Contemplating the Future of Pediatric Anesthetic Neurotoxicity Research

Cameron Bosinski, MD, MS

Whether anesthesia has any negative effect on the developing brain has been a longstanding question in the field of anesthesiology. Anesthetic neurotoxicity investigators Vesna Jevtovic-Todorovic, MD, PhD, Ashok Panigrahy, MD, Viola Neudecker, MD, and Caleb Ing, MD, MS, illustrated new directions in pediatric anesthetic neurotoxicity research from varying perspectives and research focuses during the SmartTots session,  “New Directions in Pediatric Anesthetic Neurotoxicity Research,” held Sunday, March 20 at the IARS 2022 Annual Meeting. The speakers contemplated the future of this important research and how to eventually answer this quandary.

This fascinating session on current outcomes and new directions for pediatric anesthetic neurotoxicity was facilitated by Dean B. Andropoulos, MD, MHCM. Dr. Andropoulos serves as the SmartTots Medical Officer and is Anesthesiologist-in-Chief and Chair of Texas Children’s Department of Anesthesiology and Vice Chair, Baylor College of Medicine Department of Anesthesiology and Professor of Anesthesiology and Pediatrics, Baylor College of Medicine.

Drs. Neudecker and Ing presented data from animal and human studies, which indicated that even a single exposure to anesthesia early in life may have a negative effect on the social development of children. Dr. Panigrahy presented some preliminary data from preclinical studies on volatile anesthetic agent exposure in infants and neonatals while Dr. Jevtovic-Todorovic envisioned some possible solutions to this difficult dilemma, examining the potential for neurosteroids to induce anesthesia without the associated neurotoxicity.

Viola Neudecker, MD, Associate Research Scientist at Columbia University Medical Center, explored non-human animal study outcomes with a talk entitled, “Behavioral Phenotype After Anesthetic Exposure in Non-Human Primates.” There is controversy as to whether anesthetics affect the neurobehavioral development of children, she explained. While some studies have found an increased risk of developing learning disabilities, lower intelligence and an increased risk of attention deficit hyperactivity disorder (ADHD) after exposure to anesthesia, other studies have not found this relationship. A recent meta-analysis by Dr. Caleb Ing did not find any reductions in IQ after a single exposure to anesthesia, but exposure did increase parental reports of behavioral problems and executive function impairment. Studies in nonhuman primates (NHP) have found deficits in learning, memory and behavioral regulation after exposure to anesthesia early in life. These studies also found apoptosis after exposure to anesthesia, but it is unknown whether these structural changes in the brain are persistent.

In her study, Dr. Neudecker exposed postnatal day 6 NHP to 5 hours of 1.8% isoflurane. A subset of these animals had repeat exposure to isoflurane on postnatal days 9 and 12. Control animals were exposed to 30% oxygen over these 3 days. At one year of age, control and experimental animals were transferred to new social groups and were followed for another year, which corresponds to age 6 in humans. Spontaneous behavior was assessed and classified into 4 categories: close social behavior, aggressive behavior, anxiety behavior, and appeasement behavior. Animals exposed to isoflurane spent less time in close social contact compared to controls. NHP were assessed at 2 years of age for chronic astrocyte activation identified by increased GFAP expression. Animals exposed to isoflurane were found to have increased astrogliosis in the primary visual cortex and the amygdala. These studies indicate that even a single exposure to anesthesia in early life may impair the development of social behaviors.

Continuing the discussion on behavioral phenotypes, Caleb Ing, MD, MS, Assistant Professor of Anesthesiology at Columbia University Medical Center, focused on clinical studies. His presentation was entitled, “Behavioral Phenotype in Clinical Studies of Anesthetic Neurotoxicity.” Many studies have been performed to evaluate the long-term effects of anesthesia on the development of children. Bartels 2009 studied monozygotic twins in which one was exposed to anesthesia younger than 3 years. No difference was found in the Conner’s Teacher Rating Scale (which evaluates for externalization of behavior problems) at 12 years old. Stratmann 2014 studied children who were exposed to anesthesia at less than 1 year of age and did not have any differences when evaluated by the Child Behavior Checklist (CBCL), which asks parents to evaluate the behavioral problems of children. Bakri 2015 found that children with multiple exposures to anesthesia were 11 times more likely to develop anxiety or depression and 8 times more likely to have attention problems. Khochfe 2019 found that compared to children with regional anesthesia, children with general anesthesia were 7 times more likely to develop behavioral problems.

