Incidence of gastrointestinal bleeding after transoesophageal echocardiography in patients with gastro-oesophageal varices: a systematic review and meta-analysis

Abstract Funding Acknowledgements Type of funding sources: None. Introduction Transoesophageal echocardiography (TEE) is useful for cardiac assessment and intraoperative monitoring. However, the safety of TEE in cirrhotic patients with gastro-oesophageal varices has remained uncertain. Purpose This meta-analysis aims to determine the incidence of gastrointestinal bleeding after TEE in patients with varices. The secondary objectives are to compare the bleeding risks between patients with and without varices; and to determine the incidences of TEE-related oesophageal perforation and mortality. Methods Systematic literature search was conducted on MEDLINE, EMBASE and Cochrane Database using the terms "Transoesophageal echocardiography", "Varices", "Bleeding", and related terms. Articles describing the incidence of post-TEE bleeding in patients with varices were included. Non-English articles were excluded. Risk of bias was assessed through a validated scale for bias assessment in prevalence studies. The level of evidence was assessed using the Grades of Recommendation, Assessment, Development, and Evaluation approach. Pooled weighted incidence of gastrointestinal bleeding and risk difference in bleeding were calculated with a random effects model. Results 569 articles were identified initially, and 10 articles (comprising of 908 patients) were included. The incidence of post-TEE bleeding in patients with varices was 0.84% (95% CI 0.34%-1.56%). When stratified by TEE indication, the pooled incidence of bleeding was 0.68% (95% CI 0.11%-1.63%) in intraoperative TEE; and 1.03% (95% CI 0.23%-2.29%) in diagnostic TEE. No cases of oesophageal perforation or mortality were reported. 6 studies included a comparator group of patients without varices, and the bleeding risk was comparable between patients with and without varices (risk difference 0.26%; 95% CI -0.80%-1.32%; I2 = 0%, p = 0.88). Eight studies had moderate or high risk of bias, and overall level of evidence was low. Conclusion TEE appears to be associated with low gastrointestinal bleeding incidence in patients with gastro-oesophageal varices. Nonetheless, results should be treated with caution due to bias and low level of evidence. Large-scale high-quality studies will be required to confirm the safety of TEE in patients with gastro-oesophageal varices. Abstract Figure. Study selection process Abstract Figure. Forest plot of pooled bleeding incidence