3.4.5 Thyroid disease during pregnancy

Thyroid disorders are common. The prevalence of hyperthyroidism is around 5/1000 in women and overt hypothyroidism about 3/1000 in women. Subclinical hypothyroidism has a prevalence in women of childbearing age in iodine-sufficient areas of between 4% and 8%. As these conditions are generally much more common in females, it is to be expected that they will appear during pregnancy. Developments in our understanding of thyroid physiology (1) and immunology (2) in pregnancy, as well as improvements in thyroid function testing (3), have highlighted the importance of recognizing and providing appropriate therapy to women with gestational thyroid disorders. Before considering the clinical entities occurring during and after pregnancy it is useful to briefly review thyroid physiology and immunology in relation to pregnancy.

The recommended daily iodine intake in pregnancy has been increased to 250 μg/day which implies a urinary iodine excretion of 150–250 μg/day as being adequate (4) Urinary iodine excretion in pregnancy is maximal in the first trimester followed by a decline in the second and third trimesters. Often there is an increase in urinary iodine in the first trimester compared to control nonpregnant women, but where the population has a high median iodine concentration this difference may not occur.

Iodine deficiency during pregnancy is associated with maternal goiter due to the imbalance between the intake and increased requirements for iodine during gestation and results eventually in a reduced circulating maternal thyroxine (T₄) concentration. This gestational goitrogenesis is preventable by iodine supplementation (5) not only in areas of severe iodine deficiency (24-h urinary iodine less than 50 μg) but also in areas where trials have shown clear beneficial effects on maternal thyroid size. Clinical studies of children born to mothers with known iodine deficiency clearly showed impaired neurointellectual development, sometimes to the extreme of cretinism in severely deficient states. These defects can be corrected by iodine administration before and even during gestation and this should be performed in areas of moderate to severe iodine deficiency.

Pregnancy is associated with significant, but reversible changes in thyroid function (Table 3.4.5.1). Thyroid hormone transport proteins, particularly thyroxine-binding globulin, increase due to enhanced hepatic synthesis and a reduced degradation rate due to oligosaccharide modification. Serum concentrations of free thyroid hormones are decreased, increased, or unchanged during gestation depending on the assays used. Nevertheless, there is a transient rise in free T₄ in the first trimester due to the relatively high circulating human chorionic gonadotropin (hCG) concentration and a decrease of free T₄ in the second and third trimester. Because of these variations (Fig. 3.4.5.1) there is a need for normative trimester-specific reference ranges for thyroid hormones (6). In iodine-deficient areas (including marginal iodine deficiency seen in many European countries), pregnant woman may become significantly hypothyroxinaemic with preferential triiodothy-ronine (T₃) secretion. The thyroidal ‘stress’ is also evidenced by a rise in the median thyroid-stimulating hormone (TSH) and serum thyroglobulin.

The fetal thyroid begins concentrating iodine at 10–12 weeks gestation and is under the control of fetal pituitary TSH by about 20 weeks gestation. Although there is no functioning fetal thyroid in early pregnancy, thyroid hormone is important in the development of many organs including the brain. Maternal circulating T₄ crosses the placenta into the fetus at all stages of pregnancy by incompletely understood mechanisms but involving both the type 2 and type 3 deiodinase enzymes, both expressed in the placenta. Type 3 deiodinase (D3), which degrades thyroid hormones (7), is also expressed in pregnant uterus, placenta, and fetal and neonatal tissues, and may act as a ‘gatekeeper’ to prevent too much thyroid hormone transport. Type 2 deiodinase, also located in the uterus and other parts of the genital tract, degrades T₄ in the fetus to provide T₃ for tissue growth and differentiation and may have a role in fetal implantation.

Human immune regulation involves homeostasis between T helper 1 (Th1) and T helper 2 (Th2) activity, with Th1 cells driving cellular immunity and Th2 cells humoral immunity (2). In pregnancy there is a bias towards a Th2 lymphocyte response evidenced by the fetal/placental unit producing Th2 cytokines, which inhibit Th1. Th1 cytokines are potentially harmful to the fetus as, e.g. interferon-α is a known abortifacient.

Pregnancy also has a significant effect on the immune system in order to maintain the fetal–maternal allograft and prevent rejection (Box 3.4.5.1). The trophoblast does not express the classic major histocompatibility complex (MHC) class Ia or II which are needed to present antigenic peptides to cytotoxic cells and T helper cells, respectively. Instead HLA-G, a nonclassic MHC Ib molecule is expressed which may be a ligand for the natural killer (NK) cell receptor so protecting the fetus from NK cell damage; it may also activate CD8⁺ T cells that may have a suppressor function. Human trophoblasts also express abundant Fas ligand, thereby contributing to the immune privilege by mediating apoptosis of activated Fas-expressing lymphocytes of maternal origin (8).

