Abstract
Background:
Recent studies report conflicting results regarding labour epidural analgesia (LEA) in mothers and neurodevelopmental disorders in the offspring. This study evaluates behavioural and neuropsychological test scores in children of mothers who used LEA.
Methods:
Children enrolled in the Raine Study from Western Australia and delivered vaginally from a singleton pregnancy between 1989 and 1992 were evaluated. Children exposed to LEA were compared to unexposed children. The primary outcome was the parent-reported Child Behaviour Checklist (CBCL) reporting total, internalizing, and externalizing behavioural problem scores at age 10. Score differences, an increased risk of clinical deficit, and a dose-response based on duration of LEA exposure were assessed. Secondary outcomes included language, motor function, cognition, and autistic traits.
Results:
Of 2180 children, 850 (39.0%) were exposed to LEA. After adjustment for covariates, exposed children had minimally increased CBCL Total scores (+1.41 points; 95% CI 0.09, 2.73; p = 0.037), but not Internalizing (+1.13 points; 95% CI −0.08, 2.34; p = 0.066) or Externalizing (+1.08 points; 95% CI −0.08, 2.24; p = 0.068) subscale subscores. Increased risk of clinical deficit was not observed for any CBCL score. For secondary outcomes, score differences were inconsistently observed in motor function and cognition. Increased exposure duration was not associated with worse scores in any outcomes.
Conclusions:
LEA exposure was associated with slightly higher total behavioural scores, but there was no difference in subscores, increased risk of clinical deficits, or dose-response relationship. These results argue against LEA exposure being associated with consistent, clinically significant neurodevelopmental deficits in children.
Keywords: labour epidural analgesia, neurodevelopment, behaviour, neurotoxicity, the Raine Study
Introduction
Labour epidural analgesia (LEA) is the gold standard for maternal pain relief during labour and delivery, with current rates ranging between 60 and 76% in the United States.1 Neuraxial labour analgesia is a safe analgesic approach for mothers and is associated with reduced maternal morbidity and postpartum hemorrhage.2,3 The long-term impact of in-utero exposure to LEA on children has not been conclusively determined.
A recent controversial study suggested an association between LEA for vaginal delivery and an increased incidence of autism spectrum disorder (ASD) in offspring.4 Methodological issues were raised, particularly surrounding residual confounders not accounted for regarding family history of ASD, parental social characteristics, and pregnancy-related complications. Subsequent studies have shown mixed results. Some studies examining longitudinal birth cohorts found no association between LEA and ASD in children,5–7 while others identified small increased risks.8–12 Most other studies evaluating other neurodevelopmental delays and learning disabilities after LEA exposure found no association,7,13–15 but one found elevated risks of neurodevelopmental delays throughout early childhood.16 Most studies evaluated administrative datasets, which often lack patient-level familial or perinatal characteristics that may confound the association between maternal LEA and child outcomes. In several studies identifying positive results, associations were attenuated when evaluating siblings discordant on exposure status, suggesting that residual confounding may have been partially responsible for the observed associations.7,10,11
Diagnoses for ASD and other neurodevelopmental delays were commonly evaluated as outcomes. However, children with ASD often have a range of signs, symptoms, and levels of severity, but remain lumped under one diagnosis. This evaluation of ASD as a categorical outcome is considered a major limitation of ASD research,17 and may result in under-identification of meaningful toxic effects, misclassification, and bias.18 As such, continuous neuropsychological outcomes are considered to be superior assessments of development than binary outcomes such as an ASD diagnosis.19
In addition, few studies evaluated post-exposure mediating factors associated with LEA that may clarify the causality of observed results. Maternal intrapartum fever and oxytocin use during labour are strongly correlated with LEA, but results concerning their mediating influence on LEA’s association with neurodevelopment are mixed.4,12,20
In this study, behavioural problems in children with in-utero exposure to maternal LEA as measured by the Child Behaviour Checklist (CBCL) were evaluated primarily after accounting for sociodemographic, medical, psychosocial, antenatal, perinatal, and foetal characteristics. In the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, disorders such as ASD and attention deficit hyperactivity disorder (ADHD) are classified in the same neurodevelopmental cluster, and the CBCL has been used to assess symptoms of ASD and ADHD.21–25 As deficits in other neurodevelopmental domains have also been reported,16 a range of other domains were also evaluated.
Methods
This study was approved by the Institutional Review Board at Columbia University Irving Medical Center, with the requirement of written informed consent waived. Data collection and storage were approved by Ethics Committees at King Edward Memorial Hospital, Perth; Princess Margaret Hospital, Perth; and the University of Western Australia. The manuscript was written in accordance with STROBE guidelines.
The Raine Study
The Raine Study is a prospective multigenerational cohort in Perth, Australia, created to study the life-course of a cohort of Western Australians. Generation 1 consists of mothers who delivered between 1989 and 1992, Generation 2 their 2868 children. Mothers received prenatal care at clinics associated with King Edward Memorial Hospital and were expected to deliver at that hospital. Mothers and children in the Raine Study are representative of the general Australian population,26 making the results of the Raine Study externally generalizable.27
Labour epidural analgesia exposure and covariates
Detailed sociodemographic and medical data were collected for mothers and children, including a range of antenatal and perinatal data. In-utero exposure to maternal LEA and the duration of exposure were determined from data extracted from maternal medical records collected at the time of delivery. Vaginally delivered children from a singleton pregnancy were included for analysis, as indications for and factors present during caesarean birth would introduce additional confounding.
76 covariates were selected a priori to control for confounding. Sociodemographic characteristics included maternal age, race, religion, birth country, time in Australia, substance use, education, and BMI; paternal age, time in Australia, education, and cigarette use; whether the parents live together; family’s primary language; family income; and the number of recent moves. Maternal medical characteristics included diagnoses of asthma, epilepsy, hypertension, heart disease, thyroid disease, hepatitis, diabetes, anaemia, platelet disease, or psychiatric disease. Maternal psychosocial characteristics included deaths of close relatives or friends; problems with children, marriage, or money; physical or emotional trauma during pregnancy; and number of emotional upsets during pregnancy. Antenatal characteristics included gravidity, parity, whether the pregnancy was planned, attendance of antenatal classes, substance use, antenatal admissions or surgery, antenatal urinary tract infection or renal disease, pre-eclampsia, gestational diabetes, and prescription medication or acetaminophen use. Perinatal characteristics included antepartum bleeding, prostaglandin use, onset of labour, oxytocin for induction, opioids during labour, duration of the first stage of labour, preterm labour, and tocolytic use. Neonatal characteristics included sex, birth weight, and congenital abnormalities.
