The first clinical reports of a placental thyroid stimulator arose from studies indicating that placental tumors that secrete very large amount of human chorionic gonadotropin (hCG), hydatidiform moles and choriocarcinomas, were associated with hyperthyroidism [summarized by Yoshimura and Hershman (1)]. In 1967, Burger reported that impure, commercial hCG had thyroid-stimulating activity in a mouse bioassay (2). Several years later my laboratory (3) and that of Nisula (4) reported that highly purified hCG had activity in this bioassay. We found that the molar TSH was hCG (4). Because hCG varies in its carbohydrate composition, there has been considerable controversy in regard to the potency of these variants. hCG lacking its C-terminal tail, therefore closely resembling LH, is a more potent TSH than native hCG (5). Desialylated basic variants of hCG are more potent in binding to the TSH receptor than fully sialylated acidic hCG (6). However, the acidic hCG has a longer half-life in the circulation than the more basic partially desialylated hCG, and this may contribute to greater thyrotropic potency in humans (7). These studies mainly arose from patients with placental tumors or from a few patients with gestational thyrotoxicosis associated with high hCG levels. The possibility that hCG plays a role in regulation of the thyroid gland during normal pregnancy was first considered when there were more sensitive immunoassays for TSH. Braunstein and Hershman (8) reported that there was an inverse relationship between TSH and hCG at about 10–12 wk of pregnancy, the time of peak hCG levels. Harada et al. (9) further advanced this concept by showing the inverse relationship more clearly and amassing data showing an increase of free T4 and free T3 associated with the peak hCG. We postulated that hCG caused an increase of free thyroid hormone levels, thus suppressing TSH, albeit within the physiological range of serum TSH. Glinoer et al. (10), in a landmark study using a more sensitive TSH measurement, further refined the mirror image temporal relationship between hCG and TSH in normal pregnancies. Other reports also contributed to the concept that hCG is a thyroid stimulator in the early part of pregnancy when hCG levels are very high (11). In this issue of the Journal, Haddow et al. (12) have done a complex analysis of the relationship between hCG, TSH, and free T4 during the first trimester (wk 11–13) and second trimester (wk 15–18) in 9562 women with singleton pregnancies. They found that higher hCG levels were associated with a lower body mass index. This fits with the observation of Goodwin et al. (13) that hCG is higher in women with nausea of pregnancy and hyperemesis gravidarum than in women who do not have any nausea, assuming that weight gain is restricted by nausea. More importantly, the data of Haddow et al. further strengthen the role of hCG as a thyroid stimulator in normal pregnancy. At wk 11–13, higher hCG correlates with higher free T4 levels, although the relationship is a shallow slope, confirming the data of Glinoer et al. (10) that in normal pregnancy hCG only modestly increases free T4. The relationship between hCG and TSH is more difficult to appreciate in the complex analysis by Haddow et al. (12). It shows clearly that low TSH is associated with high levels of hCG, but when TSH is in the higher deciles, there is no correlation with hCG. My interpretation is simplistic. Placental secretion of hCG is not regulated by thyroid hormones. However, when hCG levels are high, they increase thyroid hormone secretion, and the negative feedback at the level of the pituitary suppresses TSH release. When TSH levels are in the higher part of the normal range, or perhaps slightly elevated for pregnancy, the contribution of hCG to increase free T4 and T3, resulting in lower TSH levels, is not sufficient to lower serum TSH significantly. Perhaps evolution designed the hCG system as a backup for stimulating the thyroid gland to produce essential amounts of thyroid hormone, but this backup system did not fully evolve to perfection.
Read full abstract