Congenital hypothyroidism (CH) is defined as thyroid hormone deficiency present at birth. Thyroid hormone deficiency at birth is most commonly caused by a problem with thyroid gland development (dysgenesis) or a disorder of thyroid hormone biosynthesis (dyshormonogenesis). These disorders result in primary hypothyroidism. Secondary or central hypothyroidism at birth results from a deficiency of thyroid stimulating hormone (TSH). Congenital TSH deficiency may rarely be an isolated problem (caused by mutations in the TSH β subunit gene), but most commonly it is associated with other pituitary hormone deficiencies, as part of congenital hypopituitarism. Peripheral hypothyroidism is a separate category resulting from defects of thyroid hormone transport, metabolism, or action.

Congenital hypothyroidism is classified into permanent and transient CH. Permanent CH refers to a persistent deficiency of thyroid hormone that requires life-long treatment. Transient CH refers to a temporary deficiency of thyroid hormone, discovered at birth, but then recovering to normal thyroid hormone production. Recovery to euthyroidism typically occurs in the first few months or years of life. Permanent CH can be further classified into permanent primary and secondary (or central) CH; transient primary CH has also been reported. In addition, some forms of CH are associated with defects in other organ systems; these are classified as syndromic hypothyroidism.

The underlying etiology of CH typically will determine whether hypothyroidism is permanent or transient, primary, secondary, or peripheral, and whether there is involvement of other organ systems (see section on Etiology for details). The primary emphasis of this review is a discussion of primary CH, but there also will be some discussion of secondary or central CH. It should be borne in mind that an underlying etiology may not be determined for many cases of CH. Further, while the exact cause of some cases of thyroid dysgenesis is known, e.g., a mutation in the TTF-2 gene, mutations in genes encoding such transcription factors important in thyroid gland development have been found in only 2% of cases. Thus, an exact cause for the vast majority of cases of thyroid dysgenesis remains unknown. This has not been a significant issue, however, as management of CH is based on restoring thyroid function to normal, not necessarily knowing the exact underlying cause.

Prior to the onset of newborn screening programs, the incidence of congenital hypothyroidism, as diagnosed after clinical manifestations, was in the range of 1;7,000 to 1:10,000 [1]. With the advent of screening of newborn populations, the incidence was initially reported to be in the range of 1:3,000 to 1:4,000 [2]. With more experience from state, regional, and national screening programs, it has become apparent that the incidence varies by geographic location. A report from the French newborn screening program summarizing a 20 year period found the incidence of permanent hypothyroidism to be 1:10,000 [3], whereas a report from the Greek Cypriot population over an 11 year period found the incidence in newborns to be 1:800 [4].

The clinical features of congenital hypothyroidism are often subtle and many newborn infants remain undiagnosed at birth [8, 9]. This is due in part to passage of maternal thyroid hormone across the placenta. This is measured in umbilical cord serum to be 25-50 percent of normal [10]. This provides a protective effect, especially to the fetal brain [11]. Also, the most common form of congenital hypothyroidism has some moderately functioning thyroid tissue [12]. The slow development of obvious clinical symptoms [13], coupled with the importance of early treatment led to the implementation of widespread newborn screening for this condition [2]. However, newborn screening for hypothyroidism is not done in many third world countries. Only an estimated 1/3 of the worldwide birth population is screened. It is therefore important that clinicians are able to recognize and treat the disorder.

Up to one third have a birth weight greater than the ninetieth percentile [8]. On initial examination, the most common signs are umbilical hernia, macroglossia and cold or mottled skin [14]. Thyroid hormone is also important in the formation and maturation of bone [15, 16].This can lead to a wide posterior fontanel of greater than 5 mm. This, along with persistent jaundice and poor feeding are the most striking clinical features [12]. A few infants with congenital hypothyroidism may have a palpable goiter. This is usually found in thyroid dyshormonogenesis where there is a defect in thyroid hormone production. Pendred syndrome mentioned below can present with deafness and a palpable goiter. Other forms of dyshormogenesis are due to defects in enzyme function within the thyroid gland and are discussed further in the section on etiology.

The typical appearance of a hypothyroid infant before the advent of newborn screening is shown in the infant in Figure 1. Features include jaundice, a puffy face and a wide posterior fontanelle with open sutures. The nasal bridge is flat and the eyes exhibit pseudohypertelorism. The mouth may be slightly open revealing macroglossia. Further examination would reveal bradycardia and a protuberant abdomen with a large umbilical hernia. Neurologic examination findings include hypotonia with delayed reflexes. Skin may be cool to touch and mottled in appearance reflecting circulatory compromise [8]. X-rays can reveal absent femoral epiphyses in up to 54% [4]. Figure 2 shows typical radiographs of epiphyseal dysgenesis.

Congenital hypothyroidism appears to be associated with an increased risk of congenital malformations. In one study of 1420 infants with congenital hypothyroidism, extra thyroidal congenital malformations had a prevalence of 8.4%. Of these, the majority were cardiac [17]. Other associated malformations include spiky hair, cleft palate, neurologic abnormalities and genitourinary malformations [17–19]. Also, the incidence of congenital hypothyroidism is increased in patients with Down's Syndrome [20].

Gene mutations causing congenital hypothyroidism can be a rare cause of distinct clinical phenotypes. Most well known is Pendred's syndrome. Affected patients have sensorineural deafness, hypothyroidism and goiter. This syndrome is due to a defect in pendrin, which is a transmembrane chloride-iodide transporter expressed in both the thyroid gland and the inner ear [21]. A mutation in thyroid transcription factor 2 (TTF-2) causes a syndrome of thyroid dysgenesis, choanal atresia, cleft palate and spiky hair also known as Bamforth-Lazarus syndrome [22] (Figure 3). Mutations in NKX 2.1 causes congenital hypothyroidism associated with respiratory distress and neurologic problems such as benign hereditary chorea and ataxia [23–25].

