REGULATION OF THYROID FUNCTION IN NORMAL PREGNANCY
Thyroidal Economy in Normal Pregnancy
Numerous hormonal changes and metabolic demands occur during pregnancy, resulting in profound and complex effects on thyroid function. Because thyroid diseases are common in women during the childbearing period, it is important to envisage both the expected changes in thyroid function tests taking place during a normal pregnancy and how pregnancy may affect the presentation and course of preexisting diseases such as thyroiditis, hypothyroidism, and Graves' disease. Over the past decade, new information regarding the relationship between pregnancy and the thyroid has permitted a major clarification of the complex interactions that take place between gestational processes and regulation of the thyroid system, both in normal individuals and in patients with thyroid disorders. Because the expecting mother is the “natural vector” allowing a future child to be born, a better understanding of the complexities in maternal-fetal interrelationships is primordial to ensure the best possible health status of both the mother and progeny. Normal changes in thyroid function are discussed in both iodine-sufficient and iodine-deficient women. In addition, those aspects of the diagnosis and treatment that must be kept in mind in the management of pregnant patients with thyroid disease are emphasized.
Early in pregnancy, renal blood flow and glomerular filtration increase, leading to an increase in iodide clearance from the plasma, and resulting in a decreased plasma iodide and an increased requirement for iodide in the diet (1,2). In pregnant women with iodine sufficiency, there is little impact of the obligatory increase in renal iodine losses, because the intrathyroidal iodine stores are plentiful at the time of conception and they remain unaltered throughout gestation (3). In pregnant women with iodine restriction, even though urinary iodine concentrations (UICs) may transiently show an early increase, there is, thereafter, a steady decrease in UIC from the first to third trimesters of gestation, hence revealing the underlying tendency toward iodine deficiency associated with the pregnant state (4).
Human chorionic gonadotropin (hCG) is a peptide hormone composed of two subunits termed the α and β chains. Whereas the α subunit is identical to that of thyrotropin (TSH), the β chains differ between both molecules and confer ligand specificity. The partial structural homology between hCG and TSH explains why hCG may act as a thyrotropic hormone, by overlap of their natural functions (“spillover” mechanism). Current evidence has clearly established that hCG is a direct thyroid stimulator and perhaps even a thyroid growth promotor (5,6,7,8). During normal pregnancy, the stimulatory effect of hCG on the thyroid induces a small and transient increase in free thyroxine (T4) near the end of the first trimester and in turn a partial TSH suppression (9,10,11). The thyrotropic role of hCG during normal pregnancy is illustrated in Figure 80.1, depicting the inverse relationship between serum hCG and TSH concentrations. At the end of the first trimester, there is a mirror image between serum TSH (nadir) and hCG (peak) levels. In one fifth of otherwise euthyroid pregnant women at this time of gestation, serum TSH may even be transiently lowered below the normal range. The figure also shows the direct relationship between circulating free T4 and hCG concentrations at the time of peak hCG values. The parallelism between peak hCG levels, the increases in free T4 and triiodothyronine (T3), and the TSH suppression provide compelling arguments for the thyrotropic role of hCG during normal pregnancy. An interesting study of desialylated and deglycosylated hCG, in a serum-free culture system using human thyroid follicles, showed that removal of sialic acid or carbohydrate residues from native hCG transformed such hCG variants into thyroid-stimulating superagonists (12). Further evidence supporting a pathophysiologic role of hCG in stimulating the human thyroid gland was found in studies of patients with hydatidiform mole and choriocarcinoma (13). In these conditions, clinical and biochemical manifestations of thyrotoxicosis often occur and, as expected, the abnormal stimulation of the thyroid is rapidly relieved after appropriate surgical treatment.
FIGURE 80.1. Top: Serum thyrotropin (TSH) and human chorionic gonadotropin (hCG) as a function of gestational age in a large group of healthy pregnant women. Between 8 and 14 weeks' gestation, the changes in hCG and TSH levels are mirror images of each other, and there is a significant negative correlation between the individual TSH and hCG levels ( < 0.001). Scattergram of free thyroxine (TpBottom:4) 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 graph shows the direct relationship with progressively increasing free T4 levels in healthy pregnancies. (Reproduced from Glinoer D, De Nayer P, Bourdoux P, et al. Regulation of maternal thyroid function during pregnancy. 1990;71:276–287, with permission.)J Clin Endocrinol Metab
The increased plasma concentration of thyroxine-binding globulin (TBG), together with the increased plasma volume, results in a severalfold increase in the total thyroxine pool during pregnancy. Although the changes in TBG are most dramatic during the first trimester, the increase in plasma volume continues until delivery. Thus, if the free T4 concentration is to remain unaltered, the T4 production rate must increase (or its degradation rate decrease) to allow the additional T4 to accumulate (2,10,14). In a situation where the T4 input is constant, it would be predicted that there is an iterative increment in T4 as TBG increases as a result of the reduced availability of T4 to degradative enzymes. The evidence that levothyroxine requirements are markedly enhanced during pregnancy in hypothyroid women clearly indicates that not only is T4 degradation decreased during early pregnancy but also that increased T4 production must occur throughout gestation.
The placenta contains high concentrations of D3, the type 3 iodothyronine deiodinase (15,16,17). Inner ring deiodination of T4 catalyzed by this enzyme is the source of high concentrations of reverse T3 present in amniotic fluid. The enzyme may function to reduce T3 and T4 concentrations in fetal circulation, although fetal tissue T3 levels can reach adult levels as the result of D2's local action. D3 also may indirectly provide a source of iodide for the fetus through iodothyronine deiodination. Despite the presence of placental D3, transplacental passage of maternal T4 occurs, and fetal serum T4 levels are about one third of normal (18). Thyroxine is also detectable in amniotic fluid before the onset of fetal thyroid function (19). Thus, the placental barrier to maternal iodothyronines, even as late as the third trimester of gestation, appears not to be impermeable to the passage of maternal thyroid hormones. Although quantitatively small, such hormone concentrations may qualitatively represent an extremely important source of thyroid hormones to ensure the adequate development of the fetomaternal unit (20,21,22,23).
The only direct measurements of thyroxine turnover rates in pregnancy were obtained more than 30 years ago, and the T4 turnover rates were estimated not to differ significantly from those of nonpregnant subjects (24). From more recent work, however, compelling arguments strongly suggest that T4 production rates truly are enhanced during pregnancy. Evidence supporting this concept was developed from analyses of the repercussions of pregnancy on levothyroxine treatment in patients with primary hypothyroidism. In women with primary hypothyroidism who were receiving stable doses of levothyroxine, there was a significant increase in serum TSH during gestation, which required a compensatory increase in levothyroxine dosage to restore euthyroidism (25,26). These results indicate that there is an increase in T4 requirements, beginning early in gestation, and persisting until the time of delivery. From the point of view of maternal thyroid function during pregnancy, it is now widely accepted that there is a 30% to 50% increase in T4 production during gestation (2).
Thyroid Function Parameters in Normal Pregnancy
The increase in total serum T4 and T3 that occurs during pregnancy is due to an increase in serum TBG. This change appears early; TBG concentrations double by 16 to 20 weeks' gestation. The cause of the marked increase in serum TBG may be multifactorial. Early studies showed an increase in serum TBG synthesis in primary cultures of hepatocytes from rhesus monkeys primed with estradiol (27). There is also an increase in the fraction of more heavily sialylated, and therefore more negatively charged, fractions of TBG in the sera of pregnant or estrogen-treated individuals. Because this increase in sialic acid content of TBG inhibits the uptake of the protein by specific receptors on hepatocytes, the more heavily sialylated proteins from pregnant sera have a longer plasma half-life (28). Thus, in addition to stimulatory effects of estrogen on TBG synthesis, a major contribution to the increased TBG concentration during pregnancy is the reduced clearance of the more highly sialylated forms of the protein. In addition to the two- to threefold increases in serum TBG, modest decreases in both serum transthyretin (TTR) and albumin are also commonly found in pregnancy (29).
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. Because the reduction in the free fraction of T3 is approximately equal to that of T4, the standard approach for these determinations using 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 the free hormone fraction) 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 usually does not match the quantitative decrease in the T4 and T3 free fractions estimated directly, and in some sera the free T4 index or estimate will end up being slightly elevated relative to the true free T4 or T3 (30). Direct measurements of free T4 using 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 that commonly occurs in pregnancy. Nowadays, however, many direct assays are routinely available that provide accurate estimates of the free hormone concentrations (31).
As already noted, serum TSH concentrations may be transiently lowered (and hence become subnormal) during the first trimester in about one fifth of pregnant women in response to elevations of hCG. Thus, such a lowering in serum TSH should not lead automatically to a diagnosis of thyrotoxicosis. During the second and third trimester, serum TSH returns to the normal range, most often between 0.4 and 2.5 mU/L.
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 thyrotoxicosis are frequently present, including increased pulse pressure, tachycardia, heat intolerance, and decreased peripheral vascular resistance.
Thyroid Function in Pregnant Women with Iodine Deficiency
Epidemiology of Iodine Deficiency during Pregnancy
In countries such as the United States, Japan, and a limited number of European countries where national programs of dietary iodine supplementation have been in place for many years, iodine deficiency (ID) disorders are believed not to present problems. However, this global view is probably optimistic. For instance, a recent national survey in the United States showed that the average iodine intake had markedly decreased, compared with a similar survey conducted three decades ago: the median urinary iodine excretion was 145 µg/L, compared with over 300 µg/L in the previous period (32). Thus, even though the present iodine intake levels in the United States may at first glance appear to be comfortably above the recommended minimum, this survey showed that perhaps as many as 15% of the women in childbearing age, and almost 7% of them during a pregnancy, had iodine excretion levels into the range of moderate ID, namely below 50 µg/L (33). Along the same lines, a recent study showed that despite national efforts to implement the mandatory use of iodized salt in Switzerland for many years, mild ID still prevailed in the Berne area (34). Another consideration is that the risk for iodine deprivation during pregnancy needs to be assessed locally, because mild to moderate ID occurs in areas that are not immediately recognized as iodine deficient. For instance, a study in the southwest of France showed that over 75% of pregnant women in the area had iodine excretion levels below 100 µg/L (35). A third concept relates to the notion that the iodine intake may vary unpredictably within a given country: this occurs particularly in regions with mild to moderate ID, because of significant variations in the natural iodine content of food and water. This was illustrated by a Danish study: pregnant women without iodine supplements had a median iodine excretion level of 62 µg/g creatine in Copenhagen, compared with only 33 µg/g creatine in East Jutland; furthermore, these striking differences were not alleviated in pregnant women who did receive iodine supplements in the same areas (36). A final concept is that ID requires constant monitoring, even after the implementation of iodine supplementation during pregnancy. Since our initial studies initiated in the early 1990s in Belgium, the majority of pregnant women nowadays receive multivitamin pills, containing 100 to 125 µg of iodine as a supplement. Despite this public health effort and improved medical awareness, a recent study of neonates in Brussels showed that their iodine nutrition status, albeit improved, was not yet normalized (37).
ID becomes significant during pregnancy when the iodine intake falls below 100 µg/day and the most recent recommended daily allowance indicates that 200 to 250 µg/day of iodine is the ideal supply for pregnant and lactating women (38). The degree of ID should be assessed in each concerned area specifically and the local situation correctly evaluated before embarking on medical recommendations for adequate iodine supplementation programs.
Enhanced Thyroidal Stimulation as the Maternal Consequence of Iodine Deficiency
Reduced iodine intake during pregnancy leads to chronically enhanced thyroidal stimulation through the pituitary–thyroid feedback mechanism and is frequently accompanied by thyroidal alterations, mainly relative hypothyroxinemia and goitrogenesis (Fig. 80.2).
FIGURE 80.2. From physiologic adaptation to pathologic alterations of the thyroidal economy during pregnancy. This figure illustrates the sequence of events occurring for the maternal thyroidal economy, emphasizing the role of iodine deficiency to enhance the stimulation of the thyroid gland. (Modified from Glinoer D. The regulation of thyroid function in pregnancy: pathways of endocrine adaptation from physiology to pathology. 1997;8:404–433, with permission.)Endocr Rev
In clinical practice, enhanced glandular stimulation associated with iodine restriction can be assessed using simple biochemical parameters. The most useful parameters are (a) relative hypothyroxinemia, with serum free T4 frequently found near the lower limit of normality; (b) changes in serum TSH (usually remaining within the normal range), with a frequent doubling of initial TSH values near term; (c) changes in thyroglobulin (Tg) concentrations, with a progressive increase as gestation progresses; and (d) urinary iodine concentrations. Examples of the successive steps in the formation of a vicious circle are given in the different graphs shown on Figure 80.3. Figure 80.3A shows the progressive lowering in free T4 in pregnant women with ID in Sudan, compared with unaltered free T4concentrations in pregnant women in Sweden (39). Figure 80.3B illustrates changes in serum TSH in pregnant women from one area with ID in Sicily, compared with iodine-sufficient pregnant women from another Sicilian area (40). Figure 80.3C illustrates the progressive increase in serum Tg levels in Belgian pregnant women who did not receive iodine supplements (10). Figure 80.3D illustrates the inverse correlation between the degree of glandular hypertrophy and the severity of iodine restriction in pregnant women in the southwest of France (35).
