Extended follow-up of the PRAGUE-17 trial suggests left atrial appendage closure (LAAC) remains noninferior to direct-acting oral anticoagulants (DOACs) with regard to major cardiovascular and neurologic events in high-risk patients with atrial fibrillation.
At a median follow-up of 3.5 years, the annualized rate of the primary outcome – a composite of stroke, transient ischemic attack (TIA), systemic embolism, cardiovascular death, clinically relevant bleeding, or significant procedure- or device-related complications – was 8.6% in patients who underwent LAAC and 11.9% in those managed with DOACs (P value for noninferiority = .006).
The study was not powered to assess the individual components, but most were similar between the LAAC and DOAC groups, including cardiovascular death (20 vs. 30 events) and all stroke/TIA (16 vs. 18 events).
Nonprocedural clinically relevant bleeding was lower with LAAC (23 vs. 40 events; annualized rate, 3.4% vs. 5.9%; P = .039), said Pavel Osmancik, MD, PhD, Charles University and University Hospital Kralovske Vinohrady, both in Prague.
The data were presented at the Transcatheter Cardiovascular Therapeutics annual meeting, held virtually and live in Orlando and sponsored by the Cardiovascular Research Foundation, and was published simultaneously in the Journal of the American College of Cardiology.
The results are generally in line with those reported in 2020 with an average follow-up of 20 months, when the annualized primary event rate was 11% with LAAC and 13% with DOACs, also known as novel OACs (NOACs).
The trial randomly assigned 415 patients to LAAC with the Amulet (Abbott Vascular) or Watchman/Watchman FLX devices (Boston Scientific) or to rivaroxaban, dabigatran, or preferably apixaban (96%). The modified intention-to-treat analysis included 201 patients in each group, with follow-up extending to 4.3 years in the LAAC group and 4.2 years in the DOAC group.
Dr. Osmancik said the trial enrolled a very-high-risk atrial fibrillation cohort, citing a CHA2DS2-VASc score of 4.7 in both groups and a HAS-BLED score of 3.0-3.1. More than half of the LAAC group (54.2%) and 47.3% of the DOAC group had a history of bleeding or bleeding predisposition.
During a discussion of the results, the panel questioned whether the continuing divergence of the primary event curves at 4 years was potentially related to the effect of noncompliance to the NOACs over time.
Dr. Osmancik replied: “We didn’t do any medication look among the patients, but I don’t think that the number of patients who stopped the NOAC treatment was too high because the rate of strokes was very similar to that in the NOAC trials.”
He reported that 26 patients in the DOAC group permanently stopped their DOAC during follow-up; 15 (58%) because of clinically relevant bleeding, and 13 crossed over to LAAC. Of the 13 patients, 12 cases were successful with dual antiplatelet therapy for 3 months.
In the LAAC group, 17 patients started a DOAC during follow-up. Of these, three (18%) initiated DOAC treatment because of device-related thrombus (DRT) on transesophageal echocardiography, three (18%) because of a peridevice leak (PDL), and five (29%) because of a stroke/TIA.
“Given the observed trend of significantly increased nonprocedural clinically relevant bleeding in the NOAC arm, it is likely that late bleeding events will increasingly favor LAAC over time,” Faisal Merchant, MD, Emory University, Atlanta, wrote in an accompanying editorial.
NOACs, he noted, have important indications beyond prevention of left atrial appendage thrombi, including prevention of non-LAA sources of stroke/systemic embolism (SSE) and treatment of venous thromboembolism. “If significant numbers of patients treated with LAAC end up on anticoagulation in the long run, the benefits of LAAC are likely to be attenuated.”
Although PRAGUE-17 provides some insights into the longer-term indications for resuming anticoagulation in patients previously treated with LAAC, Dr. Merchant said the trial is a “real missed opportunity” in terms of understanding late device-associated risks. Unfortunately, two-thirds of the follow-up transesophageal echocardiograms were canceled because of the COVID-19 pandemic.
“Although the incidence of late DRT and PDL isn’t known, the longer-term PRAGUE-17 data are helpful in demonstrating that rates of SSE remain similar in the LAAC and NOAC groups over time, without any obvious signal of late ischemic events in the LAAC group,” he wrote.
The editorialist also called attention to the “often overlooked” issue of aspirin adherence in long-term medical therapy. Although patients treated with LAAC typically remain on aspirin indefinitely, the percentage who discontinue long-term aspirin is not well described and is not reported in PRAGUE-17. In the AVERROES trial, comparing aspirin with apixaban in patients with atrial fibrillation, however, 20.5% of patients permanently discontinued aspirin at 2 years, compared with only 17.9% on apixaban.
“It is plausible that discontinuation of aspirin may contribute to late ischemic events in patients treated with LAAC, potentially by increasing the risk of late DRT or through other mechanisms,” Dr. Merchant wrote. “Adherence to, and the impact of, long-term antiplatelet therapy should be a focus of future LAAC studies.”
The study was funded by a research grant from the Ministry of Health, Czech Republic. Dr. Osmancik reported occasional speaking honoraria from Bayer and Abbott. Dr. Merchant disclosed no relevant financial relationships.
A version of this article first appeared on.