At the 2019 European Society of Cardiology (ESC) Congress (31 August–4 September, Paris, France), Pavel Osmancik (Charles University and University Hospital Kralovske Vinohrady, Prague, Czech Republic) presented the results of PRAGUE-17. These showed that left atrial appendage closure (LAAC) was non-inferior to non vitamin-K-antagonists (NOACS) for the prevention of major cardiovascular and neurological events in patients with non-valvular atrial fibrillation. However, Osmancik observed that “safety issues” with LAAC remain.
He commented that both LAAC and NOACS have been found to have advantages over vitamin K antagonists (aka warfarin)—the traditional first-line therapy—for reducing thromboembolic events in patients with atrial fibrillation. Studies have indicated that LAAC is associated with reduced rates of intracranial haemorrhage (and a similar reduction in stroke) and that NOACS are associated with less intracranial haemorrhage, all stroke, and mortality. However, Osmancik added: “The relative safety and efficacy of LAAC vs. NOAC is unknown.”
In PRAGUE-17, 201 patients with non-valvular atrial fibrillation (at high risk of stroke) were randomised to undergo LAAC with either the Amplatzer Amulet device (Abbott) or the Watchman/Watchman Flex device (Boston Scientific); another 201 were randomised to receive a NOAC (apixaban preferably, but could also be rivaroxaban or dabigatran). In the LAAC group, patients received clopidogrel and aspirin (dual antiplatelet therapy) for three months before undergoing transoesophageal echocardiogram (TOE). At this time point, clopidogrel was discontinued but the aspirin was continued indefinitely. However, this regimen could be individualised (i.e. shortened) if the patient had a high risk of bleeding etc.
The primary endpoint was a component of stroke or transient ischaemic attack, systemic embolism, clinically significant bleeding, or significant periprocedural or device-related complications. Follow-up was at six weeks, three months, nine months, and 12 months (and then every six months). There were no significant differences between groups in the cumulative incidence function for the primary endpoint (p=0.004 for non-inferiority of LAAC vs. NOACS). This was the case for both the per-protocol populations and for the intention-to-treat populations.
However, there were nine complications in the LAAC group; this included one procedure-related death (groin haematoma) and one device-related death (late pericardial tamponade). Furthermore, of those originally assigned to LAAC, 7% did not undergo the procedure because of anatomical considerations, pericardial effusion, possible infective endocarditis, or patient refusal.
Osmancik concluded: “In the current NOAC era, LAAC may be considered with an adequate rationale for a non-pharmacological alternative; similar overall outcomes are expected with each of these strategies.” However, he added that safety issues remained with LAAC “warranting further refinements in both operator techniques and device technology.”