This retrospective analysis was undertaken to determine whether selective adenomectomy for Cushing's disease can achieve acceptable cure rates while causing minimal pituitary dysfunction. Tumour size, histology and pituitary function were evaluated in 34 consecutive patients (26 F:8 M, mean age 33.6 years) undergoing transsphenoidal adenomectomy for Cushing's disease from 1975 to 1992. Follow-up averaged 5.8 years. Cure was defined as resolution of symptoms and signs and normalization of urinary cortisol excretion. Sixty-three per cent of patients achieved cure after selective adenomectomy; repeat adenomectomy cured an additional four patients. Twenty-eight per cent required bilateral adrenalectomy and/or pituitary irradiation. Postoperative pituitary function remained completely intact in 81%. Secondary hypogonadism occurred in 8%, hypothyroidism in 15% and permanent diabetes insipidus in 4%. There was recurrence in 26% after a mean of 4.6 years (range 1-7). It is concluded that selective adenomectomy can achieve acceptable cure rates with a low prevalence of postoperative hypopituitarism, although an increase of recurrence may be the result of conservative surgery.