To the Editor.—Ìý Fang et al (Archives 1982;39:1217) raised the important point that plasma cortisol concentrations of 5 to 7 ug/dL after the dexamethasone suppression test (DST) may not be pathologic if the laboratory carrying out the assay uses an unstandardized radioimmunoassay (RIA) method. We have used the DST to examine the plasma cortisol response in depressed patients, normal controls, and patients with other psychiatric illnesses.¹ Initially, we used a commercially available RIA kit produced in Europe (kit A; details of suppliers are available on request). The results of our clinical investigations indicated a prevalence of abnormal DST responses in depressed patients similar to that reported by Carroll et al.² However, after the completion of our study, the manufacturers of kit A withdrew their product from the market and substituted a new kit (kit B), which they claimed was "better" than kit A, as they had found that