Children undergoing complex cardiac surgery are exposed to substantial cumulative doses of sedative medications and volatile anesthetics and are more frequently anesthetized with ketamine, compared with healthy children. This study hypothesized that greater exposure to sedation and anesthesia in this population is associated with lower neurodevelopmental scores at 18 months of age.
A secondary analysis was conducted of infants with congenital heart disease who participated in a prospective observational study of environmental exposures and neurodevelopmental outcomes to assess the impact of cumulative volatile anesthetic agents and sedative medications. Cumulative minimum alveolar concentration hours of exposure to volatile anesthetic agents and all operating room and intensive care unit exposures to sedative and anesthesia medications were collected before administration of Bayley Scales of Infant and Toddler Development, 3rd edition (Bayley III), at 18 months of age.
The study cohort included 41 (37%) single-ventricle and 69 (63%) two-ventricle patients. Exposures to volatile anesthetic agents, opioids, benzodiazepines, and dexmedetomidine were not associated with abnormal Bayley III scores. At 18-month follow-up, after adjusting for confounders, each mg/kg increase in ketamine exposure was associated with a 0.34 (95% CI, −0.64 to −0.05) point decrease in Bayley III motor scores (P = 0.024).
Total cumulative exposures to volatile anesthetic agents were not associated with neurodevelopmental impairment in infants with congenital heart disease undergoing various imaging studies and procedures, whereas higher ketamine doses were associated with poorer motor performance.
Preclinical data suggest that cumulative exposure to general anesthetics in the early postnatal period can induce lasting impairment of neurodevelopment
Children with congenital cardiac abnormalities undergo prolonged diagnostic and surgical procedures requiring multiple anesthetic exposures early in life
The question of whether greater exposure to general anesthesia in children with congenital heart disease is associated with worse neurodevelopmental outcomes is incompletely explored
In this secondary analysis of a prospective study evaluating neurodevelopmental outcomes of pediatric patients with congenital heart disease, no association was found between total cumulative exposures to volatile anesthetics and Bayley III scores assessed at 18 months of age, whereas higher ketamine doses were associated with poorer motor performance
The past two decades have seen intense discussions about the potential long-term deleterious effects of anesthetic and sedative exposures in young children on brain development.1,2 Dendritic and synaptic alterations, immature neuronal cell death and long-term impairment of memory, learning, and behavioral development have consistently been described in newborn animals, including nonhuman primates, for all clinically used γ-aminobutyric acid agonists and N-methyl-d-aspartate antagonists.3,4 However, clinical studies in human infants and young children have not consistently demonstrated neurodevelopmental abnormalities after brief volatile anesthetic exposures for noncardiac surgery, and a randomized controlled trial and a well designed ambidirectional cohort trial have found no cognitive abnormalities.5–7 Importantly, however, exposure times in these clinical studies have largely been limited to less than 2 h, while most animal studies have used exposures of 4 h or more. While the existence and exact timing of a potential vulnerable period of the developing animal and human brain for the deleterious effects of anesthetic exposure remain controversial, it is generally felt that susceptibility may be highest in younger children and primates under 4 to 5 yr of age.1,2
Neurodevelopmental disability is the most common complication for survivors of congenital cardiac surgery, and long-term studies in this vulnerable patient population have consistently reported significant cognitive and motor delays.8–10 Children requiring cardiac surgery may be particularly susceptible to potentially deleterious effects of anesthetics and sedatives, as they frequently undergo prolonged procedures and sometimes multiple anesthetic exposures early in life, often within the neonatal period. Furthermore, children requiring cardiac surgery are frequently treated with sedative medications postoperatively, which due to their similar mechanism of action compared to volatile anesthetics may contribute to subsequent neurocognitive impairment.11 Thus, we conducted a secondary analysis of the impact of exposure to volatile anesthetic agents and sedative medications in infants with congenital heart disease who were enrolled in a prospective observational study evaluating the relationship between environmental exposures and neurodevelopmental outcomes.12 The primary aim of this analysis was to describe the association between neurodevelopmental outcomes and cumulative inpatient exposure to sedative and anesthetic medications during the first 18 months of life in children with congenital heart disease. The primary hypothesis was that greater exposure to sedation and anesthesia in infants undergoing cardiac surgery would be associated with lower neurodevelopmental scores at 18-month follow-up. A secondary aim was to examine and describe patterns of sedation and anesthesia administration in infants with congenital heart disease over the first 18 months of life.
