Abstract
BACKGROUND: Observational data comparing warfarin with no treatment for patients with non-valvular atrial fibrillation (NVAF) and severely reduced glomerular filtration rate (GFR) are conflicting and randomized controlled trials (RCTs) are lacking. Most studies do not provide information on warfarin treatment quality, making them difficult to compare. METHODS: This national cohort study investigates the risk of ischaemic stroke and major bleeding during warfarin treatment compared with no oral anticoagulants in patients with NVAF, GFR category 3-5 (G3-G5) or on dialysis (G5D), with kidney transplant recipients excluded, between 2009 and 2018. Data extracted from high-quality Swedish national healthcare registries, including the Swedish Renal Registry, AuriculA-the Swedish national quality registry for atrial fibrillation and anticoagulation-and the Stroke Registry. RESULTS: At enrolment of 12 106 patients, 21.4% were G3, 43.5% were G4, 11.6% were G5 and 23.6% were G5D. The mean time in the therapeutic range was 70%. Warfarin compared with no treatment showed a lower risk for ischaemic stroke for G3 {hazard ratio [HR] 0.37 [95% confidence interval (CI) 0.18-0.76]}, G4 [0.53 (0.38-0.74)] and G5 [0.49 (0.30-0.79)] and an increased risk of major bleeding in G4 [HR 1.22 (1.02-1.46)], G5 [1.52 (1.15-2.01)] and G5D [1.23 (1.00-1.51)]. All-cause mortality was more than halved on warfarin compared with no treatment in all GFR categories. CONCLUSIONS: Warfarin treatment is associated with a lower risk of ischaemic stroke for patients with NVAF and G3, G4 and G5D at the cost of a higher risk of major bleeding for G4-G5D. Existing observational data are conflicting, stressing the need for RCTs on warfarin compared with no treatment in G4-G5D. Awaiting RCTs, it seems reasonable to treat selected patients on dialysis and NVAF with warfarin.