Current issues of ACP Journal Club are published in Annals of Internal Medicine


Stroke prevention with anticoagulants was cost effective in patients with atrial fibrillation

ACP J Club. 1993 May-June;118:92. doi:10.7326/ACPJC-1993-118-3-092

Source Citation

Gustafsson C, Asplund K, Britton M, et al. Cost effectiveness of primary stroke prevention in atrial fibrillation: Swedish national perspective. BMJ. 1992 Dec 12;305:1457-60.



To estimate the costs and savings of primary stroke prevention with anticoagulants or aspirin in patients with atrial fibrillation in the Swedish population.


Cost-effectiveness analysis based on 4 previously published randomized controlled trials evaluating the effectiveness and hemorrhagic side effects of primary stroke prevention, 2 observational studies assessing the prevalence of atrial fibrillation and the incidence of stroke, and estimates of costs derived from Swedish health care data.




83 000 persons aged 50 to 89 years were estimated to have atrial fibrillation in Sweden. The prevalence varied from 0.5% (aged 50 to 59 years) to 8.8% (aged 80 to 89 years). 22 000 patients were estimated to be candidates for anticoagulant treatment and 55 000, for aspirin.


Warfarin only, warfarin and aspirin (aspirin for patients not eligible for warfarin), and aspirin only. Warfarin was assumed to reduce the risk of stroke by 64% and aspirin by 25%. 50% of patients were assumed to be eligible for warfarin and 90%, for aspirin.

Main costs and outcome measures

Number of strokes prevented, direct and indirect savings from prevented strokes, and costs of preventive treatment were estimated.

Main results

The risk for stroke was assumed to be 5% annually. Depending on the annual risk of intracerebral hemorrhage caused by anticoagulants (0.3% to 2.0%), 34 to 83 patients would need to be treated annually with anticoagulants to prevent 1 additional stroke (ischemic and hemorrhagic). 83 patients would need to be treated annually with aspirin to prevent 1 additional stroke. The number of strokes prevented ranged from 1300 (warfarin and aspirin, risk for intracerebral hemorrhage on warfarin 0.3%) to 260 (warfarin only risk for intracerebral hemorrhage on warfarin 2%). With warfarin only, the total cost (treatment cost minus direct and indirect savings) per stroke prevented varied from a saving of 99 Swedish kronor (Kr) {U.S. $16}* (intracerebral hemorrhage 0.3%) to a cost of Kr 147 {U.S. $24}* (intracerebral hemorrhage, 2%). Aspirin saved only Kr 262 {U.S. $43}* per stroke prevented.


Primary stroke prevention with anticoagulants or, if contraindicated, aspirin resulted in a net savings, provided the risk for hemorrhagic stroke from anticoagulants was low.

Sources of funding: FONUS; County Council in Skaraborg; Serafimer Hospital Foundation.

For article reprint: Dr. C. Gustafsson, Senior Registrar, Department of Medicine, Mariestad Hospital, Box 411, S-54224 Mariestad, Sweden. FAX 46-501-62845.

*Conversion rate used was the average for 1991 [U.S. $1 = 6.053 Kr].


Like most economic evaluations, the data about efficacy and costs used in the study by Gustafsson and colleagues were derived from different sources and are of varying quality. The most reliable datum concerns the efficacy of warfarin, which has been consistently shown in 5 randomized trials (1). In this study the cost effectiveness of warfarin depended on the rate of intracerebral hemorrhage, but even at a rate of excess hemorrhage considerably higher than that seen in the randomized trials (1.3%), money was saved for each stroke prevented. It is relatively rare for new medical interventions to actually save money (2), and this makes warfarin therapy for atrial fibrillation appear both clinically and economically attractive.

The lifetime cost of stroke that was used in this study was derived from unpublished data, which is cause for some concern. It would have been reassuring to see the results of a sensitivity analysis done on this and other variables.

As the authors point out, there is considerable uncertainty about the effectiveness of aspirin. Their conclusions about aspirin, therefore, must be interpreted with caution until the results of ongoing studies comparing the efficacy of warfarin and aspirin are reported.

Andreas Laupacis, MD
Ottawa Civic HospitalOttawa, Ontario, Canada


1. Laupacis A, Albers G, Dunn MI, Feinberg WM. Antithrombotic therapy in atrial fibrillation. Chest. 1992;102(Suppl):426S-33S.

2. Laupacis A, Feeny D, Detsky A, Tugwell P. How attractive does a new technology have to be to warrant adoption and utilization? Tentative guidelines for using clinical and economic evaluations. Can Med Assoc J. 1992;146:473-81.