Clinical and Health Affairs
The Effect of Maternal Diabetes during Pregnancy on the Neurodevelopment of Offspring
By Michael K. Georgieff, M.D.
Abstract
As many as 10% of pregnancies are complicated by maternal glucose intolerance. With the risk of diabetes and gestational diabetes rising because of the obesity epidemic, that figure is likely to rise. Many obese individuals suffer from the metabolic syndrome, which makes them more prone to glucose intolerance when they are pregnant. Among the potential risks posed by poor maternal glucose control are those to the developing fetal brain. The goal of this article is to acquaint physicians with the results and clinical implications of studies conducted at the University of Minnesota on outcomes of infants of diabetic mothers and, in particular, on the role of iron deficiency in differential brain processing.
Maternal glucose intolerance complicates up to 10% of pregnancies because of either pre-existing diabetes mellitus or disease that begins during gestation.1 Advances in the understanding of the pathophysiologic changes that take place during diabetic pregnancies have resulted in improved outcomes for the mothers, their fetuses, and their newborn infants. Nevertheless, these pregnancies are still considered higher-risk than routine pregnancies, and for good reasons. The complex metabolic changes that take place in the mother pose potential long-term risks to her offspring. Indeed, the National Institutes of Health recently acknowledged these risks by issuing a request for grant proposals for research dealing with the long-term outcomes of the offspring of women whose pregnancies are complicated by diabetes mellitus or obesity.
Foremost among the potential risks are those to the developing fetal brain. It has long been recognized that maternal diabetes can affect the development of the fetal and neonatal nervous system, with the result being poor neurologic outcomes in offspring.2 Some of these risks have been well-documented in the literature and include birth asphyxia from macrosomia and neonatal hypoglycemia from islet cell hyperplasia and hyperinsulinemia. Others are less well-known and include chronic fetal hypoxia and brain-iron deficiency.3,4 All are a function of the lack of adequate maternal glycemic control during pregnancy, particularly during the last trimester. In fact, it matters little whether the glucose intolerance is caused by pregestational or gestational diabetes. Some factors, such as macrosomia, hypoglycemia, and iron deficiency result from worsening glycemic control late in pregnancy, during the two weeks prior to delivery.
The goal of this article is to acquaint physicians with the results and clinical implications of studies conducted at the University of Minnesota on outcomes of infants of diabetic mothers (IDMs) and, in particular, on the role of iron deficiency in determining these outcomes.
Metabolic Changes in the Fetus during Diabetic Pregnancies
Health care professionals have worked hard during the last 40 years to encourage women with diabetes mellitus to pay particularly close attention to their glycemic control during gestation. As a result of those efforts, maternal and fetal/neonatal morbidity from this disease has decreased markedly. For mothers without pre-existing diabetes mellitus, glucose screening at 28 weeks’ gestation to diagnose gestational diabetes has become routine. This test does not identify all cases, however, as some women become glucose-intolerant later in pregnancy. Moreover, the risk of gestational diabetes is likely rising because of the obesity epidemic among teenagers and young adults. Many of these individuals suffer from the metabolic syndrome, which is characterized by increased insulin resistance, making them more prone to glucose intolerance when they are pregnant.
The fetus undergoes significant metabolic changes when the mother is hyperglycemic. Glucose traverses the placenta easily; thus, if the mother has frequent or constant hyperglycemia, so does the fetus. The fetus, however, has a normal pancreas after 20 weeks gestation and thus is capable of mounting a hyperinsulinemic response to control the hyperglycemia.
In sheep, both fetal hyperglycemia and hyperinsulinemia increase the metabolic rate by about 30%, requiring the fetus to consume oxygen at a 30% higher rate.5,6 Neither the ovine nor the human placenta is capable of upregulating oxygen delivery to meet this higher demand, rendering the fetus chronically hypoxic. Studies of the fetal lamb and rhesus monkey with indwelling catheters demonstrate profound fetal hypoxemia.3,7 The evidence of a similar process in humans is provided by elevated cord serum erythropoietin concentrations and a 20% incidence of polycythemia in newborn IDMs.8 This chronic fetal hypoxia could exist for weeks prior to delivery and potentially affect fetal brain development. Recently, Raghu Rao, M.D., at the University of Minnesota Medical School and Lakshmi Raman, M.D., at Hennepin County Medical Center, detailed the neurochemical and structural abnormalities in the hippocampus of chronically hypoxic developing rats, suggesting that this area of the brain, which is central to recognition memory and learning, is altered by long-term hypoxia.9 In particular, the truncation of neuronal dendritic structure was remarkable. One, therefore, has to consider that a human IDM who is similarly chronically hypoxic may also suffer from altered fetal brain development.
