ENDOCRINE ASPECTS OF HUMAN DECIDUA AND EARLY PREGNANCY FAILURE

Abstract

Prolactin (PRL) producing capacity, its regulation and function were studied in explants of decidua compacta and decidua spongiosa obtained from 61 patients undergoing termination of pregnancy at 6 to 12 weeks of gestation. In vitro PRL producing capacity, expressed as mIU/g protein, of decidua compacta was significantly higher (p < 0.05) than those of decidua spongiosa. Production of PRL increased with gestation from 6 to 12 weeks, with a more rapid rate at the later gestation. The pattern of increase fitted significantly (p < 0.0001) to the exponential model for both decidua compacta and decidua spongiosa. The exponential regression equations for decidua compacta and decidua spongiosa were In y = 4.21 + 0.20x and In y = 2.77 + 0.31 x respectively. Hence, although both had a similar pattern of increase, the rate of increment was significantly greater in decidua spongiosa than in decidua compacta. These findings suggest that separating decidua compacta and decidua spongiosa of the first trimester would reduce the heterogeneity of decidual explants. We employed this approach to study the regulation of decidual PRL production that followed. Our data show that first trimester decidua compacta was more endocrine-sensitive than decidua spongiosa. Both human chorionic gonadotrophin (hCG) and oestradiol (E2) have significant inhibitory effects on PRL secretion by decidua compacta but not by decidua spongiosa. On the other hand, progesterone (P4) and 170:-hydroxyprogesterone (17OHP) have no effect on the decidual PRL secretion. The functions of decidual tissue, PRL and its interactions with trophoblast tissue were also determined in this study. We noted that trophoblast P4 production was significantly modulated by PRL (stimulatory, p < 0.01) and decidual factors (inhibitory, p < 0.001). In addition, trophoblast hCG production were significantly stimulated by decidual tissue or decidual factors (p < 0.05). Later experiment confirmed that decidual PRL was one of the stimulators. Hence, we offer strong evidence that trophoblast P4 and hCG production is modulated in part by decidual tissues, and decidual PRL production is modulated in part by trophoblast tissues. This proves that paracrine regulation of decidua and trophoblast exists. In the second part of this thesis, serum levels of hCG, PRL, 170HP, P4 and E2 were measured in 474 women with either normal pregnancy or pregnancy complicated with vaginal bleeding (i.e. threatened, incomplete and missed abortions) and ectopic pregnancy. The hCG, P4 and E2 measurements were also combined to determine the impact of applying additional diagnostic tests. This was done by multiplication of these hormone concentrations. Since these hormones were of lower values in threatened and ectopic pregnancies (Barnea et al 1986; Guillaume et al 1987, 1990), this approach permitted detection of greater deviations in hormone levels in normal and abnormal pregnancies. To account for the variation in the hormone levels measured al 5 to 13 weeks of gestation, all hormone concentrations and their multiplication results were expressed as multiples of the median (MoM). The endocrine profile of those who had first trimester vaginal bleeding but went on to delivery were noted to be normal, except that mean P4 levels were slightly lower than those in normal gestation. Significantly lower levels of hCG, 170HP, P4 and E2 were observed in spontaneous abortions, miscarriages and ectopic pregnancies (p<0.01). To determine clinical usefulness of these estimations in the prediction of early pregnancy outcome, frequency distribution curves and receiver operating characteristic (ROC) curves which uses the concept of sensitivity and specificity were derived for the respective hormones measured. Receiver operating characteristics analysis indicated that the best discriminatory hCG concentration in the prediction of miscarriages was 0.21 MoM, which gave a sensitivity of 90.3% and a specificity of 86.4%. A cut-off level of hCG < 0.47 MoM gave a very high sensitivity (96.1%) but the specificity was sacrificed (75.9%) in the prediction of ectopic pregnancy. In contrast with other authors' findings (Witt et al 1990; Cowan et al 1992; Stern et al 1993), our data show that E2 estimations were better than P4 estimations, having a higher sensitivity (91 .4% and 88.2% for miscarriage and ectopic pregnancy, respectively) and specificity (86.0% and 88.2% for miscarriage and ectopic pregnancy, respectively). However, serum 17OHP and also PRL were found to have no diagnostic significance. This study indicates that a combination of two of three tests (e.g. hCGxP4) was better than a single hCG, P4 or E2 test. However, ROC analysis demonstrated that the two best tests were hCGxE2 and hCGxP4xE2. HCGxE2 test was preferred since it is easier to perform and less expensive. Thus, we conclude that hCGxE2 measurement during first trimester may offer a valuable adjunct to existing methods of diagnosis of non-viable pregnancy and ectopic gestation.