A randomized double-blind intervention trial was done in Huixian, People's Republic of China, a population with a high incidence of esophageal cancer. The aim of the trial was to determine whether a once-a-week treatment with retinol (15 mg or 50,000 IU), riboflavin (200 mg), and zinc (50 mg) could result, after 1 year, in a lower prevalence of precancerous lesions of the esophagus in the group receiving the active treatment as compared with the prevalence in the group receiving a placebo. The results of the trial, published elsewhere, indicated that the treatment had no effect on the prevalence of precancerous lesions of the esophagus. In determining whether an effect could be detected when earlier end points are used, the prevalence of micronuclei was evaluated in exfoliated cells from the esophagus and from the buccal mucosa in the present study. In a subsample of 200 out of the original 610 study subjects, smears were taken from the buccal mucosa before and after treatment, and in 170 subjects esophageal smears were obtained during endoscopy only after treatment. The smears were fixed and kept at room temperature over 1 year before being evaluated for the presence of micronuclei by means of 4'-6-diamidino-2-phenylindole fluorescent staining. Smears from approximately half of the subjects were considered suitable for evaluation. No statistically significant difference in the prevalence of micronuclei in the buccal mucosa cells was observed before and after treatment (the mean percentage of micronucleated cells in the vitamin group upon first examination, before treatment started, 0.35%; 1 year after treatment, 0.31%) or between the treatment and the placebo group at the final examination. (The mean percentage of micronucleated cells in the vitamin-treated group was 0.31 and 0.39% in the placebo group.) However, a statistically significant reduction (P = .04) was observed in the prevalence of micronuclei in esophageal cells in the treatment group as compared to the placebo. (The mean percentage of micronucleated cells in the vitamin-treated group was 0.19%; it was 0.31% in the placebo group.