One consequence of chronic intrauterine hypoxia is alteration of fetal iron metabolism and subsequent brain-iron deficiency. Iron deficiency has long been known to alter neurodevelopment both during the time the patient is iron deficient and up to 20 years after iron repletion. Work by Amos Deinard, M.D., and Pi-Nian Chang, Ph.D., at the University of Minnesota in the 1980s demonstrated the adverse effects of low dietary iron during toddlerhood on developmental test scores.10 More than 40 subsequent studies have shown the negative short- and long-term effects of postnatal dietary iron deficiency on development.
Fetal brain iron deficiency occurs as a consequence of the combination of chronic intrauterine hypoxia and rapid somatic growth—both sequelae of poor maternal glycemic control. A basic principle in all mammals that have been studied is that iron is prioritized to the red blood cells over all other organs for the synthesis of hemoglobin. Initially, when the demand for iron exceeds the supply, pools of stored ferritin-bound iron located predominantly in the liver are catabolized in order to provide a steady supply to all organs that require iron including the muscles, heart, and brain. However, when iron stores are reduced by more than 80%, the body prioritizes where to send the remaining stored iron as well as dietary iron. The red cells are at the top of the hierarchy, followed by the brain, the heart, and the skeletal muscles.
We have measured iron concentrations at autopsy in infants born to iron-sufficient diabetic mothers and found a complete loss of hepatic storage iron, a 55% reduction in heart iron, and a 40% reduction in brain iron.4 These infants represent the far end of the spectrum. Assaying serum ferritin concentrations from cord blood in newborn IDMs from unselected, consecutive pregnancies, we found that 65% had ferritin concentrations below the 5th percentile (60 mcg/L in newborns). The mean value was 26 mcg/L, and many had ferritin concentrations of 0 mcg/L.11 We have been able to estimate from existing nomograms that newborn ferritin concentrations less than 35 mcg/L represent a risk to the developing brain and have used that value as a cutoff in our neurodevelopmental studies of IDMs.12
We subsequently assessed the amount of iron in the red cells of IDMs and determined that the infants with the lowest ferritin concentrations had the highest hematocrits at birth.8 Furthermore, we and others have shown that the neonatal hematocrit in IDMs is directly related to poor maternal glycemic control. In our study, we measured glycosylated fetal hemoglobin to assess how much glucose the fetus was chronically exposed to and found the values directly correlated with elevations in red-cell iron content, cord erythropoietin concentrations, and degree of iron abnormalities.8
The Effect of Diabetes on Neurodevelopment
Knowing that early postnatal iron deficiency has a significant negative effect on brain development, our group has sought multiple grants from the National Institutes of Health over the past 11 years to study memory function and hippocampal development in IDMs and in a rat model of 40% fetal/neonatal brain-iron deficiency.
The results of our study of recognition memory in newborn infants were fascinating. The advantage of being able to study memory ability during the newborn period is that it firmly links any neurologic abnormalities to fetal and neonatal events, as opposed to testing children years later after many confounding variables have been introduced. The neural circuitry underlying the recognition memory system is well-developed at birth, and its integrity can be tested behaviorally and electrophysiologically using standard EEG montages.
