There are compelling preclinical data that common general anesthetics cause increased neuroapoptosis in juvenile animals. Retrospective studies demonstrate that young children exposed to anesthesia have school difficulties, which could be caused by anesthetic neurotoxicity, perioperative hemodynamic and homeostatic instability, underlying morbidity, or the neuroinflammatory effects of surgical trauma. Unnecessary procedures should be avoided. Baseline measures of blood pressure are important in determining perioperative blood pressure goals. Inadvertent hypocapnia or moderate hypercapnia and hyperoxia or hypoxia should be avoided. Pediatric patients should be maintained in a normothermic, euglycemic state with neutral positioning. Improving outcomes of infants and children requires the collaboration of anesthesiologists, surgeons, pediatricians and neonatologists.
Key points
- •
There are compelling preclinical data that commonly used general anesthetics cause neuroapoptosis in juvenile animals in vitro and in vivo.
- •
Many but not all retrospective clinical studies in children demonstrate an association between anesthesia exposure at a young age and later neurocognitive difficulties.
- •
The only prospective randomized trial to date did not find an association between general anesthesia exposure in early infancy for inguinal hernia repair and neurocognitive difficulties.
- •
However, this study did not address the effects of prolonged anesthetic exposure during early childhood. Perioperative factors such blood pressure, arterial carbon dioxide tension, temperature, glycemic state, arterial oxygen saturation, and positioning are modifiable during anesthesia and, if abnormal, can lead to possible brain injury.
Concerns about long-term anesthetic-induced neurotoxicity have permeated the pediatric anesthesia community since landmark laboratory reports demonstrated neurodegeneration as well as behavioral deficits in juvenile animals exposed to general anesthetics and sedatives. , These preclinical findings led the US Food and Drug Administration in 2016 to issue a safety warning as well as change the labeling of general anesthetics and sedatives to reflect these concerns. The safety warning states that “repeated or lengthy use of general anesthetic and sedation drugs during surgeries or procedures in children younger than 3 years or in pregnant women during their third trimester may affect the development of children’s brains.” Regulatory bodies and medical societies in other countries have declined to issue specific warnings about the safety of general anesthetic exposure in young children, reflecting the lack of compelling evidence that general anesthetic exposure at a young age leads to clinically relevant anesthetic-induced neurotoxicity.
Measuring the impact of general anesthesia on children undergoing procedures is difficult owing the many confounding factors. These confounders include the underlying pathology of the child which leads to the need for anesthesia for diagnostic tests and procedures as well as the neuroinflammatory effects of surgery. This neuroinflammation has been associated with postoperative neurocognitive deficits in older adults. There has been only a single randomized prospective trial published examining the effects of general anesthesia compared with regional anesthesia for hernia repair on young infants. Difficulty enrolling patients in a control group has limited the ability to conduct more research. , There is an ongoing international randomized controlled trial known as the TREX trial comparing sevoflurane alone with low-dose sevoflurane, remifentanil and dexmedetomidine for procedures lasting longer than 2 hours.
One of the concerning negative effects of general anesthesia is neurotoxicity leading to cognitive dysfunction. However, other factors such as environmental factors and genetic background have a significant impact on neurodevelopment and are beyond the control of the clinician during a perioperative encounter. , In addition, the impact of surgery and anesthesia can provoke homeostatic variability and instability that can lead to brain injury. Factors that can contribute to this instability include alterations in blood pressure, Pa co 2 , blood glucose levels, temperature, Pa o 2 , and head positioning during procedures. In this review, we discuss the preclinical as well as clinical data available about the neurotoxic effects of general anesthesia on the developing brain as well as clinical data linking abnormal homeostasis perioperatively or in the intensive care unit with abnormal neurologic development.
