Chapter 12: Hypoglycemia

The transition from intrauterine to extrauterine life requires prompt adaptations to maintain glucose homeostasis. Increased rates of glucose utilization in infants and children compared to adults heighten dependency on coordinated glycogenolysis, gluconeogenesis, and fatty acid oxidation to prevent hypoglycemia during early life. Consequently, hypoglycemic disorders often present in the neonatal period or in infancy during transitions to longer feeding intervals. Since the brain during this time is particularly susceptible to injury due to glucose deprivation, prompt recognition and treatment of hypoglycemia is paramount.

Glucose is the preferred substrate for cerebral metabolism and accounts for almost all of its energy requirements. Since the brain can neither synthesize nor efficiently store glucose, its function and health depend on a continuous glucose supply. Metabolic demands of the developing brain raise glucose utilization rates during infancy and early childhood, when the brain is a relatively larger contributor to body mass. Thus, to avoid neurological damage, diagnosing and appropriately treating hypoglycemia is particularly critical during infancy and early childhood, when adequate glucose delivery is essential for neural growth and cognitive development. To promote the prompt diagnosis and treatment of hypoglycemia by the primary care physician and endocrinologist, this chapter integrates current molecular and clinical knowledge.

Normal Pathways ofGlucose Production

Regulation of blood glucose levels in the normal state relies on a delicate balance between nutrient intake and energy expenditure. Nutrition comprises the fundamental processes by which essential nutrients are absorbed by the gut and transferred through the bloodstream and across membranes for transport into cells for metabolism. All metabolic pathways which generate adenosine triphosphate (ATP) pass through glucose or key amino acids. Thus, abnormalities in the absorption, transport, storage, and metabolism of glucose provide the mechanisms for diseases resulting in low levels of glucose in blood (hypoglycemia). Clinically, low glucose levels in key body fluids (notably blood and the cerebrospinal fluid) allow initial identification of such diseases. Beyond infancy, these disorders present rarely; thus, the astute clinician recognizes that, while many patients attribute their symptoms and problems to hypoglycemia, actual hypoglycemia is seldom documented adequately and its prevalence often overestimated.¹

Adaptation to separation from energy supplied via the placenta to self-sufficiency at birth requires rapid activation of three counter-regulatory mechanisms: glycogenolysis, gluconeogenesis, and lipolysis. Thesemechanisms are immature in the newborn and even more rudimentary in premature, asphyxiated, hypoxic, or small-for-gestational age (SGA) infants. The rate of glucose turnover in newborns is 6 mg/kg/min, approximately three times the adult rate. Although the demand for glucose is high, the activities of several liver enzymes involved in glucose production (eg, glucose-6-phosphatase) and ketogenesis are diminished, particularly in preterm infants. Consequently, before the first feeding, approximately 10% of normal term newborns and approximately 67% of preterm SGA infants display blood glucose levels below 30 mg/dL in the first hours of life. These data highlight the importance of administering the initial feeding, typically the fat-rich colostrum, as soon after birth as possible.

Proper storage, oxidation, and production of glucose require intact systems for glycolysis, glycogen synthesis, glycogenolysis, and gluconeogenesis coordinated through activity of the Krebs cycle and the pentose phosphate shunt. The normal integrated regulatory effects of key hormones (eg, insulin, glucagon, and epinephrine), neural pathways, and metabolic substrates precisely match glucose utilization with the sum of exogenous glucose delivery and endogenous production. In the immediate postprandial period, exogenous glucose is absorbed in the gastrointestinal tract via sodium-dependent mechanisms, at a rate reflecting carbohydrate content and gastric emptying, and is transported into liver and pancreas via insulin-independent GLUT-2 transporters, which, unlike other GLUT transporters, are central to glucose homeostasis. In response, rising insulin levels inhibit endogenous glucose production, stimulate peripheral glucose uptake via upregulation of GLUT-4 expression in muscle and fat, and increase muscle and liver glycogenolysis. Simultaneously, inhibition of carnitine palmitoyltransferase-1 reduces peripheral free fatty acid (FFA) oxidation and hepatic β-oxidation. In this setting of increased glucose and insulin concentrations, and low FFA and ketone levels, glucose is transported into cells by GLUT-1, -3, and -4 transporters. When sensitivity to and supply of insulin is normal, glucose fluctuations before and after a meal are relatively small (eg, 60-140 mg/dL), and insulin and glucose levels return to baseline within 2 to 3 hours. Apart from periods of active carbohydrate absorption from the gut (~30 minutes after ingestion), the majority of life is spent in the fasting state. The most immediate adaptation to fasting involves release of glucose from hepatic glycogen stores, which can typically provide substrate for up to 8 hours in older children, but for much shorter periods of time in neonates and ill patients. In the fed state, glycogen only comprises about 5% of liver weight, so that in a 30-kg child energy derived from glycolysis of a total glucose supply of roughly 45 g of glucose could meet basalmetabolic demands for 5 to 6 hours. Thereafter, gluconeogenesis maintains homeostasis, and adequate 3-carbon precursors are essential as substrates for generating glucose—particularly lactate, pyruvate, alanine, glutamine, and glycerol from muscle. Both liver and kidney are capable of releasing glucose into the circulation because these tissues express key enzymes for gluconeogenesis (pyruvate carboxylase, phosphoenolpyruvate carboxykinase [PEPCK], fructose-1-6-bisphosphatase, and glucose-6-phosphatase). Gluconeogenesis is relatively impaired until late childhood as a result of the decreased availability of substrates, but, by that time, can provide 50% to 60% of glucose production after an overnight fast. Active recycling of intermediate metabolites occurs within the mitochondria of the hepatocyte and myocyte during glucose metabolism. The balance between glycolysis and gluconeogenesis appears to depend on the ambient flux of ATP in the hepatocyte, which is determined, under normal conditions, primarily by the extracellular glucose concentration. The maximal hepatic glucose output is approximately 8 mg/kg/min, which is close to the normal glucose requirement in infancy.

