Abstract Body: Introduction: The ARCADIA trial intention to treat (ITT) analysis did not show decreased stroke recurrence using apixaban over aspirin in patients with recent cryptogenic stroke (CS) and evidence of atrial cardiopathy. One reason for trial neutrality could be that recurrent strokes were often non-CS and thus unlikely to be reduced by apixaban. Details of the recurrent stroke events are thus described.
Methods: ARCADIA stroke outcome events were adjudicated via review of a packet of medical records by non-study stroke neurologists. Stroke types and subtypes were determined by 2 study neurologists (WL, DT) via information from packets and adjudicators’ notes. Stroke subtypes were classified using a modified TOAST scheme classified as: large artery atherosclerotic (LAA), small vessel disease (SVD), or cardioembolic (CE) if an evident finding was identified OR otherwise CS, even if workup was incomplete.
Results: Adjudicated strokes were recorded in 82 ARCADIA patients. Two events classified as CS occurred after study closure and were not included in the ITT analysis; patients were off study medication. Of the 80 strokes included in the ITT analysis, 4 patients never took study medication, with strokes including 2 CS, 1 SVD and 1 LAA. Of the 76 patients that took study medication for some duration, 20 had discontinued study medication when they had their stroke; strokes were 14 (70%) CS, 3 (15%) non-ischemic, 2 (10%) CE and 1 (5%) SVD. Of the remaining 56 patients who were on study medication when stroke occurred, 22 were in the apixaban group and 34 in the aspirin group; strokes were 44 (79%) CS, 4 (7%) LAA, 4 (7%) CE, 3 (5%) non-ischemic, and 1 other (endocarditis and LAA) without heterogeneity of recurrent stroke types between randomized treatments. Overall, 60/80 (75%) of the recurrent strokes included in the ITT analysis were CS.
Conclusions: The ARCADIA trial hypothesis is that recurrent CS in randomized patients could be reduced by apixaban over aspirin, as the stroke mechanism is hypothesized to be like that in atrial fibrillation. The vast majority (75%) of recurrent stroke events included in the ARCADIA ITT analysis were CS, so recurrent stroke being non-CS is not an explanation for the overall neutral ARCADIA results. Only 56/80 (70%) of recurrent strokes occurred while patients were taking study medication, with numerically fewer in the apixaban than the aspirin arm; further analysis of this on-study-medication group are ongoing.