Objective. To determine the potential contribution of initial hypoglycemia to the development of neonatal brain injury in term infants with severe fetal acidemia.
Methods. A retrospective chart review was conducted of 185 term infants who were admitted to the neonatal intensive care unit between January 1993 and December 2002 with an umbilical arterial pH <7.00. Short-term neurologic outcome measures include death as a consequence of severe encephalopathy and evidence of moderate to severe encephalopathy with or without seizures. Hypoglycemia was defined as an initial blood glucose ≤40 mg/dL.
Results. Forty-one (22%) infants developed an abnormal neurologic outcome, including 14 (34%) with severe hypoxic ischemic encephalopathy who died, 24 (59%) with moderate to severe hypoxic ischemic encephalopathy, and 3 (7%) with seizures. Twenty-seven (14.5%) of the 185 infants had an initial blood sugar ≤40 mg/dL. Fifteen (56%) of 27 infants with a blood sugar ≤40 mg/dL versus 26 (16%) of 158 infants with a blood sugar >40 mg/dL had an abnormal neurologic outcome (odds ratio [OR]: 6.3; 95% confidence interval [CI]: 2.6–15.3). Infants with abnormal outcomes and a blood sugar ≤40 mg/dL versus >40 mg/dL had a higher pH (6.86 ± 0.07 vs 6.75 ± 0.09), a lesser base deficit (−19 ± 4 vs −23.8 ± 4 mEq/L), and lower mean arterial blood pressure (34 ± 10 vs 45 ± 14 mm Hg), respectively. There was no difference between groups in the proportion of infants who required cardiopulmonary resuscitation (7 [46%] vs 15 [57%]) and those with a 5-minute Apgar score <5 (11 [73%] vs 22 [85%]). By multivariate logistic analysis, 4 variables were significantly associated with abnormal outcome: initial blood glucose ≤40 mg/dL versus >40 mg/dL (OR: 18.5; 95% CI: 3.1–111.9), cord arterial pH ≤6.90 versus >6.90 (OR: 9.8; 95% CI: 2.1–44.7), a 5-minute Apgar score ≤5 versus >5 (OR: 6.4; 95% CI: 1.7–24.5), and the requirement for intubation with or without cardiopulmonary resuscitation versus neither (OR: 4.7; 95% CI: 1.2–17.9).
Conclusion. Initial hypoglycemia is an important risk factor for perinatal brain injury, particularly in depressed term infants who require resuscitation and have severe fetal acidemia. It remains unclear, however, whether earlier detection of hypoglycemia, such as in the delivery room, in this population could modify subsequent neurologic outcome.
Glucose is the primary essential energy substrate in the developing brain, and under steady-state conditions, oxidative metabolism accounts for almost all glucose uptake by the brain.1,2 Because the major source of brain glucose is via the blood supply, if the blood glucose concentration should become reduced, then there is an increased risk for brain injury. One important cause of hypoglycemia is asphyxia, during which glucose is rapidly metabolized anaerobically to minimize cellular energy depletion in all tissues, including brain.2 Indeed, in an experimental model of asphyxia, the increase in the glycolytic substrates distal to the phosphofructokinase step that result from accelerated anaerobic glycolysis as noted in the normoglycemic state are absent with hypoglycemia.3 Both mature and immature animal models in addition to in vitro laboratory experiments have demonstrated a critical role for hypoglycemia in the pathogenesis of brain injury. Despite this, there are few observations in human infants to critically assess the role of glucose under conditions of hypoxia-ischemia or asphyxia. The objective of this retrospective study was to determine the potential contribution of initial hypoglycemia to the development of neonatal brain injury in the term infant with severe fetal acidemia.
