The Thyroid and its Diseases
	Chapter 14                                                               HOME Thyroid Regulation and Dysfunction in the Pregnant Patient Revised October 9, 2004 by Dr. Daniel Glinoer

MATERNAL THYROID PHYSIOLOGY

Numerous hormonal changes and metabolic demands occur during pregnancy, resulting in profound and complex effects on thyroid function. As thyroid diseases are, in general, much more prevalent in women during the childbearing period (than in men), it is not surprising that thyroid disorders such as chronic thyroiditis, hypothyroidism, Graves' disease, etc are relatively common in pregnant women. To facilitate our understanding of pathologic processes that affect the thyroid gland, it is important to understand first the normal physiologic processes that take place in the pregnant state such as, for instance, the changes in thyroid function tests, thyroid volume, immune modulation, thyroidal economy in relation with the iodine nutrition status, etc. Over the past 10-15 years, a profusion of relevant new information regarding the relationships between pregnancy and the thyroid gland has allowed to clarified many aspects of the interactions between gestational processes and regulation of the thyroid system, both in normal individuals and patients with thyroid disorders. This field has expanded rapidly to become a normal part of texbooks in endocrinology, gynecology-obstetrics, and thyroidology, as well as a customary topic in major international congresses. Despite the new insights gained over the last decade, many uncertainties remain and important questions remain incompletely understood. Finally, it is important to consider that the expecting mother is the natural vector allowing for a future child to be born. Hence, a better understanding of the complex maternal-fetal interrelationships related to the ongoing thyroid processes must remain our constant quest, in order to ensure the best possible health status of mother and progeny.

Global regulation of the thyroid in normal pregnancy

Table 14-1 summarizes the main changes that occur during a normal pregnancy, and which relate to thyroid function or thyroid function testing. Early in pregnancy there is an increase in renal blood flow and glomerular filtration which lead to an increase in iodide clearance from plasma ^1-5. This results in a fall in plasma iodine concentrations and an increase in the iodide requirements from the diet ¹. In women with iodine sufficiency there is little impact, in terms of thyroid function, of the obligatory increase in renal iodine losses, because the intrathyroidal iodine stores are plentiful at the time of conception and remain unaltered throughout gestation. As an example, in a collaborative study between the Universities of Massachusetts (USA) and Santiago (Chili), iodine metabolism was investigated in the 1^st, 2^nd, 3^rd trimesters of gestation and again 1-10 months after delivery. Plasma inorganic iodide (PII) concentrations, urinary iodide levels (24-hr urine collections) and thyroid function (total T4 or FTI, TG, and TSH) were measured in 16 pregnant women. The results showed a wide variability in PII values (even within the same individuals) and in urinary iodide concentrations, but there was no trend for PII concentrations to be depressed during pregnancy. Mean iodine excretion ranged from 459 to 786 µg/day, respectively in postpartum and 3^rd trimester. The authors concluded that in iodine-sufficient regions, pregnancy does not have a major influence on circulating iodine concentrations. It should be noted, however, that the levels of iodine excretion were unusually elevated in this study, reaching 4-8 fold the iodine levels found in European non iodine-supplemented pregnant women, even when they reside in regions such as Finland and Ireland, which are known to be iodine-sufficient ⁶.

In regions where the iodine supply is borderline or low, the situation is clearly different and significant changes occur during pregnancy ^2,3,5. While 24-hr radioiodine uptake determinations are not performed in the pregnant state, past studies have shown this is increased ⁷. In addition, there is a further increment in the iodine requirements, due to transplacental iodide transport necessary for iodothyronine synthesis by the fetal thyroid gland, which becomes progressively functional after the 1^st trimester. The need for increased iodine requirements has been convincingly demonstrated in several studies that have well documented the notion that, when a pregnancy takes place in conditions with borderline iodine adequacy, significant increments in both the maternal and fetal thyroid volume occur, if no supplemental iodine is given during early pregnancy ^8-10. This is in sharp contrast with the notion that in iodine-sufficient areas, there is little, if any, change in thyroid size during pregnancy ^11,12.

