Clarithromycin resistance of Helicobacter pylori is relatively frequent in France and is assumed to be the main cause of failure of the proton pump inhibitor-amoxicillin-clarithromycin (PPI-AC) therapy, which is the first-line regimen in our country. We determined the respective effects of clarithromycin primary and secondary resistances on efficacy of the PPI-AC regimen and examined whether failures were associated with persistence of the same strain or with emergence of a new one. Hundred and twenty three H. pylori-infected patients were treated for seven days with omeprazole 20 mg b.d., amoxicillin 1 g b.d., and clarithromycin 500 mg b.d. Eradication was assessed by breath test in 102 patients. MICs of clarithromycin were determined by E-test. Strain genotyping was performed by random amplified polymorphic DNA. The pre-treatment and post-treatment prevalences of clarithromycin resistance were 18.7% (23/123) and 69.2% (9/13), respectively. The rates of eradication were 67.6% (69/102), 78.8% (67/85), and 11.8% (2/17) for all, susceptible and resistant strains, respectively. The post-treatment isolate was available for six patients with a susceptible pre-treatment isolate and a persistent infection; resistance emerged in two patients and was associated with persistence of the pre-treatment strain in one and with selection of a new strain in the other. In conclusion, in our hospital, failures of the PPI-AC therapy are related to both clarithromycin primary and secondary resistances but emergence of secondary resistance does not explain all failures in the initial clarithromycin-susceptible group. In that group a new strain can emerge after failure. (C) 2001 Editions scientifiques et medicales Elsevier SAS.