The definition of clinically significant hypoglycemia remains one of the most confused and contentious issues in contemporary neonatology. In this article, some of the reasons for these contentions are discussed. Pragmatic recommendations for operational thresholds, ie, blood glucose levels at which clinical interventions should be considered, are offered in light of current knowledge to aid health care providers in neonatal medicine. Future areas of research to resolve some of these issues are also presented.
Hypoglycemia is one of the common metabolic problems in contemporary neonatal medicine.1 In the majority of healthy neonates, the frequently observed low blood glucose concentrations are not related to any significant problem and merely reflect normal processes of metabolic adaptation to extrauterine life. However, when the low blood glucose levels are prolonged or recurrent, they may result in acute systemic effects and neurologic sequelae.2–4 Unfortunately, untoward long-term outcomes in infants with one or two low blood glucose levels have become the grounds for litigation and for alleged malpractice, even though the causative relationship between the two is tenuous at best.5 For these reasons, the management of low blood glucose in the first postnatal days assumes considerable focus in newborn nurseries and postnatal wards worldwide. However, it is not possible to define a blood glucose level that requires intervention in every newborn infant because there is uncertainty over the level and duration of hypoglycemia that cause damage, and little is known of the vulnerability, or lack of it, of the brain of infants at different gestational ages for such damage. The purpose of this review, prepared by a group of investigators in the field, is to provide a consensus statement of pragmatic recommendations for the operational thresholds of blood glucose concentrations in the day-to-day management of the newborn infant.
Between 1925 and 1960, the only responses to a low plasma or blood glucose concentration that were recognized in the neonate were clinical manifestations including tremor, sweating, lethargy, floppiness, coma, and seizures.6–8 The level of glucose at which clinical manifestations occurred6,,8,9 determined a working definition for significant hypoglycemia. Because similar manifestations can occur with a variety of other neonatal problems, for example, perinatal asphyxia, sepsis, or other metabolic abnormalities, the sine qua non for the diagnosis of significant neonatal hypoglycemia must be to satisfy Whipple's triad10: 1) the presence of characteristic clinical manifestations, 2) coincident with low plasma glucose concentrations measured accurately with sensitive and precise methods, and 3) that the clinical signs resolve within minutes to hours once normoglycemia has been reestablished.
Only when all 3 requirements are met is it possible to be confident of the possibility of the diagnosis of clinically significant hypoglycemia. This glucose level may be quite different from the level and duration that may be associated with long-term damage. Defining these long-term sequelae will require new, carefully conducted and controlled prospective clinical trials in populations of infants with different clinical and metabolic circumstances.
Of additional importance is the fact that this definition based on signs did not include infants with no clinical manifestations, even with extremely low plasma glucose concentrations: so-called asymptomatic hypoglycemia. Substantial controversy remains as to whether asymptomatic hypoglycemia actually causes brain damage.11–14
With the availability of microtechniques to measure circulating concentrations of hormones and substrates in plasma, it became possible to enhance the definitions and develop protocols to examine the possible mechanism that caused the low blood glucose levels. Thus, a fall in plasma glucose concentration provoked a measurable counterregulatory hormonal response in both adults and children. The increases in epinephrine, growth hormone, cortisol, and glucagon, together with substrates such as free fatty acids, glycerol, and ketone bodies could also be defined. In addition, several studies demonstrated the ability of the fetal and neonatal brain to use alternate fuels, eg, ketone bodies and lactate, for oxidative metabolism.15–17
More recently, with the understanding and documentation of neuroglycopenic responses to induced hypoglycemia, it has been possible to record transient abnormalities in brainstem auditory and sensory evoked potentials.18 Changes have also been described during hypoglycemia in cerebral blood flow in preterm newborn infants (as measured by positron emission tomography scan).19,,20 However, to attribute these abnormalities to significant hypoglycemia alone, a reversal of these abnormalities in response to glucose administration needs to be documented. More research is needed to correlate acute changes in function with long-term damage.
