Does anticoagulation improve survival in people aged 75 and over in atrial fibrillation? A comparison of results from randomised and non-randomised patients in the BAFTA trial in Int. J. Stroke

J. Mant, Kate Fletcher, Richard Hobbs, Rebecca Kandiyali, Andrea Roalfe

Research output: Contribution to conferenceConference Abstractpeer-review

Abstract

Introduction: Non-randomised studies have shown that warfarin
improves survival in atrial fibrillation (AF) relative to aspirin, but
not randomised trials. Our aim was to compare survival of people in
AF who did and did not take part in a randomised trial (BAFTA),
and the impact that treatment had in these two populations.
Method: Extended follow up (median 4.5 years) of mortality of a
randomised and non-randomised cohort (n = 1440) age 75+ recruited
from primary care and treated with warfarin (target INR 2.5) or
aspirin (usual dose 75 mg). Treatment effects were adjusted for age,
risk of stroke (CHADS2 score) and recency of AF diagnosis.
Results: Non-randomised patients prescribed warfarin were younger
than those prescribed aspirin (mean age 80.3 vs. 83.1), but at
higher risk of stroke. There was no evidence of a treatment effect
on mortality in the randomised cohort (29% vs. 30% at 4 years,
HR 0.98 (0.79–1.21), but there was in the non-randomised cohort
(22% vs. 35% at 4 years) favouring anticoagulation, which
remained significant for vascular mortality after adjustment (HR:
0.46, 0.23–0.89) but not for all cause mortality (HR: 0.76, 0.49–
1.17).
Conclusion: Warfarin is unlikely to increase survival in elderly people
with AF. Apparent improvements in survival in non-randomised
studies are likely to be the result of confounding by unmeasured factors.
Non-randomised designs can give substantively different
answers to randomised trials, even when carried out in the same
population using identical methods of ou
Original languageEnglish
Publication statusPublished - 2010

Fingerprint Dive into the research topics of 'Does anticoagulation improve survival in people aged 75 and over in atrial fibrillation? A comparison of results from randomised and non-randomised patients in the BAFTA trial in Int. J. Stroke'. Together they form a unique fingerprint.

Cite this