摘要

Objectives This meta-analysis aims to systematically measure the potential diagnostic value of anti-Helicobacter pylori IgG in urine for infection diagnosis, using all eligible studies published in English and Chinese languages. Design The random effect model was used to analyse the pooled sensitivity, specificity, positive likelihood ratio (PLR), negative LR (NLR), diagnostic OR (DOR), together with the summary receiver operator characteristic curve. Setting Literature searches of databases including PubMed, EMBASE, MEDLINE, Web of Science, Chinese National Knowledge Infrastructure and Wanfang databases were performed to retrieve studies evaluating the diagnostic value of urine IgG antibody for H. pylori infection. Primary outcome measure Twenty-three studies with 4963 subjects were included in the current meta-analysis. Results The pooled sensitivity, specificity, PLR, NLR, DOR and area under the curve (AUC) were 0.83 (95% CI 0.82 to 0.85), 0.89 (95% CI 0.88 to 0.90), 8.81 (95% CI 6.37 to 12.2), 0.13 (95% CI 0.09 to 0.2), 73 (95% CI 46.45 to 114.74) and 0.9551, respectively. Subgroup analyses showed that diagnostic accuracy of the urine IgG assay was no different in age, region, study population and assay method. Conclusions Anti-H. pylori antibody in urine might serve as a good marker in diagnosing H. pylori infection. However, further validation based on a larger sample is still required.