Other perinatal or neonatal factors or complications may also occur following LEA that could affect neurodevelopment. If these factors occur because of exposure, controlling for them may result in overadjustment bias. Intrapartum fever requiring antibiotics, oxytocin for augmentation of labour after LEA placement (separate from oxytocin for induction of labour, which was considered a confounder), and foetal heart rate (FHR) abnormalities were identified as post-exposure covariates and evaluated as potential mediators. A directed acyclic graph (eFigure 1) identifies confounders, mediators, and additional factors for which this study was unable to control.
Neurodevelopmental outcomes
Neurodevelopmental outcomes in children were prospectively collected in the Raine Study for reasons unrelated to LEA. Primary and secondary outcomes were selected prior to starting any analyses. The Child Behaviour Checklist (CBCL), administered at age 10, was chosen as the primary outcome. The CBCL is a 118-item parental questionnaire reporting a child’s internalizing (e.g. questions about anxiety, depression, somatic symptoms) and externalizing (e.g. questions about aggression, tantrums, hyperactivity) behavioural problems. The CBCL reports Total scores, as well as Internalizing and Externalizing subscale subscores; higher scores indicate more behavioural problems. Scores are normed by age and gender for a mean of 50 and standard deviation of 10, with worse scores indicative of a possible disorder.15,22–25,28
Six additional assessments examining other neurodevelopmental domains were evaluated as exploratory secondary outcomes. Language was evaluated using the Clinical Evaluation of Language Fundamentals (CELF) and the Peabody Picture Vocabulary Test (PPVT). The CELF reports scores for receptive, expressive, and total language abilities, and the PPVT is a receptive listening vocabulary test. Development of fine and gross motor function was assessed with the McCarron Assessment of Neuromuscular Development (MAND). Cognition was evaluated using Raven’s Coloured Progressive Matrices (CPM) and the Symbol Digit Modality Test (SDMT). CPM assesses global cognitive performance, abstract reasoning, visuospatial functions, and nonverbal intelligence. The SDMT tests visual tracking, attention, and fine motor skills and is administered both written and orally. These assessments were administered by neuropsychologists to children at age 10. In these assessments, higher scores indicate better performance.
The final exploratory secondary outcome, the Autism Spectrum Quotient (AQ), is a fifty-item self-report questionnaire that solicits the degree of autistic traits exhibited by the subject and has been validated in Australian populations.29 AQ was completed by children at age 20. Higher scores indicate more autistic traits.
Statistical analysis
Mean scores for neurodevelopmental outcomes in exposed and unexposed groups were initially evaluated in unadjusted analyses. Missing covariate values were imputed using multiple imputation, generating five imputed datasets for analysis.30 To account for differences in exposed children, the predicted probability of LEA exposure, conditional on all covariates, was applied to Inverse Probability of Treatment Weights (IPTW). Appropriate balance was evaluated by confirming that standardized differences in all covariate means were below 0.1 following IPTW.31 To account for missing outcome data, censoring conditional on all covariates and exposure status was applied to Inverse Probability of Censoring Weights (IPCW). IPCW compensates for censored subjects by giving more weight to subjects with similar characteristics who were not censored.32 Adjusted analyses on each imputed dataset were pooled using Rubin’s rules, generating 95% confidence intervals (CI) reporting confidence ranges around each estimate.33
In the primary analysis, linear regression with IPTW and IPCW estimated mean score differences between LEA-exposed and unexposed children. As a sensitivity analysis, IPTW was used to adjust for confounders, but without IPCW accounting for censoring.
Several secondary analyses were performed. First, modified Poisson regression with robust variance, adjusted by IPTW and IPCW, assessed the adjusted relative risk of crossing a score threshold for clinical deficit. For the CBCL, the threshold used was 60, one standard deviation above the population mean score.34,35 The threshold used for AQ was 29, a pre-established cutoff.29 For other exploratory secondary outcomes, the threshold was defined as one standard deviation below the mean in this cohort, approximately representing the worst 16th percentile of scores. The purpose of this secondary analysis was to determine whether any identified difference in outcome scores was associated with a significantly higher proportion of children classified as having a clinical deficit in each given score.
Second, a dose-response was assessed in LEA-exposed children. The association between LEA exposure duration, the time between LEA placement and birth, and outcome scores was evaluated using multivariable linear regression, conditional on all 76 covariates. As a sensitivity analysis, assessment scores following short vs. long exposures were evaluated, short exposures being less than the median exposure duration and long exposures greater than or equal to the median.
Finally, mediation analyses, adjusted with IPTW and IPCW, were performed for each identified post-exposure covariate on outcomes showing significant score differences between exposed and unexposed children. These analyses decompose the total observed effect of exposure to LEA on outcome scores into the direct effect of LEA not transmitted through the mediator, and the indirect effect transmitted through the mediator. Since mediation analysis may be susceptible to bias in the presence of exposure-mediator interaction, interactions between the mediator and exposure variables were included in regression models.36–38 In these mediation analyses, FHR abnormalities were encoded two ways: all FHR abnormalities, and only severe abnormalities associated with foetal distress, including severe variable decelerations, late decelerations, variable and late decelerations, flat tracing, and foetal bradycardia.
SAS 9.4 was used for all statistical analyses, with mediation analysis performed using the CAUSALMED procedure.
Results
Of the 2868 children in the Raine Study, 2180 (76.0%) were delivered vaginally from a singleton pregnancy. Of these children, 850 (39.0%) were exposed to LEA (Figure 1).
Figure 1.
Flowchart of children included for analysis in this study, listing completion rates of primary and secondary outcomes by exposure status.
Abbreviations: CBCL: Child Behaviour Checklist; CELF: Clinical Evaluation of Language Fundamentals; PPVT: Peabody Picture Vocabulary Test; MAND: McCarron’s Assessment of Neuromuscular Development; CPM: Coloured Progressive Matrices; SDMT: Symbol Digit Modality Test; AQ: Autism Spectrum Quotient.
