Data from a double-blind intervention trial in China are reanalyzed to explore auxiliary information. The trial had shown that in a high-risk area for esophageal cancer the dietary supplementation of apparently healthy individuals with a combination of retinol, riboflavin, and zinc did not lead to a different prevalence of precancerous lesions of the esophagus among those receiving the active treatment compared to a placebo group. However, improvement of blood retinol and zinc levels were also observed in the placebo group. The logistic regression analysis presented in this paper illustrates that those individuals who showed large increases in retinol, riboflavin, and zinc blood levels were more likely to have a histologically normal esophagus at the end of the trial. This effect is clearer for retinol than for riboflavin and zinc and it is independent of whether the change was caused by the active treatment or occurred otherwise.
Presented at the Fifth Symposium on Epidemiology and Cancer Registries in the Pacific Basin, November 16–21, 1986, Kauai, HI.