disseminated cSS). Over a median follow-up of 2.23 years, 11 patients had cerebrovascular events (15 ICHs, 3 ischemic strokes), all in the untreated group (p=0.041 for ICH; p=0.037 for any event). Mortality did not differ (p=0.634). cSS progressed in 2/6 (33%) treated versus 4/7 (57%) untreated patients. Baseline VWI detected inflammation in 7/9 patients (78%), which regressed after corticosteroids. Corticosteroid therapy was associated with fewer cerebrovascular events in high-risk CAA, supporting prospective controlled validation.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.