Liberation and utilization of fat as an energy source limits catabolism of lean body mass, while meeting energy needs in a more prolonged fasting state. Derived from adipose tissue, FFA circulates to hepatocytes for conversion to ketones: acetone, acetoacetate (AcAc; reported by most clinical laboratories unless requested otherwise), and β-hydroxybutyrate (β-OHB). Over periods of prolonged fasting, FFA oxidation accounts for almost 80% of the body’s energy sources. In contrast to gluconeogenesis, ketogenesis occurs more rapidly and to a higher level in children than in adults. After fasting for 20 hours, the ketone body turnover rate in young children is comparable to that of adults who have fasted for several days. Most body tissues (including cardiac and skeletal muscle) can use FFA for oxidative metabolism, with the notable exception of the brain, which requires β-OHB, the ketone species that predominates in human plasma (typically 10-fold over AcAc or acetone) and also readily crosses the blood-brain barrier. Ultimately, these adaptations to prolonged fasting decrease glucose turnover by approximately 50% and lead to a gradual decrease in circulating glucose concentrations.

In summary, finely tuned orchestration of glycogenolysis, gluconeogenesis, and fat oxidation normally ensures a critical arterial plasma glucose concentration which, facilitated by insulin-independent glucose transport proteins (GLUT-1 and -3), can provide a continuously adequate supply of glucose to the brain. Importantly, GLUT-1 and -3 proteins are upregulated by chronic cerebral hypoglycemia, but glucoregulatory factors, including insulin, do not affect glucose uptake into the brain.

Regulation of Insulin Secretion

Released from β-cells in the pancreatic islets of Langerhans, insulin is the only endocrine hormone which significantly and directly lowers the circulating glucose level. It does so by mediating glucose uptake into tissues. Pancreatic β-cells and the liver integrate the major sensors responsible for detecting high glucose and stimulating insulin release. At the β-cell level of glucose-sensing, any acute rise in the intracellular concentration of ATP or free calcium induces insulin release (Figure 12-1). Glucokinase constitutes the primary sensor for high glucose within the β-cell because its enzymatic action is the rate-limiting step for glycolysis, which results in ATP production. Oxidation of many intermediate fuels, such as glutamate, results in insulin release by raisingintracellular ATP. Glutamate dehydrogenase (GDH) reversibly oxidizes glutamate to α-ketoglutarate, a direct intermediate of the Krebs cycle, for ATP production. This GDH-catalyzed reaction is enhanced by positive allosteric regulators such as L-leucine and inhibited by guanosine triphosphate (GTP).²

The primary pathway for the regulation of insulin secretion in humans is the sulfonylurea receptor (SUR1)-potassium channel complex, which accounts for more than 90% of glucose-stimulated insulin release. In humans, this complex consists of four subunit transporter proteins which bind ATP (ABCC8) and four inward-rectifying subunits regulating potassium (K) flux (Kir6.2 or K-ATP channel). Derived from glycolysis or other pathways, ATP binding at the intracellularnucleotide-binding site of the trans plasma membrane of SUR1 will close the associated inward-rectifying K channel. Closure to K flux depolarizes the membrane, which activates voltage-gated, L-type calcium channels. Influx of calcium raises free calcium levels intracellularly, leading to insulin release from storage granules via a pathway regulated by calmodulin-dependent kinase II.

Counter-regulation to maintain normoglycemia

Multiple counter-regulatory pathways have evolved to maintain normal circulating glucose concentrations (also known as euglycemia or normoglycemia) by opposing insulin action during the stress response.

Glucagon.

In response to falling plasma glucose levels, glucagon released from pancreatic α-cells traverses the portal circulation to act primarily on 6-phosphofructo-2-kinase/fructose-2,6-bisphophatase in hepatocytes. The kinase and phosphatase portions of this bifunctional enzyme regulate, respectively, the key steps controlling glycolysis and gluconeogenesis. Glucagon favors the dephosphorylated state, resulting in activation of the kinase and inhibition of the phosphatase, enhancing gluconeogenesis while simultaneously inhibiting glycolysis.

Catecholamines.

Catecholamines acutely rise in response to hypoglycemia. Adrenergic α-receptor binding enhances amino acid uptake into the liver, which increases the availability of substrates for gluconeogenesis while β-adrenergic receptor activation stimulates lipolysis by inducing phosphorylation of hormone-sensitive lipase (the enzyme that cleaves triglycerides into FFA and glycerol within adipose tissue). There, catecholamines stimulate glycogenolysis via both the α₁- and β-receptor mechanisms; in muscle, stimulation of glycogenolysis occurs solely by β₂-receptor activation and requires the presence of glucocorticoids to antagonize the effects of circulating insulin. Catecholamines also enhance FFA entry and mobilize triglycerides from skeletal muscle, but inhibit the release of amino acids from skeletal muscle, preserving this resource in muscle during stress. Effects on insulin secretion are complex and incompletely understood; β₂-adrenergic activation appears to stimulate insulin and glucagon secretion while α-receptor stimulation suppresses insulin secretion and stimulates glucagon secretion. Finally, catecholamines acutely inhibit insulin-mediated glucose uptake via β-adrenergic stimulation.

Growth hormone.

Growth hormone (GH) is released in response to stress, fasting, hypoglycemia, and rapid declines in blood glucose concentrations. In the neonatal period, GH provides a primary defense against hypoglycemia via its inhibition of peripheral insulin-stimulated glucose uptake, perhaps by action of specific insulin-like growth factor (IGF)-binding proteins to induce insulin resistance. GH also stimulates release of FFA from adipose tissue stores, thereby providing an alternate fuel to glucose during prolonged states of hypoglycemia.

Hypoglycemia of significant duration (eg, hours) is a more potent stimulus than is acute hypoglycemia for GH, ACTH, or cortisol release, but this may not be the case when hypoglycemia is persistent or recurrent. For that reason, a low serum GH level found at the time of spontaneous, as opposed to acute induced, hypoglycemia does not necessarily imply GH deficiency as a cause of the hypoglycemia.

Glucocorticoids.

