Anticoagulation Linked to Lower Mortality in Adults 80 and Older With Atrial Fibrillation

Anticoagulation was associated with lower all-cause mortality among adults aged 80 years or older with atrial fibrillation (AF), despite more gastrointestinal bleeding, according to a retrospective cohort study published in the Journal of General Internal Medicine.

The investigators noted that few randomized clinical trials have enrolled adults in this age group, leaving the benefits and risks of anticoagulation for them poorly characterized. They analyzed 45,305 patients with AF in MDClone, a synthetic electronic health record database derived from the BJC HealthCare network, a large US health system. Patients were grouped as continuing, initiating, discontinuing, or never receiving anticoagulation; those with mitral stenosis, mechanical heart valves, cirrhosis, or end-stage renal disease were excluded. Anticoagulants included warfarin, apixaban, rivaroxaban, edoxaban, and dabigatran.

One-year mortality was 8.7% among patients who continued anticoagulation and 17.7% among those never anticoagulated. In the primary time-varying Cox model, used to address immortal time bias, anticoagulation was associated with a 25% lower hazard of all-cause mortality (hazard ratio, 0.75; 95% CI, 0.71-0.79). In a multivariable model, the initiation, discontinuation, and never groups each had higher 1-year mortality than the continuation group.

The association persisted after propensity score matching of the continued and never groups and in a 90-day landmark analysis, although the landmark estimate for the never group was more modest (hazard ratio, 1.16; 95% CI, 1.03-1.30). By 5 years, mortality had converged across groups (27.5% to 30.2%). Gastrointestinal bleeding occurred in 13.0% of the continuation group and 11.9% of the initiation group, compared with 6.5% of the discontinuation group and 6.6% of the never group. Intracranial hemorrhage rates were low (1.3% to 1.6%) and did not differ significantly between groups. Secondary outcomes were compared without adjustment.

The authors noted that the observational design cannot establish causality, that unmeasured confounders may remain, and that selection bias is likely because clinicians and patients, not randomization, decided who received treatment.  Cause-specific mortality and INR values were unavailable. The authors concluded that the findings support continuing anticoagulation in carefully selected older patients instead of withholding it because of age alone. 

Source: Kaya G, Sanchez Garcia J, Schneider M, et al. Anticoagulation and all-cause mortality in adults aged 80 years and older with atrial fibrillation: a retrospective cohort study. J Gen Intern Med. Published online October 27, 2026. doi:10.1007/s11606-026-10920-7