Part G: Endocrine

INTRODUCTION

This chapter outlines the epidemiology, etiology, pathophysiology, differential diagnosis, and management approach to the neonate with hypoglycemia, here defined as 40 mg/dL or less (≤2.2 mmol, the fifth percentile for age) during the first 2 days of life and 50 mg/dL or less (≤2.8 mmol) of whole blood glucose thereafter, with or without suggestive symptoms. Note that plasma glucose concentration is about 10%–15% greater than whole blood glucose concentration, so criteria for hypoglycemia based on plasma glucose measurements must be appropriately adjusted to be greater. Recent advances in our understanding of the biochemistry, physiology, and molecular biology regulating prenatal and postnatal glucose homeostasis combine to provide a rational basis for defining, identifying, diagnosing, and treating hypoglycemia in the newborn to enable normal neurodevelopment.^{1, 2} These considerations provide a systematic approach to the problem of neonatal hypoglycemia and argue for glucose measurements to become part of routine care in all neonates prior to discharge from the newborn nursery and in all sick neonates even after discharge from the newborn nursery.

EPIDEMIOLOGY

Hypoglycemia is a relatively common and highly important problem in the newborn.^{3, 4, and 5} Precise data on incidence are unavailable and depend in part on the definition of hypoglycemia, an area of ongoing discussion, as well as the degree of gestational maturity and condition of the newborn.^{1, 2, and 3} Based on a meta-analysis,⁵ it has been proposed that, in full-term normal newborns, blood glucose of 40 mg/dL or less in the first 48 hours and 48 mg/dL or less between 48 and 72 hours represents less than the fifth percentile for age. A minor modification to these data permits defining hypoglycemia as 40 mg/dL or less in the first 2 days of life and 50 mg/dL or less thereafter. By the fourth day of life, normal infants usually maintain average blood glucose values greater than 60 mg/dL, approaching values of healthy children and adults. Using these criteria, the reported incidence of hypoglycemia in the first days of life varies from approximately 70% in infants who are small for gestational age (SGA),⁶ 20%–50% in those large for gestational age (LGA) but otherwise-normal newborn infants born to nondiabetic mothers, and greater than 50% in those born to infants of diabetic mothers⁷; only about 2% of term infants born to nondiabetic mothers after normal pregnancy develop hypoglycemia.⁸

The importance of neonatal hypoglycemia lies in its association with potential impairment of neurocognitive development⁶ because glucose is the preferential energy substrate of the newborn's brain, which normally utilizes greater than 90% of normal basal glucose turnover and has been shown to be about 4–8 mg/kg/min.⁹ Although the brain of a neonate can use other energy substrates, such as lactate and β-hydroxybutyrate, in hypoglycemia secondary to hyperinsulinism (HI), the production of glucose (by glycogenolysis or gluconeogenesis) is suppressed, as is lipolysis and ketogenesis, while glucose utilization and storage are increased. Hence, in HI, the brain is effectively deprived of all sources of nutrients. The consequences of such hyperinsulinemic hypoglycemia, the most common form of persistent hypoglycemia of the newborn, include high rates of seizures, mental retardation, neuromuscular spasticity, and other disturbances in psychomotor development.^{10, 11, and 12}

In SGA and preterm infants, hypoglycemia that is recurrent or repetitive is a more predictable factor for long-term neurodevelopmental deficits than the severity of a single hypoglycemia episode.⁶ Thus, the degree and duration of hypoglycemia, as well as associated conditions such as hypoxia and seizures, add to the risk of neural and intellectual impairment. Moreover, the younger the infant, the earlier and more persistent the hypoglycemia, and the greater the difficulty to control seizures by medical means alone, the more likely it is that long-term neural and intellectual development will be adversely affected.^{10, 11, and 12} There is some evidence that the neonatal brain may be more resistant than the adult brain to hypoglycemia, especially because of its ability to better use lactate and β-hydroxybutyrate as alternate fuels.¹³ Precisely for these reasons, HI is the most concerning of the syndromes of hypoglycemia of the newborn because generation of alternate fuels also is diminished or abolished by the HI.

Based on these findings, we suggest that an operational threshold of whole blood glucose of 50 mg/dL or less should be viewed with concern for monitoring and intervention. In making this recommendation, we recognize that we are “raising the bar” from previous standards. However, it is now established that persistent hypoglycemia is harmful, even if a precise glucose concentration below which harm ensues cannot be individually determined.^{10, 11, and 12} Hence, we urge earlier investigation and intervention as a precautionary measure. When symptoms of hypoglycemia, especially seizures, accompany the low glucose measurements, adverse neurodevelopmental outcomes should be anticipated.

PATHOPHYSIOLOGY

Hypoglycemia

Neonatal hypoglycemia has its origins in utero and in the dramatic transition to independent existence following separation from the placenta. In utero, the fetus receives all of its nutrients from the mother via the umbilical vein, and gluconeogenesis is absent or minimal until birth.^{1, 2} During the third trimester, weight doubles from approximately 1700 g at 32 weeks' gestation to approximately 3500 g at term, representing in large part accretion and deposition of glycogen, fat, and muscle tissue. The degree of fetal insulin secretion, though modest, is sufficient to stimulate cognate insulin receptors and permit partitioning of nutrients for immediate energy use and for tissue growth. Excessive nutrient transfer, as occurs to a variable degree in all diabetic pregnancies, promotes secretion of insulin, a growth-promoting hormone, which leads to macrosomia in utero.^{14, 15} Transient neonatal hypoglycemia, lasting 1-3 days, occurs after umbilical cord cutting, when glucose supply is interrupted, whereas excessive insulin continues for some time. Similar considerations determine macrosomia resulting from genetic defects causing HI in utero, but in these cases, postnatal hypoglycemia persists.^{16, 17} The severity of the hypoglycemia depends on the degree of HI and the duration between feedings, that is, “fasting.”

The term hyperinsulinism denotes excessive insulin action relative to the prevailing blood glucose concentration. This is reflected in the metabolic profile of HI-associated hypoglycemia caused by suppressed glucose production concurrent with increased glucose utilization, exceeding the usual range of 4–8 mg/kg/min,⁹ suppressed lipolysis leading to low circulating free fatty acids (FFAs), and suppressed ketogenesis leading to low circulating concentrations of β-hydroxybutyrate. The degree of HI may not be reflected in the circulating concentration of insulin, in part because insulin is secreted into the portal vein, and hepatic extraction may exceed the theoretical 50% reported in normal adults, and in part because activating mutations in elements of the insulin-signaling pathway will produce an identical clinical and biochemical phenotype (macrosomia, hypoglycemia, low FFA, low β-hydroxybutyrate) but with insulin levels that are low (<2 mIU/mL) or even below the limits of assay detection.¹⁸ Thus, the key to differentiating and managing the causes of neonatal hypoglycemia is the metabolic profiles of glucose, FFAs, and β-hydroxybutyrate, which are more readily available in most laboratories than accurate ultrasensitive assays for measurement of hormones such as insulin, cortisol, and growth hormone (GH).

Umbilical cord cutting at birth triggers a series of coordinated hormonal and enzymatic changes that mobilize the fuel stores deposited in utero to sustain energy needs until feeding is established.¹⁹ Epinephrine and glucagon concentrations increase 3- to 5-fold within minutes and can act quickly via cyclic adenosine monophosphate (cAMP) to mobilize glucose by glycogenolysis. GH and cortisol concentrations in blood also are high in the initial hours after birth. Cortisol promotes gluconeogenesis, which takes several hours to be initiated and several days before it is fully established, whereas GH in concert with epinephrine acts to induce lipolysis with provision of fatty acids for oxidation, and the high concentrations of glucagon induce ketogenesis in mitochondria. A key enzyme for gluconeogenesis, phosphoenol pyruvate carboxykinase (PEPCK) is not expressed at birth and, under the influence of glucocorticoids, matures by about 24 hours.^{19, 20} Likewise, a key enzyme for ketogenesis, carnitine palmitoyl transferase 1 (CPT-1) is not expressed initially, and transcription is activated in part by long-chain fatty acids that are contained in colostrum, so that expression is evident by about 24 hours.^{20, 21}

Thus, immediately after birth, all infants initially must rely on glucose derived from glycogen and calories received from feedings of colostrum or milk. Moreover, whereas the four energy-mobilizing hormones are increased, insulin concentrations initially decline in the first few hours after delivery, and insulin secretion responds only sluggishly to glucose stimulation. As a result, the pattern in a normal full-term infant is for glucose to decrease to approximately 50 mg/dL in the first few hours after delivery and to return to average concentrations of 60 mg/dL or higher by day 2 and certainly by day 3 of life.⁵ Thus, conditions that diminish nutrient reserves for mobilization, such as intrauterine growth retardation or premature birth, will impair the ability to maintain glucose concentration that is appropriate for a normal newborn and hence predispose to hypoglycemia.

It has long been known that, in the first 8 hours after birth, almost one-third of term infants who are adequate for gestational age (AGA) and almost one-half of preterm or immature infants will have blood glucose concentrations less than 50 mg/dL.²² These are the normal adaptations in the transition from intrauterine life and dependence on maternal glucose supply to extrauterine existence with integration of glucose homeostasis during feeding and fasting that may be associated with “transitional” forms of hypoglycemia during adaptation. However, by day 3 of life, virtually no term AGA infant and less than 1% of all those in other categories have a plasma glucose concentration below 50 mg/dL.^{5, 22}

Hypoglycemia, as we have defined it, represents a mismatch between the ability to provide glucose either entirely by endogenous production or supplemented by feeding and factors that enhance its utilization. Hence, the ability to maintain glucose concentration may be impaired if feeding is delayed; if enzymatic pathways are not established fully, as occurs in the premature infant; or if there is transient persistence of excessive insulin secretion, as occurs in infants born to mothers with diabetes during pregnancy.²³ Hypoglycemia will persist if there are defects in the ability to metabolize glycogen to glucose, to initiate gluconeogenesis, or to convert fatty acids to ketones as a source of energy. Higher-than-normal glucose utilization rates can be anticipated with conditions such as sepsis, fever, or seizures, all of which also prevent normal feeding, or if hypoxia or other intrauterine stress has depleted glycogen stores. However, a glucose utilization rate in excess of 10 mg/kg/min, as reflected in the glucose infusion rate needed to maintain blood glucose above 60 mg/dL, almost invariably indicates a state of HI.^{1, 2, 16, 17, and 18} In an otherwise-healthy full-term AGA infant, delay in initiating feeding via breast or bottle does not in itself cause hypoglycemia, but it hastens the unmasking of an underlying defect, such as HI, deficiency of cortisol or GH, or even more rarely, severe defects in gluconeogenesis or fatty acid oxidation.^{23, 24, and 25} Less-severe disorders of glycogen breakdown and gluconeogenesis are not likely to manifest while the baby is being fed at 2- to 4-hour intervals, and with rare exceptions the same holds for disorders of fatty acid oxidation. These tend to be manifested later when the interval between feedings is lengthened (eg, at weaning), and the increased period of fasting unmasks an inborn error of metabolism. GH deficiency and cortisol deficiency, either as isolated defects or as part of hypopituitarism,²⁶ also may be masked by frequent feedings and unmasked by fasting.^{24, 25, and 26} The same holds true for transient or permanent HI, the most important and frequent cause of persistent hypoglycemia in a newborn child.^{1, 2} The more severe the defect, the sooner the clinical manifestations will appear in an otherwise-normal full-term newborn even with a normal feeding schedule.

In the absence of predisposing factors such as diabetes mellitus in the mother, premature birth, or intrauterine growth retardation, hypoglycemia in the first hours or days of life is almost always caused by deficiencies of GH or cortisol,^{24, 25} HI associated with perinatal asphyxia/hypoxia,^{27, 28} or HI due to genetic defects in the regulation of insulin secretion, manifesting as “fasting hypoglycemia” when the baby cannot feed or the interval between feedings is lengthened.^{29, 30, and 31}

Mechanisms of Central Nervous System Damage

Neuroglycopenia of short duration may be reversible. However, prolonged severe hypoglycemia, when glucose concentration is 1 mM (18 mg/dL) or less, results in neuronal death from cellular energy failure and the release of the excitatory amino acid aspartate into the extracellular space. Interaction of aspartate with its receptor on central nervous system (CNS) cells leads to membrane disruption, influx of calcium, activation of enzymes such as phospholipase, changes in cellular redox state, and cell death primarily affecting the posterior cerebral cortex but sparing the cerebellum and brainstem.³² In addition, hypoglycemia is reported to activate A1 adenosine receptors, permitting influx of calcium and activation of the proapoptotic enzyme caspase-3; blockade of adenosine A1 receptors ameliorated neuronal damage in the mouse model used.^{33, 34} These findings suggest the possible future use of adenosine A1 receptor antagonists to mitigate neuronal damage in newborns with severe hypoglycemia.^{33, 34} They also may explain the patterns of injury in severe hypoglycemia reported in magnetic resonance imaging (MRI) studies.

Magnetic Resonance Imaging in Neonatal Hypoglycemia

Some investigators reported specific patterns of MRI changes in newborns with severe hypoglycemia that differed from findings reported in hypoxic-ischemic brain injury. Diffuse cortical and subcortical white matter damage affecting primarily the parietal and occipital lobes is typically seen with severe hypoglycemia of less than 20 mg/dL and may be associated with bilateral occipital lobe cortical atrophy, a possible distinguishing feature not seen in hypoxic-ischemic brain injury. These changes may later manifest as cortical blindness. In some reports, changes observed initially seemed to resolve over time.^{35, 36, 37, and 38}

Diagnostic Approach

The symptoms and signs of hypoglycemia in a newborn are predominantly neurological, due to neuroglycopenia, as listed in Table 45-1. Most are nonspecific and could be caused by other conditions, such as sepsis in the newborn. Hence, eternal vigilance and consideration for the possibility of hypoglycemia as the cause of the symptom are keys to early diagnosis and favorable outcome. Therefore, we recommend that hypoglycemia be considered in any newborn with the symptoms listed. A blood glucose concentration of less than 40 mg/dL on days 1–2 of life, less than 50 mg/dL on day 3, and less than 60 mg/dL thereafter should be followed at a minimum by a formal blood glucose measurement in a laboratory together with the concentration of FFAs and β-hydroxybutyrate. If all 3 are suppressed, HI is the most likely diagnosis. If glucose is low but β-hydroxybutyrate and FFAs are normal or elevated, a defect in the ability to produce glucose is most likely. If the concentrations of both glucose and β-hydroxybutyrate are low but that of FFA is elevated, then a defect in fatty acid oxidation should be suspected. Sufficient blood should be made available to measure insulin, GH, and cortisol; a more detailed analysis of the “critical sample” taken at the time of hypoglycemia is discussed in the section on the diagnostic approach and treatment.

CLASSIFICATION OF NEONATAL HYPOGLYCEMIA

A classification of neonatal hypoglycemia is presented in Table 45-2. Routine screening of blood glucose concentration is not indicated in a healthy term infant after a normal pregnancy and delivery and with no hint of perinatal asphyxia. Blood glucose measurements and monitoring are promptly indicated in any infant who manifests the signs and symptoms listed in Table 45-1. Such infants also require measurements of FFAs and β-hydroxybutyrate and determination of insulin, GH, and cortisol.

Transitional Hypoglycemia (Days)

Transitional hypoglycemia that resolves within days occurs in newborn infants who are SGA or born prematurely. Such newborns should be monitored for hypoglycemia (<40 mg/dL) within 4 hours of birth and treated by feeding and intravenous glucose as needed to maintain plasma glucose concentrations equal to or greater than 45 mg/dL before each feeding. The target glucose should be greater than 45 mg/dL before routine feedings during the first 4 to 48 hours of life and greater than 50–60 mg/dL thereafter. Late preterm infants, defined as infants at 34 to 37 weeks' gestation, are likely to resolve their hypoglycemia within days as they have transitional hypoglycemia due to delay in enzyme maturation and diminished nutrient reserves.

Infants born to mothers with diabetes pre-dating pregnancy or developing during pregnancy also should be monitored as they are likely to need frequent feeding or continuous intravenous glucose infusion at rates of 5 to 7 mg/kg/min. These infants typically are LGA and have a surfeit of nutrients that they cannot mobilize due to hyperinsulinemia and diminished glucagon. Also glycogen may have been depleted if there was prolonged labor resulting in stress-activated epinephrine secretion. Meticulous glycemic control during pregnancy minimizes macrosomia, hypoglycemia, and other perinatal problems.^{39, 40} During labor and delivery of a pregnant woman with diabetes mellitus, intravenous glucose infusion should be avoided if possible as glucose crosses the placenta to the fetus and will stimulate insulin secretion that predisposes to hypoglycemia after placental separation. Treatment of such newborn infants also should avoid bolus glucose infusions; they stimulate surges of insulin secretion and cause rebound hypoglycemia. Most infants of diabetic mothers are stabilized within 3 to 5 days of birth using frequent feedings supplemented by intravenous glucose.