One clinical manifestation of long standing congenital hypothyroidism is the Kocher-Debre- Semelaigne syndrome. This presents as promixal muscle weakness associated with calf hypertrophy and resolves with thyroid hormone treatment [26]. Other clinical syndromes that include congenital hypothyroidism are included under "Syndromic hypothyroidism" in Table 3.

Transient hypothyroidism may be caused by maternal or neonatal factors. Maternal factors include antithyroid medications, transplacental thyrotropin receptor blocking antibodies and exposure to iodine deficiency or excess. Neonatal factors include, neonatal iodine deficiency or excess, congenital liver hemangiomas and mutations in the genes encoding for DUOX and DUOXA2 (see Table 3).

In iodine sufficient countries, 85% of congenital hypothyroidism is due to thyroid dysgenesis. This term refers to an aberration of the embryological development of the thyroid gland. The remaining 10-15% of cases can be attributed to the inborn errors of thyroid hormone synthesis, also called dyshormonogenesis, or to defects in peripheral thyroid hormone transport, metabolism, or action [27].

Thyroid dysgenesis presents in three major forms: thyroid ectopy, athyreosis and thyroid hypoplasia. Thyroid ectopy refers to an ectopic location of the thyroid gland. This accounts for two-thirds of congenital hypothyroidism due to thyroid dysgenesis and is twice as common in females [28]. In these cases, a thyroid remnant is usually found along the normal pathway of the thyroglossal duct. This represents the path taken by the developing thyroid as it descends from the base of the tongue to its final location in the neck [28, 29]. In one study done on hypothyroid neonates, ectopic thyroid tissue was found inferior and superior to the hyoid bone, and above the thyroid cartilage [30]. Athyreosis refers to the complete absence of thyroid tissue. Athyreosis and thyroid hypoplasia account for the remaining one third of thyroid dysgenesis.

Thyroid dysgenesis is generally thought to be sporadic in occurrence. However, recent evidence points to the possibility of a genetic component. One study of all cases of thyroid dysgenesis found that 2% were familial in occurrence [31]. Additional studies also showed that 7.9% of first degree relatives of infants with congenital hypothyroidism had a thyroid developmental anomaly [32]. There was some speculation as to a possible seasonal variation in the incidence of congenital hypothyroidism; however, this topic is still under debate [33, 34].

In contrast, PAX8 mutations seem to cause thyroid dysgenesis in the absence of other congenital anomalies [35–37]. However, given that PAX8 is also expressed in the mesonephros and ureteric buds [40], this may explain the increased incidence of genitourinary malformations in patients with congenital hypothyroidism [19].

There are several forms of TSH resistance. Mutations in the TSH receptor gene leading to thyroid hypoplasia have been found [41]. Another form of TSH resistance is dominantly inherited and has been linked to the long arm of chromosome 15 [42]. Resistance occurs in the absence of a TSH receptor mutation and can again cause thyroid hypoplasia [43]. Pseudohypoparathyroidism type 1a, caused by mutations in the alpha subunit of the stimulatory guanine nucleotide binding protein (Gs alpha), results in defective TSH signaling [44].

Hereditary defects in virtually all the steps of thyroid hormone biosynthesis and secretions have been described and account for 10-15% of permanent congenital hypothyroidism. These are generally transmitted in an autosomal recessive manner, but at least one condition has autosomal dominant inheritance. Dyshormonogenesis leads to goitrous hypothyroidism; however, this is rarely seen in babies detected by newborn screening [45].

Most commonly, dyshormonogenesis is due to defects of thyroid peroxidase activity [46]. Thyroid peroxidase uses hydrogen peroxide to couple iodine to thyroglobulin within the thyroid gland, forming T3 and T4. Severe defects in this enzyme lead to total iodide organification defects (TIOD). This diagnosis is made by showing high radioactive iodine (RAI) uptake of the thyroid gland followed by more than 90% release after sodium perchlorate administration [47] (see section on diagnosis). Less severe mutations cause partial iodide organification defects (PIOD). Pendred's syndrome is a well known form of syndromic hypothyroidism and is characterized by a triad of hypothyroidism, goiter and deafness. This syndrome is caused by a genetic defect in the transmembrane protein pendrin (encoded on 7q31), which acts as a chloride-iodide transporter in both in the thyroid gland and the inner ear. Defects in pendrin lead to impaired iodide organification and these patients have a positive perchlorate discharge test [21]. More recently, mutations in the enzyme dual oxidase 2 (known as DUOX2 or THOX2) have been found. They lead to dyshormonogenesis from deficient hydrogen peroxide generation and can be autosomal dominant. Their phenotype is heterogeneous and can be permanent or transient and cause either total or partial iodide organification defects [48]. Mutations in the dual oxidase maturation factor (DUOXA2) gene also lead to deficient iodide organification through similar mechanisms and can cause partial iodide organification defects [49]. Other, rare causes of dyshormonogenesis include defects in sodium/iodide transport, resulting from a mutation in the gene encoding the sodium-iodide symporter [50], and defective thyroglobulin action, resulting from a mutation in the gene encoding thyroglobulin [51]. A defect in the enzyme iodotyrosine deiodinase which aids in the peripheral conversion of T4 to T3 has been shown in hypothyroid individuals. This can be due to homozygous mutations in the genes DEHAL1 or SECISBP2 [52, 53].