FIGURE 80.3. Top: Schematic representation of the formation of a vicious circle, when pregnancy takes place in conditions with iodine deficiency. The progressive lowering in free thyroxine (TA:4) concentrations in Sudanese pregnant women with iodine deficiency (•), compared with unaltered serum free T4in pregnancies in Sweden with an adequate iodine nutrition status (○). The increase in serum thyrotropin (TSH) () in Sicilian pregnancies from an area with iodine deficiency, compared with unaltered TSH levels in Sicilian pregnant women from an area with iodine sufficiency (). The progressive increase in serum thyroglobulin () levels during pregnancy in Belgian pregnant women without iodine supplementation, in first trimester (), at the end of second trimester (), and immediately after parturition (). The inverse correlation between glandular hypertrophy (%) and the severity of iodine restriction (based on urinary iodine concentration) in pregnant women in the southwest of France. [Modified from Glinoer D. Feto-maternal repercussions of iodine deficiency during pregnancy. 2003;64:37–44, with permission.]B:circlestrianglesC:TGABCD:Ann Endocrinol (Paris)
Goitrogenesis in Mothers and Fetus and Its Prevention
Gestational goiter formation affects both the mother and progeny (Fig. 80.4). In areas with a low iodine intake, several studies have shown that thyroid volume (TV) increases by 20% to 35% on the average, and approximately 10% of pregnant women develop a goiter (2,10,14,41). Concerning the progeny, it was shown that the mean TV was 40% larger in newborns from women without iodine supplementation, compared with a treated group of pregnant women. Moreover, glandular hyperplasia was already present at birth in 10% of the newborns from nonsupplemented mothers, whereas no instance of neonatal goiter occurred in the newborns from an iodine-supplemented group (Fig. 80.4B). These findings indicate that the process of goitrogenesis occurs early during fetal development in association with a low iodine intake, perhaps even as soon as the fetal thyroid gland starts to develop (42).
FIGURE 80.4. A: Thyroid volume (TV) determined by echography in 10 selected pregnant women at initial presentation in first trimester (), at delivery (), and 12 months postpartum (). The dotted line represents the upper limit of TV normality (i.e., 22 mL). (Reproduced from Glinoer D, Lemone M, Bourdoux P, et al. Partial reversibility during late postpartum of thyroid abnormalities associated with pregnancy. 1992;74:453–457, with permission.) Distribution frequencies of TVs in neonates born to mothers without (on the left) and with (on the right) iodine supplementation received during pregnancy. Iodine supplementation allowed for a marked 38% average reduction in mean TV and for the complete prevention of neonatal thyroid hyperplasia. The upper limit of normal TV in newborns is indicated by the vertical dotted lines (at 1.5 mL). (Reproduced from Glinoer D, De Nayer P, Delange F, et al. A randomized trial for the treatment of mild iodine deficiency during pregnancy: maternal and neonatal effects. 1995;80: 258–269, with permission.) Progressive changes in TV between group 0 (nulliparous) and group IV (Group I, one pregnancy; group II, two pregnancies; group III, three pregnancies; group IV, four or more pregnancies) in a retrospective study of the changes in TV in relation to parity in Italy (Naples). (Reproduced from Rotondi M, Caccavale C, Di Serio C, et al. Successful outcome of pregnancy in a thyroidectomized-parathyroidectomized young woman affected by severe hypothyroidism. 1999;9:1037–1040, with permission.) The effect of parity and smoking on thyroid size in women in Denmark. (Reproduced from Knudsen N, Bülow I, Laurberg P, et al. Parity is associated with increased thyroid volume solely among smokers in an area with moderate to mild iodine deficiency. 2002;146:39–43, with permission.)left set of barsmiddle set of barsright set of barsJ Clin Endocrinol MetabB:J Clin Endocrinol MetabC:ThyroidD:Eur J Endocrinol
An important question concerns the long-term evolution of a goiter formed during pregnancy. A gestational goiter may regress only partially after parturition; therefore, pregnancy represents one of the environmental factors that may help explain the higher prevalence of goiter and thyroid disorders in women compared with men (43,44) (Fig. 80.4A). The hypothesis of a direct relationship between parity and thyroid size has now been confirmed in a study of TV in women from a moderately iodine-deficient area in southern Italy. The researchers were able to show retrospectively a significant correlation between parity and thyroid volume (45) (Fig. 80.4C). In another recent study from Denmark, it was shown that, in addition to the effects of parity, active smoking intensified the differences in TV (46) (Fig. 80.4D). These results indicate that several environmental factors may play a role to explain goiter formation in women, that together tend to reinforce each other: iodine deficiency, successive pregnancies, and smoking habits.
In order to prevent gestational goitrogenesis, women should ideally be provided with an adequate iodine intake (≥150 µg/day) long before they become pregnant, because it is only by reaching a long-term steady state of intrathyroidal iodine stores that triggering of the thyroid machinery can be avoided, once gestation begins. To achieve such a goal, national public health authorities need to develop iodine supplementation programs of the population's diet. In the meantime, the most appropriate preventive and therapeutic approach to avoid gestational goitrogenesis is to systematically increase the iodine supply as early as possible during gestation and to continue after parturition in the mothers who plan to breast-feed. This can easily be achieved by the use of multivitamin pills containing appropriate amounts of iodine supplements. When iodine supplementation is implemented early and maintained throughout gestation, it allows for the correction and almost complete prevention of gestational goitrogenesis (42,47,48,49,50). How much supplemental iodine should be given to prevent goiter formation remains a matter of local appreciation, and depends mainly on the extent of the preexisting iodine deprivation. The ultimate goal is to restore and maintain a balanced iodine status, and this can be achieved with 100–200 µg iodine given as a daily supplement. It should be remembered, however, that with long-standing iodine restriction in the diet before pregnancy, a lag period of about one trimester is inevitable before benefits of iodine supplementation to improve thyroid function are observed.
AUTOIMMUNE THYROID DISEASE AND PREGNANCY
Successful implantation of the fetal allograft is permitted by specific modulation in the maternal immune surveillance system. Progesterone (P) and estrogen (E) exert these immunomodulating effects. Whereas progesterone decreases reactivity of both the humoral and cellular arms of the immune system, estrogen exerts opposite effects. Because pregnancy is characterized by an overall increase in the P:E ratio, the reactivity of both arms of the immune system is dampened (51). These and other not yet completely elucidated factors lead to partial tolerance and a dominant T-helper subset 2 (Th2) immune profile, which explain the positive influence of pregnancy on the clinical course of many (albeit not all) autoimmune diseases and, characteristically, the generalized improvement of thyroid autoimmune diseases during pregnancy. The precise mechanism by which thyroid autoantibodies as well as those directed against other tissues, partially suppressed during pregnancy, are exacerbated after delivery remains obscure. Presumably, the rapid reduction in immune suppressor functions following delivery leads to the reestablishment and exacerbation of these conditions (52,53,54). The postpartum rebound of thyroid autoimmune diseases is a striking example of this phenomenon (55).
Risk for Miscarriage in Women with Thyroid Autoimmunity
The association between thyroid autoimmunity (TAI) and increased fetal loss has been investigated in a large number of studies performed during the past decade in three continents, with over 5500 women investigated, both as study cases and controls (56). The main information from these studies is summarized in Table 80.1. The results concur that TAI, without overt thyroid dysfunction, is associated with a three- to fivefold increase in the rate of miscarriage (57,58,59,60,61,62,63). To find an association does not, however, imply a causal relationship, and the precise etiology of pregnancy loss in women with TAI remains largely unknown. Three hypotheses have been proposed. The first hypothesis holds that pregnancy loss is not directly related to the presence of circulating thyroid antibodies. In this view, TAI would only represent a marker of an underlying (yet to be defined) more generalized immune imbalance that, in turn, would explain a greater rejection rate of the fetal graft. The second hypothesis holds that despite apparent euthyroidism, the presence of TAI could be associated with a subtle deficiency in thyroid hormone concentrations or with a lesser ability of the thyroid economy to adapt to the necessary changes associated with the pregnant state, because of the reduced functional reserve characteristic of chronic thyroiditis. The third hypothesis is that TAI could act by delaying the occurrence of pregnancy, because of its frequent association with subfertility. Women with thyroid antibodies tend to become pregnant at an older age (average of 3–4 years later) and are therefore more prone to pregnancy loss. These hypotheses are not in contradiction with one another, and it remains plausible that the increased risk for pregnancy loss associated with TAI is multifactorial, eventually resulting from the combination of several independent deleterious factors (64,65).
TABLE 80.1. MISCARRIAGES IN WOMEN WITH POSITIVE THYROID ANTIBODIES
Miscarriage Rate (%)in
No. of Subjects
Positive Thyroid Antibodies (%)
Ab Neg.(or Control Women)
Characteristics of Selection of the Study Groups
Unselected population study
Unselected population study
Unselected population, before 14 wks' gestation
Recurrent spontaneous abortions
Pregnant with assisted reproductive techniques
Recurrent spontaneous abortions
Unselected population study
Recurrent pregnancy loss
Two or more consecutive abortions
Pregnant with assisted reproductive techniques
Failure to conceive after three cycles of IVF
Recurrent spontaneous abortions
Unselected population study
Ab neg., antibody negative; Ab pos., antibody positive; IVF, fertilization; NA, not applicable.in vitro
Modified from Poppe K, Glinoer D. Thyroid autoimmunity and hypothyroidism before and during pregnancy. 2003;9:149–161, with permission.Hum Reprod Update
If increased pregnancy loss is due to an underlying generalized immune dysregulation, and the presence of TAI merely represents an indirect marker of the immune condition, then there is no proven medical intervention that can presently be proposed. Short-term administration of steroids or injection of immunoglobulins has been used in a few patients with recurrent abortions to modulate immune responses, with variable success. If mild thyroid underfunction plays a significant role, this would constitute an argument for the systematic screening of women (before conception if they express the desire of being pregnant or as soon as a pregnancy is ongoing) for the presence of TAI or mild thyroid insufficiency, and give such patients the potential benefit of levothyroxine treatment (66). To date, only one such prospective trial has been reported, with promising results. In this study, women with TAI and recurrent early miscarriages were given levothyroxine treatment, both before and during pregnancy, and the researchers reported a significant reduction in the rate of miscarriages: 81% of the women who received thyroid hormone had live births, compared with only 55% in those given immunoglobulin injections (67). Obviously, conclusions must be considered with caution and balanced with the small number of patients investigated and the fact that there was no strict randomization. Despite these limitations, the study constitutes the first therapeutic intervention trial showing an effect of thyroid hormone administration in women who were habitual aborters. If delayed conception plays a role to decrease fertility in women with TAI, this would also constitute a strong argument for systematically screening infertile women for thyroid underfunction, particularly when seeking medical advice before fertilization procedures. Finally, women with TAI could be advised to plan for pregnancy at a younger age, although this type of medical advice is more easily written than practically applicable.
Effects of Pregnancy on Thyroid Function in Women with Thyroid Autoimmunity
When systematically screened in the early stages of pregnancy, 5% to 10% of women have thyroid autoantibodies with normal thyroid function. Despite the expected decrease in the titers of thyroid antibodies during gestation, the parameters of thyroid function show a gradual deterioration toward subclinical hypothyroidism in a significant fraction of women with TAI. Already in the first trimester, serum TSH, albeit normal, is shifted to slightly higher values, compared with antibody-negative pregnant controls. Thus, women with TAI usually are able to maintain normal thyroid function during early gestation, as a result of sustained thyrotropic stimulation. Thereafter, a marked reduction in free T4 values is often observed with the progression of gestation, and at delivery, almost one half of women with TAI have free T4 values in the hypothyroid range, hence confirming their reduced functional thyroid reserve (68). At the individual level, it is possible to predict the risk for progression to subclinical hypothyroidism, based on serum TSH levels and thyroid peroxidase (TPO) antibody titers: when serum TSH is greater than 2.0 mU/L and/or TPO antibody titers are greater than 1250 U/mL before 20 weeks' gestation, these markers reveal the propensity to develop hypothyroidism before the end of pregnancy (69). These observations provide clinicians with simple tools to help identify those women who carry the highest risk during early gestation. As a consequence, thyroid function can be more closely monitored and preventive levothyroxine treatment administered to avoid potentially deleterious effects of hypothyroxinemia on both maternal and fetal outcome.
