Individuals with atrophic gastritis (n = 863) were recruited to participate in a chemoprevention trial in Nariño, Columbia. The volunteers were randomly assigned to intervention therapies, which included antibiotic treatment for Helicobacter pylori infection, and then daily dietary supplementation with antioxidant micronutrients in a 2(3) factorial design. Biopsies were obtained according to a specified protocol from designated areas in the stomach for each individual at baseline (before intervention therapy), at year 3, and at year 6. A systematic sample of 160 participants was selected from each of the eight treatment combinations, and the first exon of KRAS was examined for mutations. At year 3, the data indicated that individuals with KRAS mutations in their baseline premalignant stomach biopsies were 3.74 times as likely to progress to a higher premalignant stage than those who lacked baseline mutations (P = 0.04; C. Gong et al., Cancer Epidemiol. Biomark. Prevy. 8:167-171, 1999). However, after 6 years, baseline KRAS mutations failed to predict histological progression. Also, KRAS mutation in 72-month biopsies did not predict histological progression.