Glucose

Perinatal Glycemic Adaptations

The primary source of energy for the fetus is glucose, which is received from its mother down a concentration gradient. Insulin is secreted at lower glucose concentrations in utero compared with postnatally, as its function during this time is primarily related to fetal growth rather than glucose regulation. Just before delivery, maternal and fetal glucose concentrations increase until clamping of the umbilical cord, when maternal glucose supply to the infant is interrupted and neonatal glucose concentrations decrease, reaching a nadir between 1 to 3 hours after birth. In response to falling glucose concentrations, the infant’s insulin secretion should decrease while secretion of glucagon and catecholamines increases, stimulating glucose production through gluconeogenesis and glycogenolysis. Typically, these physiological transitions, which involve “resetting” the threshold for glucose-stimulated insulin secretion by the pancreatic beta cell to the higher, adult glucose concentration range, occur over the first 48 to 72 hours after birth, resulting in infants achieving normoglycemia by adult definitions by this time point ( Fig. 29.1 ).

Statistical Thresholds

Historically, the designation of hypoglycemia has usually been based on statistical measures (i.e., marked deviation from normal blood glucose levels). Previous determinations of such blood glucose levels in the newborn were derived from infants generally not fed in the first hours of life. Such determinations led to the definition of significant hypoglycemia in the newborn as a whole blood glucose concentration less than 30 mg/dL in the term infant and less than 20 mg/dL in the preterm infant. Subsequent reports in healthy infants, initially breast-fed within the first 3 hours, show that blood glucose concentrations as low as 30 mg/dL are observed in some infants within 1 to 2 hours after birth and are usually transient, asymptomatic, and considered to be part of normal adaptation to postnatal life. In term appropriate–for–gestational age healthy newborns, blood glucose concentrations ranged between 25 and 110 mg/dL within the first few hours after birth; however, by about 72 hours of age, glucose concentrations typically reach at least 60 to 100 mg/dL. The mean plasma glucose increased from 59 ± 11 mg/dL in the first 48 hours to 83 ± 14 mg/dL after 72 hours.

Operational Thresholds

Cornblath and coworkers suggested that the term hypoglycemia is not readily defined for individual patients and that operational thresholds (i.e., “a concentration of blood or plasma glucose at which clinicians should consider intervention”) should be established. An operational threshold by itself is an indication for action and is not diagnostic of disease or predictive of adverse neurological sequelae. These thresholds were defined as less than 45 mg/dL (2.5 mmol/L) for “the infant with abnormal clinical signs” and less than 36 mg/dL (2.0 mmol/L) for the asymptomatic infant and the infant “at risk for hypoglycemia” (without regard to other influencing factors mentioned previously). Harris and colleagues studied infants greater than 35 weeks of gestational age (GA) and noted hypoglycemia (blood sugar < 47 mg/dL [2.6 mmol/dL]) in 51% of high-risk infants (SGA or LGA infants, infants of diabetic mothers, or late preterm infants), and 19% of these infants experienced a blood sugar less than 36 mg/dL (2 mmol/L) (see Table 29.1 ). Clearly the incidence of hypoglycemia will vary substantially depending on the applied operational threshold (see later; also see Table 29.1 and Fig. 29.2 ). The operational threshold still focuses, however, on individual glucose concentrations and does not address whether the threshold level of blood glucose represents the threshold level for neuronal injury. Hawdon makes the plausible argument that hypoglycemia should perhaps be defined as a persistently low blood glucose level in a baby at risk for impaired metabolic adaptation but with no abnormal clinical signs or a single low blood glucose level in a baby presenting with abnormal clinical signs.

Neurophysiological Threshold

The complexity of defining hypoglycemia is further illustrated by the lack of consistency in a particular threshold value and outcome (see next). Blood glucose concentrations less than 47 mg/dL have been associated with prolongation of BSAER latencies, although only 5 of the 17 infants studied were newborns and not all infants had altered responses below this threshold ( Fig. 29.3 ). A later report of a single infant did not detect prolongation of latency to wave V until blood glucose fell to 25 mg/dL. Other relevant physiological measures include CBF and cerebral glucose metabolism. Two studies of human infants showed that CBF increases at glucose concentrations less than 30 mg/dL and that transport becomes limiting for cerebral glucose utilization at a glucose level less than approximately 54 mg/dL.

