“…[T]here may be risk associated with exposure to general anesthesia…[but] surgery without adequate anesthesia and analgesia is simply not an option.”
TWENTY-FIVE years ago, surgery was commonly performed on awake or paralyzed infants and children without the benefit of analgesics, sedatives, or anesthetics. This practice was justified by a fear that these drugs would increase the risk of adverse surgical outcomes and the belief that infants did not experience pain.1 Subsequent clinical studies have demonstrated that inadequate sedation, anesthesia, and analgesia may actually increase the risk of adverse postoperative outcomes in infants and children and have led to a paradigm shift such that the use of general anesthetics and analgesics is now routine in the humane practice of pediatric anesthesiology.1–3 This does not mean that doing so is without risk or controversy. Seven years after this change, questions began to arise from the laboratory of John Olney, M.D., who was well known for his work on fetal alcohol syndrome, suggesting that exposure to anesthetics and sedatives with mechanisms of action similar to those of alcohol may result in adverse outcomes in neonatal animals.4–6 These findings have been replicated in multiple laboratories, using a number of animal species, and demonstrate that prolonged or repeated administration of anesthetics and sedatives to prenatal and neonatal animals is associated with abnormal brain development and an increased risk of cognitive and behavioral abnormalities in later life.4 In this issue of Anesthesiology, Dr. Jevtovic-Todorovic, a prominent investigator in the field, reviews the basic science literature7 that served to suggest to the U.S. Food and Drug Administration (FDA; Silver Spring, Maryland) a need for a warning on the use sedatives and anesthetics during pregnancy and early childhood.8
Dr. Jevtovic-Todorovic’s review of the topic is a significant contribution to the literature as it comprehensively describes the work performed in rodents and subhuman primates, demonstrating that treatment with anesthetics and sedatives in utero and during early postnatal development is associated with enhanced apoptosis (neurons and glia) and alterations in neurogenesis, gene expression, cytokine expression, synaptic transmission, and lipid metabolism that are, in turn, associated with cognitive and behavioral abnormalities in later life.7 Perhaps the most interesting of these findings is that they appear to be dependent on developmental stage and duration of treatment. Administration of a single anesthetic to rodents or subhuman primates for long durations (6 h or more) or multiple shorter treatments during critical periods of development are crucial for the development of these changes. Studies in rodents that have looked at shorter exposures to general anesthetics during these same critical developmental periods have found acute but not long-lasting changes in synapse density or function, neurogenesis, and cell survival and differentiation, and no differences in cognitive performance in adulthood.9 Similarly, while a recent primate study demonstrated enhanced apoptosis after a 3-h treatment with isoflurane, another noted no long-term cognitive consequences after a single 5-h treatment with isoflurane.10,11
Caution is warranted when extrapolating data derived from rodents or subhuman primates to humans as the developmental stages vary in both timing and length such that anesthetic exposure for 3, 6, or 24 h in a rodent or subhuman primate during the same critical developmental period may represent a relatively longer ratio of treatment time to critical developmental period time and may result in different outcomes. In addition, most surgical procedures in children are less than 6 h, whereas the majority of the preclinical studies were performed using longer durations of treatment and multiple exposures on consecutive days. Indeed, the issue of whether administration of general anesthetics to human fetuses, neonates, infants, and young children is safe is controversial in part because there was no clinical suspicion of neurocognitive deficits associated with general anesthetics until preclinical studies suggested that there may be a problem. To date there are insufficient human data to conclude that there are adverse effects of anesthesia in the developing human brain, as suggested by the reviews of clinical data by Drs. Davidson and Sun in this month’s issue of Anesthesiology,12 and by others.13–16
There are also issues around the design of the preclinical studies that make it difficult to translate them to the clinical realm primarily because the majority of the studies have been designed to demonstrate toxicity as evidenced by the utilization of multiple drugs, high dosages, and prolonged treatment periods. Baxter and Alvarado13 eloquently describe this point: “Because of the time and expense involved in a prospective study with nonhuman primates, investigators have based their anesthesia protocols on durations of exposure that are known to result in increased neuro- and glioapoptosis, a candidate mechanism for the long-term effects of pediatric anesthesia on behavior. Because these durations tend to be longer than those commonly encountered in the pediatric operating room, many of these studies face the criticism that the anesthesia exposures are not clinically relevant. Single anesthesia exposures in infants that have been investigated recently are of short duration relative to those in the studies to date with monkeys.” However, their conclusions put the prolonged exposure studies into perspective and point a way forward: “Although this is a limitation of the extant preclinical studies, these exposure protocols establish boundary conditions upon which future work can build, to carry out more finely grained analyses of duration and frequency of anesthesia exposure.” They are not the only authors to have made suggestions that changes in the design of the preclinical studies need to be considered. Disma et al.14 have suggested the need for well-controlled preclinical studies investigating issues such as multiple exposures, multiple types of anesthetics, durations of exposure, critical stages of development, and the use of appropriate physiologic monitoring and behavioral and cognitive testing. However, given the prospective clinical studies published to date, the results of properly controlled preclinical studies investigating whether toxicity occurs under clinically relevant conditions will likely be negative and highly criticized for failing to demonstrate the toxicity associated with prolonged or multiple exposures. In the words of Paracelsus, “What is there that is not poison? All things are poison and nothing is without poison. Solely the dose determines that a thing is not a poison.”17