Infants recognize their mother’s voice as evidenced by increased sucking and heart rate changes (alas, this is not true for the father’s voice). The assumption is that the fetus encodes the mother’s voice in utero. Charles Nelson, Ph.D., a child psychologist now at Harvard but who was a faculty member at the University of Minnesota’s Institute of Child Development for 19 years and Raye-Ann deRegnier, M.D., a neonatologist who trained at the University of Minnesota and is now at Northwestern University, pioneered the use of event-related potentials (ERPs) in neonates. ERPs are microvolt changes in brain electrical activity that can be detected against the background of the general millivolt brain activity seen on typical EEGs. The pattern of these microvolt changes differs based on the familiarity or novelty of a presented memory stimulus. This group had previously used visual ERP paradigms in children as young as 4 months of age to test recognition memory. In these trials, an infant would be shown pictures of his or her mother interspersed with pictures of an unknown woman. Electrodes placed on the infant’s skull recorded the EEG. The researchers noted a unique and reproducible response within hundreds of milliseconds of the infant being shown each of the pictures. Less activity was seen when the infants processed their mother’s face presumably because it is familiar and needs no further updating. Greater activity was seen when they were shown the unfamiliar face, indicating that more processing was taking place.
Neonates don’t see particularly well at birth, nor do they hold their heads up without support. For those reasons, Nelson and deRegnier adapted the visual ERP paradigm to an auditory ERP—using the mother’s voice as the familiar stimulus. Two- to 10-day-old IDMs and control infants were played 50 trials of their own mother’s voice interspersed randomly with 50 trials of an unfamiliar woman’s voice.13 The same type of differential brain processing occurred with the auditory stimuli in the control newborns as was seen with the visual stimuli in the older infants. However, the IDM group did not show evidence of discrimination, suggesting abnormal hippocampal processing. Further testing revealed that it was the iron-deficient IDMs (ie, those born with ferritin concentrations less than 35 mcg/L) that showed no ability to discriminate, while iron-sufficient IDMs processed the stimuli in the same way as did the controls.12
The infants with low ferritin concentrations spontaneously replete their iron status within their first 9 postnatal months whether they are breastfed or formula-fed.14 Nevertheless, neurodevelopmental follow-up of these infants at 6 months, 8 months, 1 year, and 3 years reveals that they continue to process recognition memory events differently and less efficiently than the controls.15-17 Interestingly, the researchers found the iron-deficient IDMs do not have lower developmental quotients (the infant equivalent of IQ) nor do they exhibit any particular behavioral abnormalities. Yet, clearly, they are “wired” differently, and their activity as they grow older bears watching, particularly in light of the long-standing literature about poorer development in IDMs.
When these findings first appeared in press, there was concern that the newborn iron-deficient IDMs did not “recognize” their mothers. It was important for us to emphasize that we did not have behavioral evidence of that, rather that the brains of the IDMs appeared to work differently than those of the control infants. Indeed, electrophysiologic maps constructed from 128-lead studies performed in 3-year-olds who had been iron deficient or iron sufficient at birth continue to demonstrate this processing difference.
We are just beginning new follow-up studies on this original cohort of children, who are now 7 to 9 years of age, using much more sophisticated neuropsychological test batteries and functional MRI to understand how their brains process memory and learning events.
Whether neonates may be candidates for iron therapy (beyond the usual American Academy of Pediatrics recommendations) is the subject of another new study involving newborn IDMs at the University of Minnesota Children’s Hospital-Fairview, St. Paul Children’s Hospital, and Hennepin County Medical Center.
What is the Neurobiology?
We wondered how the hippocampus of the iron-deficient newborn might be structurally, neurochemically, and physiologically different than that of an iron-sufficient neonate. In order to better understand the potential differences, we used a rat model of fetal/neonatal iron deficiency (a 40% reduction in total brain iron) similar to the deficiencies seen in the iron-deficient IDMs in our study. In the rat model, iron deficiency was induced through the mother’s diet as opposed to through diabetes mellitus (a difficult model in the pregnant rat).
