In 32 patients with macroprolactinomas or functionless pituitary macroadenomas biochemical and clinical data were correlated with PRL immunocytochemistry. Serum PRL levels revealed a positive correlation with tumour PRL content. Hyperprolactinaemia of 3000 mU/l or more was found only in patients with PRL-positive tumours. In 15 patients with borderline hyperprolactinaemia (below 3000 mU/l), 7 PRL-positive and 8 PRL-negative macroadenomas were found, and in 9 normoprolactinaemic patients 4 PRL-positive and 5 PRL-negative macroadenomas. Patients with PRL-immunostainable tumours had significantly higher median basal serum PRL (P less than or equal to 0.05) than patients with PRL-negative tumours. PRL stimulation after TRH, basal and GnRH-stimulated FSH and LH did not show significant differences between the two groups. A discriminant analysis using 6 biochemical variables was attempted to differentiate between PRL-negative and -positive tumours, which would be helpful in patients with borderline hyperprolactinaemia. Dopamine agonist therapy led to suppression of serum PRL with few exceptions in patients with PRL-positive and -negative tumours, whereas shrinkage was only observed in PRL-immunostainable tumours with high serum PRL levels (over 18,000 mU/l). All patients with PRL-negative tumours showed no change or even growth of the tumour despite dopamine agonist therapy. Our observations indicate that a pituitary macroadenoma associated with serum PRL of more than 3000 mU/l is most probably a prolactinoma (tumour immunostainable for PRL). Dopamine agonist therapy is effective in PRL suppression and tumour shrinkage in most of these patients. Macroadenomas without hormone hypersecretion or with borderline hyperprolactinaemia below 3000 mU/l may or may not contain PRL-immunostainable cells.(ABSTRACT TRUNCATED AT 250 WORDS)