Congenital secondary or central hypothyroidism generally results from defects of TSH production; most commonly, it is part of a disorder causing congenital hypopituitarism. Congenital hypopituitarism often is associated with midline defects such as septo-optic dysplasia or cleft lip and/or palate and can be part of a larger genetic syndrome. Mutations in genes regulating pituitary gland development, which include HESX1, LHX3, LHX4, PIT1 and PROP1 have been reported to be a cause of familial hypopituitarism. Besides TSH deficiency, other pituitary hormones are often deficient, including growth hormone, adrenocorticotrophic hormone and antidiuretic hormone. Rarely, specific gene defects lead to central hypothyroidism. These include isolated TSH deficiency (autosomal recessive, caused by mutations in the TSH β subunit gene), and thyrotropin releasing hormone (TRH) resistance, resulting from mutations in the TRH receptor gene (see Table 3).

Passage of thyroid hormone into cells is facilitated by thyroid hormone plasma membrane transporters. A mutation in a gene encoding monocarboxylase transporter 8 (MCT8) has been reported in five boys as a cause of X-linked hypothyroidism associated with mental retardation and neurologic abnormalities including quadriplegia. The defective transporter appears to impair the passage of T3 into neurons and is characterized by elevated serum T3 levels, low T4 and normal TSH [54]. This is also known as Allan-Herndon-Dudley syndrome.

Peripheral resistance to the action of thyroid hormone has been described. This is due in 90% of cases to mutations in genes encoding for thyroid hormone receptor β (TR β). These mutations are dominantly inherited and affected individuals are generally euthyroid, however some hypothyroid individuals have been described. Circulating T3 and T4 are mildly elevated without suppression of TSH. Thus these infants are usually not detected by newborn screening [55].

Transient congenital hypothyroidism is found to be more common in Europe (1:100) than the United States (1:50,000) [3]. In a report of over twenty years in the French newborn screening program, the incidence of transient congenital hypothyroidism was found to be 40 percent [3]. Causes of transient congenital hypothyroidism include:

In those countries with newborn screening programs in place, essentially all infants with congenital hypothyroidism are diagnosed after detection by newborn screening tests. Screening programs have been developed in Canada, the United States, parts of Mexico, Western Europe, Japan, Australia, New Zealand, and Israel, and they are under development in parts of many countries in Eastern Europe, Asia, South America and Africa. Of the worldwide birth population of 127 million, it is estimated that 25 percent undergo screening for congenital hypothyroidism. In infants born in locations without newborn screening programs, diagnosis may be made after development of clinical manifestations of hypothyroidism.

The specimen used for newborn screening tests is blood from a heel-prick collected on special filter paper cards. The specimen is routinely collected between two and five days of age (or at discharge from the hospital, if this occurs earlier); some programs use cord blood for screening. In addition, some programs also routinely obtain a 2 ^nd specimen between two and six weeks of age. The filter paper cards are then sent to a centralized laboratory for testing. Early in the experience of screening, most programs undertook an initial T4 test, with a follow-up TSH test on infants below a specified T4 cutoff [2]. With increasing accuracy of TSH measurement, many screening programs now carry out an initial TSH test to detect congenital hypothyroidism. Each program must develop its own T4 and TSH cutoff for recall of infants with abnormal test results (see Figure 4, Diagnostic algorithm). As there are rapid changes in TSH and T4 in the first few days of life, many programs have developed age-related cutoffs. Following are examples of typical cutoffs for T4 and TSH:

• Initial T4 < 10 ^th percentile →follow-up TSH test

• Initial TSH > 30 mU/L serum (> 15 mU/L whole blood)); some programs use an upper percentile TSH cutoff, e.g., > 97 ^th percentile

Once an infant has been detected with abnormal thyroid screening tests, they should be recalled immediately for examination, and a venapunctue blood sample should be obtained for confirmatory serum testing (see Figure 4, Diagnostic algorithm). Confirmatory serum is tested for TSH and either free T4 or total T4 combined with some measure of binding proteins, such as a T3 resin uptake.

The finding of an elevated serum TSH level and a low free T4 or total T4 confirms the diagnosis of primary hypothyroidism. The finding of an elevated serum TSH with a normal free T4 or total T4 is consistent with subclinical primary hypothyroidism. Because of the dependence of the developing brain on optimal concentrations of thyroid hormone, we recommend treating infants with subclinical hypothyroidism.

It is now recognized that preterm infants or acutely ill term infants with primary hypothyroidism may not show an elevated TSH level on the 1 ^st screening test. Thus, many programs undertake a routine 2 ^nd screening test in preterm and acutely ill term infants. Such testing leads to the detection of infants with "delayed TSH rise", which occurs in approximately 1:18,000 newborns [64].

Those programs that undertake a primary T4 test and recall infants with persistently low T4 screening levels, e.g. on the routine 1 ^st and 2 ^nd specimens (without elevated TSH levels) detect some infants with secondary or central hypothyroidism [65]. Confirmatory serum testing will show a low free T4 or total T4, with either a low TSH or an "inappropriately normal" TSH level. In infants with central hypothyroidism, the TSH deficiency may be isolated, as in infants with TSHβ gene mutations. In most cases of central hypothyroidism, however, TSH deficiency is associated with other pituitary hormone deficiencies. Such infant should be investigated for other pituitary hormone deficiencies, particularly if they have features such as hypoglycemia, which is suggestive of growth hormone (GH) and/or ACTH deficiency, or micropenis and undescended testes in a male, which are suggestive of gonadotropin (LH, FSH) deficiencies.