The animals were studied in the Center for Magnetic Resonance Research at the University of Minnesota by Rao using high-field (9.4 Tesla) magnetic resonance spectroscopy. This technique allows us to visualize the neurochemistry of the developing hippocampus in the same animal over time. As we studied the rat pups from postnatal day 7 (equivalent to a premature infant born at 34 weeks’ gestation) to postnatal day 28 (equivalent to a toddler), we found significant changes in metabolites that are markers of neurotransmission (glutamate and GABA), myelination, and energy metabolism in the iron-deficient pups.18 Interestingly, these processes all rely on iron-dependent proteins in their metabolic pathways. Structurally, the hippocampal neurons visualized under the microscope were abnormal with truncated, gnarled dendrites.19 Signaling molecules such as alpha-CaMKinase II, which are important for memory processing, were reduced or abnormally localized. Long-term potentiation is the electrophysiologic footprint of memory formation recorded from the pyramidal cells of area CA1 in the hippocampus. Its developmental appearance was delayed and was ultimately lower in the iron-deficient rats.20 Behaviorally, the iron-deficient rats could not learn the mazes as quickly or efficiently as those that were iron-sufficient.21
Our multi-tiered approach convinced us that fetal/neonatal iron deficiency caused fundamental changes in the way the rat hippocampus developed, resulting in abnormal recognition memory behaviors while the rats were iron-deficient and long after repletion. It suggested that the developmental trajectory had been thrown off at a critical period, thus permanently altering this type of memory processing. Most concerning to us was the fact that starting iron treatment at postnatal day 10 (the hippocampal equivalent of full-term birth in humans) did not result in complete neurobehavioral recovery in spite of complete recovery of iron status.
Does this mean that all of these biochemical and structural abnormalities are present in the brains of iron- deficient newborn IDMs? Of course, we don’t know the answer to that. Some of those possibilities may be studied using high-end neuroimaging techniques that are becoming available at the Center for Magnetic Resonance Research. Ultimately, however, the relationship between the degree of iron abnormalities and cognitive function at birth and poor maternal glycemic control is striking. It once again underscores what health care practitioners have emphasized for decades—that keeping pregnant women with diabetes under good glycemic control is crucial to reducing risks to both the mother and the fetus. Our work emphasizes that in the fetus, significant neurologic and hematologic risks can be prevented. Given the possibility that postnatal iron treatment may not induce full recovery, an ounce of prevention truly is worth a pound of cure.
Clinical Implications
We hope physicians take away two messages related to clinical practice from this article. First, maternal gestational glucose control is crucial to reducing neurodevelopmental risks to the fetus and the newborn. To that end, we need to be more vigilant in screening for gestational diabetes, given the current obesity epidemic. We may ultimately need to consider rescreening pregnant women who are at higher risk (eg, obese) for glucose intolerance after the usual assessment at 28 weeks of gestation. Second, health care practitioners who care for pregnant women with diabetes could bolster the argument for tight glucose control by informing them of the risks of fetal hypoxia and iron deficiency to their offspring, particularly as these conditions relate to their infant’s long-term development. MM
The author wishes to thank all of his collaborators, especially Charles Nelson, Ph.D., Raghavendra Rao, M.D., and Raye-Ann deRegnier, M.D., the technicians and students, and the General Clinical Research Center staff who made this research possible. The author acknowledges the agencies that funded this work, including the National Institutes of Health (grants HD-29421, NS-34458, RR- 00400), the Minnesota Medical Foundation, and the Viking Children’s Fund.
Michael Georgieff is a professor of pediatrics and child development and director of the Center for Neurobehavioral Development at the University of Minnesota.
References
1. Nold J, Georgieff MK. Infants of Diabetic Mothers. In Rademacher R, Kliegman R, Eds. Pediatric Clinics of North America. Philadelphia: W.B. Saunders, 2004;51:619-37.
2. Ornoy A, Ratson N, Greenbaum C, et al. School-age children born to diabetic mothers and to mothers with gestational diabetes exhibit a high rate of inattention and fine and gross motor impairment. J Pediatr Endocrinol Metab. 2001;14(Suppl 1):
681-9.
3. Widness JA, Susa JB, Garcia JF, et al. Increased erythropoiesis and elevated erythropoietin in infants born to diabetic mothers and in hyperinsulinemic rhesus fetuses. J Clin Invest. 1981;67:637-42.
4. Petry CD, Eaton MA, Wobken JD, Mills MM, Johnson DE, Georgieff MK. Iron deficiency of liver, heart, and brain in newborn infants of diabetic mothers. J Pediatr. 1992;121(1):109-14.