Effects of human chorionic gonadotropin on thyroid function

Human chorionic gonadotropin (hCG) is a member of the glycoprotein hormone family that is composed of a common a-subunit and a hormone-specific b-subunit, non-covalently associated. The a-subunit of hCG consists of a polypeptide chain of 92 amino acid residues containing two N-linked oligosaccharide side-chains. The hCG b-subunit consists of 145 residues with two N-linked and four O-linked oligosaccharide side-chains. The TSH b-subunit is composed of 112 residues and one N-linked oligosaccharide. The b-subunits of both possess 12 half-cysteine residues at highly conserved positions. Three disulfide bonds form a cystine knot structure, which is identical in both TSH and hCG and is essential for binding to their receptor (LH and hCG bind to the same receptor) ^13bis,241. A single gene on chromosome 6 encodes for the common a-subunit, while the genes that encode for the b-subunits are in a cluster on chromosome 19, with seven genes (but only three actively transcribed) coding for b-hCG ^14bis.

The partial structural homology between hCG and TSH provides an indication that hCG may act as a “thyrotropic” hormone, by overlap of their natural functions. It has indeed been established that hCG does possess an intrinsic (albeit weak) thyroid-stimulating activity and perhaps even a direct thyroid-growth-promoting activity ^13-20. During a normal pregnancy, the direct stimulatory effect of hCG on the thyroid induces a small and transient increase in free thyroxine levels near the end of the 1^st trimester (peak circulating hCG) and in turn a partial TSH suppression ^2,5,21-27. In bioassays, hCG is only about 1/10⁴ as potent as TSH during normal pregnancy. This weak thyrotropic activity explains why, in normal conditions, the effects of hCG remain largely unnoticed and thyroid function tests unaltered.

The thyrotropic role of hCG in normal pregnancy is illustrated in Figure 14-1. The upper figure shows the inverse relationship between serum hCG and TSH concentrations, with a mirror image between serum TSH (nadir) and peak hCG levels et the end of the 1^st trimester. The lower figure shows that the rise in free T4 is proportional to peak hCG values. At this time of gestation, one fifth of otherwise euthyroid pregnant women may have a transiently lowered serum TSH, even below the lower limit of the normal range ^2,20,28.

Figure 1. Upper graph, Serum TSH and hCG as a function of gestational age in 606 healthy pregnant women. Between 8 and 14 weeks gestation, the changes in in hCG and TSH levels are mirror images of each other, and there is a significant negative correlation between the individual TSH (nadir) and peak hCG levels (P<0.001). Lower graph, Scattergram of free T4 levels in relation to hCG concentrations in the first half of gestation. When peak hCG values are plotted for 10.000 IU/L increments in circulating hCG, the figure shows the direct relationship with progressively increasing free T4 levels in healthy pregnancies. (Reproduced by permission of Glinoer et al.; Journal of Clinical Endocrinology and Metabolism 71:276, 1990; Ref 2)

Recently, an interesting study of desialylated and deglycosylated hCG, in an experimental setting using T3 secretion as the response parameter (in a serum-free culture system with human thyroid follicles), showed that the removal of the sialic acid or the carbohydrate residues from native hCG transformed such hCG variants into thyroid stimulating superagonists ²⁹. Further evidence supporting a pathophysiological role of hCG to stimulate the human thyroid gland is found in studies of patients with hydatidiform mole and choriocarcinoma (see Chapter 13). In these conditions, clinical and biochemical manifestations of hyperthyroidism often occur and, as expected, the abnormal stimulation of the thyroid is rapidly relieved after appropriate surgical treatment ^30-32.