Glucocorticoids, synthesized in the adrenal cortex upon ACTH stimulation, enter the adrenal medulla and stimulate the activity of phenylethanolamine-N-methyltransferase (PNMT), which catalyzes the N-methylation of norepinephrine to epinephrine. ACTH and cortisol deficiency, therefore, result in a decrease in immediate counter-regulation by epinephrine production and secretion. Furthermore, cortisol promotes gluconeogenesis via mobilization of amino acids from muscle and activation of glucose-6-phosphatase and phosphoenolpyruvate carboxykinase.

In summary, the main insulin counter-regulatory hormonal action during stress arises from adrenomedullary epinephrine, with contributions from adrenocortical cortisol, pancreatic glucagon, and pituitary GH. Although rare tumors can secrete IGFs which cross-react with insulin receptors, IGFs typically exert only permissive actions on glucose transport and metabolism. Regulation of insulin-mediated glucose uptake is the primary mechanism upon which the clinician must focus. Diseases that impact this process involve the hormones of the stress response as well as inborn errors of insulin secretion, glucose transport, and ATP metabolism.

Hypoglycemia is a common clinical problem with serious neurological sequelae when detection is delayed or treatment inadequate. Even though low blood sugar may often be transient, hypoglycemia itself is never physiological (in other words, never normal) and should not be disregarded when documented adequately. Various approaches have been applied to define an abnormally low level of circulating glucose.³ The most pragmatic relies on glucose level–dependent appearance and resolution of clinical signs upon treatment (Whipple triad, Box 12-1) (also see “Signs and Symptoms” later on in the chapter for specifics). Older epidemiological studies of at-risk groups (infants not fed promptly and frequently after birth) were troubled by ascertainment bias are not representative of a truly healthy, normal population. Physiological fasting studies, which integrate sensitive and specific endocrine assessments with safe clinical outcomes, provide the most logical and compelling information.

Decades of such physiological work confirm that blood glucose at any age is normally above 70 mg/dL (3.9 mmol/L).⁴ While many complain that this cut-off triggers too many interventions, abundant evidence suggests that the developing brain is more susceptible to low glucose levels, which behooves clinical vigilance and use of a threshold (ie, 70 mg/dL) above which no counter-regulatory response typically is triggered to maintain glucose homeostasis. Thus, the prudent clinician feeds a newborn infant with a blood sugar less than 70 mg/dL and reassesses the blood glucose level every 15 minutes until clinical recovery is confirmed. While no specific blood glucose level will be universally accepted as the dividing line between hypo- and normoglycemia, use of 70 mg/dL as a target for normoglycemia is safe and supported by physiology.⁴ Three important caveats should be considered. First, this recommendation applies to the symptomatic infant with documented recurrent and/or persistent hyperinsulinemic hypoglycemia, but it may not apply to newborns with a single episode of low plasma glucose by less reliable point-of-care testing. Second, any documented plasma glucose concentration less than 60 mg/dL for infants and children should prompt thorough evaluation. Third, for reliable diagnostic information from the “critical blood sample,” a simultaneous plasma glucose £ 60 mg/dL (and preferably less than or equal to 50 mg/dL) is recommended.

Casual measurement of blood glucose by finger-stick is common clinically. Blood glucose should be assessed with any new-onset seizure disorder (preferably at the time of the convulsion) at any age, or when the patient demonstrates supraphysiological glucose utilization to maintain blood glucose greater than 70 mg/dL (ie, an intravenous glucose infusion rate > 5 to 8 mg/kg/min in infants or > 3 mg/kg/min in adolescents and adults). It is also important to assess blood glucose levels in high-risk infants, including those born SGA, premature, or large-for-gestational age (given their high incidence of known or unknown diabetic mothers).

Although portable glucose meters are convenient and relatively inexpensive, their advantages are offset by their inaccuracy (up to 20% off the true laboratory glucose reading) and unreliability (simultaneous results often vary by 10%). Point-of-care glucometers generally provide interpretable readings between 70 and 500 mg/dL. Glucometer measurements at either extreme raise concern and require further assessment and intervention. Definitive diagnosis of hypoglycemia requires proper specimen collection and prompt handling. The ideal specimen is whole blood drawn from a free-flowing intravascular catheter into a sodium fluoride (NaF)-containing, gray-top tube, which is then processed immediately by the hospital laboratory. (NaF inhibits glycolysis by erythrocytes which, prior to centrifugation, can decrease the glucose level by as much 20 mg/dL/h.) Plasma glucose results obtained promptly by a glucose oxidase assay are typically accurate and reliable within ±1 mg/dL.

Hypoglycemia results from one or more of the following:

1. Excessive glucose utilization owing to excess insulin or insulin-like action

2. Deficient energy production owing to defective glycogenolysis, gluconeogenesis, ketogenesis, or ketone utilization

3. Energy needs which exceed carbohydrate reserves (eg, illness with prolonged vomiting and anorexia, SGA infant)

4. Exogenous medications or toxic ingestions

5. 5 Reactive hypoglycemia

A detailed list of etiologies of childhood hypoglycemia is presented in Table 12-1.

Excess Insulin or Insulin-Like Action

Transient congenital hyperinsulinism occurs most commonly secondary to maternal factors. Hyperglycemia as a result of uncontrolled gestational diabetes induces fetal β-cell hyperplasia and hyperinsulinism, often leading to macrosomia, risk for shoulder dystocia, and hypoglycemia within a few hours of birth. This usually resolves within days to 2 weeks of birth. Hyperinsulinism may also occur as a result of intravenous glucose administered during labor and delivery; secondary to tocolytic and other medications administered to the pregnant woman (terbutaline, propranolol, oral sulfonylureas, and other hypoglycemic agents used to treat gestational or type 2 diabetes); and from perinatal stress (birth asphyxia, low birth weight, preeclampsia, maternal toxemia, premature labor, or premature birth).

