Background: Left atrial appendage (LAA) occlusion is offered for stroke prevention in patients with atrial fibrillation (AF) who are unsuitable for lifelong oral anticoagulation. Medium to long term follow up (FU) data for this high-risk cohort are sparse.
Methods: We report 4-year FU outcomes in 73 patients (median age 78 [73-82] years; 34.2% female) who underwent successful LAA occlusion with the Amulet device at our centre between November 2014 and March 2017.
Results: Median CHA2DS2-VASc and HAS-BLED scores were 4 (4-5) and 2 (2-3), respectively; 69 (94.5%) patients had a previous major haemorrhagic event (intra-cranial haemorrhage in 57.5%, gastro-intestinal bleed in 27.4% and other sites 9.6%). Other co-morbidities included hypertension (71.2%), diabetes mellitus (37.0%), congestive cardiac failure (17.8%), coronary artery disease (11.0%), peripheral vascular disease (11.0%) and carotid artery disease (2.7%). Imaging at FU was performed in 67 (91.8%) patients, including trans-oesophageal echocardiography in 54 and contrast-enhanced cardiac computed tomography in 13 patients. 14 (19.2%) had evidence of a minor peri-device leak of <5 mm, and none had a major leak of >5 mm. One (1.4%) patient had a device-related thrombus soon after the procedure that resolved with a short-course of apixaban, with no late sequelae. Over the median FU period of 46 (19-56) months, seven patients suffered an ischaemic stroke or transient ischaemic attack and three suffered from haemorrhagic stroke. Five (6.8%) patients suffered from a major extracranial bleeding event, including three gastro-intestinal bleeds, one haemoptysis and one epistaxis. A total of 29 (39.7%) patients died during this period (Figure). The antithrombotic regime in the remainder (n=44) was aspirin only in 16 (36.4%), clopidogrel only in 5 (11.4%), apixaban in 2 (4.5%), dual-antiplatelet therapy in 1 (2.3%) and no antithrombotic in 20 (45.5%). The rate of ischaemic stroke with LAA occlusion was significantly lower than that predicted without treatment based on the CHA2DS2-VASc score (9.6% vs 24.9%). Using multivariable analyses, independent predictors of mortality were age >80 years (HR 2.42 [95% CI, 1.03 – 5.68]) and diabetes mellitus (HR 2.66 [95% CI, 1.13 – 6.25]), after adjusting for other risk factors.
Conclusion: Left atrial appendage occlusion using the Amulet device appears to be effective at reducing the risk of ischaemic stroke in high-risk AF patients who are deemed unsuitable for oral anticoagulation. These patients have a high rate of mortality over the medium term, and an ongoing risk of thrombotic and bleeding events. This has cost efficacy implications, and underscores the need for careful patient selection, especially in patients >80 years of age and those with diabetes mellitus.
