Measurement of serum prolactin levels is useful in differentiating prolactinoma from other causes of increased prolactin concentrations. Distinguishing between prolactinoma and a nonfunctioning pituitary adenoma, or even nonadenomatous disease, is important because dopamine agonist therapy may reduce tumor size in most macroprolactinomas but is largely ineffective in nonsecreting adenoma and nonadenomatous lesions [1]. Solid-phase sandwich radioimmunoassay (immunoradiometric assay) is frequently used to measure prolactin levels. Results from immunoradiometric assay kits may be subject to the well-documented, high-dose "hook effect," or prozone effect, where extremely high serum concentrations of an antigen may paradoxically produce a lower response than expected and, on dilutions, produce a higher assay response [2]. We present a case that shows this phenomenon.

The patient was a 37-year-old man who developed visual disturbances and was found, using a high-resolution computerized tomographic scan, to have a large infiltrating pituitary adenoma of 5.5 x 4 x 3 cm. Before surgery, 17 serial prolactin measurements in duplicate were determined by a one-step immunoradiometric assay (Immunocorp; Montreal, Canada) and showed values from 41 to 48 µg/L (normal, <14 µg/L). The patient had incomplete transsphenoidal resection of his large cystic tumor. Pathological and immunocytochemical studies showed typical characteristics of prolactinoma, where all cells were positive with anti-prolactin serum. Dilution studies done on the patient's serum obtained before surgery showed an initial prolactin level of 31 795 µg/L. After 15 months of treatment with bromocriptine, 2.5 mg thrice daily, the tomographic scan showed a marked decrease in the size of the adenoma, and prolactin levels diminished from 11 171 to 227 µg/L.

The development of newer techniques for measuring prolactin, such as the immunoradiometric assay, has emphasized the gain in efficiency that occurs when separate incubation steps followed by washes are eliminated. In our experience, the accuracy and reproducibility of this assay system appears satisfactory, although results using immunoradiometric assay kits may be subject to the hook effect, a problem of which users may be unaware [2]. The effect is thought to be caused by an excess of antigen that saturates the recognition sites on both the unlabeled and labeled antibodies, inhibiting formation of the antigen-antibody sandwich. Because the amount of antigen (prolactin) in the complex is assumed to be proportional to the amount of radioactivity present, falsely low results would be obtained. The "hook effect" has been recently documented with other immunoradiometric assays [3, 4]. A case report illustrating high-dose hook effect for prolactin determination has not been previously reported. However, Haller and colleagues [5] showed that this effect, when induced in vitro, can be eliminated by using a two-step assay or by doing the analysis at two sample dilutions [5]. Such a procedure should be done in patients with a prolactinoma.