Machine Learning-Identified Potential Interaction Between Clazosentan and Nicardipine in Patients with Subarachnoid Hemorrhage

Authors:
Open Access
Publisher website
Google Scholar
Abstract
Background/Objectives: Subarachnoid hemorrhage (SAH) is frequently complicated by cerebral vasospasm (VS). Clazosentan has reduced VS in Japanese studies but shown inconsistent efficacy in Western trials. We hypothesized that clinical and pharmacologic interactions may influence its effectiveness. Methods: We analyzed the multicenter “Database of Cohort Study for Outcome of SAH In Japan” (DCI Japan) registry, prospectively collected from 2020 to 2023, to assess associations between clazosentan use, VS prevention, functional outcomes, and potential interactions in adults with aneurysmal SAH (aSAH) treated by surgical clipping or endovascular coiling within 4 days of onset. Outcomes included angiographic VS (AVS), symptomatic VS (SVS), cerebral infarction, and modified Rankin Scale (mRS) scores at discharge and at 6 months. Predictors and interactions were first screened using univariable analysis and Light Gradient Boosting Machine, then evaluated via multivariable logistic regression. Results: Among 544 patients (mean age 65.2 ± 14.2 years; 71.5% female), 34.0% received clazosentan. AVS, SVS, and cerebral infarction occurred in 20.6%, 16.0%, and 22.4%, respectively. Poor outcomes (mRS 3–6) were observed in 48.8% at discharge and 33.7% (137/406) at 6 months. Clazosentan use was associated with reduced odds of AVS (OR 0.27, 95% CI [0.11–0.69]), SVS (OR 0.15 [0.04–0.64]), and poor 6-month outcome (OR 0.08 [0.01–0.68]). A potential interaction with nicardipine was linked to higher odds of AVS (OR 1.85 [1.43–2.65]). Conclusions: Clazosentan was associated with reduced VS and improved 6-month outcomes after aSAH, although concomitant nicardipine may attenuate its prophylactic effectiveness against AVS.
Funding Information
  • Japan Society for the Promotion of Science Grant-in-Aid for Scientific Research (23K08521)
  • European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant (101034252)
  • Research Ireland (12/RC/2289_P2)

This publication has 49 references indexed in Scilit: