Abstract Details

The ARCADIA trial (n=1015), a secondary stroke prevention study, evaluated the efficacy of apixaban vs aspirin in patients with CS and AC. We hypothesized that the rate of neurocognitive decline would be slower among subjects randomized to apixaban due to fewer silent infarcts.

To determine if apixaban prevents cognitive decline in patients with cryptogenic stroke (CS) and atrial cardiopathy (AC).

Subjects randomized in ARCADIA who were actively taking study drug were approached for enrollment into this ancillary study. Participants underwent serial neurocognitive testing administered centrally by telephone. The baseline was obtained >3 months after the index stroke and follow-up yearly thereafter. Testing consisted of a 5-test battery and patients’ overall cognitive functioning was summarized as a composite Z-score based on scores from the five tests. Patients with a baseline composite Z-score > 1.5 SDs below norms were excluded from the analyses. Trajectories of the composite Z-score and on each individual test were compared between treatment arms using a mixed-effect model.

Of the 310 enrolled in ARCADIA-CSI, 14 (5%) did not complete any neurocognitive batteries. Of the 296 subjects with baseline data, 47 (16%) were excluded due to impaired baseline scores. In the remaining 249 patients (apixaban 128, aspirin 121), 582 cognitive assessments were obtained. Baseline characteristics were balanced between the apixaban and aspirin arms. Mean age was 66.7 years, median mRS 1, 52% female and 19% black patients. During median follow-up of 378 days, the annual change in the overall standardized composite score was similar in the aspirin arm (0.08; 95% CI:0.02-0.15) and the apixaban arm (0.11; 95% CI:0.04-0.17). There were also no differences between treatment arms when comparing cognitive trajectories of each individual neurocognitive test.

Among patients with CS and AC, the rate of cognitive change was not improved with apixaban compared to aspirin.