P458 Diagnostic accuracy of patency capsule and cross-sectional imaging for capsule endoscopy retention: A systematic review and meta-analysis

Abstract

Abstract Background Although small bowel capsule endoscopy (SBCE) has been widely used to evaluate small bowel lesions including Crohn’s disease, there is a risk of capsule retention that can cause obstructive symptoms potentially necessitating surgical treatment. We conducted a systematic review and meta-analysis to evaluate the accuracy of patency capsule (PC) and cross-sectional imaging (CSI) for predicting capsule retention. Methods MEDLINE/PubMed, EMBASE, Web of science, and the Cochrane library databases were searched for articles that invested the diagnostic accuracy of PC or CSI for predicting capsule retention up to August 15, 2023. Pooled sensitivity and specificity were calculated using a bivariate random-effects model. Considering that SBCE was generally not performed when capsule retention was suspected on PC, the false-negative rates of PC and CSI were separately analyzed and compared using a single-proportion meta-analysis based on random-effects modeling. Quality assessment of the selected studies was performed using the revised Quality Assessment of Diagnostic Accuracy Studies-2 tool. Results A total of 22 peer-reviewed articles, including 17 articles for PC (covering 2,234 patients) and 7 articles for CSI (covering 302 patients) were included in this systematic review. The pooled sensitivity and specificity for predicting capsule retention with CSI were 64% [95% confidence interval (CI), 31%–87%] and 85% [95% CI, 62%–95%], respectively. The pooled sensitivity and specificity for predicting capsule retention with PC were 75% [95% CI, 43%–92%] and 94% [95% CI, 90%–96%], respectively (Figure 1, Table 1). Although there was no statistically significant difference in prediction accuracy between PC and CSI on the bivariate random-effects model (P=0.14), the pooled false-negative rate was significantly lower in the PC compared to CSI (2.5% [95% CI, 1.4%–4.7%] vs. 11.0% [95% CI, 5.2%–21.8%]; P=0.003) (Table 1). Conclusion Cross-sectional imaging showed good specificity, but showed moderate sensitivity in predicting capsule retention. In contrast, patency capsule showed good sensitivity and excellent specificity, with a significantly lower false-negative rate compared to cross-sectional imaging. Therefore, in patients with clinically suspicious small bowel stenosis, the patency capsule is a more reliable modality for predicting capsule retention.