Characteristics of exposed and unexposed children
Mothers with LEA (exposed children) were more likely to be primiparous, attend antenatal classes, have an antenatal hospital admission, smoke cigarettes while pregnant, and have comorbid medical conditions such as asthma, anaemia, or pre-eclampsia. During delivery, mothers with LEA were more likely to receive prostaglandins or oxytocin for induction of labour and have a longer first stage of labour. Exposed children also had a higher neonatal birth weight. A selected list of covariates by exposure status is included in Table 1, while a full list is included in eTable 1.
Table 1.
Selected covariates for labour epidural analgesia (LEA)-exposed and unexposed children. (The full list of covariates is available in eTable 1 in the Supplementary Appendix)
| No LEA (%) n = 1330 |
LEA (%) n = 850 |
|
|---|---|---|
| Parental sociodemographic characteristics | ||
| Mother’s age ≥ 35 years | 166 (12.5) | 85 (10.0) |
| Father’s age ≥ 35 years | 312 (23.5) | 141 (16.6) |
| Mother’s race white | 1153 (86.7) | 772 (90.8) |
| Mother native to Australia | 775 (58.3) | 518 (60.9) |
| Family income ≥ $24k | 673 (50.6) | 456 (53.6) |
| Maternal health characteristics | ||
| Treated for asthma | 180 (13.5) | 156 (18.4) |
| Treated for epilepsy | 22 (1.7) | 26 (3.1) |
| Treated for anaemia | 339 (25.5) | 280 (32.9) |
| Maternal antenatal characteristics | ||
| Mother primiparous | 506 (38.0) | 550 (64.7) |
| Attended antenatal classes | 518 (38.9) | 491 (57.8) |
| Hospital admission during pregnancy | 243 (18.3) | 258 (30.4) |
| Pre-eclampsia during pregnancy | 240 (18.0) | 260 (30.6) |
| Used prescription medications during pregnancy | 215 (16.2) | 184 (21.6) |
| Used paracetamol products during pregnancy | 150 (11.3) | 113 (13.3) |
| Alcohol use during pregnancy | 721 (54.2) | 463 (54.5) |
| Cigarette use during pregnancy | 281 (21.1) | 234 (27.5) |
| Recreational drug use during pregnancy | 145 (10.9) | 101 (11.9) |
| Maternal perinatal characteristics | ||
| Received prostaglandins during labour | 31 (2.3) | 66 (7.8) |
| Labour onset via induction of labour | 355 (26.7) | 418 (49.2) |
| Received oxytocin for induction of labour | 198 (14.9) | 322 (37.9) |
| Duration of first stage of labour ≥ 12 hours | 43 (3.2) | 125 (14.7) |
| Child neonatal characteristics | ||
| Child sex male | 666 (50.1) | 436 (51.3) |
| Neonatal birthweight ≥ 4000 grams | 107 (8.0) | 92 (10.8) |
| Post-exposure mediators | ||
| Antibiotics for fever during labour | 11 (0.8) | 58 (6.8) |
| Oxytocin for augmentation of labour after LEA | 56 (4.2) | 157 (18.5) |
| Any neonatal foetal heart rate abnormality | 257 (19.3) | 424 (49.9) |
| Serious neonatal foetal heart rate abnormalitya | 174 (13.1) | 263 (30.9) |
Excluding mild-to-moderate variable decelerations, foetal tachycardia, or other unspecified abnormality.
Unadjusted analyses
CBCL scores were available in 68.6% (n = 583) of exposed children and 72.0% (n = 927) of unexposed children (Figure 1). eTable 2 compares covariate values between children who did and did not complete the CBCL. Mean scores and standard deviations of normative populations for primary and secondary outcomes are provided in eTable 3. The mean CBCL Total score and Internalizing and Externalizing subscores for this cohort were 47.4, 49.2, and 47.5, similar to the population means of 50. Prior to covariate adjustment, exposed children had higher, indicating worse, CBCL Total scores (+1.79 points; 95% CI 0.66, 2.92; p = 0.002) and higher Internalizing (+1.62 points; 95% CI 0.56, 2.69; p = 0.003) and Externalizing (+1.30 points; 95% CI 0.22, 2.37; p = 0.018) subscale subscores.
Availability of secondary outcomes ranged from 54.6% (n = 464) for the CPM to 43.8% (n = 372) for the AQ in exposed children and from 58.6% (n = 780) for the CPM to 45.3% (n = 602) for the AQ in unexposed children (Figure 1). In unadjusted analyses, statistically significant mean score differences were not observed in any secondary outcomes (Table 2).
Table 2.
Mean score differences between labour epidural analgesia (LEA)-exposed and unexposed children.
| Domain | Outcomea | Mean Scores | Unadjusted Score Differenceb | Adjusted Score Differencec | Adjusted Score Differenced | ||||
|---|---|---|---|---|---|---|---|---|---|
| No LEA | LEA | Estimate (95% CI) | p | Estimate (95% CI) | p | Estimate (95% CI) | p | ||
| Primary Outcome | |||||||||
| Behaviour | CBCL Total | 46.77 | 48.55 | 1.79 (0.66, 2.92) | 0.002 | 1.31 (0.16, 2.47) | 0.026 | 1.41 (0.09, 2.73) | 0.037 |
| CBCL Internalizing | 48.56 | 50.18 | 1.62 (0.56, 2.69) | 0.003 | 1.07 (−0.02, 2.15) | 0.054 | 1.13 (−0.08, 2.34) | 0.066 | |
| CBCL Externalizing | 47.00 | 48.30 | 1.30 (0.22, 2.37) | 0.018 | 1.00 (−0.07, 2.07) | 0.066 | 1.08 (−0.08, 2.24) | 0.068 | |
| Secondary Outcomes | |||||||||
| Language | CELF Total | 95.49 | 96.48 | 0.99 (−0.79, 2.76) | 0.28 | −1.42 (−3.20, 0.37) | 0.12 | −1.15 (−2.90, 0.60) | 0.20 |
| CELF Receptive | 100.93 | 101.90 | 0.97 (−0.84, 2.79) | 0.29 | −1.34 (−3.16, 0.48) | 0.15 | −1.41 (−3.19, 0.37) | 0.12 | |
| CELF Expressive | 90.81 | 92.28 | 1.47 (−0.29, 3.23) | 0.10 | −0.75 (−2.53, 1.02) | 0.40 | −0.20 (−1.96, 1.57) | 0.83 | |
| PPVT | 103.82 | 104.66 | 0.83 (−0.61, 2.28) | 0.26 | −0.73 (−2.11, 0.64) | 0.30 | −0.81 (−2.20, 0.58) | 0.26 | |
| Motor | MAND | 95.20 | 93.67 | −1.54 (−3.13, 0.06) | 0.059 | −1.78 (−3.32, −0.25) | 0.023 | −1.34 (−2.87, 0.19) | 0.085 |
| Cognition | CPM Total | 31.18 | 31.06 | −0.11 (−0.53, 0.31) | 0.60 | −0.39 (−0.80, 0.02) | 0.065 | −0.38 (−0.79, 0.03) | 0.071 |
| SDMT Written | 34.49 | 34.62 | 0.12 (−0.73, 0.98) | 0.78 | −0.82 (−1.69, 0.05) | 0.065 | −1.13 (−2.02, −0.24) | 0.013 | |
| SDMT Oral | 43.78 | 43.53 | −0.24 (−1.42, 0.93) | 0.68 | −1.23 (−2.40, −0.05) | 0.041 | −1.48 (−2.67, −0.30) | 0.014 | |
| Autism | AQ | 15.29 | 15.35 | 0.06 (−0.66, 0.78) | 0.87 | 0.34 (−0.39, 1.07) | 0.36 | 0.52 (−0.22, 1.27) | 0.17 |
Abbreviations: LEA: labour epidural analgesia; CI: confidence interval; CBCL: Child Behaviour Checklist; CELF: Clinical Evaluation of Language Fundamentals; PPVT: Peabody Picture Vocabulary Test; MAND: McCarron’s Assessment of Neuromuscular Development; CPM: Coloured Progressive Matrices; SDMT: Symbol Digit Modality Test; AQ: Autism Spectrum Quotient.
