UNLABELLED This study was performed to investigate the correlations between preoperative prolactin (PRL) plasma values, immunohistochemical picture and the clinical course in growth hormone (GH) secreting pituitary adenomas. In 47 patients (19 males and 28 females; mean age 40 years; range 13 - 70 years), we measured GH, IGF-1 and prolactin plasma values both before and after transsphenoidal surgery, and basal IGF-1 and GH after an oral glucose tolerance test (OGTT) during four years of follow-up. We considered those patients as "controlled" who presented an undetectable growth hormone after OGTT (GH < 1 microg/l), IGF-I plasma values in the normal range, matched for age and sex, and no clinical activity or neuroradiological recurrence after a four-year follow-up. We considered patients as "poorly controlled" who still showed elevated GH and IGF-I plasma levels, uninhibited GH after OGTT (GH > 1 microg/l), presence of clinical activity and/or radiological signs of adenoma recurrence, even if a reduction of tumor size had been demonstrated. RESULTS Controlled patients (n = 22) exhibited mean preoperative PRL levels (+/- SEM) lower than the group of poorly controlled (n = 25) ones (21.40 +/- 5.51 vs. 38.44 +/- 5.16 microg/l; p < 0.03). From 3 to 12 months after surgery, postoperative PRL levels were also lower in the controlled patients compared to the poorly controlled ones (8.31 +/- 1.20 vs. 25.32 +/- 3.20 microg/l; p < 0.0001). Eighty percent (20/25) of poorly controlled patients showed both PRL and GH positivity after immunostaining. Only 3/22 (13.6 %) of controlled patients showed the same double positivity. In conclusion, preoperative hyperprolactinemia identifies a group of acromegalic patients at elevated risk of disease persistence after surgery. We hypothesize that most of these high-risk patients may have more aggressive mixed GH-PRL secreting adenomas.