There have been 4 approaches to the definition of hypoglycemia: 1) an approach based on clinical manifestations (Table 1), 2) the epidemiologic approach based on measured range of glucose values, 3) an approach based on acute changes in metabolic and endocrine responses and on neurologic function, and 4) an approach based on long-term neurologic outcome.21 None of these has been entirely satisfactory, and all have often been misinterpreted. The approach based on clinical signs is flawed because similar manifestations can occur with a variety of other neonatal problems. The findings of the epidemiologic approach have been erroneously interpreted and used to define cutoff points between normoglycemia and hypo- or hyperglycemia rather than recognizing that hypoglycemia reflects a continuum of biological abnormalities ranging from mild to severe.5,,6,22,23 The third approach is based on relatively few data in small groups of subjects.18 Finally, the data correlating neonatal hypoglycemia with neurologic outcome are limited because of a lack of suitable non-hypoglycemic controls, a failure to consider other pathology, and the small number of asymptomatic infants followed.24
METABOLISM AND EXTRAUTERINE ADAPTATION
At birth, with the sudden discontinuation of the nutrient and other supplies from the mother, the neonate mounts adaptive responses including the mobilization of glucose and fatty acids from glycogen and triglyceride depots, to meet the energy demands.25–28 All these responses are so well orchestrated and integrated under the influence of postnatal hormonal surges and timely expression of genes for key regulatory enzymes that in the majority of neonates, an uncomplicated transition to the extrauterine environment occurs.
The umbilical venous plasma glucose concentration at birth is 60% to 80% of that of the maternal venous glucose concentration. During the first 2 hours of postnatal life, there is a decline in plasma glucose levels followed by a rise, reaching a steady-state glucose concentration by 2 to 3 hours after birth. This adaptation is associated with the hepatic release of glucose at the rate of 4 to 6 mg·kg· −1min−1. Tracer isotope studies have shown that a steady-state rate of glucose production can be measured in normal infants, infants of diabetic mothers, and small for gestational age infants by 3 to 4 hours after birth. Whether such a rate of glucose production is evident earlier than this period has not been confirmed due to limitations of the methodology. In addition to the immediate postnatal stimulation of glycogenolysis after birth, the induction of gluconeogenesis has also been documented at this age.26
Deviations or perturbations in these adaptive responses can occur as a result of a wide range of factors and circumstances, including antepartum metabolic/nutritional events, intrapartum clinical management of the mother, congenital disorders, postnatal complications, and, possibly, the impact of ontogeny, ie, prematurity. Thus, infants born prematurely or following intrauterine malnutrition, maternal diabetes, endogenous fetal hyperinsulinism, etc., may develop abnormally low plasma glucose concentrations for a prolonged period as a result of the failure to mount an appropriate and adequate counterregulatory metabolic and endocrine response.
Although the major oxidative fuel of the brain is glucose, fetal and neonatal brains also have the capacity to oxidize ketone bodies, lactate, and possibly amino acids.16,,17 Therefore, a full understanding of the significance of the blood glucose concentration requires a concurrent understanding of the contribution of other metabolic fuels as well, and the metabolic processes of all the constituent cells of the nervous system.26–28 Several studies have documented a shift in energy metabolism in the immediate newborn period from a respiratory quotient of ∼1.0 at birth to ∼0.85 by 2 hours, suggesting a shift from glucose to a significant contribution by fat (almost 50%) toward oxidative metabolism. This shift in energy metabolism may not occur in situations like hyperinsulinemia resulting from maternal diabetes.26
The ability of the infant to mobilize fat has also been demonstrated by quantifying the rates of lipolysis using tracer isotope studies. This has been measured in normal infants, infants of diabetic mothers, and small for gestational age infants.29–33 Although there may be some qualitative differences, the data in general show that when the neonates are deprived of nutrients by fasting, they are able to increase blood ketone body levels. This has been true for both preterm and term infants. In addition, on day 1 after birth, in full-term infants, fasting (calorie deprivation) is associated with a significant increase in ketone body turnover (17 μmol/kg·min).30 Such a high turnover rate is observed in adults only after several days of fasting. A significant correlation has been demonstrated between ketone body levels and free fatty acid levels. Thus, the data in the literature support the concept that a healthy full-term neonate, as well as a preterm neonate, can readily mobilize fatty acids and form ketone bodies. Evidence for an increased rate of lipolysis and fatty acid oxidation has also been presented in full-term small for gestational age infants.33 Although the blood lactate levels in the immediate hours after birth are high compared with older infants and adults, the size of the lactate pool (maximal estimate 2.4 mM) could not substitute for glucose for neuronal metabolism over a period of time. Furthermore, glucose is the primary source of lactate (and pyruvate) and, therefore, lactate flux is likely to be low during hypoglycemia. Data in adults have shown that even after a very prolonged fast in an adult man, neither ketone bodies nor lactate can fully substitute for glucose, which accounted for 60% of the energy consumption by the brain.34
An operational threshold is that concentration of plasma or whole blood glucose at which clinicians should consider intervention, based on the evidence currently available in the literature. Significant hypoglycemia is not and can never be defined by a single number that can be applied universally to every individual patient. Rather, it is characterized by a value(s) that is unique to each individual and varies with both their state of physiologic maturity and the influence of pathology. It can be defined as the concentration of glucose in the blood or plasma at which the individual demonstrates a unique response to the abnormal milieu caused by the inadequate delivery of glucose to a target organ (for example, the brain). At present, no simple bedside measures exist that can determine these values and hence provide an absolute indication for an intervention in any individual infant. Furthermore, no data exist that define the concentration of plasma glucose or its duration that causes damage. The impact of the availability of alternate fuels as a potentially protective mechanism remains to be determined. Bedside micromethods of accurate analysis of ketone bodies and lactate are not readily available or are not in use. It follows, therefore, that all that can be proposed are pragmatic intervention thresholds that also provide a margin of safety. More precise recommendations await the results of prospective randomized clinical trials.
