Clinical Manifestations
Clinical evidence of hypothyroidism is usually difficult to appreciate in the newborn period. Many of the classic features (large tongue, hoarse cry, facial puffiness, umbilical hernia, hypotonia, mottling, cold hands and feet and lethargy), when present, are subtle and develop only with the passage of time. This is illustrated in Figure 15-11 (below) which compares the findings in a baby with untreated congenital hypothyroidism diagnosed clinically with an infant in whom the diagnosis was made by newborn screening. In addition to the aforementioned findings, nonspecific signs that should suggest the diagnosis of neonatal hypothyroidism include: prolonged, unconjugated hyperbilirubinemia, gestation longer than 42 weeks, feeding difficulties, delayed passage of stools, hypothermia or respiratory distress in an infant weighing over 2.5 kg. A large anterior fontanelle and/or a posterior fontanelle > 0.5 cm is frequently present in affected infants but may not be appreciated. In general, the extent of the clinical findings depends on the cause, severity and duration of the hypothyroidism. Babies in whom severe feto-maternal hypothyroidism was present in utero tend to be the most symptomatic at birth. Similarly, babies with athyreosis or a complete block in thyroid hormonogenesis tend to have more signs and symptoms at birth than infants with an ectopic thyroid, the most common cause of congenital hypothyroidism. Unlike acquired hypothyroidism, babies with congenital hypothyroidism are of normal size. However, if diagnosis is delayed, subsequent linear growth is impaired. The finding of palpable thyroid tissue suggests that the hypothyroidism is due to an abnormality in thyroid hormonogenesis or in thyroid hormone action, or that it will be transient.
 |
| Figure 15-11. (Upper panel) Infant with severe, untreated congenital hypothyroidism diagnosed prior to the advent of newborn screening. (Lower panel) Infant with congenital hypothyroidism identified through newborn screening. Note the striking difference in the severity of the clinical features. |
Infants detected by newborn screening should be evaluated without delay, preferably within 24 hours. The diagnosis of neonatal hypothyroidism is confirmed by the demonstration of a decreased concentration of T4 (<6.5 m g/dL; 3.7 nmol/L) and an elevated TSH level (> 20 mU/L) in serum. As noted previously, most infants with permanent abnormalities of thyroid function have a serum TSH concentration >40 mU/L. Physicians should be aware that the serum T4 concentration is much higher in full term infants in the first 2 months of life (6.5 - 16.3 m g/dL; 3.7 - 210 nmol/L) than in adults for whom reference values are given in most laboratories) (171). Normal values for thyroid function in the neonatal period are given Table 15-8. Measurement of T3 is of little value in the diagnosis of congenital hypothyroidism.
A bone age is often performed as a reflection of the duration and severity of the hypothyroidism in utero. A radionuclide scan (either 123I or 99mTcO4) provides information about the location, size and trapping ability of the thyroid gland; ectopic thyroid glands, frequently sublingual, may be located anywhere along the pathway of thyroid descent from the foramen cecum to the anterior mediastinum. Thyroid imaging is helpful in verifying whether a permanent abnormality is present and aids in genetic counselling since thyroid dysgenesis is a almost always sporadic condition whereas abnormalities in thyroid hormonogenesis are autosomal recessive. Scintigraphy with 123I, if available, is usually preferred because of the greater sensitivity and because, 123I, unlike pertechnetate is organified. Therefore, imaging with this isotope allows quantitative uptake measurements and tests for both iodine transport defects and abnormalities in thyroid oxidation. The lowest possible dose of 123I, usually 25 mCi, should be used. Advantages of pertechnetate, on the other hand, are that it is cheaper and more widely available. There is some disagreement as to whether a thyroid scan should be performed in all babies because of the unknown risk of radiation exposure, particularly in centers where only 131I is used and relatively large doses of isotope are administered.
If there is no uptake on thyroid imaging, an ultrasound study should be performed to confirm the absence of thyroid tissue. Ultrasonography is also helpful as an alternative to thyroid scintigraphy to verify the presence of a eutopic thyroid gland if a transient abnormality is suspected or if a thyroid gland is palpable on clinical examination; this procedure is less sensitive than a radionuclide scan, however 71.
