Effect of pretreatment antibiotic resistance to metronidazole and clarithromycin on outcome of Helicobacter pylori therapy: a meta-analytical approach.

Our purpose was to define the effect of pretreatment Helicobacter pylori resistance to metronidazole or to clarithromycin on the success of antimicrobial therapy. We used 75 key words to perform a literature search in MEDLINE as well as manual searches to identify clinical treatment trials that provided results in relation to H. pylori susceptibility to metronidazole and clarithromycin or both during the period 1984-1997 (abstracts were not included). Meta-analysis was done with both fixed- and random-effect models; results were shown using Galbraith's radial plots. We identified 49 papers with 65 arms for metronidazole (3594 patients, 2434 harboring H. pylori strains sensitive to metronidazole and 1160 harboring resistant strains). Metronidazole resistance reduced effectiveness by an average of 37.7% (95% CI = 29.6-45.7%). The variability in the risk difference for metronidazole was 122.0 to -90.6 and the chi-square value for heterogeneity was significant (P<0.001). Susceptibility tests for clarithromycin were performed in 12 studies (501 patients, 468 harboring H. pylori strains sensitive to clarithromycin and 33 harboring resistant strains). Clarithromycin resistance reduced effectiveness by an average of 55% (95% CI = 33-78%). We found no common factors that allowed patients to be divided into subgroups with additional factors significantly associated with resistance. In conclusion, metronidazole or clarithromycin pretreatment resistant H. pylori are the main factors responsible for treatment failure with regimens using these compounds. If H. pylori antibiotic resistance continues to increase, pretherapy antibiotic sensitivity testing might become necessary in many regions.