The bacteriumHelicobacter pylori (HP) has been implicated in the etiology of precancerous lesions of the stomach and there is evidence suggesting that it may influence the efficacy of chemoprevention of gastric cancer with vitamin C. Eradication seldom has been attempted in populations from developing countries, with a high prevalence of HP frequently resistant to metronidazole. A randomized, double-blind, controlled trial to evaluate the efficacy of colloidal bismuth subcitrate (120 mg q.i.d.) and amoxycillin (500 mg q.i.d.) in eradicating HP was conducted in 220 subjects drawn from a population with a high prevalence of metronidazole-resistant HP in Tachira state, Venezuela. One month after completion of two weeks' treatment, eradication rates of 6.5 percent in the treatment group and two percent in the placebo group were estimated on the basis of HP diagnosis in biopsies, and of 13.9 percent compared with 3.9 percent on the basis of a14C-urea breath test, although the negative predictive value of the breath test was very low compared with HP diagnosis in biopsies. In the treatment group, particularly among males, a significant decrease in bacterial load was detected. Reasons for failure of treatment in high HP-prevalence areas are discussed, and it is suggested that primary prevention of HP infection may be the optimal approach to reducing levels of stomach cancer in these high-risk groups.