Congenital hyperinsulinism (CHI) is the most common etiology of hypoglycemia persisting beyond the immediate neonatal period (first hours of life). Several inborn errors of pancreatic β-cell function disrupt the normal coupling of insulin release to the ambient glucose concentration. At least eight gene mutations have been associated with CHI. These include diffuse or focal mutations of the SUR1 or Kir6.2 genes; mutations of the gene encoding glutamate dehydrogenase (autosomal dominant gain-of-function GLUD1); glucokinase deficiency (autosomal dominant GK gene mutation); select fatty acid oxidation gene defects (eg, SCHAD); defective glycoprotein synthesis (congenital disorder of glycosylation); mutations in transcription factor genes (hepatic nuclear factor 4α, or HNF4α); and loss-of-function mutations in mitochondrial uncoupling protein 2 (UCP2). Genetic tests currently available can identify diagnostic mutations in 40% to 50% of patients affected by these disorders.

Autosomal recessive defects caused by either homozygous mutations in the SUR1 gene (ABCC8) or in the Kir6.2 gene (KCNJ11) which render critical components of the K_ATP channel inactive (ie, constitutively closed) result in diffuse β-cell dysfunction and intractable severe hypoglycemia soon after birth. Other sporadic variants may create focal adenomatous hyperplasia (40%-60% of all cases of K_ATP-CHI hypoglycemia) caused by loss of the maternal allele imprinted at 11p15 in a patient harboring an SUR1 mutation on the paternal allele. It is postulated that the loss of heterozygosity either unmasks a recessively inherited SUR1 or Kir6.2 mutation located on the paternal allele or leads to loss of suppressor genes which normally would inhibit islet cell growth-promoting actions of IGF-II. The significance of identifying diffuse versus focal disease lies in the potential of curative, focused partial versus 95% pancreatectomy. Over the past decade, PET scans with 18-fluoro-l-3,4-dihydroxyphenylalanine (18F-DOPA) have been shown to accurately discriminate focal CHI from diffuse CHI.²

Four autosomal dominant forms of hyperinsulinism tend to have milder clinical presentations and sequelae than autosomal recessive forms. Infants are frequently not large for gestational age and may not present symptomatically until later in infancy or even childhood. Of these four forms, the most common activating mutation in GLUD1, the gene for mitochondrial glutamate dehydrogenase, causes relatively mild hyperinsulinemia accompanied by persistent mild hyperammonemia (blood ammonia levels ~100-200 μmol/L, or three-five times normal). These mutations in GLUD1 cause impaired sensitivity to inhibition by GTP, resulting in gain-of-function and excessive sensitivity to activation by leucine (ie, “leucine sensitivity”). Increased conversion of glutamate to α-ketoglutarate ultimately increases the intracellular ATP to ADP ratio, stimulates insulin release, and yields excessive ammonia (which does not depend on concurrent protein intake).

The second autosomal dominant form of CHI results from defective glucokinase (GCK), causing increased affinity of the enzyme for glucose and an abnormally low “set point” for glucose-stimulated insulin release. Importantly, a defect in GCK can be missed if the critical sample is obtained at a blood sugar below the threshold for insulin secretion.

The third autosomal dominant form of CHI is caused by mutated HNF4α, a transcription factor involved in development and function of the pancreas, liver, and kidney. In the pancreatic β-cell, HNF4α regulates the genes involved in glucose-stimulated insulin secretion. These mutations have previously been linked to early-onset, monogenic, autosomal dominant, non-insulin- dependent diabetes (MODY1). These mutations can lead to presentation with macrosomia and hyperinsulinism during the first week of life (followed bydiabetes later in adolescence or adulthood).

The fourth autosomal dominant form of CHI derives from a loss-of-function mutation in the gene encoding uncoupling protein 2 (UCP2).¹¹ UCP2 is a mitochondrial carrier protein involved in oxidative phosphorylation of ATP synthesis, which ultimately acts as a downregulator of insulin secretion in β-cells. Mutation of UCP2 leads to insulin hypersecretion.^{11, 13} Patients described to date have shown resolution of symptoms by approximately 6 years of age (and some as early as 7 months).

GLUD1, GCK, HNF4α, or UCP2 mutations causing CHI are typically distinguished by a favorable therapeutic response to diazoxide. A description of these various causes of CHI with their genetic and clinical features appears in Table 12-2.

Hypoglycemia also occurs in 50% of infants with Beckwith-Wiedemann syndrome. This most likely results from disordered imprinting at chromosome 11p15 encompassing the IGF2, SUR1, and Kir6.2 genes which gives rise to islet cell hyperplasia (nonimprinted IGF2 produces excessive growth in utero) and dysregulated insulin release from abnormal SUR1 function.

Counter-regulatory hormone deficiencies clinically (although not biochemically) can resemble CHI during the first month of life, since glycogen reserves may be inappropriately conserved during hypoglycemia in both the infants with CHI and those with either GH deficiency or hypopituitarism, and ketone production in response to hypoglycemia is limited in neonates, complicating the distinction between “ketotic” and “nonketotic” disorders at this age.

Postsurgical dumping syndrome¹⁴ typically occurs after Nissen fundoplication or Roux-en-Y gastric bypass. In normal circumstances, when nutrient absorption is complete, plasma glucose levels frequently decrease to below the premeal value. Overly rapid gastric emptying of elemental formulas leads to rapid transluminal absorption of glucose and rapid glucose-stimulated insulin release; the circulating insulin level climbs so high that it does not fall quickly enough to coincide with the empty gut’s lack of ongoing glucose absorption, leading to postprandial hyperinsulinemic hypoglycemia. Classical symptoms of dumping syndrome are loss of consciousness or mild neuroglycopenic signs within 3 hours after each meal, which are rapidly relieved by carbohydrate. Nonclassical symptoms include borborygmi (loud, small bowel sounds) and adrenergic signs as the hypoglycemia resolves in response to secretion of catecholamines.

Iatrogenic hyperinsulinemic hypoglycemia may occur as a result of abruptly discontinued hypertonic dextrose infusions, including total parenteral nutrition. Excessive glucose utilization leading to hypoglycemia may also result from surreptitious insulin or sulfonylurea administration, either unintentionally from insulin overdose or, rarely, as part of homicidal, suicidal, or Munchausen-by-proxy behaviors.