Materials and Methods
Ethics Approval and Reporting Guidelines
The Children’s Hospital of Philadelphia Institutional Review Board approved this study and waived the requirement for written informed consent. Informed consent was obtained for the parent study. This study adheres to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.13
This retrospective cohort study represents a secondary analysis of a prospective observational study evaluating environmental exposures and neurodevelopmental outcomes in pediatric patients with congenital heart disease undergoing surgery at Children’s Hospital of Philadelphia (Philadelphia, Pennsylvania), an urban, quaternary care pediatric hospital.12 The prospective observational study did not include the retrospective study variables of cumulative in-hospital anesthetic and sedative agent exposures from birth until the neurodevelopmental evaluation at 18 months of age, which were therefore collected from medical and anesthetic records.
This study’s inclusion and exclusion criteria were the criteria from the primary study, and the analysis plan was established before the data were accessed.12 Inclusion criteria for this study included the following: infants with congenital heart disease necessitating surgery with cardiopulmonary bypass, date of surgery before 44 weeks postconceptional age, biologic parents of the infant with congenital heart disease necessitating surgery with cardiopulmonary bypass, and parental/guardian permission (informed consent). The inclusion criteria for the primary study specified biologic parents because the study involved trio whole-exome sequencing. Exclusion criteria included a known genetic syndrome, a major extracardiac anomaly, and primary language other than English spoken in the home. The patients were evaluated by a genetic dysmorphologist. Neonatal recognition of dysmorphic features may be difficult; therefore, some patients for whom the diagnosis of a genetic syndrome was made at a later evaluation were enrolled. Genetic testing was performed as clinically indicated. The results of the genetic evaluations were classified as normal if no genetic or chromosome abnormality was demonstrated, abnormal if a specific diagnosis was confirmed, and suspect if there was evidence of a genetic syndrome that could not be confirmed.
Maternal education and the child’s race were determined by parental report. Familial socioeconomic status was assessed by parental report according to the Hollingshead scale.14
Cardiac diagnosis was coded according to a previously described classification incorporating anatomy and perioperative physiology that has been shown to be predictive of perioperative mortality.15 Class I represents two ventricles with no aortic arch obstruction, class II is two ventricles with aortic arch obstruction, class III is a single ventricle with no arch obstruction, and class IV is a single ventricle with arch obstruction. Accordingly, patients with tetralogy of Fallot and transposition of the great arteries are class I, for example, whereas patients with hypoplastic left heart syndrome represent class IV.
Follow-up Assessment at 18 Months
Medical follow-up included obtaining growth measurements and an interim medical history followed by a physical and neurologic examination by a team of developmental pediatricians and a nurse practitioner. All patients underwent evaluation with the Bayley Scales of Infant and Toddler Development, 3rd Edition (Bayley III), an individually administered instrument that assesses the developmental functioning of infants and young children between 1 month and 42 months of age.16 The Bayley III assesses development across five domains: cognition, language, motor, social–emotional, and adaptive. The composite scores have a mean of 100 and a SD of 15. The Bayley III cognitive, language, and motor composite scores were used as the primary outcomes for this study.
Measurements and Data Handling
The electronic health record and anesthesia information management system databases (CompuRecord, Phillips Medical System, USA; Epic Systems Corporation, USA; ChartMaxx, Quest, USA) were queried to obtain the patient characteristics and intensive care unit and perioperative data for patients meeting the inclusion criteria. Data filtering and validation were performed to identify patients incorrectly classified as eligible patients and to exclude from analysis any patients who met the exclusion criteria. The authors performed manual chart review and data validation for every study patient. Exposure data including patient encounters, presence and absence of bolus, and continuous infusions medication during encounters, continuous infusion calculations, and medication sum calculations were audited at multiple time points over the course of the study. All audits demonstrated reliable data with error rates of 0 to 5%. The errors were corrected before statistical analysis.