Preclinical data
The quest for determining whether commonly used general anesthetics in pediatrics causes long-term harm began about 20 years ago with reports that infant mice exposed to general anesthesia had increased neuronal apoptotic cell death and long-term behavioral and cognitive deficits. Apoptotic cell death is a normal developmental process whereby an organism rids itself of unnecessary cells. Both in vitro and in vivo studies have shown that very young animals exposed to either GABA agonists (eg, volatile anesthetics, benzodiazepines, barbiturates and propofol) as well as N-methyl- d -aspartate (NMDA) antagonists (nitrous oxide and ketamine) develop abnormally high levels of neuronal apoptosis. The vast majority of the animal studies have been on very young murine species who show an increased susceptibility to neuroapoptosis at postnatal day 7 of life. Slightly older murine species (postnatal days of life 14–20) exposed to these same general anesthetics develop increased but abnormal dendritic growth. , The clinical significance of this abnormal dendritic formation is unknown, but some human psychiatric and neurologic disorders such as autism spectrum disorder and schizophrenia are associated with abnormal dendritic formation and synaptogenesis. Although most of these studies have been performed in rats and mice, there is also a convincing body of evidence that rhesus monkeys have increased neuroapoptosis when exposed to anesthetic drugs as fetuses or within the first 5 days of life. ,
Anesthetic-induced neuroapoptosis has been found throughout the central nervous system and regional changes occur at different time points during development. Exposure to general anesthetics also affects neuronal generation by causing abnormal neurogenesis in rats and mice. Glial cell formation is decreased and maturation is delayed by general anesthetic exposure.
Laboratory animals such as rats, mice, guinea pigs, and rhesus monkeys exposed to anesthetic drugs in the neonatal period develop impairments in learning, memory, attention, emotional behavior, psychomotor speed, concept formation, motivation, emotional behaviors, and motor function. Studies in rhesus monkey have shown that exposure to general anesthesia during infancy leads to animals with heightened emotional reactivity to perceived threats such as contact with strangers, as well as overall increased anxiety. In addition, exposure to general anesthesia may lead to late findings in monkeys. Repeated exposure to general anesthetics in rhesus monkeys can lead to visual memory deficits that become apparent only after the first year of life.
One of the possible mechanisms is that neuronal cells during exposure to general anesthesia are deprived of stimulating input and thus deprived of trophic support. This trophic or growth support is necessary for neurogenesis, synaptogenesis, and other context-sensitive modulation necessary for neuronal plasticity. This trophic factor dysregulation leads vulnerable cells toward an apoptotic pathway. General anesthetics can also cause mitochondrial dysfunction in immature animals, which can stress neuronal cells. Aberrant cell cycle entry, where neurons abnormally enter the mitotic cell cycle and perish, is another mechanism observed with exposure to NMDA receptor antagonists. Volatile anesthetics induce excitotoxicity and seizures by binding to immature GABA receptors and activating the cotransporter protein NKCC1, leading to chloride influx and neuronal depolarization. With maturity, this protein is replaced by KCC2, which actively transports chloride out of the cellular membrane causing inhibition of depolarization. In humans, this protein receptor starts to switch around 15 weeks after birth, but is not complete until 1 year of age. KCC2 expression determines the developmental stage-dependent impact of propofol anesthesia on the brain growth spurt. Seizures can occur as well as increased neuroapoptosis in neonatal rats exposed to sevoflurane. This neurotoxicity has been successfully mitigated with pretreatment with bumetanide, which blocks the NKCC1 channels. In addition, epigenetic modulation of transcription can be altered by general anesthetics, which can impede neurodevelopment.
In addition to bumetanide, there have been several other nonspecific medications, which have been shown to ameliorate the effects of general anesthesia in animal models. These drugs include estradiol, erythropoietin, melatonin, l -carnitine, lithium, xenon, and dexmedetomidine. Currently, dexmedetomidine is the only medication commonly used as a sedative or adjunctive anesthetic in pediatric medicine that has been shown to mitigate anesthetic-induced developmental neurotoxicity Dexmedetomidine has been found in animal studies to mitigate the effects of isoflurane in rodent models, but high doses of dexmedetomidine in these same models induces a modest increase in neuroapoptosis. ,
Although the preclinical animal data are compelling, a meta-analysis of more than 400 preclinical reports found no clear exposure duration threshold below which no injury or subsequent cognitive abnormalities followed. This meta-analysis did not clearly identify a specific age beyond which anesthetic exposure did not cause any structural or functional abnormalities. The majority of animal studies that demonstrated increased histologic and behavioral derangements involve exposures of 4 hours or more, which has been associated with severe physiologic changes that can trigger neurodegenerative processes. , It is also important to note that several animals studies have compared general anesthetic exposures with no exposures in groups of young animals who were in either controlled or enhanced environments. , The effects of enhanced environments (cages with wheels and other toys) to a very large degree mitigated the neurotoxic effects of general anesthesia in rodent models.
