CONTEXT AND OBJECTIVE The role of cabergoline in Cushing's disease (CD) remains controversial. The experience is limited to case reports and few open studies that report the effects determined after ≥1 month of treatment. In prolactinomas and dopamine-responsive GH-secreting tumours, effects of cabergoline are seen within days or weeks. Here, we searched for short-term effects of cabergoline in CD. DESIGN Twenty patients (19 naïve and one recurrent) were included in a prospective study. Cabergoline was administered in increasing doses of 0.5-5 mg/week over 6 weeks. METHODS Urinary free cortisol (UFC) 24 h, morning cortisol and ACTH, and salivary cortisol at 0800, 1600 and 2300 h were determined once weekly throughout. Diurnal curves (six samples) of serum cortisol were measured at start and end. RESULTS At study end, the median cabergoline dose was 5 mg, range 2.5-5 mg/week. The prolactin levels, markers of compliance, were suppressed in all patients. During the treatment, hypercortisolism varied, gradual and dose-dependent reductions were not seen. Five patients had a >50% decrease of UFC, three had a >50% rise of UFC. Salivary cortisol at 2300 h showed a congruent >50% change with UFC in two of the five cases with decreased UFC, and in one of the three cases with increased UFC. One patient with decreases in both UFC and 2300 h salivary cortisol also had a reduction in diurnal serum cortisol during the course of the study. CONCLUSIONS Cabergoline seems to be of little value in the management of CD. Only one patient had a response-like pattern. Given the known variability of disease activity in CD, this might represent a chance finding.