Some infants who undergo serum testing because of "low T4, non-elevated TSH" screening test results will be discovered to have thyroxine binding globulin (TBG) deficiency. Serum testing will show a low total T4, but a normal free T4 and normal TSH level. TBG deficiency can be confirmed by finding a low serum TBG level. TBG deficiency is an X-linked recessive disorder that occurs in approximately 1:4,000 infants, primarily males [66]. These infants are euthyroid and treatment is not necessary.

Either iodine-123 (I-123) or sodium pertechnetate 99 m (Tc99 m) are preferred for thyroid uptake and scan in neonates to minimize the radioactivity exposure; I-131 delivers a higher dose to the thyroid and total body and should not be used. Radionuclide uptake and scanning generally are the most accurate tests in defining some form of thyroid dysgenesis, e.g., an ectopic gland, thyroid hypoplasia (decreased uptake in a eutopic location), or thyroid aplasia [67] (Figure 5). Absence of radionuclide uptake should be confirmed by an ultrasonography. Absence of uptake can also be seen with TSHβ gene mutations, TSH receptor inactivating mutations, iodide trapping defects, and with maternal thyrotropin receptor blocking antibodies (TRB-Ab); thyroid ultrasonography and other studies, such as measurement of serum Tg or TRB-Ab will help to separate these etiologies from thyroid aplasia (see below).

If radionuclide studies show a large gland in a eutopic location with increased uptake, these findings are suggestive of one of the dyshormonogeneses beyond trapping. In such cases, I-123 uptake can be followed by a perchlorate discharge test. If there is defective oxidation and organification of iodide, it will not be attached to tyrosine on thyroglobulin, and so it will be rapidly "discharged" from the thyroid gland when high doses of perchlorate are given. Genetic studies for a mutation in thyroid peroxidase, the enzyme responsible for oxidation and organification, can confirm this inborn error of thyroid hormone biosynthesis.

Thyroid ultrasonography is accurate in confirming true thyroid aplasia. In the situations described above where radionuclide uptake and scan show absent uptake, but a gland is actually present (TSHβ gene mutations, TSH receptor inactivating mutation, iodide trapping defect, maternal TRB-Ab), ultrasonography may show a thyroid gland in a eutopic location. Ultrasonography generally is not as accurate as radionuclide scan in demonstrating ectopic glands [68]. More recent studies report that color flow doppler ultrasonography is able to detect ectopic thyroid tissue in 90 percent of infants with ectopic glands detected by radionuclide imaging [69]. Thyroid ultrasonography can confirm a large gland, suggestive of dyshormonogenesis.

Serum thyroglobulin levels reflect the amount of thyroid tissue and generally are elevated with increased thyroid activity, as when TSH is elevated. In addition, with inflammation, more thyroglobulin "leaks" into the circulation. One study showed that serum thyroglobulin levels were lowest in neonates with thyroid aplasia (mean 12 ng/mL, range 2 to 54 ng/mL), intermediate with ectopic glands (mean 92 ng/mL, range 11 to 231 ng/mL), and highest in cases associated with large glands (mean 226 ng/mL, range 3 to 425 ng/mL) [68]. Thus, while these groups could be separated by their serum thyroglobulin levels, given the degree of overlap it could not be used to diagnose the etiology in individual cases. In cases of true thyroid aplasia, serum thyroglobulin levels are absent if measured a few weeks after birth. Serum thyroglobulin determinations can be useful in cases of absent radionuclide uptake. If the thyroglobulin level is increased, this suggests that the thyroid gland is present, and that the neonate may have a TSH receptor inactivating mutation [70], a trapping defect, or maternal TRB-Ab, rather than aplasia.

Maternal autoimmune thyroid disease may be associated with production of a thryotropin receptor blocking antibody (TRB-Ab). This antibody will cross to the fetus and block TSH binding, inhibiting fetal thyroid gland development and function. Maternal autoimmune thyroid disease is relatively common, as approximately 5 percent of women of reproductive age have either anti-thyroglobulin or thyroid peroxidase antibodies [71]. However, maternal TRB-Ab is relatively rare, causing transient congenital hypothyroidism in approximately1:100,000 neonates [57]. Thus, we only recommend TRB-Ab determinations in a case where a previous child has had transient congenital hypothyroidism, and mother has known autoimmune thyroid disease and is pregnant again. TRB-Ab can be screened for by a thyrotropin binding inhibitor immunoglobulin (TBII) determination in mother.

If an infant with congenital hypothyroidism is born in an area of endemic iodine deficiency, or if there is a history of excess iodine exposure, measurement of urinary iodine may confirm either iodine deficiency or excess. 24 hour urinary iodine approximates iodine intake; the normal range in neonates is approximately 50 to 100 mcg.

Testing for specific genetic mutations generally is only considered after other studies point to a specific defect, e.g, a mutation in the gene for thyroid peroxidase in a neonate with a goiter, increased radionuclide uptake, and positive perchlorate discharge test. As noted previously, mutations in TTF-1, NKX2.1, or PAX-8 genes are found in only 2% of cases of thyroid dysgenesis. Thus, in the vast majority of cases of thyroid dysgenesis, the underlying cause remains unknown.

In cases where an infant with congenital hypothyroidism is detected by newborn screening tests and the diagnosis is confirmed by serum thyroid function tests, a clinical differential diagnosis is not considered. As described under confirmatory serum thyroid testing (above), results will lead to a diagnosis of primary congenital hypothyroidism, subclinical hypothyroidism, and, in some programs, secondary or central hypothyroidism.