Translating the findings of the preclinical studies to young children is problematic. The sensory and social deprivation seen in the rodent experimental paradigms, which use controlled environments (sterile cage, solitary confinement), is not a natural environment and thus calls into question the validity of developmental studies on rodents in such environments. It is also unknown whether the prolonged neurodevelopment of human beings as well as the generally enriched environments that children live in compared with other species may allow for restoration of function of any damage caused by general anesthetic. It is also unknown which neurodevelopmental domains would be affected in humans compared with animals.
Clinical studies
There are many human studies that show an association between anesthetic exposure at a young age and later learning or behavioral difficulties. In contrast, there are also many studies that show no such associations. The outcome measures used include comprehensive neurocognitive testing, parental reports of school and medical issues, behavioral or medical diagnoses in large databases often derived from billing codes, or school reports of academic performance or readiness. The age of exposure has also varied in these reports ranging from infancy to about 5 years of age. Another difficulty encountered is that some of the neurocognitive domains such as executive function and some social emotional skills are not reliably tested in humans until early adolescence. Finally, it is important to consider the type of studies that have been reported. The vast majority have been retrospective in nature, although there have been some ambidirectional trials in which a retrospective cohort is developed and then prospectively evaluated and a single randomized control trial. In this review, we highlight some of the important trials that address the effects of anesthesia on neurodevelopment.
Retrospective Trials
One of the first studies published about this topic was in 2009 was from the Mayo Clinic. Because of the stability of the population of Olmsted County, Minnesota, a wealth of information is known about school age children, including their health and school records. The first retrospective birth cohort study of children born between 1976 and 1982 found that more than 1 exposure to general anesthesia before the age of 4 years was associated with almost double the cumulative risk of a school referral for a learning difficulty. A single exposure was not associated with an increased risk. This study was criticized because the children mostly received halothane (an anesthetic generally not used anymore) and intraoperative pulse oximetry was not yet available. A similar cohort was developed examining a later birth cohort born between 1996 and 2000 who were exposed to contemporary anesthetics and monitoring in children less than 3 years of age. Again, the findings were that a single exposure to general anesthesia did not confer an increased risk of school difficulties, but further exposures did.
The Western Australia Pregnancy (Raine) cohort, which was originally developed as randomized controlled trial to test the effects of fetal ultrasound examination on birth outcomes, has become one of the largest prospective cohorts of pregnancy and childhood. Extensive secondary analysis has been done on this cohort looking at the effects of early anesthetic exposure and neurocognitive outcomes. From the original cohort of 2868 children, 148 children were anesthetized before the age of 3 years. When tested at age 10 years, the exposed children were found to have increased risks of language deficits and overall intelligence, even after a single exposure. Further studies were done on this cohort using latent class analysis, in which the authors defined 4 classes of neurodevelopmental outcomes: normal or nearly normal testing, language and cognitive deficits, behavioral deficits, and global deficits. They found that only language and cognitive deficits were associated with exposure to general anesthesia before the age of 3 years.
There have been 2 large database studies from Canada looking at school readiness using the Early Developmental Instrument, a 104-component questionnaire encompassing 5 developmental domains to assess developmental skills. , Both studies found a modest increase in risk for early developmental vulnerability for single anesthetic exposures in children, but only when the exposure was between the ages of 2 and 4 years. Exposure before the age of 2 did not confer any increased risk. A large Swedish cohort of more than 2 million individuals born between 1973 and 1993 found a modest decrease in overall IQ after single or multiple exposures to general anesthesia before the age of 4 years in military recruits who were tested at age 18 years.
At least 2 retrospective studies used siblings as a control group to attempt to control for inherited traits, socioeconomic, and cultural factors. One of the studies published in 2009 examined the neurocognitive and academic outcomes of identical twins in the Netherlands from the National Twin Registry. The authors examined general anesthetic exposure before the ages of 3 and 12 years. This study found that anesthetic exposure was a risk factor for poorer academic outcomes, but in discordant twin pairs of whom one was exposed and the other nonexposed, the academic outcomes were identical. The authors concluded that it was unlikely that general anesthesia exposure per se was a cause of poor academic outcomes. Another sibling study using the cohort developed in Ontario, Canada, found that there was no increased risk of developmental vulnerabilities in siblings exposed to general anesthesia compared with nonexposed siblings.
