Two hundred and thirty-eight residents of Linxian, China, with histologically confirmed mild or moderate esophageal squamous dysplasia (a precursor to esophageal cancer), were randomly assigned to receive, in double-blind fashion, 200 mcg/day of selenium (as selenomethionine), 200 mg twice a day of celecoxib, both, or placebo for ten months, in a 2 x 2 factorial design. Esophagogastroduodenoscopy was performed before and after ten months of treatment. Selenium treatment, as compared with no selenium, resulted in a trend toward increased dysplasia regression (43% vs. 32%) and decreased dysplasia progression (14% vs. 19%) (p = 0.08 for the combined endpoint). In unplanned stratified analysis, compared with no selenium, selenium favorably affected the change in dysplasia grade among 115 subjects with mild dysplasia at baseline (39% vs. 21% for regression, 19% vs. 36% for progression; p = 0.02 for the combined endpoint), but not among 123 subjects with moderate dysplasia at baseline (p = 1.00).

Comment: This study suggests that selenium has a protective effect on disease progression in people with mild, but not moderate, esophageal squamous dysplasia. The results should be interpreted with caution, because they were obtained in a post-hoc stratified analysis, which is less reliable than a pre-planned analysis. Nevertheless, the results are consistent with many other studies showing an anticancer effect of selenium. Selenium prophylaxis would be expected to be most effective when begun early in the course of the disease; for example, at the time someone is offered his or her first cigarette or boiling hot cup of tea.