In the absence of a history of documented maternal diabetes, infants born LGA should be suspected of having HI and monitored accordingly. Such infants, along with infants who are SGA, should not be discharged from the hospital without demonstrating that their blood glucose is maintained at equal to or greater than 60 mg/dL at least 4 hours after a feeding for 3 feeding cycles. Blood glucose should be measured one-half hour before the next scheduled feeding. If feeding cannot be established and intravenous glucose infusion rates exceed 5–7 mg/kg/min to maintain a glucose concentration above 60 mg/dL, HI should be suspected. If glucose infusion rates of 10 to 12 mg/kg/min or greater are necessary to maintain blood glucose greater than 60 mg/dL, the cause is most likely HI.

Transient Hypoglycemia (Days–Weeks)

Transient hypoglycemia due to HI that lasts days to weeks is also reported in newborn infants with a history of perinatal stress, including maternal hypertension, preeclampsia, intrauterine growth retardation (IUGR), hypoxia, and cesarean delivery.^{27, 28} Onset may be sudden, unanticipated, and severe, with glucose concentrations plunging to levels below 20 mg/dL; these concentrations are associated with subsequent neurological devastation. Onset may occur after discharge from the newborn nursery and may be heralded by the baby having “feeding difficulties.” Insulin concentrations are modestly elevated but inappropriate for the degree of hypoglycemia; FFA and β-hydroxybutyrate concentrations are lower than normal. The majority of affected infants respond to diazoxide at a dose of 5–15 mg/kg/d, with the median effective dose 8 mg/kg/d in 1 reported series.²⁷ Some 20% of such affected newborns require frequent feedings or supplementation with intravenous glucose until hypoglycemia resolves. A variant of this hyperinsulinemic syndrome is associated with lactic acidosis and is thought to be associated with defects in the pyruvate dehydrogenase complex.²⁸ Diazoxide is the treatment of choice, and resolution usually occurs within 3–4 weeks. The mechanism responsible for malfunction of insulin secretion in these syndromes is unknown but likely involves the potassium channel regulated by adenosine triphosphate (K_ATP).

Persistent Hypoglycemia in Infancy

The causes of persistent hypoglycemia in infancy are listed in Tables 45-2 and 45-3. Their onset may be in the newborn period, but unlike the causes described previously, they persist if left untreated and can lead to neurological sequelae. Three separate reviews in the literature all pointed to early onset, prolonged hypoglycemia, and inability to achieve target blood glucose concentrations of greater than 60 mg/dL by medical means alone as poor prognostic signs.^{10, 11, and 12}

The persistent forms of hypoglycemia in infancy include deficiency of the counterregulatory hormones GH and cortisol, either alone or in combination, as occurs in hypopituitarism. Clues for the diagnosis of hypopituitarism include midline facial defects such as cleft palate, holoprosencephaly, and nystagmus as a manifestation of the septo-optic dysplasia syndrome.²⁶ Jaundice with both cholestatic and inflammatory features is frequently found in hypopituitarism and is likely due to GH and cortisol deficiency; treatment with GH plus cortisol reverses the hepatic defects.⁴¹ In boys, micropenis and small undescended testes provide important clinical clues that gonadotropin deficiency also exists. In normal full-term boys, testosterone concentrations in the first few days of life are as high as they are in the midteen years, and GH concentrations average 30–40 ng/dL in the first days of life. Therefore, stimulation tests for GH or testosterone are unnecessary; low concentrations in association with hypoglycemia are strong clues to the existence of hypopituitarism. If prolactin is elevated while GH, cortisol, and testosterone are low, the primary defect is in the hypothalamus, which normally restrains prolactin through the secretion of dopamine, previously called prolactin-inhibiting factor (PIF). MRI with contrast may reveal an ectopic posterior pituitary bright spot with interruption or absence of the pituitary stalk.^{26, 42} The deficiencies of GH and cortisol permit the unopposed action of insulin so that hypoglycemia occurs during fasting, which may represent only brief periods beyond the customary 2 to 3 hours between feedings. Cortisol should be replaced at 10 to 15 mg/m²/d, and GH may be given at a dose of 20–40 μg/kg/d if hypoglycemia persists and GH deficiency is documented.

Clues to the existence of an adrenal disorder causing cortisol deficiency may include ambiguous genitalia, hyponatremia, hyperkalemia, and metabolic acidosis. These findings are secondary to lack of aldosterone, and abnormal patterns of steroid precursors define the site of the biosynthetic block in congenital adrenal hyperplasia, the most common of which is 21-OH deficiency.

Bilateral adrenal hemorrhage should be considered after difficult delivery; in boys, congenital adrenal hypoplasia, a contiguous gene defect on the X chromosome also associated with Duchenne muscular dystrophy, and deficiency of glycerol kinase should be considered. Elevated serum levels of creatine phosphokinase (CPK) and triglycerides rapidly identify this condition. Adrenoleukodystrophy, another X-linked condition, usually does not present in the newborn period. Congenital unresponsiveness to ACTH (corticotropin) and the triple A syndrome (alacrima, achalasia, adrenal insufficiency; also known as Allgrove syndrome) can affect both sexes and should be considered in the differential diagnosis of primary adrenal insufficiency (low cortisol with high ACTH) associated with hypoglycemia, although presentation does not generally occur in the newborn period.⁴³

Inborn errors of metabolism affecting glycogen breakdown or synthesis, gluconeogenesis, and fatty acid oxidation are unlikely to become apparent unless feeding is delayed for periods of much longer than the customary schedule of feedings every 2 to 4 hours. These defects generally become apparent after 3 to 6 months when the feeding schedules are progressively delayed and the child is being weaned so that the interval between feedings becomes 4 to 6 hours rather than the previous 2 to 4 hours.^{1, 2} Fatty acid oxidation defects, unless severe, do not manifest in the immediate newborn period when feedings are frequent. Difficulties with initiating breast-feeding may unmask any of the aforementioned entities, including severe fatty acid oxidation defects that can result in hypoglycemia.^{21, 23} Moreover, enhanced neonatal genetic screening now identifies infants with defects in fatty acid oxidation so that the diagnosis is known within days after discharge from the newborn nursery. Note again that all these syndromes are manifestations of the inability to adapt to fasting, and that there is an inverse relationship between the duration of fasting necessary to provoke hypoglycemia and the severity of the defect(s) in counterregulatory hormones, in glycogen breakdown, in gluconeogenesis, in fatty acid oxidation defects, and in HI.

Persistent Hyperinsulinemic Hypoglycemia of Infancy

Persistent hyperinsulinemic hypoglycemia of infancy is now generally known as congenital HI and is the most common form of persistent hypoglycemia of infancy, caused by one of several genetic entities (Table 45-3). Understanding the biochemical and molecular basis has revolutionized the care of the newborn with HI and is essential for appropriate management and optimum outcome in affected newborns.^{1, 16, 17, 29, 30, and 31} The biochemistry of HI, discussed in the previous material, is summarized in Table 45-4. The glycemic response to glucagon is a rapid means to confirm that the liver contains glycogen, whose breakdown is inhibited by the excess actions of insulin. This restraint can be overcome by the intravenous or intramuscular injection of 0.5–1.0 mg glucagon, resulting in a rise of glucose concentration greater than 30 mg/dL above the baseline 30 minutes after injection. In other fasting forms of hypoglycemia, liver glycogen stores are less responsive or unresponsive to glucagon due to a block in glycogenolysis, as occurs in glycogen storage disease, or because glycogen has been depleted by fasting, as occurs in defects of gluconeogenesis or fatty acid oxidation; hence, these entities respond with an increment in glucose less than 30 mg/dL.

A genetic basis for HI was suspected because of its familial nature, especially in certain ethnic groups, and distinct patterns of autosomal-recessive or autosomal-dominant transmission. Genetic linkage studies localized some forms of HI to chromosome 11.⁴⁴ In 1995, the sulfonylurea receptor 1 (SUR1) gene and its association with familial HI was discovered⁴⁵ and shortly thereafter was shown to be closely linked functionally to an inward-rectifying potassium channel (Kir6.2), which together constituted the K_ATP channel.⁴⁶ The genes for both the SUR1 (ABCC8) and Kir6.2 (KCNJ11) were closely associated on chromosome 11. Mutations in the K_ATP were rapidly shown to be responsible for several forms of HI, and several other genetic defects have subsequently been discovered,^{47, 48} permitting the construct of a model, summarized in Figure 45-1. As shown in Figure 45-1, the model explains the following:

1. It indicates the mechanism by which the chemical energy of glucose is converted to an electrical signal that acts like a rheostat to close channels and cause insulin secretion when ATP concentration rises after glucose or amino acids are metabolized and gradually decreases insulin secretion as ATP generation falls when glucose or amino concentration falls. Insulin secretion generally ceases altogether at glucose concentrations of less than 40–50mg/dL. The normal (active) state of the K_ATP channel is to be open; inability to close the channel results in diminished insulin secretion, which is responsible for neonatal diabetes or diabetes in later life depending on the severity of the defect. Thus, inactivating mutations of either ABCC8 or KCNJ11, which inappropriately maintain the channel closed even at low glucose concentrations, cause HI, whereas activating mutations in these genes, which keep the channel open despite high glucose concentration, cause neonatal diabetes.^{47, 48} Likewise, activating mutations of the enzyme glucokinase (GCK) will promote insulin secretion at low glucose concentrations and result in hypoglycemia.⁴⁹

2. The model shows the site and potential mechanism by which drugs such as sulfonylurea act to close the channel causing insulin secretion, whereas diazoxide opens the channel and therefore limits insulin secretion. Somatostatin, a hormone that inhibits various secretory processes in tissues, including the endocrine pancreas, may be effective in the short term to inhibit insulin secretion in HI,⁵⁰ but it should be used with caution as a case of necrotizing enterocolitis (NEC) has been attributed to its use.⁵¹

3. The model shows the mechanisms of amino acid–stimulated insulin secretion, including leucine-stimulated insulin secretion, in which leucine acts as an allosteric positive stimulus for the enzyme glutamate dehydrogenase (GDH), enhancing the metabolism of amino acids and explaining the entity previously known as leucine-sensitive hypoglycemia.^{52, 53}

4. The model shows locations within the cell of some of the other recognized genetic defects responsible for HI.⁵⁴

Despite these advances, the molecular basis for some 30% to 50% of HI remains unknown and under investigation.^{1, 16, 17, 31} It should be noted that there are different potassium channels in which mutations have been shown to be involved in neonatal human epilepsy,⁵⁵ in vasospastic angina,⁵⁶ and serving a protective role in hypoxia-induced generalized seizures.⁵⁷ The mechanisms responsible for hypoglycemia in mutations of short-chain hydroxyl-acyl dehydrogenase (SCHAD),⁵⁸ hepatic nuclear factor 4 alpha (HNF4α),⁵⁹ and mitochondrial uncoupling protein 2 (UCP2)⁶⁰ (Figure 45-1) are not completely defined, so these entities are only briefly mentioned. There is, however, evidence that the enzyme L-3-hydroxycacyl-coenzyme A dehydrogenase (HADH) (shown in Table 45-3), which is mutated in the clinical entity SCHAD, normally restrains the activity of the enzyme GDH, so that mutations in SCHAD remove this restraint and result in activation similar to the constitutive activation of the GDH syndrome. Major clinical relevance resides in the K_ATP channel itself with mutations in the ABCC8 gene, which specifies the protein SUR1, and the KCNJ11 gene, which specifies the protein Kir6.2 (potassium inward-rectifying channel); ABCC8 mutations are more common in HI.

Familial forms of K_ATP channel defects are common in some defined populations, with an incidence that may be as high as 1:2500 to 1:3000 live births.^{1, 31} Among the general population, however, the incidence of K_ATP mutations is only 1:30,000 to 1:50,000 live births, and most are sporadic rather than familial. Most of the sporadic mutations for the recessive forms of HI occur in the ABCC8 gene responsible for SUR1; mutations are predominantly located within certain sites of the molecule, particularly the nuclear-binding fold-2 and some transmembrane domains.⁴⁸ In patients of Ashkenazi Jewish descent, almost 90% of reported cases had 1 of 2 mutations in the nuclear-binding fold-2 region (3993-9G>A; ΔF1388). The severity of clinical manifestations depends on the site of the mutation and the extent of impaired function, which may be in part caused by the chance distribution of normal and abnormal subunits because each K_ATP channel contains 4 SUR1 and 4 Kir6.2 subunits (48) (Figure 45-1).

Among the sporadic forms, 2 types of histopathological lesions, a focal and a diffuse form, are found in approximately equal distribution.⁶¹ Distinguishing between the focal and diffuse forms is critical for guiding the surgeon preoperatively to perform a local excision of the focal form, which is curative for hypoglycemia, while retaining sufficient pancreatic endocrine tissue to enable normal insulin secretion and avoid hyperglycemia-diabetes in the future.^{61, 62} By contrast, the diffuse forms require near-total pancreatectomy to control hypoglycemia, which may not be curative of hypoglycemia and yet predispose to diabetes in later life.^{61, 62}

The focal forms of HI are characterized by focal adenomatous hyperplasia of islet-like cells with small β-cell nuclei packed closely together.⁶¹ This focal defect results from a “2-hit” mechanism of molecular changes.⁶³ The first consists of a loss of a maternal allele in the p15 region of chromosome 11, with the loss expressed only in certain regions of the pancreas. This locus on the maternal allele contains an imprinted region containing H19 and p57^KIP, which normally suppress growth.⁶³ The paternal allele contains a mutation in the ABCC8 gene in a locus normally associated with the expression of insulinlike growth factor 2 (IGF-2), which promotes growth but now is no longer balanced by the normal copy of the maternal allele that suppresses growth.⁶³ Hence, the focal lesion within the pancreas represents somatic reduction to homozygosity for a paternal ABCC8 mutation promoting growth within the affected β-cell area of the pancreas and causing focal hyperplasia. In the diffuse form of HI, heterozygosity is maintained within each cell because the lesion results from inheritance of both a paternal and a maternal mutation. Distinction between these two currently is made by positron emission tomographic (PET) scanning using ¹⁸F-L-dopa, which is taken up by pancreatic β cells, providing a remarkably accurate picture when overlain on a computed tomographic (CT) image and pinpointing the site and size of a lesion as focal or as diffuse (Figure 45-2); however, occasional difficulties in resolving these forms of diffuse or focal lesions remain.⁶²

Nevertheless, in most cases when a focal lesion is identified via PET scanning, the surgeon can be guided preoperatively with the extent of resection. Molecular diagnostic confirmation of the type of the lesion can be obtained by testing the surgical specimen for the presence of p57^KIP, the imprinted region of the maternal gene missing in the focal lesion, because it only contains the paternal allele with the ABCC8 mutation. The paternal copy of this region of chromosome 11 expresses the imprinted IGF-2 gene, which promotes growth without the restraint imposed by p57^KIP and adjacent H19 regions of the maternally imprinted genes, thereby explaining the mechanism for the focal hyperplasia.^{63, 64}

Alterations in these imprinted gene regions, including abnormal methylation patterns, are also responsible for the various defects found in Beckwith-Wiedemann syndrome, which also may be associated with perinatal hyperinsulinemic hypoglycemia.^{65, 66} Beckwith-Wiedemann syndrome is characterized by other abnormalities, including macroglossia, hemihypertrophy, distinctive earlobe fissures, omphalocele or other umbilical anomalies, and a propensity for tumors.^{65, 66}

The ability to distinguish the localized from the diffuse form preoperatively has revolutionized the care of the newborn or infant with HI and may require the transfer of infants affected by HI to centers capable of performing the PET scans and with the requisite surgical and histopathology expertise to successfully treat affected infants.^{67, 68}

AUTOSOMAL-DOMINANT HYPERINSULINISM

Patients who have autosomal-dominant HI tend to have milder clinical presentations than those who have the autosomal-recessive or focal lesions; these autosomal forms include activating mutations of GCK and GDH, as well as inactivating mutations in ABCC8 and KCNJ11.⁶⁹ Macrosomia is less common at birth, and symptoms may occur only after the neonatal period when weaning has occurred and the duration of fasting between feedings is extended to 6 to 8 hours. Some of these mutations do involve the K_ATP channel, primarily the potassium pore itself (ie, the Kir6.2 protein encoded by KCNJ11). Note again that the mutations of the K_ATP channel involving ABCC8 and KCNJ11 are inactivating mutations. By contrast, activating mutations of GCK and GDH cause neonatal HI^{69, 70} (see Figure 45-1).

