Purpose: Recent guidelines suggest that a single prolactin measurement is adequate to confirm hyperprolactinaemia. This may lead to unnecessary investigation of artefactual hyperprolactinaemia. Prolactin measurement drawn from an indwelling cannula after rest removes stress as a confounding variable. The objective was to determine the frequency of true hyperprolactinaemia amongst patients referred following a single prolactin measurement.
Methods: A cannulated study was considered if prolactin on referral ('Referral Prolactin') was <5,500 mU/L (260 ng/mL) but >410 mU/L (19 ng/mL) in males or >510 mU/L (24 ng/mL) in females, irrespective of clinical context. Case-notes of 267 patients undergoing cannulated prolactin measurement over a 10-year period (2000-2010) were reviewed. Pre-existing pituitary disease, dopamine antagonist use, and macroprolactinaemia were excluded. Morning ante-cubital vein cannulation was followed immediately by withdrawal of 'Repeat Prolactin' sample. After 120-min bed-rest, 'Resting Prolactin' was withdrawn through the cannula.
Results: 235 patients were included for analysis. 64 (27%) were within normal range; following Repeat Prolactin in 41 (17%) and Resting Prolactin in 23 (9%) cases. Referral Prolactin was higher in patients with true hyperprolactinaemia, 1,637 ± 100 mU/L (77.2 ± 4.7 ng/mL) than with artefactual hyperprolactinaemia, 1,122 ± 68 mU/L (52.9 ± 3.2 ng/mL; P < 0.001) but there was substantial overlap. 21 out of 171 cases (12%) with true hyperprolactinaemia had a macroadenoma. Presenting symptoms did not predict true hyperprolactinaemia. Referral Prolactin of 2,000 mU/L (94 ng/mL) had 97% specificity to identify true hyperprolactinaemia.
Conclusions: Reliance on a single, non-rested prolactin value may lead to over-diagnosis of hyperprolactinaemia. A resting sample should be considered with random values <2,000 mU/L (94 ng/mL).