Neurological Outcome Thresholds

A careful epidemiological study suggested a deleterious effect on subsequent cognitive development in preterm infants whose plasma glucose levels were less than approximately 47 mg/dL on at least one occasion on 3 or more separate days. Abnormalities in arithmetic and motor scores persisted at 7.5 to 8 years. Conversely, in a subsequent study that attempted to duplicate these findings, 47 of 566 infants of GA less than 32 weeks with a blood glucose level less than 47 mg/dL on at least 3 days were matched with hypoglycemia-free infants and followed up through 15 years. The investigators found no difference in physical disability or developmental progress at 2 years or in psychometric assessment at 15 years, although the study had very low power to detect anything but large differences between groups. In a recent retrospective review of 1943 infants (23 to 42 weeks GA), early transient newborn hypoglycemia was noted in 19.3% using a value of 45 mg/dL (2.6 mmol/L), in 10.3% using a value of 40 mg/dL (2.3 mmol/L), and in 6.4% using a value of less than 35 mg/dL (2.0 mmol/L). In this study, transient hypoglycemia was associated with a decreased probability of proficiency on literacy and mathematics fourth-grade achievement tests at 8-year testing (see later). In the largest study to prospectively examine school-age outcomes ( n = 473) in infants at risk for hypoglycemia, infants in the lowest quintile for interstitial glucose concentrations (using continuous glucose monitoring) in the first 12 hours after birth had increased risk of neurosensory impairment at age 4.5 years. The association of neonatal hypoglycemia with outcomes was dose dependent, with the greatest risk of a low executive function score and a low visual motor integration score in children exposed to severe (<36 mg/dL) or recurrent (≥3 episodes) hypoglycemia. Clinically undetected hypoglycemia, determined to be present only by post hoc blinded continuous glucose monitoring, was associated with a higher risk of executive dysfunction. Interestingly, these findings did not persist in later life, with investigators reporting no difference in academic achievement, executive function, or visual-spatial outcomes at 9 to 10 years between children who experienced hypoglycemia and those who did not. Severity or duration of hypoglycemia also did not affect academic achievement outcomes. Rates of low educational attainment were unexpectedly high in this cohort (hypoglycemia, 47%; not exposed to hypoglycemia, 48%), leading authors to suggest that underlying factors predisposing to hypoglycemia may influence later outcomes.

A pragmatic randomized trial ( n = 278) of two thresholds (36 mg/dL vs. 47 mg/dL) for treatment of mild to moderate hypoglycemia that was identified within 24 hours after birth did not find that maintaining blood glucose at the higher threshold improved outcomes at 18 months. However, this study is limited by lack of later outcomes, when many of the developmental domains thought to be associated with hypoglycemia can be assessed, and by exclusion of infants with severe hypoglycemia, potentially biasing results. The extant literature is limited by the low number of quality studies, small sample sizes, lack of follow-up beyond midchildhood, and minimal information about maternal metabolic status. It is impossible to determine whether the outcomes were related to the in utero environment, perinatal risk factors, neonatal hypoglycemia, or all of these. All of these studies were conducted in populations considered “at risk” for hypoglycemia, so generalizability of these associations beyond these populations is also limited. Thus the controversy around definitions of hypoglycemia continues without data to support the effects of treatment at various thresholds on long-term outcomes ( Table 29.2 ).

In summary, defining hypoglycemia is highly complex given the uncertainty around thresholds for neurodevelopmental outcomes, variable susceptibility factors, such as brain energy reserves, hepatic glycogen reserves, gluconeogenic capacity, prior or concomitant hypoxic-ischemic insults, and seizures. To arrive at a generalizable consensus, national organizations have attempted to bridge the outcome-based, neurophysiologic, and statistical approaches to arrive at “operational thresholds.” The variation across these thresholds recommended by different organizations is further evidence of the limitations of available data (see Fig. 29.2 ). These and related issues are discussed later.