Ambidirectional cohort studies
The Pediatric Anesthesia & NeuroDevelopment Assessment (PANDA) study tested the neurocognitive outcomes of 105 sibling pairs who were born within 36 months of each other and in which one of the siblings was exposed to general anesthesia for inguinal herniorrhaphy. There were no meaningful differences in IQ, the primary outcome in this cohort, who were tested between ages 8 and 15 years. There were also no differences in mean scores between sibling pairs in memory/learning, motor/processing speed, visuospatial function, attention, executive function, language, or behavior. This trial highlights the difficulties in doing ambidirectional trials. Almost 10,000 sibling pairs were screened for this study, but only 105 sibling pairs were enrolled.
The Mayo Anesthesia Safety in Kids (MASK) study enrolled a total of 997 children: 411 were unexposed, 380 were singly exposed, and 206 were exposed more than once before the age of 3 years to determine whether anesthetic exposure was related to poor neurocognitive outcomes. The children were propensity matched and underwent neuropsychologic testing between ages 8 to 12 years or 15 to 20 years. There were no significant differences found in the primary outcome measure, the full-scale IQ as determined by the Wechsler Abbreviated Scale of Intelligence. For secondary outcomes, children who had multiple anesthetic exposures were at higher risk for decreased processing speed and fine motor abilities. Parental reports of the exposed children revealed an increased risk of difficulties with executive function as well as reading for children with both single and multiple exposures. The children of this cohort were also tested by the operant test battery, in which subjects operate a lever to complete a task and are rewarded with a small treat for task completion. This type of testing has been done with nonhuman primates exposed to ketamine in early infancy who demonstrated deficits in motivation as well as deficits in accuracy of task performance and response speed. The operant test battery did not detect a difference between the control and exposed groups in the MASK cohort.
Randomized, Controlled Trials
The General Anesthesia versus Spinal Anesthesia (GAS) trial, a prospective randomized controlled equivalence trial, compared the neurocognitive outcomes of infants less than 60 weeks postmenstrual age who were exposed to either sevoflurane general anesthesia or bupivacaine regional anesthesia for inguinal hernia repair. An interim analysis done at age 2 and using the Bayley III developmental test found no differences between the 2 groups. The full-scale IQ using the Wechsler Primary and Preschool Scales of Intelligence was performed at age 5 years and also found no differences between the 2 groups. The data was analyzed on a per-protocol and intention-to-treat basis. Criticisms of this trial include the medium length of exposure (55 minutes), mostly male enrollees (>80%), and the time of general anesthetic exposure was during a relatively narrow time during development.
The TREX trial is currently prospectively enrolling subjects in the United States, Australia, New Zealand, and Great Britain. In this trial, patients are randomized to receive either standard dose sevoflurane as an aesthetic agent or low-dose sevoflurane, remifentanil, and dexmedetomidine for procedures expected to last longer than 2 hours in children less than 2 years of age. Neurocognitive testing is being done at age 3 years.
Perioperative factors associated with adverse pathophysiology in pediatric patients
There are many factors that may lead to neurologic injury in susceptible patients undergoing procedures. Some of these factors may not be easily modifiable, such as the effects of neuroinflammation or abnormal circulatory patterns owing to congenital heart disease, but many factors are modifiable. Altered cerebral circulation can occur from high intrathoracic pressure or head positioning, leading to decreased cerebral venous drainage. Hypercapnia or hypocapnia cause cerebral arterial vasodilation or vasoconstriction. Low blood pressure can lead to inadequate cerebral perfusion pressure. Metabolic cellular insufficiency can result from inadequate metabolic fuel (hypoglycemia or hypoxia) or from unmet metabolic demand such as occurs with pain, stress, fever, or seizures. Neurotoxic mediators produced from hypoxic or ischemic neurons, as well as free radicals produced during periods of hyperoxia, can promote brain injury.
It is important to recognize that there may be some children who may be at higher risk for perioperative brain injury. There are many illnesses and conditions that put pediatric patients at risk for neuronal injury perioperatively. We discuss a few of these conditions to elucidate the manner in which a pediatric patient’s underlying preoperative state may influence the anesthesiologist on methods to optimize the conduct of the general anesthetic.