The Term Infant
Healthy full-term infants born after an entirely normal pregnancy and delivery and who have no clinical signs do not require monitoring of glucose concentrations. Routine measurements of blood glucose concentration should only be undertaken in normal full-term infants who have clinical manifestations or who are known to be at risk of compromised metabolic adaptation.
Breastfed term infants have lower concentrations of blood glucose but higher concentrations of ketone bodies than formula-fed infants.27,,35 Those breastfed infants who appear to lose the most weight postnatally have the highest ketone body concentration. These data suggest that the provision of alternate fuels constitutes a normal adaptive response to transiently low nutrient intake during the establishment of breastfeeding. Therefore, the operational thresholds recommended here may not be applicable to breastfed infants. These infants may well tolerate lower plasma glucose levels without any significant clinical manifestations or sequelae.
The Infant With Abnormal Clinical Signs
If any infant shows clinical manifestations compatible with a significant low blood glucose concentration (symptomatic infant), the plasma glucose concentration should be measured (Table 1). If the value is <45 mg/dL (2.5 mmol/L), clinical interventions aimed at increasing the blood glucose concentration are indicated. Other underlying pathologic processes should be considered as well, and especially if the clinical symptoms do not ameliorate in response to improvement in glucose concentration.
Infants With Risk Factors for Compromised Metabolic Adaptation
Routine measurements of blood glucose concentration should be undertaken in infants known to be at risk as a result of alteration in maternal metabolism, intrinsic neonatal problems, or anticipated or perceived endocrine or metabolic disorders (Table 2). Glucose monitoring can be initiated as soon as possible after birth, and within 2 to 3 hours after birth and before feeding, or at any time there are abnormal signs. If the plasma glucose concentration is less than 36 mg/dL (2.0 mmol/L), a close surveillance should be maintained, and intervention is recommended if plasma glucose remains below this level, if the level does not increase after a feed, or if abnormal clinical signs develop.
At very low glucose concentrations (<20–25 mg/dL, 1.1–1.4 mmol/L), intravenous glucose infusion aimed at raising the plasma glucose levels above 45 mg/dL (2.5 mmol/L) is indicated. It should be underscored that the therapeutic objective (plasma glucose >45 mg/dL, 2.5 mmol/L) is quite different from the operational threshold for intervention (36 mg/dL, <2.0 mmol/L). The higher therapeutic goal is chosen to include a significant margin of safety in the absence of any data evaluating the correlation between glucose levels in this range and long-term outcome in full-term infants. Although the recommendation for maintaining therapeutic levels in excess of 60 mg/dL (3.3 mmol/L) may be indicated in the symptomatic infant with documented profound, recurrent or persistent hyperinsulinemic hypoglycemia,36it should not be the therapeutic goal for the vast majority of newborns with transient or brief episodes of low plasma glucose concentrations that are less than the operational thresholds recommended in this article.
There are no recent data to support the adoption of lower operational thresholds for the preterm infant. Previous observational data suggesting lower plasma glucose in the preterm reflected the prevailing nutritional management of these small infants.37 One retrospective study of preterm infants has suggested a cutoff value of 47 mg/dL (2.6 mmol/L). Values below 47 mg/dL (2.6 mmol/L), if present on 5 different days during the first 2 months of life, showed a significant statistical correlation with abnormal neuromotor and intellectual performance at 18 months of age.11 However, a longer follow-up showed only a decrease in arithmetic and motor test scores at 7½ to 8 years of age, suggesting that the earlier abnormalities had been transient.12 The observed abnormalities at 18 months could also result from the difficulty in making such an evaluation at the younger age. As suggested by the authors,11,,12 a prospective controlled study to test this hypothesis is necessary to confirm these findings.