In the absence of newborn screening programs, the diagnosis of congenital hypothyroidism is made after development of clinical manifestations. Since symptoms and signs develop gradually after birth, the diagnosis of hypothyroidism may be difficult at first. The timing of clinical features will vary depending on the severity of hypothyroidism. The myxedematous facial features, flat nasal bridge, macroglossia, and hypotonia may suggest Down syndrome or a metabolic storage disease. Prolonged jaundice and a protuberant abdomen may suggest a congenital liver disorder such as biliary atresia. Slow linear growth, a large head with immature body proportions, and radiological features of epiphyseal dysgenesis may be mistaken for a skeletal dysplasia or pituitary dwarfism. Eventually the clinical symptoms and signs suggest congenital hypothyroidism, and appropriate thyroid function tests confirm the diagnosis.

The most common cause of congenital hypothyroidism, thyroid dysgenesis, is typically a sporadic disorder, and so there is no recurrence risk with future pregnancies. The sporadic nature is supported by twin studies, which show a discordance for thyroid dysgenesis in both monozygotic and dizygotic twins [74]. There is evidence of a familial component in some cases of thyroid dysgenesis (aplasia, hypoplasia and ectopic glands). In a report of 2,472 patients with congenital hypothyroidism identified by the French newborn screening program and shown to be a result of thyroid dysgenesis, 48 (2 percent) appeared to be familial (typically occurring in siblings or cousins, but also some mother/father-daughter/sons) [28].Further evidence for a familial component comes from a French study which reported that 21.4 percent of first degree relatives of patients with congenital hypothyroidism had asymptomatic thyroid developmental anomalies, such as thyroglossal duct cysts, hemiagenesis, or a pyramidal lobe, as compared to 0.9 percent in a control population [32]. This study suggests a common genetic component between thyroid dysgenesis and these developmental anomalies. Rare cases of apparent thyroid agenesis have been reported in patients with loss-of-function mutations of the TSH receptor [70]. These cases appear to have an autosomal recessive pattern of inheritance. In summary, if a patient is detected with congenital hypothyroidism and imaging studies show some form of thyroid dysgenesis, the families can be counseled that the recurrence risk appears small, around 2 percent.

A minority of patients develop congenital hypothyroidism as a result of a hereditary defect in thyroid hormone biosynthesis, one of the dyshormonogeneses. Dyshormonogenesis may be suspected in an infant detected with congenital hypothyroidism and a goiter. These inborn errors of thyroid hormone biosysnthesis are the result of mutations in the sodium-iodide symporter, thyroid peroxidase, thyroglobulin, or iodotyrosine deiodinase genes. All of these inborn errors are autosomal recessive, and so they carry a 25 percent recurrence risk in future pregnancies. One specific disorder, Pendred's syndrome consists of sensorineural deafness, goiter, and impaired iodide organification. Pendred's syndrome is also an autosomal recessive disorder, linked to chromosome 7q22-33.1, and results from a mutation in a chloride-iodide transport protein expressed in the thyroid and inner ear [21]. While some patients with Pendred syndrome may develop hypothyroidism at birth [75], the majority are clinically and biochemically euthyroid.

There is a high recurrence risk of hypothyroidism in babies born to mothers with autoimmune thyroid disease associated with a thyrotropin receptor blocking antibody (TRB-Ab). Mothers should be investigated for TRB-Ab in cases of recurrent congenital hypothyroidism in siblings. The TRB-Ab will cross the placenta and block fetal thyroid gland development. Although this is an uncommon cause of congenital hypothyroidism [57], mothers should be counseled that as long as they have a high concentration of TRB-Ab, future pregnancies are at risk.

As congenital hypothyroidism is most commonly not a heritable disorder, and the majority of cases are sporadic, it is not possible to identify a population of pregnant women who are at high risk for fetal hypothyroidism. Specific pregnancies may be determined to be at risk based on a family history of a previous infant with congenital hypothyroidism, for example resulting from dyshormonogenesis or maternal TRB-Ab. Most cases, however, are not familial and are discovered when routine ultrasonography discloses a fetal goiter [76]. In addition, if a pregnant woman with Graves' disease is treated with antithyroid drugs, the fetus is at risk for hypothyroidism. Further, if a pregnant woman inadvertently receives radioactive iodine (RAI) after 8-10 weeks gestation, the fetal thyroid will trap the RAI, resulting in thyroid ablation and hypothyroidism.

There are rare case reports of subsequent pregnancies in families where a previous sibling had a familial form of congenital hypothyroidism [77]. In cases where a fetus is at risk for hypothyroidism, e.g., a previous infant with congenital hypothyroidism caused by dyshormonogenesis, or one of the rarer forms of familial thyroid dysgenesis, or a defect in thyroid hormone transport, where a genetic defect has been identified, genetic testing on fetal cells obtained by amniocentesis will determine whether or not the current pregnancy is also affected (for a list of defects, see 6. Genetic mutations, under "Diagnosis"). In these recessive disorders, with a recurrence risk of 25 percent, subsequent affected pregnancies may be suspected based on ultrasound findings of a fetal goiter, along with increased amniotic fluid and fetal bradycardia. In general, measurement of amniotic fluid TSH or thyroid hormone levels are not reliable, and sampling of fetal umbilical cord blood is necessary to diagnose fetal hypothyroidism. One case discovered by routine antenatal ultrasonography reported simultaneous amniotic fluid and fetal cord TSH measurements [78]. At 32 weeks gestation, amniotic fluid TSH was 8.76 mU/L while fetal cord blood TSH was 231.00 mU/L. With intra-amniotic injections of levothyroxine (l-thyroxine), the amniotic fluid TSH fell to 0.95 mU/L, while the cord blood TSH fell to 1.20 mU/L. The authors felt in this case that the initial amniotic fluid TSH was elevated (normal range 0.15-1.7 mU/L) and diagnostic of fetal hypothyroidism [78]. In general, genetic testing on fetal cells obtained by amniocentesis is a more direct and safer method of diagnosis than fetal cord blood sampling.

