Serum specimens for 190 women with early stage breast cancer, 190 women with high-risk types of benign breast disease, and 190 women with low-risk types of benign breast disease,as well as samples to be used for quality control purposes, were sent to four different laboratories for measurement of estrogens (total estradiol, non-SHBG bound estradiol, estrone, and calculated SHBG-bound estradiol, estrone sulfate), SHBG, androgens (testosterone, androstenedione, DHEA, and androstenediol), serum c-peptide of insulin, fructosamine, total insulin-like growth factor-1, total insulin-like growth factor-11, and insulin-like growth factor binding protein-1. Coefficients of variation (CV) for each assay were calculated using quality control specimens from the first three batches. If CVs were satisfactory, laboratories were instructed to complete the remaining assays. CVs were satisfactory for all assays except those for estradiol, estrone, and estrone sulfate. Changes in the laboratory procedures for these assays were implemented. Assay results after the change in laboratory procedures have not been fully evaluated.