Whether in isolation or as part of multiple endocrine neoplasia type 1 (MEN1), islet cell adenomas (insulinomas) are rare in children, and rarer still in infants. Hypoglycemia may result from tumors that overutilize fuel substrates or that produce IGFs. Such nonislet cell tumor hypoglycemia (NICTH) results from excessive free IGF concentrations acting via insulin and hybrid insulin-IGF receptors at the cell surface. These tumors release immature IGF molecules which nevertheless retain binding properties of the mature forms. The immature forms are often fully glycosylated, raising molecular weight and thus labeled as “big” IGFs. These variant IGF molecules may not be detected on routine commercial assays, requiring consultation with an academic center or a reference laboratory for special testing.

Deficient Energy Production Caused by Defective Glycogenolysis, Gluconeogenesis, Ketogenesis, or Ketone Utilization

Insufficient production of glucose during the postabsorptive state results from either: (1) intrinsic enzymatic abnormalities in glycogenolysis or gluconeogenesis; (2) deficient mobilization of glucose precursors for otherwise normal gluconeogenesis; or (3) inability to effectively produce and/or utilize fatty acids for energy.

Disorders of glycogenolysis are more accurately described as glycogen release diseases (GRD, listed in Table 12-1), rather than the traditional designation as glycogen storage diseases. These GRD patients share the clinical feature of usually minimally ketotic hypoglycemia and markedly impaired fasting tolerance (usually < 6 hours to onset of symptoms related to tissue ATP deficiency). Patients with GRD type 1 present most often after 6 months of age with growth retardation, cherubic facies, and hepatomegaly. Laboratory tests reveal fasting hypoglycemia, lactic acidemia, and hyperlipidemia (reflecting intact counter-regulatory hormone responses), hyperuricemia (from increased nucleic acid catabolism and decreased renal tubular secretion), and a markedly blunted glucose response to glucagon challenge. Prolonged fasting (typically over 8-12 hours) or a vigorous catecholaminergic response to hypoglycemia is usually required in patients with GRD or defective gluconeogenesis to elicit ketonuria. Neurological symptoms are usually absent in patients with GRD type 1 because the brain (and other tissues) utilize high lactate levels as an alternate fuel for ATP (an important distinction from lack of alternate substrate with hyperinsulinemic hypoglycemia). Other forms of GRD (see Table 12-1) show similar, but milder, manifestations (GRD types 3 and 6).

Disorders of gluconeogenesis include autosomal recessive hereditary fructose intolerance (HFI; fructose-1-phosphate aldolase deficiency) and fructose-1,6-diphosphatase deficiency. HFI presents with vomiting, hypoglycemia, and seizures after initiation of fructose ingestion. Continued fructose exposure can lead to failure to thrive, jaundice, hepatosplenomegaly, proteinuria, aminoaciduria, and death. Fructose 1,6-diphosphatase deficiency results in fasting hypoglycemia, lactic acidosis, hepatomegaly, and ketonemia. Patients with galactosemia, an inherited autosomal recessive deficiency of galactose-1-phosphate uridyl transferase (included routinely in newborn screening programs), after exposure to lactose (hydrolyzed to glucose and galactose), develop vomiting, failure to thrive, cataracts, susceptibility to infection, and hepatic and renal dysfunction.

Deficient mobilization of substrate required for gluconeogenesis can also lead to fasting hypoglycemia. Beyond the immediate neonatal period, the second most common cause of infantile hypoglycemia arises from deficiencies of the counter-regulatory (ie, energy substrate mobilizing) actions of GH and/or cortisol. This spectrum includes disorders associated with isolated GH deficiency or panhypopituitarism including congenital malformations of the hypothalamus and pituitary; disorders of the adrenal cortex; and inborn errors of cortisol synthesis. Isolated GH deficiency may have an idiopathic basis or, rarely, be due to either autoimmune hypophysitis or a GH gene mutation. Congenital hypopituitarism results from malformations, such as holoprosencephaly or lissencephaly; mutations in any of the pituitary gland transcription factors (PROP1, POU1F1, LHX3, LHX4, PITX1, PITX2, SOX2, or SOX3) that can cause failure of either somatotroph (GH) and/or corticotroph (ACTH) development; and optic nerve hypoplasia/septo-optic dysplasia (associated, in a small number of cases, with a HESX1 mutation). Acquired GH deficiency can occur secondary to tumor, radiation, infection, and trauma. Several defects inadrenal development or steroid synthesis can cause cortisol deficiency, including both congenital adrenal hyperplasia (most commonly caused by 21-hydroxylase deficiency, but also seen with many other rarer varieties) or hypoplasia (caused by DAX1 mutation).

The most common cause of (noniatrogenic) hypoglycemia in the preschool-aged child is described simply as “ketotic hypoglycemia.” This entity usually presents with early morning symptoms of hypoglycemia (from lethargy to seizures to coma) in a child exposed to an exaggerated fasting stress of 12 to 24 hours (eg, decreased intake due to intercurrent illness or a missed evening meal after a strenuous day). Typically, but not always, the child is light in weight for age and the clinical presentation of hypoglycemia during fasting stress occurs beyond the age of infancy, but prior to age 6 years. This diagnosis of exclusion is made after careful consideration of other inborn errors of metabolism or hormonal deficiencies associated with fasting hypoglycemia with ketonemia (eg, glycogen release disease types 0 and IX, and short-chain fatty acid oxidation defects).

The precise pathophysiology of ketotic hypoglycemia, also referred to as “fasting functional hypoglycemia,”“accelerated starvation,” or “transient intolerance of fasting,” remains unclear. As noted earlier, muscle tissue plays an important role in maintaining normal glucose concentrations through its release of substrates for gluconeogenesis, especially alanine and pyruvate (which can be reduced to lactate and transaminated to form alanine). Low alanine levels in ketotic hypoglycemia have been hypothesized to reflect decreased muscle mass, decreased protein catabolism, and decreased muscle glucose uptake (leading to inhibition of alanine release by increased FFA and ketone production). Infusion of alanine into affected children has been shown to increase glucose levels appropriately, confirming normal underlying capacity for gluconeogenesis. Others have demonstrated reduced catecholamine response in somecases, but this observation is complicated by the adaptive elevation in the threshold for epinephrine release that is seen in normal subjects who have had prior episodes of hypoglycemia. Since approximately 25% of children will develop hypoglycemia after a fast of 24 to 36 hours, ketotic hypoglycemia is perhaps best described as a developmental disorder of homeostasis representing one extreme end of the spectrum of alanine or lactate recycling required to sustain gluconeogenesis in the small child.

