Anticoagulation in atrial fibrillation over 80 was linked to lower mortality despite more GI bleeding

Design

Retrospective cohort with time-varying Cox model and propensity matching

Population

45,305 adults aged ≥80 with atrial fibrillation, US health system

Primary outcome

All-cause mortality

Effect

HR 0.75 (0.71–0.79); 1-year mortality 8.7% continued vs 17.7% never anticoagulated

Patients over 80 are the ones most often denied anticoagulation for atrial fibrillation, and the least represented in the trials. This retrospective cohort, published in the Journal of General Internal Medicine on 27 October, used electronic records from a large US health system for 45,305 adults aged 80 or older with atrial fibrillation, excluding mitral stenosis, mechanical valves, cirrhosis and end-stage kidney disease. Patients were grouped as continuing, starting, stopping or never taking anticoagulation, and immortal time bias was addressed with a time-varying model, a landmark analysis and propensity matching.

One-year mortality was 8.7% in those continuing anticoagulation against 17.7% in those never anticoagulated. Overall, anticoagulation was associated with a 25% lower hazard of death (HR 0.75, 95% CI 0.71–0.79). Compared with continuing, mortality was higher in those who stopped (HR 1.64) or never started (HR 2.22). Gastrointestinal bleeding was more frequent on anticoagulation (about 12–13%), while intracranial haemorrhage was low and similar across groups (1.3–1.6%). By five years, mortality had converged.

This is observational, and frailer patients are more likely to be left off anticoagulation, which would exaggerate the apparent benefit. Even so, the direction matches trial evidence in older subgroups, and the absence of excess intracranial bleeding addresses the fear that usually drives the decision. Current guidelines already recommend anticoagulation based on stroke risk, not age.

• Do not withhold or stop anticoagulation in atrial fibrillation on the grounds of age alone.
• When stopping is considered, weigh specific reasons — recurrent major bleeding, falls with head injury, limited life expectancy — and record them.
• Anticipate gastrointestinal bleeding: review aspirin and NSAID co-prescription and consider gastroprotection where indicated.
• Check the dose against renal function and weight; underdosing does not remove bleeding risk.
• Revisit the decision at discharge from any admission, when anticoagulation is often stopped and not restarted.

Why it matters

Age is the commonest reason anticoagulation is withheld, and this links that habit to higher mortality without lower brain-bleed risk.

Don't overread it

This is observational; part of the mortality difference may reflect who was chosen for anticoagulation, not the drug.

The statistics, in plain English

A hazard ratio of 0.75 means a 25% lower rate of death over follow-up among those anticoagulated. Because patients were not randomised, healthier people may have been more likely to receive anticoagulation; the authors used several methods to reduce this bias, but cannot remove it.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.