Occasionally, apparent thyroid agenesis is due to the presence of maternal TSH receptor blocking antibodies, which, if present in a sufficiently high titer, completely inhibit TSH-induced thyroidal uptake of radioisotope 165. The presence of autoimmune thyroid disease in the mother or a history of a previously affected sibling should alert the physician to the possibility of this diagnosis but this information is not always known and should not be relied upon. A radio-receptor assay is appropriate for screening; bioassay can be done later if desired to demonstrate the biological action of the antibodies. This topic is discussed in further detail below. In cases of TSH receptor antibody-induced congenital hypothyroidism, the blocking activity is extremely potent, half-maximal TSH binding-inhibition being reported with as little as a 1/20 to 1/50 dilution of serum; a weak or borderline result should cause a reconsideration of this diagnosis. Similarly, thyroid peroxidase (TPO) antibodies, although frequently detectable in babies with blocking antibody-induced congenital hypothyroidism, are neither sensitive nor specific in predicting the presence of transient congenital hypothyroidism 165.
Other disorders that may mimic thyroid agenesis on thyroid scintigraphy include loss of function mutations of the TSH receptor, iodine excess, or an iodide concentrating abnormality. As noted above, potential clues to the diagnosis of a loss of function mutation of the TSH receptor include a normal thyroglobulin and/or evidence of a thyroid gland on ultrasound examination despite the failure to visualize thyroid tissue on imaging studies 148. Ultimately verification of this diagnosis resides in the demonstration of a genetic abnormality in the TSH receptor gene. Measurement of urinary iodine is helpful if a diagnosis of iodine-induced hypothyroidism is suspected. An iodide- concentrating defect should be suspected in patients with a family history of congenital hypothyroidism, particularly if an enlarged thyroid gland is present. The detailed evaluation of infants suspected of having this and other abnormalities in thyroid hormonogenesis has been described in Chapter 16b and elsewhere 71,141,142.
Serum thyroglobulin concentration, sometimes measured in the evaluation of babies with congenital hypothyroidism, reflects the amount of thyroid tissue present and the degree of stimulation. For example, thyroglobulin is undetectable in most patients with thyroid agenesis; thyroglobulin may be elevated in patients with abnormalities of thyroid hormonogenesis not involving thyroglobulin synthesis and secretion, and is intermediate in babies with an ectopic thyroid gland. Considerable overlap exists, however, limiting the value of thyroglobulin measurement in the differential diagnosis of the aforementioned disorders. Measurement of thyroglobulin is most helpful when an inactivating mutation of the TSH receptor or a defect in thyroglobulin synthesis or secretion is being considered. In the latter condition the serum thyroglobulin concentration is low or undetectable despite the presence of an enlarged, eutopic thyroid gland.
In babies in whom hypothyroxinemia unaccompanied by TSH elevation is found, a free T4 should be measured, preferably by a direct dialysis method and the TBG concentration should be evaluated as well. The finding of a low free T4 in the presence of a normal TBG may suggest the diagnosis of 2o or 3o hypothyroidism. In these cases, TRH testing (TRH, 7 mg/kg IV) will help to distinguish whether the abnormality is at the level of the pituitary gland or hypothalamus. Pituitary function testing and brain imaging should also be performed in these infants. In premature, low birth weight or sick babies in whom a low T4 and "normal" TSH" are found, the free T4 when measured by a direct dialysis method, frequently is not as low as the total T4. In the latter infants T4 (and/or free T4), and TSH should be repeated every 2 weeks until the T4 normalizes because of the rare occurrence of delayed TSH rise 124,172. Similarly, any baby suspected of being hypothyroid clinically should have repeat thyroid function testing because of rare errors in the screening program.
