Should we use anticoagulation for patients with chronic heart failure?

Ronald S Freudenberger* and Jonathan L Halperin  About the authors

Whether or not to recommend anticoagulation for patients with chronic heart failure (HF) is a complex decision. There are scant data concerning the incidence of and risk factors for thromboembolism, the efficacy of prophylactic treatment, and the potential for drug–drug interactions. In fact, no adequate data from prospective studies of primary or secondary prevention of thromboembolism in patients with systolic dysfunction without atrial fibrillation (AF) are available to guide this decision.

The incidence of thromboembolic events in the available investigations is poorly defined. In the population-based Framingham Heart Study,¹ the relative risk of stroke in individuals with HF was 4.1 for men and 2.8 for women, but many of these individuals had concurrent AF. In HF trials, annual stroke rates between 1.3% and 3.5% have been reported; however, all of these analyses included patients with AF. An analysis by Freudenberger et al. reported thromboembolic rates of only 1% per year in a population of patients in NYHA class II and III without AF.²

The risk of thromboembolic events can be related to patients' functional capacity, the degree of left ventricular dysfunction or to the presence of comorbid conditions that are prevalent in the HF population. A diseased ventricular myocardium might predispose patients to mural thrombus formation through mechanisms related to stasis or endocardial thrombogenesis, but pertinent information is, once again, limited. Patients in the V-HeFT trials³ who experienced ischemic events tended to have lower peak exercise oxygen consumption and a lower mean left ventricular ejection fraction than patients without ischemic events, but the differences were not statistically significant. In the SAVE trial,⁴ which involved survivors of myocardial infarction (MI), the relative risk of stroke among patients with an ejection fraction in the range 19–35% was 1.15, rising to 1.86 for those with an ejection fraction of less than 28%. Women with low ejection fractions were at a greater risk of stroke than comparable men. In the SOLVD trials,⁵ women (but not men) had a risk of thromboembolism that was inversely proportional to their ejection fraction (relative thromboembolism risk 4.2 for women with an ejection fraction below 10%). These analyses included patients with AF. A more recent analysis, excluding those with baseline AF, did not demonstrate a relationship between the degree of left ventricular dysfunction and thromboembolic events.²

In addition, conflicting analyses about the risk and benefits of prophylaxis exist. Half a century ago, three randomized studies of coumarin anticoagulants demonstrated a marked reduction in ischemic events with coumarin prophylaxis.⁶ These trials involved only hospitalized patients and included cases with rheumatic mitral valve disease and AF, for which the value of anticoagulation is now clear. Although anticoagulant therapy was not randomized in the V-HeFT trials, warfarin was notably not associated with stroke risk reduction. In two long-term studies of patients with MI, warfarin (target international normalized ratio [tINR] 2.8–4.8) was associated with 55% and 40% lower rates of stroke, compared with placebo.⁷ By contrast, another study demonstrated no significant reduction in the occurrence of stroke with anticoagulation at a similar intensity (tINR 2.7–4.5).⁷ It is important to bear in mind, however, that these studies did not target patients with chronic HF and included those with AF.

In trials reviewed by the Antithrombotic Trialists' Collaboration,⁸ allocation of high-risk patients to antiplatelet therapy reduced the combined rates of serious vascular events by about a quarter, nonfatal MI by a third, nonfatal stroke by a quarter, and vascular mortality by one-sixth. This meta-analysis led to the broad recommendation to use antiplatelet agents for secondary prevention of MI; however, the database included only two trials totaling 134 patients with HF. Despite the lack of data supporting aspirin efficacy, even in patients with ischemic cardiomyopathy, the recommendation has been generalized to patients with chronic HF, making some practitioners reluctant to allow inclusion of their patients in a trial comparing warfarin with aspirin for chronic HF. In fact, two studies, WARIS and ASPECT, found that warfarin reduced mortality (by 24%), MI (by 34–53%) and stroke (by 40–55%) over 3 years in patients with ischemic heart disease.⁷ The beneficial effects, as found by meta-analysis, were lower for aspirin, with reductions in vascular-related death, MI and stroke of 15%, 31% and 39%, respectively.⁹ A more recent meta-analysis by Leor et al.¹⁰ found no difference between warfarin and aspirin with respect to these endpoints in patients with coronary disease.

Aspirin has been implicated as a cause of adverse events in patients with HF, and there are no data to indicate that aspirin reduces the risk of thromboembolism in patients with this condition. Two recent studies, WASH and WATCH, demonstrated an increased need for hospitalization in patients with HF who received aspirin, perhaps owing to prostaglandin inhibition. In a retrospective analysis from SOLVD,⁹ patients taking aspirin had lower all-cause mortality than those who were not taking aspirin (hazard ratio 0.82). In addition, a cohort study of patients with coronary artery disease with or without HF being treated with angiotensin-converting-enzyme inhibitors, found that the use of aspirin was associated with lower mortality than angiotensin-converting-enzyme inhibitor treatment without aspirin.¹⁰ In a meta-analysis⁸ of patients with prior MI, the risk of stroke was reduced by 39% with aspirin, but no studies have addressed the effect of aspirin on stroke rate in patients with HF. Thus, the available data are insufficient to guide the use of aspirin in patients with HF.

The WATCH, WASH and WARCEF trials were developed to answer the important questions surrounding the use of anticoagulation in patients with HF. The WATCH trial, designed to compare warfarin, aspirin and clopidogrel in patients with HF, was stopped because of poor recruitment. Despite early termination, the study demonstrated a statistically significant increase in the hospitalization of patients assigned aspirin. Similarly, WASH, which was designed to test the feasibility of comparing open-label warfarin (tINR 2.5) with aspirin, demonstrated an increase in adverse secondary outcomes, particularly hospitalization, in those given aspirin. The WASH trial was a small pilot study, however, and was not powered to conclusively demonstrate such an outcome. In addition, as the larger WATCH trial was terminated prematurely, the finding of increased hospitalization (which was not a primary outcome) cannot be considered conclusive. The ongoing WARCEF trial is based on a double-blind, placebo-controlled design for primary and secondary prevention of thromboembolism in patients with HF or reduced left ventricular ejection fraction and is using mortality as the primary endpoint. Both WASH and WATCH suffered from patient recruitment well below projected goals and below that in other studies of HF treatments involving similar patient populations. Hopefully, the WARCEF trial will be completed and provide much needed data.

In the absence of such results, clinicians will be compelled to rely on extrapolation of observations in other cardiovascular disease states and on supposition to select appropriate antithrombotic treatment for patients with HF. We urge physicians to avoid succumbing to opinion in the absence of data and to support the successful completion of properly designed randomized trials.

Competing interests

The authors declared no competing interests.

ARTICLE NAVIGATION -ARTICLE TOOLS

Full text
• Download PDF
• Send to a friend
• Rights and permissions
• Order commercial reprints

ISSUE NAVIGATION - THIS ISSUE

• Table of Contents
• Previous article
• Next article