Antithyroid peroxidase antibodies (TPOAbs) are found in around 10% of otherwise normal pregnant women when measured at the end of the first trimester. The presence of TPOAbs before and during gestation have several implications. Fertility is impaired in hypothyroid women with autoimmune thyroid disease and, if such patients do achieve pregnancy, the hypothyroid state is associated with a higher incidence of miscarriage early in pregnancy. Thyroid autoimmunity, with positive TPOAbs present during early pregnancy even in the euthyroid situation, is associated with an increased risk of subsequent miscarriage (9). TPOAb-positive women miscarry at a rate of between 13% and 22% compared to 3.3–8.4% in control euthyroid antibody-negative women. One controlled trial has shown that thyroxine administration reduced the miscarriage rate in TPOAb-positive women. The association between TPOAbs and recurrent abortion is less strong than for miscarriage but one uncontrolled study reported a significant success rate with thyroxine administration (10).

Hyperthyroidism occurs in 2/1000 pregnancies, the commonest cause (85%) being Graves’ hyperthyroidism, due to thyroid stimulation by thyrotropin receptor stimulating antibodies (TRAb) (Box 3.4.5.2). Transient gestational thyrotoxicosis (due to thyroid stimulation by hCG) is seen in the first trimester and is more common in Asian women than European women. A deterioration in previously diagnosed Graves’ disease is not infrequent during the first trimester of pregnancy, and may be due to an increase in the titre of TRAb or high levels of hCG acting as a thyroid stimulator. Relapse may also be caused by impaired absorption of antithyroid medication secondary to pregnancy-associated vomiting or by reluctance to continue medication in the first trimester.

The immune status of pregnancy is a Th2 state, which allows tolerance of the fetus during pregnancy, and this is thought to be the reason why there is usually an amelioration of the severity of Graves’ hyperthyroidism (and other autoimmune diseases) after the first trimester. Graves’ hyperthyroidism before pregnancy may remit during pregnancy but will exacerbate in the postpartum period as the immune status reverts to a Th1 state.

Maternal complications of hyperthyroidism include miscarriage, placental abruption, and preterm delivery. Congestive heart failure and thyroid storm may also occur, the risk of pre-eclampsia is significantly higher in women with poorly controlled hyperthyroidism, and a low-birthweight infant may be up to nine times more likely (11). Neonatal hyperthyroidism, prematurity, and intrauterine growth retardation may be observed. There are no increased risks of subclinical hyperthyroidism. Women with thyroid hormone resistance (where thyroid hormone levels and TSH are inappropriately high not due to autoimmunity) also have a high miscarriage rate, indicating a direct toxic effect of thyroid hormones on the fetus.

There is no doubt that overt clinical and biochemical hyperthyroidism should be treated to lessen the rate of complications described above. Gestational amelioration of Graves’ disease is usually associated with a reduction in titre of TSH-receptor antibodies and sometimes a change from stimulatory to blocking antibody activity. Of neonates of mothers with Graves’ disease, 1–5% have hyperthyroidism due to the transplacental passage of maternal stimulating TRAb (even though the mother may be euthyroid and has received previous treatment for Graves’ disease). Neonatal hyperthyroidism may also be due to an activating mutation of the TSH receptor dominantly inherited from the mother. Transient neonatal central hypothyroidism is due to poorly controlled Graves’ disease leading to suppression of the fetal pituitary–thyroid axis due to placental transfer of T₄.

Undiagnosed Graves’ hyperthyroidism is present in approximately 0.15% and others will already be known to have the disease before gestation. Features such as tachycardia, palpitations, systolic murmur, bowel disturbance, emotional upset, and heat intolerance may be seen in normal pregnancy but should alert the clinician to the possibility of hyperthyroidism, particularly if goiter or more specific features of thyroid disease (weight loss, eye signs, tremor, or pretibial myxoedema) are observed.

Ideally a woman who is known to have hyperthyroidism should seek prepregnancy advice; appropriate education should allay fears that are commonly present in these women. She should be referred for specialist care for frequent checking of her thyroid status, thyroid antibody evaluation, and close monitoring of her medication needs (12).