The most common cause of primary hypothyroidism in women in reproductive age is chronic autoimmune thyroiditis. This occurs in both the goitrous and atrophic forms of the disease. One percent to 2% of pregnant women are already receiving levothyroxine therapy for hypothyroidism (70,71). Population-based studies have evaluated the prevalence of elevated serum TSH concentrations in the early part of gestation in women without apparent hypothyroidism: 2% to 2.5% of apparently healthy unselected pregnant women were found to have elevated TSH levels (68,72).
Effect of Hypothyroidism on Pregnancy Outcome
Until recently, hypothyroidism has been considered to be relatively rare during pregnancy, presumably because of the increased infertility and miscarriage rates associated with hypothyroidism (73). Nowadays, however, this view has changed. Several studies have shown that when hypothyroid women become pregnant and maintain the pregnancy, they carry an increased risk for both maternal obstetric and fetal complications. These complications are more frequent and more severe in pregnant women with overt hypothyroidism, when compared with subclinical hypothyroidism. The older data concerning obstetric repercussions refer to a period when the diagnosis of hypothyroidism was not always made before conception, when levothyroxine was given during late gestation only, when thyroid function was not adequately monitored, and finally when the adaptation of levothyroxine doses was not systematically implemented (71,74). Such studies also frequently mention the association of hypothyroidism with other conditions (hypertension or diabetes) that may also have increased the overall obstetric risks (75). There are relatively few reports on the outcome of pregnancy in untreated hypothyroid pregnant women, but available information indicates that levothyroxine therapy greatly improves, but does not entirely suppress, the frequency of obstetric complications (76,77,78). Recently it was shown that when hypothyroid patients were not rendered euthyroid, pregnancy either ended in spontaneous abortion (in >60% of the cases) or led to an increased prevalence of preterm deliveries. Conversely, in hypothyroid women with adequate treatment, the frequency of abortions was minimal and pregnancies were in general carried to term without complications (79).
In the newly diagnosed hypothyroid pregnant patient, a full replacement dose of levothyroxine should be instituted immediately, assuming there are no abnormalities in cardiac function. To normalize the extrathyroidal thyroxine pool more rapidly, therapy may be initiated by giving for 2 to 3 days a dose that is two to three times the estimated final replacement daily dose; this will allow for a more rapid return to a euthyroid state. Several studies have confirmed that levothyroxine requirements in most women with preexisting hypothyroidism must be increased by approximately 50% on average during pregnancy (25,26,69). Adjustment of levothyroxine doses should be implemented early in pregnancy, preferably in the first trimester. If the pregnancy is planned, the patient should have thyroid function tests soon after the missed menstrual period. If serum TSH is not increased at that time, tests should be repeated (at 8–12 and 20 weeks) because an increase in hormone requirements may not become apparent until later during gestation (80,81). The magnitude of the increment depends in part on the cause of hypothyroidism. Women without residual thyroid tissue (after radioiodine ablation for thyrotoxicosis or surgery for thyroid cancer) require a larger increment in levothyroxine, whereas women with with some residual functional thyroid tissue (Hashimoto's disease) require a smaller increment (26) (Table 80.2). The need to adapt levothyroxine doses is highly variable among hypothyroid patients; therefore, treatment monitoring should be tailored individually (69,82,83). Finally, women with subclinical hypothyroidism who are already taking small levothyroxine doses before pregnancy may not require a change systematically during gestation (although they frequently do) if their functional reserve remains adequate.
TABLE 80.2. DAILY THYROXINE DOSE REQUIRED FOR MAINTAINING NORMAL SERUM THYROTROPIN CONCENTRATION DURING PREGNANCY IN PATIENTS WITH PRIMARY HYPOTHYROIDISMa
Patients with Hashimoto's Disease (=15)n
Patients with Thyroid Ablation (=18)n
T4 dose (µg/day)
111 ± 25
139 ± 52
114 ± 33
166 ± 64a
T4 dose (µg/kg/day)
1.7 ± 0.6
1.9 ± 0.9
1.8 ± 0.5
2.3 ± 0.8a
Serum TSH (mU/L)
2.0 ± 1.8
1.8 ± 1.1
1.5 ± 1.3
1.9 ± 1.1
ap < 0.01.
Adapted from Kaplan MM. Monitoring thyroxine treatment during pregnancy. 1992;2:147–152, with permission.Thyroid
Fetal and Neonatal Consequences of Maternal Thyroid Underfunction
In general, infants born to hypothyroid mothers appear healthy and without evidence of thyroid dysfunction, provided that there was no severe iodine deficiency . Clearly, maternal hypothyroidism during pregnancy raises a serious concern about long-lasting psychoneurologic consequences for the progeny, due to the risk for insufficient placental transfer of maternal thyroid hormones to the developing fetus during the first half of gestation. Because the fetal thyroid gland becomes operational only after midgestation, thyroid hormones transferred transplacentally from mother to fetus are important, both before and after the onset of fetal thyroid function. Development of the fetal brain (with neuronal multiplication, migration, and architectural organization) during the second trimester corresponds to a phase during which the supply of thyroid hormones to the growing fetus is almost exclusively of maternal origin. During later phases of fetal brain development (with glial cell multiplication, migration, and myelinization), from the third trimester onward, the supply of thyroid hormones to the fetus is essentially of fetal origin. Therefore, whereas severe maternal hypothyroidism during the second trimester will result in irreversible neurologic deficits, maternal hypothyroxinemia occurring at later stages will result in less severe, and also partially reversible, fetal brain damage (18,19,21,22,23,84,85).in utero
In disease, three sets of clinical disorders ought to be considered and are schematically illustrated in Figure 80.5. For infants with a defect of thyroid gland ontogeny leading to congenital hypothyroidism, the participation of maternal hormones to the fetal circulating thyroxine environment is unaffected and the risk for brain damage results exclusively from insufficient fetal hormone production. In contrast, when the maternal thyroid is deficient (i.e., in women with TAI), it is both the severity and temporal occurrence of maternal hypothyroxinemia that drive the resulting consequences for fetal neuronal development. Finally, in iodine deficiency, both the maternal and fetal thyroid functions are affected, and it is the degree and precocity of iodine deficiency that drive the potential repercussions for fetal neurologic development (21). In 1999, the results of an important prospective investigation of neuropsychological development were reported in children 7 to 9 years of age, born to mothers with variable degrees of thyroid deficiency during pregnancy (86). For this aim, the investigators recruited children born to women who had an elevated serum TSH at 17 weeks' gestation. Detailed neuropsychological testing of the study children (and appropriate controls) showed that the study children performed less well, and had a mean intelligence quotient (IQ) score that was 4 points below that of controls. When the performances were analyzed in the children born to hypothyroid mothers who were left untreated, 9 of 15 tests were significantly lower and the average IQ's difference was 7 points, compared with the controls. In contrast, in the children born to hypothyroid mothers who were receiving levothyroxine, the mean IQs were similar to those of the controls. Finally, it is noteworthy that 77% of the mothers in this study had thyroid antibodies, confirming that chronic autoimmune thyroiditis was the most frequent underlying cause of maternal thyroid underfunction. Haddow's recent work underlines the notion that maternal thyroid underfunction (even when it is mild or “subclinical”) may be associated with an impairment of normal brain development in the offspring. When present in the first half of gestation, maternal hypothyroxinemia represents a risk for fetal brain development, because of an insufficient transfer of thyroid hormones to the fetoplacental unit. Furthermore, in most circumstances where a woman's thyroid function is defective, hypothyroxinemia is not restricted to the first trimester, and hypothyroidism tends to worsen as gestation progresses, especially when left undiagnosed and untreated (68,79). Therefore, the fetus may also be deprived of adequate amounts of thyroid hormones during later neurologic maturation and development.
FIGURE 80.5. A schematic representation of the three sets of clinical conditions that may affect thyroid function in the mother alone, the fetus alone, or the fetomaternal unit, showing the relative contributions of an impaired maternal and/or fetal thyroid function, that may eventually lead to alterations in fetal thyroxinemia. (Reproduced from Glinoer D, Delange F. The potential repercussions of maternal, fetal, and neonatal hypothyroxinemia on the progeny. 2000; 10:871–887, with permission.)Thyroid
Diagnosis and Screening of Hypothyroidism during Pregnancy
The diagnosis of primary hypothyroidism during pregnancy can be readily established by measuring serum TSH and free T4 (or a free T4 index). It should be recalled that total serum T4 should be increased 4 to 5 µg/dL (50–60 n) in the pregnant patient, whereas free TM4 remains broadly normal. The signs and symptoms of hypothyroidism are similar to those in the nonpregnant woman; however, they may be difficult to separate from nonspecific symptoms, such as the fatigue commonly associated with normal pregnancy.
Because thyroid disorders related to TAI are common in young female subjects, subclinical hypothyroidism may often remain undiagnosed. Thus, there is justification for proposing to systematically screen for hypothyroidism during pregnancy. Among different possible algorithms, a general scheme is shown in Figure 80.6 (87). Serum TSH and thyroid antibodies should be measured in early gestation. Based on recent findings that hypothyroxinemia might perhaps occur in some women without a concomitant serum TSH elevation, it appears reasonable to also include a free T4 determination (22,88). When serum TSH is elevated or free T4 is clearly below normal, and irrespective of the presence (or absence) of thyroid antibodies, the woman should be considered as highly suspect of having thyroid underfunction and treated with levothyroxine throughout pregnancy. Concerning the women with positive thyroid antibodies, medical management ought to be based on serum TSH levels before 20 weeks' gestation. When TSH is below 2 mU/L, most frequently associated with relatively low titers of thyroid antibodies, levothyroxine treatment is not warranted, and serum TSH should be monitored at the end of the second trimester. When serum TSH is still within the normal range but already between 2 and 4 mU/L in early gestation, most frequently associated with higher titers of thyroid antibodies, physicians should consider levothyroxine treatment. It is important to remember that serum TSH is down-regulated under the influence of hCG during the first half of gestation and that the thyroid deficit in women with TAI tends to deteriorate as gestation progresses. Because the potentially deleterious effects are not due to high serum TSH per se, clinical judgment should be based on serum free T4: if low or low-normal for the gestational age, treatment is probably justified. In practice, when such a scheme is systematically applied, most, if not all, of the pregnancies followed are successful and uneventful (89). Obviously, more studies are needed to assess the clinical relevance of such a scheme. Even though there is not as yet sufficient direct evidence for the advantage of treating women with subclinical hypothyroidism, many indirect arguments strongly suggest that no harm can be done and that levothyroxine treatment can only be beneficial for both mother and offspring (90,91,92). Finally, the systematic screening for TAI during early pregnancy should also allow for the delineation of a subgroup of women prone to developing thyroid dysfunction after parturition. Therefore, even when no specific treatment is warranted during gestation, the systematic screening should be of great help to clinicians for organizing a closer monitoring of thyroid dysfunction during the postpartum period (21,90,93) (see Chapter 27).
FIGURE 80.6. A proposed algorithm for the systematic screening of thyroid autoimmune disease and hypothyroidism during pregnancy, based on the determinations of thyroid perixidase antibodies, serum thyrotropin, and free thyroxine concentrations in the first half of gestation. (Modified from Glinoer D. The systematic screening and management of hypothyroidism and hyperthyroidism during pregnancy. 1998;9:403–411, with permission.)Trends Endocrinol Metab
Hyperthyroidism during pregnancy is relatively uncommon. It has been estimated indirectly that approximately 1 or 2 of 1000 pregnancies will be complicated by hyperthyroidism (91,94). The causes of hyperthyroidism include ones evident in the general population, as well as others that occur only during pregnancy. Clinical entities such as toxic adenoma, multinodular toxic goiter, subacute or silent thyroiditis, iodide-induced thyrotoxicosis, and thyrotoxicosis factitia are extremely uncommon during pregnancy. Molar disease should always be considered and can potentially lead to severe thyrotoxicosis; however, uncomplicated hydatidiform mole is now easily diagnosed in the early stages of gestation, and therefore rarely leads to severe hyperthyroidism.
