We have followed nine adult patients with congenital adrenal hyperplasia (CAH) for between 7-77 months on dexamethasone (DXM) 0.5 mg mane and 0.25 mg nocte, reducing to 0.5 mg mane. Twenty-four hour profiles of ACTH, 17-hydroxyprogesterone (17OHP), and androstenedione were performed; the areas under the curves (AUC) and the heights of the morning peaks were used to assess biochemical control. Comparisons were made between treatment before DXM (pre-DXM), 0.75 mg for 2 weeks (DXM-ST), 0.75 mg for at least 3 months (DXM-LT), and 0.5 mg for at least 3 months (DXM-0.5). None of the three males suffered any significant side-effects. All women had menstrual disturbance but in three ovulation was induced. One female developed Cushing's syndrome and two developed hirsutism which resolved on stopping DXM. Overall there was no significant difference between DXM-ST and DXM-LT (mean AUCs for ACTH: DXM-ST 660, DXM-LT 383, for 17OHP: DXM-ST 1177, DXM-LT 587, for androstenedione DXM-ST 232, DXM-LT 121). Reduction of the dose from 0.75 mg to 0.5 mg led to significant deterioration in control (Mean AUC's for ACTH DXM-0.5, 1123 (P less than 0.02), for 17OHP DXM-0.5, 2068 (P less than 0.002), for androstenedione DXM-0.5, 213 (P less than 0.5). We conclude that DXM is a satisfactory regime but the dose must be adjusted for each patient.