In addition to the questions about the appropriateness of the animal model, there are a number of variables in humans that do not occur in the preclinical laboratory that may have an effect on patient outcomes. An interesting study involving monozygotic twins nicely addressed this issue. The authors noted that children exposed to general anesthesia were more likely to have lower cognitive performance and educational achievement scores compared to children who were not.18 However, the unexposed co-twins from discordant pairs did not differ from their exposed co-twins in educational achievement and cognitive performance, suggesting that early anesthesia may be a marker of some unknown vulnerability for long-term learning problems irrespective of their exposure to anesthesia.18 In addition, human children do not live in the same social environment, eat the same food, and have the same physical activity, and they often have parents with varying degrees of education and parental involvement in their care. Each of these variables has been associated with altered cognitive and behavioral performance in offspring.19–21 Even preclinical studies suggest that this may be true, at least in rodents, where environmental enrichment, by itself, prevented cognitive decline after treatment with a general anesthetic.22 In essence, human children are not research animals, and it is currently not possible to discern whether lower cognitive or social performance after surgery and anesthesia, if they exist, are the result of the anesthetic, the surgery, the patient’s underlying condition, or some critical combination of the three. Even the FDA warning states that “it is unclear whether any negative effects seen in children’s learning or behavior were due to the drugs or to other factors, such as the underlying medical condition that led to the need for the surgery or procedure.”8
This is not the first instance in which preclinical data related to anesthetic and sedative administration have been difficult to translate into the clinical arena. For example, anesthetics and sedatives were thought to be neuroprotective in the setting of central nervous system injury for decades, as multiple preclinical models clearly demonstrated improvements in central nervous system outcomes after a variety of insults. However, when taken to clinical trials, none have been shown to improve outcomes.23 Part of this may be due to a failure on our part as scientists to develop and adhere to strict preclinical study guidelines that are relevant to the clinical scenario we are trying to replicate, causing some to question whether we should consider a requirement to register preclinical studies before they are to be performed in a manner consistent with what is required of clinical studies.14,23 Unfortunately, even this type of scrutiny may not be sufficient as it is difficult to remove potential species effects and the host of events that affect real-world patient outcomes.
The preclinical data that have led to the FDA’s warning on the use of sedatives and anesthetics during pregnancy and early life consistently identify central nervous system abnormalities and cognitive and behavioral deficits in later life, although translation of this information to the clinical realm has not yet occurred. This has led some to opine that the FDA’s warning was unexpected.24 In particular, there is no clinical evidence that exposure to general anesthetics is associated with adverse outcomes, and there are concerns that children may not receive medically appropriate interventions based solely on evidence from preclinical studies published before any hint of a clinical suspicion that they cause harm.24 Nonetheless, it is important that anesthesiologists understand the potential risks and known benefits of any anesthetic or procedure being performed on young children and ably participate in shared decision making with the parents of such children requiring treatments necessitating the use of general anesthetics. While there may be risk associated with exposure to general anesthesia, not having surgery or a procedure that requires general anesthesia may be associated with its own risks, and surgery without adequate anesthesia and analgesia is simply not an option.2,3
Dr. Culley is Executive Editor of Anesthesiology (Schaumburg, Illinois), Director of the American Board of Anesthesiology (Raleigh, North Carolina), Chair of the American Board of Medical Specialties Committee on Continuous Certification (Chicago, Illinois), an Accreditation Council for Graduate Medical Education – Residency Review Committee member (Chicago, Illinois), and an American Society of Anesthesiologists committee member (Schaumburg, Illinois), and receives grant funding from Anesthesia Patient Safety Foundation (Schaumburg, Illinois), Controlled Risk Insurance Company (Boston, Massachusetts), and National Institute of Aging (Bethesda, Maryland). Dr. Avram is Assistant Editor-in-Chief of Anesthesiology and an American Society of Anesthesiologists committee member, and receives grant funding from National Institute of Child Health and Human Development (Bethesda, Maryland).