NEW ORLEANS—Patients with atrial fibrillation who have the highest risk of ischemic stroke derive the greatest net clinical benefit from warfarin therapy to prevent thromboembolic events, reported Daniel E. Singer, MD, at the 2008 International Stroke Conference.
Current guidelines endorse risk-based use of warfarin in patients with atrial fibrillation, but the only risk explicitly considered is that for ischemic stroke. "These risk estimates are primarily based on old clinical trials and do not reflect current event rates," said Dr. Singer, Director of the Epidemiology, Outcomes, and Disease Management Unit at the Massachusetts General Hospital Clinical Research Program, Boston. "Importantly, these risk schemes do not explicitly account for the risk of major hemorrhage, particularly intracranial hemorrhage." The impact of intracranial hemorrhage is comparable to that of ischemic stroke, he added, and should therefore be included in assessing the risk:benefit ratio of warfarin in patients with atrial fibrillation.
In the AnTicoagulation and Risk factors In Atrial fibrillation (ATRIA) study, Dr. Singer and his colleagues calculated the net clinical benefit of warfarin in a cohort of 13,559 patients with nonvalvular atrial fibrillation who were enrolled in Kaiser Permanente of Northern California.
Warfarin exposure was determined using a validated algorithm based on warfarin prescriptions and serial outpatient measurement of the international normalized ratio. The overall net clinical benefit was defined as: (thromboembolic rate off warfarin minus thromboembolic rate on warfarin) minus (intracranial hemorrhage rate on warfarin minus intracranial hemorrhage rate off warfarin).
The median age of the cohort was 73; 43% were female; 9% had a prior stroke, and 51% had hypertension. The median follow-up was six years, generating 66,754 person-years of observation. During the follow-up, there were 1,092 validated thromboembolic events, mostly ischemic strokes, and 299 validated intracranial hemorrhage events.
The net clinical benefit was lower than previously reported rates of thromboembolic events prevented, for two reasons. One reason was the subtraction of intracranial hemorrhage from thromboembolic events prevented. The other was that the absolute rates of thromboembolic events in the ATRIA trial were lower than those observed in older trials.
In patients with a prior stroke, 3.96 thromboembolic events per 100 person-years were prevented and 0.27 intracranial hemorrhage events were induced, for a net clinical benefit of 3.69 events per 100 person-years. The net clinical benefit was 0.57 per 100 person-years in those without a prior stroke. The net clinical benefit of warfarin increased with advancing age. "With older age, the risk of ischemic stroke increases sharply, and the rate of thromboembolisms prevented by warfarin increases in parallel," explained Dr. Singer. "While the risk of intracranial hemorrhage increases in old age, the incremental impact of warfarin is relatively small." He noted that "these [older] patients were all presumably selected as reasonably good warfarin candidates."
Persons 85 or older had a net clinical benefit with warfarin of 2.13 events prevented per 100 person-years, those ages 75 to 84 had a net clinical benefit of 0.83, and those ages 65 to 74 had a net clinical benefit of 0.09. "The patients younger than 65 may actually do a little worse as a result of taking warfarin," he said.
The net clinical benefit of warfarin also improved with increasing ischemic stroke risk by the CHADS2 criteria, Dr. Singer noted.
—Wayne Kuznar