AQ is completed at age 19–20. All other assessments are completed at age 10. Higher scores on CBCL and AQ indicate worse scores; lower scores on CELF, PPVT, MAND, CPM, and SDMT indicate poorer performance.
Unadjusted for covariates.
Adjusted with inverse probability of treatment weights.
Adjusted with inverse probability of treatment weights and outcome-specific inverse probability of censoring weights.
Child behavioural outcomes after adjustment
Following IPTW, standardized mean differences for all covariates were below 0.1 (eFigure 2, eTable 4), indicating appropriate balance between exposed and unexposed children.39
Following adjustment for covariates and missing outcome data using IPTW and IPCW, exposed children had higher, indicating worse, scores on the CBCL Total (+1.41 points; 95% CI 0.09, 2.73; p = 0.037). No significant differences were seen with Internalizing (+1.13 points; 95% CI −0.08, 2.34; p = 0.066) or Externalizing (+1.08 points; 95% CI −0.08, 2.24; p = 0.068) subscale subscores. Sensitivity analyses implementing only IPTW demonstrated similar results (Table 2).
In secondary analyses, after adjustment using IPTW and IPCW, exposed children had no increased adjusted relative risk (aRR) of clinical deficit based on CBCL Total scores (aRR 1.11; 95% CI 0.75, 1.64; p = 0.60), Internalizing (aRR 1.25; 95% CI 0.87, 1.81; p = 0.23), or Externalizing (aRR 1.05; 95% CI 0.68, 1.63; p = 0.82) scores. Sensitivity analyses implementing only IPTW demonstrated similar results (Table 3). Primary and threshold analyses for individual imputed datasets are provided in eTable 5.
Table 3.
Risk of clinical deficit in children exposed to labour epidural analgesia (LEA) compared to unexposed children.
| Domain | Outcomea | No. with Deficit (%) | Unadjusted Relative Riskc | Adjusted Relative Riskd | Adjusted Relative Riske | ||||
|---|---|---|---|---|---|---|---|---|---|
| No LEA | LEA | RR (95% CI) | p | aRR (95% CI) | p | aRR (95% CI) | p | ||
| Primary Outcome | |||||||||
| Behaviour | CBCL Total | 107 (11.2) | 88 (15.1) | 1.35 (1.04, 1.76) | 0.025 | 1.13 (0.80, 1.60) | 0.50 | 1.11 (0.75, 1.64) | 0.60 |
| CBCL Internalizing | 113 (11.8) | 99 (17.0) | 1.44 (1.12, 1.85) | 0.004 | 1.25 (0.90, 1.74) | 0.18 | 1.25 (0.87, 1.81) | 0.23 | |
| CBCL Externalizing | 91 (9.5) | 66 (11.3) | 1.19 (0.88, 1.61) | 0.26 | 1.07 (0.73, 1.59) | 0.72 | 1.05 (0.68, 1.63) | 0.82 | |
| Secondary Outcomes | |||||||||
| Language | CELF Total | 114 (14.7) | 60 (13.0) | 0.88 (0.66, 1.18) | 0.42 | 1.10 (0.75, 1.63) | 0.62 | 1.05 (0.71, 1.56) | 0.80 |
| CELF Receptive | 103 (13.2) | 59 (12.8) | 0.97 (0.72, 1.31) | 0.85 | 1.24 (0.83, 1.84) | 0.29 | 1.15 (0.77, 1.73) | 0.49 | |
| CELF Expressive | 123 (15.9) | 67 (14.5) | 0.91 (0.69, 1.20) | 0.53 | 1.11 (0.77, 1.61) | 0.56 | 1.05 (0.72, 1.53) | 0.80 | |
| PPVT | 116 (16.0) | 64 (15.0) | 0.94 (0.71, 1.24) | 0.65 | 1.18 (0.81, 1.73) | 0.39 | 1.20 (0.82, 1.75) | 0.34 | |
| Motor | MAND | 103 (13.4) | 90 (19.6) | 1.46 (1.13, 1.90) | 0.004 | 1.52 (1.09, 2.13) | 0.014 | 1.47 (1.02, 2.11) | 0.040 |
| Cognition | CPM Total | 118 (15.1) | 76 (16.4) | 1.08 (0.83, 1.41) | 0.57 | 1.12 (0.78, 1.60) | 0.53 | 1.17 (0.81, 1.69) | 0.41 |
| SDMT Written | 123 (15.9) | 84 (18.2) | 1.15 (0.89, 1.47) | 0.30 | 1.32 (0.93, 1.86) | 0.12 | 1.45 (1.00, 2.09) | 0.049 | |
| SDMT Oral | 120 (15.5) | 75 (16.3) | 1.05 (0.81, 1.37) | 0.73 | 1.29 (0.91, 1.83) | 0.15 | 1.50 (1.04, 2.18) | 0.032 | |
| Autism | AQ | 11 (1.8) | 5 (1.3) | 0.74 (0.26, 2.10) | 0.57 | 0.44 (0.12, 1.59) | 0.21 | 0.74 (0.18, 2.99) | 0.68 |
Abbreviations: LEA: labour epidural analgesia; RR: relative risk; aRR: adjusted relative risk; CI: confidence interval; CBCL: Child Behaviour Checklist; CELF: Clinical Evaluation of Language Fundamentals; PPVT: Peabody Picture Vocabulary Test; MAND: McCarron’s Assessment of Neuromuscular Development; CPM: Coloured Progressive Matrices; SDMT: Symbol Digit Modality Test; AQ: Autism Spectrum Quotient.
