Purpose: Using data from the largest multinational prospective registry in atrial fibrillation (the Global Anticoagulant Registry in the FIELD–Atrial Fibrillation, GARFIELD-AF), we examined whether treatment effects of oral anticoagulation (OACs) vs no OAC and, non-vitamin K antagonist oral anticoagulants (NOACs) vs VKAs on all-cause mortality are manifest within 3 months of diagnosis of AF in patients with a CHA2DS2-VASc score ≥2 (including gender).
Methods: Patients were enrolled consecutively into GARFIELD-AF and followed prospectively. The analyses were conducted in patients enrolled between Apr-2013 and Sep-2016, during which time NOACs became available in many countries. Within each comparison, overlap probability weighting was used to evaluate the adjusted associations between drug use (at baseline) and outcomes within 3 months. Weights were applied to Cox proportional hazards models to estimate the effects for OAC vs no OAC and NOAC vs VKA use for each endpoint, respectively.
Results: The study population comprised 20,457 anticoagulated patients (10,330 [NOACs]; 10,127 [VKAs]) and 8,019 patients without OAC (of the latter, 60% received anti-platelets). 442 patients died within 3 months of diagnosis of AF; the Kaplan-Meier survival rate at 3 months was 98%. The causes of death were: cardiovascular (in 41% of cases), non-cardiovascular (36%) and unknown (22%). Congestive heart failure (16%), cancer (8%) and respiratory failure (6%) were the most common known causes of death followed by: myocardial infarction (5%), ischaemic stroke (5%), sudden death (5%), infection (5%) and sepsis (5%). After weighting, standardised differences showed an accurate balance between the 29 baseline variables and drug use. Weighted hazard ratios (HR) for all-cause mortality were: 0.57 (95% CI, 0.46–0.71); P<0.001 for the comparison of OAC vs non OAC; and 0.69 (95% CI, 0.52–0.92); P=0.010 for NOACS vs VKAs (figure). Stroke/systemic embolism (117 events, overall) and major bleeding (99 events) were not significantly different for the treatment comparisons.
Conclusion: GARFIELD-AF reveals significant early mortality in patients with newly diagnosed AF and significant mortality differences in favour of OACs, even after adjustment for 29 baseline variables. These differences are manifest within 3 months after diagnosis (number needed to treat, NNT=80). Strokes constituted only a minority of the events prevented. The benefit of NOACs (relative to VKAs) on mortality was also observed by 3 months (NNT=188). These findings extend the evidence from randomised trials with data from a prospective multinational registry population with newly diagnosed AF and raise questions about the impact of anticoagulation, beyond stroke prevention.