The data of all term infants who were admitted to the neonatal intensive care unit (NICU) at Parkland Hospital between January 1993 and December 2002 with severe fetal acidemia (ie, an umbilical arterial pH <7.00) were retrieved from a prospectively collected resuscitation registry. The registry contains the following information: maternal history, including complications during pregnancy, labor complications (eg, fetal heart rate abnormalities), mode of delivery, and use of medications; sequence of resuscitation interventions; birth weight; gestational age (a pediatric assessment as determined by the Dubowitz examination4); Apgar scores at 1, 5, and 10 minutes; initial rectal temperature; initial blood sugar; initial mean blood pressure (obtained with the Dinamap method); and cord arterial pH, Pco2, and base deficit. Arterial sampling immediately after delivery is routinely obtained on all deliveries at our institution from double-clamped sections of umbilical cord.5 Short-term neurologic outcome measures include death as a consequence of severe encephalopathy and evidence of moderate to severe encephalopathy with or without seizures. The Sarnat classification was used to grade the severity of encephalopathy as previously reported.6–9 Hypoglycemia was defined as an estimated blood sugar <40 mg/dL.10 Any infant with a blood sugar ≤40 mg/dL had a serum level drawn to confirm the hypoglycemia. The estimated blood sugar was the value entered into the database. The initial blood sugar was obtained within the first 30 minutes after birth. Hypoglycemia was treated with an intravenous bolus of 2 mL/kg of a 10% dextrose solution. In general, if the hypoglycemia persisted after a second bolus of glucose, then the glucose infusion rate was increased accordingly before subsequent boluses. Cardiopulmonary resuscitation (CPR) was defined as the use of chest compressions with or without epinephrine.8 Clinical chorioamnionitis was defined as a maternal temperature >38°C associated with fetal or maternal tachycardia, uterine tenderness, or foul-smelling amniotic fluid.11 Intrauterine growth retardation was considered when the birth weight was less than the 10th percentile for gestational age.
Differences were assessed by χ2 analysis or Fisher exact test for categorical data and t test or Mann-Whitney rank sum test when normality failed for continuous data. Analyses were performed comparing infants with normal and abnormal neurologic outcome in general, as well as infants with an abnormal neurologic outcome with an initial blood sugar ≤40 mg/dL versus >40 mg/dL. Because of the multiple variables that were associated with adverse outcome, a multivariate logistic analysis was used to ascertain independent predictors of an abnormal outcome (death and/or neurologic outcome as described above). Thus, the predictive potential of birth weight, gestational age, 5-minute Apgar score ≤5 versus >5, the requirement for intubation with or without CPR versus neither, a blood sugar ≤40 mg/dL versus >40 mg/dL, temperature ≥36.5°C versus <36.5°C, pH ≤6.90 versus >6.90, base deficit ≥20 versus <20 mEq/L, and initial mean blood pressure ≤35 mm Hg versus >35 mm Hg was ascertained. Odds ratio (OR) estimates and 95% confidence intervals (CIs) are provided. All other data are expressed as mean ± standard deviation where appropriate. Statistical analysis was performed using Sigmastat (by SPSS Inc, Chicago, IL).
During the 10 years of review, 185 term infants were admitted to the NICU with a cord pH <7.00. Forty-one (22%) of the 185 infants developed an abnormal neurologic outcome, including 14 (34%) with severe hypoxic ischemic encephalopathy who died, 24 (59%) with moderate to severe hypoxic ischemic encephalopathy, and 3 (7%) with seizures as the major manifestation. In 2 of these 3 cases, neuroimaging revealed intracranial hemorrhage. Twenty-seven (15%) of the 185 infants had an initial blood sugar ≤40 mg/dL, and in 158 (85%) infants, the blood sugar was >40 mg/dL. Seven (3.7%) of the 185 infants had evidence of intrauterine growth retardation; none had an initial blood sugar ≤40 mg/dL.
Characteristics of Infants With Abnormal Neurologic Outcome
There were no differences in birth weight (3450 ± 708 g vs 3375 ± 590 g) or gestational age (39.3 ± 1.5 weeks vs 39.7 ± 1.4 weeks) between infants with abnormal and normal neurologic outcomes. Infants with an abnormal versus normal outcome exhibited more fetal heart rate abnormalities (13 [32%] vs 21 [15%]; P = .02), were more likely to be delivered via emergent cesarean section (23 [56%] vs 50 [35%]; P = .01), were more acidotic (pH 6.79 ± 0.10 vs 6.91 ± 0.07; P = .00005), had a larger base deficit (22 ± 4 vs 17 ± 3.8 mEq/L), required more CPR (22 [53%] vs 4 [2.7%]; P = .00001), had a higher proportion of 5-minute Apgar score ≤5 (33 [80%] vs 27 [19%]; P = .00001), had a lower mean arterial blood pressure (41 ± 13 vs 46 ± 10 mm Hg; P = .008), had a lower rectal temperature on NICU admission (36.3°C ± 1.1 vs 36.8°C ± 0.7; P = .001), and had a lower initial blood sugar on NICU admission (76 ± 58 mg/dL vs 108 ± 53 mg/dL; P = .00004; Table 1).