AQ is completed at age 19–20. All other assessments are completed at age 10.
Score threshold above (CBCL, AQ) or below (all other assessments) which a deficit is defined. Thresholds for CBCL and AQ based on normalizing populations, and for all other assessments are one standard deviation worse than the mean score in this cohort.
Unadjusted for covariates.
Adjusted with inverse probability of treatment weights.
Adjusted with inverse probability of treatment weights and outcome-specific inverse probability of censoring weights.
Of the 850 exposed children, 828 had a recorded duration of LEA exposure. No association between LEA exposure duration and CBCL Total (+0.07 points hour−1; 95% CI −0.25, 0.40; p = 0.66), Internalizing (+0.03 points hour−1; 95% CI −0.29, 0.36; p = 0.84), or Externalizing (+0.10 points hour−1; 95% CI −0.21, 0.41; p = 0.52) scores was observed. Sensitivity analyses comparing short vs. long exposures demonstrated similar results (Table 4).
Table 4.
Mean score differences based on duration of labour epidural analgesia (LEA) exposure for exposed children.
| Domain & Outcomea | Participants (%) n = 828 |
Adjusted score difference for every one-hour increase in LEA exposureb | Adjusted score difference in long vs. short exposurec | |||
|---|---|---|---|---|---|---|
| Estimate (95% CI) | p | Estimate (95% CI) | p | |||
| Primary Outcome | ||||||
| Behaviour | CBCL Total | 573 (69.2) | 0.07 (−0.25, 0.40) | 0.66 | 0.03 (−2.01, 2.08) | 0.98 |
| CBCL Internalizing | 573 (69.2) | 0.03 (−0.29, 0.36) | 0.84 | −0.24 (−2.29, 1.81) | 0.82 | |
| CBCL Externalizing | 573 (69.2) | 0.10 (−0.21, 0.41) | 0.52 | 0.36 (−1.55, 2.27) | 0.71 | |
| Secondary Outcomes | ||||||
| Language | CELF Total | 454 (54.8) | 0.69 (0.14, 1.25) | 0.013 | 3.74 (0.31, 7.18) | 0.033 |
| CELF Receptive | 456 (55.1) | 0.79 (0.21, 1.37) | 0.008 | 4.34 (0.73, 7.96) | 0.018 | |
| CELF Expressive | 454 (54.8) | 0.52 (−0.03, 1.07) | 0.063 | 2.96 (−0.48, 6.41) | 0.092 | |
| PPVT | 420 (50.7) | 0.20 (−0.23, 0.63) | 0.36 | 1.35 (−1.35, 4.04) | 0.33 | |
| Motor | MAND | 453 (54.7) | −0.06 (−0.57, 0.45) | 0.81 | −2.18 (−5.40, 1.03) | 0.18 |
| Cognition | CPM | 457 (55.2) | 0.05 (−0.09, 0.18) | 0.49 | 0.45 (−0.39, 1.29) | 0.29 |
| SDMT Written | 455 (55.0) | 0.08 (−0.20, 0.35) | 0.59 | 1.45 (−0.24, 3.13) | 0.092 | |
| SDMT Oral | 454 (54.8) | 0.05 (−0.31, 0.40) | 0.80 | 1.15 (−1.05, 3.35) | 0.31 | |
| Autism | AQ | 368 (44.4) | 0.06 (−0.20, 0.32) | 0.65 | 0.00 (−1.56, 1.56) | 0.99 |
Abbreviations: CI: confidence interval; CBCL: Child Behaviour Checklist; CELF: Clinical Evaluation of Language Fundamentals; PPVT: Peabody Picture Vocabulary Test; MAND: McCarron’s Assessment of Neuromuscular Development; CPM: Coloured Progressive Matrices; SDMT: Symbol Digit Modality Test; AQ: Autism Spectrum Quotient.
AQ is completed at age 19–20. All other assessments are completed at age 10. Higher scores on CBCL and AQ indicate worse scores; lower scores on CELF, PPVT, MAND, CPM, and SDMT indicate poorer performance.
Adjusted for all covariates.
Adjusted for all covariates. Estimates represent changes in score for children with long (greater than the median 4.6-hour LEA exposure) relative to children with short (less than the median) exposures.
All post-exposure covariates—intrapartum fever requiring antibiotics, oxytocin for augmentation of labour, and FHR abnormalities—were more common in exposed children (eTable 1). None of the post-exposure covariates mediated observed CBCL Total score differences, and interactions between LEA and these post-exposure covariates also were not associated with CBCL Total scores (eTable 6).
Exploratory secondary neurodevelopmental outcomes after adjustment
After adjustment by IPTW and IPCW, exposed children had lower, indicating worse, scores on the SDMT Written (−1.13 points; 95% CI −2.02, −0.24; p = 0.013) and Oral (−1.48 points; 95% CI −2.67, −0.30; p = 0.014) assessments (Table 2). In sensitivity analyses implementing only IPTW, exposed children scored lower, indicating worse, scores on the MAND and SDMT Oral assessments (Table 2).
After adjustment by IPTW and IPCW, exposed children had an increased relative risk of clinical deficit on the MAND (aRR 1.47; 95% CI 1.02, 2.11; p = 0.040), SDMT Written (aRR 1.45; 95% CI 1.00, 2.09; p = 0.049), and SDMT Oral (aRR 1.50; 95% CI 1.04, 2.18; p = 0.032) assessments, but in sensitivity analyses using only IPTW, an increased risk of deficit was seen only on MAND. (Table 3). Increased exposure duration was not associated with worse performance in any secondary outcome (Table 4). Intrapartum fever, oxytocin for augmentation of labour, and FHR abnormalities after LEA did not mediate observed differences in SDMT scores in exposed children, and no exposure-mediator interactions were significant (eTable 6).
