The importance of sleep monitoring in the diagnosis of hyperprolactinemic syndromes

Prolactin (PRL) has a particular circadian secretory pattern characterized by a sleep-dependent rising plasma level. A relationship with non-rapid eye movement sleep (NREMS) has been found with a progressive secretion increase during sleep despite the waning occurrence of NREMS in late sleep. However, it is not clear if the PRL rises are correlated with stages I and 2, stages 3 and 4 (slow wave sleep, SWS), or with NREMS as a whole. In hyperprolactinemic syndromes neuroradiological and dynamic secretory invesiigations do not always give a clear etiological diagnosis. The aim of this study was to show that in these syndromes, in addition to the above investigations, the evaluation of PRL circadian plasma levels associated with sleep monitoring can improve the diagnostic possibilities. Moreoever, we have attempted to ascertain to which stages of NREMS the PRL rises are linked. Subjects and methods: Thirty-four subjects were examined (mean age, 35 years; range, 18 - 53 years): 11 healthy volunteers (3 males, 8 females) and 23 hyperprolactinemic patients (5 males, 18 females). All the patients underwent extensive clinical and neuroradiological examinations. Immunohistochemical and microscopic investigations were also carried out on biopsy or post-mortem specimens. After two adaptation nights, sleep was polygraphically monitored (EEG, EOG, EMG) from 22.00 h to 07.00 h and scored according to standard criteria. Each subject underwent hourly (day) and half-hourly (night) blood sampling; PRL plasma levels were determined by radioimmunoassay. The PRL secretive curves were calculated by the single cosinor and group mean-cosinor methods. Such methods (24 h cosine function of a continuous curve fitted by least-squares analysis) provide both the properties of the circadian parameters and the curve statistical signifìcance (i.e., the presence of a rhythm) in the case of an individual or a group respectively. Results: Of all the subjects, 24 were considered to have had an effìcient sleep showing an effìciency index (i.e., the ratio: SWS + REMS/total sleep time) greater than 0.35 (normal range, 0.35-0.60). Furthermore, they allow the detection of a PRL circadian rhythm (F-value).On the basis of a normal or increased PRL plasma level and according to the presence or absence of a PRL circadian rhythm, three groups were distinguished. Group I was represented by 8 volunteers considered as controls. The patients of group II were 8 hyperprolactlnemics with PRL rhythm and were represented by: 3 with a non-secreting pituitary adenoma with extrasellar extension, 2 with a probable post-traumatic lesion of the stalk, 2 with involvement of the stalk by granulomatous nodules in the course of sarcoidosis and histiocytosis X respectively, and 1 with a primary empty-sella syndrome. In addition to hyperprolactinemia all these patients had diabetes insipidus, suggesting a disconnection of the hypothalamic-pituitary axis due to stalk pathology. An analysis of variance showed a significant difference (F 6.96, df = 1,14, P < 0.02) between group l and group II as far as acrophase was concerned, with a mean delay of 161 min in group II. The patients of group III were hyperprolactinemics without PRL rhythm. In all these-patients there was neuroradiological evidence of microadenoma and a histological diagnosis of PRL-secreting tumours. The last ten subjects (3 volunteers and 7 hyperprolactinemics) did not have an efficient sleep (i.e., the efficiency index was inferior to 0.35). Furthermore, they did not present a PRL rhythm and therefore were distinguished from the others. In volunteers an acrophase shifting and an amplitude reduction were also observed. Discussion: Our data demonstrates the need for a polygraphic study of sleep to show the presence of a PRL circadian rhythm. In the 3 normal subjects in whom such a rhythm did not reach significance the sleep efficiency index was not higher than 0.35. The value of such an index can only be estimated with polygraphic recording of all the stages of sleep. In view of the efficiency index parameters and that the PRL rises are related to NREMS, it is reasonable to deduce that such rises are quite probably connected with SWS. In hyperprolactinemic syndromes if the PRL rhythm is preserved but with a significant delay of acrophase, one can infer a stalk pathology. Such a delay may be explained by the hypothalamic PRL inhibiting factor (PIF) reaching the adenohypophysis via the systemic circulation. On the other hand, hyperprolactinemia associated with a lack of PRL rhythm points to a diagnosis of prolactinoma, which is characterized by an autonomous secretion as can be also deduced by unresponsiveness to both stimulatory and inhibitory agents. In our series of patients with hyperprolactinemia the PRL mean plasma levels were shown to be higher in subjects with stalk pathology than in those with microadenoma. This was due to a patient with stalk pathology (non-secreting pituitary adenoma with extrasellar extension diagnosed with histological, immunohistochemical and ultrastructural methods) having a particularly elevated PRL level (mesor: 1.314 ng/ml). For unknown reasons very high PRL levels are sometimes observed in such a pathology, as reported also by others, even if in this case the value of the hyperprolactinemia is exceptional.