PRL exists in different forms in human serum. The predominant form is little PRL (molecular mass 23 kDa) with smaller amounts of big PRL (molecular mass 50--60 kDa) and at times big big or macroprolactin (molecular mass 150--170 kDa). The frequency and clinical consequences of macroprolactinemia have not been clearly established, mainly because of difficulty in identifying these patients biochemically. This previously required the use of gel filtration chromatography, which could not be used routinely. Recently, a screening test using polyethylene glycol (PEG) has been used to identify macroprolactin in serum. Consequently, this study was designed to examine the use of PEG precipitation in the identification of patients with a predominance of macroprolactin and to establish the clinical characteristics of such a cohort. Over 12 months, 18,258 requests for serum PRL were received and of these 1225 patients had a serum PRL more than 700 mU/L. A total of 322 of these patients (26%) had a percentage recovery after PEG precipitation of less than 40%, thus indicating the presence of a predominance of macroprolactin. Fifty-five of these patients were referred for detailed clinical assessment. Symptoms typical of hyperprolactinemia were not common in this cohort. None had sustained amenorrhea and eight have had oligomenorrhea at age less than 40 yr. One had galactorrhea. All had pituitary imaging, and four had a microadenoma with none having a macroadenoma. PEG precipitation allows easy identification of macroprolactin in routine clinical practice. As the clinical consequences of this entity at this stage seem relatively benign, referral and intensive investigation of these patients may not be necessary. However, follow-up of a large cohort is required to ensure that the long-term outlook is likewise benign. This would have important implications for both patients and healthcare systems.

The 1-3% detection rates of three recently published series have not been realized at our center where we studied a group using criteria making patients more likely to have hypercortisolism. Our results do not support the validity of screening patients without clinical features of Cushing's syndrome in the diabetes clinic.

During prolonged follow-up, no symptomatic progression was noted in any of our patients. This study suggests that patients with macroprolactinemia and normal concentrations of monomeric prolactin can be reassured, and extended endocrine review of such patients is not required.

(n = 10). Each patient's series of serum aliquots was thawed and assayed in triplicate in one assay run with dilution in zero standard for samples with free {j-hCG greater than 50 lUlL. RESULTSOur results (Table 1 and Fig. 1) showed that the free (j-hCG concentration increased significantly when samples were left unseparated for 6 h or more. An average increase of 14% over basal levels was seen at 24 h with the rise reaching 43% for 4-day-old specimens. Such artefactual increases could lead to significant errors in reported Blood samples were collected from informed pregnant women attending a routine out-patient clinic, and each specimen was ali quoted into six plain glass bottles and allowed to clot. The first aliquot was centrifuged after 30 min and the serum was separated and stored at -20 D C . The remaining samples were left to stand at room temperature for 6, 24, 48, 72 or 96 h after the time of collection before centrifugation and freezing of the serum.Serum free {j-hCG was measured by a commercial solid phase two-site immunoradiometric assay [CIS (UK) Ltd, High Wycombe, UK]. The working range of the assay is 0'1-50,0 lUlL and the intra-assay coefficient of variation was found to be 5' 60/0 at a mean value of 0·79 lUlL (n = 6) and 2·3% at a mean value of 29·3 lUlL SUBJECTS AND METHODS Measurement of the free {j-subunit of human chorionic gonadotrophin (hCG) rather than the total or intact hormone in serum has recently been employed as a tumour marker,I as a prognostic indicator in in vitro fertilization programmesand as a maternal serum marker in prenatal screening for Down's syndrome.t? In 1991 Knight and Cole" suggested that serum free {j-hCG levels can rise due to slow dissociation of hCG in 'poorly stored samples of pregnancy serum', but few publications have adequately documented this phenomenon.To assess the suitability of maternal blood samples posted from general practitioners to this laboratory as part of a prenatal screening programme for Down's syndrome, we have investigated the stability of free {j-hCG concentrations in blood samples left unseparated for various lengths of time. Additional key phrases: prenatal screening; Down's syndrome; tumour markerCorrespondence: Mr Brian Sheridan.

No abstract