3.4.7 Thyroid disease in newborns, infants, and children

There are good reasons to describe congenital hypothyroidism and hyperthyroidism separately from acquired thyroid diseases because the risks of a disturbed thyroid hormone supply in young children are clearly different from the risks in older children or adults. For adequate metabolism, vertebrates with a higher degree of development, or a more complex ontogeny, are highly dependent on thyroid hormone. Nevertheless, humans appear to be able to ‘vegetate’ for years in the absence of this hormone. After resumption of hormone supply the metabolism normalizes again. However, brain development in young children does not. With the exception of the development of the neural tube, thyroid hormone is involved in regulation of later events, such as cell migration and the formation of cortical layers, and in neuronal and glial cell differentiation. Thyroid hormone also controls differentiation of not only neurons and oligodendrocytes, but also astrocytes and microglia (1).

The important role of the thyroid in brain development had already been recognized by 1850 when the British surgeon Curlings reported two mentally impaired children with large tongues, who appeared to have no thyroid gland at obduction. Later, more detailed publications about congenital hypothyroidism patients appeared and, in 1871, the British internist Fagge described some of his patients as extremely small (adult height less than 100 cm), with short broad hands and feet, a broad face with a flat root of the nose, thick nostrils, a large open mouth and thick lips, swollen skin, mental impairment, and often deaf. Osler, in 1897, called these patients ‘pariahs of nature’. Although at that time a relation was suggested between this striking disease and the absence of the thyroid, the function of this organ was still completely unknown. Remarkably, by the 1890s it was known that administration of (animal) thyroid preparations to children with congenital hypothyroidism improved their clinical condition markedly.

The belief that endemic cretinism, characterized by neurological problems such as mental impairment, deafness, pareses, spasticity, and squint, and endemic goiter might be caused by lack of iodine dates back to the 1850s. In the following century awareness gradually developed that the aforementioned cretinoid features in the offspring are the result of impaired thyroid hormone synthesis during pregnancy. In the event of long-standing iodine deficiency, neither the pregnant woman nor her fetus are able to make sufficient thyroxine (T₄) to prevent cerebral damage. Since 1989 it has been clear that the amount of T₄ that the healthy pregnant woman donates to her baby is usually sufficient to secure fetal brain development, even if the fetus itself is unable to produce T₄ (2). The fundamental value of an adequate maternal thyroid function during pregnancy is well illustrated by case histories of both severely impaired maternal and fetal T₄ production due to a dominant POU1F1 mutation (see section on disturbances in thyrotropin synthesis and regulation) and due to thyrotropin binding inhibiting immunoglobulins; in both instances children developed severe cognitive and motor disability, in spite of immediate postnatal T₄ therapy (3). Moreover, recent cohort studies have demonstrated that when women have a moderately impaired thyroid function, or just low to normal plasma free T₄ levels during early pregnancy, the mean IQ in the offspring is slightly impaired (4, 5).

The major problem in congenital hypothyroidism, disturbance of brain development resulting in life-long cognitive and motor problems, appears to be dependent on the severity and duration of the hypothyroid condition in the postnatal phase. Administration of T₄ to the affected neonate as soon as possible will largely prevent this problem (6). Because clinical signals are often lacking or are not recognized at that time, neonatal screening has been introduced in many countries. Diagnosis by means of such a mass-screening programme demands an essentially different approach to that used in individual symptomatic thyroid problems. Knowledge about the cause of congenital hypothyroidism is not only scientifically important, but also gives indispensable support to the treatment, (genetic) counselling, and knowledge about the long-term prognosis of the patient.

Throughout gestation the thyroid hormone supply of the fetal tissues is a subtle interplay between the fetal thyroid and its regulatory system, the maternal thyroid and its regulatory system, the various deiodinating enzymes and thyroid hormone receptors in the placenta, and the fetal target organs. This interplay brings about correct thyroid hormone status (optimal thyroid hormone receptor occupancy) in the different tissues, including the brain, in the different phases of development.

The thyroid develops primarily as a ventral bulge of the endoderm, located between the first and second branchial arches. Sometimes, in later life, a remnant of this median anlage is recognizable as the foramen caecum of the tongue. About 17 days after conception the human primordial thyroid can be detected close to the developing heart, and around day 30 a hollow bilobate structure is formed. Both lobes then fuse with the ultimobranchial bodies (lateral anlagen), developed from the fourth branchial pouches. The calcitonin-secreting cells (C cells) of the thyroid originate from these ultimobranchial bodies.

The thyrocytes are organized into tubes 8 weeks after conception, and 2 weeks later intercellular follicles form and iodine can be bound, indicating that the thyrocytes are able to synthesize thyroperoxidase and thyroglobulin, and to transport these thyroid-specific proteins into the follicular lumen by exocytosis. For some time, the number of follicles is increased by budding from the primary follicles; later on, the thyroid growth is mainly due to the increasing volume of existing follicles (7).

Near the end of the first trimester (free) T₄ and thyroxine-binding globulin become detectable in the fetal circulation, in very low concentrations compared to normal values for infants and adults (Fig. 3.4.7.1). Subsequently, the concentrations of thyroid-stimulating hormone (TSH), thyroxine-binding globulin, and T₄ increase more or less arithmetically, while free T₄ increases geometrically; all reach adult values at about 36 weeks. Until about 30 weeks gestation, fetal plasma triiodothyronine (T₃) is hardly detectable. It then increases geometrically, although the concentration at term is still very low compared to the normal values for infants and adults (Fig. 3.4.7.1) (8). In contrast, the prenatal levels of reverse T₃ are high (9). The fetus cannot produce its own T₄ until about midgestation and so is completely dependent on the maternal hormone supply. Thyroid hormone synthesis presumably increases gradually in the second half of gestation, since at term the infant provides its own T₄ supply completely.