An activating mutation of GCK causes HI because GCK serves as the sentinel of the β-cell glucose-sensing apparatus. Thus, inactivating mutations of GCK require a higher glucose concentration before adequate phosphorylation of glucose can occur; therefore, in the heterozygous state, they cause a form of monogenic diabetes of youth termed MODY 2 (mature-onset diabetes of the young). Homozygous inactivating mutations of GCK cause permanent neonatal diabetes. However, activating mutations of GCK permit glucose to be phosphorylated and lead to closure of the K_ATP channel at lower-than-customary glucose concentrations. Most of the GCK mutations present after the age of 3 to 6 months because milder mutations will be masked by the frequent feeding patterns of infancy. Hence, it would require a severe activating mutation of GCK to manifest as neonatal hypoglycemia, an event that is rare but reported. These activating mutations decrease the threshold for insulin release to approximately 2 mM in comparison to the normal 4 to 5 mM in normal subjects. Marked clinical heterogeneity has been reported in families with several members affected by the same mutation, demonstrating that the mutation itself cannot be the entire explanation for the heterogeneity within families.⁷¹

Another form of autosomal-dominant HI involves an activating mutation of the GDH gene (Figure 45-1). Here, increased activity of the enzyme in β cells leads to increased glutamate oxidation, generating ATP, which induces insulin secretion, as illustrated in the figure. Leucine is an allosteric activator of this enzyme, whereas normally, guanosine triphosphate (GTP) inhibits its activity. This is the mechanism responsible for the previously termed condition leucine-sensitive hypoglycemia and could manifest in the newborn period if infants are fed formula milk.⁷² Generally, such affected infants present in the first year of life with milder episodes of hypoglycemia that generally respond to diazoxide. A biochemical feature is the presence of moderate hyperammonemia, in the range of 80 to 200 μmol/L, but patients lack symptoms of hyperammonemia, such as lethargy and coma, found with the inborn errors of the urea cycle. A report of a patient with GDH mutation and with seizures in the absence of hypoglycemia suggests that this mutation may induce hyperexcitability of the CNS. Clearly, in leucine-sensitive hypoglycemia, dietary restriction of L-leucine-containing nutrients should be used in conjunction with diazoxide to avoid hypoglycemia.⁷²

A third form of activating mutation responsible for HI in the newborn is HNF4α.⁷³ HNF4α is a transcription factor normally involved in the secretion of insulin, so that affected individuals present later in life with a form of monogenic diabetes, labeled MODY1. Nevertheless, a number of patients with this entity have been described as having both macrosomia and hypoglycemia at birth and only later in life proceeding to hyperglycemia and diabetes.⁵⁹ Hence, a family history of hypoglycemia in infancy followed by hyperglycemia later in life should alert the clinician to the possibility of this entity. A similar pattern of HI representing a dominant heterozygous mutation E1506K in the ABCC8 gene was described in a Finnish family; this pattern was characterized by hypoglycemia in infancy, loss of insulin secretory capacity in early adulthood, and diabetes in middle age.⁷⁴

Recently, an activating mutation in the AKT-2 component of the insulin-signaling pathway was reported in 3 subjects whose metabolic profile was consistent with a hyperinsulinemic effect but in whom insulin concentrations were below the limits of detection.^{18, 70} Thus, an activating mutation in the insulin receptor or its transduction-signaling cascade can result in the same phenotype as excessive insulin secretion.^{17, 18, 70} The mechanism is fundamentally the same: increased insulin action, in this instance due to activation of the insulin receptor effect despite low insulin levels. It is likely that other similar patients will be described in the future.

OTHER CAUSES OF NEONATAL HYPOGLYCEMIA

Hypoglycemia is a feature of the hyperviscosity syndrome seen in neonates with a hematocrit of 65% or higher. Mechanisms are unclear, with excessive glycolysis in red blood cells implicated; treatment is symptomatic.^{75, 76} Hyperglycemia also occurs frequently in neonatal sepsis when increased metabolic rates require glucose production that cannot be sustained, especially in the absence of feeding, which may be limited by the ongoing septic process; treatment is symptomatic.⁷⁷ Maternal use of propranolol or salicylates may inhibit gluconeogenesis in the mother and fetus, hence leading to hypoglycemia in the newborn.^{78, 79} Drug-induced hypoglycemia with insulin, or oral hypoglycemic agents, is virtually unknown in the newborn infant but might result from a pharmaceutical therapeutic error.⁸⁰ Distinction between these possibilities relies on the fact that exogenous insulin administration suppresses endogenous insulin secretion so that C-peptide levels will be low, whereas with an insulin-stimulating agent such as a sulfonylurea, both insulin and C-peptide are elevated.⁸¹

A defect in the Glut-1 glucose transporter has been described in 2 newborn infants with a seizure disorder.⁸² Their cerebrospinal fluid (CSF) glucose concentration was low despite normal plasma glucose concentration; CSF lactate also was low, indicating that this was not a bacterial septic process.⁸² By definition, these children did not have hypoglycemia, although they did have hypoglycorrhachia. These infants were successfully treated with a ketogenic diet, which reduced the severity of seizures by providing ketones to the brain as a fuel source capable of bypassing the metabolic effect of diminished glucose transport into the brain.⁸²

Hyperinsulinism has been described as a presenting sign in phosphomannose isomerase deficiency.⁸³ However, this manifestation of a carbohydrate-deficient glycoprotein syndrome, treatable by oral mannose supplementation, did not present until 3 months of age and therefore is not, strictly speaking, a disorder of the newborn.

Galactosemia may be associated with hypoglycemia after milk feeding and can be diagnosed by the presence of reducing substances but not glucose in the urine, by the presence of hepatomegaly and jaundice, and by enzyme activity of Gal1P-uridyltransferase measured as part of neonatal screening programs.^{84, 85}

DIAGNOSTIC APPROACH AND TREATMENT

An algorithm for approaching the newborn with suspected hypoglycemia is shown in Figures 45-3 and 45-4. Clearly, the appropriate treatment of neonatal hypoglycemia depends on the establishment of the specific diagnosis, and making that diagnosis is facilitated by careful review of the history during labor and delivery; by noting any of the nonspecific symptoms or signs listed in Table 45-1 and considering hypoglycemia to be their cause; by obtaining a family history, which may provide clues to the diagnosis of a familial disorder; by appropriate physical examination (noting jaundice, midline facial defects, umbilical hernia/omphalocele, microphallus or cryptorchidism in a male); and by performing key laboratory investigations, including the metabolic fuel profile at the time of hypoglycemia and key hormone measurements. This constitutes the critical sample, with emphasis on the concentration of glucose, β-hydroxybutyrate, and FFAs as well as insulin, GH, and cortisol, as listed in Figure 45-5. A single low level of cortisol or GH in the absence of other confirmatory features may not be diagnostic of adrenal-pituitary disease, especially beyond the first few days of life.⁸⁶ In the newborn, GH is constitutively high, averaging approximately 40 ng/dL in the initial days of life, so values below 10 ng/mL suggest GH deficiency in the newborn.

When hypoglycemia is considered, a bedside point-of-care glucose measurement is performed, and values below 50 mg/dL on days 1 and 2 or less than 60 mg/dL thereafter are confirmed via a laboratory glucose measurement before treatment is undertaken. In the absence of macrosomia, perinatal asphyxia, or features of hypopituitarism, it is reasonable to treat an infant, particularly if premature, with 0.5 g glucose per kilogram given as a bolus over 5 minutes in the form of 5 mL 10% dextrose in water per kilogram and retesting the infant 30 minutes later. Persistence of hypoglycemia should be treated via an infusion of intravenous glucose providing 4–8 mg/kg/min and glucose monitored hourly. Recurrence or persistence of hypoglycemia despite these measures or the need to increase the glucose infusion rate to 10 mg/kg/min or greater to maintain blood glucose above 60 mg/dL should alert the clinician to the possibility of hypopituitarism or HI and require the performance of a critical sample as described previously and in Table 45-4.

A critical sample should be taken immediately from any infant with glucose below 50 mg/dL on day 3 of life; the metabolic profile and hormone measurements may define the problem as hypopituitarism, which can be treated with cortisol 10–15 mg/m²/d in 3 divided doses or GH 20–40 μg/kg/d given by subcutaneous injection once daily. MRI of the brain and pituitary should be performed to define the lesion(s),^{26, 87} and biochemical/genetic testing should be undertaken to examine the cause of primary adrenal insufficiency determined by the presence of high plasma concentrations of ACTH, in contrast to the low concentration of ACTH in hypopituitarism.⁴³ Consultation with a pediatric endocrinologist facilitates diagnosis and management.

Hyperinsulinism should be suspected as the likely cause of hypoglycemia in a full-tem infant without any predisposing factors and certainly if the infant is macrosomic without confirmed maternal diabetes, if the infant has a glucose level below 20 mg/dL at any time, or if there has been perinatal asphyxia. The critical sample should reveal low glucose plus low FFA plus low β-hydroxybutyrate. Defects in glucose production via glycogen breakdown or gluconeogenesis^{88, 89} will be revealed by normal or elevated β-hydroxybutyrate and lactate, whereas defects in fatty acid oxidation are likely if FFA levels are elevated while β-hydroxybutyrate is low; these should become apparent only in an infant who is not feeding well because normal feeding beyond day 2 of life should avoid these manifestations when there is available glucose from feedings every 2–3 hours. In these circumstances, the acylcarnitine profile and genetic testing become essential to establish the defect and manage accordingly.^{21, 90}

Hyperinsulinism, the most common cause of persistent neonatal hypoglycemia, is managed medically, with surgery reserved for those who do not respond to medical management.³¹

Medical Management

Initial medical management is the provision of glucose by intravenous infusion at rates to maintain a glucose level above 60 mg/dL; this may require a central line and rates of 10–20 mg/kg/min (occasionally higher) or supplemental milk feedings “thickened” with polycose to provide more than the customary 20 calories/ounce. Diazoxide, a drug that inhibits insulin secretion by maintaining the K_ATP channel in an open state, can be used at 5–20 mg/kg/d given orally in 3 divided doses; increments should be made at 2-day intervals to permit equilibration of the drug's long half-life and metabolic effects. Side effects include fluid retention, which may precipitate congestive heart failure if cardiac reserve is already impaired; thrombocytopenia, neutropenia, anemia, hyperuricemia, and pancreatitis also are described but occur only rarely. In long-term use, hirsutism most evident on the back and limbs as well as coarsening of facial features occur but are reversible with discontinuation of the drug. Doses above 15 mg/kg/d require more careful monitoring of electrolyte and hematological side effects. Fluid retention can be minimized by the concurrent use of oral chlorothiazide at doses of 7–10 mg/kg/d given in 2 divided doses.

If these measures do not control blood glucose above 60 mg/dL, temporary use of glucagon, infused intravenously or subcutaneously at a constant rate to deliver 1 mg per day (24 hours), may help to maintain the desired glucose concentration because glycogen is available for release and the restraint imparted by the excess insulin is overcome by the opposing glycogenolytic effect of glucagon. Another possibility is the use of octreotide, a long-acting somatostatin analogue given at a starting dose of 3–5 μg/kg/d divided into subcutaneous injections every 6–8 hours. Rapid desensitization to its effect of inhibiting insulin secretion requires gradual increases in the dose to a maximum of 15 μg/kg/d. In addition, NEC has been reported with its use, so caution is strongly recommended.⁵¹ While these measures are being provided, molecular diagnostic testing should be performed on the patient and parents in a laboratory certified by CLIA (Clinical Laboratory Improvement Amendments)⁹¹ while retaining some DNA for potential investigation in research laboratories for mutations yet to be discovered.

Mutations in GDH or GCK often can be managed medically and hence have a better prognosis for normal intellectual development; they may be new mutations or transmitted in an autosomal-dominant mode from a parent. Heterozygous mutations of the ABCC8 or KCNJ11 genes in both parents suggest the infant has homozygous or compound heterozygous mutations, which are associated with diffuse pancreatic hyperplasia and likely will require subtotal pancreatectomy as medical measures often fail. A paternal mutation without a maternal mutation raises the possibility that the infant may have focal disease, which needs to be confirmed by ¹⁸F-L-dopa PET scanning, which may then permit curative local excision. Single-parental mutation also suggests the possibility of an autosomal-dominant form of HI resulting from these K_ATP genes, and these also tend to have milder features with good response to medical measures. There are only a limited number of centers capable of performing ¹⁸F-L-dopa scans at present.

Success with medical management to enable discharge from the hospital is defined by the ability to maintain blood glucose above 60 mg/dL with diazoxide plus feeding of milk or fortified formula without the use of intravenous glucose infusion; the infant should be able to maintain glucose above 60 mg/dL even if a feeding is missed (ie, 4–8 hours of fasting on 2 occasions). The use of cornstarch in the management of HI is generally considered ineffective.

Failure of medical treatment is assumed if maximal doses of diazoxide with enhanced feeding and the use of temporary glucagon infusion fail to consistently maintain blood glucose at greater than 60 mg/dL. In general, the younger the onset of the clinical manifestations and the more severe and prolonged the hypoglycemia, the more likely it is that the infant will not respond to medical measures and will require definitive therapeutic surgery to avoid neurodevelopmental impairment. The decision regarding the need for surgery, or transfer to a unit capable of performing the requisite investigation and surgical management, should not be deferred if hypoglycemia remains a recurring problem in the neonatal intensive care unit (NICU). Medical approaches succeed in approximately 20%–30% of cases; surgery is required in the majority, especially when the cases first become apparent in the NICU.^{31, 71}

Surgical Management

Surgery must be undertaken in institutions with trained pediatric surgical and anesthesia teams together with requisite expertise in histopathological interpretation of biopsy specimens obtained during the operation, which may guide the extent of surgical resection.^{64, 67, 68} Laparoscopic approaches enable resection of preoperatively defined focal lesions that result in cure of hypoglycemia and preservation of exocrine and endocrine function; traditional open abdominal approaches are needed for suspected diffuse lesions. Remarkably, despite wide resection of the pancreas in diffuse lesions, hypoglycemia may persist or recur, and pancreatic exocrine function remains paradoxically normal; these phenomena are not understood. Hyperglycemia or frank diabetes are more likely after extensive pancreatic resections but may not be apparent until puberty with its attendant resistance to insulin.⁶¹

FUTURE DIRECTIONS

Improvements in laparoscopic instrumentation, particularly in video magnification, may permit pancreatic biopsies as a means to more rapidly identify and distinguish focal as opposed to diffuse disease without the need for preoperative PET scanning via ¹⁸F-L-dopa. The era of whole-genome or exome sequencing, soon to be ushered in to routine practice, will permit identification of the genetic defect(s) of the 50%–60% of newborns with HI in whom the cause is currently unknown.⁹² It is conceivable that pancreatic β cells isolated from the resected pancreas of patients with diffuse hyperplasia could be cultured, stored, and reimplanted at some future date into the same patient to “cure” their diabetes because rejection of their own tissue would not be a consideration.

REFERENCES

DEFINITION/CLASSIFICATION

Congenital adrenal hyperplasia (CAH) is a family of inherited, autosomal recessive disorders of adrenal steroidogenesis due to an abnormality in a step necessary for conversion of cholesterol to cortisol in the adrenal cortex (Figure 46-1). Of all CAH cases, 95% are caused by 21-hydroxylase deficiency. There are 2 forms of 21-hydroxylase deficiency: classic and nonclassic CAH (also called late-onset CAH). Classic CAH comprises the salt-wasting form (cortisol deficiency and aldosterone deficiency) and simple virilizing form. The salt-wasting form comprises 75% and simple virilizing form 25% of patients with 21-hydroxylase deficiency. The classic form is associated with severe enzyme deficiency, leading to prenatal virilization in girls; the nonclassic form has mild enzyme deficiency that causes postnatal hyperandrogenism but no prenatal virilization.

ETIOLOGY

The most common form of CAH is caused by mutations in CYP21A2, the gene encoding the adrenal steroid 21-hydroxylase enzyme (P450c21).¹ The steroid 21-hydroxylase enzyme catalyzes conversion of 17-hydroxyprogesterone (17-OHP) to 11-deoxycortisol and progesterone to deoxycorticosterone, which are precursors of cortisol and aldosterone, respectively. In 21-hydroxylase deficiency, because of deficient cortisol synthesis, progesterone and 17-hydroxyprogesterone are shunted to the androgen synthetic pathway (androstenedione and testosterone). These adrenal androgens are oversecreted in utero, resulting in variable degrees of virilization in the female fetus.

EPIDEMIOLOGY

The incidence of classic CAH in the general population, as shown by newborn screening in different populations worldwide, ranges from 1:10,000 to 1:20,000, with an overall incidence of 1:15,000.^{2, 3} The incidence of CAH in specific populations ranges from 1 in 5000 live births in Saudi Arabia to 1 in 21,270 live births in New Zealand. The newborn screening incidence in the United States and Canada is 1 in 14,203 live births, with Brazil at 1 in 1863 live births and Japan at 1 in 18,827. A high frequency of CAH exists among Yupik Eskimos from western Alaska: 1 in 282 live births.

The prevalence of nonclassic 21-hydroxylase deficiency CAH in the general heterogeneous population of New York City was estimated to be 1 in 100. Ashkenazi Jews have the highest prevalence at 1/27. Other ethnic groups with a high prevalence of nonclassic CAH include Hispanics (1/40), Slavs (1/50), and Italians (1/300).¹

CLINICAL MANIFESTATIONS

The cardinal feature of classic CAH in newborn females is ambiguous genitalia (Figure 46-2). Clitoromegaly occurs as a result of adrenal androgens binding to genital skin androgen receptors. Females may present with a urogenital sinus (Figure 46-3). Urogenital sinus due to high levels of circulating adrenal androgens beginning at about 7 weeks’ gestation prevent the formation of separate vaginal and urethral canals. Females with classic CAH have normal Müllerian structures (uterus, fallopian tubes). The girl with classic CAH has ambiguous or male-appearing external genitalia with perineal hypospadia, chordee, and cryptorchidism. The severity of virilization is graded using Prader staging, a 5-point scale developed by Prader, with stage 1 indicating mild clitoromegaly and stage 5 representing a cryptorchid male with penile urethra.

