Stein et al[28] reported that upper gastrointestinal endoscopy had lower accuracy and predictive value than scintigraphy or selleck gastric pH monitoring in the assessment of DGR. We can find duodenogastric reflux under direct visualization in our daily clinical work. But this is only a temporary phenomenon for the most part, and not on behalf of the patient��s disease status. The chief source of error in this technique is the possible effect of the intubation in either promoting or hindering reflux. Therefore, endoscopic findings only give us an intuitive, subjectivity evidence of the bile reflux and the test is largely a qualitative one. In addition to the observation of DGR situation, we did pathological examinations during the routine gastroscopy examination.

It has been demonstrated in animal experiments that duodenal fluid caused an increase in inflammatory cells in the gastric mucosa, a decrease in parietal cells, a hyperplasia of mucous cells and changes in glandular morphology. The important factor is the antrum which serves to protect the mucosa of the gastric body from the toxic effects of DGR. In our research, we found the atrophic gastritis was very common in patients with severe reflux in endoscopy, and we didn��t found the histopathology significantly different between DGR and atrophic gastritis, which was consistent with the previously research[29]. The assessment of gastric fluid, an important work in the endoscopic progress, is another important impact in the diagnosis of DGR. The surfactant effect of bile acids is closely related to their hydrophobic-hydrophilic balance.

Bile acids have a surfactant effect for lipid absorption[30], and they may have a cytotoxic action if the surfactant effect is too strong[31,32]. Indeed, Heuman reported Brefeldin_A that the hydrophilic-hydrophobic balance of bile acids correlates with their toxicity, and increasing hydrophobicity was associated with increasing cytotoxicity towards the gastrointestinal epithelium[33]. Therefore, the bile acids may also have some roles in the formation of duodenogastric gastritis and in the diagnosis of DGR. In our study, we found there was a good correlation between TBA and DBIL, TBA and TBIL, DBIL and TBIL in DGR group. When we used Fisher��s linear discriminant analysis to analyze the three indexes in the determination of DGR, we found TBA was the most important factor in the diagnosis and created two formats to discriminant the diagnosis of DGR. The consistency between the direct vision of endoscopy and gastric juice examination was nearly 84%. By this method, the sensitivity and the specificity was separately 83.8% and 84.