Birth induces a number of changes in thyroid hormone production and metabolism within a short period (Fig. 3.4.7.1). This adaptation process starts with an acute surge of TSH into the circulation. About 30 min after birth, plasma TSH reaches its maximum level. Thereafter, it gradually decreases and stabilizes within 1–2 days at slightly higher values than those in adults (10). Immediately after birth a rapid and substantial surge of the plasma T₃ concentration takes place. The TSH surge significantly increases the thyroid production of T₄ and thyroglobulin. Plasma T₄ and T₃ reach maximum levels approximately 24 h after birth, while plasma thyroglobulin level peaks about 3 days after birth. In the first week after birth, plasma reverse T₃ concentration decreases rapidly, caused by the loss of placental and hepatic type deiodinase activity (T₄ to reverse T₃ conversion) (9).

The functional maturation of the thyroid in preterm infants at birth is incomplete. Timing of the TSH surge is similar to that of term neonates, but quantitatively lower, especially in preterm infants with respiratory distress syndrome. During the first day following the TSH surge, increasing plasma T₄ and T₃ concentrations can indeed be observed, but the T₄ and T₃ peak levels are lower as the pregnancy is shorter and in the case of complications, such as intrauterine growth retardation and respiratory distress syndrome, a nadir is observed at about 1 week after birth, followed by a second TSH increase. On average the plasma thyroglobulin concentrations in preterm infants are higher than in term infants and are highest in preterm infants with respiratory distress syndrome, in spite of the lower TSH surge (11). Although premature neonates temporarily have lower postpartum free T₄ levels than would be normal for intrauterine life at the same age, administration of T₄ immediately after birth has no significant influence on mortality and morbidity, except for extremely preterm infants (less than 27 weeks gestation) in whom such a bridging T₄ supplement may be beneficial for brain development (12).

Maternal–fetal T₄ transfer has been described from the second month of pregnancy. Initially the transfer takes place via the coelomic cavity and yolk sac. After approximately 8–10 weeks gestation nuclear T₃ receptors are detectable in the embryonic tissues. Thereafter, the T₃-receptor concentrations increase strongly (13). In children who are unable to produce any thyroid hormone by themselves, T₄ concentrations in term cord plasma are 30–70 nmol/l, which is 25–50% of the normal cord plasma concentrations (2). This can only be of maternal origin. These thyroid hormone concentrations appear to be high enough to prevent cerebral damage (almost) completely (6, 14).

Although the maternal contribution to the fetal thyroid hormone provision is indispensable, a free placental transfer of T₄ and T₃ may have disadvantages. As far as can be deduced from the course of the fetal plasma (free) T₄ concentration during the first trimester, and of the (free) T₃ concentration throughout the whole gestation, a partial barrier to T₄ and T₃ between the maternal and fetal circulations is maintained.

Newborns of women with untreated hyperthyroidism during pregnancy have been found to have inappropriately low free T₄ concentrations during the first weeks to months of life, without a concomitant increase in the secretion of TSH (fulfilling the criteria of central hypothyroidism) (15). Since this phenomenon is not reported to occur in the offspring of treated euthyroid pregnant women with Graves’ disease, it is less likely that maternal antibodies are primarily responsible. Apparently, the severely hyperthyroxinaemic environment of the fetus, during at least the third trimester, may override the placental barrier and inhibit the maturation of thyrotropic cells in the fetal pituitary, or alter the set point for thyroid hormone homeostasis.

The clinically detectable consequences of congenital hypothyroidism are mainly dependent on the severity and duration of the hypothyroid state. Furthermore, the variability in expression between individuals is considerable. At early ages the external signs are only recognizable in cases of severe congenital hypothyroidism (Box 3.4.7.1); milder types may remain undetected for years. Questioning of the parents of neonates with congenital hypothyroidism, detected by screening, showed that subtle signs of hypothyroidism had been observed in the first weeks of life (16).

In only a minority of cases with thyroid dyshormonogenesis is the neonate’s thyroid clearly visible or palpable. There is no clearly observable correlation between severity of the defect and neonatal goiter size. Goitrogenesis rarely leads to airway obstruction. Depending on the aetiology of the congenital hypothyroidism, there may be other subtle signs and symptoms (Box 3.4.7.1) (7, 17).

Transient congenital hypothyroidism, usually of short duration and often accompanied by other paediatric problems, usually escapes clinical detection. In such cases the hypothyroid state forms a complicating factor and will be an extra threat to the sick newborn. Data from the (maternal) medical history dealing with, for instance, maternal thyroid disease, use of thyroid-influencing medication, iodine-containing radiographic contrast agents, and disinfectants should draw attention to the neonate’s thyroid function.

Starting administration of T₄ to congenital hypothyroidism patients shortly after birth will prevent (postnatal) cerebral damage. Unfortunately, congenital hypothyroidism in neonates is difficult to recognize. In 1974 it became possible, on a large scale, to determine T₄ and TSH in just a few drops of blood, obtained by a heel puncture, and absorbed in filter paper. Since then many countries have introduced neonatal mass-screening procedures.

While most European countries have chosen to determine TSH, the Netherlands opted for the North American method of screening based on determination of T₄. Later, the procedure was modified to reduce the number of false-positives: TSH is determined in the 20% of samples with the lowest T₄ concentrations, and thyroxine-binding globulin in the samples with the 5% lowest T₄ levels from which T₄/thyroxine-binding globulin can be calculated. The long-term results of the Dutch screening method are that probably all cases with permanent primary congenital hypothyroidism are diagnosed at an early stage (incidence in Dutch newborns between 1 April 2002 and 31 May 2004 is 1:2 400) and probably more than 90% of cases with permanent secondary/tertiary congenital hypothyroidism (incidence 1:21 000 to 1:16  400) (Table 3.4.7.1) (18–20).

Estimates from a number of international screening reports in areas without endemic iodine deficiency give the mean incidence of permanent primary congenital hypothyroidism as roughly 1:3500 newborns, with considerable ethnic differences (extremes are 1:30 000 among African-Americans in the USA and 1:900 among Asian groups in the UK).