Males with classic CAH do not have ambiguous genitalia but may present with subtle scrotal hyperpigmentation because of corticotropin (ACTH) hypersecretion or phallic enlargement because of excessive adrenal androgens in utero.

Neonates with classic salt-wasting CAH typically manifest salt-wasting crisis during the first 1–3 weeks of life. Salt-wasting crisis may occur as late as a few months of age. This manifests as poor feeding, vomiting, weak cry, failure to thrive, loose stools, dehydration, and lethargy. The biochemical sine qua non of salt-wasting crisis is hyponatremia, hyperkalemia, and metabolic acidosis with or without hypoglycemia. Without treatment, circulatory collapse, arrhythmias from hyperkalemia, and death may occur within days or weeks.

Postnatally in untreated boys and girls, excessive adrenal androgen secretion leads to rapid somatic growth, bone age advancement, progressive penile or clitoral enlargement, and premature appearance of pubic hair, axillary or facial hair, and acne. Without cortisol treatment, early epiphyseal closure occurs as a result of aromatization of adrenal androgens (androstenedione and testosterone to estrone and estradiol, respectively) and results in short adult height.

DIAGNOSIS

CAH is the most common cause of ambiguous genitalia in the newborn, especially if no gonads are palpable in the scrotum or inguinal canal. Timely diagnosis is needed to prevent sex misassignment of a female newborn as male. If gonads are palpable in the scrotum or inguinal canal, workup for 46 XY disorder of sex differentiation or other causes of ambiguous genitalia should be pursued.

Immediate steroid treatment is necessary to prevent salt-wasting crisis in the salt-wasting form of CAH. Karyotype will identify chromosomal sex, and pelvic ultrasonography will help identify Müllerian structures (ovaries and uterus). Consultation with an endocrinologist should be sought. Adrenal steroids should be measured by blood sample on day 2–3 of life. A markedly elevated level of 17-OHP (>10,000 ng/dL or > 300 nmol/L) in a virilized 46,XX female with nonpalpable gonads is most likely due to CAH.

Diagnosis of 21-hydroxylase deficiency is most accurately established by measuring 17-OHP before and 30 or 60 minutes after intravenous injection of 0.25 mg of Cosyntropin (ACTH1-24). Basal and 60-minute 17-OHP are plotted on a nomogram and distinguish normal individuals and patients with nonclassical and classical CAH.

In affected females, androstenedione and testosterone levels are high. In males, testosterone is usually elevated because of minipuberty in the first 6 months of life and is thus not helpful. ACTH levels are high but are not needed for diagnosis. In salt-wasting CAH, plasma renin is elevated and serum aldosterone is inappropriately low for the renin level.

TREATMENT

Cortisol deficiency is treated with glucocorticoids, usually in the form of hydrocortisone. Glucocorticoid treatment also suppresses high adrenal androgens, preventing further virilization, rapid somatic growth, and early epiphyseal fusion.⁴

Hydrocortisone tablets at a dose of 10–15 mg/m²/d in 2 or 3 divided doses are the preferred treatment. Infants with CAH require higher doses, 50 mg/m²/d divided in 2 to 4 doses during the first few weeks of life to suppress ACTH hypersecretion. Patients with salt-wasting CAH require mineralocorticoid replacement with fludrocortisone 0.1–0.3 mg daily in 1 or 2 divided doses and NaCl supplementation at 1–3 g/d. NaCl tablets are available, and a 1-g tablet is equivalent to 17 mEq of NaCl. Table salt (one-quarter teaspoon = 575 mg NaCl) can also be added to formula if NaCl tablets are not available.

Stress dosing (3–5 times the maintenance dose) is needed in febrile illness (>38.5°C), gastroenteritis with dehydration, surgery accompanied by general anesthesia, and major trauma. Parenteral Solu-Cortef^® (hydrocortisone sodium succinate) 25 mg IM or IV should be given for severe vomiting, dehydration, surgery, or severe trauma.

Glucocorticoid treatment of CAH patients involves the delicate balance of adrenal androgen suppression against iatrogenic Cushing syndrome. Glucocorticoid excess suppresses growth hormone secretion and impairs growth. Inadequate glucocorticoid causes high 17-OHP, which is shunted to androstenedione and testosterone; the latter is aromatized to estradiol, the principal hormone that closes epiphyses. An acceptable 17-OHP range in the treated patient is 100–1000 ng/dL.

SURGERY

Surgery for significantly virilized females early in childhood is no longer considered conventional treatment in many tertiary care centers. Surgical procedures have traditionally included both vaginoplasty to repair the urogenital sinus and reduction clitoroplasty to feminize the appearance of the clitoris. Vaginoplasty should be individualized and performed by an experienced surgeon, usually before puberty. Numerous studies have revealed poor sexual functioning in adult women with CAH who underwent feminizing surgery as a child, suggesting that clitoroplasty should be restricted to fully informed patients.⁵ A growing number of subspecialists are emphasizing functional outcome over cosmetic appearance. We recommend consideration of clitoroplasty be delayed until the age of consent, when the young woman can make an informed decision about the risks and benefits of surgery. In the meantime, parents can be assured that the severity of clitoromegaly will decrease with suppressive glucocorticoid therapy during the first 6–12 months of treatment.

PROGNOSIS

Untreated salt-wasting CAH can lead to circulatory collapse, arrhythmia, and death. Untreated nonclassical and simple virilizing CAH can lead to excessive adrenal androgen secretion, rapid bone age advancement, and short adult stature. The majority of females with classic CAH have female gender identity and behavior, and most women can bear healthy offspring, thus providing the rationale for female sex assignment.

PREVENTION

Because CAH is common and potentially fatal, it is a disease that is ideal for newborn screening. As of 2009, all 50 states in the United States and 12 other countries screened for CAH. Early recognition and treatment can prevent morbidity and mortality.

CAH is an autosomal recessive condition. If a mother has a previously affected child with CAH and is pregnant with a child from the same father, the fetus has a 1 in 4 chance of having CAH. Prenatal treatment of 21-hydroxylase deficiency with dexamethasone has been shown to prevent virilization should the fetus be an affected female.⁶ Dexamethasone at a dose of 20 μg/kg/d in 3 divided doses is administered to the pregnant mother before 10 weeks’ gestation. Dexamethasone does not bind cortisol-binding globulin in the maternal blood and is not metabolized by placental 11β-hydroxysteroid dehydrogenase; therefore, it crosses the placenta, suppresses fetal ACTH secretion, and reduces fetal androgen production. Prenatal therapy is considered experimental and should be pursued through protocols approved by institutional review boards at specialized centers.

ADRENAL INSUFFICIENCY IN THE PREMATURE INFANT

Hypocortisolemia in premature infants is related to immaturity of the hypothalamic-pituitary-adrenal (HPA) axis. Activation of the HPA axis with sufficient cortisol secretion is critical in maintaining hemodynamic homeostasis. Relative adrenal insufficiency (RAI) refers to inadequate cortisol production for the degree of stress or illness and has been described in acutely stressed adults, children, and premature and term newborns with cardiovascular instability and shock. RAI is a contributing factor to cardiovascular instability in the sick premature newborn. Many ill premature newborns have inadequate cortisol production due to immaturity of the HPA axis and lagging expression of adrenal steroidogenic enzymes.

Normal Adrenal Development

At 4–5 weeks’ gestation in humans, cells from the intermediate mesoderm migrate retroperitoneally to the upper pole of the mesonephros and develop into adrenal cortex. At 7–8 weeks, neural crest-derived sympathetic cells infiltrate the adrenal gland and give rise to the adrenal medulla. By 7 weeks’ gestation, the fetal hypothalamus is visible, and it contains releasing hormones by 8–10 weeks. ACTH secretion is evident in pituitary cells by 5–8 weeks. ACTH is suppressible by transplacental dexamethasone by 7–8 weeks, and the ACTH-adrenal feedback arc is functional by 9 weeks.

At 12 weeks, vascular connections to the pituitary are present, but the definitive portal system develops in the third trimester. ACTH is measurable in fetal plasma by 16 weeks. Corticotropin-releasing factor (CRF) from the hypothalamus stimulates pituitary ACTH release by 14 to 20 weeks. Therefore, the HPA axis has emerged by 20 weeks’ gestation.⁷

Adrenal Function in the Fetus and Premature Infant

The adrenal cortex in preterm newborns does not express 3-β-hydroxysteroid dehydrogenase (3β-HSD) before 23 weeks and thus has limited capacity for de novo cortisol synthesis until 30 weeks.⁸ However, prior to 30 weeks, the fetal adrenal gland uses placental progesterone to bypass 3β-HSD and produce cortisol. Adrenal steroid precursor concentrations are higher in premature infants compared to term infants, indicating the need for 3β-HSD and 11-β-hydroxylase for cortisol synthesis.

The serum cortisol concentrations in ill preterm infants are variable due to intrinsic and extrinsic factors; therefore, assessment of serum cortisol concentration and definition of normal and insufficient amounts are equally problematic. Intrinsic (eg, immaturity of the HPA axis, physiologic decline of serum cortisol in immediate postpartum period, effect of antenatal steroids on HPA axis) and extrinsic factors (eg, vaginal delivery, low Apgar scores, respiratory distress syndrome, mechanical ventilation, hypoglycemia, infection) affect serum cortisol concentrations during the first few weeks of life.⁹

Newborn infants lack mature circadian rhythm for cortisol regulation, but pulsatility of cortisol secretion exists, and random cortisol levels vary widely. Jett et al found cortisol levels of 2 to 54 μg/dL (55–1489 nmol/L) in extremely low birth weight (ELBW) infants less than 28 weeks, with minimal variability over time.¹⁰

For a more extensive discussion of the physiology and adrenal development of the fetus and premature newborn, see an excellent review by Watterberg.¹¹

Definition of Relative Adrenal Insufficiency in the Premature Infant

Relative or transient adrenal insufficiency in premature infants was initially described in abstracts by Colasurdo in 1989 and Ward in 1991 in ELBW infants presenting with “Addisonian”-like crisis: hypotension, oliguria, hyponatremia, hyperkalemia, and cortisol less than 15 μg/dL (<413 nmol/L). Since these initial reports, several case series have been published describing clinical variables and response to glucocorticoid treatment.¹²

Most infants treated with steroids were less than 30 weeks’ gestation and presented with volume and pressor-resistant hypotension either within a few hours of life or as late as 3 weeks of life. These infants rapidly responded to dexamethasone (0.25 mg/kg) or hydrocortisone (stress dose 25–60 mg/m²/d). The rise in blood pressure occurred as early as 30 minutes and within 8 hours to maintain normal mean arterial pressure without needing pressor support. However, hydrocortisone treatment failure occurred in some infants, suggesting other causes of hypotension. In some infants, a second course of stress steroids was needed due to persistent hypotension.

ACTH Testing in Premature Infants

Adrenal function testing in the premature infant is complicated by lack of consensus about the appropriate dose of cosyntropin (ACTH), different gestational and postnatal ages of ACTH testing, severity of illness, and criteria of appropriate cortisol response after ACTH stimulation. Different ACTH (1-24 corticotropin) doses (0.1 μg/kg, 1 μg/1.73 m², 0.2 μg/kg, 1 μg/kg, up to 36 μg/kg) have been administered to premature infants at different postnatal ages using different criteria for adequate cortisol response, yielding a wide range (10%–56%) of reported RAI.¹²

Watterberg’s study of cortisol response of 276 ELBW infants weighing 500–999 g comparing 1 μg/kg vs 0.1 μg/kg showed that the higher dose (1 μg/kg) yielded higher median cortisol values, 24.6 μg/dL (678 nmol/L) vs 17.4 μg/dL (480 nmol/L), and fewer negative responses (2% vs 21%).¹³ The 1-μg/kg dose showed different responses for girls vs boys, infants receiving enteral nutrition vs not, infants exposed to chorioamnionitis vs not, and those receiving mechanical ventilation vs not. A response curve for the 1-μg/kg dose for infants receiving enteral nutrition (surrogate marker of clinically well infants) showed a cortisol response at the 10th percentile of 16.96 μg/dL (468 nmol/L). Infants who stimulated less than 17 μg/dL (468 nmol/L) showed an increased incidence of bronchopulmonary dysplasia (BPD) and length of hospital stay.

Based on Watterberg’s study, we conclude that 1 μg/kg ACTH to test adrenal function is the most discriminating in neonates in whom adrenal insufficiency is suspected.

Management

There is presently no consensus and insufficient evidence to support the routine use of glucocorticoids in patients with suspected adrenal insufficiency of prematurity. Current therapy for refractory hypotension includes volume expanders and vasopressors.

Two randomized controlled trials showed the benefit of glucocorticoids for preventing and treating refractory hypotension during the first week of life.^{14, 15} These studies showed that this regimen effectively reduces the use of volume expansion, dopamine, and dobutamine for blood pressure support compared to placebo. Significantly more very low birth weight infants who were treated with hydrocortisone were weaned off pressors within 3 days.

A meta-analysis examined the relationship between hydrocortisone administration, hypotension, and vasopressor requirements and found that hydrocortisone significantly increased blood pressure and reduced vasopressor requirements in hypotensive and vasopressor-dependent preterm infants.¹⁶

A treatment algorithm has been proposed in critically ill infants with cardiovascular instability and hypotension⁸ (Figure 46-4). If hydrocortisone treatment is being considered for treatment of refractory hypotension, a serum cortisol value should be obtained before administering empiric hydrocortisone. Caution should be taken not to administer hydrocortisone treatment with indomethacin or ibuprofen to avoid the risk of spontaneous gastrointestinal perforation. If a random cortisol value is high (>15 to 20 μg/dL; 413–551 nmol/L), discontinue glucocorticoid if there has been no cardiovascular response. If serum cortisol is less than 15 μg/dL and the patient responds to glucocorticoid treatment, reassessment of cardiovascular function after 3–5 days of hydrocortisone treatment should be done.

The goal is to minimize prolonged treatment with hydrocortisone because the long-term clinical benefit of hydrocortisone therapy has yet to be established.

Outcome and Follow-up

Adrenocortical insufficiency in the preterm infant is transient. The majority of preterm newborns respond to exogenous CRF and ACTH testing by 2–3 weeks of life, suggesting increased expression and activity of 3β-HSD and 11-β-hydroxylase enzymes and maturity of the HPA axis.¹⁷

An association between adrenocortical dysfunction, cardiovascular instability, and BPD has been reported and led to studies testing the hypothesis that hydrocortisone prophylaxis of early adrenal insufficiency prevents BPD.^{18, 19}

Watterberg’s study of neurologic outcome at 18 to 22 months corrected age of 360 preterm infants at 25 weeks’ gestational age showed that infants treated with prophylactic hydrocortisone compared to placebo did not have an increased incidence of neurologic impairment.¹⁸ Peltoniemi et al reported follow-up data from a randomized trial of early hydrocortisone vs placebo given for 10 days and found neurological and neuropsychological examination at 2 years corrected age were similar.¹⁹ However, more randomized, controlled studies with longer follow-up are needed.

REFERENCES

INTRODUCTION

Disorders of the thyroid are one of the most common reasons for pediatric endocrinology consultations in the newborn period. Physical symptoms of thyroid disease are often absent, but the effects of thyroid dysfunction in the neonatal period could result in significant deleterious health effects later in life. This chapter reviews disorders of the thyroid most likely to be encountered in the newborn nursery and neonatal intensive care unit (NICU).

THYROID DEVELOPMENT AND FUNCTION

Prenatal Development of the Hypothalamic-Pituitary-Thyroid Axis

The hypothalamic-pituitary-thyroid system develops during the first trimester of gestation. The hypothalamus forms from the neural plate by gestational week 6.¹ The anterior pituitary is derived from oral ectodermal tissue that becomes Rathke’s pouch. Studies have demonstrated formation of the pouch as early as 8.5 days and separation from the ectodermal tissue by 12.5 days of gestation.^{2, 3} Several transcription factors are responsible for formation of the anterior pituitary cells, including HESX1, sonic hedgehog (SHH), SIX3, ZIC1, and others. HESX1 mutations have been reported in septo-optic dysplasia, which results in pituitary hypoplasia.⁴ Mutations in SHH, SIX3, and ZIC1 are associated with holoprosencephaly.^{5, 6, and 7} Pit1 is a major transcription factor responsible for formation of thyrotrophs (thyroid-stimulating hormone-producing cells), somatotrophs (growth hormone-producing cells), and lactotrophs (prolactin-producing cells).⁸ This factor interacts with the promoter regions of genes responsible for production of the β subunit of thyroid-stimulating hormone (TSH) and the thyrotropin-releasing hormone (TRH) receptor and is involved in thyroid hormone receptor formation.^{9, 10}

The thyroid gland is derived from components of the pharyngeal floor and the fourth pharyngobronchial pouch.¹ Follicular cells migrate from the pharyngeal floor along the thyroglossal duct until they merge with the parafollicular C cells formed from the fourth pharyngobronchial pouch. The thyroid is fully formed and located in its normal position in the anterior neck by 50 days’ gestation.^{1, 11} Morphologic errors that occur during these 50 days can result in ectopic thyroid tissue anywhere along the migratory tract, including sublingually, high in the neck, or even in the mediastinum or the heart.¹ Rarely, thyroid tissue may be found in a thyroglossal duct cyst, when the thyroid does not migrate appropriately and the thyroglossal duct simultaneously does not appropriately degenerate.