Premature infants are at high risk for perinatal brain injury for many reasons, and the effects of intraoperative instability may increase this risk. These infants are born with much lower levels of neuroprotective hormones such as increased placental levels of estrogen, progesterone, and oxytocin, which increase during the third trimester of development and peak around 40-weeks of gestation. Levels of progesterone and estrogen increase 100-fold during the last trimester of fetal development. Estrogen acts as an antioxidant; promotes growth of dendrites, axons, and synapses; promotes the expression of neurotrophic factors; and acts as an antiapoptotic agent. Progesterone decreases postischemic inflammation and induces the production of brain-derived neurotrophic factor. Oxytocin in fetal rats promotes an excitatory to inhibition switch in GABA actions, further protecting the developing brain.
Premature infants are at risk for ventricular intravascular hemorrhage because of the immature and fragile nature of the choroid plexuses in the lateral cerebral ventricles. Alterations in cerebral perfusion can lead to rupture of these delicate blood vessels, which ultimately lead to a loss of white matter if there is intracerebral parenchymal hemorrhage or hydrocephalus if there is obstruction to cerebral spinal fluid flow. The time of maximum vulnerability for this injury is the first 72 hours of life, when the infant is transitioning from fetal circulation to extrauterine circulation, but the first 10 days of life is a time of heightened risk for this type of injury.
Premature infants are also at extremely high risk for leukomalacia or white matter injury and the majority of infants born before 30-weeks of gestation demonstrate some white matter injury. There are several important factors that lead to this heightened risk. The arterial vascular supply to the brain is immature in premature infants, leading to a vulnerability to white matter injury in the end zones of the deep penetrating arteries and in the watershed areas between the deep and short penetrating arterial end zones. In addition, the estimated cerebral blood flow through the white matter is about one-tenth of that seen in that of the overall adult brain, suggesting that the premature brain has a narrower safety margin in terms of adequate cerebral blood flow compared with the mature adult brain. , The precursor oligodendrocytes or glial cells are at risk for both necrotic and apoptotic cell death in premature infants. These cells are exquisitely sensitive to the effects of oxygen free radicals, which are produced during conditions of ischemia, hypoxia, or hyperoxia.
Former premature infants or infants that are greater than 37 weeks postmenstrual age but who were born prematurely may also be a high-risk group because of ongoing and recent brain injury and neuroinflammation. Diseases of prematurity such as necrotizing enterocolitis or sepsis can lead to an inflammatory state. This state puts the infants at risk for perioperative injury in at least 2 ways. The ongoing inflammatory state can lead to baseline hypertension, which may or may not be recognized preoperatively, as well as lead to neuroinflammation.
Another group of susceptible neonates include those born with complex congenital cardiac disease. The fetal cerebral circulation of these infants is abnormal, which leads to a delay in maturation of the brain. It is estimated that the delay in maturation for an infant born at term is about 4 weeks putting these infants at high risk for brain injury, especially white matter injury.
There are also many neurologic conditions that put the developing brain at risk during general anesthesia, including poorly controlled seizure disorders, traumatic brain injury, anatomic brain abnormalities, and tumors. Careful collaboration with the patient’s care team including neurologists, neonatologists, and pediatric and intensive care specialists is needed to craft an anesthetic plan to optimize the brain health of these pediatric patients.
