Usefulness of prolactin levels in predicting the etiology of hyperprolactinemia in a cohort of 770 patients.

Objective: Determining the etiology of hyperprolactinemia is fundamental for selecting the most appropriate treatment strategy. The aim of this study was to evaluate the usefulness and accuracy of prolactin levels in predicting the etiology of nonphysiological hyperprolactinemia. Subjects and methods: In this retrospective study, we reviewed medical records of patients with nonphysiological hyperprolactinemia seen at two neuroendocrine reference centers located in Recife, Brazil, from January 2000 to December 2019. Results: The study included 770 patients aged 12-73 years (65% female). The three most frequent etiologies of hyperprolactinemia were prolactinomas (n = 263; 34.2%), drug-induced hyperprolactinemia (n = 160; 20.8%), and macroprolactinemia (n = 120; 15.6%). The highest mean prolactin levels were observed in cases of prolactinomas and idiopathic hyperprolactinemia. Most patients with hyperprolactinemia due to other etiologies had prolactin levels < 100 ng/mL, but these levels were also found in 16.5% of patients with microproplactinomas and in 20% of those with idiopathic hyperprolactinemia. Likewise, prolactin levels largely overlapped among patients with microprolactinomas, macroprolactinemia, and drug-induced hyperprolactinemia. Notably, prolactin levels > 250 ng/mL enabled a clear distinction between the etiologies of macroprolactinoma and nonfunctioning pituitary adenoma. Moreover, prolactin levels > 500 ng/mL were highly suggestive of macroprolactinomas, although they were also found in very few patients (<2%) with microprolactinomas or drug-induced hyperprolactinemia. Conclusion: Despite considerable overlap in prolactin levels among the different etiologies of hyperprolactinemia, values > 250 ng/mL allowed a clear distinction between macroprolactinomas and nonfunctioning pituitary adenomas. Furthermore, prolactin levels > 500 ng/mL were almost exclusively found in patients with prolactinomas.