Multiple homeobox genes and transcription factors are involved in normal development and migration of the thyroid. Thyroid transcription factor 1 (TTF-1 or Nkx2.1) is necessary for survival and proliferation of primitive thyroid cells.¹ Thyroid transcription factor 2 (TTF-2 or FOXE-1) is responsible for proper migration of thyroid tissue to its normal position in the neck.^{12, 13} Pax8 is important for follicular cell development.^{13, 14} Other genes, such as various Hox, Hex, FOXE-2, and Eya subtypes, interact with TTF-1, TTF-2, and Pax8 to form the thyroid gland.¹³ Although critically important in thyroid development, mutations in TTF-1, TTF-2, and Pax8 genes are responsible for only 2% of thyroid dysgenesis in newborns with congenital hypothyroidism (CH).¹⁴

Maturation of the Fetal Thyroid

The fetal thyroid gland begins to function at 14 weeks’ gestation, which is when it is able to trap iodine. Thyroid hormone production is minimal until 18–20 weeks’ gestation, when iodine uptake is increased and thyroxine (T₄) levels rise throughout the remainder of gestation.^{15, 16, 17, and 18} Triiodothyronine (T₃) concentrations remain low until about 30 weeks’ gestation and then rise slowly until term. This is due in part to low tissue concentrations of type I iodothyronine monodeiodinase, which converts T₄ to T₃, and high concentrations of type III monodeiodinase in the placenta and many fetal tissues, which converts T₄ to the inactive reverse T₃ (rT₃). As the fetal liver matures, hepatic type I monodeiodinase increases T₄ to T₃ conversion, resulting in rising T₃ levels.¹⁹

Pituitary secretion of TSH begins in the 14th week of gestation. TSH concentrations begin to rise significantly between 16 and 18 weeks’ gestation until a TSH surge occurs immediately after birth.²⁰ The pituitary begins responding to TRH by 25 weeks’ gestation. Pituitary production of TSH seems to initially occur independently of T₄ production, as the fetal TSH concentration continues to rise throughout gestation despite normal adult concentrations of T₄ by the third trimester.¹⁸ However, the ratio of TSH to free T₄ progressively decreases into extrauterine life, suggesting that the negative-feedback system matures as TSH-producing cells in the pituitary become more sensitive to circulating T₄ concentrations.²¹

Thyroid Hormone Metabolism and Transport

Eighty percent of T₃ found in the body is derived from T₄ that undergoes deiodination in peripheral tissues.²² T₄ is produced exclusively by the thyroid gland. The peripheral conversion of T₄ to T₃ is controlled by 2 enzymes, type I and type II monodeiodinase.^{23, 24} Type I monodeiodinase is found predominantly in the liver, kidney, and thyroid itself, whereas type II monodeiodinase is found in the brain, pituitary, placenta, skeletal muscle, heart, thyroid, and brown adipose tissue. Type I monodeiodinase can be inhibited by propylthiouracil (PTU), whereas type II cannot.^{1, 23} Type II monodeiodinase is the major enzyme involved in peripheral conversion of T₄ to T₃.^{22, 23, and 24} It is particularly important in making sure the brain is provided with sufficient T₃ for development.^{25, 26}

The thyroid hormones can also be irreversibly inactivated by deiodination of T₄ and T₃ into rT₃ and diiodothyronine (T₂), respectively, which are biologically inactive.²² This inactivation can be catalyzed by type I monodeiodinase but is mostly controlled by a third enzyme, type III monodeiodinase. This enzyme is found in the brain, placenta, pregnant uterus, and most fetal tissues.²⁷ Type II and type III deiodinase work to regulate the amount of active thyroid hormone to which the fetus is exposed from the mother.^{27, 28} All 3 enzymes are regulated by the thyroid status of the fetus.¹

T₃ and T₄ travel in the blood bound to proteins, primarily thyroid-binding globulin (TBG). Forty-nine to 64% of total T₄ and 80% of total T₃ are bound to TBG.²⁹ Fetal TBG production by the liver increases in the second half of gestation and is stimulated by placental estrogen.²³ TBG and other transport proteins are discussed in greater detail further in the chapter (in the section on disorders of thyroid hormone transport proteins).

At the time of birth in full-term infants, there is a large increase in TRH and TSH that results in the so-called TSH surge. TSH levels peak to as high as 100 mU/L by about 30 minutes after birth and then decrease progressively to normal levels by 3 to 5 days of life.^{21, 30, 31} T₄ also rises, in response to this surge, and free T₄ levels remain elevated for up to 10 weeks after birth.³² T₃ levels also rise due to increased thyroidal production in response to the surge and increased activity of hepatic type 1 monodiodinase.^{28, 33}

Placental Iodine and Thyroid Hormone Transport

The biologically active thyroid hormones T₃ and T₄ are formed by coupling of tyrosine rings and organification of iodide. The placenta transports iodide to the fetus via a Na⁺/I^- symporter.^{34, 35} The iodide can then be used to synthesize thyroid hormones. Severe CH occurs in populations of mothers with insufficient iodine intake and may affect the brain development in the fetus.^{13, 36, 37} The mechanism for hypothyroidism is 2-fold. In the first trimester, prior to the fetal thyroid gland functioning, the fetus is dependent on maternal T₄. When maternal iodine intake is insufficient, she becomes hypothyroid, and the fetus does not have sufficient T₄. The second is maternal and fetal thyroid competition for iodide.³⁸ When iodide concentrations are low in the maternal serum, the fetal thyroid is unable to increase its affinity for iodide despite upregulation of the Na⁺/I^- symporter, resulting in underproduction of fetal T₄.^{32, 34}

Influence of Maternal Thyroid Function on the Fetus

While the fetal thyroid begins to function at 14 weeks’ gestation, the fetal brain T₃ receptors are detectable by 10 weeks’ gestation.^{16, 39} The placenta contains monodeiodinases that convert most maternal T₄ and T₃ to the inactive thyroid hormones rT₃ and T₂, respectively.^{13, 27} However, T₄ and T₃ have been found in fetal tissues as early at 9 weeks’ gestation, and by 13 weeks, T₃ concentrations in the cortex have been found to be 50%–60% of those found in adults.⁴⁰ Children born to mothers with undiagnosed hypothyroidism in the first trimester have been found to have cognitive impairment.^{41, 42, 43, and 44} Newborns without functional thyroid glands have low levels of circulating T₄ at birth of maternal origin.^{45, 46} Presence of circulating maternal T₄ in utero is neuroprotective and in the newborn is, at least partially, responsible for the frequently asymptomatic nature of CH early in postnatal life. Figure 47-1 summarizes maternal-placental-fetal interactions with respect to the hypothalamic-pituitary-thyroid axis.⁴⁷

DISORDERS OF THYROID HORMONE TRANSPORT PROTEINS

The majority of thyroid hormone, present in serum, is bound to proteins, including thyroid TBG, albumin, and T₄-binding prealbumin, also known as transthyretin (TTR).⁴⁸ Of these proteins, TBG carries the majority of T₄ and T₃.⁴⁹ In general, patients with defects of thyroid hormone transport proteins are euthyroid, with normal measurement of free T₄ and TSH, but have measured abnormalities in the total T₄.

Both complete and partial TBG deficiency have been described. TBG deficiency is X linked; consequently, it is more common in male infants. However, due to variable X inactivation, female infants can also express TBG deficiency. Prevalence estimates range from 1 in 5000 to 1 in 12,000 newborns.¹

Excess of TBG, which is also X linked, is less common, occurring in 1 in 6000 (Birmingham, UK) to 1 in 40,000 (New York, USA) newborns on screening. The variability in the reported prevalence may be related to transient TBG excess. Those with inherited forms of TBG excess have levels that are 3 to 4.5 times the normal range in affected males and intermediate between normal and affected males in female carriers.⁵⁰

Familial dysalbuminemic hyperthyroxinemia (FDH), which is inherited in an autosomal dominant fashion, is the most common cause of euthyroid hyperthyroxinemia, occurring in up to 1.8% of Hispanics but thought to be less common in other ethnic groups.^{48, 51} Individuals with FDH are clinically euthyroid with elevated total T₄ levels.

Rarely, TTR variants have been described. Autosomal dominant mutations in the TTR gene have been described that increase the affinity for T₄ and lead to detection of low levels of total T₄ in the serum in otherwise-euthyroid individuals.⁵² Likewise, mutations of the TTR gene that decrease the affinity for T₄ and lead to detection of elevated levels of total T₄ in the serum in otherwise-euthyroid individuals have been described. Individuals who are homozygous and heterozygous for TTR mutations have been reported.⁵³

MEDICATIONS AFFECTING THYROID FUNCTION

Although uncommon in the United States, iodine deficiency remains the most common cause of hypothyroidism worldwide and can have an impact on both a pregnant mother and her fetus. Somewhat paradoxically, iodine excess also acutely leads to hypothyroidism via the Wolff-Chaikoff effect, presumably as a feedback mechanism to prevent hyperthyroidism. High levels of dietary iodine, for example in Asian cultures with heavy consumption of seaweed, have been associated with both hypothyroidism and hyperthyroidism.⁵⁴ Rare cases of CH related to excessive iodine intake in nutritional supplements by a pregnant mother have also been reported.⁵⁵ Transient hypothyroidism from topical iodine exposure has been reported and should be in the differential diagnosis but is not a common cause of transient neonatal CH in North America.⁵⁶ The premature infant is at particularly high risk because of the increased frequency of procedures and the immature skin that allows increased absorption. Although there is the potential for iodinated contrast during computed tomographic (CT) studies also to negatively affect the developing thyroid gland of the fetus, a retrospective study of 21 pregnancies in which contrast CT studies were performed between 8 and 37 weeks’ gestation (mean 23) revealed no evidence of hypothyroidism in the offspring.⁵⁷

Dopamine and glucocorticoids interfere with TSH release from the pituitary and therefore could contribute to central hypothyroidism or mask TSH rise in primary CH. Antithyroid medications used in Graves disease are discussed further in the chapter (in the neonatal hyperthyroidism section).

Several other drugs are well known to have an impact on the thyroid axis; however, many are not routinely utilized in pregnancy or the neonatal period. For example, lithium is associated with thyroid dysfunction; however, it is a class D drug in pregnancy because of its link with congenital heart disease, and there are no indications for its use in neonates. The antiarrhythmic amiodarone alters thyroid hormone secretion but is uncommonly used in pregnant woman and neonates. Transient hypothyroidism has been reported in infants born to mothers treated with amiodarone.⁵⁸ The uses of other medications, such as interferon α, in pregnancy are at the case report level.⁵⁹

NEONATAL HYPOTHYROIDISM

Congenital hypothyroidism is a common preventable cause of mental retardation. Newborn screening programs for CH have been a huge public health success, such that severe mental retardation as a consequence of CH has been essentially eliminated.

Primary Congenital Hypothyroidism

Epidemiology

Prior to the onset of newborn screening, the incidence of primary CH was estimated to be 1 in 7000 births.⁶⁰ With the advent of newborn screening, the incidence was initially lowered to the 1:3–4000 range.⁶¹ More recent data suggest the incidence may be even higher. In the United States, the incidence of primary CH increased from 1:4098 in 1987 to 1:2370 in 2002. In New York State, the reported rates have been even higher,⁶² with 1:2278 in 1978 to 1:1414 in 2005. The near doubling of primary CH over the past 2–3 decades is not limited to the United States.⁶³ Several proposed reasons likely together contribute to the increasing incidence of primary CH, including changing trends in ethnic populations, earlier screening, higher proportion of low birth weight infants, and possibly additional prenatal environmental exposures.^{64, 65} Perhaps most significantly, lower TSH cutoffs have increased identification of mild hypothyroidism.^{66, 67}

Higher incidences of primary CH have been reported in Hispanic, Asian, and Native American ethnic groups, with lower incidences among whites and non-Hispanic blacks.^{63, 64, 68} There is an unexplained higher incidence of females with primary CH. An odds ratio of 1.56 was reported in the United States from 1991 to 2000. The sex-specific odds ratio is increased in certain ethnic groups and has been reported as high as 2.6 in Hispanic newborns in California.⁶⁴

Pathophysiology and Genetics

Congenital hypothyroidism can result from several distinct pathophysiological mechanisms. Primary hypothyroidism describes the condition in which there is inadequate thyroid hormone synthesis or release in response to TSH, leading to the distinctive laboratory feature of elevated TSH. The 2 major causes of primary hypothyroidism are thyroid dysgenesis, in which the thyroid gland does not properly form, and dyshormonogenesis, when the gland forms normally and in the correct location but does not function appropriately. Together, they make up the majority of primary CH cases.⁶⁹ Table 47-1 summarizes the known causative genes associated with primary CH.^{47, 70, 71, and 72}

Thyroid Dysgenesis

Developmental disorders of the thyroid ranging from complete absence of thyroid gland development to hypoplasia to ectopic placement historically make up the majority of cases of primary CH. Thyroid dysgenesis is typically considered sporadic, such that the risk of recurrence in a future pregnancy is only modestly different from the general population. Familial cases have been shown to occur in 2%–3% of cases, and causative mutations in genes encoding transcription factors critical to thyroid development and differentiation, including PAX-8, TTF-1, TTF-2, and GL-153, have been identified.^{71, 73, 74, and 75} While thyroid dysplasia has been estimated to constitute approximately 75% of CH cases, this may no longer hold true with the recent change in incidence driven by increasing identification of mild cases. An Italian study identified 435 cases of CH with a cutoff of 10 mU/L over a 7-year period, calculated as an incidence of 1:1446. They then grouped the cases by etiologies, athyreosis/profound hypoplasia, ectopy/hemiagenesis, or gland in situ and determined that 68% were classified as gland in situ, that is, not thyroid dysgenesis, conflicting with the long-held postulate about frequency of different etiologies of CH.⁷⁶

Dyshormonogenesis

Disorders in which the thyroid gland fails to concentrate iodide, form thyroid hormone, and secrete it into the circulation collectively are referred to as thyroid dyshormonogenesis. Key steps in thyroid hormone synthesis and secretion have been reviewed previously; defects in any of these steps can lead to CH. Included among these are defects of the TSH receptor, sodium/iodide symporter, peroxidase system, thyroglobulin, and deiodinases.⁷⁷

Transient and Mild Hypothyroidism

While thyroid dysgenesis and dyshormonogenesis are considered to have established pathophysiological mechanisms, the diagnoses of transient and mild hypothyroidism are somewhat controversial. Transient elevations in TSH can have identifiable etiologies, such as maternal antithyroid drugs, maternal thyrotropin receptor-blocking antibodies (TRAbs), or exposure to excessive iodine.^{78, 79} Genetic etiologies have rarely been described, but it is known that patients heterozygous for DUOX2 and THOX2 mutations can have transient or mild CH.^{80, 81} Most often, the cause of transient CH is not definitively determined. In addition, as one cannot always predict with certainty whether an infant has permanent or transient hypothyroidism, clinicians will often begin treatment in the neonatal period, thereby precluding the opportunity to observe the natural history. Meanwhile, the cutoff TSH value for upper range of normal is arbitrary, and no single feature reliably consistently distinguishes mild hypothyroidism from normal physiology.

Controversy exists regarding whether to treat infants with “borderline” or “subclinical” hypothyroidism: those with normal T₄ but TSH levels persistently above the upper limit of normal but less than 10 μIU/mL. As TSH is considered the most sensitive indicator for hypothyroidism, it is possible that the mild elevation is an indicator of insufficient T₄, even if the T₄ is within the normal reference range for the assay. There are no controlled studies in otherwise-healthy infants to evaluate outcomes in those treated with levothyroxine (L-T₄) for mild TSH elevations vs those left untreated. A randomized, double-blind, placebo-controlled study of children with Down syndrome and borderline TSH values in the newborn period showed that those who received L-T₄ in the first 2 years of life had better motor development and a trend toward, but not significantly, better cognitive development. The L-T₄-treated group also had a better length at 2 years of age when compared to the placebo-treated group, suggesting that treatment may benefit growth.⁸² TRH tests have been used to identify infants with mild elevations of TSH who are most likely to benefit from treatment.⁸³

Newborn Screening

Newborn screening for CH to prevent the outcome of mental retardation fits Wilson and Jungener’s 1968 classical principles for screening for a medical condition in every respect.⁸⁴

Infants born with CH may be indistinguishable from those without CH on physical examination; however, a small sample of blood obtained by heel stick is sufficient to identify the condition. The most significantly affected newborn will achieve normal or near-normal cognitive outcomes when treatment is initiated early. Newborn screening for CH first became available in the 1970s with pilot programs originating in Quebec, Toronto, and Pittsburgh, Pennsylvania.⁸⁵ Since then, newborn screening for CH has been established as a mainstay of newborn care in the United States and the industrialized world.