The major cause of thyrotoxicosis in women of childbearing age is Graves' disease. In recent years, another cause has been characterized, resulting from the direct stimulation of the thyroid gland by hCG, which can induce a transient form of thyrotoxicosis during the first half of gestation. This syndrome is referred to as gestational transient thyrotoxicosis (GTT), and it occurs more frequently, but usually less severely, than Graves' disease. GTT differs from Graves' disease in that it is not of autoimmune origin and the course, fetal risks, and management and follow-up of both entities are different (2,10,11,95,96). This section focuses primarily on Graves' disease and GTT.
The historical clues and physical findings of hyperthyroidism in pregnant patients are the same as those occurring in nonpregnant patients. Their recognition is sometimes more difficult because of the similarity of the symptoms of normal pregnancy and those of thyrotoxicosis (97) (Table 80.3). Fatigue, palpitations, anxiety, heat intolerance, and diaphoresis are all symptoms that can be found in the pregnant female without thyrotoxicosis. A useful symptom is that, instead of the customary weight gain, patients report weight loss or, perhaps more frequently, the absence of weight gain despite an increased appetite unless there is associated excessive vomiting. Nausea or emesis (“morning sickness”) frequently occurs during a normal pregnancy; however, the occurrence of hyperemesis gravidarum leading to weight loss always must raise the possibility of thyrotoxicosis.
TABLE 80.3. CLINICAL FEATURES SUGGESTING THE POSSIBILITY OF HYPERTHYROIDISM DUE TO GRAVES' DISEASE IN THE PREGNANT STATE
1. Prior history of thyrotoxicosis or autoimmune thyroid disease in the patient or her family
2. Presence of typical symptoms of thyrotoxicosis including weight loss (or failure to gain weight), palpitations, proximal muscle weakness, or emotional lability
3. Symptoms suggesting Graves' disease such as ophthalmopathy or pretibial myxedema
4. Thyroid enlargement
5. Accentuation of normal symptoms of pregnancy such as heat intolerance, diaphoresis, and fatigue
1. Pulse rate > 100 beats/min
2. Widened pulse pressure
3. Eye signs of Graves' disease or pretibial myxedema
4. Thyroid enlargement especially in iodine sufficient geographical areas
Adapted from Glinoer D. Thyroid regulation and dysfunction during pregnancy. In: DeGroot L, ed. (update 2003). Accessed at Thyroid disease managerwww.thyroidmanager.org , with permission
The diagnosis of thyrotoxicosis can readily be confirmed by appropriate and simple laboratory tests. Virtually all patients with significant clinical symptoms have serum TSH values less than 0.1 mU/L as well as concurrent elevations in serum free T4 and T3 levels. In the first trimester, serum TSH may be transiently suppressed (< 0.2 mU/L) at the time of peak hCG levels in up to 20% of euthyroid women. Therefore, the degree and duration of TSH suppression during the first trimester must be considered in making the diagnosis. Because most patients with Graves' disease test positive for thyroid autoantibodies, antibody presence should alert the clinician to the possibility that autoimmune thyroid disease is the cause of symptoms evoking thyrotoxicosis.
Graves' Disease in Pregnancy
Three clinical situations are important to consider: (a) women with active Graves' disease diagnosed before pregnancy and who are receiving antithyroid drug (ATD) treatment, (b) women who are in remission or considered cured after prior treatment, and finally (c) women in whom the diagnosis of Graves' disease has not been established before the onset of pregnancy but who have thyrotropin receptor (TSHR) antibodies. An important concept is that both maternal and fetal outcome is directly related to adequate control of hyperthyroidism. Several reports have identified severe obstetric consequences, such as preeclampsia, fetal malformations, premature delivery, and low infant birth weight when thyrotoxicosis remains uncontrolled, usually following poor compliance with therapy (98,99,100). For patients in whom the diagnosis is correctly made early in pregnancy and treatment is promptly started, the prognosis for both the mother and the offspring remains excellent. The overall goal of therapy is to keep the patient at high euthyroid or borderline hyperthyroid level of thyroid function throughout pregnancy, using the lowest possible dose of antithyroid drug (ATD) (101,102).
Another important consideration is that TSHR antibody titers may remain elevated, even after prior thyroidectomy or thyroid ablation using radioiodine or the apparent cure of hyperthyroidism with ATD several years before pregnancy. The risk for fetal and neonatal hyperthyroidism is negligible in euthyroid women not currently receiving ATD treatment, but who had received antithyroid drugs previously for Graves' disease; therefore, systematic measurements of TSHR antibody are not mandatory. For a euthyroid woman (with or without thyroid hormone substitution therapy) who has previously received radioiodine therapy or undergone thyroid surgery for Graves' disease, the risk for fetal and neonatal hyperthyroidism depends on the level of TSHR antibody in the mother. As a result, antibodies should be measured early in pregnancy to evaluate this risk. For a pregnant woman who takes ATDs for active Graves' disease (assuming that therapy was started before or early during pregnancy), TSHR antibodies should be checked again in the last trimester. If the antibody titers have not substantially decreased during the second trimester, the possibility of fetal hyperthyroidism should be considered (20,100,103,104,105).
Hyperthyroidism due to Graves' disease usually tends to progressively improve during the course of gestation, although exacerbations can be observed in the first months (106). Several reasons may explain this spontaneous improvement: (a) partial immunosuppression (characteristic of pregnancy) with a significant decrease in TSHR antibody titers; (b) marked increase in serum TBG levels, which tends to reduce the free T4 and T3 fractions; (c) obligatory iodine losses specific to pregnancy, which may constitute, paradoxically, an advantage for pregnant patients with Graves' disease; (d) suggestion that the balance between TSHR antibody–blocking and–stimulating activities may be modified in favor of blocking autoantibodies; and finally (e) changes in cytokine production, with an impaired cross-regulation of interleukin 12 (IL-12) by IL-10 between normal and Graves' disease pregnant women (107,108,109).
Pregnant patients with Graves' disease should be treated exclusively with ATD, unless the severity of the condition justifies (exceptionally) a more radical approach by surgery, which is then preferably performed during the second trimester. The dosage of ATD should be maintained at a minimum and the drugs discontinued as soon as possible (often after 4–6 months' gestation). Combined administration of ATD and levothyroxine to the mother to maintain euthyroidism in the fetus should be avoided because transplacental passage of ATDs is relatively high, but negligible for thyroid hormones. All ATDs cross the placenta and may therefore affect fetal thyroid function (16,99,110). The thiourea drugs, propylthiouracil (PTU), methimazole (MMI), and carbimazole (CMI, which is rapidly metabolized to MMI), have been compared with respect to their use during pregnancy. PTU is more water soluble and is therefore less well transferred from maternal to fetal circulation as well as from the maternal circulation into breast milk (110,111), although perfusion of the placenta did not demonstrate any difference in the transport of PTU or MMI. This has led to the recommendation that PTU should be used in preference to MMI or CMI during pregnancy, unless a specific therapy is directed to suppress also thyroid function in the fetus. However, MMI and CMI are commonly used during pregnancy in many countries where PTU is not commercially available, and this without particular problems, except for the rare occurrence of gut embryopathy during MMI or CMI use (112,113,114). We therefore believe the recommendation for the sole use of PTU is not fully justified. Concerning the risk for fetal hypothyroidism induced by maternal treatment with ATD, most studies have concluded that there is little reason to choose PTU over MMI (99,115,116,117). The main guidelines for the treatment of pregnant patients with Graves' disease are summarized in Table 80.4.in vitro
TABLE 80.4. GUIDELINES FOR THE MANAGEMENT OF GRAVES' DISEASE DURING PREGNANCY
1. Monitor pulse, weight gain, thyroid size, free T4 and T3, and TSH at monthly intervals.
2. Use the lowest doses of ATD that will maintain the patient in a euthyroid or mildly hyperthyroid state, but not higher than ~300 mg PTU (or ~20 mg MMI)
3. Communicate regularly with the obstetrician, especially with respect to fetal pulse and growth.
4. One should not attempt full normalization of serum TSH. Serum TSH concentrations between 0.1 and 0.4 mU/L are generally appropriate, but lower levels are acceptable if the patient is doing well clinically.
5. PTU may be preferred to MMI, but both ATDs can be used.
6. Although even as little as 100–200 mg of PTU/day may affect fetal thyroid function, doses as high as 300 mg PTU (~20 mg MMI) have been used. Iodides should not be used during pregnancy except to prepare patients for surgery.
7. Indications for surgery are:
a. requirements for high doses of PTU (>300 mg) or MMI (>20 mg) with inadequate control of clinical hyperthyroidism.
b. poor compliance with resulting clinical hyperthyroidism.
c. the appearance of fetal hypothyroidism (retarded bone age, bradycardia) at ATD doses required for control of disease in the mother.
8. Usually the dose of ATD can be adjusted downward after the first trimester and often discontinued during the third trimester.
9. ATDs often need to be reinstituted or increased after delivery.
ATD, antithyroid drug; MMI, methylmercaptoimidazole; PTU, propylthiouracil; T3, triiodothyronine; T4, thyroxine; TSH, thyrotropin.
Adapted from Glinoer D. Thyroid regulation and dysfunction during pregnancy. In: DeGroot L, ed. (update 2003). Accessed at Thyroid disease managerwww.thyroidmanager.org , with permission
Fetal and Neonatal Aspects
Graves' disease during pregnancy can affect the fetus in several ways. First, when maternal TSHR antibodies are elevated during early gestation and the antibody titers have not substantially decreased during the second trimester, fetal (and neonatal) hyperthyroidism constitutes a real risk. This risk can be assessed by ultrasonographic data indicating the presence of a fetal goiter, tachycardia (>160 beats/ min), growth retardation, increased fetal motility, or accelerated bone maturation. In selected cases, fetal cord blood sampling may be required for diagnosis or for monitoring therapy (118,119). If persistent fetal tachycardia is present, it is reasonable to initiate ATD therapy with 200 to 400 mg of PTU per day (or 20–30 mg MMI), with levothyroxine supplementation to maintain maternal euthyroidism when needed. Second, both the hyperthyroidism itself and the administration of ATD to the expecting mother may raise concern related to potential teratogenicity of the disease or the drug. To date, it remains uncertain whether untreated Graves' disease is associated with a higher frequency of congenital abnormalities. Drug-related congenital abnormalities include aplasia cutis congenita (absence of skin and accessory structures, usually over the scalp) and severe embryopathy in rare instances (113,114). The evidence linking aplasia cutis to maternal MMI is not conclusive, but it is also not sufficient to rule out a causal role (99,117). So far aplasia cutis has not been reported in mothers receiving PTU. In view of the potential dangers (for both mother and offspring) of not treating active Graves' disease during pregnancy, the concerns of rare congenital anomalies would not, in our opinion, justify withholding ATD administration. Third, the administration of ATD to pregnant women with Graves' disease may induce fetal hypothyroidism. As alluded to in an earlier section of this chapter, this should certainly be avoided in view of the potentially deleterious consequences on the neuro-psychointellectual development, and it can be avoided in practice by keeping maternal circulating thyroid hormone levels in the upper part of the normal range (115,120,121). Fourth, both the disease and its treatment can induce a goiter in the fetus. In these circumstances, fetal goiter may result from growth-stimulating effects of maternal TSHR antibodies or from direct ATD effects on the fetal thyroid. Fifth, unrecognized fetal hyperthyroidism may be followed by neonatal hyperthyroidism at birth. Neonatal hyperthyroidism is usually considered to be uncommon, occurring in approximately 1% of pregnancies in patients with Graves' disease in North America. However, in an update of this question, it was suggested that the actual frequency of neonatal hyperthyroidism may be significantly higher, perhaps on the order of 2% to 10% (105). The risk is highest in the offspring of women with Graves' disease in whom the hyperthyroidism is not well controlled, as well as in those women with the highest TSHR antibody titers. The relationship between the risk for neonatal hyperthyroidism and the antibody levels is logical, in that higher titers of antibodies are more likely to be transmitted from mother to fetus. Usually, neonatal Graves' disease is diagnosed at or shortly following delivery, after the maternal ATDs have been cleared from the neonatal thyroid gland and serum (122,123,124). Signs of hyperthyroidism in the neonate include congestive heart failure, goiter, proptosis, jaundice, hyperirritability, failure to thrive, and tachycardia. Once considered, the diagnosis is easily confirmed. Cord serum free T4 and TSH determinations should be performed in all deliveries of mothers with a history or the presence of Graves' disease. Treatment should be performed in conjunction with the neonatologist and may include iodide or iopanoic acid, ATD, glucocorticoids, digoxin, and β-adrenergic blocking agents, depending on the cardiovascular status.
