Concern regarding hydrogen peroxide (H(2)O(2)) carcinogenicity arises from its ability to act as a strong oxidizing agent. In short-term genotoxicity tests, H(2)O(2) has given predominantly positive results; however, these assays have been performed using either bacterial strains engineered to be exquisitely sensitive to oxidant damage, or mammalian cells deficient in antioxidant enzymes. Significantly, the addition of antioxidant protective measures (normally present in vivo) to these assay systems protects against H(2)O(2) genotoxicity. In most whole animal studies, H(2)O(2) exposure neither initiates nor promotes tumors. In mice, however, 0.4% H(2)O(2) in drinking water was reported to induce hyperplastic lesions of the duodenum and to erode areas in the glandular stomach epithelium. Owing to the chemistry of dilute H(2)O(2) solutions and the anatomy/physiology of the gastrointestinal tract, it is unlikely that orally ingested H(2)O(2) reaches the duodenum. Instead, greatly decreased water consumption and the resultant abrasion of the luminal lining on ingestion of pelleted dry rodent chow is the most likely cause of the observed gastric and duodenal lesions following H(2)O(2) administration in drinking water. Significantly, when hamsters received high doses of H(2)O(2) by gastric intubation (and water intake was not affected), the gastric and duodenal epithelia appeared normal. In-depth analysis of the available data supports the conclusion that oral ingestion of H(2)O(2) should not be considered a carcinogenic threat.