Risk of Thromboembolism in Heart Failure


Synopsis: In patients with reduced EF, the risk of thromboembolic events is low (2%), and EF is independently associated with these events only in women.

Source: Dries DL, et al. J Am Coll Cardiol 1997; 29:1074-1080.
The risk of thromboembolic events in heart failure patients is difficult to interpret because of confounders such as atrial fibrillation and left ventricular aneurysm. However, the risk of thromboembolic complications is important because of its therapeutic implications. Thus, Dries and colleagues performed a retrospective analysis of the 6378 patients in the Studies of Left Ventricular Dysfunction (SOLVD) in sinus rhythm at the time of enrollment to test the hypothesis that LV dysfunction is an independent risk factor for embolic events. The events studied included strokes, pulmonary emboli, and peripheral emboli, and the analyses were stratified by gender. The annual incidence of thromboembolic events over the 40-month average follow-up was 2.4% in women and 1.8% in men.

In both groups, the majority of events were strokes, but women had more pulmonary emboli than men (24 vs 14%; P = 0.01). Univariate analysis showed that ejection fraction (EF), history of hypertension, diabetes, history of stroke, and drug therapy were associated with thromboembolic risk. Multivariate analysis showed that a decline in EF was an independent risk factor in women but not men. Historical factors such as hypertension, diabetes, and prior stroke increased the risk in men. Diabetes (increased incidence) and antiplatelet drugs (decreased incidence) were also important in women. When pulmonary emboli were excluded, which were more common in women, the association with EF was reduced to a trend (P = 0.07). Dries et al conclude that in patients with reduced EF, the risk of thromboembolic events is low (2%) and EF is independently associated with these events only in women.


The Agency for Health Care Policy and Research (AHCPR) Guidelines on the management of heart failure (1996) clearly stated that routine anticoagulation use is not recommended unless other specific indications are present, such as atrial fibrillation. Their reason was the absence of prospective randomized data showing the efficacy of anticoagulant therapy to prevent emboli in patients with left ventricular dysfunction. Such data still do not exist, but this study sheds light on the issue by defining the incidence of embolic events in patients with moderate-to-severe left ventricular dysfunction in normal sinus rhythm. An overall 2% per year incidence is not impressive, especially since only about two-thirds of the events were strokes. Thus, these data do not seriously challenge the AHCPR recommendations.

Many cardiologists prescribe anticoagulation for heart failure patients in sinus rhythm under the belief that the lower the EF, the worse the risk of stroke; adjusting for other covariates, they arrive at a risk/benefit ratio decision that is usually favorable for patients with the lowest EFs (< 20%). Surprisingly, this hypothesis that the lower the EF, the higher the risk was not proven in this study in the men, who comprised most of the patients (86%). There was only a trend toward proving the hypothesis in women when the risk of systemic emboli was considered (P = 0.07). If pulmonary emboli, which were more common in women, are included, there was a relation to EF in women, but prevention of pulmonary emboli in women is not the usual rationale for the practice of prescribing anticoagulants in heart failure patients.

There are several limitations to this study in addition to being a retrospective, observational study. Cerebrovascular events were not required to be fully evaluated with CAT scans, etc., so we don’t know how many were truly embolic. Since the women were older and had a higher incidence of hypertension and diabetes, perhaps more of their cerebral events were non-embolic, which would further weaken the EF hypothesis in them. Also, the presence of sinus rhythm and drug therapy were only categorized at intake. No adjustment for changes over time was made. In addition, for those on warfarin (about 10%) no INR data are available. Perhaps anticoagulant therapy was not associated with embolic event incidence because it was ineffective. Finally, there was no breakdown of dilated cardiomyopathy vs. ischemic or whether aneurysms were present. The new ACC/AHA practice guidelines for acute myocardial infarction lists patients with extensive wall motion abnormalities as a class IIa indication for anticoagulation. (Ryan TJ, et al. Circulation1996;94:2341-2350.) Also, these guidelines list patients with severe LV dysfunction or heart failure as class IIb indications for anticoagulation. Clearly, some experts believe the ischemic LV dysfunction patient is different with regard to anticoagulant therapy.

Atrial fibrillation trials have shown a significant bleeding rate on oral anticoagulants of about 1% a year, which, compared to the estimated stroke risk of 5% a year, seems to be an acceptable risk. In the case of LV dysfunction, the stroke risk is less than 2% annually, and we do not know if bleeding complications would be similar in heart failure patients. In fact, alterations in hepatic flow could make bleeding complications more common in heart failure patients. Thus, in the absence of a prospective therapeutic trial, there seems to be little to recommend anticoagulant therapy as a routine treatment in heart failure patients. Whether a high risk group could be culled out, such as patients with diabetes, hypertension, prior stroke, and very low EF, is unclear. At present, the only definite indications for anticoagulation in patients with heart failure, are indications independent of heart failure such as atrial fibrillation, presence of LV thrombus, history of pulmonary embolus, etc.