Discussion
In-utero exposure to LEA was associated with slightly higher, indicating worse, behavioural scores on the CBCL Total scale, but not Internalizing or Externalizing subscales. The small difference in CBCL Total scores was not associated with an increased risk of scoring more than one standard deviation above the normative mean for CBCL, suggesting that LEA did not significantly increase the number of children with clinical deficit. Similarly, longer exposures were not associated with worse scores. Intrapartum fever, oxytocin for augmentation of labour, and FHR abnormalities did not mediate the observed difference. The lack of an increased risk of clinical deficit suggests that this score difference is clinically insignificant, and the lack of dose-response between exposure duration and CBCL scores suggests that local anaesthetics may not be responsible for the observed difference.
Among the exploratory secondary outcomes, exposed children scored lower on the MAND and SDMT, indicating poorer performance, although these differences were inconsistent across sensitivity analyses. Increased duration of LEA exposure was not associated with worse scores in any secondary outcomes, and score differences were not secondary to mediation by intrapartum fever, labour augmentation, or FHR abnormalities. In this study, LEA exposure was not associated with language development or autistic tendencies.
That intrapartum fever did not account for observed differences is consistent with prior research,4 and the fact that FHR abnormalities did not mediate score differences suggests that any effect of LEA exposure on FHR abnormalities is not associated with worse neurodevelopmental outcomes in exposed children. These results also did not support a mediating role for oxytocin for augmentation of labour in the association between LEA exposure and neurodevelopment, in contrast to a recent study.12
This study’s findings are consistent with animal studies showing behavioural and motor deficits, but not cognitive changes, after local anaesthetic exposure.40–42 However, given the high administered doses and inconsistencies in the animal studies, they should be interpreted with caution.43 Furthermore, for LEA exposure to be causally connected to paediatric neurodevelopment, a mechanism of injury is needed. Potential hypotheses include apoptosis of neurons induced by intrinsic caspase, PI3K, and MAPK pathways,44–46 but at present, the mechanism for toxicity in preclinical studies remains unclear.
Strengths and Limitations
A major strength of this study is the availability of continuous neuropsychological outcomes, which offer a broad neurodevelopmental assessment of LEA-exposed children that is considered superior to the binary diagnoses used in prior studies.15,17–19 Additionally, covariate data in this prospective birth cohort were more detailed than that in most prior studies, particularly surrounding pre-existing medical conditions,8,10,11,13–16 maternal psychosocial characteristics,5,6,8–11,13–16 antenatal exposures and diagnoses,6,10,11,13,15,16 and perinatal exposures and complications.6,10,11,13,15 Finally, the implementation of IPCW allowed this study to better control for attrition bias.
This study, however, has several limitations. First, multiple secondary outcomes were evaluated. Given the possibility of a Type I error, results from the exploratory secondary outcomes should be interpreted with caution. Regarding one secondary outcome, the AQ is used to screen for autistic traits self-reported by children, and it does not establish a formal diagnosis of ASD. Similarly, the threshold analyses for all outcomes should not be interpreted as the adjusted relative risk of a formal psychiatric diagnosis. Second, despite the availability of a range of covariates, this study was observational, and unmeasured confounders are likely present. Similarly, while IPCW is a powerful tool to minimize censoring bias, unmeasured factors affecting differential censoring may also still be present. A sibling analysis such as that employed in prior studies would be helpful; however, only 12 children in the Raine Study were siblings.
Third, this cohort of children exposed to LEA was born between 1989 and 1992, and since then, obstetric anaesthesia practice has evolved. During the study period, intermittent boluses were the most common technique used to deliver LEA, a method now largely replaced with continuous infusions or low-dose programmed intermittent epidural bolus supplemented by patient-controlled epidural analgesia. Our dose-response analysis defines the cumulative dose as the duration between epidural placement and birth. While longer durations of epidural placement are likely associated with more local anaesthetic boluses and therefore higher doses, duration may be an imperfect estimation of total dose. Our duration analysis may also be limited by inadequate power or the presence of a ceiling effect.
Since the study period, obstetric anaesthesia practice has also moved towards a reduction of epidural local anaesthetic dosing. Based on discussions with obstetric anaesthesiologists practicing in Perth during that period, LEA routinely employed 0.25% to 0.5% bupivacaine, compared to the contemporary practice of 0.0625% bupivacaine. Consequently, children in this cohort were certainly exposed to significantly higher doses of local anaesthetic agents than doses currently used, and they may also have been exposed to more difficult deliveries due to increased maternal motor block. Yet we did not find an increased risk of clinically significant deficits in the primary or most secondary outcomes, which is reassuring.
In addition to progress in obstetric anaesthesia practice, obstetric and paediatric care have evolved, which could influence outcomes in children following LEA exposure.
While this study’s observational nature limits the ability to make causal conclusions, a study randomising which mothers receive LEA is likely unfeasible in contemporary obstetric and anaesthesia practice. Additionally, observational studies investigating associations between perinatal exposures and outcomes in twenty-year-old children must analyse twenty-year-old clinical practices. Both limitations are likely unavoidable in clinical studies on this topic.
Finally, since our results reflect exposures and outcomes in children born at a single centre in Perth, different results may be found in children receiving care elsewhere. Additionally, as our study design only included vaginal deliveries from singleton pregnancies, the results may not be generalizable to multiple gestation or caesarean deliveries. As the decision for the mode of delivery may occur after LEA placement, and LEA can influence the delivery mode, including only vaginal deliveries may also introduce selection bias.
Conclusions
In this observational cohort study of children born in Perth, Western Australia between 1989 and 1992, LEA exposure was associated with slightly worse total behavioural scores at age 10, but no difference in behavioural subscores and no increased risk of crossing a predefined threshold for clinical deficit, even at the significantly higher doses of local anaesthetics employed at the time. Although no differences were seen in autistic tendencies, inconsistent differences in motor and cognitive neurodevelopmental domains were observed. However, these findings have unclear long-term relevance and must be interpreted with caution. While this study is not without limitations, the results argue that in-utero LEA exposure is not associated with consistent, clinically significant neurodevelopmental deficits.