Disorders of fatty acid oxidation/ketone production or utilization predispose the patient to fasting hypoglycemia owing to ongoing glucose utilization in the absence of an ability to use ketones as an energy source. Metabolic errors throughout the pathway of fatty acid uptake, activation, and β-oxidation have been described (caused by mutations in the genes encoding MCAD, LCAD, VLCAD, and SCHAD; see Table 12-1); these tend to share clinical features and to occur during prolonged fasting when normal reliance on fatty acid metabolism is maximal. Defective ketogenesis (eg, carnitine deficiency) fails to yield circulating and urinary ketones during hypoglycemia, and abnormal dicarboxylic acids are present in the urine. In addition to skeletal and cardiac muscle and liver dysfunction, a distinctive and threatening feature of some fatty acid oxidation disorders (eg, MCAD) is early profound neurological depression and subsequent damage that is out of proportion to the degree of hypoglycemia. This is thought to reflect direct central nervous system (CNS) toxicity of very high levels of abnormal fatty acid metabolites. Fortunately, newborn screening using mass spectrometry is increasingly allowing early detection of most of these disorders. The two known ketone utilization defects (β-ketothiolase deficiency and succinyl-CoA transferase deficiency) typically present with persistent, marked ketoacidosis in the first months of life. Defective gluconeogenesis arises from a defect in one of several enzymatic actions (in mitochondria and in the cytosol) which create glucose from noncarbohydrate carbon substrates (including lactic acid, glycerol, and amino acids [other than lysine or leucine]).

Energy Needs Which Exceed Carbohydrate Reserves—Other Causes

Hypoglycemia may occur when glucose needs exceed the patients’ ability to meet them for reasons not discussed above. Occasionally, energy requirements associated with intercurrent illness can exceed carbohydrate reserves (eg, prolonged vomiting or anorexia). Hypoglycemia owing to a complex and multifactorial etiology is also a complication of intrauterine growth restriction (IUGR). During the first several days of extrauterine life, without the placental source of nutrients, maintenance of normal glucose levels requires adequate glycogen stores; intact glycogenolytic, gluconeogenic, and lipogenic mechanisms; and appropriate counter-regulatory hormone responses. Hypoglycemia in infants born IUGR is typically attributable to decreased glycogen reserves, with a possible contribution from diminished liver gluconeogenesis derived from alanine and lactate substrates. Growth-restricted infants also have limited fat stores, appear to not oxidize FFA and triglycerides effectively, and may show decreased counter-regulatory hormone responses to falling blood glucose levels. In addition, hyperinsulinism or excessive sensitivity to insulin appears to exacerbate hypoglycemia and, in some cases, leads to the prolonged need for high glucose infusion rates. Short-term treatment with diazoxide to reduce insulin secretion can be helpful in weaning these particular infants from dependency on intravenousglucose or continuous nasogastric feedings.¹⁵

Exogenous Medicationsor Toxic Ingestions

Most hypoglycemic agents impair hepatic gluconeogenesis, stimulate insulin release, or block ATP metabolism. Table 12-1 (subheading 7) lists the agents most commonly associated with hypoglycemia, highlighted by insulin, salicylates, antimalarials, oral hypoglycemics (sulfonylureas and acarbose), and the rodenticide Vacor.

In diabetic patients, hypoglycemia most commonly results from exogenous insulin overdose, delayed feeding, increased physical exertion without decreasing insulin dose, or rarely malabsorption (eg, acute diarrhea, sprue, etc). Despite the overwhelming advantage of improved glycemic control, intensive insulin therapy carries higher risk of hypoglycemia. Recurrent hypoglycemia blunts the threshold for counter-regulatory responses, resulting in hypoglycemic unawareness. With critical illness, hypoglycemia associated with autonomic failure may impair cognition. This sets in motion a vicious cycle of more frequent and severe hypoglycemia, which over time leads to neurological injury ranging from memory loss to motor and cognitive impairments.

Reactive Hypoglycemia

Reactive hypoglycemia is a widely used term for a poorly defined and controversial condition of postprandial adrenergic symptoms in the absence of surgery or apparent pathology. Hypoglycemia cannot always be documented. Empirical therapy includes frequent small meals with limited intake of carbohydrates with high glycemic index. Acarbose may help those who fail dietary interventions.

Signs and Symptoms

Symptoms of hypoglycemia are nonspecific, so that a low threshold for assessing blood glucose levels is often the key to early recognition and diagnosis (Figure 12-2). Mild hypoglycemia may present as headache, hunger, anger, irritability, shakiness, absentmindedness, or overt cognitive impairment. These symptoms are age-dependent. Hypoglycemic infants typically display jitteriness, hunger, and irritability, symptoms that are usually relieved promptly by oral intake. Symptoms tend to be most severe with hyperinsulinism, in which the formation of alternate fuels such as FFA and ketones is inhibited. Alternatively, chronic and persistent hypoglycemia can lessen clinical symptoms and hypoglycemic awareness, an adaptation thought to reflect upregulation of CNS glucose transporters.