Dr. Ing and colleagues (Ing 2021) performed a systematic review and meta-analysis of three prospective studies of children after a single anesthetic exposure. Dr. Ing referenced the PANDA, MASK and GAS studies for correlating evidence in this area. The PANDA (Sun 2016) and MASK (Warner 2018) were both observational studies. The GAS study (McCann 2019) was performed as a randomized control trial (RCT) of general versus regional anesthesia for hernia repair. Dr. Ing found that after a single exposure to anesthesia there were no significant differences in IQ. Although he did find more externalizing and internalizing behavioral scores on the CBCL and worse executive function, which trended towards significance. Dr. Walkden and colleagues (Walkden 2020) performed an observational study and found that exposure of children younger than 4 years to anesthesia was associated with worse motor, sociobehavioral and linguistic function, but not with educational or cognitive function. In Dr. Ing’s 2021 study, he examined prenatal exposure and found that exposed children had more externalizing behavior problems. Multiple studies have found a 30-40% increased risk of ADHD after a single exposure to anesthesia (Sprung 2012, Hu 2017, Tsai 2018, Ing 2017, Ing 2019, Shi 2021). This risk increased with multiple exposures to anesthesia. More randomized controlled trials are needed to study this problem and assess the particular deficits that may occur after exposure to anesthesia in childhood.

Transitioning into discussions of preliminary preclinical studies in neonates and infants, Ashok Panigrahy, MD, Professor of Radiology at University of Pittsburgh and Radiologist-in-Chief and Vice Chair Clinical and Translational Imaging Research at UPMC Children’s Hospital of Pittsburgh, expanded on new directions for the research area, focusing on “Imaging Biomarkers of Anesthetic Exposure to Neonates and Infants.” During his presentation, he covered known outcomes related to imaging biomarkers of volatile anesthetic agent (VAA) exposure in neonates and infants and outlined safe infant neuroimaging approaches that allow researchers to conduct studies in this area.

Dr. Panigrahy reviewed the findings in several recent studies focused on anesthetic exposure in neonates and infants. Like Dr. Ing, Dr. Panigrahy also referred to the GAS, PANDA and MASK studies. The GAS (McCann et al 2019) and PANDA (Sun et al 2016) studies showed that exposure of short duration does not adversely affect neurocognitive outcomes in children. However, the MASK study (Warner et al 2018) showed an association of multiple anesthetic exposures and decreases in cognitive ability. According to Dr. Panigrahy, most neuroimaging correlates volatile anesthesia agent (VAA) neurotoxicity research has consisted of neuroimaging multimodal studies that have been performed on older children and adolescents.

However, some studies have related sedatives to neonatal neuroimaging outcomes, particularly the relationship between midazolam and hippocampus abnormality in structures and similarly a relationship between opioids exposures and cerebellar abnormalities have been shown. Dr. Panigrahy pointed out a major gap in knowledge in this area of research – very few, if any, studies show correlations between VAA exposure and neonatal and infant human neuroimaging biomarkers, particularly with magnetic resonance imaging (MRI).

Dr. Panigrahy posed the question, “Can we perform safe neonatal and infant neuroimaging for research?” Encouragingly, he answered in the affirmative. He provided an example on one of his presentation slides of some of the neuroimaging tools investigators can use to take an infant and scan them during natural sleep, including using ear protection and different feed and swaddling techniques. One image showed how researchers are able to map brain development over the first two years of life.

Next, he highlighted the outcomes of a recent preclinical study (BJA, 2021) that showed the cumulative effect of repeated exposure to ketamine, tiletamine and zolazepam, and isoflurane on early brain development in the developing rhesus monkey. In this study, performed by the Styner Research Group, the investigators took data that was collected as part of imaging studies performed on these developmental non-human primates and showed that increased exposure to these volatile anesthetic agents is correlated with abnormal microstructure as measured with diffusion tensor imaging (DTI). He displayed color maps that showed dose-dependent effects of these anesthetic agents on the microstructure in different areas of the brain in a regional manner.

To conclude his presentation, Dr. Panigrahy reviewed an ongoing human congenital heart disease (CHD) study (Panigrahy et al 2015) that his research group is performing, and shared some preliminary data that demonstrated a similar approach to the Styner Research Group, only focusing on a preclinical model using human imaging in infants, specifically in infants with congenital heart disease. Infants with congenital heart disease often require numerous, lengthy anesthetics during infancy during surgery with multiple repetitive procedures, presenting as an ideal group to study to help answer this research question. He shared examples of a pilot study and multicenter study that recapitulates the impact of VAA on white matter microstructure, using diffuser tensor imaging. He noted that human studies are confounded by many aspects of heterogeneity related to technique but also with other clinical covariants. Dr. Panigrahy then reviewed the harmonization tools his group is developing to address this concern. Using these techniques, he noted, investigators may control for clinical heterogeneity in particular social demographics, genetics and other factors that could be related to tying a neuroimaging biomarker like DTI to neurodevelopmental outcome in the setting of prolonged VAA exposure.

