Of all the medications in my clinical repertoire, none elicits more groans from my patients than warfarin. As soon as I mention it, most of them are vaguely aware of what’s coming: burdensome INR checks, frequent dose adjustments, dietary restrictions, and risk for bruising or something worse. In short, warfarin’s bruised reputation precedes itself. That’s why results from the PROTECT AF trial, published in Circulation, piqued my interest.
In this trial, 707 patients with nonvalvular atrial fibrillation (AF) and at least one CHADS2 risk factor were randomized to undergo percutaneous left atrial appendage (LAA) closure with the Watchman filter device or to receive ongoing warfarin therapy. The primary composite endpoint was the incidence of stroke, systemic embolism, or cardiovascular death. Mean follow-up was 2.3 years. The device maker funded the trial.
The compelling results:
1. LAA closure with the Watchman device was noninferior to warfarin. Incidence of the primary endpoint, per 100 patient-years, was 3.0% in the device group and 4.3% in the warfarin group (relative risk, 0.71; 95% CI, 0.44%–1.30%).
2. The device’s failure to achieve superiority was related to acute, procedure-related stroke events. The primary adverse-outcome rate was 5.5% for Watchman and 3.6% for warfarin (RR, 1.53; 95% CI, 0.95–2.70); most of the events occurred early. Otherwise, rates remained lower in the device group than in the warfarin group throughout follow-up.
3. Superiority of the device was found in secondary analyses aimed at isolating the effect of LAA closure from that of transient concomitant antithrombotic therapy. The adverse events, mainly bleeding, were similar in the two groups while on antithrombotic therapy. However, after patients completed interim antithrombotic therapy (warfarin for 45 days, then clopidogrel for 4.5 months after device deployment), major bleeding was significantly less common with the device than with warfarin (RR, 0.35; 95% CI, 0.16–0.79).
4. LAA closure had a sustained benefit in secondary prevention of stroke. The event rate was 5.3% for Watchman versus 8.2% for warfarin (RR 0.64; 95% CI, 0.24–1.74)
Given these results, we cardiologists now have three options to offer patients with nonvalvular AF: warfarin, novel anticoagulants (dabigatran, rivaroxaban, and apixaban), and the Watchman device. As a consultative cardiologist, I have three main thoughts about this new reality.
First, I find the prospect of avoiding frequent INR checks, constant dosage adjustments, and concern about sub- and supratherapeutic INR levels very attractive. However, we can avoid these problems using the novel anticoagulants without exposing patients to the risks of an invasive procedure. Thus, if avoiding the hassles of warfarin administration is the objective, I will have a frank discussion with my patients about treatment risks and benefits, as well as their own expectations. I suspect some patients will shy away from the procedure while others will do anything to avoid taking another pill. So I’m not sure which option will ultimately be favored most often.
Second, I foresee a potential niche for the Watchman device (not unlike the niche for bioprosthetic valves) in patients at high risk for bleeding. Consider the patient scheduled for an elective surgery that can be delayed until after the 6-month anticoagulation period; or the young patient who wishes to avoid long-term anticoagulation; or someone whose very active lifestyle predisposes him to potential trauma.
Third, we need further exploration of the risk of performing direct cardioversion (DCV) in people who undergo the Watchman procedure. Will DCV be safe in a patient off warfarin with a Watchman device in place? Will a patient who requires frequent DCV need to repeatedly take warfarin for brief periods?
The Watchman device now has a class IIb recommendation in the 2012 focused update of the European Society of Cardiology’s atrial fibrillation guidelines. Given the data from PROTECT AF and this new recommendation, I pose three questions to you:
1) Will you begin using the Watchman device as an alternative to warfarin in most patients with nonvalvular AF?
2) Will you use the device in your patients with contraindications to warfarin, a group not evaluated in this trial?
3) Will you let your patients make the Watchman vs. warfarin choice themselves?