A suggested approach to the investigation of infants with abnormal results on newborn thyroid screening is presented in Figure 15-10:
 |
| Figure 15-10. |
Replacement therapy with levothyroxine sodium should be begun as soon as the diagnosis of congenital hypothyroidism is confirmed, and the parents should be counselled regarding the causes of congenital hypothyroidism, the importance of compliance and the excellent prognosis in most babies if therapy is initiated sufficiently early and is adequate. Educational materials should be provided. Treatment need not be delayed in anticipation of performing thyroid imaging studies as long as the latter are done within 5-7 days of initiating treatment (before suppression of the serum TSH). An initial dosage of 10-15 m g/kg is recommended so as to normalize the T4 as soon as possible. Babies with compensated hypothyroidism may be started on the lower dosage, while those with severe CH (e.g., T4<5 mg/dL (64 nmol/L)) such as those with thyroid agenesis should be started on the higher dosage. Thyroid hormone may be crushed and administered with juice or formula, but care should be taken that all of the medicine has been swallowed. Thyroid hormone should not be given with substances that interfere with its absorption, such as iron, soy, or fiber. Many babies will swallow the pills whole or will chew the tablets with their gums even before they have teeth. Unfortunately, liquid preparations are unstable.
The aims of therapy are to normalize the T4 as soon as possible, to avoid hyperthyroidism and to promote normal growth and development. When the aforementioned amount of levothyroxine is used, the T4 will normalize in most infants within 1 week and the TSH will normalize within 1 month. Subsequent adjustments in the dosage of medication are made according to the results of thyroid function tests and the clinical picture. Often small increments or decrements of L-thyroxine (12.5 mg) are needed. This can be accomplished by 1/2 tablet changes, by giving an alternating dosage on subsequent days, or by giving an extra tablet once a week. Some infants will develop supraphysiologic serum T4 values on this amount of thyroid replacement but the serum T3 concentration usually remains normal, affected infants are not symptomatic, and available information suggests that these short-term T4 elevations are not associated with any adverse effects on growth, bony maturation, or cognitive development. In rare infants, normalization of the TSH concentration may be delayed because of relative pituitary resistance. In such cases, characterized by a normal or increased serum T4 and an inappropriately high TSH level, the T4 value is used to titrate the dosage of medication, but noncompliance should be excluded 81. One usually aims at maintaining the T4 above 10 m g/dL (128.7 nmol/L) and the TSH at less than 10 mU/L. Close follow-up is necessary. Current recommendations of both The American Academy of Pediatrics and the American Thyroid Association are to repeat T4 and TSH at 2 and 4 weeks after the initiation of L-thyroxine treatment, every 1-2months during the first year of life, every 2-3 months between 1 and 3 years of age, and every 3-12 months thereafter until growth is complete 124. In hypothyroid babies in whom an organic basis was not established at birth and in whom transient disease is suspected, a trial off replacement therapy can be initiated after the age of 3 years when most thyroxine-dependent brain maturation has occurred.
Whether or not premature infants with hypothyroxinemia should be treated remains controversial at the present time. Early retrospective investigations failed to document a difference in cognitive outcome in premature infants with hypothyroxinemia as compared with controls, but small numbers were studied. More recently, several retrospective, cohort studies have documented a relationship between severe hypothyroxinemia and both developmental delay and disabling cerebral palsy in preterm infants <32 weeks gestation 60,160,173. Whether or not the poorer prognosis in these infants is causal or coincidental cannot be determined, however, since the serum T4 in premature infants, as in adults, has been shown to reflect the severity of illness and risk of death. In other studies, investigators have evaluated the effect of therapeutic intervention with T4 or T3 not only on neuro-cognitve outcome, but on mortality rate and respiratory function as well. A variety of dosage schedules have been used, but once again conflicting results have been obtained. In the most thorough study to date, Van Wassenaer et al carried out a placebo-controlled, double-blind trial of T4 treatment, 8 mg /kg per day for 6 weeks in 200 infants less than 30 weeks gestation. Although overall no difference in cognitive outcome was found, there was an 18-point increase in the Bayley Mental Development Index score in the subgroup of T4-treated infants<27 weeks gestation 174. Of some concern was the additional finding that treatment with T4 was associated with a 10-point decrease in mental score (p=0.03) in infants >27 weeks gestation. While further studies are needed, it would seem reasonable at the present time to treat any premature infant with a low T4 and elevated TSH and to consider treatment of any infant <27 weeks with a low T4 whether or not the TSH is elevated. A dosage of 8 mg/kg/day in the latter group of infants has been recommended. Whether or not to treat older premature infants with hypothyroxinemia and what dosage to use remains uncertain.
