DESPITE the many advances in cardiac surgery, neurologic complications continue to be recognized postoperatively. Cognitive deficits appear in about one-half of adults after coronary artery bypass grafting and in as many as one-third of children after neonatal heart surgery. 1,2Preoperative, intraoperative, and postoperative episodes of hypoxia-ischemia all seem to contribute to these complications. Hyperglycemia has been shown to worsen neurologic injury in adult ischemia models. 3Given the risk of ischemic neurologic injury in neonatal heart surgery and the role of hyperglycemia in ischemic brain injury in adults, de Ferranti et al. 's examination of the relationship of blood glucose to neurologic outcome after neonatal heart surgery, published in this issue of the Journal, addresses an important and timely question. 4
To appreciate the distinction between neonates and adults, it is useful to briefly review their differences in whole body and brain glucose metabolism. During development, brain metabolism changes markedly. Glucose crosses the blood-brain barrier through transporter proteins (GLUT1), and then enters the cell through a second glucose transporter system (GLUT3). Glycolysis then begins with the phosphorylation of glucose by hexokinase I. GLUT3 and hexokinase I increase fivefold from neonate to adult as cerebral metabolic rate increases. 5The developmental increase in cerebral glucose metabolic rate corresponds with an increase in synaptic activity, synaptogenesis, and myelination of specific brain regions.
Cerebral glucose metabolism yields adenosine triphosphate, which provides energy to maintain ion gradients, support synaptic activity, and preserve cellular homeostasis. Unlike the adult brain, the neonatal brain is able to metabolize ketone bodies (acetoacetate and D-3-hydroxybutyrate) and free fatty acids to generate adenosine triphosphate under physiologic conditions. The neonatal brain is also able to metabolize lactate to generate adenosine triphosphate for up to 60% of its energy requirements. 6Lactate permeability across the blood-brain barrier is greater in neonates compared with adults, thus supporting brain lactate metabolism and limiting its build-up. 7During ischemia, the neonatal brain is able to use alternative substrates such as lactate and glycogen for energy. 8
A wealth of information from animal models and clinical studies implicates hyperglycemia to be detrimental to the adult brain during global and focal ischemia. 3Although hyperglycemia supports adenosine triphosphate production through glycolysis and delays cellular energy failure during ischemia, the resultant lactic acidosis seems to be toxic to several intracellular processes, thereby hastening cell death and poisoning the repair mechanism of surviving cells.
In contrast to the adult, hyperglycemia in the neonate seems to protect the brain from ischemic damage. In a neonatal rat model of hypoxia-ischemia, Vannucci et al. found that low-dose glucose treatment yielding mild hyperglycemia (270–360 mg/dl) did not exacerbate brain damage; unexpectedly, glucose treatment yielding moderate hyperglycemia (630–720 mg/dl) ameliorated the brain damage in this model. 9Studies in neonatal pigs involving hypothermic low-flow cardiopulmonary bypass or deep hypothermic circulatory arrest also demonstrated less brain damage with higher glucose levels. 10
There are several reasons why hyperglycemia may help the neonatal brain. 11,12First, hyperglycemia increases cerebral high-energy reserves and glycogen stores. As a result, high-energy phosphates are sustained longer during ischemia in hyperglycemic compared to normoglycemic neonatal animals. Second, glucose uptake and metabolism is slower and lactate accumulates slower in the neonatal brain compared with the adult brain. Third, lactate clearance is enhanced, thereby avoiding the toxicity of lactacidosis.
Although many studies have related serum glucose levels to ischemic neurologic outcome in adults, only one clinical study pertains to cardiac surgery. Ceriana et al. found that hyperglycemia was associated with adverse neurologic outcome in adults undergoing aortic arch reconstruction. 13As a result, many cardiac anesthesiologists treat hyperglycemia based on clinical studies of stroke or cardiac arrest and animal studies of ischemia. For pediatric cardiac surgery, the role of hyperglycemia in neurologic injury is even less clear. At the same time, neonates are at additional risk for hypo glycemic neurologic injury.
