Cost-Effectiveness of Therapy in Nonvalvular Atrial Fibrillation

Source: Catherwood E, et al. Ann Intern Med 1999; 130:625-636.

Catherwood and colleagues performed a cost-effectiveness analysis of treatment strategies in patients with nonvalvular atrial fibrillation. Catherwood et al expanded a previously published Markov decision analysis (), which favored cardioversion plus amiodarone by including health-related costs of the treatments and outcomes in the analysis. Eight different potential strategies were considered. These included rate control with either metoprolol or diltiazem, initial cardioversion followed by aspirin or warfarin at the time of relapse, initial cardioversion followed by quinidine or amiodarone at relapse, and quinidine or amiodarone along with cardioversion at presentation. The Markov model was based on a population of 70-year-old patients of both genders with nonvalvular atrial fibrillation. Patients were assumed to be hemodynamically stable with acceptable symptoms with only rate control. Probabilities of various outcomes for each strategy were estimated from data in the literature. Outcomes and costs were recalculated at three-month intervals over a five-year period. Results are reported as expected costs, increase in quality-adjusted life-years, and incremental cost-effectiveness.

Strategies involving initial cardioversion alone were most effective and less costly than those not involving this option. The cost-effectiveness of the various options after initial cardioversion could be predicted by the patient’s risk of stroke. Among high- and moderate-risk patients, initial cardioversion followed by amiodarone and repeat cardioversion upon relapse were the preferred strategies. Costs were estimated to be $9300 and $18,900 per quality-adjusted life-year increments in these two cohorts, respectively. Among patients thought to be at low risk for stroke since they had no risk factors other than age, cardioversion followed by aspirin therapy on relapse was most cost effective. Sensitivity analysis showed that baseline risk for stroke, estimated stroke rate in sinus rhythm, efficacy of warfarin, and costs of warfarin and amiodarone were the major factors influencing the analysis. Catherwood et al conclude that cardioversion alone should be the initial therapy of non-valvular atrial fibrillation, with amiodarone reserved for relapses

Comment by John P. DiMarco, MD, PhD

Atrial fibrillation is the most commonly encountered sustained arrhythmia at the present time. Catherwood et al present an analysis of the cost-effectiveness of various treatment strategies for patients presenting with new onset persistent atrial fibrillation. Although their model is very complex, it still does not take into account several important factors that may influence clinical decision making. Catherwood et al assumed that their hypothetical patients were asymptomatic and hemodynamically stable in atrial fibrillation on rate-controlling agents only. This is probably the case for those patients in whom atrial fibrillation is discovered by chance but frequently is not the case in patients who present with new onset arrhythmia. Evaluating the efficacy of rate control and the subtle effects of atrial fibrillation on quality of life can be difficult and these problems were not considered here. Problems with chronic warfarin anticoagulation are also frequent. The event rates used here are derived from large, randomized trials but these trials excluded up to one-third of the patients they screened because of real or perceived contraindications to long-term warfarin. Finally, only a minority of atrial fibrillation patients present with persistent atrial fibrillation without any other history of arrhythmia or heart disease. A careful clinical analysis to detect prior episodes of self-terminating episodes of arrhythmia, evidence for associated sinus node dysfunction, or structural heart disease can be helpful for predicting an individual’s risk of arrhythmia recurrence.

The analysis here, therefore, strictly applies to only a narrow subset of atrial fibrillation patients. However, the clinician can use these data as a guideline when he or she is deciding on care for an individual patient. Considerations of symptoms, risk of therapy, and estimates of recurrence rate should be used in making the final clinical decision.


1. Disch DL, et al. Ann Intern Med 1994;120:449-457.

Dr. DiMarco is Professor of Medicine, Division of Cardiology, University of Virginia, Charlottesville.

Cost-effectiveness in the management of nonvalvular atrial fibrillation is enhanced by:

a. initial cardioversion.

b. chronic warfarin therapy.

c. initial amiodarone therapy.

d. None of the above.