Volatile anesthetic agent (halothane, sevoflurane, isoflurane, and desflurane) exposure data for all subjects’ procedural and imaging encounters at our institution were retrieved from the anesthesia information management system and electronic health record databases (CompuRecord, Phillips Medical System; Epic Anesthesia, Epic) from the time of first anesthetic exposure to the date of neurodevelopmental assessment. All exposures to volatile anesthetic agents were recorded at intervals ranging from 15 s to 1 min throughout each anesthesia encounter, with end-tidal exhaled volatile agent concentrations used for exposure calculations. Because volatile anesthetic agent administration is frequently continued until time of departure from the anesthetizing location, elimination assumptions were developed and used if the last recorded expired concentration of volatile anesthetic agent in the anesthetizing location was greater than 0 so as to account for total volatile anesthetic agent exposure.11 Specifically, end-tidal concentrations of all volatile anesthetic agents were assumed to drop by 50% at 5 min after cessation of administration. A concentration drop to 0 in 10 min was assumed for sevoflurane and desflurane, and a drop to 0 in 30 min was assumed for halothane and isoflurane.11 Total volatile anesthetic agent exposure was then calculated and converted to age-adjusted minimum alveolar concentration hours (MAC-h).17–19 Total volatile anesthetic agent exposure in cumulative MAC-h was then determined for each patient inclusive of all known anesthetic exposures before the developmental assessment at the 18-month follow-up visit, standardized, and summed between different gases. Of note, end-tidal concentration for volatile anesthetic exposure during cardiopulmonary bypass could not be quantified and was therefore not included in this analysis.
Data on subjects’ exposures to sedation, anesthesia, and opioid medications in the cardiac intensive care unit (ICU) were abstracted from electronic health record systems (ChartMaxx; Epic). Each subject’s daily weight was obtained to calculate weight-adjusted doses. Cardiac ICU exposures were tallied hourly and included boluses and continuous infusions of all sedatives, anesthetics, and analgesics given during hospital admissions before the 18-month follow-up assessment. Perioperative sedative exposures were also collected for all surgical procedures that occurred before the 18-month follow-up visit. All administered opioid doses of fentanyl, hydromorphone, demerol, oxycodone, and remifentanil were converted to weight-adjusted intravenous morphine equivalents.20 All benzodiazepines including administered doses of midazolam, diazepam, and lorazepam were converted to weight-adjusted intravenous midazolam equivalent total doses.21 For each patient, all doses were summed, and weight-adjusted totals were generated for exposures of ketamine (mg/kg), dexmedetomidine (μg/kg), opioids (mg/kg), and benzodiazepines (mg/kg) until the 18-month follow-up.
There was no statistical power analysis conducted before the study. The intent was to recruit as many eligible patients as possible from the primary study to establish a representative sample; thus, the sample size was based on a convenience sample. The primary outcomes of interest were the Bayley III motor, language, and cognitive scores. The primary exposures of interest included the following: exposures to volatile anesthetic agents, total operating room and ICU dexmedetomidine, total opioids, total benzodiazepines, and total ketamine. The patient characteristics considered as potential confounders included genetic anomaly, sex, race, mother’s education level, socioeconomic class, need for extracorporeal membrane oxygenation (ECMO) therapy, birth weight Z score, and hospital length of stay.