Numerous studies have been performed to evaluate the cognitive outcome of babies with congenital hypothyroidism detected on newborn screening. In the initial reports, despite the eradication of severe mental retardation, the intellectual quotient (IQ) of affected infants was nonetheless 6-19 points lower than control babies 72. Though this IQ deficit was small, it was nonetheless significant as judged by a 4-fold increase in the need for special education in affected children. In addition, sensorineuiral hearing loss, sustained attention problems, and various neuropsychological variables were noted in some patients, although the frequency and severity of these abnormalities were much less than in the pre-screening era. Those babies most likely to have permanent intellectual sequelae were infants with the most severe in utero hypothyroidism as determined by initial T4 level (<5mg/dL (64 nmol/L)) and skeletal maturation at birth. These findings led to the widely-held conclusion at the time that some cognitive deficits in the most severely affected babies might not be reversible by postnatal therapy 72.
In the initial programs, a levothyroxine dosage of 5mg-8mg/kg was used and treatment was not initiated until 4-5 weeks of age. In contrast, accumulating data from a number of different studies have demonstrated that when a higher initial treatment dose is used (10mg-15mg/kg) and treatment is initiated earlier (before 2 weeks) this "development gap" can be closed, irrespective of the severity of the congenital hypothyroidism at birth 174. In the most compelling study to date, Bongers-Schokking et al have shown recently that even babies with severe congenital hypothyroidism can achieve normal psychomotor development at 10 to 30 months as long as treatment is initiated before 13 days of age and an initial dose of >9.5 mg/kg/day is used. However, if treatment is delayed or a lower dose is used, a 20 point deficit in both mental and psychomotor development is observed 174a. This conclusion is consistent with the original findings of the New England Hypothyroid Screening Collaborative. The latter program reported over 20 years ago that when a starting dose of (10 mg/kg) of thyroid replacement was used, the mean IQ of the congenital hypothyroid babies was indistinguishable from control infants, a finding that has remained true for 14 years of follow up 81,83. In addition to adequate dosage, assurance of compliance is essential for an optimal developmental outcome 81. For patients treated in the original screening programs, the long term problems appear to be in the areas of memory, attention and visual spacial problems 174b.
Unlike congenital hypothyroidism which usually is permanent, neonatal hyperthyroidism almost always is transient and results from the transplacental passage of maternal TSH receptor stimulating antibodies. Hyperthyroidism develops only in babies born to mothers with the most potent stimulatory activity in serum 175,176. This corresponds to 1-2% of mothers with Graves’ disease, or 1 in 50,000 newborns, an incidence that is approximately four times higher than is that for transient neonatal hypothyroidism due to maternal TSH receptor blocking antibodies.
Some mothers have mixtures of stimulating and blocking antibodies in their
circulation, the relative proportion of which may change over time. Not surprisingly,
the clinical picture in the fetus and neonate of these mothers is more complex
and depends not only on the relative proportion of each activity in the maternal
circulation at any one time but on the rate of their clearance from the neonatal
circulation postpartum. Thus, one affected mother gave birth, in turn, to a
normal infant, a baby with transient hyperthyroidism, and one with transient
hypothyroidism 177. In another
neonate, the onset of hyperthyroidism did not become apparent until 1-2 months
postpartum when the higher affinity blocking antibodies had been cleared from
the neonatal circulation 178.
In the latter case, multiple TSH receptor stimulating and blocking
antibodies were cloned from the maternal peripheral lymphocyties. Each monoclonal antibody recognized
different antigenic determinants ("epitopes") on the receptor and
had different functional properties 179.
