S595 Efficacy and Safety of the Push and Pull Method for Treatment of Food Bolus Impaction: A Systematic Review and Meta-Analysis

Abstract

Introduction: Esophageal food impactions (EFI) contribute significantly to morbidity and health expenditures. Professional recommendations for endoscopic management have advised a pull or bolus extraction method, noting a risk of perforation with the alternative push or gastric advancement method. Recent studies have suggested non-inferior safety of the push method compared to traditional piecemeal extraction. We sought to systematically compare the rate of significant adverse events of the push and pull techniques for EFI. Methods: MEDLINE and Embase were searched from inception to September 2021. Studies with over five adult participants that reported endoscopic outcomes for EFI were selected. The primary outcomes were success and adverse event rates of endoscopic foreign body removal (FBR) via the pull vs push method. All outcomes were assessed with pooled event rates (ER) and 95% confidence intervals (CI) using a random-effects model, and groups were compared in a mixed-analysis model, with p < 0.05 considered significant. Results: After reviewing 1567 publications, 17 studies with 3296 patients were included. Of these, 1522 cases were treated with the push method and 983 with the pull method. There were no significant differences in demographic data between the two groups. Our analysis of overall adverse events showed that the pull method had ER of 0.046 (95% CI 0.032-0.067), and push ER of 0.058 (95% CI 0.044-0.076), p= 0.798 for the difference. There was a trend towards increased success rates with the push method [pull 0.860 (95% CI 0.749-0.927), push 0.933 (95% CI 0.982-0.997), p= 0.185]. ER of perforation was 0.009 in the pull group (95% CI 0.004-0.019) and 0.012 in the push group (95% CI 0.007-0.023), p= 0.507. ER of bleeding after pull was 0.039 (95% CI 0.024-0.062) and push 0.057 (95% CI 0.040-0.079). For aspiration, the pull ER was 0.024 (95% CI 0.013-0.042) and push 0.014 (95% CI 0.008-0.025), p= 0.536. Overall success rate for all studies for endoscopic treatment of EFI was 0.985 (95% CI 0.935 - 0.996). (Table) Conclusion: Endoscopic therapy for relieving EFI is effective with low rates of adverse events overall. Based on limited data available, there was a trend for increased success rates with the push technique, and a trend for lower adverse events with the pull method. To our knowledge, this is the first meta-analysis to demonstrate that the push method for treatment of EFI is non-inferior to the pull method with respect to both success rates and safety profile. Table 1. - Pooled event rates of outcomes among patients undergoing treatment of esophageal food impaction with the push vs pull method Outcome Event Rate 95% Confidence Interval No. of Studies (No. of patients) I Squared Value P-value Success, overall 0.985 (0.938-0.996) 14 (2978) 0 Adverse events, overall 0.009 (0.005-0.015) 17 (3296) 0 Success, push 0.993 (0.982-0.997) 7 (1479) 81.151 0.185 Success, pull 0.860 (0.749-0.927) 7 (1479) 81.151 0.185 Adverse events, push 0.058 (0.044-0.076) 14 (2908) 57.927 0.798 Adverse events, pull 0.046 (0.032-0.067) 14 (2908) 57.927 0.798 Perforation, push 0.012 (0.007-0.023) 11 (2127) 0 0.507 Perforation, pull 0.009 (0.004-0.019) 11 (2127) 0 0.507 Aspiration, push 0.014 (0.008-0.025) 10 (1915) 0 0.536 Aspiration, pull 0.024 (0.013-0.042) 10 (1915) 0 0.536 Bleeding, push 0.057 (0.040-0.079) 10 (1686) 63.013 0.949 Bleeding, pull 0.039 (0.024-0.062) 10 (1686) 63.013 0.949