Objectives: To select rifaximin-resistant mutants of Escherichia coli and to establish the frequency of mutation, cross-resistance with other antimicrobial agents and the stability of the mutants obtained. Methods: Four E. coli isolates [two enteroaggregative E. coli (EAEC) and two enterotoxigenic E. coli (ETEC)] were used to obtain rifaximin-resistant mutants. The frequency of mutation in the presence of rifaximin, rifampicin and ciprofloxacin was established by growth on plates containing serial dilutions of antibiotic above the bacterial MIC. To determine the stability of rifaximin resistance, 28 selected resistant mutants were grown for 20 consecutive cultures on antibiotic-free plates. Every 10 days, the MICs of rifaximin, chloramphenicol, nalidixic acid and ciprofloxacin were established. Results: The frequency of mutation in the presence of rifaximin ranged between 5.7 × 10-7 and 1.6 × 10-6 in the case of the ETEC isolates, and between 2.0 × 10-8 and 9.3 × 10-8 in the case of the EAEC isolates; the frequency of mutation in the presence of rifampicin was in the order of 10-8 and no mutant in the presence of ciprofloxacin was obtained. Twenty-six out of 28 selected mutants exhibited resistance levels around or higher than 256 mg/L. In all cases, the resistance was stable, and no reversion towards the original parental MIC was observed. In no case was the MIC of chloramphenicol, nalidixic acid or ciprofloxacin affected. Conclusions: Rifaximin has a low level of resistance selection, although it may select stable highly resistant mutants in a single step. Periodical surveillance of the levels of rifaximin resistance is required to detect the possible appearance of rifaximin-resistant clinical isolates. Further studies to characterize in-depth the mechanisms of stable resistance to rifaximin are necessary. © The Author 2008. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved.