Transsphenoidal pituitary surgery was performed in 64 patients with Cushing's disease in search of a corticotroph adenoma. In 4 patients, profuse local bleeding prevented completion of the exploration. Of the 60 patients who had an adequate exploration, 4 could not be followed after surgery. Short term assessment of the surgical outcome (3-6 months postoperatively) was performed on 60 patients, including the 4 who had incomplete pituitary exploration. Forty-two patients (70%) were judged as immediate successes [urinary cortisol excretion, less than 90 micrograms (less than 248 nmol)/day]; the mean urinary cortisol excretion and plasma ACTH level fell from 463 +/- 70 (+/- SE) to 26.7 +/- 3.6 micrograms/day (1277 +/- 193 to 74 +/- 10 nmol/day; n = 33) and from 111 +/- 33 to 36 +/- 14 pg/mL (24 +/- 7 to 8 +/- 3 pmol/L; n = 23) in patients who had both measurements pre- and postoperatively. Eighteen patients (30%) were judged as immediate failures; neither urinary cortisol excretion nor plasma ACTH levels changed significantly in patients who had both measurements pre- and postoperatively. The preoperative epidemiological, clinical, hormonal, and radiological characteristics of the 2 groups were similar. Histological examination of pituitary fragments removed in 58 of the 60 patients evaluated postoperatively revealed the presence of tumoral tissue in a higher percentage of patients in the immediate success group (72%) than in the immediate failure group (24%; P less than 0.01). The 42 patients in the immediate success group were followed from 6 months to 7 yr (median, 2 yr); 6 patients had recurrences from 2-3 yr after operation. Actuarial analysis indicates that the probability of a patient remaining well 6 yr after surgery is 72 +/- 20% (95% confidence limit). Most of the patients in the immediate success group had transient ACTH deficiency preceding a progressive return to normal pituitary-adrenal function.