Hyperemesis gravidarum

Hyperemesis gravidarum (HG) occurs in about 1.5 % of pregnancies and is probably more prevalent in Asian women than in Caucasians. It is characterized by prolonged and severe nausea and vomiting in early pregnancy that leads to a loss of 5 % body weight, dehydration and ketosis. There are many reports of series of hyperemesis patients whose thyroid function has been studied (see recent review by Hershman) ^13bis. Even though all studies did not show an association between HG and an abnormal thyroid function, many of them did, and we can therefore conclude that HG is significantly associated with abnormal increases in free T4 and free T3, and also with abnormally elevated serum levels of hCG ^33-39. Perhaps up to two thirds of pregnant women presenting HG (in some reports, at least) have biochemical features suggesting hyperthyroidism and the more seriously ill among them also show an elevation in serum free T3 ^35,36. This leads to transient thyrotoxicosis which, together with the generalized illness associated with this syndrome, may make the differentiation of this transient thyrotoxic condition from Graves' disease difficult ^40,41,42.

The etiology of HG is unknown. It may be related to the high hCG level and possibly some action of hCG that is still unclear. The increased thyroid function results from an exaggerated stimulation of the thyroid gland, mainly in the 1^st (and sometimes also the 2^nd) trimester, and is attribulted to the effects of hCG to stimulate directly the TSH receptor. This may even extend to an elevation in serum free T3, despite the fact that these patients are often nutritionally compromised (by abundant vomiting) and one would expect to have a low T3 due to impairment of T4 to T3 conversion.

In virtually all patients, appropriate fluid replacement will lead to resolution of the clinical symptoms and, as gestation proceeds and hCG levels fall, normal thyroid function is progressively resumed. In severe (but rare) cases, antithyroid drug treatment may be required (as described in more detail below). In occasional patients, hyperemesis is a symptom of coexistent thyrotoxicosis due to Graves' disease ^37,40-43. In such cases, appropriate treatment for hyperthyroidism is obviously required. Several investigators have observed that there may be an even more subtle form of hyperthyroidism associated with morning sickness ^35,36,41. They showed that the severity of emesis correlated with the levels of free T4 and hCG (and TSH suppression), suggesting strongly that hyperemesis reflects the extreme of the spectrum of the physiological changes that occur in normal pregnancy (see Figure 14-2). One tempting hypothesis to conceptually correlate the two is to consider that higher hCG levels cause both increased estrogen secretion as well as thyroid hyperfunction, hence explaining the coexistence of nausea and vomiting on the one hand and hyperthyroidism on the other ⁴⁰.

A recent study compared the charge-isoforms profiles of circulating hCG in pregnant women from different ethnic backgrounds (Samoan vs European) with hyperemesis gravidarum (HG) ⁴⁴. The results confirmed an increase in total serum hCG concentrations as well as an increase in the proportion of acidic hCG variants in women suffering from HG, compared with matched control subjects. The same study also confirmed the association between the hCG concentrations reached in early pregnancy and the elevations in thyroid hormone levels ⁵. Finally, while there was no major association between HG and ethnic background, the authors observed a high prevalence of recurrent HG and a familial predisposition for this condition, suggesting that either long-term environmental factors or genetic factors may play a crucial role in the pathogenesis of HG and gestational transient non autoimmune thyrotoxicosis.