Modifiable Factors
Blood pressure
Blood pressure goals during anesthesia should be defined before the initiation of general anesthesia, especially in vulnerable patient populations. There is no generally accepted definition of hypotension in pediatric practice. Many pediatricians consider a blood pressure below the 10th percentile for age in a child that is clinically ill or distressed as hypotension. In addition, there is no absolute agreement on the best method to improve blood pressure in pediatric patients who are hypotensive. Many algorithms include fluid boluses of 10 to 20 mL/kg followed by the initiation of vasoactive agents. The maximal allowable decrease in blood pressure during anesthesia to maintain cerebral perfusion in pediatrics is even less well-defined, but there are a few studies that can guide practice. Many practitioners strive to keep the blood pressure above the lower limit of cerebral autoregulation, because there is less risk of inadvertent ischemia. Although there are some research methods that can approximate the lower limit of autoregulation in infancy and childhood using Doppler technology and near infrared spectroscopy, for most pediatric anesthesia cases using these methods is not feasible. These methods have demonstrated that every patient has their own individual autoregulatory curve, which can change depending on their physiologic status. The lower limit of autoregulation in children from age 6 months to 14 years was found to be approximately 60 mm Hg and did not vary with age using Doppler technology. The baseline mean arterial pressure (MAP) for infants is lower than that for older children, indicating that infants have less autoregulatory reserve. The maximal allowable decrease in MAP in infants less than 6 months of age compared with older children undergoing sevoflurane anesthesia with end-tidal CO 2 levels controlled was found to be 20% in the young infants versus 40% in the older infants and children. Another study found that, in infants less than 6 months of age undergoing similar anesthesia, there was adequate cerebral blood flow and oxygenation at a MAP of 45 mm Hg but when the MAPs were less than 35 mm Hg, there was both decreased cerebral blood flow and decreased cerebral oxygenation. A root cause analysis of a very small series of infants less than 4 kg in weight undergoing 2 to 3 hours of general anesthesia who developed postoperative seizures and evidence of cerebral ischemic hypoperfusion injury in MRI and computed tomography imaging determined that mean MAPs between 32 and 38 mm Hg were a contributing cause.
Hypocapnia or hypercapnia
There is a linear relationship between the tissue partial pressure of arterial carbon dioxide and cerebral blood flow as estimated by the tissue oxygen index using near infrared spectroscopy values over a large range in neonates. In children, the cerebral blood flow decreases by up to 3% for every 1 mm Hg decrease in the tissue partial pressure of arterial carbon dioxide. Mechanical ventilation to hypocapnic levels can independently decrease cerebral perfusion to ischemic ranges in critically ill children with preserved autoregulation. Many of the studies examining the effects of hypocapnia have been done in neonates who have sustained a hypoxic ischemic birth injury. A large multicenter trial examining hypothermia as a treatment for neonatal encephalopathy found that both the minimum Pa co 2 value as well as the cumulative exposure to Pa co 2 were risk factors for death and disability. The incidence of PVL is also increased in preterm infants who have had known hypocapnia of less than 35 mm Hg.
Mild hypercapnia may be beneficial to some neonates requiring ventilatory support. In a small randomized trial of extreme premature infants, it was found that mild hypocapnia (Pa co 2 of 45–55 mm Hg) led to fewer days requiring respiratory support. Patients with congenital diaphragmatic hernia also have been shown to benefit from mild hypercapnia versus normcapnia with decreased mortality rates. However, moderate hypercapnia (Pa co 2 of >55 mm Hg) has not been beneficial over normocapnia or mild hypercapnia in very premature infants. ,
Hypoxia or hyperoxia
Hypoxic ischemic injury is the most common cause of death and disability in the neonate. The disabilities incurred include persistent motor, sensory, and cognitive impairment. Every effort should be made to ensure that there is adequate oxygen delivery to the brain to meet metabolic demand.
Hyperoxia can increase the risk of retinopathy of prematurity, chronic lung disease, and brain injury in premature infants. The generally accepted target goal for blood oxygen saturation levels in premature infants to minimize these risks is between 90% and 94%. Term and older infants are at less risk for retinopathy of prematurity, so many practitioners aim for an oxygen saturation of greater than 95%. Regional anesthesia in young infants carries less of a risk of either hyperoxia or hypoxia compared with general anesthesia. Oxygen free radicals generated from hyperoxia can lead to both necrotic and apoptotic neuronal cell death. Oxidative stress is particularly harmful to neonates because the scavenging systems that detoxify reactive oxygen species are poorly developed in neonates.
Glucose homeostasis
Metabolic requirements decrease during general anesthesia for infants and older children but because the energy requirements of neonates is up to 6 times greater per body weight than adults, they may be at risk for intraoperative hypoglycemia. The risk of hypoglycemia is greatest during the first hours of life before the first feeding with almost 5% of normal neonates having a glucose concentration of 28 mg/dL or less. Then, after the first feeding, the glucose typically increases to near adult levels in the range of 70 to 100 mg/dL.