A clear diagnosis is required to decide upon the optimal treatment and to evaluate the risk of other (endocrine) defects or complications, the risk of recurrence in the family, and the possibilities of prenatal diagnosis and treatment (21). It may also be possible to judge the longer term consequences of congenital hypothyroidism for the patient, especially the risk of a delay in cognitive and motor development.

A clinicopathological approach is the main method used, so that diagnosis is as efficient as possible. The starting point is to produce an aetiological description (‘clinicopathological entity’), that is as detailed as possible, for every case of congenital hypothyroidism (22). The gene structures and coding sequences of several proteins involved in T₄ synthesis have been explained in recent years. Nevertheless, cDNA containing a novel mutation usually has to be expressed and its function tested before the mutation can be established as the primary cause. At present it is possible to establish this in only a minority of patients with congenital hypothyroidism. This implies that the ‘classic’ diagnostic methods, such as plasma TSH, free T₄, thyroglobulin, and thyroid autoantibody determination, ultrasound imaging, radio-iodide uptake with a perchlorate test, measurement of the urinary excretion of iodine and iodotyrosines, radio-iodide saliva/blood ratio, and the mode of inheritance will still be needed in the foreseeable future. For the list of known clinicopathological entities, we have developed a set of diagnostic profiles, each representing the combined data of this series of determinants (Table 3.4.7.2) (22). By combining the measurements for series of determinants, each of which alone yields little specific data, the most likely aetiology can be established.

The actual stimulatory activity of TSH is of great importance to the determinants representing the thyroid’s action. For instance, generally the plasma thyroglobulin concentration and the radio-iodide uptake are related to plasma TSH concentration, TSH bioactivity, TSH-receptor responsiveness, and the amount of thyroid tissue present, but in the case of a thyroglobulin synthesis defect the plasma thyroglobulin concentration is unusually low. Ultrasound imaging is a useful, fast, and noninvasive diagnostic technique for localizing the thyroid and measuring its volume, but it does not detect small remnants. These, however, are easily visualized with ¹²³I (23). The most sensitive determinant for detecting traces of thyroid tissue is the plasma thyroglobulin concentration (24).

Congenital hypothyroidism resulting from thyroid disorders are due to two main causes: disturbances in the thyroid’s ontogeny, making up the major portion of defects, and inborn errors in the thyroid’s hormonogenesis (extensively reviewed elsewhere) (22).

Congenital hypothyroidism caused by disturbances in the development of the thyroid gland may result in mild to very severe hypothyroidism. The thyroid gland may be completely absent (agenesis) or remnants of variable size may be present along the tract of the thyroglossal duct (Fig. 3.4.7.2). These structures, called dystopic (synonym: ectopic) remnants, are often localized in the sublingual area. Agenesis is characterized by complete absence of any thyroid tissue (indicated by ¹²³I and ultrasound imaging), and complete inability to produce thyroid hormone and thyroglobulin (2, 22). However, patients with a negative ¹²³I scintigram and (almost) complete absence of circulating thyroid hormone, but with clearly measurable plasma thyroglobulin levels, have been described (24). As the thyrocytes are the only cells able to produce thyroglobulin, this cell type has to be present, although it cannot be localized. We introduce the term ‘cryptic thyroid remnant’ to describe this type of disorder.

Why the migration and development of the thyroid becomes disturbed is still unexplained. Studies in mice showed the involvement of the four transcription factors TITF1/NKX2-1, PAX8, FOXE1 (formerly called TTF2), and NKX2-5 (7, 25). Mice missing the TTF1/NKX2-1 gene were stillborn, lacked thyroid, pituitary, and lung parenchyma, and had extensive defects in brain development; the heterozygous animals were phenotypically normal. Mice lacking PAX8 only had a rudimentary thyroid gland, almost completely composed of calcitonin-producing C cells. Mice missing the FOXE1 gene had dystopic thyroid tissue and cleft palate, and their pituitary responded normally to the decreased plasma free T₄ levels, whereas with TITF1/NKX2-1, heterozygous mice showed normal thyroid function. Mouse embryos missing the NKX2-5 gene appeared to have a smaller thyroid bud.

In contrast to the findings in knockout mice, in patients with thyroid dysgenesis only monoallelic inactivating mutations in PAX8 have been found. In humans homozygous missense mutations in the forkhead domain of FOXE1 were shown to be associated with congenital hypothyroidism, cleft palate, and choanal atresia (7). Missense mutations in NKX2-5 have been found in three patients with dystopic thyroid remnants (thyroid ectopy) and in one patient with thyroid agenesis (25).

Although there are strong indications that transcription factors encoded by TITF1/NKX2-1, PAX8, FOXE1, NKX2-5, and the TSH receptor play a role in the ontogeny of the human thyroid, only a small minority of the patients with thyroid dysgenesis mutations in these transcription factors has been found (7). This accords with the observations worldwide that familial occurrence of thyroid dysgenesis is rare. It is puzzling, too, why a dysgenic thyroid remnant hardly develops after the embryonic phase, while its hormone production seems to be adequate for the amount of tissue, especially in view of the impressive growth capacity of normally developed thyroids under similar TSH stimulation. Because the more caudally located remnants are usually the larger ones, it is likely that common factors are responsible for both the impaired growth potential, the insufficient ‘descendance’, and the absence of bifurcation into two lobes.

Inborn errors can occur in all regulatory and metabolic steps involved in the synthesis of thyroid hormone.