Basic summary statistics were generated, as appropriate for the data, such as counts and percentages or means and standard deviations or median and interquartile range for all primary outcomes, primary confounders, and patient characteristics. These were followed by scatter plots and bivariate analysis between each combination of outcome and independent variable. The bivariate analyses utilized Spearman correlations for continuous variables and chi-square or Fisher’s exact test, as applicable, for categorical variables. Last, general linear models were run to assess the effect of each primary exposure of interest on each primary outcome of interest while adjusting for potentially confounding patient characteristics, with and without statistical outliers. Separate regressions were run for each primary drug exposure of interest and each outcome. Any observation beyond the 99.7% confidence limits was considered a statistical outlier. A two-sided P value of less than 0.05 was used as the criterion for statistical significance, and all results were considered exploratory. The data analysis was performed using SAS (USA) software version 9.4.22
Of the 140 patients enrolled in the initial study, 10 patients died, and 20 were lost to follow-up and excluded from the primary study. All patients in the primary study met the inclusion criteria for this study. Thus, the primary study’s cohort and this study’s cohort consisted of 110 patients enrolled between September 1, 2011, and August 31, 2015, who completed neurodevelopmental evaluations at the 18-month follow-up visit. Table 1 summarizes the demographic and clinical characteristics of the study cohort and patients who were alive yet lost to follow-up (“live nonreturners”).12 The median birth weight was 3.3 kg (interquartile range, 3.0 to 3.6 kg). A total of 91 subjects were admitted on the first day of life, and the remaining 19 subjects were admitted at ages 2 to 11 days. Median length of stay was 15 days (interquartile range, 11 to 23 days). The cohort included 41 (37%) single-ventricle patients and 69 (63%) two-ventricle patients.
Table 2 shows the counts and percentages of patients exposed to sedation and anesthesia at the 18-month follow-up visit. Weight-adjusted anesthetic and sedative exposures at 18-month follow-up are listed in table 3. During the first hospitalization, infants received at least one anesthetic or sedative drug for a median of 12 days (interquartile range, 9 to 19 days), which on average comprised 81% (95% CI, 77 to 84%) of the initial length of stay. Concurrent administration of two or more sedative agents (i.e., polypharmacy) in the intensive care unit occurred for a mean 38% (95% CI, 34 to 42%) of the initial length of stay. By the 18-month follow-up, over half of the study cohort had been exposed to two or more of the following agents: volatile anesthetic, fentanyl, ketamine, morphine, oxycodone, pentobarbital, midazolam, and dexmedetomidine. When compared to patients with two ventricles, single-ventricle patients had been exposed to polypharmacy to a greater degree at 18-month follow-up (table 2), and single-ventricle patients received higher doses across all anesthetic and sedative agents (supplemental table 1, https://links.lww.com/ALN/D233).
Bayley III motor, cognitive, and language scores at the 18-month follow-up visit for the entire study cohort and grouped by cardiac class are shown in table 4. The Bayley III motor, cognitive, and language scores for subjects who did not receive ECMO (n = 107) were 92.7 ± 1.1, 94.0 ± 1.1, and 92.9 ± 1.7, respectively, while the Bayley III motor, cognitive, and language scores for those needing ECMO (n = 3) were 66.3 ± 6.3, 65.0 ± 6.8, and 59.0 ± 10.0, respectively. Two-ventricle patients (cardiac classes I and II, n = 69) had significantly higher Bayley III motor scores than single-ventricle patients (cardiac classes III and IV, n = 41; 94.4 ± 9.3 and 87.8 ± 14.0; P = 0.004), while Bayley III cognitive and language scores did not significantly differ across the two-ventricle and single-ventricle groups (cognitive, two-ventricle: 94.3 ± 11.1, single ventricle: 91.4 ± 14.9; P = 0.25; language, two-ventricle: 93.0 ± 18.4, single ventricle: 90.2 ± 17.7; P = 0.44).