Occasionally, neonatal hyperthyroidism may even occur in infants born to hypothyroid
mothers. In these situations, the maternal thyroid has been destroyed either
by prior radioablation, surgery or by coincident destructive autoimmune processes
so that potent thyroid-stimulating antibodies, present in the maternal circulation
are silent in contrast to the neonate whose thyroid gland is normal 179.
Although maternal TSH receptor antibody-mediated hyperthyroidism may present in utero, most often the onset is during the first week of life. This is due both to the clearance of maternally-administered antithyroid drug (propylthiouracil, PTU, methimazole or carbimazole) from the infant's circulation and to the increased conversion of T4 to the more metabolically active T3 after birth. Rarely, as noted earlier, the onset of neonatal hyperthyroidism may be delayed until later if higher affinity blocking antibodies are also present. Fetal hyperthyroidism is suspected in the presence of fetal tachycardia (pulse greater than 160/min) especially if there is evidence of failure to thrive. In the newborn infant, characteristic signs and symptoms include tachycardia, irritability, poor weight gain, and prominent eyes. Goiter, when present, may be related to maternal antithyroid drug treatment as well as to the neonatal Graves' disease itself. Rarely, infants with neonatal Graves' disease present with thrombocytopenia, hepatosplenomegaly, jaundice, and hypoprothrombinemia, a picture that may be confused with congenital infections such as toxoplasmosis, rubella, or cytomegalovirus 180. In addition, arrhythmias and cardiac failure may develop and may cause death, particularly if treatment is delayed or inadequate. In addition to a significant mortality rate that approximates 20% in some older series, untreated fetal and neonatal hyperthyroidism is associated with deleterious long-term consequences, including premature closure of the cranial sutures (cranial synostosis), failure to thrive, and developmental delay 181.
The half-life of TSH receptor antibodies is 1 to 2 weeks 182. The duration of neonatal hyperthyroidism, a function of antibody potency and the rate of their metabolic clearance, is usually 2 to 3 months but may be longer.
Because of the importance of early diagnosis and treatment, fetuses and infants
at risk for neonatal hyperthyroidism should undergo both clinical and biochemical
assessment as soon as possible. Situations that should prompt consideration
of neonatal hyperthyroidism are listed in Table
15-9. A high index of suspicion is necessary in babies of women who have
had thyroid ablation because in them a high titer of TSH receptor antibodies
would not be evident clinically. Similarly, women with persistently elevated
TSH receptor antibodies and with a high requirement for antithyroid medication
are at an increased risk of having an affected child.
Table 15- 9. Situations That Should Prompt Consideration of Neonatal Hyperthyroidism |
|
1. Unexplained tachycardia, goiter or stare |
Table 15-10. Differential Diagnosis of Juvenile Hypothyroidism |
|
1o HYPOTHYROIDISM Congenital Abnormality Iodine Deficiency (endemic goiter) Drugs or Goitrogens Miscellaneous 2o OR 3o HYPOTHYROIDISM Congenital Abnormality Surgery Radiation |
The diagnosis of hyperthyroidism is
confirmed by the demonstration of an increased concentration of circulating
T4 (and free T4, and T3, if possible) accompanied by a suppressed TSH level
in neonatal or fetal blood. The latter can be obtained by cordocentesis if someone
experienced in this technique is available. Results should be compared with
normal values during gestation (Figure
15-8) 26 Fetal
ultrasonography may be helpful in detecting the presence of a fetal goiter and
in monitoring fetal growth. Demonstration in the baby or mother of a high titer
of TSH receptor antibodies will confirm the etiology of the hyperthyroidism
and, in babies whose thyroid function testing is normal initially, indicate
the degree to which the baby is at risk.