The question of the safety of lactation during ATD therapy arises frequently. In the past, women receiving ATD have been advised against breast-feeding because of the fear that ATD, concentrated in milk, might affect the infant's thyroid function. Both PTU and MMI are secreted in human milk, although the secretion of PTU may be less pronounced because of its increased binding to albumin. In one study evaluating the effects of 15 mg of CMI or 150 mg of PTU on infants of nursing mothers, there was no evidence of neonatal hypothyroidism during the first 21 days of life. In another study, serum MMI levels were measured in breast-fed infants of thyrotoxic mothers who received 20 to 30 mg MMI/day: 2 hours after MMI ingestion, serum MMI levels in the babies were extremely low, far below the therapeutic range (125). Thus, with both PTU and MMI, only limited quantities of drugs are concentrated into the milk and, as long as MMI or PTU doses can be kept moderate (< 20 mg/day MMI or < 150 mg/day PTU), the risk for the infant is practically negligible and there is no “evidence-based” reason to advise mothers against nursing when they take ATD (98,112,126). Periodic monitoring of the infant's thyroid function is necessary during ATD administration to the mother. There is also a possibility that allergic reactions associated with ATD (agranulocytosis or rash) may occur. Although these reactions are rare, they should be kept in mind when evaluating a febrile infant or presence of a rash. Thus, within the limitations outlined above, the use of ATD in lactating mothers does not pose a risk to the neonate and appears to be safe.
Thyrotoxicosis during Postpartum
Because Graves' disease is an autoimmune disorder, and owing to the profound autoimmune modifications occurring after the delivery, it is understandable that the postpartum period has been associated with a greater frequency of onset, recurrence, or exacerbation of thyrotoxicosis resulting from Graves' disease (106,127,128,129). This occurrence seems to be more prevalent among women with Graves' disease in Japan, where up to 40% of new cases diagnosed in women of childbearing age develop Graves's disease during the postpartum period (130). It should be mentioned, however, that such a high rate of association between the onset of Graves' disease and the postpartum period is not universally observed. The reason for such a discrepancy is not clearly apparent, but may be related to genetic differences or iodine intake differences between the populations.
Differentiation of postpartum Graves' disease from the early thyrotoxic phase related to postpartum thyroiditis (PPT) may be difficult (131). Two features have been advocated to distinguish between these entities: the radioiodine uptake (high in thyrotoxicosis resulting from Graves' disease and low in the “destructive” thyroiditis related to PPT) and TSHR antibodies (positive in Graves' disease and negative in PPT) (129,132). However, the clinical diagnosis based on these features often remains complex. Differential diagnosis may even be further complicated by the possibility that both conditions may coexist in the same woman (133). Making the distinction is important because the management of both conditions is strikingly different: although ATDs are not indicated in women with thyrotoxicosis due to PPT, ATD therapy (or other more radical approaches) is the treatment of choice for thyrotoxicosis due to Graves' disease (See Chapters 27 and 45).
GESTATIONAL TRANSIENT THYROTOXICOSIS
Gestational transient thyrotoxicosis (GTT) is defined as nonautoimmune hyperthyroidism of variable severity that occurs in women with a normal pregnancy, typically in association with hyperemesis (2,112,134). GTT differs from Graves' disease in that it occurs in women who have no history of thyrotoxicosis and in the absence of detectable TSHR antibodies. Its etiology is directly related to the thyrotropic stimulation of the thyroid gland associated with hCG (96,135,136,137). Recent prospective studies have indicated that the prevalence of GTT may represent 2% to 3% of all pregnancies, which is 10-fold more frequent than hyperthyroidism resulting from Graves' disease (2,10,11). The prevalence of GTT appears to be surprisingly highly variable in different parts of the world, as low as 0.3% in Japan and as high as 11% in Hong Kong; the reason for such a discrepancy in frequency is presently not understood (138,139).
Owing to its transient nature, clinical manifestations of GTT are not always apparent or routinely detected in GTT (69,87,112). Symptoms compatible with hyperthyroidism, including weight loss or the absence of weight increase, tachycardia, and unexplained fatigue are found in half of the women with GTT. Hyperemesis is frequently associated with the most severely thyrotoxic cases, and the symptoms may become sufficiently alarming to require hospitalization for treatment. In these relatively rare instances, the severity of clinical presentation may require treatment with PTU or MMI, usually administered for only a few weeks, because free T4 normalizes in parallel with the decrease in hCG concentrations. In most cases of GTT, no specific treatment is required and symptoms can be relieved by the administration of β-adrenergic blocking agents for a short period. GTT is not associated with a less favorable outcome of the pregnancy. Finally, it should also be noted that, by coincidence, GTT can occur in women with preexisting thyroid disorders such as glandular autonomy, autoimmune thyroiditis, or cryptic Graves' disease.
The precise pathogenic mechanisms underlying GTT are still not fully understood. It remains possible that abnormal molecular variants of hCG are produced, with a prolonged half-life explaining sustained high circulating hCG levels or variants of hCG with a more potent thyrotropic activity (96,134,135,136). It has also been hypothesized that a dysregulation of β-hCG production may transiently take place. A study of a group of women with twin pregnancies reinforced the concept that normal women may develop thyrotoxicosis transiently in association with abnormally elevated hCG levels, particularly when the hCG elevation is maintained for a prolonged period (140). Comparing twin with single pregnancy, we have shown that peak hCG values are significantly higher (almost double) and also of much longer duration in twin pregnancy. Whereas hCG values above 75,000 U/L lasted less than a week in a single pregnancy, hCG levels above 100,000 U/L (and often reaching or exceeding 200,000 U/L) lasted up to 6 weeks in some of the twin pregnancies. The latter were also associated with a more profound and frequent blunting of serum TSH, as well as free T4 values that transiently rose above normal. Based on GTT in twin pregnancy, a quantitative direct effect of elevated hCG may presumably be sufficient to explain GTT in most pregnant women, provided that hCG values remain above 75,000 to 100,000 U/L for a sufficiently long period (69,87,97,140,141). Finally, increased placenta weight is associated with elevated HCO levels and may cause hyperthyroidism (141a). The present concept therefore is that GTT is directly related to both the amplitude and duration of peak hCG values.
The effects of hCG to stimulate the thyroid gland can best be explained by the marked homology that exists between the hCG and TSH molecules, as well as between the luteinizing hormone/chorionic gonadotropin and TSH receptors (141,142). GTT can be considered an example of an endocrine “spillover” syndrome, a concept based on molecular mimicry between hormone ligands and their receptors (96). An interesting, but unresolved, question is whether the thyroid gland is a “passive bystander” (or a victim) of an abnormal thyrotropic hCG activity in GTT, or whether the gland itself, via variable degrees of sensitivity of the TSH receptor (TSHR), plays a role in its responsiveness to the effects of hCG. So far, only one clinical example has been reported with a substantially increased sensitivity of the TSHR to stimulation by hCG, due to a single mutation (K183R) in the extracellular domain of the TSHR (143). The mutant TSHR was more sensitive than the wild-type receptor to hCG, thus accounting for recurrent hyperthyroidism in pregnancy in the presence of normal hCG levels. This finding raises the possibility that some women who develop GTT may have an abnormality at the level of the thyroid follicular cell. In elaborate studies of the structure-phenotype relationship of the TSHR, the group of Gilbert Vassart performed site-directed mutagenesis, substituting lysine 183 in the ectodomain of the TSHR by a variety of amino acids expressing different physicochemical properties (144,145). They obtained unexpected results since all TSHR mutants displayed a widening of their specificity toward hCG stimulation. Modeling of the mutated receptors indicated that the increased gain in sensitivity might have resulted from the release of a nearby glutamate residue (in position E157) from a salt bridge formed with K183.
Gestational transient thyrotoxicosis is often associated with morning sickness, increased vomiting, and hyperemesis gravidarum, a severe condition requiring hospitalization and definitive treatment (134,146,147,148). There is a good correlation between the intensity of emesis and the abnormalities of thyroid function, and it is also known that women with a twin pregnancy experience severe vomiting more often in early gestation. Because there is no indication of increased vomiting among pregnancies with Graves' disease, hyperemesis appears to be associated with hCG-induced thyrotoxicosis, although not all cases of vomiting during early pregnancy are related to disturbances of thyroid function (149). The most likely explanation is that sustained elevations of circulating hCG levels promote estradiol production in these women. By a mechanism that is not yet fully understood, the combination of high hCG and estradiol levels, as well as increased free T4 concentrations transiently promotes emesis near the period of peak hCG (95,98,135,150).
SYSTEMATIC SCREENING FOR HYPERTHYROIDISM DURING PREGNANCY
Even though thyrotoxicosis is relatively uncommon during pregnancy, it probably occurs more frequently than usually believed. Taken together, the prevalence of Graves' disease and GTT may represent 2% to 4% of all pregnancies. A three-branch screening algorithm for the diagnosis of thyrotoxicosis is proposed in Figure 80.7. The overall strategy is based on the availability of serum TSH values and TPO antibody titers in early gestation. If TSH is suppressed (or blunted) and TPO antibodies are positive, serum free T4 and TSHR antibodies are determined. This first branch in the algorithm allows for the diagnosis of cases with unsuspected underlying thyrotoxicosis of autoimmune origin (i.e., Graves' disease), broadly corresponding to three to five cases per year given a hospital setting delivering 1000 to 1500 babies per year. If TSH is suppressed or blunted but TPO antibodies are undetectable, free T4 and hCG concentrations should be determined. The second branch in the algorithm allows for the diagnosis of nonautoimmune, hCG-induced thyrotoxicosis (GTT), corresponding to approximately 20 to 30 cases per year given a similar hospital setting. In women with a history of Graves' disease (present or past, active or considered cured), TSHR antibody titers and free T4 concentrations should be determined when women have a history of Graves' disease before conception. In women with both active disease or “metabolically” cured Graves' disease, but in whom high titers of Graves' IgGs are found, TSHR antibody titers should be monitored at a later stage during gestation, preferably near the end of the second trimester. This strategy allows for the diagnosis of maternal immunity related to Graves' disease and for the potential suspicion of fetal thyrotoxicosis, thus indicating the need for a careful monitoring of fetal development. Finally, all cases with thyrotoxicosis of autoimmune origin diagnosed during pregnancy require a close monitoring of thyroid antibody titers (both TPO antibodies and TSHR antibodies) and thyroid function during the first year postpartum because of the significant risk for developing postpartum thyroiditis or postpartum exacerbation of thyrotoxicosis.
FIGURE 80.7. A proposed three-step algorithm for the systematic screening of thyroid hyperfunction during pregnancy. The first step allows for the diagnosis of unsuspected thyrotoxicosis of autoimmune origin; the second step for the diagnosis of gestational transient thyrotoxicosis (); the third step concerns patients with Graves' disease. HR, hyperthyroidism. (Reproduced from Glinoer D. Thyroid hyperfunction during pregnancy. 1998;8:859–864, with permission.)GTTThyroid
NODULAR THYROID DISEASE
Risk for Thyroid Cancer
A thyroid nodule or a dominant nodule within a multinodular gland may be recognized for the first time during pregnancy (151,152). Evaluation of a nodule during pregnancy relies on the use of ultrasound and fine-needle aspiration biopsy (FNAB). FNAB is indicated in all pregnant women with a nodule larger than 1 cm, or enlarging during gestation, or when it is associated with palpable cervical lymph nodes. If FNAB is highly suggestive of malignancy, thyroid surgery should be performed even though the patient is pregnant, although surgery can be safely deferred until after delivery (153). If FNAB results show a possible follicular neoplasm or is mildly suspicious, surgery is usually deferred until after the pregnancy and breast-feeding periods (154,155,156). There is no clear evidence that the natural history of any form of thyroid cancer is significantly modified by pregnancy (157,158,159). The outcome of thyroid cancer diagnosed during pregnancy was reevaluated recently in a study with a median follow-up period of 14 years (160). The researchers showed that differentiated thyroid cancer presenting in pregnancy generally has an excellent prognosis. When the disease is discovered early in pregnancy, surgery should be considered in the second trimester. Radioiodine scintigraphies and treatment can safely be deferred until after delivery. Treatment, however, should not be delayed more than a year.