Characteristics of Infants With Initial Blood Sugar ≤40 mg/dL as Compared With Infants With a Blood Sugar >40 mg/dL
Infants with a blood sugar ≤40 mg/dL as compared with >40 mg/dL were of a larger birth weight (3628 ± 707 g vs 3355 ± 592 g; P = .03) but comparable gestational age (39.8 ± 1.6 vs 39.6 ± 1.3 weeks), respectively. Complications during labor were common in both groups, although differences were noted. For example, clinical chorioamnionitis was more common in normoglycemic versus hypoglycemic infants (33 [21%] vs 0 [0%]; P = .03), respectively, whereas diabetes (class B or greater) was more common in the hypoglycemic versus normoglycemic infants (6 [22%] vs 3 [2%]; P = .0003), respectively. There were no differences in the mode of delivery between the groups. Infants with a blood sugar ≤40 versus >40 mg/dL were more likely to receive CPR (8 [30%] vs 18 [11%]; P = .03) and have a lower initial mean blood pressure (37 ± 11 mm Hg vs 46 ± 10 mm Hg; P = .0003), respectively. No differences in arterial pH, base deficit, and initial temperature were noted (Table 2).
Abnormal Neurologic Outcome in Infants With Initial Blood Sugar ≤40 mg/dL as Compared With Infants With an Initial Blood Sugar >40 mg/dL
Fifteen (56%) of the 27 infants with a blood sugar ≤40 mg/dL versus 26 (16%) of the 158 infants with a blood sugar >40 mg/dL had an abnormal outcome (P = .00003; OR: 6.3; 95% CI: 2.6–15.3). The predominant complication during labor in both groups was fetal heart rate abnormalities with meconium-stained amniotic fluid. Infants with a blood sugar ≤40 versus >40 mg/dL had a higher pH (6.86 ± 0.07 vs 6.75 ± 0.09; P = .00008), a lesser base deficit (19 ± 3.7 mEq/L vs 23.8 ± 3.7 mEq/L; P = .0004), and lower mean arterial blood pressure (34 ± 10 mm Hg vs 45 ± 14 mm Hg; P = .006), respectively. The requirement for CPR (7 [46%] vs 15 [57%]), the proportion of infants with a 5-minute Apgar score <5 (11 [73%] vs 22 [85%]), and initial temperature (36.4°C ± 1.3 vs 36.3°C ± 0.7) were comparable between groups (Table 3).
Comparison of Infants Who Had Abnormal Versus Normal Neurologic Outcome and Initial Blood Sugar ≤40 mg/dL
The 15 (56%) infants who had an abnormal outcome as compared with the 12 (44%) infants who had a normal outcome and an initial blood sugar ≤40 mg/dL had a lower pH (6.86 ± 0.07 vs 6.92 ± 0.04; P = .004) but comparable base deficit (−19 ± 4 mEq/L vs −18 ± 2 mEq/L). In addition, they were more likely to require intubation and/or CPR (12 [75%] vs 4 [36%]; P = .04), and a greater proportion had a low 5-minute Apgar score (11 [73%] vs 1 [8%]; P = .001) and lower mean arterial blood pressure (34 ± 10 vs 42 ± 10; P = .006). However, there were no differences in initial temperature (36.3°C ± 0.7 vs 36.7°C ± 0.9) or initial blood sugar (18 ± 15 vs 17 ± 11) between the 2 groups. For all 27 hypoglycemic infants, a blood sugar <40 mg/dL persisted for 140 ± 88 minutes. There was no relationship between the duration of low blood sugar and abnormal versus normal outcome (120 ± 88 vs 200 ± 69 minutes), respectively (Table 4).
