Anticoagulation May Not Be Necessary When Replacing Aortic Valve With Bioprosthesis, Meta-analysis Suggests

Anticoagulation with warfarin in the months following aortic valve replacement (AVR) with a bioprosthesis—a practice that receives a tepid recommendation in current guidelines—is associated with an elevated bleeding risk but appears to have no impact on thromboembolic events, a new meta-analysis shows.

Although long-term anticoagulation with warfarin is routinely used following implantation of mechanical prosthetic heart valves, what should be done with bioprosthetic valves is less clear.

Valvular heart disease guidelines from the American College of Cardiology and American Heart Association say anticoagulation with a vitamin K antagonist may be reasonable for the first 3 months after bioprosthetic AVR, but there is substantial variation in practice, according to Haris Riaz, MD (Cleveland Clinic, Cleveland, OH), and colleagues.

“The rationale for anticoagulation stems from microthrombi that are occasionally observed on valves during early postoperative imaging,” they note in a paper published online May 10, 2016, ahead of print in Circulation: Cardiovascular Quality and Outcomes. But, they say, “many investigators have questioned this approach in the absence of other risk factors predisposing to thromboembolism, and anticoagulation for bioprostheses is not generally practiced.”

Recently, however, attention has increased on the potential need for anticoagulation after bioprosthetic AVR because of the release of data showing that anticoagulation is associated with a lower rate of reduced leaflet motion, which is possibly related to subclinical valve thrombosis, in patients undergoing either TAVR or surgical AVR.

To explore the issue, Riaz and colleagues pooled data from 13 studies of patients who underwent bioprosthetic AVR, including 6,431 patients who received warfarin anticoagulation and 18,210 who received aspirin or placebo. All of the studies were observational (four retrospective and nine prospective). The duration of anticoagulation was 3 months in most of the studies, although it ranged from 1 month to 2 years.

Warfarin use was associated with a greater likelihood of overall bleeding (OR 1.38; 95% CI 1.07-1.78) and a trend toward more bleeding at 3 months (OR 1.26; 95% CI 0.97-1.64) compared with aspirin or placebo.

That was not accompanied by a lower rate of the composite of venous thromboembolism, stroke, or transient ischemic attack either at 3 months (OR 1.01; 95% CI 0.56-1.84) or overall (OR 1.08; 95% CI 0.61-1.91).

“This is a critically important clinical finding that is contrary to current guideline recommendations,” the authors write.

Questions Remain

The researchers acknowledge that their findings are limited, however, because they are based on observational data subject to selection bias and because there was substantial heterogeneity across studies.

It also remains unknown whether the results are applicable to patients undergoing TAVR, they say. “Most investigators think that transcatheter valves are comparable [to] surgically implanted valves and that potential mechanisms of thrombosis should be similar,” they write.

The best way to determine whether anticoagulation should be used following TAVR or surgical AVR “would be with a randomized controlled trial that would ideally be triple armed (antiplatelet versus anticoagulant versus placebo),” they add. “Until that time, our analysis suggests that anticoagulation of bioprosthetic valves in the absence of other indications for thrombosis prevention seems to be a dangerous practice.”

Josep Rodés-Cabau, MD (Laval University, Quebec City, Canada), who commented on the study for TCTMD, said the findings of the meta-analysis are not very surprising, adding that a study would need larger numbers of patients to demonstrate a difference in thromboembolic events, which are infrequent following AVR. Although anticoagulation probably does have some beneficial effects, he said, “we have to see whether or not these effects will overcome the bleeding problems.”

Further, adequately powered studies are needed to address that question, Rodés-Cabau said. But in the meantime, he added, the need for anticoagulation perhaps should be made on a case-by-case basis. “Maybe this should be more individualized and restricted to patients with a relatively low bleeding risk,” he suggested.

Related Stories:

• Post-TAVR Thrombus May Be More Common Than Thought, Often Subclinical
• Reduced Leaflet Motion in Bioprosthetic Aortic Valves Not Uncommon But Remedied by Anticoagulation
• New Findings on Bioprosthetic Valve Thrombosis Have Implications for Surveillance, Research