Diagnostic Performance of Conventional X-ray for Detecting Foreign Bodies in the Upper Digestive Tract: A Systematic...

Foreign body (FB) ingestion is a common clinical problem in acute settings. Detecting FBs in the upper digestive tract is challenging. The conventional X-ray is usually the first-line imaging tool to detect FBs. However, its diagnostic performance is inconsistent in the literature. In this study, we performed a systematic review and meta-analysis to determine the diagnostic performance of the conventional X-ray for detecting FBs in the upper digestive tract. We conducted a systematic search of PubMed, Embase, Cochrane Library, Web of Science, and Scopus until 1 August 2020. Prospective or retrospective studies investigating the diagnostic accuracy of conventional X-rays for detecting FBs in the upper digestive tract in patients of all ages were included. The Quality Assessment of Studies of Diagnostic Accuracy-2 tool was used to review the quality of included studies. We used a bivariate random-effects model to calculate diagnostic accuracy parameters. Heterogeneity was assessed using I² statistics. We included 17 studies (n = 4809) in the meta-analysis. Of the 17 studies, most studies were rated as having a high risk of bias. Conventional X-rays had a pooled sensitivity of 0.58 (95% confidence interval [CI] = 0.36-0.77, I² = 98.52) and a pooled specificity of 0.94 (95% CI = 0.87-0.98, I² = 94.49) for detecting FBs in the upper digestive tract. The heterogeneity was considerable. The area under the summary receiver operating characteristic curve was 0.91 (95% CI = 0.88-0.93). Deek's funnel plot asymmetry test results revealed no significant publication bias (p = 0.41). The overall sensitivity and specificity of conventional X-rays were low and high, respectively, for detecting FBs in the upper digestive tract. Hence, conventional X-rays to exclude patients without upper FBs in the digestive tract are not recommended. Further imaging or endoscopic examinations should be performed for at-risk patients.