Objective We aimed to evaluate the prevalence of H.pylori infection and the prevalence of cagA+ strains in patients with and without Barrett’s esophagus. Methods A full literature search to February 2008 was conducted in PubMed, MEDLINE and EMbase databases to identify case-control studies or cohort studies evaluating the prevalence of H.pylori in patients with or without Barrett’s esophagus. Summary odds ratios (OR) and 95% confidence interval (CI) were calculated by RevMan 4.2.8. Results Nineteen studies were identified (16 case-controlled studies and 3 cohort studies). In case controlled studies, the prevalence of H.pylori infection significantly decreased in patients with Barrett’s esophagus as compared subjects with normal endoscopic appearance, with a overall OR of 0.56 (95%CI 0.40 to 0.79). The prevalence of H.pylori infection was no statistically significant difference in patients with Barrett’s esophagus as compared to those with gastroesophageal reflux disease, with a overall OR of 0.86 (95% CI 0.74 to 1.00). In cohort studies, the prevalence of H. pylori was no statistically significant difference in patients with Barrett’s esophagus as compared to patients with normal endoscopic appearance or patients with gastroesophageal reflux disease, with a overall OR of 1.12 (95%CI 0.77 to 1.61) and 1.10 (95%CI 0.32 to 3.83). When the analysis was stratified by the status of cagA, the prevalence of cagA positive strains significantly decreased in patients with Barrett’s esophagus as compared both to subjects with normal endoscopic appearance with OR 0.30 and 95% CI 0.12 to 0.74, and to those with gastroesophageal reflux disease (OR 0.55; 95%CI 0.33 to 0.94). Irrespective of the presence of intestinal metaplasia, similar magnitude for the reduction of H.pylori infection was observed for patients with Barrett’s esophagus and those with normal endoscopic appearance. While accompared with the presence of intestinal metaplasia, Barrett’s esophagus was associated with a significantly reduction as compared to the patients with gastroesophageal reflux disease (OR 0.81, 95%CI 0.68 to 0.98). When stratified analyses were performed, a significant reduction of H.pylori infection was observed only in patients with long-segment Barrett’s esophagus (OR 0.54; 95%CI 0.35 to 0.82), but not in those with short-segment Barrett’s esophagus (OR 0.72; 95%CI 0.43 to 1.20). Conclusion This meta-analysis indicated that the prevalence of H.pylori infection, especially the prevalence of cagA positive strains was significantly lower in patients with Barrett’s esophagus than in subjects with normal endoscopic appearance. However, the prevalence of H. pylori infection was no statistical difference in patients with Barrett’s esophagus as compared to those with gastroesophageal reflux disease. Colonization with cagA positive strains may be protective against the formation of Barrett’s esophagus.
Objective To compare the risk of bone fractures in proton pump inhibitor users and nonusers, so as to evaluate the effects of proton pump inhibitors on the risk of bone fractures. Methods We searched PubMed, MEDLINE and EMbase databases to March 1st 2011 to identify case-control studies or cohort studies evaluating the risk of fracture in proton pump inhibitor users and nonusers. We conducted systematic review and meta-analysis according to the fracture site, duration of exposure, average daily dose and time of last use. Summary odds ratios (OR) and 95% confidence interval (CI) were calculated by RevMan 5.0.25 software. We also calculated and looked for heterogeneity. Results Eleven studies were identified from ten literatures, including seven case-control studies and four cohort studies. In case-control studies, the risk of total fractures increased by 36% in proton pump inhibitor users as compared with nonusers (OR=1.36, 95%CI 1.20 to 1.55). The risk of hip fracture increased by 39% (OR=1.39, 95%CI 1.13 to 1.71). In cohort studies, the risk of total fractures increased by 59% (OR=1.59, 95%CI 1.47 to 1.73). The risk of non-hip fractures increased by 65% (OR=1.65, 95%CI 1.47 to 1.85). As compared with nonusers, fracture risk increased by 41% in current users and by 38% in past users whose last use was at least 1 year ago. There was no significant difference between the two groups. Conclusion Proton pump inhibitors increase the risk of fracture to a certain degree. The effect does not fade away by discontinuation of PPI use for at least one year. Stricter clinical trials are needed to exclude confounding factors.