Abstract
BACKGROUND: ARCADIA (Atrial Cardiopathy and Antithrombotic Drugs in Prevention After Cryptogenic Stroke), a secondary stroke prevention study comparing apixaban versus aspirin for cryptogenic stroke and biomarkers of atrial cardiopathy, ended prematurely for futility. In the ARCADIA-Cognition study we hypothesized that cognitive decline would be slower in the apixaban arm due to less microembolization. METHODS: ARCADIA subjects on study drug and eligible for magnetic resonance imaging participated in ARCADIA-CSI (ARCADIA-Cognition and Silent Infarction). Cognitive tests were administered centrally by telephone ≥3 months after the ARCADIA index stroke and yearly thereafter. Composite Z scores were the means of 5 standardized cognitive tests. Trajectories of the composite Z score and the individual test scores were compared between treatment arms using a mixed-effects model. RESULTS: Of 799 screened patients at 75 sites, 310 were enrolled in ARCADIA-CSI. Of these, 296 completed at least 1 cognitive exam, and 47 subjects were excluded from the primary analysis for baseline dementia. For the 249 subjects included in the analysis, there were 582 cognitive assessments. Baseline characteristics were balanced between the apixaban (n=128) and aspirin (n=121) arms. Mean age was 68 (SD: 10.4) years, median modified Rankin Scale score 1 (interquartile range, 0-2), 52% female, and 19% Black. During median follow-up of 378 (interquartile range, 183-735) days, the annual change in the overall standardized composite score was 0.084 (95% CI, 0.017-0.149) in the aspirin arm and 0.107 (95% CI, 0.041-0.174) in the apixaban arm (P=0.62). CONCLUSIONS: Cognitive trajectories did not differ between apixaban and aspirin. Further studies should address infarct location and volume and concurrent pathology to determine optimal treatment to mitigate cognitive decline with atrial disease. REGISTRATION: URL: http://clinicaltrials.gov; Unique Identifier: NCT03192215.