Background-Studies of selected populations suggest that anticoagulatio
n in atrial fibrillation is underused and that nonclinical factors inf
luence the use of this stroke-preventing therapy. We wished to examine
recent national trends and predictors of warfarin sodium use in atria
l fibrillation. Methods and Results-A nationally representative sample
of office visits from the 1989 to 1996 National Ambulatory Medical Ca
re Surveys was used. We selected 1125 visits by patients with atrial f
ibrillation, including 877 visits to cardiologists and primary care ph
ysicians in which apparent contraindications for anticoagulation were
absent. The principal outcome measure was the proportion of visits wit
h warfarin reported. We analyzed trends in warfarin use and statistica
lly evaluated the predictors of warfarin use. Warfarin use increased f
rom 13% of atrial fibrillation visits in 1989 to 40% in 1993 (P for tr
end <.001) in patients without contraindications. Between 1993 and 199
6, however, there was no change in warfarin use. Independent of other
factors, warfarin was significantly more likely to be reported in pati
ents with a history of stroke and in patients residing outside of the
South. Conclusions-Warfarin use in atrial fibrillation has not increas
ed recently, indicating inadequate implementation of this highly effec
tive therapy. Barriers to anticoagulation in real-world clinical pract
ice need to be identified and addressed.