Fourteen patients with Helicobacter pylori infection were treated with 20 mg omeprazole, 1 g amoxycillin and 400 mg metronidazole bd for 7 days (OAM), and 16 patients were treated with 20 mg omeprazole, 250 mg clarithromycin and 400 mg metronidazole bd for 7 days (OCM). Saliva, gastric biopsies and faecal samples were collected before, during (day 7) and 4 weeks after treatment in order to analyse alterations of the normal microflora and to determine antimicrobial susceptibility. Both treatment regimens resulted in marked quantitative and qualitative alterations. A selection of resistant streptococcal strains were noticed in both treatment groups, most apparent in the OCM group where a shift from susceptible to resistant strains was recorded. In the OAM group, six patients had overgrowth of resistant enterobacteriaceae during treatment compared with none in the OCM group, in the gastric microflora. The MICs for Enterococcus spp. and Enterobacteriaceae in faeces increased significantly during treatment in both groups. Nine patients in the OAM group became intestinally colonized by yeasts during treatment. The total anaerobic microflora was strongly suppressed in both treatment groups, although most pronounced in the OCM group, where the frequency of clarithromycin-resistant bacteroides strains increased from 2 to 76% during treatment, and remained at 59% 4 weeks post-treatment. Even if the treatment outcome was better in the OCM group (100%) than in the OAM group (71%), the amoxycillin-based treatment might be preferable from an ecological point of view, since the qualitative alterations in terms of emergence and persistence of resistant strains seemed to be most pronounced in the clarithromycin-treated group.