In:
Stroke, Ovid Technologies (Wolters Kluwer Health), Vol. 50, No. Suppl_1 ( 2019-02)
Abstract:
Introduction: Accurately balancing risks of interventional treatment and intracranial hemorrhage (ICH) in the untreated course of unruptured brain arteriovenous malformation (uBAVM) patients is critical for optimal management. Observational studies of uBAVM afford an opportunity to address important gaps and previous criticisms of randomized data to help guide treatment decisions and assess long-term risk/benefit ratio. We previously published an individual patient data meta-analysis (IPDMA) in 4 MARS cohorts and identified ICH presentation and increasing age as significant predictors of ICH during follow-up, but was not powered to detect predictors in those with uBAVM. Thus, the goals of this project are to identify risk factors for ICH in uBAVMs, to estimate the risk with precision, and to create personalized, risk-prediction models. Methods: MARS is an international, multi-center study of 11 cohorts with target enrollment of 4,500 uBAVM patients ascertained through population-based or referral-based studies. We are harmonizing both retrospective and prospective data, including clinical, lifestyle, imaging, and angiographic factors. Treatments, complications, and functional outcomes (physician and patient-reported) will be updated annually between 2018-2022. We propose to: 1) identify predictors of outcome in uBAVMs using IPDMA; 2) test whether long-term outcomes differ by treatment using statistical approaches for causal inference and unbiased estimates; 3) compare treatment outcomes in randomized and non-randomized data to address generalizability; and 4) validate models and provide a novel tool for calculating individualized risks. Eligible sites have: a) prior BAVM publications; b) data from a minimum of 100 uBAVM patients; c) agree to random outcome adjudication; and d) agree to MARS data sharing policy. Discussion: The NIH-funded MARS consortium will provide important and comprehensive characterization of the untreated and treated course of uBAVMs, using large observational datasets and sophisticated epidemiological approaches. The data and models generated will be a useful resource for decision-making. We welcome participation from eligible sites with longitudinal data on uBAVM patients
Type of Medium:
Online Resource
ISSN:
0039-2499
,
1524-4628
DOI:
10.1161/str.50.suppl_1.TP585
Language:
English
Publisher:
Ovid Technologies (Wolters Kluwer Health)
Publication Date:
2019
detail.hit.zdb_id:
1467823-8
