Two percent of adults 65 to 75 years of age will develop atrial fibrillation, and the risk for developing this cardiac irregularity increases with age. Studies have shown that warfarin (Coumadin) therapy in patients with atrial fibrillation reduces the concomitant risk of stroke. Warfarin therapy presents challenges because of its narrow therapeutic window and its risk of overcoagulation, which increases the risk for hemorrhage. Maintaining warfarin therapy so that the international normalized ratio (INR) is between 2 and 3 is the most effective and safe level. Reynolds and colleagues performed a systematic review and meta-analysis on the safety and efficacy of warfarin therapy in nonvalvular atrial fibrillation. They assessed the risk for bleeding associated with overcoagulation (INR greater than 3) and the risk of stroke and other thromboembolic events with undercoagulation (INR less than 2).

A systematic review of the literature was performed using all relevant studies published in English between January 1, 1985, and October 30, 2002. In addition, the authors searched the reference lists of the retrieved articles to identify other relevant studies. Studies were included if they enrolled patients with nonvalvular atrial fibrillation who were receiving warfarin therapy. The studies also must have reported stroke or bleeding events based on the therapeutic INR range or time spent above, at, or below the INR therapeutic range. Studies that used combination therapy were not included in the analysis. Data extracted from the studies included patient and treatment characteristics, INR information, and number of patients with adverse events of interest by INR level.

There were 21 studies that met the inclusion criteria, with a total of 6,248 patients enrolled. The meta-analysis of the studies comparing ischemic events when the INR was less than 2, compared with when it was 2 or greater, showed an increase in the odds ratio of 5.07 (95% confidence interval [CI], 2.92 to 8.80). If the INR was greater than 3, the odds ratio for bleeding events was 3.21 (95% CI, 1.24 to 8.28) when compared with an INR of 3 or less. Four studies reported time spent in therapeutic range. The results showed an INR between 2 and 3 was maintained 61 percent of the time, exceeded the range 13 percent of the time, and was below the range 26 percent of the time.

The authors conclude that the risk for ischemic stroke in patients receiving warfarin therapy for nonvalvular atrial fibrillation is significantly higher in patients who are receiving insufficient anticoagulation (i.e., INR less than 2). They also note that the risk for adverse events is higher in patients who are overcoagulated (i.e., INR greater than 3). The authors note that the published data on this issue are sparse, heterogeneous, and primarily reported from clinical trials.