Several cases of hypothyroidism diagnosed antenatally have undergone treatment via intra-amniotic injections of l-thyroxine [79]. Typically, 250 mcg of l-thyroxine (range 250 to 600 mcg) has been injected weekly into the amniotic fluid. Subsequent dosing was based on the treatment effect in reducing the size of the fetal goiter and on repeat fetal cord blood thyroid tests. In general, such antenatal monitoring and treatment is well tolerated, although the risks of amniotic fluid injections and fetal cord blood sampling include premature labor, bleeding, and infection. While most cases report good psychomotor developmental outcome, there have not been any systematic studies of antenatal treatment of fetal hypothyroidism. Considering that the majority of infants with congenital hypothyroidism do well if detected by newborn screening with treatment started within the first 2-4 weeks of life, it is unclear if antenatal treatment is necessary for optimal neurocognitive outcome, though it clearly is successful in shrinking fetal goiter.

Congenital hypothyroidism is one of the most common treatable causes of mental retardation. Studies have shown that the timing of therapy is crucial to neurologic outcome. Indeed, there is an inverse relationship between intelligence quotient (IQ) and the age at diagnosis [13, 80]. Even when diagnosed early, neurologic development may suffer if treatment is not optimized in the first two to three years of life [81]. It is therefore important for these patients to receive early treatment and close follow up.

The overall goal of therapy is ensure that these patients are able to have growth and mental development that is as close as possible to their genetic potential. This is achieved by rapidly restoring the free T4 and the TSH to the normal range and then maintaining clinical and biochemical euthyroidism.

Levothyroxine (l-thyroxine) is the treatment of choice. Although triiodithyronine (T3) is the biologically active form of the hormone, most T3 in the brain is formed from local deiodination of T4; thus, T3 replacement is not needed for normal neurologic functioning. In a study of forty seven infants given varying treatment doses of l-thyroxine, serum T3 normalized and remained normal regardless of the treatment dose used, again suggesting that treatment with l-thyroxine alone is adequate [82]. Treatment should be initiated in any infant with a positive screening result, right after confirmatory tests are drawn but before results are available [82, 83].

Currently, only l-thyroxine tablets are approved for use in the United States. Thyroid suspensions prepared by individual pharmacies may result in unreliable dosing. In Europe, however, l-thyroxine drops have been successfully used [83]. The l-thyroxine tablet should be crushed, mixed with breast milk, formula or water and fed to the infant. The tablet should not be mixed with soy formula as this has been shown to interfere with absorption. One retrospective study of 78 patients with congenital hypothyroidism showed that infants fed soy formula took significantly longer to achieve a TSH under 10 mU/l (150 days versus 40 days) [84]. Should the infant require soy formula, l-thyroxine should be given halfway between feeds and thyroid function should be monitored carefully [80]. Some nutritional supplements or drugs are known to interfere with absorption of l-thyroxine.

These include:

Finally, prolonged heat exposure may reduce the efficacy of l-thyroxine tablets.

The dose and timing of thyroid hormone replacement are important in achieving optimal neurocognitive outcome. A delay in serum T4 normalization over one week can result in lower intelligence scores [81]. In one study, T4 normalization beyond two weeks resulted in patients scoring lower on behavioral and cognitive testing than patients who normalized in less than two weeks [85]. Thus, the goal of treatment should be to restore the serum T4 to > 129 mmol/L (> 10 μg/dl) as rapidly as possible. There is an inverse relationship between the starting l-thyroxine dose and the time to achieve the goal serum T4 concentration [85]. The recommended initial l-thyroxine dose set forth by the American Academy of Pediatrics (AAP) [82] and the European Society for Paediatric Endocrinology (ESPE) is 10-15 mcg/kg per day. In term infants this amounts to an average of 37.5 to 50 mcg per day [86]. In the aforementioned study, infants who received 50 mcg per day (12-17 mcg/kg/day) versus 37.5 mcg (10-15 mcg/kg/day) achieved higher performance scores for behavior, reading, spelling and math. Full scale IQ scores were 11 points higher in those started on 12- 17 mcg/kg/day [85]. Thus, for infants with low serum T4 levels, less than 5 μg/dl we would recommend using a dose of 12-17 mcg/kg/day, which is slightly above the higher end of the AAP recommended dose.

Infants with severe congenital hypothyroidism are at greater risk for developmental delay. This has been illustrated in studies done in both Europe, the United States and Canada [80, 85][87–89]. Therefore, rapid replacement with adequate doses of l-thyroxine is particularly important. This point was illustrated in one study of 83 infants who were assigned to receive three different starting doses of thyroid hormone at birth. The first group received 6.0-8.0 mcg/kg/day, group 2 received a dose of 8.1-10.0 mcg/kg/day and Group 3, a dose of 10.1-15.0 mcg/kg/day. These infants were then followed for growth and intellectual outcome at four years of age. In this study, the infants with severe congenital hypothyroidism achieved the highest intellectual scores when started at the highest dose [90]. Another study done in 61 infants compared early versus late treatment with low versus high dosing. Results showed that in infants with severe congenital hypothyroidism, only those treated early (< 13 days) and with higher doses (>9.5 mcg/kg/day) achieved normal psychomotor development at 10-30 months of age [81].

It is important to note that the time for TSH normalization is inversely related to neurodevelopmental outcome [86]. One study done in 45 children compared intellectual outcome at 2 and 6 years of age with variance of serum T4 and TSH. These investigators found that scores on the mental development index (MDI) and verbal IQ were predicted by mean T4 and TSH during the first year of treatment [91]. Therefore it is important to closely monitor these infants and adjust the l-thyroxine dose frequently until the desired level is achieved.

