Despite major advances in the management of heart failure with angiotensin-converting–enzyme inhibitors, beta-blockers, and resynchronization therapy, there are more than 1 million hospitalizations for heart failure in the United States each year, and mortality remains high. As compared with the general population, patients with heart failure have an increased risk of stroke and of systemic thromboembolic events, which are believed to arise from within the heart as a result of left ventricular stasis, endocardial dysfunction, and a systemic hypercoagulable state. Heart failure is a risk factor for atrial fibrillation, which, even if asymptomatic, further increases the risk of stroke. Consequently, one would expect that patients with heart failure could benefit from oral anticoagulant therapy. However, until now, the hypothesis that patients with heart failure benefit from long-term anticoagulant therapy has been tested only in modest-sized randomized, controlled trials.

In the Warfarin versus Aspirin in Reduced Cardiac Ejection Fraction (WARCEF) trial reported by Homma et al., 2305 patients (mean age, 61 years - relatively young for a population with heart failure) who had severe left ventricular dysfunction (mean left ventricular ejection fraction of 25%) were randomly assigned to receive warfarin (with a target international normalized ratio [INR] of 2.0 to 3.5) or aspirin (at a dose of 325 mg per day), with treatment continued for a mean of 3.5 years. Only 43% of patients had evidence of underlying ischemic heart disease. This, together with the exclusion of patients with known atrial fibrillation, meant that the trial was primarily testing whether anticoagulant therapy for the prevention of an embolism emanating from the left ventricle or caused by subclinical atrial fibrillation would lead to a reduction in the composite end point of stroke or death.

The careful conduct of this blinded trial, in which patients in the warfarin group had good control of INR levels (mean time in the therapeutic range after a 6-week period of dose adjustment, 62.6%) and which included more than 600 primary outcome events, has provided clinicians with clear answers. As compared with aspirin, warfarin did not significantly reduce the rate of the primary composite outcome of stroke (ischemic or hemorrhagic) or death.

The rate of ischemic stroke in the aspirin group was 1.36 per 100 patient-years (about one half to one third of the rate one would expect in aspirin-treated patients with established atrial fibrillation), and the rate with warfarin, as compared with aspirin, was reduced by almost 50% (0.72 events per 100 patient-years vs. 1.36 events per 100 patient-years, P=0.005); these findings are consistent with those of the previously published Warfarin/Aspirin Study in Heart Failure (WASH) pilot study and the Warfarin and Antiplatelet Therapy in Chronic Heart Failure (WATCH) trial. The rates of death were similar with warfarin and aspirin (6.63 deaths per 100 patient-years and 6.52 deaths per 100 patient-years, P=0.91), a finding that is also consistent with the lack of an effect of anticoagulants on mortality in previous trials of anticoagulant therapy in patients with heart failure. The lack of an effect of warfarin on mortality suggests that most of the deaths in these patients with heart failure, who had severe impairment of left ventricular function, are unrelated to thromboembolism and, instead, are most likely due to pump failure or ventricular arrhythmias.

The benefit of warfarin in reducing the rate of ischemic stroke was offset by an increase in the rate of major bleeding (1.78 events per 100 patient-years with warfarin vs. 0.87 events per 100 patient-years with aspirin, P<0.001; number needed to harm, 110), although there was no excess of intracerebral bleeding events (0.12 events per 100 patient-years with warfarin and 0.05 events per 100 patient-years with aspirin, P=0.35). Patients with atrial fibrillation place a higher value on prevention of stroke than on reduction of the risk of bleeding events,9 but the modest absolute reduction in the rate of stroke (number needed to treat, 156) and the lack of mortality benefit in the WARCEF trial provide little support for the use of warfarin in preference to aspirin in patients with heart failure.

Could there be some patients with heart failure who would benefit from warfarin anticoagulant therapy? Those most likely to benefit are patients with heart failure who also have atrial fibrillation and those with a history of cardioembolic stroke or formation of left ventricular thrombus. Patients with atherothrombotic coronary artery disease - the most common underlying cause of heart failure and a disease process that is responsive to anticoagulant therapy - might also benefit. Warfarin would be expected to reduce the rates of both ischemic stroke and nonfatal or fatal myocardial infarction in patients with heart failure who also have coronary artery disease, because warfarin is highly effective for the prevention of major cardiovascular events in survivors of myocardial infarction.10 We believe that any future evaluation of anticoagulants in patients with heart failure should focus on patients with underlying coronary heart disease who do not have advanced systolic dysfunction.

In conclusion, the results of the WARCEF trial are consistent with those of three previous smaller randomized, controlled trials in showing that warfarin anticoagulant therapy, as compared with aspirin, is not associated with a reduction in mortality among patients with heart failure. The WARCEF trial provides clear evidence that anticoagulant therapy prevents stroke, probably embolic stroke, in patients with heart failure who have severe systolic dysfunction, but the rates of stroke are too low to justify the routine clinical use of warfarin in most patients with heart failure, in light of the increase in the risk of bleeding.

###

John W. Eikelboom, M.B., B.S., and Stuart J. Connolly, M.D.
N Engl J Med 2012; 366:1936-1938

Source Information
From the Department of Medicine, McMaster University, Hamilton, ON, Canada.