It is known that in acromegalic patients the GH secretion exhibits a pattern of response to various stimuli that differs from that observed in normal subjects. We have evaluated the paradoxical GH response to 1-Dopa in acromegaly, in order to clarify the relationships between this datum and the tumour size, the basal GH secretion, the GH response to TRH and the long-term prognosis. MATERIALS AND METHODS. 34 acromegalic patients, 12 men and 22 women, aged 35-70 yr, participated in this study. They were divided in 4 classes, according to Hardy's classification of pituitary neoplasias: (1) intrasellar microadenomas, < 10 mm (n = 8); (2) intrasellar macroadenomas, > 10 mm (n = 14); 3) adenomas with local expansion (n = 9); (4) adenomas with extrasellar expansion (n = 3). All patients underwent a 1-Dopa-test; 18 of them underwent a TRH-test. Basal postoperative GH and basal pre- and postoperative PRL levels were also determined. RESULTS. The basal preoperative GH values in all patients ranged between 16 and 278 ng/ml. 22 patients showed a paradoxical response to 1-Dopa, 12 were non-responders. The following results were observed in the different classes: [table: see text] Moreover, 10 subjects (56%) were TRH-responders (9 of them were also 1-Dopa-responders, 1 was non-responder) and 8 were TRH non responders (6 were 1-Dopa non-responders, 2 were 1-Dopa responders). Basal postoperative GH values were > 5 in 17 1-Dopa responders and in 9 non responders. Basal preoperative PRL levels were > 25 ng/ml in 7 1-Dopa responders and in 2 non responders. After surgery, basal PRL became normal in 5 responder patients. DISCUSSION. In a previous work we have correlated the paradoxical GH response to TRH with tumour size and GH levels, observing a higher percentage of paradoxical response in patients in class I and II and, postoperatively, lower GH levels in preoperative TRH responders. So, we have underlined the good prognostic feature of a preoperative paradoxical response. In this paper we have evaluated the paradoxical GH response to 1-Dopa in the different Hardy's classes and compared it with the GH levels and the GH response to TRH. The results show that a paradoxical response can be observed more frequently in small (class I) adenomas than in greater size ones, and in presence of lower GH basal levels. Moreover, a concordance between 1-Dopa and TRH tests can be observed. CONCLUSIONS. The results clearly indicate that the responses to dynamic GH tests should be evaluated considering the anatomic characteristics of the neoplasias. It could be suggested that the paradoxical response can be expressed only when hypothalamus-pituitary interactions are intact.