Multivariate Analysis for Variables Independently Associated With Abnormal Outcome
Four variables were significantly associated with an abnormal outcome. These were initial blood glucose ≤40 versus >40 mg/dL (P = .001; OR: 18.5; 95% CI: 3.1–111.9), cord arterial pH ≤6.90 versus >6.90 (P = .003; OR: 9.8; 95% CI: 2.1–44.7), a 5-minute Apgar score ≤5 versus >5 (P = .006; OR: 6.4; 95% CI: 1.7–24.5), and the requirement for intubation with or without CPR versus neither (P = .02; OR: 4.6; 95% CI: 1.2–17.9; Table 5).
The data contained in this report are consistent with experimental observations and support an important role for hypoglycemia in the genesis of perinatal brain injury in the term infant with severe fetal acidemia. Thus, the overall incidence of abnormal neonatal neurologic outcome for this high-risk group was ∼20%, whereas it increased markedly to >50% in infants with concomitant hypoglycemia. This translates as an additional 18.5-fold estimated risk for brain injury.
The potential deleterious role of glucose in the genesis of neonatal brain injury has been the subject of intense investigation for many years.2 Experimental data indicate a complex process with developmental differences. Thus, the immature as opposed to mature brain is more resistant to the damaging effects of hypoglycemia.2,3,12,13 Postulated mechanisms that contribute to this phenomenon include an increase in cerebral blood flow and cerebral glucose extraction, enhanced ability to utilize alternate energy substrates and, in particular lactate, and low cerebral energy demands.14–19 Conversely, the concomitant presence of hypoxemia, ischemia, or asphyxia alters this resistance and markedly increases the vulnerability of newborn brain to hypoglycemia. Moreover, the extent of brain injury is related to the method of inducing hypoglycemia. Thus, brain damage is greatest with insulin-induced hypoglycemia as compared with injury in fasted animals.20 Presumably in fasted animals, enhanced ketogenesis serves as an important energy source.21 Regarding the adverse effects of hypoxemia and hypoglycemia, newborn rats experienced a 5-fold reduction in survival, an effect that was minimized with previous glucose administration.22 This deleterious effect is enhanced further when hypoglycemia is associated with ischemia and/or asphyxia.3,22,23 Thus, newborn rats that were subjected to hypoxia-ischemia with concomitant hypoglycemia had more brain injury than control animals, and newborn hypoglycemic piglets that were subjected to a broad range of reduced cerebral blood flow exhibited lower concentrations of cerebral β-adenosine triphosphate as compared with hyperglycemic animals.23,24 It is interesting that in the latter model, hyperglycemia was associated with lower intracellular pH. The more severe acidosis with hyperglycemia is likely related to increased adenosine triphosphate hydrolysis and more H+ production.25 When hypoglycemia is coupled with asphyxia, the cerebral metabolic effects of the latter are magnified.3 Specifically, the increased glycolytic substrates distal to the phosphofructokinase step that result from accelerated anaerobic glycolysis as noted in the normoglycemic state are absent in the hypoglycemic group. These deleterious effects extend at least through the first 4 hours after an asphyxial event if hypoglycemia continues.26 Thus, in newborn asphyxiated lambs, cerebral oxygen consumption remained significantly below control values up to 4 hours with hypoglycemia, whereas in the hyperglycemic group, recovery had occurred by 1 hour.26 In addition to the above cerebral metabolic effects, hypoglycemia has been associated with an impairment of autoregulation.14,27 This state increases vulnerability to ischemic brain injury particularly with concomitant hypotension.2