Effect of Pregnancy on Nodule Formation
In a euthyroid patient with a diffuse or nodular goiter, there is good evidence that new nodules frequently tend to form and, when already present before, to increase in both size and number during pregnancy (44,58). Our personal data on nodule formation during pregnancy were obtained in studies performed in an area with a mild to moderate iodine deficiency. Therefore, the most probable explanation was that nodule formation and growth was part of the goitrogenic stimulus associated with pregnancy. However, other studies have been conducted in areas without iodine deficiency and also showed that pregnancy had an effect on nodule formation. For instance, the prevalence of thyroid nodularity was 9% in nulliparous compared with 25% in multiparous women in Germany (161). Similarly, in Hong Kong, investigators recently reported that 15% of pregnant women have a nodular thyroid at the onset of pregnancy (with mainly micronodules identified by ultrasonography) (162). During gestation, the nodules tended to increase in size and new nodules were formed in 11% of the women with preexisting nodular disease. Altogether, the prevalence of nodularity increased from 15% at onset of pregnancy to 24% in the early postpartum period. FNAB was performed in a limited number of cases and disclosed no single case with a malignancy. This study confirmed the significant association between gravidity and frequency of thyroid nodules. Thus, pregnancy is associated with an increase in the size of preexisting nodules as well as the development of new nodules, and must therefore be considered as one of the factors predisposing women to nodular goiter.
Several physiologic complex changes take place during a pregnancy, which together tend to modify the global economy of the thyroid gland and have variable impacts at different time points during gestation. Among the main physiologic changes for the thyroidal economy during pregnancy, there is a marked increase in serum TBG and in the extrathyroidal T4 distribution space, taking place in the first half of gestation. In order to maintain the homeostasis of free T4, the thyroid machinery must produce more thyroxine until a new steady state is reached around midgestation. Thereafter, changes in peripheral thyroid hormone metabolism explain the sustained need of increased T4 production to maintain unaltered serum free T4 concentrations.
Iodine deficiency occurs when pregnancy takes place in areas with only mild iodine restriction. Because this occurs at a time when hormone requirements are increased, ID induces a vicious circle that leads to enhanced glandular stimulation, relative hypothyroxinemia, and gestational goitrogenesis, affecting both mother and fetus. The ideal iodine supply during pregnancy is 200 to 250 µg/day. This can be provided in iodine supplements given in the form of multivitamin pills especially designed for pregnancy purposes. In areas with severe iodine deficiency, prophylactic and therapeutic iodine supplementation becomes an emergency to avoid endemic cretinism.
Pregnancy dampens the immune system, leading to a pattern where both arms of the immune responses (i.e., cell-mediated and humoral) are reduced. The rapid reduction in immune suppressor functions following the delivery leads to the reestablishment and exacerbation of these conditions.
Concerning the thyroid and infertility, there is good evidence to suggest that TAI is involved in sub(in)fertility and that TAI constitutes a useful marker of the underlying abnormality, independently of overt thyroid dysfunction. With regard to pregnancy loss, the majority of studies have clearly established that TAI is associated with a significant increase in the risk for miscarriage. With regard to repercussions of positive thyroid antibodies during gestation, the main risk is the occurrence of maternal hypothyroidism, with its potentially deleterious effects for both mother and fetus. This can be prevented by the systematic screening for thyroid dysfunction and antibody presence in early gestation, followed by the administration of levothyroxine treatment.
Undisclosed hypothyroidism with various degrees of severity (from subclinical to overt disease) is present in 2% to 4% of unselected women entering pregnancy. When hypothyroid women do become pregnant and maintain the pregnancy, they carry an increased risk for obstetric and fetal complications. Most available information indicates that an adequate treatment with thyroid hormone greatly improves the frequency of these abnormalities. In most, if not all, women with preexisting hypothyroidism, levothyroxine requirements increase significantly during pregnancy, on the average of approximately 50% above the preconception dosage requirements. Because of the high frequency of autoimmune thyroiditis in young females, because subclinical hypothyroidism often remains undiagnosed, because potential obstetrical consequences are associated with untreated hypothyroidism, and finally because of the potential consequences of maternal hypothyroxinemia on fetal development, there is a justification to propose a systematic screening for TAI and hypothyroidism in pregnancy.
Recent studies have underlined the notion that maternal thyroid underfunction (hypothyroxinemia), even when it is mild or considered subclinical, and especially when it occurs in early gestation, may be associated with an impairment of the normal neuropsychointellectual outcome in the progeny.
Concerning Graves' disease during pregnancy, three clinical situations are important to consider: women with active disease under ATD treatment, women in remission or considered cured after a prior treatment, and women with undiagnosed Graves' disease before pregnancy who carry TSHR antibodies. The maternal and fetal outcome of pregnancy is directly related to an adequate control of thyrotoxicosis: the obstetric consequences of poorly controlled thyrotoxicosis include preeclampsia, fetal malformations, premature delivery, and low birth weight. The main principle of ATD treatment during pregnancy is to administer the lowest dose needed for regulating clinical symptoms, accepting mild degrees of thyrotoxicosis as long as pregnancy progresses satisfactorily. Both PTU and MMI can be used. With regard to lactation in mothers who take ATDs, there is no evidence-based reason to advise mothers against nursing, as long as the ATD doses can be kept moderate.
Graves' disease can affect the fetus in several ways. When maternal TSHR antibodies have not substantially decreased during the second trimester, fetal (and neonatal) thyrotoxicosis constitutes a real risk. Thyrotoxicosis itself and the administration of ATD may raise concern related to potential teratogenicity of the disease or the drug (aplasia cutis; severe embryopathy), but these rare congenital anomalies do not justify withholding ATD administration. ATDs may induce fetal hypothyroidism; this can be avoided by keeping maternal circulating thyroid hormone levels in the upper part of the normal range. Fetal goiter may result from the growth-stimulating effects of maternal TSHR antibodies and from the direct effect of ATD on the fetal gland. The actual frequency of neonatal thyrotoxicosis may be higher than previously considered, and is highest in the offspring of women whose disease is not well controlled and those with the highest TSHR antibody titers. Neonatal Graves' disease is usually diagnosed at or shortly following delivery, after the maternal ATDs have been cleared from the neonatal thyroid gland and serum.
Human chorionic gonadotropin possesses intrinsic thyroid-stimulating activity, leading transiently to a partial TSH suppression near the end of first trimester in approximately 20% of all pregnancies. In 10% of the latter, serum free T4 levels may become transiently elevated and, in turn, these women may develop GTT. GTT is defined as a transient increase in thyroid hormone production of nonautoimmune origin, leading to variable degrees of thyrotoxicosis, frequently associated with hyperemesis. The prevalence of GTT is 2% to 3% of pregnancies. A quantitative direct effect of elevated hCG to stimulate the thyroid gland explains most GTT cases; GTT is directly related to both the amplitude and duration of peak hCG. GTT is an example of endocrine spillover syndromes, a concept based on molecular mimicry between the hormone ligands (TSH and hCG) and their receptors.
Hyperemesis gravidarum is often present during the first months of gestation. A significant fraction of women with hyperemesis gravidarum may have biochemical features suggesting hyperthyroidism, which results from excessive hCG-induced thyroidal stimulation. Concerning emesis, the most likely explanation is that elevated hCG levels promote estradiol production and that the combination of high hCG, estradiol, and free T4 concentrations promotes emesis near the period of peak hCG.
With regard to the systematic screening for hyperthyroidism, relatively simple algorithms are available, especially when data from the systematic screening for hypothyroidism are already available.
Thyroid nodules in euthyroid pregnant patients should be aspirated for cytologic diagnosis. If a malignancy is highly suspected, surgery should be performed with the timing depending on the nature of the lesion and the time of parturition. Pregnancy itself does not adversely affect the natural history of thyroid carcinoma.
1. Burrow GN. Thyroid function and hyperfunction during gestation. 1993;14:194–202.Endocr Rev
2. Glinoer D. The regulation of thyroid function in pregnancy: pathways of endocrine adaptation from physiology to pathology. 1997;18:404–433.Endocr Rev
3. Liberman CS, Pino SC, Fang SL, et al. Circulating iodide concentrations during and after pregnancy. 1998;83:3545–3549.J Clin Endocrinol Metab
4. Brander L, Als C, Buess H, et al. Urinary iodine concentration during pregnancy in an area of unstable dietary iodine intake in Switzerland. 2003;26:389–396.J Endocrinol Invest
5. Hershman JM. Role of human chorionic gonadotropin as a thyroid stimulator. 1992;74:258–259.J Clin Endocrinol Metab
6. Kimura M, Amino N, Tamaki H, et al. Physiologic thyroid activation in normal early pregnancy is induced by circulating hCG. 1990;75:775–778.Obstet Gynecol
7. Ballabio M, Poshyachinda M, Ekins RP. Pregnancy-induced changes in thyroid function: role of human chorionic gonadotropin as putative regulator of maternal thyroid. 1991;73:824–831.J Clin Endocrinol Metab
8. Arturi F, Presta I, Scarpelli D, et al. Stimulation of iodide uptake by human chorionic gonadotropin in FRTL-5 cells: effects on sodium/iodide symporter gene and protein expression. 2002;147:655–661.Eur J Endocrinol
9. Guillaume J, Schussler GC, Goldman J. Components of the total serum thyroid hormone concentrations during pregnancy: high free thyroxine and blunted thyrotropin (TSH) response to TSH-releasing hormone in the first trimester. 1985;60:678–684.J Clin Endocrinol Metab
10. Glinoer D, De Nayer P, Bourdoux P, et al. Regulation of maternal thyroid function during pregnancy. 1990;71:276–287.J Clin Endocrinol Metab
11. Glinoer D, De Nayer P, Robyn C, et al. Serum levels of intact human chorionic gonadotropin (hCG) and its free α and β subunits, in relation to maternal thyroid stimulation during normal pregnancy. 1993;16:881–888.J Endocrinol Invest
12. Kraiem Z, Lahat N, Sadeh O, et al. Desialylated and deglycosylated human chorionic gonadotropin are superagonists of native human chorionic gonadotropin in human thyroid follicles. 1997;7:783–788.Thyroid
13. Yoshimura M, Pekary AE, Pang X-P, et al. Thyrotropic activity of basic isoelectric forms of human chorionic gonadotropin extracted from hydatidiform mole tissues. 1994;78:862–866.J Clin Endocrinol Metab
14. Glinoer D. The thyroid in pregnancy: a European perspective. 1995;18:1–11.Thyroid Today
15. Roti E, Fang SL, Green K, et al. Human placenta is an active site of thyroxine and 3,3′,5-triiodothyronine tyrosyl ring deiodination. 1981;53:498–501.J Clin Endocrinol Metab
16. Roti E, Gnudi A, Braverman LE. The placental transport, synthesis, and metabolism of hormones and drugs which affect thyroid function. 1983;4:131–149.Endocr Rev
17. Hidal JT, Kaplan MM. Characteristics of thyroxine 5′-deiodination in cultured human placental cells. 1985; 76:947–955.J Clin Invest