Supplementary Material
Acknowledgements:
We would like to acknowledge the Raine Study participants and their families for their ongoing participation in the study and the Raine Study team for study co-ordination and data collection. We also thank the NHMRC for their long-term contribution to funding the study over the last 30 years. The core management of the Raine Study is funded by The University of Western Australia, Curtin University, Telethon Kids Institute, Women and Infants Research Foundation, Edith Cowan University, Murdoch University, The University of Notre Dame, Australia, and the Raine Medical Research Foundation. The Raine Study Gen1- and Gen2-antenatal and perinatal data collections are funded by the Raine Medical Research Foundation. The Raine Study Gen2-10-year follow-up and the autism quotient data of the Gen2-20-year follow-up are funded by the NHMRC and the Raine Medical Research Foundation.
Funding:
The Raine Study is funded by project and program grants from the National Health and Medical Research Council of Australia (NHMRC, Canberra, Australia). Core management funding is provided by the Raine Medical Research Foundation, the Telethon Kids Institute, the University of Western Australia (UWA), the Women and Infants Research Foundation, Curtin University, Murdoch University, Edith Cowan University, and the University of Notre Dame Australia. Dr. Caleb Ing is supported by the Agency for Healthcare Research and Quality (AHRQ) under award number R01HS026493. The content is solely the responsibility of the authors and does not necessarily represent the official views of the AHRQ. Professor Andrew Whitehouse is supported by an Investigator Grant from the National Health and Medical Research Council (APP1173896). Professor Britta S. von Ungern-Sternberg is supported by the Stan Perron Charitable Foundation and through a National Health and Medical Research Council Investigator Grant (2009322).
Footnotes
Prior presentations: Preliminary findings of this work were presented at the April 2023 AUA/IARS/SOCCA Annual Meeting in Denver, CO.
Conflicts of Interest & Disclosures: None
Data Sharing:
The data that support the findings of this study are available from The Raine Study. Restrictions apply to the availability of these data, which were used under license for this study. Data are available with the permission of The Raine Study from https://rainestudy.org.au/information-for-researchers/available-data.
References
- 1.Butwick AJ, Wong CA, Guo N. Maternal Body Mass Index and Use of Labor Neuraxial Analgesia: A Population-based Retrospective Cohort Study. Anesthesiology. 2018;129(3):448–458. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Hawkins JL. Epidural analgesia for labor and delivery. N Engl J Med. 2010;362(16):1503–1510. [DOI] [PubMed] [Google Scholar]
- 3.Guglielminotti J, Landau R, Daw J, Friedman AM, Chihuri S, Li G. Use of Labor Neuraxial Analgesia for Vaginal Delivery and Severe Maternal Morbidity. JAMA Netw Open. 2022;5(2):e220137. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Qiu C, Lin JC, Shi JM, et al. Association Between Epidural Analgesia During Labor and Risk of Autism Spectrum Disorders in Offspring. JAMA Pediatr. 2020;174(12):1168–1175. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Wall-Wieler E, Bateman BT, Hanlon-Dearman A, Roos LL, Butwick AJ. Association of Epidural Labor Analgesia With Offspring Risk of Autism Spectrum Disorders. JAMA Pediatr. 2021;175(7):698–705. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Mikkelsen AP, Greiber IK, Scheller NM, Lidegaard Ø. Association of Labor Epidural Analgesia With Autism Spectrum Disorder in Children. Jama. 2021;326(12):1170–1177. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Ren T, Zhang J, Yu Y, et al. Association of labour epidural analgesia with neurodevelopmental disorders in offspring: a Danish population-based cohort study. Br J Anaesth. 2022;128(3):513–521. [DOI] [PubMed] [Google Scholar]
- 8.Murphy MSQ, Ducharme R, Hawken S, et al. Exposure to Intrapartum Epidural Analgesia and Risk of Autism Spectrum Disorder in Offspring. JAMA Netw Open. 2022;5(5):e2214273. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Straub L, Huybrechts KF, Mogun H, Bateman BT. Association of Neuraxial Labor Analgesia for Vaginal Childbirth With Risk of Autism Spectrum Disorder. JAMA Netw Open. 2021;4(12):e2140458. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Hegvik TA, Klungsøyr K, Kuja-Halkola R, et al. Labor epidural analgesia and subsequent risk of offspring autism spectrum disorder and attention-deficit/hyperactivity disorder: a cross-national cohort study of 4.5 million individuals and their siblings. Am J Obstet Gynecol. 2022. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Hanley GE, Bickford C, Ip A, et al. Association of Epidural Analgesia During Labor and Delivery With Autism Spectrum Disorder in Offspring. Jama. 2021;326(12):1178–1185. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Qiu C, Carter SA, Lin JC, et al. Association of Labor Epidural Analgesia, Oxytocin Exposure, and Risk of Autism Spectrum Disorders in Children. JAMA Netw Open. 2023;6(7):e2324630. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Kearns RJ, Shaw M, Gromski PS, Iliodromiti S, Lawlor DA, Nelson SM. Association of Epidural Analgesia in Women in Labor With Neonatal and Childhood Outcomes in a Population Cohort. JAMA Netw Open. 2021;4(10):e2131683. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Flick RP, Lee K, Hofer RE, et al. Neuraxial labor analgesia for vaginal delivery and its effects on childhood learning disabilities. Anesth Analg. 2011;112(6):1424–1431. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Sim MA, Shen L, Ti LK, et al. Association between maternal labour epidural analgesia and autistic traits in offspring. J Clin Anesth. 2023;89:111162. [DOI] [PubMed] [Google Scholar]
- 16.Shima M, Tokuda N, Hasunuma H, et al. Association of epidural analgesia during labor with neurodevelopment of children during the first three years: the Japan Environment and Children’s Study. Environ Health Prev Med. 2022;27:37. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.London EB. Categorical diagnosis: a fatal flaw for autism research? Trends Neurosci. 2014;37(12):683–686. [DOI] [PubMed] [Google Scholar]
- 18.Sagiv SK, Kalkbrenner AE, Bellinger DC. Of decrements and disorders: assessing impairments in neurodevelopment in prospective studies of environmental toxicant exposures. Environ Health. 2015;14:8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Rauh VA, Margolis AE. Research Review: Environmental exposures, neurodevelopment, and child mental health - new paradigms for the study of brain and behavioral effects. J Child Psychol Psychiatry. 2016;57(7):775–793. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Jansen S, Lopriore E, Naaktgeboren C, et al. Epidural-Related Fever and Maternal and Neonatal Morbidity: A Systematic Review and Meta-Analysis. Neonatology. 2020;117(3):259–270. [DOI] [PubMed] [Google Scholar]