Primary T₄ Screening

Using filter paper blood spots, an initially low or low-normal T₄ measurement is followed up with a TSH measurement. A major advantage of the primary T₄ screening test is that it has the potential to detect infants with both primary and central hypothyroidism. In addition, it will detect infants with TBG deficiency, and if high values are reported, T₄ screening can detect hyperthyroxinemia. A potential challenge with interpretation is with infants who have an initially normal T₄ concentration, especially if samples are collected close to birth, when the T₄ surge may give a false-negative result. For this reason, special consideration should be given to children with Down syndrome in regions with primary T₄ screens and a TSH sample should be obtained in the newborn period given the high incidence of primary CH in this population of children.⁸⁶ T₄ screening is also associated with a higher false-positive rate, generating additional recall visits.⁶¹

Primary TSH Screening

As the primary condition to identify by newborn screen is primary hypothyroidism, for which the most sensitive laboratory finding is hyperthyrotropinemia, it follows that screening by TSH has significant appeal. With improvement in the TSH assay on dried spot filter paper samples coupled with the recognition that a primary TSH screen has a lower false-positive rate than primary T₄ screening, there have been several newborn screening programs that have transitioned to the primary TSH screen since the beginning of the 2000s. Data from the National Newborn Screening and Genetics Resource Center revealed, in 1998, strictly a primary T₄ test was used by 75% of states. A decade later,⁸⁷ the percentage had decreased to 50%. Finally, a few states have programs that simultaneously assay both TSH and T₄, but such testing is considerably more expensive.⁸⁸

Primary TSH screening has the disadvantage of failing to identify infants with central hypothyroidism. In addition, it creates a challenge to identify those with a late rise in TSH.⁸⁹ Some screening programs routinely collect a second newborn screening specimen at 1–3 weeks of age in all infants. The justification for a second screen is not restricted to identifying CH with late rise in TSH, as other metabolic diseases may have false-negative results on the initial newborn screen. The trend toward early discharge of uncomplicated term newborns, thereby shifting the collection time for the initial newborn screen sample, has driven the increasing use of a second newborn screen to maintain high sensitivity. With CH, the TSH surge shortly after birth leads to rapid changes in the normal reference interval, creating challenges in defining appropriate cutoffs for follow-up. Second screens are currently not routine, as it has yet to be definitively established that they improve outcomes to balance the increased cost.⁹⁰

Follow-up Thyroid Function Testing

Results suspicious for CH by filter paper newborn screen require confirmatory serum testing. Therefore, individuals with the pattern of grossly elevated TSH (eg, > 40 mU/L) and low T₄ should immediately receive confirmatory serum testing and initiation of L-T₄. When the elevation in TSH is less pronounced, a repeat newborn screen is often helpful. Most screening programs have established protocols for contacting the provider to request a repeat newborn screen sample. If the time frame for receiving results from a repeat screen extends beyond the infant’s second week, obtaining serum thyroid function tests to expedite the results may be more reasonable. Infants with prolonged stays in the NICU require routine monitoring of thyroid function tests.⁹¹

Imaging

Imaging studies of the thyroid gland can be useful for diagnosing the underlying etiology for primary CH, thereby allowing appropriate family counseling. It remains controversial whether imaging has an impact on clinical decision making or outcomes, and imaging studies have yet to be routinely recommended by the American Academy of Pediatrics (AAP).⁶¹ One group has promoted a treatment paradigm factoring in the etiology for CH that thus requires imaging⁹²; however, it could be argued that severity could be assessed by TSH alone. Two complementary modalities are routinely available for imaging: radionuclide scans and ultrasound.

Radionuclide Scans

There are 2 commonly used radionuclide scans for the evaluation of primary CH, sodium technetium 99m pertechnetate (^99mTc) and iodine 123 (¹²³I). ¹²³I and ^99mTC scan can identify thyroid aplasia, hypoplasia, and an ectopic thyroid gland. A large gland with increased early ¹²³I uptake can be consistent with dyshormonogenesis.⁶⁸ Current guidelines from the AAP state that radionuclide imaging is optional, at least in part because of potential risks from exposure to low doses of ionizing radiation.⁶¹

Ultrasound

Thyroid ultrasound can be useful as both a primary diagnostic test and in conjunction with a radionuclide scan. In the absence of radionuclide uptake, an ultrasound would confirm thyroid aplasia or dysplasia. If the thyroid gland is noted to be in a normal location with normal size by ultrasound, the absence of radionuclide uptake suggests TSH receptor-inactivating mutations, iodine-trapping defects, maternal thyroid-blocking antibodies, and TSH-β gene mutations as possible underlying etiologies.⁶⁸

As a primary diagnostic tool for detecting an ectopic thyroid gland, color Doppler ultrasonography is reported to be less sensitive than radionuclide scan, detecting only 90% of ectopic thyroid glands at 1 center.⁹³ The advantages of ultrasound are that it is not impacted by exogenous thyroid treatment and there is no ionizing radiation exposure.

Treatment, Management, and Outcomes

To ensure the best neurodevelopmental outcome, once the diagnosis of CH has been made, treatment with L-T₄ should be started immediately. L-T₄ is the treatment choice for hypothyroidism because of its longer half-life and local conversion to T₃ by deiodination. The goal of therapy is to normalize thyroid function tests as soon as possible, with AAP policy goals of normalization of T₄ by 2 weeks of life and TSH within 1 month.⁶¹ An initial dose of 10–15 μg/kg/day of L-T₄ is recommended, with a higher initial dose for those with severely low T₄ values less than 5 μg/dL. At 4 years, children treated with an initial dose of 10–15 μg/kg had higher IQ than those treated with either 6–8 or 8–10 μg/kg/d.⁹⁴ In a randomized-controlled trial, full-scale IQ testing was significantly higher in those treated with high initial dose (average 14.5 μg/kg/d) compared with those treated with a lower dose (average 10.5 μg/kg/d); a nonsignificant trend for verbal and performance IQ was also reported.⁹⁵ Synthetic L-T₄ should be administered in the form of a pill crushed and suspended in a few milliliters of formula, breast milk, or water. L-T₄ should not be administered with soy, iron, calcium, or fiber as they can interfere with absorption. Suspensions of L-T₄ and alternative thyroid hormone replacement that are not regulated by the US Food and Drug Administration are unreliable and should not be used.⁶¹

After initiation of treatment with L-T₄, the AAP, American Thyroid Association, and Pediatric Endocrine Society joint statement recommends infants have repeat TSH and serum T₄ or free T₄ measurements at 2 and 4 weeks, then every 1 to 2 months during the first 6 months of life and every 3 to 4 months between 6 months and 3 years of life.⁶¹ Repeat TSH and serum T₄ or free T₄ should be tested 4 weeks following dose adjustments or changes in medication source. More frequent testing is also recommended if there are concerns for compliance.⁶¹ A more recent study suggested monthly monitoring continue until 12 months of life.⁹⁶

At 3 years of age, once the critical period of neurodevelopment dependent on thyroid hormone is complete, if a child does not have evidence of permanent CH, a trial without L-T₄ can be undertaken. TSH and free T₄ should be measured 4 weeks following the trial.⁹⁷ Evidence for permanent hypothyroidism, including thyroid scan revealing agenesis or an ectopic thyroid gland and rise in TSH above 10 mIU/mL after the first year of life, can be taken as evidence of permanent hypothyroidism, and such a trial should be deferred.

Infants with CH appropriately identified and treated achieve near-normal cognitive function. Gruters et al reviewed 8 prospective studies with IQ scores in patients with CH and controls.⁹⁸ Small differences of global IQ scores ranging from 2 to 10 were reported, although in only 2 of these studies did the difference reach statistical significance. The IQ scores of the patients with CH in all studies were in the normal range. Current guidelines are more aggressive in targeting early correction than in the past. While there is a rationale that this will lead to better prognoses, it may be difficult to demonstrate a difference in studies given the already-normal IQ scores. It should be recognized that up to 10% of appropriately identified patients with CH demonstrate developmental and neurological symptoms. Potential explanations include more severe hypothyroidism leading to a greater impact during fetal development, genetic etiologies that also influence central nervous system development independent of the thyroid, suboptimal treatment, and poor compliance.

The Premature and Ill Infant

Prematurity presents an additional challenge in the interpretation of thyroid function tests and the decision about whether to intervene with L-T₄. As discussed in the previous section, the hypothalamic-pituitary-thyroid axis continues to mature throughout gestation and into the early neonatal period. Therefore, the normal values for preterm infants are different from those in term infants. While the incidence of permanent CH is not different in preterm infants, transient primary hypothyroidism and late rise in TSH are both seen more frequently in preterm and low birth weight infants.⁹⁹

Hypothyroxinemia of Prematurity and Sick Euthyroid Syndrome

Preterm infants have a blunted TSH surge after birth and consequently blunted T₄ and T₃ rise as well.¹⁰⁰ In general, levels of both T₄ and T₃ are low in preterm infants, hence the term hypothyroxinemia of prematurity. The difference in biologically active free thyroid hormones is attenuated as levels of TBG are also low in premature infants. For that reason, measurements of free T₄, preferably by equilibrium dialysis, are recommended. On the other hand, TSH levels are typically normal or low in preterm infants; therefore, elevations of TSH are still interpreted as evidence of primary hypothyroidism.

There has been interest in whether infants with typical hypothyroxinemia of prematurity benefit from exogenous thyroxine to achieve thyroid hormone levels typical of term infants. To date, there is no conclusive evidence for a clinical benefit of thyroxine treatment, and it has been recommended that such treatment only be performed in the setting of a clinical trial, although a significant proportion of neonatologists reported treating regularly in the past.^{101, 102, and 103} While a persistently low T₄ level with normal TSH can simply be consistent with hypothyroxinemia of prematurity, it is the same pattern of thyroid function tests seen with central hypothyroidism. Central hypothyroidism without other pituitary hormone deficits is exceedingly uncommon, and the clinical picture should shape the level of concern for central hypothyroidism.

Sick euthyroid syndrome (SES), a condition in which the peripheral conversion of T₄ to T₃ is inhibited and TSH is low, is more common in ill preterm infants. Infants with a history of birth trauma, hypoxia, acidosis, infection, hypoglycemia, and hypocalcemia are at increased risk for SES.¹⁰⁴ There is no benefit to treating SES as the cause is generally nonthyroidal in nature.¹⁰⁵

Late Rise in TSH

The phenomenon in which an infant initially has a low T₄ and normal TSH that later evolves to an elevated TSH is referred to as late rise in TSH. The mechanism producing the delayed rise is not well defined, but late rise in TSH is much more frequent in low birth weight or critically ill infants.⁸⁹ A single newborn screen may miss this diagnosis; therefore, protocols have been adjusted to monitor for it. In a retrospective NICU study of 736 patients, Hyman et al reported the frequency of late rise of TSH, as defined as a TSH greater than 10 μIU/mL following a normal TSH, was 1.4%. Of those with late rise in TSH, 54% required treatment, at least transiently, with L-T₄. Risk factors associated with late rise in TSH were surgery, sepsis evaluation, dopamine treatment, gastrointestinal disorders, low birth weight, and prematurity.⁸⁹ While there is no standard protocol for repeating thyroid screening tests in premature infants, it has been suggested that routine serial thyroid screening tests be performed in premature and sick infants to detect late rise in TSH.⁹¹ With a primary T₄ screen, the T₄ value is characteristically low, triggering measurement of TSH, which returns normal. This in turn should prompt an additional newborn screen sample to be sent. A primary TSH screen would not identify an abnormal result and therefore requires specific plans to rescreen those at high risk.

Central (Secondary and Tertiary) Congenital Hypothyroidism

Epidemiology

Central CH (CCH) is much less common than primary CH,¹⁰⁶ with prevalence rates ranging from 1:16,000 to as high as 1:160,000. Reports from the Dutch screening program indicated rates of CCH are 1:21,000–22,000 in an iodine-replete population.¹⁰⁷

Pathophysiology, Genetics, and Clinical Manifestations

Central CH is caused by a dysfunction of either the hypothalamus or the pituitary gland, which leads to inadequate stimulation of the normal thyroid gland by TSH. CCH is frequently associated with genetic mutations, which are summarized in Table 47-2.^{106, 108} Noninherited causes of CCH can be caused by insult to the hypothalamus or pituitary gland, such as can occur with hypoxic events. Patients with CCH, when untreated, present similarly to those with primary CH, with prolonged jaundice, coarse cry, poor growth, hypotonia, umbilical hernia, macroglossia, temperature instability, constipation, and skin mottling.¹⁰⁶ In addition, they may also exhibit symptoms of other pituitary hormone deficiencies, including growth failure and hypoglycemia.

Laboratory Evaluations

In the setting of CCH, free T₄ will be low with an inappropriately low TSH. TRH stimulation testing allows distinguishing hypothalamic from pituitary causes of central hypothyroidism.^{109, 110} In addition to evaluating thyroid function, infants should be screened for other pituitary hormone deficiencies. In the newborn period, secondary adrenal insufficiency and growth hormone deficiency should be evaluated in any infant with suspected CCH. Gonadotropins may also be evaluated in the first few months of life.¹¹¹

Imaging

Infants with the diagnosis of CCH should have imaging of the hypothalamus and pituitary gland to assess for structural abnormalities. The preferred technique for evaluation is magnetic resonance imaging (MRI).¹¹² An ectopic pituitary gland or other midline defects, such as absence of the corpus callosum, can help confirm the diagnosis of CCH; however, the absence of pathology on MRI does not exclude the diagnosis.

Treatment

As with primary CH, infants should be treated with L-T₄. Prior to starting treatment, secondary adrenal insufficiency should be excluded. If adrenal insufficiency is present, hydrocortisone replacement should be initiated prior to starting L-T₄. For neonates, the initial starting dose of L-T₄ is 10–15 μg/kg/day.⁶¹ Monitoring the dose is more challenging than in primary CH because TSH is not a good marker for adequacy of dose. Measurements of free T₄ should ideally be just prior to the daily dose of L-T₄. The goal should be to keep the free T₄ in the central part of the reference range for the patient’s age. In adult studies, an unsuppressed TSH above 1 mU/L may be associated with insufficient L-T₄, but this has not been studied specifically in infants or children.¹¹³ The L-T₄ dose may need to be increased when treatment with recombinant human growth hormone is indicated.¹¹⁴

NEONATAL HYPERTHYROIDISM

Neonatal hyperthyroidism is most commonly due to maternal active or inactive Graves disease. Rarely, neonatal hyperthyroidism is associated with mutations leading to activation in the TSH receptor or resistance to the negative feedback of T₄ on the hypothalamus and pituitary.

Neonatal Graves Disease

Epidemiology

Neonatal Graves disease occurs in 1% to 5% of newborns born to mothers with Graves disease. As Graves disease is relatively rare in the pregnant population, the occurrence is estimated to be between 1:25,000 and 50,000 births.⁴⁷

Pathophysiology and Clinical Manifestations

Neonatal Graves disease is caused by the transplacental passage of maternal TSH receptor-stimulating antibodies (TRAb). During the second and third trimesters, TRAb can stimulate the fetal thyroid gland. This leads to excess fetal thyroxine, ultimately causing symptoms of fetal hyperthyroidism, which is manifest most commonly as fetal tachycardia and enlargement of the fetal thyroid gland. Women with an intact thyroid gland can exhibit symptoms of hyperthyroidism, including tachycardia, poor weight gain, goiter with or without bruit, fine tremor, exophthalmos, and hyperemesis. Symptoms may be worse during the first trimester due to high levels of human chorionic gonadotropin (hCG). Mothers with a history of Graves disease who had ablative therapy prior to pregnancy remain at risk for having a child with neonatal Graves disease as the TRAb can remain elevated for months following ablation.¹¹⁵ Table 47-3 lists the potential complications associated with maternal Graves disease.^{116, 117} Maternal hyperthyroidism, alone, has been associated with a 13% increased risk of congenital anomalies in the United States.¹¹⁸ Both methimazole (MMI) and PTU, the medications used to treat maternal Graves disease, have been implicated in congenital anomalies. MMI, in the first trimester, has been shown to have a rare but significantly increased risk for cutis aplasia. PTU has been associated with an increased risk for maternal hepatotoxicity.¹¹⁹ The American Thyroid Association recommends that pregnant women requiring treatment with antithyroid medications receive PTU during the first trimester and then MMI during the second and third trimesters due to the risk for maternal liver dysfunction associated with PTU.¹²⁰

Neonatal Graves disease onset can vary from shortly after birth to up to a few months after birth. The timing of the development of symptoms is dependent on the presence of maternal thyroid-blocking antibodies, as well as the presence of antithyroid hormone medication, which can take days to dissipate. In infants whose mothers are not on treatment with antithyroid medications, symptoms typically develop shortly after birth. Most commonly, symptoms develop in the second week of life in those infants born to mothers treated with antithyroid medications.¹ In the rare cases of mothers with thyroid-blocking antibodies, it can take up to several weeks until symptoms of neonatal Graves disease develop in the infant.^{121, 122} Symptoms of neonatal Graves disease include tachycardia (heart rate > 160 beats/minute), goiter, tachyarrhythmias, congestive heart failure, hyperactivity, restlessness, hyperreflexia, insomnia/poor sleep, jaundice, hyperphagia, vomiting, poor weight gain, exophthalmos, hepatosplenomegaly, liver dysfunction, thrombocytopenia, and lymphadenopathy.⁴⁷ In addition, infants can have an advanced bone age, craniosynostosis, frontal bossing, and triangular facies. Neonatal Graves disease typically resolves in 3 to 12 weeks as maternal TRAbs degrade. The half-life of TRAb is approximately 12 days.