 |
| Figure 15-8. Cord blood concentrations of T4, FT4, T3, and TSH in premature infants according to gestational age. Lines indicate the mean, 5th, and 95th percentile. Hatched areas represent lower limits of sensitivity for the assays. (From Thorpe-Beeston et al (26), as modified by Franchi (60), with permission). |
As noted in the case of TSH receptor blocking antibody-induced congenital hypothyroidism, the radioreceptor assay is the most cost-effective, rapid and technically feasible approach. If desired, bioassay can performed subsequently to demonstrate the biological activity of the antibodies if the radioreceptor assay is positive. In general, babies likely to become hyperthyroid have the highest TSH receptor antibody titer whereas if TSH receptor antibodies are not detectable, the baby is most unlikely to become hyperthyroid 175,183,184. In the latter case, it can be anticipated that the baby will be euthyroid, have transient hypothalamic-pituitary suppression or have a transiently elevated TSH, depending on the relative contribution of maternal hyperthyroidism versus the effects of maternal antithyroid medication, respectively 183. Therapy is rarely necessary. This is true whether TSH receptor antibodies are measured by radioreceptor assay or by bioassay. On the other hand, if TSH receptor antibody potency is intermediate, it is likely that the baby will be euthyroid, have a transiently elevated T4 or have transient hypothalamic pituitary suppression 168,183,184. It is important to appreciate that the sensitivity of TSH receptor assays in different laboratories varies. Therefore, specific values that are recommended in the literature should be interpreted with caution and, ideally, each laboratory should determine its own range. Close follow up of all babies with abnormal thyroid function tests or detectable TSH receptor antibodies is mandatory.
In the fetus, treatment is accomplished by maternal administration of antithyroid medication. The minimal dosage of PTU (or MMI) necessary to normalize the fetal heart rate and render the mother euthyroid or slightly hyperthyroid is usually chosen. In the neonate, treatment anticipates that the disease is self limiting. Either PTU (5 to10 mg/kg/day) or MMI (0.5 to 1.0 mg/kg/day) can be used initially in 3 divided doses. If the hyperthyroidism is severe, a strong iodine solution (Lugol's solution or SSKI, 1 drop every 8 hours) is added to block the release of thyroid hormone immediately because the effect of PTU and MMI may be delayed for several days. Therapy with both PTU and iodine is adjusted subsequently, depending on the response. Propranolol (2 mg/kg/day in 2 or 3 divided doses) is added if sympathetic overstimulation is severe, particularly in the presence of pronounced tachycardia. If cardiac failure develops, treatment with digoxin should be initiated, and propranolol should be discontinued. Rarely, prednisone (2 mg/kg/day) is added for immediate inhibition of thyroid hormone secretion. Alternately, sodium ipodate (0.5 gm every 3 days), an iodine-containing radiocontrast material that inhibits both thyroid hormone secretion and the conversion of T4 to T3, has been used successfully as the sole treatment of neonatal hyperthyroidism 185. Measurement of TSH receptor antibodies in treated babies may be helpful in predicting when antithyroid medication can be safely discontinued 176. Lactating mothers on antithyroid medication can continue nursing as long as the dosage of PTU or MMI does not exceed 400 mg or 40 mg, respectively. Since the milk/serum ratio of PTU is 1/10 that of MMI, a consequence of pH differences and increased protein binding, PTU is preferable to MMI. At higher dosages of antithyroid medication, close supervision of the infant is advisable.
Rarely, neonatal hyperthyroidism is permanent and is due to a germline mutation in the TSH receptor resulting in its constitutive activation 186-190. A gain of function mutation of the TSH receptor should be suspected if persistent neonatal hyperthyroidism occurs in the absence of detectable TSH receptor antibodies in the maternal circulation. Most cases result from a mutation in exon 10 which encodes the transmembrane domain and intracytoplasmic tail of the TSH receptor, a member of the G-protein coupled receptor superfamily 186-189. Less frequently, a mutation encoding the extracellular domain has been described 190. An autosomal dominant inheritance has been noted in many of these infants; other cases have been sporadic, arising from a de novo mutation. Early recognition is important because the thyroid function of affected infants is frequently difficult to manage medically 186,188,189, and, when diagnosis and therapy is delayed irreversible sequelae, such as cranial synostosis and developmental delay may result 186. For this reason early, aggressive therapy with either thyroidectomy or even radioablation has been recommended.