Changes in circulating thyroid hormone binding proteins

The increase in total serum T4 and T3 that occurs during pregnancy is due to an increase in serum thyroxine binding globulin (TBG) concentrations. Changes in TBG happen early and, by 16-20 weeks of gestation, TBG concentrations have doubled (see Figure 14-3) ^2,5. The cause of the marked increase in serum TBG is probably multifactorial. The early studies showed that TBG biosynthesis was increased in primary cultures of hepatocytes from Rhesus monkeys, when primed with estradiol ⁴⁵. However, the lack of increase of other binding proteins (such as CBG and SHBG) by estrogen in HEP-G2 cells raised the possibility that other factors might be operative in the pregnant state. Studies of the changes in the glycosylation patterns of TBG, induced by estrogen, have indicated that the increase in circulating levels of TBG was probably be due in a large part to a reduction of its plasma clearance (see also Chapters 3, 5) ⁴⁶. There is a marked increase in the more heavily sialylated fractions of TBG in the sera of pregnant or estrogen-treated individuals. This increase in TBG’s sialic acid content inhibits the uptake of the protein by specific asialylo-glycoprotein receptors on hepatocytes, and the more heavily sialylated proteins from pregnant sera have therefore a longer plasma half-life ⁴⁷. Such alterations in sialylation is not found in TBG isolated from patients with congenital TBG elevation, the latter being due to a true over-production of the protein ⁴⁸. Thus, in addition to the stimulatory estrogen effects of estrogen on TBG synthesis, a major contribution to the increased TBG concentration during pregnancy is the reduced clearance of the protein. This explanation is attractive since it would also also account for the increases observed in concentrations of other circulating glycoproteins in hyperestrogenemic states. Delivery leads to a rapid reversal of this process and serum TBG concentrations return to normal within 4-6 weeks. With that, serum T4 and T3 also return to pregestational serum levels. In addition to the 2 to 3-fold increase in serum TBG, modest decreases in both serum transthyretin (TTR) and albumin are commonly found in pregnancy, but the physiological impact of these changes, if any, is unknown ⁴⁹.

An interesting case report was published at the end of last year by Zigman and coworkers. It relates the story of a 42-year-old woman, with both established hypothyroidism and TBG deficiency, who the authors have followed through two full-term pregnancies. This patient had a baseline TBG that was approximately 70% below the average baseline for non-TBG-deficient women. During pregnancies, her TBG rose, although her absolute TBG level remained only half the increase usually associated with a normal pregnancy. The authors also noted that despite the patient’s low baseline TBG level and blunted pregnancy-associated rise in TBG, the patient required increases in thyroxine replacement dosage that mirrored those found in non-TBG-deficient hypothyroid pregnant women. They suggested therefore that an increase in TBG concentration is not the key determinant for the increase in thyroxine requirement in pregnancy ^49bis. In a follow up letter to the Editor in the same Journal, an alternative explanation was proposed by us ^49ter. In the normal situation before pregnancy, homeostasis of thyroid function is ensured by the following parameters: a serum total T4 of ~100 nmol/L and TBG concentration of ~260 nmol/L. This equilibrium implies, in turn, that ~75 % of the circulating T4 is bound to TBG and also that ~35-40 % of the circulating TBG is saturated by T4. During a normal pregnancy, the extracellular TBG pool expands from ~3,000 to ~7,000 nmol/L. Thus, for homeostasis of free thyroid hormones to be maintained, the extrathyroidal total thyroxine pool must follow the expansion in parallel, and this can only be achieved by the thyroid gland filling up the progressively the increasing hormonal pool during the first half of pregnancy. In the unusual case reported by Zigman et al., when this partially TBG-deficient patient was not pregnant, her serum total T4 was ~70 nmol/L and TBG ~80 nmol/L, indicating that her TBG was almost fully saturated by T4 (around 90% saturation), because of her severe TBG binding capacity restriction. However in the non pregnant condition, only a relatively small fraction of the patient’s circulating T4 could be bound to TBG: ~50 %. When this patient was pregnant, her TBG deficiency was still partially responsive to estrogen induction and her TBG increased 3-fold to ~240 nmol/L and her total T4 to ~90 nmol/L. In other words, her total T4 concentrations had to be raised by ~30 % (by an increase in her thyroxine replacement dose), hence allowing to restore a TBG binding saturation level by T4 of ~35 %, that is equivalent to what is observed at the onset of pregnancy in an non-TBG-deficient woman. There is therefore, in our opinion, no real surprise that the increment required in her l-T4 dosage was also 30 %, that is exactly the same proportion that what could have been anticipated from the rise in serum TBG during pregnancy.