Laboratory Evaluations

When neonatal Graves disease is suspected, thyroid function tests should be obtained. Serum T₄ and T₃ will be higher than normal, and the TSH will be lower than normal for age. Obtaining TRAb can be helpful to confirm diagnosis. The newborn TRAb titer has been reported to predict the likelihood of symptoms developing.¹²³ Maternal TRAb titers are not as predictive of onset, but elevations of more than 3 to 5 times the upper limit of normal in the third trimester have been associated with neonatal Graves disease as well.^{124, 125}

Imaging

Thyroid ultrasound may help to confirm enlargement of the thyroid gland and can be useful for assessing the potential impact of a goiter on the other structures in the neck, particularly the trachea. A consultation with a pediatric cardiologist and an echocardiogram are also recommended to assess for congestive heart failure.

Treatment

Prompt initiation of treatment can be lifesaving once the diagnosis of neonatal Graves disease has been confirmed. MMI is the first-line medication in the newborn period because of the relatively less-severe side effect profile when compared to PTU, which has higher rates of hepatotoxicity.¹²⁶ MMI is dosed at 0.5–1 mg/kg divided into 3 doses, every 8 hours, daily. If, after 24–36 hours, a therapeutic response is not observed, then the dose should be titrated up. Iodide can also be added to the treatment to inhibit thyroid hormone release. Iodide is administered as Lugol solution (126 mg iodine/mL) 1 drop (8 mg) by mouth every 8 hours.¹ Anti-inflammatory doses of glucocorticoids can also be added to inhibit thyroid hormone secretion and decrease peripheral conversion of T₄ to T₃. The infant’s thyroid function tests should be monitored frequently, and in the event the infant is overtreated and develops hypothyroidism, L-T₄ should be added temporarily to prevent side effects of hypothyroidism. TRAb titers should also be monitored, and when normalized, treatment can be weaned.

As the antithyroid medications can take several days to correct the hyperthyroidism, tachycardia should be treated with a β-adrenergic blocker, such as propranolol, which is dosed at 2 mg/kg divided into 3 doses daily, every 8 hours. Once the infant is euthyroid, treatment with the β-blocker can usually be discontinued. If the infant already has signs of congestive heart failure, digitalization, with or without sedation, is indicated.

Outcomes

When treatment is appropriately started, most infants become euthyroid within a few days. There are few studies that evaluated long-term outcomes in children with neonatal Graves disease. A study of 9 children diagnosed with hyperthyroidism in the newborn period and followed up at 5 months to 10 years of age reported 3 of the children had borderline intelligence, 1 had moderately severe psychomotor retardation, 1 had above-average intelligence, and the remaining children were reported as average.¹²⁷

Infants Born to Mothers Treated for Hyperthyroidism During Pregnancy

The majority of infants born to mothers diagnosed with hyperthyroidism do not develop neonatal Graves disease. However, there are other potential complications associated with maternal hyperthyroidism in the newborn. Children born to mothers with inadequate treatment of gestational Graves disease have been noted to have higher rates of central hypothyroidism and thyroid gland dysmorphology, mainly decreased gland volume. The cause is speculated to be secondary to excess fetal exposure to thyroid hormone, which led to inhibition of fetal TSH during a critical period of development of the pituitary-thyroid axis.¹²⁸ Infants born to mothers with thyroid-blocking antibodies and mothers treated during pregnancy with antithyroid medication are also at risk for hypothyroidism and goiter development.¹²⁹ At times, the goiter can become so large that it compromises the airway, presenting a medical emergency.

REFERENCES

INTRODUCTION

Normal sexual development requires a series of sequential and highly regulated events early in fetal development. This sexual differentiation process is regulated through complex interactions from both genetic and hormonal signals. Any disruption in this developmental pathway can lead to a disorder of sex differentiation (DSD).

Neonates born with ambiguous genitalia present a daunting challenge to all involved in health care delivery to the individual patient. An urgent evaluation by appropriate disciplines with discrete and delicate interaction with the family is mandatory. Older and potentially negative and confusing terminology has been replaced and should not be used. A new nomenclature for disorders of sex development, introduced at the 2005 Chicago Consensus Conference,¹ is more precise, integrates better the progress in our understanding of the molecular basis of development, and is more sensitive to the concerns of the family and patients (Table 48-1).

NORMAL SEXUAL DIFFERENTIATION AND DEVELOPMENT

Undifferentiated Stage of Sexual Differentiation

Normal fetal development of the gonads and the genitalia proceed through 3 sequential stages. The first stage involves the formation of the bipotential gonads from the adrenogonadal ridge at approximately 4 to 5 weeks’ gestation. During this undifferentiated stage, identical structures form in both the XY and XX embryo (Figure 48-1). These structures emanate from the mesonephros and coelomic epithelium. They are the wolffian (precursors to the formation of the seminal vesicles, epididymis, and vas deferens) and müllerian ducts (precursors to the fallopian tubes, uterus, and vagina).

During this phase, the cloaca, which is the terminal portion of the hindgut limited by the cloacal membrane, is partitioned into an anterior urogenital sinus and a posterior anorectal canal. Subsequently, the urogenital sinus becomes the bladder in both sexes and is the precursor for the prostate and proximal urethra in males and the entire urethra and vagina in females.

Gonadal Differentiation

Testis Formation

By the seventh week of gestation, the undifferentiated gonads in a male embryo develop into a testis under the influence of the SRY (sex determining region of the Y chromosome) gene located on the short arm of the Y chromosome. The SRY gene expression causes the primitive sex cords to differentiate into Sertoli cells, which when organized form the seminiferous cords. Primordial germ cells populate near the basement membrane of the cords, while Leydig cells appear in the interstitium between the cords around 8 weeks.

Ovary Formation

In the female embryo, the undifferentiated gonad does not contain the Y chromosome. Thus, there is no SRY gene production with no resultant differentiation of cells into Sertoli cells. The primitive sex cords degenerate, and secondary sex cords formed from the genital ridge join with primordial germ cells to form the ovarian follicles. At this time, the germ cells differentiate into oogonia and become oocytes following the first meiotic division. Many oocytes at this time are surrounded by a single layer of granulosa cells, becoming primary follicles at around 15 weeks’ gestation. Further meiosis is arrested under puberty when under hormonal influence meiosis resumes. The total population of germ cells reaches an apex of several million by 5 months’ gestation. However, by the time of birth, most cells have undergone apoptosis, leaving approximately 150,000 oocytes in each ovary at birth.

Other Genes Involved in Gonadal Differentiation

DAX1

The DAX1 gene has been implicated in several cases of XY sex reversal.² Screening of XY females with a normal SRY gene detected the presence of a submicroscopic duplication of a region designated DDS (dosage-sensitive sex reversal). Within this region, the DAX1 gene was identified. Investigation using a transgenic, XY genotype mouse carrying extra copies of the DAX1 gene led to a phenotypically female mouse in the setting of low expression of the SRY gene.³

SOX9

The SOX9 gene was found to play a role in sex determination through investigation of patients with campomelic dysplasia (CD).⁴ In this rare skeletal disorder, a high preponderance of patients were found to have male-to-female sex reversal. Loss-of-function mutations of SOX9 were found in XY females with CD, pointing to a critical role SOX9 likely plays in sex determination.⁵ While the exact targets of SOX9 are unclear, it is structurally similar to SRY and has been proposed to be responsible for an SRY-negative, XX male patient.⁶

SF-1

Steroidogenic factor 1 (SF-1) is a transcription factor that plays a key role in the regulation of endocrine function and development of the adrenal glands and the gonads in both sexes. The results of SF-1 knockout mice models demonstrated a failure of adrenal and gonadal development with XY sex reversal with persistence of the müllerian ducts in males. Ovaries fail to develop in XX female mice lacking the SF-1 gene.⁷ Clinically, there has been a report of an XY female who was found to have adrenal failure at 2 weeks with a poorly differentiated testis, normal female genitalia, and the presence of a uterus. On evaluation, she was found to have a heterozygous mutation in the DNA-binding domain of SF-1, which was likely the cause of the clinical findings.⁸

WT1

The Wilms tumor 1 (WT1) gene was originally discovered on chromosome 11p13 during efforts to find the oncogene responsible for Wilms tumor.⁹ Following its discovery, further investigation of WT1 expression in humans and mice models demonstrated a significant relationship between WT1 protein secretion and sex determination. WT1 expression has been found to have a complex role with multiple sites of action, including modulating SF-1 expression, regulating SRY gene actions, and competing with DAX1.^{10, 11, and 12}

Mutations involving WT1 expression are linked to urinary and genital abnormalities in 3 syndromes: WAGR (Wilms tumor, aniridia, genitourinary anomalies, and mental retardation) syndrome, Denys-Drash syndrome, and Frazier syndrome. Denys-Drash syndrome is a rare genetic disorder associated with gonadal and genital abnormalities, Wilms tumor, and renal failure due to mesangial sclerosis. Frasier syndrome is characterized by male-to-female sex reversal, focal segmental glomerular sclerosis, and Wilms tumor.

Differentiation of the Genital Ducts and External Genitalia

Male Genital Duct Development

By the eighth week of gestation, the developing Sertoli cells secrete müllerian-inhibiting substance (MIS), which causes the müllerian ducts to regress rapidly by the 10th week of gestation. Between the 8th and 12th weeks, Leydig cells secrete testosterone, which stimulates the cranial aspect of the wolffian duct to transform into the epididymis and vas deferens. The caudal portion of the wolffian duct develops into the ejaculatory ducts and seminal vesicles (Figure 48-2).

Female Genital Duct Development

In the female fetus, müllerian duct formation proceeds in the absence of MIS secretion. The müllerian ducts give rise to the fallopian tubes and uterus by the 12th week of gestation. Historically, it was accepted that the upper two-thirds of the vagina originated from the müllerian ducts, with the lower third developing from the posterior urogenital sinus and fusing around 12 weeks’ gestation. Recent small-animal studies have called into question this theory, with genetic molecular work pointing to a complete müllerian vagina.¹³

Development of the External Genitalia

The early development of the external genitalia is similar in both sexes. Around the fifth week of gestation, a pair of swellings called the cloacal folds joins anterior to the cloacal membrane to form the genital tubercle. The urogenital and labioscrotal folds then appear shortly on either side of the genital tubercle.

Development of Male External Genitalia

In the male fetus, the undifferentiated stage of genital development ends around the ninth week with secretion of testosterone and local conversion to dihydrotestosterone causing an increase in distance between the genital tubercle and anal folds (Figure 48-3). The genital tubercle elongates to become the penis, with the urethral groove forming over the ventrum of the penis. Urethral folds, which are extensions of the urogenital folds, form the lateral margins of the urethral groove. The urethra is then created with the fusion of the urethral folds over the urethral groove. The labioscrotal folds then fuse in the midline to create the scrotum. The urogenital folds fuse over the newly created urethra to provide skin coverage for the penis. This process is completed by the 14th week. The phallus will continue to grow throughout the remainder of gestation. Testicular descent occurs through multiple mechanisms and is typically complete by the third trimester.

Development of Female External Genitalia

In the absence of testosterone, the genital tubercle bends inferiorly and forms the clitoris in the female fetus (Figure 48-4). The labioscrotal folds do not fuse in the midline and develop primarily in the caudal portions, forming the labia majora. The urethral folds do not fuse and become the labia minora. The anterior urethral meatus and posterior vaginal orifice form between the labia minora, with this process complete by 14 weeks.

Psychosexual Differentiation

Our understanding of human psychosexual differentiation is limited and appears complex and multifactorial. The previously accepted view that humans are psychosexually neutral at birth¹⁴ has more recently been questioned by those who support the opposing view that a “neural bias” exists determined primarily by prenatal hormonal exposure.¹⁵ Studies in mammals demonstrated a significant role of both testosterone and estrogen exposure prenatally and sexual differentiation of the brain and behavior in utero.^{16, 17} This relationship between hormone exposure in utero and psychosexual differentiation is less clear in humans. However, the accumulated evidence demonstrates that androgen exposure prenatally has more influence on gender role (aspects of behavior in which males and females appear to differ) than gender identity (identification of oneself as either male or female).¹⁸

CLINICAL EVALUATION OF INFANTS WITH AMBIGUOUS GENITALIA

Prenatal Findings

Prenatal suspicion of a DSD occurs most commonly when abnormalities of the genitalia are found on prenatal sonography or when there is a suspected genotype-phenotype mismatch with the karyotype and visualized genitalia or uterus.¹⁹ Characteristic findings are seen with ultrasound allowing for accurate prenatal sexual determination. Normal male genital findings on ultrasound include the penis, which is initially visualized as an upwardly projected phallic structure, in contrast to the clitoris, which will have a downward bend.¹⁹ Also, fetal penile length can then be measured in terms of a range throughout development.²⁰ The scrotum can also be detected as a distinct structure by the second trimester²¹ and is sonographically distinct from the developing labia.²² In addition, a nomogram exists for uterine width and circumference measurements, which allows for assessment of normal female genital tract development from the second trimester until birth.²³

In the rare situations for which genital abnormalities are suspected prenatally, it is important to follow the fetus longitudinally throughout the pregnancy. Fetal ultrasounds at 13–15 weeks are not good predictors of external genital development.¹⁹ The presence of a uterus after 19 weeks is a reliable predictor of internal reproductive organs. Karyotypic or external genital discrepancies based on the presence of a uterus should lead to cautious antenatal counseling, which can include the disciplines of neonatology, pediatric endocrinology, genetic counseling, pediatric urology, and psychology. Certainly, the assignment of sex for the child would be withheld until appropriate postnatal evaluation.

Postnatal Evaluation

Every newborn must have a careful and complete genital examination. The importance of the examination includes not only assignment of the appropriate gender, but also the accurate diagnosis of more common abnormalities, such as hypospadias, chordee, and undescended testes.

Optimal management of neonates with a suspected DSD as recommended by the Lawson Wilkins Pediatric Endocrine Society (LWPES) and the European Society for Paediatric Endocrinology (ESPE) includes the following: (1) Avoidance of gender assignment before expert evaluation is complete. (2) Evaluation and long-term management must be carried out at center with an experienced multidisciplinarian team. (3) All neonates should receive a gender assignment. (4) Open communication with patients and families is essential, and participation in decision making is encouraged. (5) Patient and family concerns should be respected and addressed in strict confidence.

Physical Examination

In the apparent male neonate, the palpability, location, and size of the testis (or ovotestis) should be noted. Ovaries do not descend. The length and diameter of the penis should be measured. Penile length is measured as the stretched length following the reduction of the prepubic fat from the pubic ramus to the tip of glans. Boys with a stretched penile length at term of less than 2 cm in an otherwise physically normal phallus have a micropenis. The location of the urethral meatus should be properly identified. The presence of ventral congenital curvature (chordee) should be documented. Neonates with severe chordee can make obtaining a true stretched penile length difficult. A flat, symmetric scrotum typically indicates bilateral undescended testes. A cleft or bifid scrotum usually occurs in conjunction with a severe penoscrotal or scrotal hypospadias and can look similar to labia.

In the apparent female neonate, the clitoris should be assessed for general size. Clitoromegaly can be dramatic, and in these patients, a stretched length and width should be measured. The urethral location and relationship to a separate vaginal orifice should be carefully assessed. In female fetuses exposed to high levels of testosterone in utero, a common distal stem off the vagina and urethra can occur (urogenital sinus). The os of the urogenital sinus can be located on the perineum or associated with the masculinized clitoris. The labia are normally unfused in females. Fusion, anterior placement of the labia with hyperpigmentation, and absence of labia minora are also signs of in utero androgen exposure. The uterus may be palpable on rectal examination as a cord-like structure anterior to the rectum.

A DSD evaluation should proceed in neonates with the physical findings of (1) overt genital ambiguity; (2) apparent female genitalia with an enlarged clitoris, posterior labial fusion, or an inguinal or labial mass; or (3) apparent male genitalia with bilateral undescended testes, micropenis, isolated perineal hypospadias, or hypospadias (Figure 48-5).

Laboratory Evaluation

Immediate laboratory evaluation of a neonate with a suspected DSD consists of measurement of serum electrolytes, testosterone, and dihydrotestosterone levels and karyotype. Serum 17-hydroxyprogesterone should not be measured until day of life 3 or 4 because elevated levels of this corticosteroid precursor can be found in normal neonates during the first few days of life due to the stress of delivery.²⁴ If 17-hydroxyprogesterone is elevated, 11-deoxycortisol and deoycorticosterone levels are required. LH (luteinizing hormone) levels or a hCG (human chorionic gonadotropin) stimulation test can be performed to determine if functioning testicular tissue is present in the neonate with bilateral nonpalpable testes.

