Several studies have suggested that selenium may help to prevent colorectal neoplasia. To investigate the relation between prediagnostic serum selenium concentrations and colorectal adenomas, we conducted a nested case-control study using data from a large, multicenter, adenoma prevention trial. Cases comprised a total of 276 patients who developed a colorectal adenoma between the year 1 and year 4 follow-up exam. Controls were 276 patients who did not develop an adenoma during this time interval, matched to case subjects on age, sex, and clinical center. Total and bound selenium concentrations were measured from baseline or year 1 serum samples using instrumental neutron activation analysis. We estimated the odds ratios of colorectal adenoma in relation to serum selenium concentrations adjusting for age, clinical center, and sex. Compared with the lowest quintile, the odds ratio for the highest quintile was 0.76 (95% confidence interval, 0.44-1.30) for total selenium and 0.60 (95% confidence interval, 0.34-1.05) for bound selenium, and there was no apparent trend in risk (P for trend = 0.50 for total selenium and P for trend = 0.20 for bound selenium). Thus, our findings do not indicate a clear association between serum selenium concentrations and adenoma recurrence.