The full counter-regulatory response to acute hypoglycemia elicits a substantial catecholamine response, generating adrenergic symptoms that warn the patient or parents of the falling blood and CNS glucose levels. Usually seen in older children, adolescents, and adults, classic adrenergic symptoms serve as early warnings and include palpitations, sweating, and anxiety. If hypoglycemia is unrecognized and/or untreated, early neuroglycopenic symptoms, such as hunger, fatigue, anger, irritability, slowed cognitive processing, will ensue. If further unrecognized, this may progress to more serious manifestations—lossof consciousness and seizure activity—potentially dangerous at any age and which certainly frighten families. Recovery from a severe or prolonged hypoglycemic episode often takes up to 6 hours, a period typically associated with nausea, vomiting, and fatigue. Diminished cognitive function may persist for several hours after a major hypoglycemic episode (eg, seizure), although physical risks from a fall or while driving are limited to the acute event. Most concerning, recurrent or severe hypoglycemia can lead to long-lasting brain damage. Recent evidence supports the possibility of impaired function of the visual cortex with ongoing, severe hypoglycemia in term neonates.¹⁶ This is especially true in the youngest infants, whose nervous system is rapidly growing and developing. Early magnetic resonance imaging (MRI) findings appear to be more predictive of neurodevelopmental outcome of term infants than the severity or duration of hypoglycemia.¹⁷

History and Physical

A history of exaggerated birth length and/or large weight for gestational age, or onset of persistent hypoglycemia within the first 12 hours of life should raise suspicion for CHI. Key information to elicit is whether the mother was diabetic and whether the parents share any consanguinity. Subsequently, infants with CHI often manifest a need for more frequent and/or voracious feeding accompanied by accelerated weight gain. However, this may not occur if CNS dysfunction begins to interfere with feeding behavior. Emergence of symptoms often corresponds with changes in feeding source (eg, breast to bottle) or schedule.

GH deficiency in the newborn period typically presents as hypoglycemia rather than growth failure. Clinical findings associated with either isolated GH deficiency or panhypopituitarism should raise suspicion for hypoglycemia. These include midline facial abnormalities (either severe, eg, holoprosencephaly, or mild, eg, single central maxillary incisor), prolonged jaundice, and/or, in boys, cryptorchidism and micropenis (stretched length of the penis < 2.5 cm in a term male newborn). Only later, after weaning from frequent feedings to overnight fasting (usually by corrected gestational age 4-6 months) does a decrease in linear growth velocity typically become apparent in a child with congenital GH deficiency or hypopituitarism (with or without a relative excess of weight for length or hepatomegaly).

Beyond the neonatal period, symptoms that correlate closely with introductions of specific foods suggest a metabolic intolerance syndrome (eg, HFI). Hypoglycemia which occurs in association with episodes of more prolonged fasting or intercurrent illness suggests disorders of impaired substrate mobilization for gluconeogenesis (eg, ketotic hypoglycemia) or inability to transition to effective fat utilization (eg, fatty acid oxidation disorders). Children with fasting functional hypoglycemia usually present with early morning symptoms of hypoglycemia (ranging from lethargy to seizures to coma) after exposure to an exaggerated fasting stress of 12 to 24 hours (eg, decreased intake due to intercurrent illness or a missed evening meal after a strenuous day). Typically, but not always, the child is light in weight for age and the clinical presentation of hypoglycemia during fasting stress occurs beyond the age of infancy, but prior to age 8 years.

A variety of clinical presentations can mimic signs and symptoms of hypoglycemia and should be considered while the blood glucose is being checked.

1. Sepsis

2. Primary neurological disorder (eg, epilepsy and catalepsy)

3. Toxic ingestion (eg, ethanol, Vacor, and illicit drugs)

4. Vasovagal syncope

5. Primary cardiac disorder (eg, “Tet” spell from tetralogy of Fallot)

6. Malnutrition

7. Cachexia

8. Panic attack

Definitive evaluation begins with procurement of a “critical” sample at the time of confirmed hypoglycemia (Table 12-3), and prior to administration of glucose. Although critical samples lack perfect sensitivity and specificity, they are useful for directing further diagnostic maneuvers and management. Serum and urinary levels of ketones and plasma FFA—obtained during the episode(s) of hypoglycemia—represent inexpensive, convenient clues to the subsequent diagnosis and management (see Figures 12-3 and 12-4).

Careful analysis of the critical blood sample usually allows separation of underlying etiologies into one of the following (Figure 12-3):

1. Disorders of glucose release caused by defective glycogenolysis/gluconeogenesis: increased anion gap acidosis, elevated lactate, ketonemia/ketonuria, appropriately elevated GH and cortisol, FFA present, and insulin undetectable

2. Disorders of substrate mobilization/counter-regulation: increased anion gap acidosis, normal/slightly elevated lactate, ketonemia/ketonuria (may be absent in neonates), FFA present, and insulin undetectable with: (a) low GH (GH deficiency); (b) low cortisol (ACTH or primary adrenal insufficiency); and (c) low alanine (likely ketotic hypoglycemia)

3. Defects in fatty acid oxidation: no/minimal acidosis, low lactate and ketones, elevated FFA, urine positive for dicarboxylic acids, and undetectable insulin

4. Excess insulin effect: no/minimal acidosis, low lactate and ketones, low FFA, insulin present, and glucose rise greater than 30 mg/dL after glucagon

Some results can be recognized immediately as distinctive; for example, the absence of ketonuria within 2 hours of a hypoglycemic seizure strongly suggests the presence of hyperinsulinism. However, abnormal values have generally been defined in the context of concurrent hypoglycemia after a formal fast, controlled in the hospital setting. As the blood glucose level falls to less than 70 mg/dL during a controlled fast in a normal child, the serum insulin level should fall, and ketones should appear in the circulation and eventually in the urine.

Although several diagnostic tests should ideally be performed on this critical specimen, such maneuvers can be delayed during initial stabilization of the patient (from seizure, obtundation, etc). At a minimum, the astute clinician should obtain 5 to 10 mL of whole blood from the acutely ill patient in a plain red-top tube (lacking anticoagulants) and immediately process the sample, by spinning to separate the serum and then freezing this serum at −20°C (or −70°C if available) pending an endocrinology consultation. Most patients are best served by this practical, but underutilized approach. In addition, beyond infancy, if the quantity of blood obtained is insufficient for analysis of all critical sample tests, quick dipstick assessment of urine for the presence or absence of ketones can aid in prioritizing test choices (eg, hyperinsulinemia is unlikely in the presence of large urine ketones).