This refers to defects in the various components of the TSH stimulation pathway. In general, a defect in TSH action is characterized by the presence of a eutopic, often somewhat undersized, thyroid gland, low to very low plasma free T₄ and thyroglobulin concentrations (especially when related to the (very) high TSH concentration), and low thyroidal radio-iodide uptake with a slow iodine turnover. Several loss-of-function mutations in the TSH-receptor gene have been described, that produce congenital hypothyroidism with strongly variable expression, ranging from subclinical to overt hypothyroidism (26, 27). A related type of TSH hyporesponsiveness has been described in patients with pseudohypoparathyroidism type 1A (Albright’s hereditary osteodystrophy). The cause of this autosomal dominant inherited disease is a mutation in the GNAS gene, coding for the α-subunit of the Gs-protein (28). Some of the patients become hypothyroid; only a minority is detected by the neonatal congenital hypothyroidism screening. In some patients with the clinicopathological characteristics of TSH hyporesponsiveness, inherited in an autosomal dominant way, no mutations in TSH-receptor or GNAS genes could be found, suggesting the presence of mutations in more distal components of the TSH signalling pathway.

The first step in thyroid hormonogenesis is the active transport of iodide into the thyrocytes. Iodide transport across the basal membrane is mediated by the sodium-iodide symporter (NIS). Currently, at least 12 iodide transport defect-causing mutations of the NIS gene have been identified. Patients are hypothyroid from birth, show gradual goitrogenesis, low or very low plasma T₄ concentrations, high plasma thyroglobulin concentrations, undetectable thyroidal radio-iodide uptake, and a radio-iodide saliva/blood ratio of about unity. Heredity is autosomal recessive (17). Both the severity of hypothyroidism and the neurodevelopmental impairment vary considerably, probably due to variations in dietary iodine intake. Treatment with large doses of iodine is possible, but therapy with T₄ is preferred, especially in young children.

Oxidation of trapped iodide and binding to tyrosine residues in proteins, particularly thyroglobulin, is commonly referred to as iodide organification. Both steps take place at the apical brush border of the thyrocyte, mainly in the follicular lumen. Oxidative coupling of iodothyronine residues requires a proper thyroglobulin structure, normal peroxidase activity, and a regulated presence of hydrogen peroxide (H₂O₂). H₂O₂ is generated by (thyroid) dual oxidase 2. Defects in any of these compounds will impair thyroid hormonogenesis, resulting in primary congenital hypothyroidism. When left untreated goitrogenesis will occur.

In patients with iodination defects, the T₄ synthesis is decreased or absent depending on whether the trapped iodide is only partially or not at all organified. As a consequence of the low plasma thyroid hormone concentrations, the TSH level is enhanced, resulting in a high (radio)iodide uptake and an elevated plasma thyroglobulin concentration. The delayed or absent iodide oxidation and organification causes a high intracellular (inorganic) iodide content, which is rapidly released after the administration of sodium perchlorate (Fig. 3.4.7.3) (see notes to Table 3.4.7.2) (29).

In (almost) all cases, total iodide organification defects are caused by mutations in the gene coding for thyroperoxidase. The defect is transmitted in an autosomally recessive way. Inactivation of thyroperoxidase is caused by several types of mutations, such as deletions, insertions, missense and nonsense mutations, and splicing defects. The most frequent mutation is a duplication of a GGCC sequence in exon 8 (30). Partial organification defects are not only caused by (heterozygous) mutations in the thyroperoxidase gene, but also by mutations in the genes encoding dual oxidase 2 and dual oxidase maturation factor 2. Biallelic as well as heterozygous mutations in the DUOX2 gene appear to result in transient congenital hypothyroidism (31, 32) The single patient with a homozygous mutation in the DUOXA2 gene had relatively mild but permanent congenital hypothyroidism (33).

A remarkable subtype of partial organification defect is Pendred’s syndrome, with an estimated prevalence of 1 in 40 000. Pendrin, encoded by the SLC26A4 gene, is a highly hydrophobic membrane protein located at the apical membrane of thyrocytes, where it could function as an iodide transporter. In the inner ear, pendrin is important for generation of the endocochlear potential. Currently, more than 150 mutations of the SLC26A4 gene have been reported (17). Most patients have a moderate to severe sensorineural hearing loss from infancy. Hypothyroidism (usually mild) and goiter may be present at birth or may develop later in life. Only a few patients with Pendred’s syndrome are detected by neonatal screening.

Thyroglobulin plays a central role in thyroid hormone synthesis. Thyroglobulin synthesis occurs exclusively in the thyrocyte. The protein is very large and is encoded by a thyroglobulin mRNA containing 8307 nucleotides. Thyroglobulin is a homodimer with subunits of 330 000 Da, each containing 60 disulfide bridges and 10% carbohydrates (34).

For maximal production of iodothyronines, mainly T₄, an optimal stereospecific configuration of thyroglobulin is required. This configuration is dependent on the primary structure, disulfide bridges, the extent of glycosylation, and possibly other processes such as phosphorylation. A distorted configuration will result in an impaired formation of iodothyronines. Patients classified under the entity ‘thyroglobulin synthesis defects’ are moderately to severely hypothyroid. In relation to the TSH concentration, the plasma thyroglobulin concentration is usually low, but there are exceptions. The processes of iodide uptake, oxidation, and organification are intact. A clinicopathological evaluation, however, cannot distinguish whether disorders in the synthesis of thyroglobulin are caused by defects in transcription, translation, or post-translational processes and transport. Therefore this entity comprises all these types of defects.

The exceptional size of the gene coding for thyroglobulin makes it difficult to identify mutations in the coding regions. In four human families and three animal strains, various mutations have been described: deletions, nonsense and missense mutations, and acceptor splice-site mutations that cause alternative splicing. The mutations are all homozygous in character; the inheritance is autosomal recessive (34).

Thyroglobulin, internalized by endocytosis from the follicular lumen into the thyrocyte, is incorporated into early and late endosomes. These organelles, containing proteolytic enzymes, hydrolyse thyroglobulin to its constituent amino acids, including the iodotyrosines monoiodotyrosine and diiodotyrosine, as well as T₄ and T₃. Subsequently, the iodotyrosines are deiodinated by specific deiodinase(s) in the thyroid and other tissues.