Figure 3. Upper graph, Serum TBG as a function of gestational age. The data were obtained from individual serum TBG measurements in 585 euthyroid healthy women with normally progressing pregnancies. Each point represents the mean (with SEM) of TBG determinations at weekly intervals in the cohort of pregnant women. There was a significant correlation between serum TBG and gestation time, from the 5th to the 20th gestational week (r=0.60; P<0.001) Lower graph, Serum E2 levels determined on a weekly basis in 246 normally progressing pregnancies, between the 5th and 12th week of gestation. Even though there was a significant correlation between serum E2 and gestation time (r=0.40; P<0.001), the figure illustrates the individual variability in the progression of E2 concentrations in the early stages of pregnancy: between 5-12 weeks, 22% of E2 levels were below 500 ng/L and 42% below 1000 ng/L. (Reproduced by permission of Glinoer; Endocrine Reviews 18:404, 1997; Ref 5).

Increased plasma volume

The increased plasma concentration of TBG, together with the increased plasma volume, results in a several-fold increase in the total T4 pool during pregnancy. While the changes in TBG are most dramatic during the first trimester, the increase in plasma volume continues until the time of delivery. Thus, for the free T4 concentration to remain unaltered, T4 production rate must increase (or its degradation rate decrease) to allow for the additional T4 to accumulate. One would predict that in a situation where the T4 input was constant, there would be an iterative increment in T4 as TBG increases, due to reduced T4 availability to degradative enzymes. The evidence that L-thyroxine requirements are markedly enhanced during pregnancy in hypothyroid treated women (see below) strongly suggests that not only T4 degradation is decreased in early pregnancy but also that an increased T4 production must occur throughout gestation, for maintaining the homeostasis of free T4 concentrations ^5,50-53.

Thyroxine production rate

The only direct measurements of T4 turnover rates in pregnancy were obtained more than 30 years ago by Dowling et al ⁵⁴. In eight pregnant subjects (4 in the 1^st half and 4 in the 2^nd half of gestation), T4 turnover rates were estimated not to be significantly different from those of non-pregnant subjects. However, based on several considerations discussed above from more recent work (see Table 14-1), we can now conclude that T4 production rates truly are enhanced during pregnancy. Good evidence supporting these conclusions has arisen, for instance, from the analysis of L-thyroxine administration in pregnant women with hypothyroidism. In 9/12 women with primary hypothyroidism who received stable L-thyroxine doses, there was a significant increase in serum TSH during gestation, requiring an compensatory increase in thyroxine dosage to restore euthyroidism ⁵¹. In this study, the few patients who did not require an increase in L-thyroxine dosage were receiving slightly excessive replacement doses prior to gestation. If the increased levothyroxine dose was maintained into the postpartum period, there was a subsequent increase in free T4 and a decrease in TSH (Figure 14-4). These results showed that there was an increase in T4 requirements, beginning already in early gestation, and which persisted until delivery. In an other study, L-thyroxine replacement was evaluated in two groups of hypothyroid patients during pregnancy; one group had Hashimoto's disease while the second had had thyroid ablation for either Graves' disease or thyroid carcinoma ⁵². While the patients with thyroid ablation (because no residual tissue was present) required a 45% increase in L-thyroxine dosage to maintain euthyroidism during gestation, only a 25% increase was necessary in those with Hashimoto's disease (because some functional thyroid tissue was still present). From the point-of-view of maternal thyroid function during pregnancy, it is now accepted that there is a 30-50% increase in T4 production during gestation (reviewed in Ref. 5).

Figure 4. Individual serum free-thyroxine indexes and thyrotropin (TSH) concentrations during the third trimester and postpartum in women with hypothyroidism. Results are shown for seven patients whose measurements were made within 10 weeks before and 20 weeks after delivery. Each patient is represented by a single line. The normal ranges for serum TSH and the free thyroxine index are indicated by the vertical bars. (Reproduced by permission of Mandel et al. ; New England Journal of Medicine 323:91, 1990; Ref 51).