At all stages of pregnancy, the use of antithyroid drugs (ATDs) is the preferred treatment option. Radio-iodine is contraindicated and surgery requires pretreatment with ATDs to render the patient euthyroid. The thionamides carbimazole (CMI), methimazole (MMI), and propylthiouracil (PTU) are all effective in inhibiting thyroidal biosynthesis of thyroxine during pregnancy. PTU is the preferred drug in pregnancy due to the possibility (albeit rare) of teratogenic effects of CMI and MMI (aplasia cutis and MMI embryopathy). There are no long-term adverse effects of ATD exposure in utero, in particular on IQ scores or psychomotor development in MMI- and PTU-exposed individuals. The starting dosage of PTU is 300–450 mg/day, up to 600 mg daily if necessary, given in two or three divided doses. Some improvement is usually seen after 1 week of treatment with ATDs but 4–6 weeks may be needed for a full effect. Once the thyrotoxicosis has been controlled, the dose needs to be gradually reduced by one-quarter to one-third every 3–4 weeks, typically to 50–100 mg twice daily. The main principle of therapy is to administer the lowest ATD dose needed for controlling clinical symptoms, with the aim of restoring normal maternal thyroid function but ensuring that fetal thyroid function is minimally affected. Maternal free T₄ levels should be kept in the upper one-third of the normal nonpregnant reference range to avoid fetal hypothyroidism, as with this management serum free T₄ levels are normal in more than 90% of neonates. The administration of l-thyroxine together with PTU as a ‘block and replace’ regimen is not advisable in pregnancy as the amount of ATD may be excessive in proportion to the amount of thyroxine which crosses the placenta, resulting in fetal goiter and hypothyroidism. Recently there has been concern expressed relating to the hepatic side effects of PTU. The current recommendation is therefore to use PTU only in the first trimester.

β-adrenergic blocking agents such as propranolol may  be used for a few weeks to ameliorate the peripheral sympathomimetic actions of excess thyroid hormone but prolonged use can result in restricted fetal growth, impaired response to hypoxic stress, together with postnatal bradycardia and hypoglycaemia. If a woman is already receiving CMI, a change to PTU is recommended. Although these patients may have received ATDs, surgery, or radio-iodine therapy and be euthyroid on or off thyroxine therapy, neonatal hyperthyroidism may still occur.

TRAb should be measured early in pregnancy in a euthyroid pregnant women previously treated by either surgery or radio-iodine (13). If the TRAb level is high at this time the fetus should be evaluated carefully during gestation by serial ultrasonography (14). Ultrasonographic evidence of fetal thyroid disease includes intrauterine growth restriction, tachycardia, cardiac failure, hydrops, advanced bone age, and goiter. In the presence of a fetal goiter, it may not be possible to distinguish fetal hyper- from hypothyroid disease on clinical grounds; fetal blood sampling may then be necessary to enable a diagnosis to be made. If fetal hyperthyroidism is diagnosed, treatment involves modulation of maternal ATDs. If fetal hypothyroidism has resulted from administration of ATDs to the mother, maternal treatment should be decreased or stopped and administration of intra-amniotic thyroxine considered. Early delivery may need to be considered in the case of fetal thyroid dysfunction, depending on the gestation at diagnosis and the severity of fetal symptoms. TRAb should be measured again in the last trimester (at about 32 weeks) and if positive the neonate needs to be checked for hyperthyroidism following delivery.

Thyroid surgery (in the second trimester) is indicated if control of the hyperthyroidism is poorly controlled on account of poor compliance, inability to take drugs, or pressure symptoms due to goiter size. The administration of radioactive iodine (¹³¹I) is contraindicated during pregnancy. Because fetal thyroid uptake of ¹³¹I commences after 12 weeks gestation, exposure before 12 weeks is not associated with fetal thyroid dysfunction and the irradiation dose is not considered sufficient to justify termination of pregnancy. However, the fetal thyroid does concentrate iodine after 13–15 weeks gestation and the fetal tissues are more radiosensitive. ¹³¹I given after this gestational age therefore potentially leads to significant radiation to the fetal thyroid, resulting in biochemical hypothyroidism and even cretinism in the neonate.

