FDA reviewed many published research studies, including both nonclinical and clinical data.
Nonclinical Studies
We reviewed published nonclinical studies showing that early exposure to anesthetic drugs can produce adverse neurotoxic effects in different species, including simple nematodes, rats, and nonhuman primates.1-20 The studies demonstrate that the use of anesthetic drugs during the period of rapid brain growth or synaptogenesis results in widespread neuronal and oligodendrocyte cell loss in the developing brain, and alterations in synaptic morphology and neurogenesis. Based on comparisons across these nonhuman species, the window of vulnerability to these changes is believed to correlate with human exposures in the third trimester of pregnancy through the first year of life, but may extend to approximately 3 years.
In pregnant primates, late in gestation during peak fetal brain development, exposure to 24 hours of ketamine in the third trimester or exposure to 5 hours of isoflurane or 5 hours of propofol increased neuronal cell loss in the fetus. In neonatal primates, exposure to 3 hours of ketamine that produced a light surgical plane of anesthesia did not increase neuronal cell loss; however, neuronal cell loss was seen with treatment regimens of 24 hours of ketamine, or 5 hours or more of isoflurane plus nitrous oxide or propofol. Data in isoflurane-treated rodents and in ketamine-treated primates suggest the neuronal and oligodendrocyte cell losses are associated with prolonged cognitive deficits in learning and memory. The clinical significance of these nonclinical (animal) findings is not known.
Clinical Studies
We also reviewed the epidemiologic literature investigating the association between childhood anesthesia exposure and adverse neurodevelopmental outcomes.21-41 The studies were published between 2009 and 2014. Some studies found no association between pediatric exposures and neurodevelopmental
outcomes,21-29 whereas others did.30-41 In particular, several of the studies have increased concerns that longer or repeated exposures may contribute to various cognitive and behavioral problems, including neurodevelopmental delay-related diagnoses, learning disabilities, and attention deficit hyperactivity disorder.32, 33, 37, 38, 41 However, it remains unclear whether these associations represent an effect of the anesthesia drugs as opposed to the surgery itself, or are the result of uncontrolled confounding related to the underlying condition or other factors. The observational studies had many limitations, including heterogeneous exposure and outcome definitions and measures, potential for selection and information biases, incomplete control for confounding, and insufficient power. Most studies included children exposed to anesthesia before age 2-3 years, but the studies varied widely in the age groups included. Some focused only on newborns and infants, while others included children with anesthesia exposures up to 12 years of age. Most of the studies were not able to determine the duration of anesthesia exposure.
More recently, interim findings from the General Anesthesia Compared to Spinal Anesthesia (GAS) trial and results from the Pediatric Anesthesia NeuroDevelopment Assessment (PANDA) Study have been
published.42. 43 The GAS trial is an ongoing international, multicenter, randomized controlled trial comparing neurocognitive outcomes following randomization to either awake-regional anesthesia or to sevoflurane-based general anesthesia in children younger than 60 weeks but born at more than 26 weeks gestation who required inguinal hernia repair. The primary outcome of the trial is Wechsler Preschool and Primary Scale of Intelligence Third Edition (WPPSI-III) Full Scale Intelligence Quotient (IQ) at age 5 years. The secondary outcome is the composite cognitive score of the Bayley Scales of Infant and Toddler Development III at age 2 years. The initial 2-year follow-up results were published in January 2016. Data were evaluated from 238 children treated with awake-regional anesthesia and 294 children administered general anesthesia. The median duration of sevoflurane in the general anesthesia cohort was 54 minutes. The authors reported no difference in Bayley III development scores between the two study arms. They suggested these data support the conclusion that sevoflurane anesthesia of less than 1 hour duration does not appear to increase the risk of adverse neurodevelopmental outcome at age 2 years compared to awake-regional anesthesia. The GAS study still needs to be completed to evaluate the primary WPPSI-III IQ outcome measure at age 5 years.42
The PANDA study is a sibling-matched observational cohort study that examined whether a single anesthesia exposure in healthy children younger than 3 years is associated with an increased risk of impaired global cognitive function (IQ) as the primary outcome, and abnormal domain-specific neurocognitive functions and behavior as secondary outcomes at ages 8 to 15 years. Exposed children (n=105) had a single episode of general anesthesia before 3 years for elective inguinal hernia surgery and were 36 weeks gestational age or older at birth. The mean duration of anesthesia in the exposed group was 84 minutes, with 17 children having exposures more than 2 hours. Ninety percent of the exposed children were boys. The unexposed cohort (n=105) were biologically related siblings closest in age (within 3 years) to the exposed child, also 36 weeks gestational age or older at birth but with no anesthesia exposure before 3 years. The study found that mean IQ scores were not significantly different between the exposed and unexposed siblings, with both groups scoring somewhat higher than average. There were no significant differences in mean scores on any of the secondary assessments, although exposed children were significantly more likely to have abnormal “internalizing” scores on the Child Behavior Checklist. This may have been a chance finding, as analyses did not adjust for multiple endpoints.43
The PANDA study addressed many of the limitations of the previous epidemiologic studies. However, as designed, the study was not able to evaluate effects of prolonged or repeated anesthesia exposures, or possible effects in more vulnerable subgroups (e.g., less healthy children). The study also may not have been sufficiently powered to evaluate meaningful differences in all secondary outcomes or for analyses examining age at exposure, duration of exposure, or gender differences.
The PANDA study, along with the preliminary GAS trial findings, provide some clinical evidence that a single, relatively brief early exposure to general anesthesia in generally healthy children is unlikely to cause clinically detectable deficits in global cognitive function or serious behavior disorders. These findings are consistent with animal data, which have not predicted an increased risk with anesthesia exposures less than 3 hours. However, the GAS trial has not yet been completed, and additional high-quality research is needed to investigate the effects of repeated and prolonged anesthesia exposures in children, including vulnerable populations. Research is also needed to explore possible subtle behavioral effects, vulnerable ages of exposure, potential gender differences, and potential variability among specific anesthetic drugs and protocols.