Thyroid function parameters in normal pregnancy

Total and free T4 and T3

The increase in TBG during gestation causes an increase in total serum thyroid hormones. To estimate the free hormone concentration, a thyroid hormone binding ratio (THBR), free T4 index, or direct free T4 measurement must be obtained ^5,22,23,55. Because the reduction in the free T3 fraction is approximately equal to that of T4, the standard approach for these determinations (employing T3 as a tracer) can still be used. However, it is important to recognize that as the free fraction is reduced, the resin T3 uptake (and similar assessments of free hormone fractions) asymptotically approach a fixed lower limit. This is not linearly related to the increase in unoccupied TBG binding sites. Thus, the decrease in the THBR does not usually match the quantitative derease in the T4 and T3 free fractions estimated directly and in some sera, the free T4 index will end up being slightly elevated relative to true free T4 or T3. Direct measurements of free T4 using the older "analogue" technologies often resulted in a decreased free T4 estimate in euthyroid pregnant subjects. These artifacts have been attributed to the influence of the physiologic serum albumin decrease which commonly occurs in pregnancy. Nowadays however, many direct assays are routinely available, which provide accurate estimates of the free hormone concentrations (see review in Ref. 5).

Serum TSH

As already noted, serum TSH values decrease slightly during the first trimester in response to elevations of hCG. In approximately one fifth of healthy pregnant women, serum TSh values at the time of peak hCG concentrations may be transiently lowered to subnormal values ^2,4,20,32. Thus, such a lowering in serum TSH should not automatically lead to a suspicion of hyperthyroidism. During the remainder of a normal pregnancy, serum TSH returns to the normal range of 0.4-4.0 mU/L.

Other tests

Isotopic tracers should not be administered during pregnancy and therefore the altered iodine kinetics in the pregnant patient will not be a source of confusion. In evaluating the clinical status during gestation, it should be recalled that many physical findings suggestive of mild hyperthyroidism are frequently present, including increased pulse pressure, tachycardia, heat intolerance, and decreased peripheral vascular resistance.

Placental metabolism of thyroid hormones

The placenta contains high concentrations of the Type 3 or inner-ring (5) iodothyronine deiodinase ^56-58. The inner-ring deiodination of T4 catalyzed by this enzyme is the source of high concentrations of reverse T3 present in the amniotic fluid. Reverse T3 levels parallel maternal serum T4 concentrations ^59-62. This enzyme may function to reduce the concentration of T3 and T4 in the fetal circulation (the latter being still contributed by 20-30 % from thyroid hormones of maternal origin at the time of parturition), although fetal tissue T3 levels can reach adult levels due to the action of the Type 2 deiodinase (see Chapter 15) ⁵⁰. The Type 3 deiodinase may also indirectly provide a source of iodide to the fetus via iodothyronine deiodination. However, despite the presence of placental Type 3 deiodinase in circumstances in which fetal T4 production is reduced or maternal free T4 markedly increased, transplacental passage occurs and fetal serum T4 levels are about one third of normal ⁶³. Thyroxine is also detectable in amniotic fluid prior to the onset of fetal thyroid function ⁶⁴.

Figure 5. The steep gradient between maternal concentrations of thyroid hormones to those measured in the coelomic fluid and in the amniotic cavity with the developing embryo during the early stages of gestation. (Adapted from Ref 66).

Figure 14-5 depicts the steep maternal to fetal gradient of total T4 concentrations in early pregnancy stages : between 6 and 12 weeks of gestation, if the maternal total T4 concentration is set to represent 100 %, the total T4 in coelomic fluid would represent 0.07 % and T4 in the amniotic cavity as little as 0.0003-0.0013 % of maternal concentrations. Thus, the placental barrier to maternal iodothyronines, even in the 3^rd trimester, appears not to be impermeable to the transplacental passage of thyroid hormones of maternal origin ^50,65. Even though very small quantitatively, such concentrations may qualitatively represent an extremely important source of thyroid hormones to ensure the adequate development of the feto-maternal unit ⁶⁶.