Extensive experience and validation of polyethylene glycol precipitation as a screening method for macroprolactinemia.

Serum human prolactin (PRL) is heterogeneous in molecular size, with the 23-kDa monomer being the predominant form in healthy subjects and patients with prolactinomas. From the point of view of molecular size, other circulating forms include the 50-kDa dimer (big-PRL) and the 150to 170-kDa form (big-big-PRL, or macroprolactin) (1). Recent publications have associated asymptomatic hyperprolactinemia with a predominance of macroprolactin in the circulation; this occurrence appears to be more common than previously thought (2–4) and can have obvious practical implications. The finding of a predominance of macroprolactinemia can change the focus of the evaluation of a patient, with the possible avoidance of more sophisticated and expensive imaging studies. We evaluated the polyethylene glycol (PEG) precipitation method to screen for the presence of macroprolactinemia in a large series of clinical samples. Serum PRL was measured by immunofluorometric assay (IFMA; reference range, 2–15 mg/L; Delfia, Wallac Oy), and samples with values $30 mg/L were studied. The value of 30 mg/L or higher, considered as overtly abnormal, was arbitrarily selected. To 250 mL of serum, we added 250 mL of a 250 g/L PEG 6000 solution (in water, kept at 4 °C), mixed them for 1 min with a vortex mixer, and centrifuged them (9500g for 5 min at room temperature). PRL was determined in the supernatant, using the same IFMA and the recovery calculated on the basis of the original serum value. Reproducibility of the PEG precipitation process was evaluated in four different serum samples, studied seven times each, on different days and in different assays. The following values were obtained: for a sample with PRL value of 32 mg/L and mean recovery of 63%, the CV was 15%; for a sample with PRL value of 45 mg/L and mean recovery of 83%, the CV was 7%; for a sample with PRL of 68 mg/L and mean recovery of 47%, the CV was 28%; and for a sample with PRL of 71 mg/L and mean recovery of 5%, the CV was 20%. The gel filtration used Sephacryl S-200 (Pharmacia) packed in a 0.9 3 30 cm column previously calibrated with monomeric I PRL (New England Nuclear); 0.5 mL of serum was applied and eluted with sodium phosphate buffer (0.05 mol/L, pH 7.4), and 0.7-mL (0.1 mL/min) aliquots were collected and analyzed for PRL. The samples were classified as having a predominance of monomeric PRL forms when the analysis of the area under the curve corresponding to monomeric PRL represented .50% of the total area of PRL elution. Among 19 228 consecutive clinical samples studied in 12 months, PRL was $30 mg/L in 1279 (6.7%). Of these, 1220 were submitted to PEG precipitation, and 171 samples were submitted to the chromatographic study. They were selected to cover the whole range of recovery. The values of the percentage of recovery plotted against the percentage of the area under the curve that was classified as high molecular weight PRL (macroprolactin) are depicted in Fig. 1. On the basis of the data presented in Fig. 1, recoveries of $65% were classified as predominantly monomeric, and recoveries of #30% as predominantly high molecular weight forms. Values between 30% and 65% were classified as indeterminate and submitted to chromatography. The data also showed negative correlation (r 5 20.84) between the percentage of recovery and the percentage of high molecular forms. Of the 1220 samples studied, recoveries were #30% in 441 (36%), $65% in 607 (50%), and 30–65% in 172 (14%). The chromatographic study of these 172 samples showed 72 (42%) with a predominance of high molecular weight forms. Our data support the findings of Bjoro et al. (5), who also reported a large proportion of macroprolactinemia in their clinical population. One of the reasons for our surprisingly high incidence of macroprolactinemia may be the kind of samples we receive. As a reference laboratory, we receive samples for confirmation of an unexpected high value detected elsewhere. The PEG precipitation technique that we use is a convenient and simple procedure to screen for the presence of macroprolactinemia. Recoveries .65% are indicative of the absence of, and those ,30% indicative of the presence of, macroprolactinemia. Samples with recoveries between 30% and 65% need gel filtration chromatography to characterize the predominant molecular form. Our data, in spite of the special characteristics of the samples studied, stress the necessity that laboratories offering PRL assay service should include, in all reported increased PRL results, some form of study for the presence of macroprolactinemia, as already suggested by Lindstedt (6).