The clinical observations in this report parallel the above experimental findings indicating an important associated role for hypoglycemia in term acidemic infants with perinatal brain injury. Moreover, this risk for brain injury was increased in hypoglycemic infants who were depressed at birth and required intensive delivery room resuscitation as compared with hypoglycemic infants who had a normal neonatal neurologic outcome. These divergent outcomes occurred despite comparable initial blood sugar values of ∼18 mg/dL that persisted for a similar duration of time. This observation strongly suggests that even when there is biochemical evidence of severe acidemia and concomitant hypoglycemia, abnormal outcome is most likely to occur when there is a reduction in cerebral perfusion at the time of the primary insult and is consistent with our previous findings in this population.7–9 Two related findings in this report are worthy of elaboration. First, although by univariate analysis the initial mean blood pressures were significantly lower in hypoglycemic infants with abnormal as compared with infants with normal outcomes, this effect was not independently associated with outcome with multivariate modeling. We speculate that this may reflect the timing of the blood pressure measurement. Thus, it is conceivable that an earlier measurement (eg, in the delivery room obtained shortly after intensive resuscitation) may have been even lower and clinically relevant. Second, the less severe acidosis (6.86 vs 6.75) observed in infants who had abnormal outcomes and hypoglycemia as compared with those who had normal blood glucose levels, although unanticipated, is similar to the experimental observations of brain biochemical alterations.24
The frequency of initial hypoglycemia in this report was 14.5%. The most common perinatal event associated with hypoglycemia was a history of fetal heart rate abnormalities accompanied by the passage of meconium. This scenario suggests depletion of glycogen stores secondary to stress of unclear duration as a mechanism for the hypoglycemia. The potential for the in utero detection of infants at highest risk for initial hypoglycemia is likely to be difficult, because the same perinatal events were present in a substantial number of infants with normoglycemia. Finally, regarding the potential relationship of glucose and brain injury, it is important to note that the majority of acidemic infants with abnormal outcome: ∼60% were normoglycemic and/or hyperglycemic. This clearly highlights the complexity of factors that contribute to perinatal brain injury, including the relationship to glucose metabolism.
In conclusion, initial hypoglycemia is an important additional risk factor for perinatal brain injury, particularly in the depressed term infant with severe fetal acidemia requiring resuscitation. However, because these data are retrospective and abnormal outcome is short rather than long term, it remains unclear whether earlier detection of hypoglycemia, such as in the delivery room, and aggressive treatment could modify subsequent neurologic outcome. These data raise this possibility and could form the basis for a multicenter, randomized, clinical study of early glucose replacement in high-risk infants with initial hypoglycemia to determine the potential impact on subsequent brain injury.
- Received August 13, 2003.
- Accepted December 29, 2003.
- Reprint requests to (J.M.P.) New York Presbyterian Hospital, Weil Medical College, 525 E 68th St, Rm N-506, New York, NY 10021. E-mail:
- ↵Jones MD, Burd LI, Makowski EI, et al. Cerebral metabolism in sheep: a comparative study in the adult, the lamb and the fetus. Am J Physiol.1975;229 :235– 239
- ↵Volpe JJ. Neurology of the Newborn. 4th ed. Philadelphia, PA: WB Saunders; 2000:497–520
- ↵Sarnat HB, Sarnat MS. Neonatal encephalopathy following fetal distress. A clinical and electroencephalographic study. Arch Neurol.1976;696– 705
- ↵Perlman JM, Risser R. Can asphyxiated infants at risk for neonatal seizures be rapidly identified by current high risk markers? Pediatrics.1996;97 :456– 462
- ↵Gibbs RS, Blanco JD, St Clair PG, et al. Quantitative bacteriology of amniotic fluid from patients with clinical intrauterine infection at term. J Infect Dis.1982;45 :1– 8
- ↵Auer RN. Progress report: hypoglycemic brain damage. Stroke.1986;17 :699– 704
- Hawkins RA, Williamson DH, Krebs HA. Ketone body utilization by adult and suckling rat brain in vivo. Biochem J.1971;122 :13– 18
- ↵Laptook AR, Corbett RJT, Arencibia-Mireles O, Ruley J. Glucose associated alterations in ischemic brain metabolism of neonatal piglets. Stroke.1992;23 :1504– 1511
- ↵Hochachka PW, Mommsen TP. Protons and anaerobiosis. Science.1983;219 :1391– 1397
- ↵Sieso BK. Hypoglycemia, brain metabolism and brain damage. Diabetes Metab Rev.1998;4 :113
- Copyright © 2004 by the American Academy of Pediatrics