To reproduce the rigor and results of his group’s pilot study, they teamed with four children’s hospitals (Texas Children’s, Pittsburgh Children’s, Los Angeles Children’s and Children’s Hospital of Philadelphia) to pool similarly prospectively acquired research in neuroimaging DTI studies across 445 infants with complex and congenital heart disease. They also collected anesthetic recorded and correlated different exposure perimeters with DTI postoperative studies. Their research showed strong correlations between their primary exposure variables, the duration of general anesthesia exposure, and the VAA MAC hours on microstructure using DTI, particularly in cortical association tracks. Due to a concern about variables between site MRI scanners, in parallel, Dr. Panigrahy’s group has been working on developing a harmonization tool, using machine learning, called Empirical Bayes or COMBAT. This type of machine learning is used in adult multicenter studies to reduce scanner variance retrospectively. They applied this to their cohort and were able to show they can reduce scanner variance when using this machine learning technique. This multicenter group is continuing their work with this cohort. Next they will be looking at the effect of VAA exposure to DTI and neurodevelopmental outcomes, controlling for scanner heterogeneity and other clinical risk factors.

Vesna Jevtovic-Todorovic, MD, PhD, Professor and Chair of Anesthesiology at the University of Colorado School of Medicine, analyzed some possible new directions for pediatric anesthetic neurotoxicity research with a presentation on “Novel Neurosteroids.” Neuroactive steroids are a novel class of anesthetics that may have several favorable properties, especially for pediatric populations. Dr. Jevtovic-Todorovic compared the efficacy and side effects of three neuroactive steroids (3β-OH, alphaxalone and CDNC24) to propofol.

Neuroactive steroids are similar in structure to progesterone and can potentiate GABA receptors or block voltage-gated calcium channels. In order to investigate the therapeutic utility of these compounds, Dr. Jevtovic-Todorovic administered three neuroactive steroids (3β-OH, alphaxalone and CDNC24) to juvenile (7 day old) rats. 3β-OH was found to have an ED50 of 4 mg/kg and a therapeutic index of 20. Alphaxalone had an ED50 of 1.58 mg/kg and a therapeutic index of 31. CDNC24 had an ED50 of 0.68 mg/kg and a therapeutic index of >88. Even at the limit of solubility, CDNC24 did not have any increase in mortality. For reference, the ED50 of propofol is 2.36 mg/kg and has a therapeutic index of 23.

In order to test the ability to induce neurotoxicity, rats were exposed to ketamine, propofol and each of the three neurosteroids. After exposure, rat brains were stained for markers of apoptosis. Ketamine and propofol were both able to induce apoptosis in the rat brain. However, none of the three neurosteroids were found to induce developmental neuroapoptosis in rats. Although no direct evidence of neurotoxicity was found, rats were further tested on measures of memory and cognition. Although deficits in the radial arm maze were seen after exposure to ketamine, deficits were not seen after the administration of 3β-OH.

This spurred Dr. Jevtovic-Todorovic’s interest in determining the mechanism by which neurosteroids can induce anesthesia. Low-voltage-gated (T-type) calcium channels are known to play an important role in controlling neuronal excitability. Inhibition of T-type calcium channels leads to analgesia in animal models. T-type calcium channels are also important for the regulation of sleep and wakefulness. Patch clamping of T-type calcium channels in the ventral basal thalamus and subiculum demonstrated that application of 3β-OH resulted in sustained closure of these channels. Potentiation of GABA_A receptors has previously been proposed as a major contributor to neurodevelopmental anesthetic toxicity. Dr. Jevtovic-Todorovic was able to show that 3β-OH, alphaxalone and CDNC24 were all able to decrease spontaneous release of GABA. In addition, alphaxalone and CDNC24 induced postsynaptic potentiation of GABA_A receptors. Further development of neurosteroids that can inhibit presynaptic and postsynaptic GABA activity may prevent developmental neurotoxicity and cognitive dysfunction in children.

Moderated by Dr. Andropoulos, the session concluded with a robust discussion with the speakers on recent findings and future directions for this important research area.