5. Philipps AF, Porte PJ, Strabinsky S, Rosenkranz TS, Raye JR. Effects of fetal hyperglycemia upon oxygen consumption in the ovine uterus and conceptus. J Clin Invest. 1984;74:279-86.
6. Milley JR, Papacostas JS, Tabata BK. Effect of insulin on uptake of metabolic substrates by the fetus. Am J Physiol. 1986:251 (3Pt1):E349-59.
7. Georgieff MK, Widness JA, Mills MM, Stonestreet BS. The effect of prolonged intrauterine hyperinsulinemia on iron utilization in fetal sheep. Pediatr Res. 1989; 26(5):467-9.
8. Georgieff MK, Landon MB, Mills MM, Hedlund BE, Faassen AE, Schmidt RL, Ophoven JJ, Widness JA. Abnormal iron distribution in infants of diabetic mothers: spectrum and maternal antecedents. J Pediatr. 1990;117(3):455-61.
9. Raman L, Tkac I, Ennis, K, Georgieff MK, Gruetter R, Rao R. In vivo effect of chronic hypoxia on the neurochemical profile of the developing rat hippocampus. Dev Brain Res. 2005;156(2):202-9.
10. Deinard AS, List A, Lindgren B, Hunt JV, Chang PN. Cognitive deficits in iron-deficient and iron-deficient anemic children. J Pediatr. 1986;108 (5Pt1):681-9.
11. Chockalingam UM, Murphy E, Ophoven JC, Weisdorf SA, Georgieff MK. Cord transferrin and ferritin values in newborn infants at risk for prenatal uteroplacental insufficiency and chronic hypoxia. J Pediatr. 1987;111(2):283-6.
12. Siddappa AM, Georgieff MK, Wewerka S, Worwa C, Nelson CA, deRegnier RA. Iron deficiency alters auditory recognition memory in newborn infants of diabetic mothers. Pediatr Res. 2004;55:1034-41.
13. deRegnier RA, Nelson CA, Thomas KM, Wewerka S, Georgieff MK. Neurophysiologic evaluation of auditory recognition memory in healthy newborn infants and infants of diabetic mothers. J Pediatr. 2000;137(6):777-84.
14. Georgieff MK, Wewerka SW, Nelson CA, deRegnier RA. Iron status at 9 months of infants with low iron stores at birth. J Pediatr. 2002;141(3):405-9.
15. Nelson CA, Wewerka S, Thomas K, Tribby-Walbridge S, deRegnier RA, Georgieff MK. Neurocognitive sequelae of infants of diabetic mothers. Behav Neurosci. 2000;114(5):950-6.
16. Nelson CA, Wewerka SS, Borscheid AJ, deRegnier RA, Georgieff MK. Electrophysiologic evidence of impaired cross-modal recognition memory in 8-month-old infants of diabetic mothers. J Pediatr. 2003;142(5):575-82.
17. DeBoer T, Wewerka S, Bauer PJ, Georgieff MK, Nelson CA. Explicit memory performance in infants of diabetic mothers at 1 year of age. Dev Med Child Neurol. 2005;47(8):525-31.
18. Rao R, Tkac I, Townsend EL, Gruetter R, Georgieff MK. Perinatal iron deficiency alters the neurochemical profile of the developing rat hippocampus. J Nutr. 2003;133(10):3215-21.
19. Jorgenson LA, Wobken JD, Georgieff MK. Perinatal iron deficiency alters apical dendritic growth in hippocampal CA-1 pyramidal neurons. Dev Neurosci. 2003;25(6):412-20.
20. Jorgenson LA, Sun M, O’Connor M, Georgieff MK. Fetal iron deficiency disrupts the maturation of synaptic function and efficacy in area CA1 of the developing rat hippocampus. Hippocampus. 2005;15(18): 1094-102.
21. Schmidt AT, Waldow KJ, Salinas JA, Georgieff MK. The long-term behavioral effects of fetal/neonatal iron deficiency on a hippocampally dependent learning task in the rat. Pediatr Res. 2004;55:279A.