Radiographic Evaluation

An abdominal and pelvic ultrasound is recommended as the first-line imaging study in the neonate with ambiguous genitalia. The presence of a uterus is accurately seen with pelvic ultrasound.²⁵ Gonads may be seen, but the specificity and sensitivity of ultrasound is poor for nonpalpable testes and the presence of ovaries.²⁶ The absence of visible gonads does not confirm their absence. In experienced hands, adrenal sonography is an accurate adjunct for the diagnosis of congenital adrenal hyperplasia (CAH).²⁷ Studies, such as pelvic magnetic resonance imaging (MRI), flush genitogram, and retrograde urethrogram, are necessary in certain situations, typically dependent on physical examination findings.

CLASSIFICATION OF DISORDERS OF SEXUAL DIFFERENTIATION

Sex Chromosome Disorder of Sex Development

45,X/46,XY (Mixed Gondal Dysgenesis)

Mixed gonadal dysgenesis (MGD) is a group of disorders characterized by a unilateral testis, which is typically undescended; an intra-abdominal contralateral streak gonad, persistent müllerian duct structures; and varying degrees of virilization of the external genitalia.²⁸ Infants typically present with ambiguous genitalia as this condition is the second-most-common cause of a DSD in the newborn period.²⁹ The unilateral testis may be descended but usually is palpable in the groin or nonpalpable when located abdominally.

The uterus is typically rudimentary with bilateral fallopian tubes. In those patients with a well-differentiated testis, the ipsilateral fallopian is commonly absent. A urogenital sinus is present in about one-half of patients.

When the unilateral testis is palpable, it is typically more normal. However, the testicular cellular architecture is consistently abnormal in the adult patient, demonstrating few germ cells and sclerosis of the tubules.³⁰ Wolffian structures are more commonly found on the side of the unilateral testis and usually include only the epididymis, with a well-differentiated vas deferens rarely present.

47,XXY (Klinefelter Syndrome and Variants)

Klinefelter syndrome is the most common abnormality of sexual differentiation, occurring in approximately 1 in every 600 live newborn males.³¹ The syndrome is classically characterized by gynecomastia, small testes, absent spermatogenesis, normal-to-moderately reduced Leydig cell function, and increased secretion of follicle-stimulating hormone (FSH).³² It is most commonly associated with the 47,XXY karyotype, but upward of 20% of patients will have higher-grade chromosome aneuploidies, 46XY/47XXY mosaicism, or structurally abnormal X chromosomes.³³

The diagnosis is made prenatally in approximately 10% of patients because of genetic screening programs.³¹ Neonatal diagnosis is rare but can occur during an evaluation for micropenis, severe hypospadias, and undescended or small testes. Typically, the diagnosis is made following puberty when androgen deficiency becomes apparent. Young men with Klinefelter syndrome have sparse body and facial hair. Approximately half of the patients will develop gynecomastia. Adult men have small testes, which rarely exceed 2 cm in length and are atrophic on microscopic examination.

45,X (Turner Syndrome and Variants)

Turner syndrome is associated with an immature female phenotype with short stature, streak gonads, and various congenital anomalies. It is classically and most frequently associated with a missing X chromosome. Variant forms occur less commonly and include partial deletions of the second X chromosome. Various mosaicisms do occur and typically involve the X chromosome. In approximately 5% of patients, a 45,X/46XY mosaicism is found.

Neonates with Turner syndrome do not have ambiguous genitalia but have other somatic features that are noted both antenatally and perinatally, which can lead to the diagnosis (Table 48-2). A clinical guideline pathway exists for those children with a new diagnosis of Turner syndrome. The workup includes a cardiology evaluation, renal ultrasound, hearing testing, and referral to the appropriate support groups. Those patients with a diagnosis missed early in life are typically diagnosed later because of short stature, primary amenorrhea, or lack of secondary sexual development.

Müllerian structures are well differentiated but remain small. The ovaries are streaks by birth with rare surviving oocytes by adolescence. A small number of pregnancies have been reported. Those children with 45,X/46XY mosaicism and occult Y chromosome material have a 7% to 30% risk of developing a gonadoblastoma.

46,XX/46,XY Chimerism or Ovotesticular Development

Ovotesticular disorder, formerly known as true hermaphroditism, is defined as an individual who has both testicular tissue with seminiferous tubules and ovarian tissue with primordial follicles. The combination of gonads in this disorder includes an ovotestis and ovary (40%), bilateral ovotestes (34%), ovotestis and testis (15%), and ovary and testis (10%).³⁴

Patients predominantly are born with ambiguous genitalia and can have varying karyotypes. Approximately 60% of patients are 46,XX, 33% are either a 46,XX/46,XY chimerism or a 46,XX/46XXY mosaicism, and 7% are 46,XY. Understanding of the signaling pathway for testicular formation in patients with a 46,XX karyotype is not completely elucidated but likely involves an X-linked or autosomal mutation, which causes testicular differentiation downstream of the SRY gene.²⁹

Most have more masculinized genitalia and are raised as males following hypospadias and chordee repair. Children raised as females usually have clitoromegaly and a urogenital sinus. Ovaries are typically in a normal location with an associated fallopian tube and usually found on the left side. Testes or ovotestes are more commonly found on the right and can be found anywhere in the normal testicular descent pathway.³⁵ Testes or an ovotestis with predominantly testicular tissue are more likely to descend. A vas deferens and epididymis are almost always present with a testis. In ovotestes, a fallopian tube is found two-thirds of the time as compared to either a vas deferens or both internal ducts in a third.

Histologically, ovarian tissue is typically normal in the ovotestis with the exception of a reduced number of primordial follicles. In contrast, the testicular component of an ovotestis has tubular atrophy with absent spermatogenesis. Gonadal tumors occur in 2%–3% of these patients.³⁶

46,XY Disorders of Sexual Development

Disorders of Testicular Development

Complete Gonadal Dysgenesis

Children with complete gonadal dysgenesis have a 46,XY karyotype, female-appearing external genitalia, normal müllerian structures, and bilateral streak gonads similar to Turner syndrome.³⁷ The majority of patients present in their teens with absent breast development and amenorrhea. Some patients demonstrate clitoromegaly, thought to be due to the elevated levels of gonadotropins.²⁹ The diagnosis is usually made from the karyotype ordered during the evaluation. Patients with complete gonadal dysgenesis have a 30% risk of developing malignant tumors in their gonads by age 30.³⁸

Partial Gonadal Dysgenesis

Patients with partial gonadal dysgenesis (PGD) present similarly to those patients with MGD except that those with PGD have 2 dysgenetic testes rather than 1 dysgenetic testis and a streak gonad. The genital phenotype can range from ambiguous to either normal male or female. This variance is due to the ability and timing of secretion of testosterone from the dysgenetic testes. Usually, persistent müllerian structures are present, but to varying degrees related to MIS secretion. Similar to patients with complete gonadal dysgenesis, those with PGD are at a high risk for gonadal malignancy.³⁸

Testis Regression and Bilateral Vanishing Testes Syndromes

The testis regression and bilateral vanishing testes syndromes encompass a wide spectrum of phenotypes that occur due to testes that “vanish” antenatally, causing varying degrees of abnormalities of both the external genitalia and internal duct development. Testis regression syndrome usually refers to a loss of testicular tissue within the first trimester, resulting in genital ambiguity. In contrast, bilateral vanishing testes syndrome is thought to occur later in utero, after male sexual differentiation of both the internal ducts and external genital anatomy.³⁹

The etiology of the vanishing testis is not elucidated, but hypotheses include bilateral torsion, a genetic mutation, or a teratogen. Neonates with bilateral vanishing testes syndrome present typically with phallic abnormalities, such as a micropenis or severe chordee, or as a phenotypically normal male with an empty scrotum.⁴⁰ Newborns have castrate levels of testosterone in the setting of elevated serum LH and FSH and a 46,XY karotype.

Patients with bilateral vanishing testes syndrome present with ambiguous genitalia or as a phenotypically normal female, depending on the in utero timing of the insult. In severe cases, a 46,XY female is diagnosed at puberty with no müllerian or wolffian structures. This is thought to occur due to the secretion of MIS from the testes prior to regression before androgen secretion.

Disorders in Androgen Synthesis or Action

Androgen Biosynthesis Defect

A defect in any of the steps involved in the synthesis of testosterone from cholesterol will result in absent or decreased production of androgens and resultant 46,XY DSD. There are 6 specific steps involved in this process, 5 involving enzymes (Figure 48-6). Three of these enzymes are required also for cortisol production (cholesterol side-chain cleavage enzyme, 17α-hydroxylase, 3β-hydroxysteroid dehydrogenase), and their defect will lead to impaired cortisol and possibly mineralocorticoid production. All 5 of these enzyme deficiencies are inherited in an autosomal recessive pattern.

StAR Deficiency (Congenital Lipoid Adrenal Hyperplasia). The first step in steroidogenesis involves the transport of cholesterol into the mitochondria. This defect in steroidogenic acute regulator (StAR) protein leads to significant feminization of the genitalia with either completely female features or ambiguous genitalia. Patients have testes located in the abdomen, inguinal canal, or labia. Rudimentary wolffian structures are present with an absence of müllerian ducts. Neonates develop adrenal insufficiency. In some cases, this is life-threatening, affecting newborns as young as 1–2 days old. Other newborns have a milder presentation and present at an older age (2–4 weeks). Patients with this disorder have an enormous accumulation of lipid in the adrenal gland as seen on abdominal imaging, alternatively giving the name congenital lipoid adrenal hyperplasia.

Cholesterol Side-Chain Cleavage Enzyme Deficiency. The second step in steroid hormone synthesis is the conversion of cholesterol to pregnenolone by cholesterol side-chain cleavage enzyme. A deficiency in this enzyme leads to ambiguous genitalia in affected XY males. Corticosteroid and mineralocorticoid production is also impaired, requiring prompt identification and treatment.

3β-Hydroxysteroid Dehydrogenase Deficiency. The 3β-hydroxysteroid dehydrogenase enzyme is found early in the pathway for steroidogenesis and is required for formation of all steroids. Its deficiency affects production of glucocorticoids, mineralocorticoids, and the sex steroids. Boys with 3β-hydroxysteroid dehydrogenase insufficiency have incomplete masculinization of the external genitalia, with findings of a small phallus with hypospadias, a urogenital sinus with blind-ending vaginal pouch, and scrotal testes on examination. Patients have normal wolffian and no müllerian ducts. Salt wasting occurs because of the impaired production of cortisol and aldosterone.

17α-Hydroxylase Deficiency. Patients with 17α-hydroxylase deficiency disorder have absent or only slight masculinization of the external genitalia. Many have normal female genitalia with a blind-ending vaginal pouch and intra-abdominal testes. The deficiency of 17α-hydroxylase leads to the accumulation of desoxycorticosterone, corticosterone, and 18-hydroxycorticosterone in the adrenals, which leads to salt and water retention, hypokalemia, and hypertension.

17,20-Lyase Deficiency. Newborns with deficiency of 17,20-lyase can have a wide spectrum of genital abnormalities, but usually present with ambiguous genitalia. If the diagnosis is not made as an infant, patients present at puberty because of a lack of secondary sexual characteristics.

17β-Hydroxysteroid Oxireductate Deficiency. Newborns with deficiency of 17β-hydroxysteroid oxireductate appear to have a normal phenotype and are typically raised female. These patients have intra-abdominal, inguinal, or labial located testes with normal wolffian and absent müllerian ducts. The diagnosis may be made during repair of an inguinal hernia but typically is made at puberty when there is significant phallic growth and development of male secondary sexual characteristics. Pubertal increases in gonadotropin production lead to a spike in androstenedione production, which overcomes the enzymatic defect and raises testosterone levels into the low-normal range.

Androgen Receptor and Postreceptor Defects

Androgen Insensitivity Syndromes. Patients with both complete and partial androgen receptor defects characteristically have a 46,XY karyotype with bilateral testes. Abnormalities of androgen receptor function represent the most common definable cause of the undervirilized male and are transmitted as an X-linked trait.²⁹ Those with complete androgen insensitivity syndrome have no associated or limited derivatives of wolffian structures.⁴¹ These patients have a female phenotype and no müllerian structures. Diagnosis is rare before puberty unless a karyotype is performed during pregnancy or if testes are found in the groin or labia on physical examination or during repair of an inguinal hernia. Otherwise, these patients present at puberty with primary amenorrhea and development of secondary female sexual characteristics due to the conversion of testosterone to estradiol.

Newborns with partial androgen insensitivity syndrome have incomplete masculinization of their external genitalia. On examination, there is a wide spectrum of findings, but patients usually have hypospadias, bilateral undescended testes, and rudimentary wolffian structures. Multiple mutations of the androgen receptor gene exist, but in general, the abnormality exists as a reduced number of normally functioning receptors or a normal receptor number but decreased binding affinity.⁴²

Disorders of Androgen Action. 5α-Reductase is an enzyme that converts testosterone to dihydrotestosterone. The absence of the enzyme is transmitted in an autosomal recessive pattern, with only homozygous males affected. These children have normal formation of testes and male internal genital duct through the action of testosterone. However, without dihydrotestosterone, normal male external genitalia do not occur. Neonates with this condition present with severe hypospadias or more commonly with ambiguous genitalia. Patients undiagnosed as newborns with bilateral nonpalpable testes can present in a very unusual fashion at puberty with significant phallic development and subsequent gender reversal due to markedly high levels of testosterone.⁴³

Luteinizing Hormone Receptor Abnormality. LH receptor abnormality or Leydig cell aplasia is found typically during an evaluation for a delay in puberty in an adolescent girl. Younger patients may present with this condition with a palpable testis in the inguinal canal or labia. This disorder is characterized by a normal 46,XY karyotype with no wolffian or müllerian structures and a short vagina. It is transmitted as an autosomal recessive trait. Laboratory examination is classic for a low basal level of testosterone, which does not respond to hCG stimulation. Testicular cellular architecture demonstrates an absence or severe reduction in the number of Leydig cells but normal Sertoli cells.

Antimüllerian Hormone and Receptor Abnormality. The condition of antimüllerian hormone and receptor (AMR) abnormality, also known as persistent müllerian duct syndrome or as classically described hernia uteri inguinale,⁴⁴ has characteristic features of a normal-appearing 46,XY male with internal müllerian structures. Patients usually have bilateral fallopian tubes, a uterus, a vagina draining into a prostatic utricle, and either unilateral or bilateral undescended testes. Most frequently, this condition is found during an inguinal hernia repair or orchiopexy when müllerian structures are encountered.

Antimüllerian hormone and receptor abnormality appears to be a heterogeneous genetic disorder affecting the production of either AMH or its receptor. Its transmission occurs as a sporadic mutation or as an X-linked trait. Several clinical forms have been described. The majority of patients (60%–70%) have bilateral intra-abdominal testes similar in location to ovaries. The next-most-common presentation (20% to 30%) is when 1 testis is partially descended and a uterus and fallopian tube are encountered during the inguinal hernia repair, the classic hernia uteri inguinale presentation. The least commonly seen group (10%) is when both testes are found in the same hernia sac with fallopian tubes and uterus due to transverse testicular ectopia.⁴⁵

46,XX Disorders of Sex Development

Disorders of Ovarian Development

Gonadal Dysgenesis

Children with 46,XX gonadal dysgenesis have findings similar to Turner syndrome: bilateral streak gonads, normal müllerian structures with absence of wolffian structures, and normal external genitalia. They differ as these patients have a normal karyotype and stature without other classic physical findings seen in Turner syndrome. Familial cases do occur, which may include include renal disorders, neurologic abnormalities, and sensorineural hearing loss.⁴⁶

Testicular Disorder of Sex Development (46,XX Male)

The classic presentation of an individual with testicular disorder of sex development (46,XX male) is a normal male phenotype with a 46,XX karyotype. Nonclassic forms occur around 10%–20% percent of the time in neonates with varying degrees of sexual ambiguity and a 46,XX karyotype.⁴⁷ Patients with this condition have normal testes at birth that under the influence of the second X chromosome, deteriorate by adulthood as demonstrated by small testicular volume, an absence of spermatogonia, and hyalinization of the seminiferous tubules. Approximately 90% of boys of with this disorder result from an abnormal Y-to-X translocation involving the SRY gene during meiosis.⁴⁸ Those with more SRY gene material are typically more virilized.

46,XX Ovotesticular Disorder

For discussion of 46,XX ovotesticular disorder, refer to the previous section on 46,XX/46,XY (chimerism or ovotesticular disorder).

Androgen Excess

Congenital Adrenal Hyperplasia

Congenital adrenal hyperplasia is the leading cause for a masculinized female and most common DSD in general. Of the 5 enzymatic deficiencies that cause CAH, only abnormalities in 21-hydroxylase (21HD), 11β-hydroxylase, and to a lesser extent 3β-hydroxysteroid dehydrogenase lead to virilization of the female genitalia.