Iodotyrosine deiodinase defects, hereditary disorders in this deiodinating system, lead to loss of the iodotyrosines from the thyroid, and rapid excretion by the kidneys. The excessive loss of iodine results in postnatal hypothyroidism and mimics hypothyroidism due to iodine deficiency. Only recently, the first homozygous missense mutations and deletion in the gene encoding iodotyrosine deiodinase (DEHAL1) were described in four patients who presented with severe goitrous hypothyroidism diagnosed in infancy and childhood. The two patients who underwent neonatal screening were not detected, and one of these two patients was found to be mentally impaired. This implies that infants with DEHAL1 defects may have normal thyroid function at birth and they may be missed by neonatal screening programmes for congenital hypothyroidism (35). Elevated urine di- and monoiodotyrosine concentrations are suggestive of the diagnosis (36). Although the heredity is autosomal recessive, one recently described heterozygous carrier of an inactivating mutation presented with overt hypothyroidism suggesting dominant inheritance with incomplete penetration (36).

While a clear distinction can be made between thyroid dysgenesis and dyshormonogenesis in the case of congenital hypothyroidism of thyroidal origin, such a distinction is less straightforward in secondary/tertiary congenital hypothyroidism. Moreover, it is difficult to discriminate, using clinicopathological criteria, between secondary (synonym: pituitary) and tertiary (synonym: hypothalamic) disorders. For that reason the entity ‘congenital hypothyroidism of central origin’ is introduced, a term that does not exclude simultaneous occurrence of other pituitary hormone deficiencies.

Most cases of congenital hypothyroidism of central origin concern developmental disturbances of the pituitary and/or hypothalamus and are easy to visualize with MRI (Fig. 3.4.7.4). In these cases the endocrine problem is not restricted to the thyrotropic axis. As in congenital hypothyroidism of thyroidal origin, there are sporadic and hereditary types of central congenital hypothyroidism, but the developmental problems that cause central congenital hypothyroidism are not just ‘sporadic’. In a recent series of patients with central congenital hypothyroidism detected through the Dutch neonatal screening programme, 53% of these patients had a so-called posterior pituitary ectopia. All of these patients had multiple pituitary hormone deficiencies, with cortisol deficiency as the most (life-)threatening problem (37). Posterior pituitary ectopia may be accompanied by other (minor) malformations, often of other cerebral structures. The underlying cause is unknown. Currently, only a small percentage of the cases of multiple pituitary hormone deficiency can be explained by mutations in genes encoding transcription factors involved in hypothalamus and pituitary development (Table 3.4.7.2) (38, 39).

Inborn errors may occur in all regulatory and metabolic steps involved in the synthesis of TSH. These may be located in the hypothalamus or pituitary gland.

Knowledge about the mechanism of thyrotropin-releasing hormone (TRH) production in the hypothalamus is limited. The tripeptide TRH is also present elsewhere in the central nervous system, indicating that neither synthesis nor action is restricted to the thyrotropic axis. Isolated TRH deficiency in humans has not yet been reported.

Diminished TRH responsiveness is only a partially explained entity. Based on the analogy of TSH hyporesponsiveness, it might be assumed that mutations in TRH receptor and defects in postreceptor processes (G-protein deficiency, etc.) may result in secondary congenital hypothyroidism. Homozygous mutations in the TRH receptor gene are found which result in the formation of receptors that are unable to bind TRH. The patients are mildly hypothyroid with complete absence of TSH and prolactin responses to TRH (40).

In another group of patients with central hypothyroidism, missense and nonsense mutations, and deletions in the genes encoding the transcription factors POU1F1 and PROP1, have been described. POU1F1 regulates the expression of the TSHb, growth hormone, and prolactin genes. Mutation of the POU1F1 gene results in combined pituitary hormone deficiencies, including complete growth hormone and prolactin deficiency as well as central congenital hypothyroidism. Most cases show autosomal recessive inheritance, but some show an autosomal dominant inheritance pattern. PROP1 is involved in the early development of several lineages of anterior pituitary cells. Mutations in the PROP1 gene cause multiple pituitary hormone deficiency that is autosomal recessive and is, in addition to central congenital hypothyroidism, associated with deficiency of luteinizing hormone/follicle-stimulating hormone, growth hormone, prolactin, and, less frequently, adrenocorticotropic hormone (39). Patients with hereditary defects in the TSHb gene are rare. The hypothyroidism may be severe, and plasma TSH may vary from undetectable to slightly raised. In the latter case the circulating TSH is biologically inactive (39).

Congenital hypothyroidism or hypothyroidism originating during the first weeks to months of life can also result from increased inactivation (infantile haemangioma expressing type 3 deiodinase) or loss of T₄ (congenital nephrotic syndrome) (41, 42). Affected patients may need a rather high T₄ dose to correct the hypothyroidism. Since the cause of the hypothyroidism is not in the thyroid gland, pituitary, or hypothalamus, the entity ‘congenital hypothyroidism of peripheral origin’ is introduced. Other forms of congenital hypothyroidism of peripheral origin are resistance to thyroid hormone (due to thyroid hormone receptor β (TRβ) gene mutations), and the recently discovered thyroid hormone transporter and thyroid hormone metabolism defects (due to monocarboxylate transporter 8 (MCT8) gene and to ‘selenocysteine insertion sequence-binding protein 2’ (SBP2) gene mutations, respectively). In all of these conditions, of which the clinical and biochemical features are extensively reviewed in Chapter 3.4.8, there is reduced sensitivity to thyroid hormone (43). T₄ treatment is probably not beneficial in these conditions.

Transient primary congenital hypothyroidism is often due to exposure of the neonate (or the fetus) to excessive quantities of iodine, e.g. iodine-containing radiographic contrast agents and disinfectants. These agents are mostly used in premature or very ill infants. Detection by the neonatal congenital hypothyroidism screening depends on the timing of the exposure. There are no data available from systematic psychological or neurological investigations to estimate the risk of brain damage in children who are perinatally exposed to iodine excess. Yet, prevention of this type of thyroid dyshormonogenesis is indicated, preferably by avoiding unnecessary use of excessive quantities of iodine, or by timely administration of T₄.