The incidence of hypothyroidism during pregnancy is around 2.5% (15) and is nearly always subclinical, which is equally as important in its adverse effects affecting mother and neonate as the full expression of the disease. The aetiology is usually autoimmune thyroiditis (TPOAb positive), but it may also be due to postoperative thyroid failure and noncompliance with existing thyroxine therapy. The symptoms of hypothyroidism, such as tiredness, are also seen in pregnancy. Many patients with subclinical hypothyroidism are asymptomatic but then notice an improvement after taking thyroid hormone therapy. Classic clinical features of hypothyroidism are described in Chapter 3.4.1. Maternal hypothyroxinaemia (without increased TSH) is also being increasingly accepted as deleterious to the neuropsychological development of the child (16). Care should be taken in the interpretation of TSH concentrations in early gestation due to the thyrotrophic effects of hCG.

Previous studies have documented the effects of hypothyroidism on maternal and fetal wellbeing, drawing attention to increased incidence of abortion, obstetrical complications, and fetal abnormalities in untreated women (Box 3.4.5.3). Women already receiving thyroxine for hypothyroidism require an increased dose during gestation (17). This is critical to ensure adequate maternal thyroxine levels for delivery to the fetus especially during the first trimester. The dose should normally be increased by 50–100 μg/day as soon as pregnancy is diagnosed; subsequent monitoring of TSH and free T₄ is then necessary to ensure correct replacement dosage.

Thyroid hormones are major factors for the normal development of the brain. The mechanisms of actions of thyroid hormones in the developing brain are mainly mediated through two ligand-activated thyroid hormone receptor isoforms. It is known that thyroid hormone deficiency may cause severe neurological disorders resulting from the deficit of neuronal cell differentiation and migration, axonal and dendritic outgrowth, myelin formation, and synaptogenesis (18). This is the situation well documented in iodine-deficient areas where the maternal circulating T₄ concentrations are too low to provide adequate fetal levels particularly in the first trimester. There is also evidence that in an iodine-sufficient area maternal thyroid dysfunction (hypothyroidism, subclinical hypothyroidism, or hypothyroxinaemia) during pregnancy results in neurointellectual impairment of the child. Haddow et al. (19) found that the full IQ scores of children whose mothers had a high TSH during gestation were 7 points lower than controls (p <0.005) and that 19% of them had scores of less than 85 compared to 5% of controls (p <0.007). Maternal hypothyroxinaemia during early gestation was shown to be an independent determinant of neurodevelopmental delay, but when free T₄ concentrations increased during gestation in women who had low free T₄ in early pregnancy, infant development was not adversely affected (20). Pop et al. (21) have also shown a significant decrement in IQ in children aged 5 years whose mothers were known to have circulating anti-TPOAbs at 32 weeks gestation and were biochemically euthyroid. The neurodevelopmental impairment is similar to that seen in iodine-deficient areas and implies that iodine status should be normalized in regions of deficiency. However, much of the USA and parts of Europe are not iodine deficient, which raises the question of routine screening of thyroid function during early pregnancy or even at preconception.

Thyroid nodules are claimed to be detected in up to 10% of pregnant women. Fine-needle aspiration biopsy is the first investigation of choice which may yield a malignancy/suspicious result in 35% (22). When malignancy is diagnosed it is usually a differentiated tumour which may be surgically resected in the second trimester or in some cases safely left until the postpartum period before therapy is started. The impact of pregnancy on thyroid cancer seems to be minimal in that there is no difference in rates of metastases or recurrence compared to nonpregnant women with the same disease. Whether women already treated for thyroid malignancy should become pregnant is of concern but current evidence suggests that differentiated thyroid cancer should not inhibit an intended pregnancy. Previous ¹³¹I therapy does not result in demonstrable adverse events in subsequent pregnancies, although miscarriage appears to be more frequent during the year preceding conception.

It is clear from the information already discussed relating to the effects of thyroid dysfunction in pregnancy on both mother and fetus together with the high prevalence of thyroid abnormalities that consideration be given to screening thyroid function in pregnancy with the aim of interventional therapy (with l-thyroxine) if necessary. The development of normative reference ranges for thyroid hormone during pregnancy would assist this process considerably. Screening is a strategy to detect a disease in asymptomatic individuals in order to improve health outcomes by early diagnosis and treatment. The current recommendation of the clinical practice guideline published under the auspices of the Endocrine Society (23) is that targeted screening should be performed in those women at high risk for thyroid disease (Box 3.4.5.4). A study to validate this strategy found that restricting screening to these groups of women would miss about one-third of women with significant thyroid dysfunction (23). A cost-effective analysis of screening pregnant women for autoimmune thyroid disease concluded that screening pregnant women in the first trimester for TSH was cost effective compared with no screening (24, 25). Screening using anti-TPOAbs was also cost effective. Until the results of carefully controlled randomized prospective outcome studies are available, the screening controversy will continue.