Infants on Parenteral Nutrition
Any newborn infant receiving parenteral nutrition (glucose and amino acids) may have persistent high plasma insulin concentrations and, therefore, will not have significant lipolysis and ketogenesis. It may be prudent to maintain higher therapeutic levels of plasma glucose (>45 mg/dL or 2.5 mmol/L) at all times in these infants.
Sick infants or those with hypoxic ischemic encephalopathy and/or sepsis may require higher concentrations of glucose to satisfy increased metabolic needs. Further research is necessary to confirm this hypothesis before definitive recommendations can be made.
Screening and Measurement of Plasma Glucose
Glucose reagent strips are commonly used in the newborn nurseries to screen for low blood glucose concentration. These methods should only be considered as a screen or an estimate because they may not be reliable and should not be used as the basis of a diagnosis.38
At least one reliable laboratory value that is significantly low should be obtained when considering the diagnosis of hypoglycemia in the newborn infant, but awaiting laboratory confirmation should not delay treatment in an infant with clinical signs. However, the diagnosis depends on the laboratory plasma glucose values.
This article was prompted by current controversies over the diagnosis and management of hypoglycemia in the newborn infant. Criteria for diagnosis must be clearly separated from therapeutic goals in the neonate. In this article, we have deliberately agreed on high operational thresholds that are applicable to any infant, whether term or preterm, and at the same time, provided a wide margin of safety as derived from analysis of the literature and our own clinical experiences. We do not imply that it is either desirable or necessary for all infants to maintain a given plasma glucose concentration during the early days of life.
It is clear that substantial uncertainties remain, and further research is urgently needed to define the interrelations between blood glucose levels, acute changes in neurologic function, brain damage, and long-term adverse outcome in neonates experiencing low plasma glucose concentrations. Of greatest practical importance is the need for each neonatal nursery to have access to the accurate methods for the measurement of plasma glucose levels, and for its medical and nursing staff to be adequately trained. The maintenance of clear and unambiguous clinical records is also essential for precise correlation with long-term follow-up parameters.
We thank Joyce Nolan for secretarial assistance.
- Received January 20, 2000.
- Accepted January 20, 2000.
Reprint requests to (S.K.) Schwartz Center, Bell Greve Building, Room G-735, 2500 MetroHealth Dr, Cleveland, OH 44109-1998.
- Pildes R,
- Cornblath M,
- Warren I,
- et al.
- Anderson JM,
- Milner RDG,
- Strich SJ
- ↵Cornblath M. Neonatal hypoglycemia. In: Donn SM, Fisher CW, eds. Risk Management Techniques in Perinatal and Neonatal Practice. Armonk, NY: Future Publishing Co, Inc; 1996:437–448
- McQuarrie I
- Lucas A,
- Morley R,
- Cole JJ
- Cornblath M,
- Schwartz R
- Griffiths AD
- Griffiths AD
- Koh THHG,
- Aynsley-Green A,
- Tarbit M,
- Eyre JA
- Pryds O,
- Christensen NJ,
- Friis-Hansen B
- ↵World Health Organization. Hypoglycaemia of the newborn: review of the literature. Geneva, Switzerland: World Health Organization; 1997. Available at: http://www.who.int/chd/pub/imci/bf/hypoglyc/hypoglyc.htm
- ↵Cornblath M, Ichord R. Hypoglycemia in the neonate. Semin Perinatol. In press
- Stanley CA,
- Anday EK,
- Baker L,
- Delivoria-Papadopolous M
- ↵Kalhan SC, Raghavan CV. Metabolism of glucose and methods of investigation in the fetus and newborn. In: Polin RA, Fox WW, eds. Fetal and Neonatal Physiology. 2nd ed. Philadelphia, PA: WB Saunders Co; 1998:543–558
- ↵Kalhan SC, Saker F. Disorders of carbohydrate metabolism. In: Fanaroff AA, Martin RJ, eds. Neonatal-Perinatal Medicine. Diseases of the Fetus and Infant. St Louis, MO: Mosby-Year Book; 1997:1439–1463
- Hawdon JM,
- Ward Platt MP, Aynsley-Green A
- Hawdon JM,
- Ward Platt MP
- Melichar V,
- Drahota Z,
- Hahn P
- ↵Cornblath M, Schwartz R. Disorders of Carbohydrate Metabolism in Infancy. 3rd ed. Boston, MA: Blackwell Scientific Publications; 1991
- Cornblath M,
- Schwartz R,
- Aynsley-Green A,
- Lloyd JK
- Copyright © 2000 American Academy of Pediatrics