Diarrhea in children is often caused by enteropathogen infections that might benefit from early empirical antibiotic therapy. However, when the definition of the pathogen requires sophisticated laboratory studies, the etiology of enteritis is not known early in illness. Empirical therapy may be dangerous if the child is infected with a Shiga toxin-producing Escherichia coli (STEC) strain because antimicrobials may increase Shiga toxin (Stx) release, resulting in increased risk of microangiopathic hemolytic anemia with acute renal failure (hemolytic-uremic syndrome [HUS]) and death. There is a need for antimicrobials that would be effective against multiple bacterial enteropathogens yet not induce Stx release or increase the risk of HUS. Rifaximin has been evaluated in adults for treatment of bacterial enteritis and has a good record for safety and efficacy, but it has not been evaluated extensively in children with gastroenteritis. We therefore evaluated rifaximin's potential for phage induction, drug-induced bacteriolysis, and toxin release in 57 STEC strains (26 O157 and 31 non-O157 strains). Growth in ciprofloxacin, a known Stx phage inducer, caused bacteriolysis and release of toxin in 25/26 (96%) O157 strains and 15/31 (48%) non-O157 strains. In contrast, rifaximin did not induce phage replication or lysis in any strain. Toxin release in the presence of rifaximin was not different from release in the absence of antibiotic. Rifaximin, unlike many antibiotics used to treat pediatric gastroenteritis, does not induce phage-mediated bacteriolysis and Stx release. Copyright © 2007, American Society for Microbiology. All Rights Reserved.