18. Vulsma T, Gons MH, De Vijlder JM. Maternal fetal transfer of thyroxine in congenital hypothyroidism due to a total organification defect of thyroid dysgenesis. 1989;321:13–16.N Engl J Med
19. Contempre B, Jauniaux E, Calvo R, et al. Detection of thyroid hormones in human embryonic structures during the first trimester of pregnancy. 1993;77:1719–1722.J Clin Endocrinol Metab
20. Burrow GN, Fisher DA, Larsen PR. Mechanisms of disease: maternal and fetal thyroid function. 1994;331: 1072–1078.N Engl J Med
21. Glinoer D, Delange F. The potential repercussions of maternal, fetal, and neonatal hypothyroxinemia on the progeny. 2000;10:871–887.Thyroid
22. Morreale de Escobar G, Obregon MJ, Escobar del Rey F. Is neuropsychological development related to maternal hypothyroidism or to maternal hypothyroxinemia? 2000;85:3975–3987.J Clin Endocrinol Metab
23. Calvo RM, Jauniaux E, Gulbis B, et al. Fetal tissues are exposed to biologically relevant free thyroxine concentrations during early phases of development. 2002;87: 1768–1777.J Clin Endocrinol Metab
24. Dowling JT, Appleton WG, Nicoloff JT. Thyroxine turnover during human pregnancy. 1967;27: 1749–1750.J Clin Endocrinol Metab
25. Mandel SJ, Larsen PR, Seely EW. Increased need for thyroxine during pregnancy in women with primary hypothyroidism. 1990;323:91–96.N Engl J Med
26. Kaplan MM. Monitoring thyroxine treatment during pregnancy. 1992;2:147–152.Thyroid
27. Glinoer D, Gershengorn MC, Dubois A, et al. Stimulation of thyroxine-binding globulin synthesis by isolated rhesus monkey hepatocytes after β-estradiol administration. 1977;100:807–813.in vivoJ Clin Endocrinol Metab
28. Ain KB, Refetoff S. Relationship of oligosaccharide modification to the cause of serum thyroxine-binding globulin excess. 1988;66:1037–1043.J Clin Endocrinol Metab
29. Bartalena L. Recent achievements in studies on thyroid hormone-binding proteins. 1990;11:47–64.Endocr Rev
30. Osathanondh R, Tulchinsky D, Chopra IJ. Total and free thyroxine and triiodothyronine in normal and complicated pregnancy. 1976;42:98–103.J Clin Endocrinol Metab
31. Roti E, Gardini E, Minelli R, et al. Thyroid function evaluation by different commercially available free thyroid hormone measurement kits in term pregnant women and their newborns. 1991;14:1–9.J Endocrinol Invest
32. Hollowell JG, Staehling NW, Hannon WH, et al. Iodine nutrition in the United States. Trends and public health implications: iodine excretion data from national health and nutrition examination surveys I and III (1971–1974 and 1988–1994). 1998;83:3401–3408.J Clin Endocrinol Metab
33. Glinoer D. Pregnancy and iodine. 2001;11:471–481.Thyroid
34. Als C, Keller A, Minder C, et al. Age- and gender-dependent urinary iodine concentrations in an area-covering population sample from the Bernese region in Switzerland. 2000;143:629–637.Eur J Endocrinol
35. Caron P, Hoff M, Bazzi S, et al. Urinary iodine excretion during normal pregnancy in healthy women living in the southwest of France: correlation with maternal thyroid parameters. 1997;7:749–754.Thyroid
36. Nöhr SB, Laurberg P, Borlum K-G, et al. Iodine deficiency in pregnancy in Denmark: regional variations and frequency of individual iodine supplementation. 1993;72:350–353.Acta Obstet Gynecol Scand
37. Ciardelli R, Haumont D, Gnat D, et al. The nutritional iodine supply of Belgian neonates is still insufficient. 2002;161:519–523.Eur J Pediatr
38. Dunn JT, Delange F. Damaged reproduction: the most important consequence of iodine deficiency. 2001;86:2360–2363.J Clin Endocrinol Metab
39. Elnagar B, Eltom A, Wide L, et al. Iodine status, thyroid function and pregnancy: study of Swedish and Sudanese women. 1998;52:351–355.Eur J Clin Nutr
40. Vermiglio F, Lo Presti VP, Castagna MG, et al. Increased risk of maternal thyroid failure with pregnancy progression in an iodine deficient area with major iodine deficiency disorders. 1999;9:9–24.Thyroid
41. Glinoer D. Feto-maternal repercussions of iodine deficiency during pregnancy. 2003;64:37–44.Ann Endocrinol (Paris)
42. Glinoer D, De Nayer P, Delange F, et al. A randomized trial for the treatment of mild iodine deficiency during pregnancy: maternal and neonatal effects. 1995;80: 258–269.J Clin Endocrinol Metab
43. Glinoer D, Lemone M, Bourdoux P, et al. Partial reversibility during late postpartum of thyroid abnormalities associated with pregnancy. 1992;74:453–457.J Clin Endocrinol Metab
44. Glinoer D, Lemone M. Goiter and pregnancy: a new insight into an old problem. 1992;2:65–70.Thyroid
45. Rotondi M, Amato G, Biondi B, et al. Parity as a thyroid size-determining factor in areas with moderate iodine deficiency. 2000;85:4534–4537.J Clin Endocrinol Metab
46. Knudsen N, Bülow I, Laurberg P, et al. Parity is associated with increased thyroid volume solely among smokers in an area with moderate to mild iodine deficiency. 2002;146: 39–43.Eur J Endocrinol
47. Romano R, Jannini EA, Pepe M, et al. The effects of iodoprophylaxis on thyroid size during pregnancy. 1991;164:482–485.Am J Obstet Gynecol
48. Kung AWC, Lao TT, Chau MT, et al. Goitrogenesis during pregnancy and neonatal hypothyroxinemia in a borderline iodine sufficient area. 2000;53:725–731.Clin Endocrinol
49. Nöhr SB, Laurberg P. Opposite variations in maternal and neonatal thyroid function induced by iodine supplementation during pregnancy. 2000;85:623–627.J Clin Endocrinol Metab
50. Antonangeli L, Maccherini D, Cavaliere R, et al. Comparison of two different doses of iodide in the prevention of gestational goiter in marginal iodine deficiency: a longitudinal study. 2002;147:29–34.Eur J Endocrinol
51. Geenen V, Perrier de Hauterive S, Puit M, et al. Autoimmunity and pregnancy: theory and practice. 2002;57: 317–324.Acta Clin Belg
52. Sridama V, Pacini F, Yang SL, et al. Decreased level of helper T cells: a possible cause of immunodeficiency in pregnancy. 1982;307:352–356.N Engl J Med
53. Stagnaro-Green A, Roman SH, Cobin RH, et al. A prospective study of lymphocyte-initiated immunosuppression in normal pregnancy: evidence of a T-cell etiology for postpartum dysfunction. 1992;74:645–663.J Clin Endocrinol Metab
54. DeGroot LJ, Quintans J. The causes of autoimmune thyroid disease. 1989;10:537–562.Endocr Rev
55. Amino N, Tada H, Hidaka Y, et al. Postpartum autoimmune thyroid syndrome. 2000;47:645–655.Endocr J
56. Poppe K, Glinoer D. Thyroid autoimmunity and hypothyroidism before and during pregnancy. 2003;9:149–161.Hum Reprod Update
57. Stagnaro-Green A, Roman SH, Cobin RH, et al. Detection of at-risk pregnancy by means of highly sensitive assays for thyroid antibodies. 1990;264:1422–1425.JAMA
58. Glinoer D, Fernandez Soto ML, Bourdoux P, et al. Pregnancy in patients with mild thyroid abnormalities: maternal and neonatal repercussions. 1991;73:421–427.J Clin Endocrinol Metab
59. Muller AF, Verhoeff A, Mantel MJ, et al. Thyroid autoimmunity and abortion: a prospective study in women undergoing fertilization. 1999;71:30–34.in vitroFertil Steril
60. Abramson J, Stagnaro-Green A. Thyroid antibodies and fetal loss: an evolving story. 2001;11:57–63.Thyroid
61. Matalon ST, Blank M, Ornoy A, et al. The association between anti-thyroid antibodies and pregnancy loss. 2001;45:72–77.Am J Reprod Immunol
62. Bussen S, Steck T, Dietl J. Increased prevalence of thyroid antibodies in women with a history of recurrent fertilization failure. 2000;15:545–548.in-vitroHum Reprod
63. Dendrinos S, Papasteriades C, Tarassi K, et al. Thyroid autoimmunity in patients with recurrent spontaneous abortion. 2000;14:270–274.Gynecol Endocrinol
64. Glinoer D [Editorial]. Thyroid immunity, thyroid dysfunction, and the risk of miscarriage. 2000,43: 202–203.Am J Reprod Immunol
65. Practice Committee of the American Society for Reproductive Medicine: aging and infertility: a committee opinion. 2002;78:215–219.Fertil Steril
66. Arojoki M, Jokimaa V, Juuti A, Koskinen P, et al. Hypothyroidism among infertile women in Finland. 2000;14:127–131.Gynecol Endocrinol
67. Vaquero E, Lazzarin N, De Carolis C, et al. Mild thyroid abnormalities and recurrent spontaneous abortion: diagnostic and therapeutic approach. 2000;43:204–208.Am J Reprod Immunol
68. Glinoer D, Rihai M, Grün JP, et al. Risk of subclinical hypothyroidism in pregnant women with autoimmune thyroid disorders. 1994;79:197–204.J Clin Endocrinol Metab
69. Glinoer D. Management of hypo- and hyperthyroidism during pregnancy. 2003 (in press).Growth Horm IGF Res
70. Lazarus JH, Othman S. Thyroid disease in relation to pregnancy. 1991;34:91–98.Clin Endocrinol
71. Mestman JH, Goodwin M, Montoro MM. Thyroid disorders of pregnancy. 1995;24:41–71.Endocrinol Metab Clin North Am
72. Klein RZ, Haddow JE, Faix JD, et al. Prevalence of thyroid deficiency in pregnant women. 1991;35:41–46.Clin Endocrinol
73. Thomas R, Reid RL. Thyroid disease and reproductive failure. 1987;70:789–798.Obstet Gynecol
74. Lowe TW, Cunningham FG. Pregnancy and thyroid disease. 1991;34:72–81.Clin Obstet Gynecol
75. Jovanovic-Peterson L, Peterson CM. clinical hypothyroidism in pregnancies complicated by type I diabetes, subclinical hypothyroidism, and proteinuria: a new syndrome. 1988;159:442–446.De novoAm J Obstet Gynecol
76. Montoro M, Collea JV, Frasier SD, et al. Successful outcome of pregnancy in women with hypothyroidism. 1981;94:31–34.Ann Intern Med
77. Davis LE, Leveno KJ, Cunningham FG. Hypothyroidism complicating pregnancy. 1988;72:108–112.Obstet Gynecol
78. Liu H, Momotani N, Noh JY, et al. Maternal hypothyroidism during early pregnancy and intellectual development of progeny. 1994;154:785–787.Arch Intern Med
79. Abalovich M, Gutierrez S, Alcaraz G, et al. Overt and subclinical hypothyroidism complicating pregnancy. 2002;12: 63–68.Thyroid
80. Costante G, Crupi D, Trimarchi F, et al. Hypothyroidism induced by pregnancy in a patient submitted to suppressive L-thyroxine therapy. 1987;10:527.J Endocrinol Invest
81. Tamaki H, Amino N, Takeoka K, et al. Thyroxine requirement during pregnancy for replacement therapy of hypothyroidism. 1990;76:230–233.Obstet Gynecol
82. McDougall IR, Maclin N. Hypothyroid women need more thyroxine when pregnant. 1995;41:238–240.J Fam Pract
83. Roti E, Minelli R, Salvi M. Management of hyperthyroidism and hypothyroidism in the pregnant woman. 1996;81:1679–1682.J Clin Endocrinol Metab
84. Porterfield SP, Hendrich CE. The role of thyroid hormones in prenatal and neonatal neurological development: current perspectives. 1993;14:94–106.Endocr Rev
85. Smallridge RC, Ladenson PW. Hypothyroidism in pregnancy: consequences to neonatal health. 2001;86:2349–2353.J Clin Endocrinol Metab
86. Haddow JE, Palomaki GE, Allan WC, et al. Maternal thyroid deficiency during pregnancy and subsequent neuropsychological development of the child. 1999;34:549–555.N Engl J Med
87. Glinoer D. The systematic screening and management of hypothyroidism and hyperthyroidism during pregnancy. 1998;9:403–411.Trends Endocrinol Metab
88. Pop VJ, Kuijpens JL, Van Baar AL, et al. Low maternal free thyroxine concentrations during early pregnancy are associated with impaired psychomotor development in early infancy. 1999;50:149–155.Clin Endocrinol
89. Rotondi M, Caccavale C, Di Serio C, et al. Successful outcome of pregnancy in a thyroidectomized-parathyroidectomized young woman affected by severe hypothyroidism. 1999;9:1037–1040.Thyroid
90. Matsuura N, Konishi J, and the transient hypothyroidism study group in Japan. Transient hypothyroidism in infants born to mothers with chronic thyroiditis—a nationwide study of twenty-three cases. 1990;37:369–379.Endocrinol Jpn
91. Becks GP, Burrow GN. Thyroid disease and pregnancy. 1991;75:121–150.Med Clin North Am
92. Man EB, Brown JF, Serunian SA. Maternal hypothyroidism: psychoneurological deficits of progeny. 1991; 21:227–239.Ann Clin Lab Ser
93. Kamijo K, Saito T, Yachi A, et al. Transient subclinical hypothyroidism in early pregnancy. 1990;37:397–403.Endocrinol J
94. Wang C, Crapo LM. The epidemiology of thyroid disease and implications for screening. 1997;26:189–218.Endocrinol Metab Clin North Am
95. Goodwin TM, Montoro M, Mestman JH. The role of chorionic gonadotropin in transient hyperthyroidism of hyperemesis gravidarum. 1992;75:1333–1337.J Clin Endocrinol Metab
96. Yoshimura M, Hershman JM. Thyrotropic action of human chorionic gonadotropin. 1995;5:425–434.Thyroid
97. Glinoer D. Thyroid regulation and dysfunction during pregnancy. In: DeGroot L, ed. (update 2003). Accessed at Thyroid disease managerwww.thyroidmanager.org.
