Hypeprolactinemia: still an insidious diagnosis

Hyperprolactinemia can have different causes: physiological, pharmacological, and pathological. When investigating the etiology of hyperprolactinemia, clinicians need to be aware of several conditions leading to misdiagnosis. The most popular pitfalls are: acute physical and psychological stress, macroprolactin, hook effect, even though antibodies interferences and biotine use have to be considered. A 52-year-old woman was referred to Endocrinology clinic for oligomenorrhoea and headache. She worked as a butcher. Hormonal evaluation showed very high PRL (305 ng/ml, reference interval: <24 ng/ml) measured with the ECLIA immunoassay analyzer Elecsys 170. The patient’s pituitary MRI was normal and macroprolactin was normal. Hormonal workup showed LH: 71.5 mU/ml (2–10.9 mU/ml), FSH: 111.4 mU/ml (3.9–8.8 mU/ml), Estradiol: 110.7 pg/mL (27–122 pg/ml). Since an interference was suspected, the sample was sent to another laboratory using a different assay. After antibody blocking tubes treatment (Heterophilic Blocking Tube, Scantibodies) PRL was 28.8 ng/ml (reference interval < 29.2 ng/ml). Analytical interference should be suspected when assay results are not consistent with the clinical picture. Endogenous antibodies (EA) include heterophile, human anti-animal, autoimmune and other nonspecific antibodies, and rheumatoid factors, that have structural similarities and can cross-react with the antibodies employed by the immunoassay, causing hyperprolactinemia misdiagnosis. The patient’s job (butcher), led us to suspect the presence of anti-animal antibodies. Clinicians should also carefully investigate the use of supplements. Biotin can falsely increase hormone concentration in competitive assays. Many clinicians are still not informed about these pitfalls that are not mentioned in some recent reviews on PRL measurement.

In healthy subjects and in prolactinomas, total circulating PRL comprises 65-85% monomeric 23-kDa PRL, 15-30% dimeric 40-60-kDa "big" PRL and <10% > 150 kDa "bigbig PRL (or macroprolactin), usually composed of a complex formed by monomeric PRL and IgGs. Macroprolactin has a lower renal clearance and it is minimally active. Since macroprolactin is variably detected by the immunoassays currently used by laboratories, high PRL concentrations can be found in normally ovulating women and don't require any treatment [1,2]. The reference method for the detection of macroprolactin is size exclusion chromatography, but this technique is time consuming and expensive. Polyethylene glycol (PEG) acts as a "sponge," which absorbs water of hydration from proteins, reducing their solubility and leading to their precipitation, and has been widely proposed and used as a screening method [5][6][7][8][9]. Many authors reported that the assays and the automated analyzers used by different laboratories differently recognize macroprolactin [10][11][12].
The "hook effect," i.e., falsely normal or mildly elevated PRL while the true PRL concentration is many fold higher than the upper limit, can be found in presence of large pituitary macroadenomas (≥3 cm) and clinical manifestations typical of prolactinoma. When this situation is suspected, clinicians should carry out a serial dilution of serum sample to eliminate the artifact. Immunoassays are usually based on capture antibodies that are immobilized in a solid phase and a second antibody that is usually labeled with a chemiluminescent or a fluorescent signal [1][2][3][4][5][6][7][8][9]. These antibodies bind to the antigen (PRL) forming a "sandwich" and a signal with an intensity proportional to the concentration of PRL. The relative antigen-to-antibody proportion influences its interaction and may hamper the appropriate formation of the immunocomplexes. The hook effect occurs when extremely high PRL concentration saturates both the capture and the labeled antibody, preventing the formation of the "sandwich" and causing falsenegative results [1][2][3][4][5][6]13].

Case report
A 52-year-old woman suffering from a 12 months oligomenorrhoea and headache was referred to Endocrinology clinic for investigations. The patient reported menarche at 14 years, two previous pregnancies, and normal menstrual cycles until 12 months before. Galactorrhea was not detected and gynecologic evaluation was normal. She worked as a butcher. Hormonal evaluation showed very high PRL (305 ng/ml, reference interval: <24 ng/ml) measured with the ECLIA immunoassay analyzer Elecsys 170 (Roche, Milan, Italy). This result was confirmed with the same assay in a sample collected on a different day avoiding venipuncture stress and repeating the sampling at 15-20 min intervals. Other laboratory exams showed LH: 71.5 mU/ml (2-10.9 mU/ml), FSH: 111.4 mU/ml (3.9-8.8 mU/ml), Estradiol: 110.7 pg/mL (27-122 pg/ml) while TSH, FT4, IGF1, GH were within the reference interval. The search for macroprolactin by PEG precipitation was negative. We excluded medication and supplements use, renal failure, and hypothyroidism as well as the other conditions known to cause hyperprolactinemia (breast stimulation, chest trauma, etc). The patient's pituitary MRI was normal (no adenoma, no empty sella, no parasellar mass, etc). Since the presence of an interference was suspected, the sample was sent to another laboratory using a different immunoassay (DxI, Beckman, Milan): PRL concentration was 30.2 ng/ml (reference interval 3.3-26.7 ng/ml) and PEG precipitation research for macroprolactin was negative. Since clinical manifestations and neuroradiology imaging were not concordant with PRL concentration, we measured again PRL using a Centaur XP analyzer (Siemens, Milan, Italy) after antibody blocking tubes treatment (Heterophilic Blocking Tube, HBT, Scantibodies, Santee, CA, USA). Macroprolactin research was again negative while PRL after HBT treatment was 28.8 ng/ml (reference interval < 29.2 ng/ml).