All three Bayley III scores demonstrated significant bivariate associations with patient race (motor: F2,108 = 6.0, P = 0.003; cognitive: F2,108 = 5.8, P = 0.004; language: F2,108 = 4.7, P = 0.011), sex (motor: F1,109 = 22.1, P < 0.001; cognitive: F1,109 = 9.2, P = 0.003; language: F1, 109 = 15.6, P < 0.001), presence of genetic anomaly (motor: F2,108 = 11.9, P < 0.001; cognitive: F2,108 = 5.0, P = 0.008; language: F2,108 = 10.5, P < 0.001), birth weight Z score (motor: F1,109 = 9.4, P = 0.003; cognitive: F1,109 = 7.5, P = 0.007; language: F1,109 = 7.1, P = 0.009), and exposure to extracorporeal membrane oxygenation (motor: F1,109 = 17.1, P < 0.001; cognitive: F1,109 = 17.6, P < 0.001; language: F1,109 = 11.2, P = 0.001; supplemental table 2, https://links.lww.com/ALN/D233). Bayley III cognitive and language scores were associated with maternal education (cognitive: F3,107 = 5.0, P = 0.003; language: F3,107 = 7.3, P < 0.001), while Bayley III motor scores were not (F3,107 = 2.1, P = 0.10). These potentially confounding patient characteristics and logical confounders, such as first admission length of stay, were adjusted for as covariates in the final models for our primary confounders: exposures to sedation and anesthesia.
Table 5 displays the analysis of neurodevelopmental outcomes based on patients’ exposures to sedation and anesthesia at first admission up to 18-month follow-up. All of the general linear model results presented in the table are individual regressions for each drug exposure separately. Adjusted for the variables listed in table 5, no statistically significant associations were observed between cumulative volatile anesthetic agent exposure up to 18-month follow-up and Bayley III motor, cognitive, and language scores. At 18-month follow-up, after adjusting for confounders, each mg/kg increase in ketamine use was associated with a 0.34-point decrease in Bayley III motor scores (Beta = −0.34; 95% CL = −0.64 to −0.05; P = 0.0242). Even after removing the statistical outliers, the effect was statistically significant (Beta = −0.36; 95% CL = −0.702 to −0.0133; P = 0.0419). Exposures to opioids, benzodiazepines, and dexmedetomidine were not associated with abnormal Bayley III scores after adjusting for multiple covariates and statistical outliers (table 5; supplemental table 3, https://links.lww.com/ALN/D233).
The principal finding of this study is that total cumulative doses of volatile anesthetic agents at 18-month follow-up were not associated with neurodevelopmental impairment in infants who received general anesthesia and intensive care unit sedation for a variety of imaging studies and surgical and interventional cardiac procedures. However, cumulative ketamine exposure was associated with diminished motor scores. As previously demonstrated, nonmodifiable patient factors such as race and presence of a genetic anomaly were associated with lower neurodevelopmental scores at 18-month follow-up.12 We specifically focused this analysis on a patient population that experiences a substantial burden of cumulative anesthetic and sedative exposures early in life, as a meta-analysis of animal and human data suggested a dose dependency of deleterious outcomes.23 In accordance with these findings, infants in this study were exposed to substantial cumulative anesthetic doses and considerable polypharmacy during perioperative and intensive care, with all receiving an inhaled anesthetic and a majority being exposed to opioids, ketamine, benzodiazepines, and barbiturates by 18 months of age. Due to the substantial disease burden in our patient cohort, these data cannot be compared with unexposed children but rather must be assessed within the cohort by looking for a dose-response relationship.
This study differs from other studies that observed associations between cumulative exposures to sedative and anesthetic agents and worse neurodevelopmental outcomes in children after cardiac surgery. Andropoulos et al.24 reported an association between volatile anesthetic agent exposure during the first 12 months of life, brain injury, ICU length of stay, and lower Bayley III cognitive composite scores. A similar study by Guerra et al.25 found no evidence of an association between dose and duration of sedation or analgesia drugs during the operative and perioperative period during the first 6 weeks of life and adverse neurodevelopmental outcomes. Our study included a larger study population and a greater variety of sedative medications (e.g., ketamine, propofol, and dexmedetomidine) in the multivariable analysis, as well as a longer study duration of cumulative sedative and anesthetic medications for 18 months of life, which may help explain the discrepant results. Like our results, Andropoulos et al.24 also did not find any effect of cumulative volatile anesthetic dose on Bayley III motor or language composite scores.
