Left atrial appendage isolation and thrombotic and thromboembolic outcomes in atrial fibrillation: A systematic review and meta-analysis

BACKGROUND Electrical left atrial appendage isolation (LAAI) may be accompanied by impaired appendage function and thromboembolic complications. We evaluated the association between LAAI and clinical stroke or transient ischemic attack (TIA), while separately examining imaging-detected left atrial appendage (LAA) thrombus and related supportive outcomes. METHODS PubMed, Embase, and the Cochrane Central Register of Controlled Trials were searched from inception to December 6, 2024 for comparative clinical studies reporting thrombotic or thromboembolic outcomes after LAAI. Clinical stroke or TIA was the primary outcome. Risk ratios (RRs) were pooled using an inverse-variance random-effects model with restricted maximum-likelihood estimation and Hartung-Knapp-adjusted confidence intervals (CIs). The preferred primary analysis excluded the smaller Rillig cohort because of possible participant overlap with the Heeger cohort. Imaging-detected LAA thrombus, anticoagulation-at-event data, and LAA functional findings were evaluated separately. RESULTS Five comparative observational studies were included. The preferred primary analysis comprised four studies and 3876 patients, including 801 who underwent LAAI and 3075 controls. Clinical stroke or TIA occurred in 42 and 66 patients, respectively. LAAI was associated with a higher cumulative risk of clinical stroke or TIA (RR = 4.96, 95% CI 1.10-22.42; P = 0.043; I2 = 55.6%). The 95% prediction interval was 0.21-115.69. Inclusion of the potentially overlapping Rillig cohort yielded an RR of 5.26 (95% CI 1.80-15.36; P = 0.013), but the prediction interval remained compatible with no association. Leave-one-out analyses were directionally consistent, although statistical significance was lost after omission of Kim, Heeger, or Romero. Comparative LAA thrombus estimates were available only from Rillig (RR = 22.31, 95% CI 1.34-370.43) and Heeger (RR = 10.31, 95% CI 2.50-42.56) and were not pooled because potential cohort overlap could not be excluded. Anticoagulation-at-event and LAA functional data were unsuitable for quantitative synthesis. CONCLUSIONS The available evidence suggests that LAAI may be associated with a higher cumulative risk of clinical stroke or TIA. However, the extremely wide prediction interval crossing the null and the loss of statistical significance after omission of several individual studies indicate considerable uncertainty regarding the magnitude and consistency of this potential association across different practice settings. Comparative imaging findings also suggested a higher frequency of postprocedural LAA thrombus, although the available evidence was limited and potentially overlapping. Independent comparative studies with standardized TEE surveillance, clearly defined anticoagulation exposure, and complete clinical and imaging outcome reporting are required.