RESUMO
OBJECTIVE: To estimate the proportion of patients with persistent normoprolactinaemia following dopamine agonist (DA) withdrawal and to identify predictors of successful withdrawal in patients with hyperprolactinaemia. DESIGN, PATIENTS, AND MEASUREMENTS: A systematic review of observational eligible studies were identified by searching PubMed and Embase. The primary outcome was the proportion of patients with normoprolactinaemia after cessation of DA treatment. Secondary outcome included the proportion of patients with normoprolactinaemia after DA withdrawal using individual patient data. Risk of bias was assessed by using Newcastle-Ottawa Scale. Pooled proportions were estimated using a random effects model in case I2 ≤ 75% or by reporting range of effects if I2 > 75%. RESULTS: Thirty-two observational studies enroling 1563 patients were included. The proportion of patients with persistent normoprolactinaemia ranged from 0% to 75% (I2 = 84%). Heterogeneity was partly explained by age with more successful withdrawal in patients of higher age. Individual patient data analyses suggested that the proportion of patients with persistent normoprolactinaemia 6 months after DA withdrawal with a low maintenance dose and full regression of the prolactinoma was 87.7% (95% confidence interval [CI] = 60.7-97.1; I2 = 0%) and 58.4% (95% CI = 23.8-86.3; I2 = 75%) for microadenomas and macroadenomas, respectively. CONCLUSIONS: The proportion of patients with persistent normoprolactinaemia following DA withdrawal treatment varied greatly, partly explained by the mean age of participants of the individual studies. Individual patient data analysis suggested that successful withdrawal was likely in patients with full regression of prolactinomas using a low maintenance dose before cessation.