Endocrine Abstracts

After bromocriptine withdrawal recurrence rates of >90% in patients with macroprolac-tinomas and >80% in those with microprolactinomas have been reported, but follow-up studies were too short to permit definitive conclusions. No systematic data on cabergoline withdrawal are currently available nor have criteria for withdrawal been established. To investigate the success rate of cabergoline withdrawal in patients with non-tumoral hyperprolactinemia (NTH), microprolactinomas and macroprolactinomas so as to suggest reliable criteria for treatment withdrawal we studied 200 patients (25 NTH, 105 microprolactinomas, 70 macroprolactinomas). Inclusion criteria were: 1) prolactin normalization and 2) magnetic resonance imaging (MRI) showing no tumor or >50% tumor reduction, >5mm distance from optic chiasm and absent invasion of the cavern-ous sinuses for at least 12 months. The Kaplan-Meier estimate of recurrence rate after five years was 24% in patients with NTH, 32.6% in those with micro- and 43.3% in those with macroprolactinomas. In none of the patients did the tumor re-grow; 10 women (22.2%) and seven men (38.9%) with recurrent hyperprolactinemia developed re-appearance of gonadal dysfunction. In all groups, prolactin levels at recurrence were significantly lower than at diagnosis (p<0.001). Patients with either macro- (77.5 vs. 32.6%, p=0.001) or microprolactinoma (41.5 vs. 26.2%, p=0.02) showing small remnant tumors at MRI during cabergoline treatment had an increased recurrence rate after five years than those achieving negative MRI. Nadir prolactin level during cabergoline was the best predictor of last prolactin levels after withdrawal (t=7.496, p<0.001). In conclusion, cabergoline withdrawal can be safely performed in patients achieving prolactin normalization and tumor disappearance. In patients showing remnant tumors at MRI, recurrence rate was higher but in no patient we observed tumor re-growth. Since the duration of follow-up is still too short to rule out delayed increases of tumor size, we suggest close follow-ups mainly in macroprolac-tinomas where tumor re-growth may compromise vision.