Brain as the Primary Determinant of Glucose Production

The pathophysiology of neonatal hypoglycemic encephalopathy has as its basis the importance of glucose as the primary metabolic fuel for the brain. Glucose for normal brain metabolism is derived from the blood, and glucose production in mammals is primarily a function of the liver. The postnatal induction of hepatic glycogenolysis and gluconeogenesis and the interplay of insulin, glucagon, catecholamines, corticosteroids, and other hormones in the regulation of hepatic glucose metabolism have been reviewed in detail by others. It need only be emphasized here that the brain appears to be the major determinant of (hepatic) glucose production. Glucose production was measured in a series of infants and children from 1 to 25 kg in body weight by a continuous 3- to 4-hour infusion of the nonradioactive tracer 6,6-dideuteroglucose, and glucose production on a body-weight basis was found to be twofold to threefold greater in newborns than in older patients. The infants clearly had disproportionately higher rates of glucose production compared with adult subjects. This observation becomes understandable when glucose production is plotted as a function of estimated brain weight ( Fig. 29.4 ). The linear relationship suggests that the disproportionately high rates of glucose production in the neonatal period relate to the disproportionately large neonatal brain. Because central nervous system (CNS) consumption of glucose accounts for 30% or more of total hepatic glucose output, at least in the preterm infant, this relationship between glucose production rate and brain weight seems reasonable.

The mechanisms by which utilization of glucose by the brain may regulate hepatic glucose output are unknown. It is possible that the effect is mediated by subtle changes in blood glucose levels acting directly on pancreatic insulin secretion or on hepatic glucose output. More provocative is the possibility that the brain mediates control over hepatic glucose production by neural or hormonal effectors originating within the CNS. This possibility leads to the interesting logical extension that disturbances of the brain may lead to disturbances in glucose output by the liver and result in hypoglycemia or hyperglycemia (see later discussion). Moreover, the size of the brain per se may also possibly lead to disturbances in glucose output secondary to changes in glucose utilization. At any rate, in the normal human, from the newborn period to adulthood, it is now clear that a very close relationship exists between brain mass and glucose production.

Glucose Metabolism in the Brain

Glucose metabolism in the brain is depicted in a simplified fashion in Fig. 29.5 . Those aspects particularly relevant to this chapter are shown; a further review of cerebral glucose and energy metabolism is contained in Chapter 16 .

Glucose as the Primary Metabolic Fuel for the Brain

The role of glucose as the primary fuel for the production of chemical energy and the maintenance of normal function in the mature brain is supported by three main facts. First, the respiratory quotient (i.e., carbon dioxide output/oxygen uptake) of the brain is approximately 1, a finding indicating that carbohydrate is the major substrate oxidized by neural tissue. Glucose is the only carbohydrate extracted by the brain in any significant quantity. Second, cerebral glucose uptake almost completely accounts for cerebral oxygen uptake. Third, CNS function is rapidly and seriously disturbed by hypoglycemia.

Current data support a similar preeminence for glucose in the immature brain . Thus studies in the newborn dog indicate that glucose consumption in the brain accounts for 95% of cerebral energy supply. Moreover, studies in term fetal sheep demonstrated that, under aerobic conditions, glucose is the main substrate metabolized for energy production. Glucose/oxygen quotients of approximately 1.1 were obtained in two different laboratories. The glucose/oxygen quotient is equivalent to the arteriovenous difference of glucose (×6) divided by the arteriovenous difference of oxygen and represents the fraction of cerebral oxygen consumption required for the aerobic metabolism of cerebral glucose. Although the data demonstrate that glucose is the primary substrate metabolized by the brain, the finding that the values for glucose/oxygen quotients are slightly but consistently in excess of 1 suggests that a portion of the glucose is used for purposes other than complete oxidation to generate high-energy phosphate bonds. Other data, based on the fate of labeled glucose in the brain, indicate that glucose is also used for the synthesis of other materials (e.g., amino acids via transaminations and lipids via appropriate biosynthetic pathways; see Fig. 29.5 ). Synthesis of membrane lipids and proteins, of course, are critical events in the developing brain and probably account for a relatively larger proportion of cerebral glucose utilization than in the mature brain.

Important regional and developmental changes in cerebral glucose utilization have been defined primarily in animals but also in human infants. Thus early in development, regional differences are relatively few, and brainstem structures generally exhibit the highest rates of glucose utilization. With development, increases in cerebral glucose utilization are most prominent, particularly in cerebral cortical regions. In the human infant, the developmental progression in the first year of life occurs first in the sensorimotor cortex and thalamus; next in the parietal, temporal, and occipital cortices; and last in the frontal cortex and association areas. Careful studies in animals, focused primarily on electrophysiological maturation of the brainstem and diencephalic structures, showed a close correlation between increases in rates of glucose utilization and the acquisition of neuronal function.