Incidentally, maternal thyroid-inhibiting antibodies may cause transient congenital hypothyroidism for several weeks or months, depending on the initial concentration of circulating antibodies (44). Rarely the use of antithyroid drugs by the mother leads to abnormal congenital hypothyroidism screening results.

Transient congenital hypothyroidism of central origin has been reported in a number of newborns of women with untreated Graves’ hyperthyroidism during pregnancy. It may be caused by exposure of the fetal hypothalamic–pituitary–thyroid system to higher than normal thyroid hormone concentrations, impairing its physiological maturation during intrauterine life (45).

The main treatment goal in congenital hypothyroidism is prevention of cerebral damage due to lack of thyroid hormone. This implies that the period with decreased levels of circulating free T₄ must be kept as short as possible by administering a T₄ dose that restores euthyroidism as soon as possible. Nowadays, after an abnormal congenital hypothyroidism screening result it should be feasible to start treatment before the age of 14 days. The short-term goals of T₄ treatment in the neonatal period are to normalize the plasma (free) T₄ and TSH concentrations within 2 and 4 weeks, respectively. Higher T₄ starting doses (i.e. more than 10 μg/kg per day) result in more rapid normalization of the plasma hormone concentrations than lower doses (46).

Long-term effect evaluation of the cognitive and motor development of patients with congenital hypothyroidism has demonstrated that timely and adequate T₄ treatment results in psychological test scores within the normal range for most children (14, 47, 48). However, even with a treatment start before the age of 14 days, patients with congenital hypothyroidism have an approximately 0.5 SD deficit in their (full-scale) IQ, with the difference being somewhat greater in patients with severe congenital hypothyroidism (e.g. caused by thyroid agenesis) (14, 49).

Over recent years it has been suggested that a high T₄ starting dose may further improve the developmental outcome, especially in patients with severe congenital hypothyroidism (50). However, in a recent Cochrane review addressing this issue it was concluded that there is inadequate evidence to suggest that a high dose is more beneficial compared to a low dose for initial thyroid hormone replacement in the treatment of congenital hypothyroidism (51). Furthermore, a relatively high T₄ starting dose (more than 10 μg/kg per day) has been associated with behavioural problems later in life (52).

With this in mind, our recommendations are:

Usually an initial T₄ dose of about 10–12 μg/kg once a day will do. In the case of a severely hypothyroid neonate, the body’s T₄ deficit can be corrected by one additional T₄ dose, 12 h after the initial dose. The supplementary dose of T₄ is mainly dependent on body mass, age, and intestinal resorption, and less so on aetiology or severity of the disease. In milder forms of congenital hypothyroidism the residual thyroid function is almost completely suppressed under treatment. Whereas adults in general are adequately supplied with a daily T₄ dose of about 1.6 μg/kg, neonates require about five to six times this dose. However, a large interindividual variability exists, which cannot be predicted when treatment starts.

In general, when treated with T₄, moderately increased free T₄ levels are necessary to suppress plasma TSH levels to values within the normal range. It may sometimes take up to 1 month to return the usually extremely high initial TSH values to normal. To prevent under- or overtreatment, the plasma free T₄ concentration should be measured weekly for the first 4 weeks. Further controls can be done once every 2 weeks, later monthly, and after the age of 6 months the frequency of controls can be gradually lowered to once every 3 months between the ages of 1 and 3 years. Because a healthy thyroid produces, besides the prohormone T₄, substantial amounts of bioactive T₃, it is debatable whether the optimal preparation would be a mixture of T₄ and T₃. However, there is no evidence of benefit of adding T₃ to the T₄ treatment.

In cases of congenital hypothyroidism of central origin plasma TSH levels are useless for therapy control, and one has to rely on free T₄ concentrations. Usually the need for T₄ per kilogram of body mass is somewhat lower than in cases of congenital hypothyroidism of thyroidal origin. Obviously, a normally developed thyroid, even in the absence of TSH stimulation, is able to produce some T₄. As a rule of thumb, a T₄ starting dose of approximately 6–8 μg/kg per day is suitable. In cases of accompanying ACTH deficiency, it is important to start cortisol supplements as soon as possible, preferably before beginning T₄ therapy. Under normal conditions, young infants need 10–12 mg/m² per day of cortisol orally in three or four divided doses. The cortisol dose must be increased immediately in case of stress (illness, pain, etc.).

Treatment of transient congenital hypothyroidism is rarely necessary because the hypothyroid phase is usually short. Finally, even in doubtful cases, adequate treatment is mandatory and should not be delayed, regardless of whether the aetiology is known.

The prevailing view is that congenital hyperthyroidism is usually caused by thyroid-stimulating antibodies of maternal origin crossing the placenta from early in gestation and stimulating the (fetal) thyroid from midgestation (21). Remarkably, only a small percentage (estimated at 2%) of the children of mothers with Graves’ disease develop congenital or neonatal hyperthyroidism, indicating that what appears to stimulate the maternal thyroid does not automatically stimulate the fetal or neonatal thyroid. Yet, in the case of very high maternal levels of thyroid-stimulating antibodies during pregnancy, the child becomes hyperthyroid. Onset and severity not only depend on the level of stimulating antibodies, but also on the presence of blocking antibodies and antithyroid drugs that may mitigate, postpone, or even overrun the excessive thyroid hormone production by the fetal/neonatal gland. The suppressive effect of the antithyroid drug taken by the mother stops within a day after birth; blocking antibodies have a longer plasma half-life and may counteract the stimulating antibodies for several weeks. Breastfeeding does not influence the child’s condition significantly (53). In summary, it is difficult to predict which child is really at risk, whereas the consequences are important.

