Jonathan Bennett, MD, is our new SPS liaison to SmartTots, a public-private partnership organization launched to facilitate and support studies of existing anesthetic drugs and their effects on childhood development. He reviews two original research articles “hot off the presses” (is that a term anyone uses anymore in this electronic age?) addressing the topic of anesthetics and their impact on neurodevelopment in children.
Neurodevelopmental outcome at five years of age after general anesthesia or awake-regional anesthesia in infancy (GAS): an international, multicentre, randomized, controlled equivalence trial, was published in The Lancet in February (Lancet 2019;393;664-77). This article publicizes the primary outcome of the GAS trial, which randomized infants less than 60 weeks postmenstrual age, undergoing inguinal herniorraphy, to either sevoflurane general anesthetic or awake-regional anesthetic.
Previously published data from the GAS consortium (Lancet 2016; 387:239-50) for subjects at two years of age had shown equivalence in the composite cognitive scores of the Bayley Scales of Infant and Toddler Development between those that underwent general vs. awake-regional anesthesia. The primary outcome of the study, full scale intelligence quotient (FSIQ) on the Wechsler Preschool and Primary Scale of Intelligence, was measured at five years of age. Similar to the Bayley results, this data showed equivalence for FSIQ for patients exposed to general anesthesia (FSIQ 98.97) and awake-regional anesthesia (FSIQ 99.08). On a range of other neuropsychological tests (measures of social perception, academic (spelling, reading, number composition), executive function, and behavior) and non-psychometric outcome data (developmental issues, behavior disorders) there seemed to be equivalency as well between the two groups.
This study is significant as it represents the only randomized trial data that addresses the impact of anesthetics on neurodevelopment in children. Results of other studies that use various methods of cohort design, or case-control design, are difficult to interpret due to the inherent problem of confounders that may influence the primary outcome. These unmeasured confounders certainly contribute to what the authors point out as “mixed and conflicting evidence for associations between exposure to anaesthesia in early childhood and neurodevelopmental outcome.” The obvious limitations of this study are that it studied only a single anesthetic agent, sevoflurane, for a relatively short duration (median 54 min, IQR 40-70), on predominantly males, and had about 25% loss to follow up for the age five testing. The testing at five years of age might also be considered by some to be a limitation. The authors discuss that despite these limitations the GAS trial demonstrates strong evidence that approximately one hour of general anesthesia in infancy does not cause significant neurocognitive or behavioral deficits.
The second article, Influence of surgical procedures and general anesthesia on child development before primary school entry among matched sibling Pairs, JAMA Pediatrics 2019:173(1):29-35, follows the more standard approach to research in this field. This retrospective cohort study looked at children in Canada exposed (or not) to surgical procedures requiring general anesthesia from birth until they completed the Early Development Instrument. This instrument is administered to all children entering public or Catholic schools in Ontario, Canada at ages five to six.
What makes this study particularly interesting is that the authors used siblings for the control (unexposed) group, with the goal of “mitigating unmeasured confounding from biologic vulnerability and environmental factors.” Basically the researchers are trying to make the exposed and unexposed groups as similar as possible except for the exposure. A similar approach was used by the PANDA study where they identified sibling pairs with history of anesthesia exposure in the past and then approached them for neurocognitive testing at ages 8-15. In that study, however, there were only 105 sibling pairs. As the authors note, “the magnitude of the risk associated with neurotoxic effects of pediatric anesthesia represents only a small fraction of the variability seen and much less than other perinatal, home environmental, and social covariates”.
As a result, larger sample populations may be required to detect potentially small differences in child development outcomes.” To that end they collected linked data form the Early Development Instrument and from population-based health and demographic databases housed at the Institute for Clinical Evaluation Sciences, Toronto. By using these population databases the authors were able to identify 10,897 sibling pairs to analyze the effect of surgery requiring general anesthesia on EDI outcomes (367 pairs where both siblings had surgery, 2,346 where one sibling had surgery, and 8184 where neither underwent surgery). After adjusting for potential confounding the authors found no significant differences in vulnerability (<10th percentile) in the major domains of the EDI or scores in major domains of EDI after exposure to surgery. In analysis of only the 2346 sibling pairs where one sibling underwent surgery, the authors also found no differences in vulnerability (<10th percentile) or estimates for any major EDI domain.
This is study is limited by its retrospective nature which, despite the sibling matching, might have allowed bias or unmeasured confounder to affect the association they were trying to measure. Children who already had EDI-recorded behavioral, learning, or developmental diagnosis were excluded from the analysis, as the timing of the diagnosis could not be determined in relation to the exposure, which could have biased the results towards the null. However a sensitivity analysis that included these children did not alter results of the primary analysis. Finally, the EDI is not a neurocognitive test and was not designed to identify specific neurodevelopmental deficits in individual children. The authors point out that it does “demonstrate moderate concurrent validity with direct measures of neurodevelopment.” The authors conclude that the study findings support that anesthesia exposure is not associated with adverse child development.
These two recently published studies provide some good news for those of us who administer potentially neurotoxic medications to children. However, work is continuing to be done on the subject, and more information is needed, particularly on potential effects at later ages.