The treatment goals as outlined by the American Academy of Pediatrics (AAP) Update of newborn screening and therapy for congenital hypothyroidism [82] are similar to published European Society for Pediatric Endocrinology (ESPE) guidelines [86] and are as follows:

Infants who have serum T4 concentrations below 10 μg/dL in the first year of life accompanied by serum TSH concentrations above 15 mU/L have been shown to have lower IQ's than infants whose serum T4 concentrations are above 10 μg/dL [91]. Also, higher doses of l-thyroxine have been associated with higher intelligence quotients at 7 and 8 years of life, especially in the areas of verbal memory and verbal comprehension [92]. Thirdly, variations in serum T4 and TSH during the first year of life have been correlated with changes in mental development index and verbal intelligence quotient [91, 93].

This would suggest that higher doses of l-thyroxine lead to better overall developmental outcomes. However, the last study also notes that children on high dose l-thyroxine treatment had significant problems with hyperactivity, delinquency and aggression [92]. Other studies have also shown that high serum T4 levels contribute to poorer attention in school aged children [94]. These highlight the dangers of overtreatment in congenital hypothyroidism.

One study followed children on high dose therapy till four years of age; this demonstrated that higher intelligence quotients were observed in patients started on 10-15 mcg/kg/day without adverse effects on growth and skeletal maturation [90]

Clinical evaluation should be performed every few months during the first three years of life along with frequent measurements of serum T4 or free T4 and TSH. The American Academy of Pediatrics recommends the following monitoring schedule [86].

More frequently if results are abnormal or non-compliance is suspected.

The serum T4 should normalize within one to two weeks and the serum TSH should become normal in most infants after one month of treatment. In some individuals, a high TSH (10-20 mU/L) may persist despite a normal serum T4 or vice versa [95]. In most cases this is due to under treatment, however there are some individuals who will have abnormal maturation of free T4 feedback control on TSH secretion [96]. This abnormality is thought to exist in about 10 percent of treated individuals with congenital hypothyroidism and may be due to resetting of the pituitary-thyroid feedback mechanism in utero [97]. In one study of 42 patients, the prevalence of pituitary thyroid hormone resistance was as high as 43 per cent in younger infants less than one year and decreased to 10 percent in children and adolescents[97]. This suggests that thyroid hormone resistance is more common in the younger age group and may resolve with age.

Some patients with a positive newborn screen for congenital hypothyroidism have transient congenital hypothyroidism. Permanent congenital hypothyroidism can be assumed if:

If permanent congenital hypothyroidism has not been established by two to three years of age the AAP and the ESPE recommend a 30 day trial off l-thyroxine therapy [86, 98].

If a low serum T4 or free T4 and an elevated TSH are found, permanent congenital hypothyroidism is confirmed and the patient is restarted on therapy. If the serum T4 or freeT4 and TSH remain normal, the presumed diagnosis is transient congenital hypothyroidism and treatment is no longer needed. However, these patients must be followed closely and monitored for signs and symptoms of hypothyroidism such as constipation, slowing of growth or decreased mentation. If these appear then serum testing of thyroid function should be performed and if inconclusive, these patients should be continued to be followed closely with a low threshold for re-testing. Subjects with presumed transient hypothyroidism are vulnerable to recurrence during puberty and pregnancy and should be retested during these times. One novel approach is the use of recombinant TSH (rhTSH) to make the diagnosis of congenital hypothyroidism without requiring withdrawal of thyroid hormone. One study done on 10 children combined the use of ultrasound, scintigraphy after rhTSH, and percholorate discharge testing. This protocol resulted in the accurate diagnosis of permanent congenital hypothyroidism in 8 of 10 cases without stopping thyroxine. This shows that rhTSH may be of use in the future confirmation of permanent congenital hypothyroidism [99].

Prior to the newborn screening era, when a diagnosis of congenital hypothyroidism was made after development of clinical manifestations, studies reported an inverse relationship between the age of diagnosis and IQ outcome. A study from Pittsburgh Children's Hospital showed that if thyroid hormone treatment was started between birth and 3 months of age, the mean IQ was 89 (range 64 to 107); if treatment was started between 3 and 6 months of age, the mean IQ was 71 (range 35 to 96), while if treatment did not start until after 6 months of age, the mean IQ dropped to 54 (range 25 to 80) [100]. A report from Sweden found that "in spite of an efficient National Health Care Program for infants, the diagnosis was delayed until after 3 months in 52 percent of cases"[1].

The advent of newborn screening programs in the mid-1970s allowed earlier detection and treatment of infants with congenital hypothyroidism. Such efforts have been successful in achieving a much-improved neurocognitive outcome. Despite this, however, not all studies report a completely normal outcome. In a recent review of 51 published reports of IQ outcome in infants with congenital hypothyroidism as compared to sibling or classmate control subjects, 18 found no significant IQ difference, while 33 found a significant difference, with IQ ranging between 5 and 25 points lower in infants with congenital hypothyroidism [80]. In evaluating important variables, there is evidence that age of onset of treatment, starting l-thyroxine treatment dose, and severity of hypothyroidism each plays an important role in neurocognitive outcome.