A single low blood glucose measurement in the first few hours of life does not put the infant at risk for long-term neurologic deficits. However, prolonged mild (glucose levels between 47–70 mg/dL) to moderate (glucose levels between 35–47 mg/dL) hypoglycemia is associated with long-term neurocognitive consequences. Infants with measured glucose levels of less than 47 mg/dL on 5 or more days had 3.5 times the incidence of cerebral palsy. Studies looking at the effects of mild hypotension and mild hypoglycemia in nonhuman primates found that the risks of neurocognitive injury is similar to the effects of moderate hypotension alone.
Temperature
The risks of hypothermia in pediatric patients during surgery are significant. In premature infants, the risks are high because of their greater surface to body mass proportion, low subcutaneous fat and keratin content, immature thermoregulatory mechanisms, and limited glycogen and brown adipose tissue stores. Radiant, conductive, and evaporative losses are most predominant in infants in the operating room. The risks to infants who are hypothermic on admission to the neonatal intensive care unit are great, with a multicenter prospective cohort study of almost 2000 premature infants finding that more than 50% of them were hypothermic on admission; hypothermia in this population was associated with an increased risk of death.
Keeping infants euthermic in the operating room is difficult and usually requires active warming measures. A quality improvement project found that raising the room temperature to 29.4°C when the infants were undraped, using warmed fluids, warming blankets and/or thermal mattresses, and keeping the transport incubator warm decreased the risk of hypothermia upon arrival to the neonatal intensive care unit by almost 4-fold.
Hyperthermia increases the patient’s metabolic rate and increases the risk of ischemic injury. Term neonates with hypoxic ischemic encephalopathy who were hyperthermic had worse neurocognitive outcomes compared with normothermic infants in the National Institute of Child Health and Human Development Neonatal Research Network randomized trial.
Positioning
The importance of intraoperative positioning is unclear. Often, because the endotracheal tube of a small neonate is prone to instability, pediatric anesthesiologists prefer to position the infant’s head in a lateral position. However, this position may not be optimal for brain health because this position can lead to obstruction to venous drainage. There is a correlation between lateral head rotation and decreased cerebral oxygen index as measured by near infrared spectroscopy and an increase in cerebral blood volume in very low birth weight infants. , Intracranial pressures are lowest in infants positioned with their heads midline with the head of the bed raised 30°. However, a Cochrane meta-analysis studying the effects of neutral head position and head elevation did not find a decreased incidence of intravascular hemorrhage or an improvement in cerebral oxygenation. Further research on this topic is needed. However, if feasible, maintaining the patient’s head in a neutral position may decrease the chance of venous outflow obstruction.
Parental counseling
The best approach to parents and pediatric care providers who are concerned about anesthetic-induced neurotoxicity is to be cognizant and conversant about the available pediatric anesthesia literature on this topic. There is little to no evidence for anesthetic neurotoxicity in prospective human studies in children exposed to short to moderate length general anesthetics. The median length of general anesthetics in children less than 18 years of age is less than 1 hour. It is also important to reassure parents that every effort will be made to ensure that their child will be stable throughout the general anesthetic and the surgeon and anesthesiologist will strive to minimize the length of exposure. Many of the procedures that are done in young children are also very important for their neurodevelopment, such as strabismus surgery to ensure binocular vison and ear surgery to ensure adequate hearing. Parents and providers should be informed that there is ongoing research examining the effects of anesthetic exposures lasting longer than 2 hours.
Summary
Several factors have an impact on neurodevelopment in pediatric surgical patients. These factors include the genetics and environment in addition to the surgical lesion and the perioperative management of the patient ( Fig. 1 ). Pediatric anesthesiologists and intensivists are tasked to minimize the risk of brain injury throughout the perioperative period. This effort requires close attention and understanding of the patient’s underlying medical condition and working closely with their entire medical team to optimize care. Unnecessary procedures and imaging studies should be avoided to limit exposure to potentially toxic general anesthetics. Baseline measures of blood pressure are important in determining perioperative blood pressure goals. Inadvertent hypocapnia or moderate hypercapnia as well as hyperoxia or hypoxia should be avoided in all patients. Pediatric patients should be maintained in a normothermic, euglycemic state with neutral positioning unless there are good reasons to deviate from these goals. Most of all, the task of improving the developmental outcomes of infants and children requires collaboration between anesthesiologists, surgeons, pediatricians, and neonatologists.