Diaz et al.11 reported that in children with hypoplastic left heart syndrome and single ventricle variants, greater cumulative duration of exposure to volatile anesthetic agents was associated with worse neurodevelopmental outcomes in certain domains at ages 4 and 5 yr. The contrasting results in our study may be attributed to the inclusion of patients with a broader range of congenital cardiac conditions; some patients might have been potentially less vulnerable to injury than the chronically cyanotic patients in the study of Diaz et al.11 For example, we observed no association between Bayley III motor scores and volatile anesthetic agent exposures in the total cohort, yet the single-ventricle patients in our study had significantly lower scores when compared to two-ventricle patients. Our study also had less variability in exposures compared to the study of Diaz et al.,11 which may have contributed to the lack of an observed effect and, in agreement with the current literature, identified maternal education as a confounder of neurodevelopmental outcome, which the Diaz study did not. Moreover, total MAC-h exposures over the 4- to 5-year study period in the Diaz study were much higher than in our study, which primarily focused on exposures in infancy and assessed patients at 1.5 yr of age. Patients undergoing staged reconstruction of single-ventricle physiology will have additional exposures, and in this small subset of single-ventricle patients, repeated anesthetic exposures and longer durations of anesthesia may exacerbate the risks of neurologic injury.11,26
In our study, dexmedetomidine did not demonstrate a neuroprotective effect as seen in some animal studies when administered during volatile anesthetic agent exposures.27,28 This might be explained by the comparatively lower perioperative dexmedetomidine doses in our clinical setting and the timing of the administration at our institution toward the end of the volatile anesthetic exposure. This agrees with our animal studies, which also did not demonstrate any protective effects of dexmedetomidine administration during sevoflurane exposure.28,29
While cautioning against the use of ketamine, our findings suggest that in critically ill children undergoing major surgical procedures, such as repair of major congenital heart disease, factors other than volatile anesthetic exposure should be considered as potentially more impactful to alleviate postoperative neurodevelopmental impairment.30 Indeed, the primary study found that greater concentrations of biomarkers of exposure to environmental chemicals, especially phthalates, were associated with poorer performance for language and motor skills at 18 months of age after adjustment for known risk factors for adverse neurodevelopment outcomes. Many of these chemicals are known endocrine-disrupting compounds and/or neurotoxicants, and we identified a greater adverse effect in girls compared to boys.12 Periodic developmental surveillance, screening, and evaluation in children with congenital heart disease may enhance identification of significant deficits and enable appropriate education and therapy to improve neurodevelopment and functioning.8 Novel intervention opportunities are being explored to improve neurodevelopmental outcomes in children with congenital heart disease during the perioperative period, through the utilization of new stratification schemata and cognitive interventions.31,32 Research is also ongoing into elucidating the social determinants of health and outcomes for children with congenital heart disease; many of these factors are modifiable and are additional avenues to improve outcomes in these vulnerable patients.33 Maternal anxiety in light of a diagnosis of congenital heart disease in their unborn child, for example, has also been demonstrated to adversely affect the baby’s neurodevelopment and may be amenable to intervention.34
While the Bayley Scales of Infant and Toddler Development are the most widely used standardized developmental test battery in medical and educational settings, a different neurodevelopmental assessment tool or testing at a later age might have resulted in alternative findings.35 Evaluation at age 18 months may not predict later deficits, and the impact of anesthetic and sedative exposure may be greater on higher cognitive tasks such as executive function and social skills, which can only be tested at an older age. A systematic review and meta-analysis of common prospectively collected outcomes showed that a single exposure to general anesthesia in early childhood was associated with statistically significant increases in parent reports of behavioral problems, but no difference in general intelligence.36 However, that study used two different measures, a full-scale intelligence quotient and parent-reported outcomes: the Behavior Rating Inventory of Executive Function and the Child Behavior Checklist. Another study reported that a correlation between learning disorders and anesthesia in children was of a lower magnitude than between learning disorders and hospitalization.37 This heterogeneity in neurodevelopmental outcome measures is important to consider when assessing studies on the topic. Moreover, it is unclear whether the neonatal and infant periods represent the most susceptible period for potentially deleterious effects of anesthetic or sedative exposure in humans and how findings in immature animals can be translated into clinical practice.38
