Warfarin underperforms in preventing stroke in the real world in patients with chronic AF

Analyses of 3 retrospective databases found that warfarin is much less effective at preventing strokes in the real world in patients with chronic atrial fibrillation (AF) than it is in clinical trials. The retrospective data were reported at the ASA's International Stroke Conference 2005.

In randomized, controlled clinical trials, warfarin reduced stroke risk by about 65% in carefully monitored patients with AF, said Brian F. Gage, MD, MSc, but the participants in these trials tended to be younger than real-world users. In addition, the clinical trials had many exclusion criteria for study entry.

Dr Gage and colleagues reviewed Medicare part A and part B claims and obtained discharge warfarin prescription information and the dates of prothrombin time monitoring of warfarin therapy for 17,272 Medicare beneficiaries.

Warfarin was prescribed for 49.7% of Caucasian patients with AF, 43.2% of African-American patients, and 40.2% of Hispanic patients. When warfarin was prescribed, African-American and Hispanic patients were 2 to 3 times more likely than Caucasian patients not to have follow-up. Gaps in INR monitoring of 90 days or more occurred in 0.8% of Caucasian patients compared with 1.9% of Hispanic patients and 3.8% of African-American patients. The mean time between INR monitoring was 26 days in Caucasian patients, 30 days in African-American patients, and 33 days in Hispanic patients.

The reduction in the risk of ischemic stroke was 40% in Caucasian patients who were prescribed warfarin, but African-American and Hispanic patients prescribed warfarin had no reduction in stroke risk compared with those not prescribed warfarin.

"Warfarin monitoring is less consistent in African-American and Hispanic beneficiaries, but this disparity may not fully explain the lower effectiveness in these populations," said Dr Gage. Other factors that may have contributed to the lack of warfarin efficacy in racial minorities may be poorer compliance with therapy, resulting in less time in the therapeutic range, and lack of access to care. In addition, some health-care systems are overburdened and unable to routinely perform INR monitoring, he said. Also, because physicians are not compensated for INR monitoring, a financial incentive exists to "stretch out the INRs."

These findings were supported by 2 other epidemiologic studies presented at the International Stroke Conference. In one study, data from 32,784 patients with chronic non-valvular AF from a geographically diverse managed-care organization were analyzed. Sixty-two percent of the patients had at least 1 prescription for warfarin; INR test results were available for about half of those who were ever prescribed warfarin.

The mean follow-up time was approximately 1.5 years. After adjusting for potential clinical and demographic confounders, warfarin exposure was associated with a 40% reduction in the risk of ischemic stroke, an effect identical to the one obtained by Dr Gage, and a 29% reduction in the risk of transient ischemic attack. Among the treated patients, INR levels were within therapeutic range 54% of the time, said lead investigator Judith J. Stephenson, SM, senior researcher at Aetna Health Information Solutions in Blue Bell, Penn.

The third retrospective analysis, this one in the general Medicare population with non-valvular AF, demonstrated no reduction in hospitalization for stroke among warfarin users, said Charles A. Herzog, MD, professor of medicine, University of Minnesota, Minneapolis, Minn.

His analysis included 618,574 Medicare patients who had a diagnosis of non-valvular AF during 1992 to 2002. Patients were considered to be receiving warfarin therapy if they had at least 3 prothrombin claims during the year.

Over the 10 years, the percentage of AF patients who received warfarin increased from 24% to 56%. According to inpatient Medicare claims, ischemic stroke hospitalizations decreased from 7% to 3%, and hemorrhagic stroke hospitalizations increased slightly from 0.5% to 1.5%.

Warfarin use, however, was not associated with the diminished occurrence of stroke. During the study period, "the number of ischemic strokes was lower on warfarin but there was a higher rate of hemorrhagic cerebrovascular accidents," said Dr Herzog. "The overall effect was null in protecting against stroke. This implies that the drug is behaving differently in the real world."

Difficulty in maintaining a target therapeutic level and the number of drugs with which warfarin interacts are potential explanations for the lack of warfarin efficacy, Dr Herzog said.