Non-invasive diagnostic tests for Helicobacter pylori infection

摘要: BACKGROUND: Helicobacter pylori (H pylori) infection has been implicated in a number of malignancies and non-malignant conditions including peptic ulcers, non-ulcer dyspepsia, recurrent ulcer bleeding, unexplained iron deficiency anaemia, idiopathic thrombocytopaenia purpura, colorectal adenomas. The confirmatory diagnosis H is by endoscopic biopsy, followed histopathological examination using haemotoxylin eosin & E) stain or special stains such as Giemsa Warthin-Starry stain. Special are more accurate than E There significant uncertainty about the diagnostic accuracy non-invasive tests for pylori. OBJECTIVES: To compare urea breath test, serology, stool antigen used alone combination, symptomatic asymptomatic people, so that eradication therapy can be started. SEARCH METHODS: We searched MEDLINE, Embase, Science Citation Index National Institute Health Research Technology Assessment Database on 4 March 2016. screened references included studies to identify additional studies. also conducted citation searches relevant studies, most recently December did not restrict language publication status, whether data were collected prospectively retrospectively. SELECTION CRITERIA: evaluated at least one index (urea test isotopes as13C or14C, serology test) against reference standard (histopathological stain, immunohistochemical stain) people suspected having infection. DATA COLLECTION AND ANALYSIS: Two review authors independently extracted data. assessed methodological quality QUADAS-2 tool. performed meta-analysis hierarchical summary receiver operating characteristic (HSROC) model estimate SROC curves. Where appropriate, we bivariate univariate logistic regression models sensitivities specificities. MAIN RESULTS: 101 involving 11,003 participants, which 5839 participants (53.1%) had prevalence ranged from 15.2% 94.7%, with median 53.7% (interquartile range 42.0% 66.5%). Most (57%) dyspepsia 53 excluded who proton pump inhibitors antibiotics.There was an unclear risk bias applicability concern each study.Of 15 compared two three tests. Thirty-four (4242 participants) serology; 29 (2988 test; 34 (3139 test-13C; 21 (1810 test-14C; (127 but report isotope used. thresholds define positivity staining techniques (reference standard) varied between Due sparse threshold reported, it possible best test.Using 99 indirect comparison, there statistical evidence difference test-13C, test-14C, (P = 0.024). odds ratios 153 (95% confidence interval (CI) 73.7 316), 105 CI 74.0 150), 47.4 25.5 88.1) 45.1 24.2 84.1). sensitivity CI) estimated fixed specificity 0.90 (median across four tests), 0.94 0.89 0.97) 0.92 0.94) 0.84 0.74 0.91) 0.83 0.73 0.90) test. This implies average, given studies), out 1000 tested infection, will 46 false positives (people without diagnosed infection). In this hypothetical cohort, give 30 58), 42 86 50 140), 89 52 146) negatives respectively whom missed).Direct comparisons based few head-to-head (DORs) 0.68 0.12 3.70; P 0.56) test-13C versus (seven 0.88 0.14 5.56; 0.84) studies). 95% CIs these estimates overlap those DORs comparison. Data limited unavailable other direct comparisons. AUTHORS' CONCLUSIONS: history gastrectomy have antibiotics ,pump inhibitors, high while less infection.This comparison (with potential due confounding), unavailable. highly variable unable specific might useful clinical practice.We need further comparative obtain reliable relative Such should representative spectrum clearly reported ensure low bias. importantly, prespecify used, avoid inappropriate exclusions.