Discussion
Analytical interference should be suspected when assay results are not consistent with clinical picture. In the reported case, the presence of high gonadotropin levels (not consistent with central hypogonadism typical of true hyperprolactinemia), the absence of a pituitary adenoma, the difference in PRL values measured by the different employed analyzers and the patient's job (butcher), led us to suspect the presence of anti-animal antibodies. These antibodies interfere with PRL measurement and can cause hyperprolactinemia misdiagnosis [14]. The currently available basic immunoassay formats for measuring hormones are two: the "sandwich" and competitive assay. Endogenous antibodies (EA) include heterophile, human anti-animal, autoimmune and other nonspecific antibodies, and rheumatoid factors that have structural similarities and can cross-react with the antibodies employed by the immunoassay causing erroneous results. Heterophile antibodies, described as antibodies against red blood cell proteins of different species (e.g., rat, sheep, horse, rabbit, cow), are low avidity antibodies that occur naturally and do not require exposure to any immunogen [15][16][17][18][19][20][21]. Human anti-animal antibodies are high avidity and species-specific antibodies, produced following acute or chronic exposure to animal proteins. Circulating anti-animal antibodies can arise as a normal response of the human immune system to an administered "foreign" protein antigen [18][19][20][21]. Therapeutic administration of animal antisera and immunoglobulins (e.g., passive immunization with horse anti-tetanus antibodies), consumption of foodstuff (bovine milk and meat), prolonged exposure to animals (e.g., house pets) and animal products (e.g., meat treated by butchers) are the most common causes of the generation of specific human antibodies against animal immunoglobulins. Human anti-mouse antibodies are formed after the administration of diagnostic or therapeutic mouse monoclonal antibodies labeled with isotopes such as 99mTc or tagged with chemotherapeutic agents. Other sources of anti-animal protein are blood transfusion and vaccination, maternal transfer across the placenta to the unborn child, and the transfer of dietary antigens across the gut wall in diseases such as celiac disease [17][18][19][20][21]. The mechanism by which EA causes interference is different depending on the type of antibody and the immunoassay format. EA can lead to both falsely high and low analyte concentrations according to the site of interference (Fig. 1). EA usually cross-link capture antibodies with detection antibodies in the absence of antigen. Therefore, in this case, the system will detect the analyte (even if there is no analyte) and there will be a falsepositive result. This type of interference is more common in sandwich assays. False-negative results are also possible: EA can reduce analyte concentration, especially in competitive assays. EA could cause interference for a number of analytes: macro-enzymes (creatine kinase, amylase), thyroid hormones (free and total forms), thyroglobulin, insulin, and testosterone [17].
Clinicians should also carefully investigate the use of supplements. In recent years many authors reported analytical interference of biotin in several immunoassays based on streptavidin-biotin capture techniques [22,23]. Biotin is included in many over-the-counter multivitamins and used at very high concentration in biotinidase deficiency and multiple sclerosis, and in lower concentrations, but sufficient for interfering, for hair loss and brittle nails. In November 2017 the FDA released a safety warning that biotin supplementation may interfere in some laboratory assays. Biotin can falsely increase hormone concentration in competitive assays and decrease concentrations in sandwich assays [22][23][24][25]. However many clinicians are still not informed about this pitfall that is not mentioned in some recent reviews on PRL measurement [2,7].

Conclusion
In case of discrepancies between imaging, clinical picture, and the laboratory data, clinicians must consider the possibility of the presence of pre-, intra-, and post-analytical interferences. Recent reviews discussed many pitfalls in hyperprolactinemia diagnosis, such as venipuncture, macroprolactinemia, and hook effect. The present case report adds further pitfalls to be considered: EA interference and biotin. The development of more automated analyzers and communication between the requesting clinician and the laboratorian are essential to reduce the possibility that erroneous laboratory results cause harmful consequences to the patients.
Acknowledgements Open access funding provided by Università degli Studi di Ferrara within the CRUI-CARE Agreement.

Compliance with ethical standards
Conflict of interest The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. This article does not contain any studies with animals performed by any of the authors.
Consent to participate The patient provided written informed consent for disclosing clinical and imaging data.
Ethics approval All procedures were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.
Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons. org/licenses/by/4.0/.