- 21.Biederman J, Petty CR, Fried R, et al. Child behavior checklist clinical scales discriminate referred youth with autism spectrum disorder: a preliminary study. J Dev Behav Pediatr. 2010;31(6):485–490. [DOI] [PubMed] [Google Scholar]
- 22.Biederman J, DiSalvo M, Vaudreuil C, et al. The child behavior checklist can aid in characterizing suspected comorbid psychopathology in clinically referred youth with ADHD. J Psychiatr Res. 2021;138:477–484. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Havdahl KA, von Tetzchner S, Huerta M, Lord C, Bishop SL. Utility of the Child Behavior Checklist as a Screener for Autism Spectrum Disorder. Autism Res. 2016;9(1):33–42. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Chen WJ, Faraone SV, Biederman J, Tsuang MT. Diagnostic accuracy of the Child Behavior Checklist scales for attention-deficit hyperactivity disorder: a receiver-operating characteristic analysis. J Consult Clin Psychol. 1994;62(5):1017–1025. [DOI] [PubMed] [Google Scholar]
- 25.Arias AA, Rea MM, Adler EJ, Haendel AD, Van Hecke AV. Utilizing the Child Behavior Checklist (CBCL) as an Autism Spectrum Disorder Preliminary Screener and Outcome Measure for the PEERS® Intervention for Autistic Adolescents. J Autism Dev Disord. 2022;52(5):2061–2074. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.White SW, Eastwood PR, Straker LM, et al. The Raine study had no evidence of significant perinatal selection bias after two decades of follow up: a longitudinal pregnancy cohort study. BMC Pregnancy Childbirth. 2017;17(1):207. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Calder SD, Boyes M, Brennan-Jones CG, Whitehouse AJO, Robinson M, Hill E. Do parent-reported early indicators predict later developmental language disorder? A Raine Study investigation. International Journal of Language & Communication Disorders.n/a(n/a). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Guerrera S, Menghini D, Napoli E, Di Vara S, Valeri G, Vicari S. Assessment of Psychopathological Comorbidities in Children and Adolescents With Autism Spectrum Disorder Using the Child Behavior Checklist. Frontiers in Psychiatry. 2019;10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Broadbent J, Galic I, Stokes MA. Validation of autism spectrum quotient adult version in an Australian sample. Autism Res Treat. 2013;2013:984205. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Lu K Number of imputations needed to stabilize estimated treatment difference in longitudinal data analysis. Statistical Methods in Medical Research. 2017;26(2):674–690. [DOI] [PubMed] [Google Scholar]
- 31.Austin PC. The performance of different propensity score methods for estimating marginal hazard ratios. Stat Med. 2013;32(16):2837–2849. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Dong G, Mao L, Huang B, et al. The inverse-probability-of-censoring weighting (IPCW) adjusted win ratio statistic: an unbiased estimator in the presence of independent censoring. J Biopharm Stat. 2020;30(5):882–899. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Marshall A, Altman DG, Holder RL, Royston P. Combining estimates of interest in prognostic modelling studies after multiple imputation: current practice and guidelines. BMC Med Res Methodol. 2009;9:57. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Sun LS, Li G, Miller TL, et al. Association Between a Single General Anesthesia Exposure Before Age 36 Months and Neurocognitive Outcomes in Later Childhood. Jama. 2016;315(21):2312–2320. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Ing C, Landau R, DeStephano D, et al. Prenatal Exposure to General Anesthesia and Childhood Behavioral Deficit. Anesth Analg. 2021;133(3):595–605. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Richiardi L, Bellocco R, Zugna D. Mediation analysis in epidemiology: methods, interpretation and bias. Int J Epidemiol. 2013;42(5):1511–1519. [DOI] [PubMed] [Google Scholar]
- 37.VanderWeele TJ. A unification of mediation and interaction: a 4-way decomposition. Epidemiology. 2014;25(5):749–761. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Valeri L, Vanderweele TJ. Mediation analysis allowing for exposure-mediator interactions and causal interpretation: theoretical assumptions and implementation with SAS and SPSS macros. Psychol Methods. 2013;18(2):137–150. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Rosenbaum PR, Rubin DB. Constructing a Control Group Using Multivariate Matched Sampling Methods That Incorporate the Propensity Score. The American Statistician. 1985;39(1):33–38. [Google Scholar]
- 40.Teiling AK, Mohammed AK, Minor BG, Järbe TU, Hiltunen AJ, Archer T. Lack of effects of prenatal exposure to lidocaine on development of behavior in rats. Anesth Analg. 1987;66(6):533–541. [PubMed] [Google Scholar]
- 41.Smith RF, Kurkjian MF, Mattran KM, Kurtz SL. Behavioral effects of prenatal exposure to lidocaine in the rat: effects of dosage and of gestational age at administration. Neurotoxicol Teratol. 1989;11(4):395–403. [DOI] [PubMed] [Google Scholar]
- 42.Golub MS, Germann SL. Perinatal bupivacaine and infant behavior in rhesus monkeys. Neurotoxicol Teratol. 1998;20(1):29–41. [DOI] [PubMed] [Google Scholar]
- 43.Butwick AJ, Abrams DA, Wong CA. Epidural labour analgesia and autism spectrum disorder: is the current evidence sufficient to dismiss an association? Br J Anaesth. 2022;128(3):393–398. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Johnson ME, Uhl CB, Spittler KH, Wang H, Gores GJ. Mitochondrial injury and caspase activation by the local anesthetic lidocaine. Anesthesiology. 2004;101(5):1184–1194. [DOI] [PubMed] [Google Scholar]
- 45.Wang Z, Shen J, Wang J, et al. Lithium attenuates bupivacaine-induced neurotoxicity in vitro through phosphatidylinositol-3-kinase/threonine-serine protein kinase B- and extracellular signal-regulated kinase-dependent mechanisms. Neuroscience. 2012;206:190–200. [DOI] [PubMed] [Google Scholar]
- 46.Haller I, Hausott B, Tomaselli B, et al. Neurotoxicity of lidocaine involves specific activation of the p38 mitogen-activated protein kinase, but not extracellular signal-regulated or c-jun N-terminal kinases, and is mediated by arachidonic acid metabolites. Anesthesiology. 2006;105(5):1024–1033. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data that support the findings of this study are available from The Raine Study. Restrictions apply to the availability of these data, which were used under license for this study. Data are available with the permission of The Raine Study from https://rainestudy.org.au/information-for-researchers/available-data.