Continued versus interrupted aspirin use and bleeding risk after endoscopic submucosal dissection of gastric neoplasms: a meta-analysis

BACKGROUND: Balancing the risk of bleeding and thromboembolic events for patients who use aspirin and need to undergo endoscopic submucosal dissection (ESD) for gastric neoplasms is a delicate process. The current guidelines from different associations provide inconsistent recommendations. METHODS: MEDLINE and EMBASE databases were searched through August 2017 for studies that compared the risk of post-ESD bleeding in patients who continued aspirin vs. those who discontinued aspirin preoperatively. Pooled odds ratios (OR) and 95% confidence intervals (CI) were calculated using a random-effect model, generic inverse variance method. The between-study heterogeneity was quantified using the Q statistic and I(2). RESULTS: A total of five studies that included 700 patients were identified. Our meta-analysis could not demonstrate a significantly increased risk of post-ESD bleeding among the aspirin-continued group compared to the aspirin-interrupted group, the pooled OR being 1.81 (95%CI 0.85-3.83). The statistical heterogeneity was insignificant, with an I(2) of 25%. Nine thrombotic events occurred in the aspirin-interrupted group whereas none occurred in the aspirin-continued group. CONCLUSIONS: This meta-analysis could not demonstrate that continuation of aspirin significantly increases the risk of post-ESD bleeding. However, the analysis was restricted by the small sample size and the observational nature of the primary studies. Randomized controlled trials are still needed to clarify this risk.