Confusion in the interpretation of prolactin levels caused by inappropriately low reference intervals.

Objective: Serum prolactin measurements are important in the differential diagnosis of pituitary masses and subfertility. We observed discrepancies in serum prolactin levels in several patients measured with different immunoassays. Despite differences in assay results, the reference intervals (RIs) derived by the manufacturers were similar. In this study, we aimed to investigate prolactin assay differences and to re-establish RIs for different prolactin immunoassays.

Methods: For the assay comparison, serum samples were collected from men and women visiting the Amsterdam UMC hospital. Prolactin levels were measured using the AtellicaTM IM analyzer (Siemens Healthineers) and the Cobas (Roche Diagnostics) immunoassay. RIs for prolactin were re-established for men, premenopausal women, and postmenopausal women for both the Atellica and Cobas immunoassay.

Results: Prolactin levels measured using the Cobas immunoassay were 1.75 times higher than those measured using the Atellica immunoassay. The re-established RIs for Atellica and Cobas confirmed these findings and were <0.32 U/L; <0.55 U/L for men; <0.64 U/L; <0.86 U/L for premenopausal women, and <0.31 U/L; <0.59 U/L for postmenopausal women, respectively, for Atellica and Cobas assays. The re-established RIs of the Atellica assay matched the current and manufacturer RIs, whereas those for Cobas differed substantially.

Conclusions: Prolactin levels are assay-dependent, and the re-established RIs are different for the Atellica and Cobas assays. We recommend that laboratory specialists and manufacturers periodically review prolactin assay RIs, as incorrect RIs can lead to misinterpretation of prolactin levels and unnecessary referrals and further laboratory testing, as we have experienced.

Plain language summary: We showed that the results of different prolactin tests disagree by 75%, which hinders correct interpretation. Thus, we established test-specific prolactin normal values. By being aware of test differences and using test-specific normal values, one can ensure correct interpretation and prevent unnecessary referrals and concern.