Fetal and neonatal hyperthyroidism are severe life-threatening conditions. The prenatal signs may be intrauterine growth retardation, microcephaly, goitrogenesis, tachycardia, and premature birth. After birth the infant may be extremely restless, irritable, with an exophthalmus-like appearance, and signs of hypermetabolism and multiple organ failure. The infant may die if treatment is not instituted immediately.

If a neonate has clinical and/or clinicochemical manifestations of hyperthyroidism, the child’s thyroid must be inhibited as soon as possible. Postnatal treatment consists of methimazole (0.5–1.0 mg/kg per day, orally in two or three divided doses) and, depending on the severity of the condition, propranolol (1–2 mg/kg per day, orally in three or four divided doses), iodide (1 drop of Lugol’s solution every 8 h after the start of antithyroid drug therapy; Lugol’s solution contains 126 mg iodine/ml), and, if necessary, corticosteroids. If heart failure is imminent, digitalization is indicated (54). After the critical condition is stabilized and the euthyroid state reached, only the antithyroid drug therapy should be continued for several months, and T₄ must be added to prevent hypothyroidism (‘block and replace’). Most infants remit by 3–4 months of age. If the fetus is found to have hyperthyroidism, the mother should be treated with an antithyroid drug, preferably propylthiouracil, while keeping her euthyroid by T₄ administration.

In cases of congenital hyperthyroidism due to a gain-of-function mutation of the TSH-receptor gene, remission will not occur. On the contrary, due to the ongoing growth of the gland the inhibiting action of the antithyroid drug tends to become less effective over time, and the only reliable long-term treatment is removal of the whole gland. Mental impairment, microcephaly, and growth problems may occur when treatment is delayed.

The prevalence of autoimmune thyroid disease(s) in children is low. Since the great majority of the paediatric patients are (post)pubertal at diagnosis there is little or no risk that brain growth and development are threatened by hypo- or hyperthyroidism caused by this disease. As in adults, autoimmune thyroid disease in children occurs predominantly in girls. Apart from consequences for growth and pubertal development, all features of autoimmune thyroid disease in childhood are similar to those in adults.

In young children the presence of thyroid autoantibodies without thyroid dysfunction or goiter is rare. In older children and adolescents the prevalence of detectable serum autoantibodies may be as high as 480 in 10 000, approximately one-third of the prevalence in adults (55).

By far the most common cause of acquired hypothyroidism in children is chronic lymphocytic thyroiditis due to autoimmune disease (Hashimoto’s disease). The incidence in children is much lower than in adults, but gradually increases with age. Recently reported incidences from Denmark are 0.08 in 10 000 in 0- to 9-year-olds and 0.4 in 10 000 in 9- to 19-year-old children and teenagers (56). Autoimmune hypothyroidism is 4–7 times more frequent in girls than in boys. The severity of the hypothyroidism varies from the inability to produce any thyroid hormone (atrophic thyroiditis) to subclinical hypothyroidism (with or without palpable goiter). As long as there is sufficient functioning, thyroid tissue remission may occur (57). It is even possible that after long-standing hypothyroidism the patient becomes euthyroid.

The most common clinical manifestations of acquired hypothyroidism in children are growth retardation and pubertal delay, accompanied by goiter. Especially in milder cases there is a poor correlation between clinical expression and plasma thyroid hormone levels. Although very rare, in infants acquired autoimmune hypothyroidism may manifest itself as a progressive delay in development (58).

The risk of developing autoimmune hypothyroidism is increased in children and adolescents with Down’s syndrome. Published prevalences vary between 0 and 660 in 10 000 (59). This high prevalence is ascribed to a greater tendency to develop autoimmunity and might be caused by overexpression of one or more chromosome 21 genes directly or indirectly influencing the immune system (60). Because the symptoms and signs of overt hypothyroidism are not always easy to recognize in Down’s syndrome, and acquired hypothyroidism occurs more frequently in infancy compared with non-Down’s syndrome children, the Committee on Genetics of the American Academy of Pediatrics recently recommended thyroid function screening at the age of 6 months, again at age 12 months, and then annually during childhood (61). In adults with Down’s syndrome, yearly to 2-yearly testing is recommended. Several other syndromes and disorders (e.g. Klinefelter’s syndrome, Turner’s syndrome, type 1 diabetes mellitus, type 1 autoimmune polyendocrinopathy, juvenile idiopathic arthritis, and coeliac disease) are also associated with a substantially higher risk of developing chronic autoimmune thyroiditis.

The only (necessary) treatment is administration of T₄. Any goiter usually shrinks somewhat (the result of decreased TSH stimulation) but often does not disappear (indicating that inflammation persists). Because autoimmune hypothyroidism may be self-limited, periodic re-evaluation of the thyroid’s hormone-producing capacity is necessary. Untreated, euthyroid patients with autoimmune goiter should have periodic measurement of their free T₄ and TSH concentrations and ultrasound imaging of the thyroid. If nodules develop, these should be examined cytologically to exclude malignant degeneration. Very rarely Hashimoto’s goiter may be painful. Prednisolone treatment is often successful in controlling this symptom, but thyroidectomy may be necessary in some patients, e.g. because of unacceptable steroid side effects.

Acquired juvenile hyperthyroidism (synonym: thyrotoxicosis) is, with very few exceptions, due to Graves’ disease. It is a rare disease with incidence figures between 0.079 and 0.65 in 10 000 children/year (62, 63). It affects mainly girls, although this predominance is less than in adult women (5:1 vs 10:1). The prevalence of Graves’ disease in children with Down’s syndrome is higher than in the general population (59).

While it is assumed that all patients with Graves’ disease have thyroid-stimulating antibodies, they are not always detected by the available assays. The aetiological diagnosis is then made on clinical grounds, usually supplemented by tests for other antibodies. In the great majority of affected children a small firm symmetrical goiter is present. Ophthalmopathy is mostly absent or very mild.