However, as the starting doses were increased, reaching the currently recommended 10-15 mcg/kg/day, programs were able to compare IQ outcomes at various starting l-thryoxine doses. A report from the Toronto screening program compared psychometric outcome in infants started on 6.4 mcg/kg/day vs. 9.0 mcg/kg/day [92]. Verbal IQ was 98.6 vs. 106.3 (p < .01), performance IQ was 103.8 vs. 108.2 (p = NS), and full scale IQ was 100.0 vs. 107.6 (p < .01) in the low vs. higher dose infants, respectively. A report from Italy compared psychometric outcome in infants started on three different l-thyroxine doses: a "low" dose 6-8 mcg/kg/day, an "intermediate" dose 8.1-10.0 mcg/kg/day, and a "high" dose 10.1-15 mcg/kg/day [90]. The verbal IQs in the three treatment groups were 92, 94, and 98, respectively (p = NS); performance IQs were 85, 95, and 98, respectively (p < .01), while the full scale IQs were 88, 94, and 98, respectively (p < .01). A report from the U.S. Northwest Regional Screening Program showed that infants started on a higher l-thyroxine dose (50 mcg/day, equivalent to 12-17 mcg/kg/day) had an IQ score 11 points higher than those started on the lower dose (37.5 mcg/day, equivalent to 9.4-12.4 mcg/kg/day) [85].

In our review of the literature, of ten studies examining the effect of different starting l-thryoxine doses on psychometric outcome, two reported no effect, six reported a 12.3 higher IQ with higher starting doses, while two actually reported a 8.6 point higher IQ with lower starting doses [80]. Overall, the preponderance of studies found that children started on the currently recommended l-thyroxine dose of 10-15 mcg/kg/day appear to achieve the best IQ outcome.

Part of the explanation for a worse psychometric outcome in the most severely affected infants may be the lower starting l-thyroxine doses used in the early history of newborn screening. The Dutch newborn screening program investigated the effect of initial starting dose (< 9.5 or > 9.5 mcg/kg/day) and age of onset of treatment (< 13 days or > 13 days of age) in infants judged to have severe vs. milder hypothyroidism (pre-treatment serum free T4 0.21 ng/dL vs. 0.67 ng/dL, respectively) [106]. In the infants with severe hypothyroidism, psychometric testing at 10 to 30 months of age showed IQ 21 to 27 points lower in the groups treated with the lower starting dose and/or later age of onset; the group started on the higher dose and at the earlier age had the best IQ outcome (IQ = 125). On the other hand, all of the infants with milder hypothyroidism did well except the group treated with the lower dose and later age of onset, which had an IQ 22-25 points lower than the other groups [106]. These results support the concept of tailoring the initial starting l-thyroxine dose to the severity of hypothyroidism [107]. When the Quebec Screening Network used a higher starting dose, averaging 11.6 mcg/kg/day, they reported that psychometric testing did not show a difference in IQ in infants judged to have severe vs. more moderate hypothyroidism (107 vs. 110, respectively) [108].

The New England Congenital Hypothyroidism Collaborative also found that noncompliance beyond the first three years of life can affect cognitive performance. In a study of 14 year old adolescents with congenital hypothyroidism, the investigators made home visits without forewarning. Testing showed that 44% were inadequately treated, as judged by serum TSH > 15 mU/L, with serum T4 < 6.6 ug/dL (< 85 nmol/L) in the majority [110]. Psychometric testing showed a mean IQ of 106. Compliance with thyroid hormone treatment was stressed (the dose was not changed), and one to two years later, when thyroid testing was now improved, repeat psychometric testing showed a mean IQ of 112 (p < .002). This study showed that noncompliance in adolescents is common, and that improvement in compliance and thyroid levels was associated with improved cognitive function.

The underlying etiology of thyroid dysgenesis remains largely unknown. As described in the Etiology, Epidemiology and Genetic Counseling sections, while dysgenesis appears sporadic, there are clues to some familial/genetic factors. Approximately 2 percent of cases of thyroid dysgenesis have been shown to result from mutations in genes that code for transcription factors important in thyroid gland development, such as TTF-2, PAX-8, and NKX2. The higher incidence in certain racial and ethnic groups, in preterm infants, in twin and multiple births, and in older mothers points toward genetic or perhaps epigenetic factors that have yet to be discovered. In addition, the approximate 2:1 female:male ratio overall, more apparent with ectopic glands than with thyroid agenesis[111] is unexplained. Again, this finding points to undiscovered genetic factors, perhaps linked to autoimmuity, which is usually more common in females.

The explanation for the apparent increase in incidence of congenital hypothyroidism over the last 20 years is unclear. Is the increase real, or is it the result of changes in screening program test cutoffs, such that infants with milder cases of hypothyroidism are now being detected? Or, are there other factors, such as changes in the racial/ethnic population mix that contribute to the increase? It is also unclear whether the additional infants now being detected, including those with mild hypothyroidism and those with "delayed TSH rise" will have permanent or transient hypothyroidism.

Management of antenatal hypothyroidism as it relates to psychometric outcome is an open question. As discussed under Antenatal diagnosis, it is relatively rare to discover hypothyroidism in utero. When such cases are discovered, clinicians may feel pressure to treat the hypothyroidism with amniotic fluid injections of l-thyroxine. However, given the good neurocognitive outcome in infants with congenital hypothyroidism detected by newborn screening programs and started on thyroid hormone treatment in the first 2 to 4 weeks of life (who are not treated in utero), and the potential risks of intra-amniotic fluid injections and fetal cord blood sampling, treatment after birth may be a reasonable course of management. It is unlikely that there will ever be enough cases to perform randomized clinical trials to address this question. Psychometric testing in infants both treated and not treated antenatally, however, may provide useful information. A case can be made to treat hypothyroidism antenatally if a significant goiter is present.

Written informed consent was obtained from the parent of the patient for publication and accompanying images. A copy of the written consent is available for review by the Editor-in-Chief of this journal.