Our study has several additional limitations. First, any retrospective study is limited by confounders and cannot establish causation. While our analysis included many relevant covariates, other factors could certainly have influenced our observations. For example, in any study of neurologic outcomes in patients with congenital heart disease, it is important to determine whether there is an underlying syndromic diagnosis that in and of itself could affect outcomes. While all our patients underwent genetic assessment and major abnormalities were excluded, other undiagnosed genetic anomalies might have affected our findings. Second, while our study was sufficiently powered to detect significant associations between neurodevelopmental scores and patient factors, it may have been underpowered to detect associations between anesthetic and sedative exposures and our outcomes of interest. A larger sample with the same exposure profile might have shown an effect. However, our results still support that patient factors were more impactful on neurocognitive outcomes than cumulative anesthetic or sedative exposures. Third, due to an inability to measure volatile anesthetic agent concentrations in the cardiopulmonary bypass oxygenator, our analysis of volatile anesthetic agent exposure excluded volatile anesthetic agents administered during cardiopulmonary bypass, thereby underestimating total volatile anesthetic agent exposure in the study cohort and serving as a source of information bias. Fourth, while we focused primarily on our analysis sans extreme outliers, these outliers highlight the importance of mindfulness when managing sedative use in infants with extremely long lengths of stay. Fifth, since the potential neurologic phenotype of anesthetic exposure remains unresolved, testing at an older age might provide different results from our assessment at 18 months, and noncognitive abnormalities were not assessed. Sixth, 30 of 140 (21%) individuals were lost to follow-up, which is a potential source of selection bias. Seventh, while we performed extensive manual review of the data set, we did not use any formal methods to address missing data; the data were assumed to be missing at random. Last, no standard pre- or perioperative assessment of baseline neurologic abnormalities was performed; these data were not collected as part of the primary study.12
In infants undergoing corrective and palliative surgery for major congenital heart disease, cumulative volatile anesthetic exposures were not associated with lower neurodevelopmental assessment scores at 18-month follow-up after adjusting for nonmodifiable patient factors and length of stay. These findings contrast with studies that have shown volatile anesthetic exposure to be associated with worse neurodevelopmental outcomes in infants. However, greater ketamine exposure was associated with lower Bayley III motor scores. Important differences between the substantial structural and functional abnormalities observed in animals after prolonged anesthetic exposure and their clinical relevance for human infants require further research. Additional studies are needed to better understand timing of exposure relative to brain developmental state and to identify modifiable factors to improve neurodevelopmental outcomes in infants with congenital heart disease. Other factors beyond anesthetic neurotoxicity should be considered, such as additional perioperative factors and social determinants of health.30
The authors thank Hailey Collier, Pharm.D., B.C.P.P.S., B.C.P.S. (Department of Pharmacy Services, Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania), for her valuable input. The authors acknowledge the significant contributions of Kevin Walsh, M.D. (Department of Emergency Medicine, University of Michigan Medical School, Ann Arbor, Michigan), in data collection and validation.
Supported by the Daniel M. Tabas and Alice Langdon Warner Endowed Chairs in Pediatric Cardiothoracic Surgery and National Institutes of Health National Center for Advancing Translational Sciences (Bethesda, Maryland) grants UL1TR000003 and UL1TR001878.
The authors declare no competing interests.
Supplemental Digital Content
Supplemental Tables, https://links.lww.com/ALN/D233
Supplemental Table 1. Summary statistics on weight-adjusted anesthetic and sedative exposures for patients grouped by cardiac diagnosis class
Supplemental Table 2. Bivariate associations between Bayley III scores and patient and maternal factors
Supplemental Table 3. Data on total perioperative and ICU opioid exposure outliers at 18-month follow-up visit (n = 3)