98. Mestman JH. Hyperthyroidism in pregnancy. 1997;40:45–64.Clin Obstet Gynecol
99. Mandel SJ, Brent GA, Larsen PR. Review of antithyroid drug use during pregnancy and report of a case of aplasia cutis. 1994;4:129–133.Thyroid
100. Davis LE, Lucas MJ, Hankins GD, et al. Thyrotoxicosis complicating pregnancy. 1989;160:63–70.Am J Obstet Gynecol
101. Mitsuda N, Tamaki H, Amino N, et al. Risk factors for developmental disorders in infants born to women with Graves' disease. 1992;80:356–364.Obstet Gynecol
102. Momotani N, Noh J, Oyanagi H, et al. Antithyroid drug therapy for Graves' disease during pregnancy: optimal regimen for fetal thyroid status. 1986;315:24–28.N Engl J Med
103. Glinoer D. Thyroid hyperfunction during pregnancy. 1998;8:859–864.Thyroid
104. Drury MI. Hyperthyroidism and pregnancy. 1986;79:317–318.J R Soc Med
105. Laurberg P, Nygaard B, Glinoer D, et al. Guidelines for TSH-receptor antibody measurements in pregnancy: results of an evidence-based symposium organized by the European Thyroid Association. 1998;139:584–586.Eur J Endocrinol
106. Amino N, Tanizawa O, Mori H, et al. Aggravation of thyrotoxicosis in early pregnancy and after delivery in Graves' disease. 1982;55:108–112.J Clin Endocrinol Metab
107. Kung AW, Jones BM. A change from stimulatory to blocking antibody activity in Graves' disease during pregnancy. 1998;83:514–518.J Clin Endocrinol Metab
108. Kung AW, Lau KS, Kohn LD. Epitope mapping of TSH receptor-blocking antibodies in Graves' disease that appear during pregnancy. 2001;86:3647–3653.J Clin Endocrinol Metab
109. Jones BM, Kwok JS, Kung AW. Changes in cytokine production during pregnancy in patients with Graves' disease. 2000;10:701–707.Thyroid
110. Gardner DF, Cruikshank DP, Hays PM, et al. Pharmacology of propylthiouracil (PTU) in pregnant hyperthyroid women: correlation of maternal PTU concentration with cord serum thyroid function tests. 1986;62:217–220.J Clin Endocrinol Metab
111. Wing DA, Millar LK, Koonings PP, et al. A comparison of propylthiouracil versus methimazole in the treatment of hyperthyroidism in pregnancy. 1994;170: 90–95.Am J Obstet Gynecol
112. Mestman JH. Hyperthyroidism in pregnancy. 1998;27:127–149.Endocrinol Clin North Am
113. Clementi M, Di Gianantonio E, Pelo E, et al. Methimazole embryopathy: delineation of the phenotype. 1999;83:43–46.Am J Med Genet
114. Johnsson E, Larsson G, Ljunggren M. Severe malformations in infant born to hyperthyroid woman on methimazole. 1997;350:1520.Lancet
115. Azizi F, Khoshniat M, Bahrainian M, et al. Thyroid function and intellectual development of infants nursed by mothers taking methimazole. 2000;85:3233–3238.J Clin Endocrinol Metab
116. Cheron RG, Kaplan MM, Larsen PR, et al. Neonatal thyroid function after propylthiouracil therapy for maternal Graves' disease. 1981;304:525–528.N Engl J Med
117. Frieden IJ. Aplasia cutis congenita: a clinical review and proposal for classification. 1986;14:646–660.Acta Dermatol
118. Skuza KA, Sills IN, Stene M, et al. Prediction of neonatal hyperthyroidism in infants born to mothers with Graves' disease. 1996;128:264–267.J Pediatr
119. Polak M, Leger J, Luton D, et al. Fetal cord blood sampling in the diagnosis and treatment of fetal hyperthyroidism in the offspring of a euthyroid mother, producing thyroid stimulating antibodies. 1997;58:338–342.Ann Endocrinol (Paris)
120. Momotani N, Noh JY, Ishikawa N, et al. Effects of propylthiouracil (PTU) and methimazole (MMI) on fetal thyroid status in mothers with Graves' hyperthyroidism. 1997;82:3633–3636.J Clin Endocrinol Metab
121. Azizi F, Khamseh ME, Bahreynian M, et al. Thyroid function and intellectual development of children of mothers taking methimazole during pregnancy. 2002;25:586–589.J Endocrinol Invest
122. McKenzie JM, Zakarija M. Fetal and neonatal hyperthyroidism due to maternal TSH receptor antibodies. 1992;2: 155–163.Thyroid
123. Tamaki H, Amino N, Aozasa M, et al. Universal predictive criteria for neonatal overt thyrotoxicosis requiring treatment. 1988;5:152–158.Am J Perinatol
124. Zakarija M, Garcia A, McKenzie JM. Studies on multiple thyroid cell membrane directed antibodies in Graves' disease. 1985;76:1885–1891.J Clin Invest
125. Lamberg BA, Ikonen E, Osterlund K, et al. Antithyroid drug treatment of maternal hyperthyroidism during lactation. 1984;21:81–87.Clin Endocrinol
126. Mandel SJ, Cooper DS. The use of antithyroid drugs in pregnancy and lactation. 2001;86:2354–2359.J Clin Endocrinol Metab
127. Amino N, Mori H, Iwatani Y, et al. High prevalence of transient post-partum thyrotoxicosis and hypothyroidism. 1982;306:849–852.N Engl J Med
128. Amino N, Tada H, Hidaka Y. Autoimmune thyroid disease and pregnancy. 1996;19:59–70.J Endocrinol Invest
129. Browne-Martin K, Emerson CH. Postpartum thyroid dysfunction. 1997;40:90–101.Clin Obstet Gynecol
130. Tada H, Hidaka Y, Tsuruta E, et al. Prevalence of postpartum onset of disease within patients with Graves' disease of child-bearing age. 1994;41:325–327.Endocr J
131. Momotani N, Noh J, Ishikawa N, et al. Relationship between silent thyroiditis and recurrent Graves' disease in the postpartum period. 1994;79:285–289.J Clin Endocrinol Metab
132. Hidaka Y, Tamaki H, Iwatani Y, et al. Prediction of post-partum Graves' thyrotoxicosis by measurements of thyroid stimulating antibody in early pregnancy. 1994;41: 15–20.Clin Endocrinol
133. Yoshida S, Takamatsu J, Kuma K, et al. Thyroid-stimulating antibodies and thyroid stimulation-blocking antibodies during the pregnancy and postpartum period: a case report. 1992;2:27–30.Thyroid
134. Goodwin TM, Montoro M, Mestman JH. Transient hyperthyroidism and hyperemesis gravidarum: clinical aspects. 1992;167:648–652.Am J Obstet Gynecol
135. Kimura M, Amino M, Tamaki H, et al. Gestational thyrotoxicosis and hyperemesis gravidarum: possible role of hCG with higher stimulating activity. 1993;38:345–350.Clin Endocrinol
136. Tsuruta E, Tada H, Tamaki H, et al. Pathogenic role of asialo human chorionic gonadotropin in gestational thyrotoxicosis. 1995;80:350–355.J Clin Endocrinol Metab
137. Ramsey PS, Ramin KD. 24-year-old pregnant woman with nausea and vomiting. 2000;75:1317–1320.Mayo Clin Proc
138. Tanaka S, Yamada H, Kato EH, et al. Gestational transient hyperthyroxinaemia (GTH): screening for thyroid function in 23163 pregnant women using dried blood spots. 1998;49:325–329.Clin Endocrinol
139. Yeo CP, Hsu Chin Khoo D, Hsi Ko Eng P, Tan HK et al. Prevalence of gestational thyrotoxicosis in Asian women evaluated in the 8th to 14th weeks of pregnancy: correlations with total and free beta human chorionic gonadotrophin. 2001;55:391–398.Clin Endocrinol
140. Grün JP, Meuris S, De Nayer P, et al. The thyrotropic role of human chorionic gonadotropin (hCG) in the early stages of twin (versus single) pregnancy. 1997;46:719–725.Clin Endocrinol
141. Vassart G, Dumont JE. The thyrotropin receptor and the regulation of thyrocyte function and growth. 1992;13: 596–611.Endocr Rev
141a. Ginsberg J, Levanczuk RZ, Honore LH. Hyperplacentosis: a novel cause of hyperthyroidism. 2001;11:393–396.Thyroid
142. Kosugi S, Mori T. TSH receptor and LH receptor. 1995;42:587–606.Endocr J
143. Rodien P, Brémont C, Sanson M-L, et al. Familial gestational hyperthyroidism caused by a mutant thyrotropin receptor hypersensitive to human chorionic gonadotropin. 1998;339:1823–1826.N Engl J Med
144. Smits G, Govaerts C, Nubourgh I, et al. Lysine 183 and glutamic acid 157 of the TSH receptor: two interacting residues with a key role in determining specificity toward TSH and human CG. 2002;16:722–735.Mol Endocrinol
145. Rodien P. Un récepteur de la TSH hypersensible à l'hCG responsible d'une hyperthyroïdie gestationelle récidivante et familiale. 2001;14:655–660.Reprod Hum Horm
146. Goodwin TM, Hershman JM. Hyperthyroidism due to inappropriate production of human chorionic gonadotropin. 1997;40:32–44.Clin Obstet Gynecol
147. Bober SA, McGill AC, Tunbridge WMG. Thyroid function in hyperemesis gravidarum. 1986;111: 404–410.Acta Endocrinol (Copenh)
148. Lao TT, Chin KH, Chang AMZ. The outcome of hyperemetic pregnancies complicated by transient hyperthyroidism. 1987;27:99–101.Aust N Z J Obstet Gynaecol
149. Wilson R, McKillop JH, McLean M, et al. Thyroid function tests are rarely abnormal in patients with severe hyperemesis gravidarum. 1992;37:331–334.Clin Endocrinol (Oxf)
150. Tan JY, Loh KC, Yeo GS, et al. Transient hyperthyroidism of hyperemesis gravidarum. 2002;109:683–688.Br J Obstet Gynaecol
151. Rosen IB, Walfish PG. Pregnancy as a predisposing factor in thyroid neoplasia. 1986;121:1287–1290.Arch Surg
152. Hamburger JI. Thyroid nodules in pregnancy. 1992;2: 165–168.Thyroid
153. Moosa M, Mazzaferri EL. Outcome of differentiated thyroid cancer diagnosed in pregnant women. 1997;82:2862–2866.J Clin Endocrinol Metab
154. Tan GH, Gharib H, Goellner JR, et al. Management of thyroid nodules in pregnancy. 1996;156:2317–2320.Arch Intern Med
155. Glinoer D. Nodule et cancer thyroïdiens chez la femme enceinte. 1997;58:263–267.Ann Endocrinol (Paris)
156. Marley EF, Oertel YC. Fine-needle aspiration of thyroid lesions in fifty seven pregnant and postpartum women. 1997;16:122–125.Diagn Cytopathol
157. Hod M, Sharony R, Friedman S, et al. Pregnancy and thyroid carcinoma: a review of incidence, course, and prognosis. 1989;44:774–779.Obstet Gynecol Surv
158. Asteris GT, DeGroot LJ. Thyroid cancer: relationship to radiation exposure and pregnancy. 1976;4:1287–1291.J Reprod Med
159. Mc Clellan DR, Francis GL. Thyroid cancer in children, pregnant women, and patients with Graves' disease. 1996;25:27–48.Endocrinol Metab Clin North Am
160. Vini L, Hyer S, Pratt B, et al. Management of differentiated thyroid cancer diagnosed during pregnancy. 1999;140:404–406.Eur J Endocrinol
161. Struve C, Haupt S, Ohlen S. Influence of frequency of previous pregnancies on the prevalence of thyroid nodules in women without clinical evidence of thyroid disease. 1993;3:7–9.Thyroid
162. Kung AW, Chau MT, Lao TT, et al. The effect of pregnancy on thyroid nodule formation. 2002;87: 1010–1014.J Clin Endocrinol Metab