In neonates, hypoglycemia during fasting or illness is well known and results from several factors. Whole body glucose metabolism corrected for body mass in neonates is up to twice as high as in adults. Hepatic glycogen stores, corrected for body mass, are less in neonates than adults. Gluconeogenic enzymes to convert amino acids to glucose are also inefficient. Neonates suffering from infections or cardiopulmonary disease are at particularly high risk for fasting hypoglycemia. Nicolson et al. randomized infants undergoing heart surgery to receive either lactated Ringer's solution or lactated Ringer's solution with 5% dextrose before cardiopulmonary bypass. 14The group not receiving dextrose had a 5% incidence of hypoglycemia, whereas infants receiving dextrose had no episodes of hypoglycemia. Similarly, de Ferranti et al. administered fluids without dextrose while monitoring serum glucose levels and identified a 9% incidence of hypoglycemia. Consequently, during neonatal cardiac surgery, glucose is often infused intravenously, or if it is not infused, glucose is closely monitored to prevent hypoglycemia.
Although prolonged hypoglycemia is known to cause brain damage, transient hypoglycemia has also been associated with neurologic injury in neonates. Kinnala et al. compared neonates with a history of hypoglycemia with matched controls and found they were four times more likely to display neurologic abnormalities on magnetic resonance imaging or ultrasound scanning. 15Thus, in neonatal heart surgery, preventing hypoglycemia may be more important to improve neurologic outcome than preventing hyperglycemia.
Despite the concern over hypoglycemia in neonates, cardiac surgery is usually associated with hyperglycemia related to the administration of glucocorticoids, hypothermia, and the stress response. Nicolson et al. found similar increases in blood glucose concentrations during cardiopulmonary bypass and following circulatory arrest in both the glucose supplemented and not-supplemented groups. 14These findings indicate that the infusion of dextrose-containing fluids decreases the incidence of hypoglycemia without significantly affecting the incidence of hyperglycemia. However, there has been concern about the resultant hyperglycemia and neurologic injury among pediatric cardiac anesthesiologists.
To address this concern, de Ferranti et al. reviewed the database of a prospective trial conducted between 1988 and 1992, which compared neurologic outcome following surgery using a low-flow cardiopulmonary bypass or deep hypothermic circulatory arrest strategy for the arterial switch operation for D-transposition of the great arteries. 4Although the effect of serum glucose on neurologic outcome was not the initial aim of the study, this database is unique in that it provides a cohort of 171 patients undergoing a similar procedure at one institution with uniform clinical practices. The study protocol included determination of serum glucose levels at specific time points, continuous electroencephalogram monitoring, and neurologic and developmental evaluations at 1, 4, and 8 yr of age. The electroencephalogram and neurologic evaluations were performed by blinded observers, and long-term follow-up was excellent. After examining their data in several different ways, the authors found no relationship between high glucose levels and poor early or late neurologic or developmental outcome. In fact, electroencephalogram activity returned more rapidly following deep hypothermic circulatory arrest in patients with higher glucose levels, and lower glucose levels after deep hypothermic cardiopulmonary bypass was correlated with an increased risk for electroencephalogram seizures, suggesting that higher glucose is better for the neonatal brain than normal or low glucose.
This study does have several weaknesses to temper these conclusions. It was observational and was not originally designed to address the impact of glucose management on neurologic outcome. Although electroencephalogram activity returned earlier in patients with higher glucose levels, the correlation coefficients, although statistically significant, were weak. Further, the relationship between return of electroencephalogram activity and neurologic outcome is uncertain in this setting. Given these issues, the conclusion of high blood glucose concentration being beneficial is tenuous. It is also possible that the serum glucose threshold of 150 mg/dl used by the authors to define hyperglycemia was not the proper “hyperglycemic” threshold to test. Management of cardiopulmonary bypass during pediatric cardiac surgery has significantly changed since this study was performed. Many centers now use regional cerebral perfusion or low-flow bypass instead of circulatory arrest, pH-stat instead of α-stat blood gas management during deep hypothermia, and higher hematocrit levels during cardiopulmonary bypass. These factors clearly reduce ischemic neurologic injury and may therefore lessen the importance of blood glucose on neurologic outcome.
Despite these limitations, this study provides evidence for the lack of association between hyperglycemia and adverse neurodevelopmental outcome after neonatal heart surgery. At the same time, hypoglycemia occurs not infrequently during neonatal heart surgery, and transient hypoglycemia poses a risk of neurologic injury to the immature brain. In light of the clinical and experimental evidence available to date, it is wise to administer dextrose-containing fluids during neonatal heart surgery or, if this is not done, to closely monitor serum glucose levels.