Full-blown Graves’ disease in children is easily recognized by the abundance of signs and symptoms, but sometimes the disease develops insidiously. One has to keep in mind that the clinical expression is extremely variable, and that there are patients with clearly elevated free T₄ levels and plasma TSH concentrations below the detection limit (usually less than 0.01 mU/l), who have hardly any complaint. The various clinical manifestations (see Chapter 3.3.1) are mostly aspecific and quite common in childhood and adolescence. For example, signs of hyperthyroidism during puberty can easily be interpreted as pubertal behavioural problems. Sometimes the clinical condition even resembles hypothyroidism. In such cases the plasma TSH level discriminates from thyroid hormone hyporesponsiveness.

Differential diagnosis in juvenile hyperthyroidism is very similar to that in adults (see Chapter 3.3.5). It is usually simple to demonstrate that Graves’ disease is the cause. However, especially in young children it may not be easy to conclude that the disorder is an acquired one, because congenital disorders such as hereditary hyperthyroidism (due to activating TSH-receptor gene mutations) and McCune–Albright syndrome may become clinically manifest several years after birth.

Although the spectrum of therapeutic possibilities for juvenile hyperthyroidism is the same as in adults, the choice may be different and depend on the patient’s age. Because of the possibility of long-term remissions, most paediatric endocrinologists recommend antithyroid drugs as the initial treatment rather than radio-iodide destruction or subtotal thyroidectomy. Given the risk of propylthiouracil-related acute liver failure (possibly 1 in 2000 in children), methimazole is the drug of first choice (64). Since it is difficult to realize permanent euthyroidism by titrating the drug on plasma free T₄ and/or TSH levels, we prefer to administer the combination of a suppressive dose of methimazole (0.5 mg/kg per day, in two divided doses) and a dose of T₄ that is adjusted primarily on plasma free T₄ concentrations. TSH secretion may remain suppressed for several months, making it an unsuitable determinant for the control of initial treatment. Unfortunately, the chance of permanent remission in children after antithyroid drug treatment is probably not much higher than 20–30%, even after treatment for longer than 1–2 years (65). When there is need for a more definitive form of treatment, radio-iodide destruction seems a safe alternative to subtotal thyroidectomy (65, 66).

Differentiated thyroid carcinoma is not uncommon in children and adolescents. It accounts for 1% of paediatric cancer cases in prepubertal children, and 7% in adolescents aged 15–19 years old (67). The overall incidence of thyroid carcinoma in children is approximately 17.5 in 10 000 with a higher incidence in girls than in boys and an approximately 5 times higher incidence in 15- to 19-year-old teenagers than in 0- to 14-year-old children (67). Differentiated thyroid carcinoma shows mostly a papillary histological pattern and usually has a good prognosis despite the clinical characteristic of rather aggressive behaviour. It is important to realize that a child with a single thyroid nodule has a 20% chance of having thyroid cancer, which is higher than in adults. Because the long-term survival of children with malignant diseases has improved impressively since the 1980s, differentiated thyroid carcinoma is becoming a rather common (second) malignancy after external radiation of the neck (malignant tumours in the cervical region, before bone marrow transplantation).

Malignant thyroid tumours tend to develop more rapidly in young children than in adolescents and adults. This has been shown clearly in the populations exposed to the radio-iodine in the fall-out after the Chernobyl disaster (see Chapter 3.2.5). Also, thyroid cancer recurrences usually occur earlier in young children. Nevertheless, long-term monitoring is required because recurrence may even arise several decades after the initial diagnosis and treatment.

In general, patients are effectively treated by surgery, (often) followed by radio-iodine therapy and suppression of TSH secretion. The surgical treatment in children is similar to that in adults (see Chapter 3.5.6). Treatment with T₄ aiming at TSH suppression, however, should not interfere with the child’s growth and pubertal development. Furthermore, it is not always possible to suppress the plasma TSH concentration permanently below the detection limit (less than 0.01 mU/l), because children continuously ‘grow out of their T₄ dosage’ unless an overdose is given. A recent T₄ treatment scheme that gives in to these ‘objections’ initially suppresses TSH levels to less than 0.1 mU/l and then allows the TSH concentration to rise to 0.5 mU/l once the child enters remission (67). The most sensitive indicators of thyroid cancer recurrence are TSH-stimulated radio-iodide whole body scanning and measurement of the plasma thyroglobulin concentration, after withdrawal of T₄ therapy or administration of recombinant human TSH.

Medullary thyroid carcinoma is an uncommon but highly malignant disease (see Chapter 3.5.7). It is, however, relevant to paediatric endocrinologists because in about 20% of cases it is part of the autosomal dominant inherited multiple endocrine neoplasia type 2 (MEN 2) syndromes: MEN 2A with medullary thyroid carcinoma, phaeochromocytoma, and hyperparathyroidism; MEN 2B with medullary thyroid carcinoma, phaeochromocytoma, multiple mucosal neuromas, and a marfanoid body habitus; and familial medullary thyroid carcinoma (FMTC) without other endocrine or neural abnormalities. In MEN 2B the medullary thyroid carcinoma can sometimes be so aggressive that widespread metastases have already occurred in childhood. All three of these syndromes stem from characteristic mutations in the RET proto-oncogene on chromosome 10, and it has become clear that most mutation carriers will develop disease sooner or later. It has also become clear that some mutations result in earlier development of medullary thyroid carcinoma than others (68). The prognosis of medullary thyroid carcinoma is much worse than that of differentiated thyroid carcinoma, especially once metastases have developed. However, prophylactic thyroidectomy before the development of medullary thyroid carcinoma probably prevents disease (69). Therefore, it is advisable to screen the offspring of patients with these syndromes at a very young age, and to thyroidectomize the children that carry the mutant allele. Nowadays, a prophylactic thyroidectomy between the ages of 1 and 6 months is recommended in case of MEN 2B, before the age of 5 years in MEN 2A, and between the ages of 5 and 10 years in FMTC (70).