The approach outlined in Table 12-3A is most suitable for the acute emergency setting. Plasma may be obtained in addition to serum. In contrast, the approach in Table 12-3B follows a controlled fast and requires careful, advanced planning and is most safely performed in the inpatient setting. As soon as sufficient hypoglycemia is attained (preferably < 40 mg/dL, but at least < 50 mg/dL), the critical sample is procured. To conclude a formal fasting study (usually confined to the second approach), the clinician should perform the diagnostic glucagon challenge (Figure 12-5). This test is predicated on the fact that, in the setting of hyperinsulinemia, there is continuous replenishment of hepatic glycogen so that there will be a significant rise in plasma glucose in response to the administered glucagon which will not occur with any other etiology of hypoglycemia (beyond the newborn period). Figure 12-5 outlines the steps of a controlled diagnostic fast.

CHI is confirmed by plasma FFA concentrations less than 1.5 mM, plasma β-hydroxybutyrate levels less than 2 mM, and an excessive glycemic rise (following 1 mg of intravenous or intramuscular glucagon) of at least 30 mg/dL from a basal level less than 40 mg/dL (detailed in Table 12-4). GH deficiency can also be associated with an inappropriate glycemic rise after glucagon, especially in infancy, but not usually in older children; however, the fasting levels of ketones are usually high with GH deficiency. Lack of a glycemic rise during a glucagon challenge excludes hyperinsulinism. Although rarely requested even by experienced endocrinologists during a glucagon diagnostic challenge, a simultaneous rise in plasma lactatelevel by at least 2 mM confirms the diagnosis of glucose-6-phosphatase deficiency (GRD type 1), which can avoid the performance of a liver biopsy. The diagnosis of hyperinsulinism is suggested when the serum insulin level is measured to be above the lower limit of detection of the assay (usually > 2 μU/mL) when the simultaneous laboratory-measured blood glucose is less than 50 mg/dL. CHI is not excluded from the differential diagnosis by the absence of absolute hyperinsulinism because the hypersecretion of insulin in CHI is erratic in both timing and amplitude with respect to the ambient glucose level. Measurement of serum IGFBP-1 during a period of severe hypoglycemia can aid diagnosis of hyperinsulinism (see Figure 12-5), since insulin suppresses hepaticIGFBP-1 release. Inappropriately low IGFBP-1 (< 80 ng/mL)while hypoglycemic suggests hyperinsulinism, since IGFBP-1 is typically greater than 270 ng/mL in children with hypoglycemia due to other causes.¹⁹

Advanced diagnostic tests

Accurate diagnosis of the presence and cause of hyperinsulinism requires advanced, highly specialized tests, necessitating referral to pediatric endocrinologists with special expertise in this area (see Figure 12-4).

Molecular discoveries about CHI yielded commercial DNA tests to enable diagnosis. Early detection of specific mutations for the patient as well as the parents can facilitate long-term prognosis, personalized management, and genetic counseling. (For example, a paternal mutation in the K_ATP channel strongly suggests focal hyperinsulinism, whereas mutations in HNF4a indicate possible early- onset type 2 diabetes). Currently, commercial testing (eg, Athena Diagnostics and Baylor Human Genetics) can be performed for the four most common genes causing CHI (ABCC8, KCNJII, GLUD1, and GCK), HNF4a, hereditary fructose intolerance, and glycogen release diseases.

In addition to the aforementioned laboratory testing, select imaging studies may assist in establishing the diagnosis. MRI of the brain (with and without gadolinium contrast) can define congenital malformations and acquired intracranial masses. Computed tomography (CT) or MRI of the pancreas rarely reveals a focal insulinoma in pediatric patients with hyperinsulinemia, although a positive result is helpful for guiding surgery. Over the past decade, minimally invasive, acute insulin response (AIR) tests have proven useful to differentiate several recognized β-cell disorders.^4,20,21 These tests take advantage of biological behaviors peculiar to the most prevalent defects in glucose-regulated insulin release. In resting, unstimulated β-cells, an open potassium channel complex (SUR1) extrudes potassium from the cell, keeping the voltage-dependent calcium channels (VDCC) closed. Defective SUR1 function will disrupt the electrostatic homeostasis by keeping the potassium channels closed (a missing channel is equivalent to a closed channel), thus activating and opening the VDCC. Infusing calcium, therefore, to defective β-cells with open VDCCs will stimulate acute insulin release. This principle underlies the diagnostic utility of the calcium AIR tests (whether using peripheral, intravenous, or intrapancreatic arterial access). In cases where GLUD1 mutations and the HI/hyperammonemia syndrome are strongly suspected, a similar test can be performed using leucine as a provocative agent.

Positron emission tomography (PET) is available only at specialized centers, but has demonstrated utility for identifying focal lesions.²² Though limited tospecialized centers, selective intrapancreatic arterial stimulation with calcium and portal (or hepatic) venous sampling (ASVS) may provide additional diagnostic utility.²¹

Guidelines for progressive treatment of hypoglycemic disorders are outlined in Table 12-5. Treatment of fasting functional hypoglycemia is focused on education and monitoring designed to prevent subsequent episodes. Avoidance of prolonged periods of fasting is paramount, often necessitating a bedtime snack (rarely in the form of uncooked cornstarch 1 g/kg) and careful attention to provision of frequent carbohydrate-rich nutrition sources during illness. Monitoring of urine for the presence of ketones is very helpful, since the appearance of ketonuria above trace levels in the morning indicates the need for increased nocturnal calorie intake and, during illness, precedes the decline in plasma glucose. If ketosis continues to worsen in spite of efforts to increase oral carbohydrate intake, administration of intravenous glucose may be needed to avert hypoglycemia.

Glucagon is indicated only for emergent treatment of known intoxication with exogenous insulin because hypoglycemia due to causes other than hyperinsulinism will not respond to glucagon.

See Box 12-2 for elements on educating patients and families.

Box 12-3 lists circumstances in which a general pediatrician or internist-endocrinologist should consult a pediatric endocrinologist, rather than manage the patient independently. For help in finding a pediatricendocrinologist in your area, please visit www.hormone.org.