Effects of dabigatran versus warfarin on 2-year cognitive outcomes in old patients with atrial fibrillation: results from the GIRAF randomized clinical trial

Background: Observational studies support a role for oral anticoagulation to reduce the risk of dementia in atrial fibrillation patients, but conclusive data are lacking. Since dabigatran offers a more stable anticoagulation, we hypothesized it would reduce cognitive decline when compared to warfarin in old patients with atrial fibrillation. Methods: The GIRAF trial was a 24-month, randomized, parallel-group, controlled, open-label, hypothesis generating trial. The trial was done in six centers including a geriatric care unit, secondary and tertiary care cardiology hospitals in Sao Paulo, Brazil. We included patients aged >= 70 years and CHA2DS2-VASc score > 1. The primary endpoint was the absolute difference in cognitive performance at 2 years. Patients were assigned 1:1 to take dabigatran (110 or 150 mg twice daily) or warfarin, controlled by INR and followed for 24 months. Patients were evaluated at baseline and at 2 years with a comprehensive and thorough cognitive evaluation protocol of tests for different cognitive domains including the Montreal Cognitive Assessment (MoCA), Mini-Mental State Exam (MMSE), a composite neuropsychological test battery (NTB), and computer-generated tests (CGNT). Results: Between 2014 and 2019, 5523 participants were screened and 200 were assigned to dabigatran (N = 99) or warfarin (N = 101) treatment. After adjustment for age, log of years of education, and raw baseline score, the difference between the mean change from baseline in the dabigatran group minus warfarin group was - 0.12 for MMSE (95% confidence interval [CI] - 0.88 to 0.63; P = 0.75), 0.05 (95% CI - 0.07 to 0.18; P = 0.40) for NTB, - 0.15 (95% CI - 0.30 to 0.01; P = 0.06) for CGNT, and - 0.96 (95% CI - 1.80 to 0.13; P = 0.02) for MoCA, with higher values suggesting less cognitive decline in the warfarin group. Conclusions: For elderly patients with atrial fibrillation, and without cognitive compromise at baseline that did not have stroke and were adequately treated with warfarin (TTR of 70%) or dabigatran for 2 years, there was no statistical difference at 5% significance level in any of the cognitive outcomes after adjusting for multiple comparisons.