Additional compelling evidence for the obligatory role of glucose in the developing brain emanates from studies of human newborns by PET. Thus values for cerebral metabolic rate for oxygen in the brains of preterm and term infants are only 3% and 28%, respectively, of the adult values. One reasonable conclusion from these data is that glucose is critical for energy production in the brain, especially in the preterm infant. The data raise the possibility that anaerobic glucose utilization is important in the neonatal brain, and because energy production is markedly less with anaerobic versus aerobic metabolism (see Chapter 16 ), glucose delivery to the brain is critical for energy production in the neonatal brain, especially in the preterm infant.

Alternative Substrates for Glucose in Brain Metabolism

Ketone Bodies

Appreciable data have accumulated to suggest that ketone bodies may be used as alternative substrates for brain metabolism in the neonatal period. Ketone bodies are taken up by the brain by a carrier-mediated transport system and are subsequently used according to the reactions outlined in Fig. 29.6 .

Energy Production

Studies of newborn infants have demonstrated that the cerebral extraction of ketone bodies from blood is markedly greater in the newborn than it is in older infants and adults. Associated with this finding is an enhanced rate of ketone body utilization in the newborn brain. Thus it was shown that ketone bodies account for approximately 12% of total cerebral oxygen consumption in infants subjected to 6-hour fasts. An enhanced capacity to use ketone bodies was also demonstrated in the human fetal brain. These data indicate relevance for animal studies that demonstrate relatively high activities for the enzymes involved in ketone body utilization in the immature versus the mature brain. These enzymatic activities have also been demonstrated in the human fetal brain.

Thus the newborn brain, at least under conditions of brief fasting, normally satisfies a small portion of its energy demands by the conversion of ketone bodies to acetyl-CoA, which then proceeds through the citric acid cycle (see Fig. 29.5 ). However, neonatal hypoglycemia within the first 48 hours appears to be a hypoketotic state, with varying levels of lactate and detectable insulin levels in many infants but undetectable ketones. A recent study of normal term neonates reported that ketones were most available on days 2 to 3 but still only contributed 7% of potential ATP. This finding suggests that during the first days after birth when hypoglycemia is most common, lactate may be more important than ketones as an alternative energy source to glucose. Moreover, experimental data in the newborn dog do not support an important role for ketone bodies in this context (see later discussion).

Limitations of Hepatic Ketone Synthesis

Utilization of ketone bodies as alternative substrates for glucose in brain energy production under conditions of glucose deprivation depends on the capacity of the liver to deliver these compounds to the blood. Data obtained in human newborn infants suggest that hepatic ketone synthesis is restricted during the early neonatal period. The findings demonstrate (1) low levels of ketone bodies, (2) failure of ketone bodies to rise with fasting (in contrast to fasting in older children), and (3) failure of ketone bodies to rise with hypoglycemia ( Table 29.3 ). In a subsequent study, relatively low plasma concentrations of ketone bodies were also documented with formula feeding. Because cerebral utilization of ketone bodies linearly depends on plasma concentrations, these data from studies of human infants suggest that limitations of hepatic ketone synthesis prevent a major role for these materials as alternative metabolic substrates in the brains of human infants with hypoglycemia. However, these data do not rule out the possibility that exogenous administration of ketone bodies or of exogenous sources of ketone bodies (e.g., fatty acids) could serve as alternative metabolic substrates. One report demonstrated cerebral uptake of exogenously administered beta-hydroxybutyrate for the management of hypoglycemic infants in the first year of life.

TABLE 29.3

Failure of Ketone Bodies to Increase in Blood With Hypoglycemia

Data from Stanley CA, Anday EK, Baker L, Delivoria-Papadopoulos M. Metabolic fuel and hormone responses to fasting in newborn infants. Pediatrics . 1979;64:613–619.

	KETONE BODIES (mmol/L)
INFANTS	BETA-HYDROXYBUTYRATE	ACETOACETATE
Normoglycemic, term, AGA	0.31 ± 0.04	0.06 ± 0.01
Hypoglycemic, term, AGA	0.16 ± 0.03	0.02 ± 0.01
Hypoglycemic, SGA	0.24 ± 0.07	0.03 ± 0.01

AGA , Appropriate for gestational age; SGA , small for gestational age.