Evidence Mounts Refuting an Association Between Epidural Use and Neurodevelopmental Adverse Outcomes in Children

Neuraxial analgesia—more commonly described as epidural pain relief—is the most effective means of managing pain during labor.¹ In 2015, nearly 3 out of 4 US women used epidural pain relief during labor.² Despite its popularity, flawed or misleading research has linked epidural pain relief to several unfavorable perinatal and postnatal outcomes, such as an increased risk of cesarean delivery or long-term back pain.³ In response, anesthesia researchers have invested considerable time and effort to assuage these claims.³

A recent population-based study of 147 895 live births in Kaiser Permanente Southern California hospitals reported that epidural pain relief is associated with a 37% increased risk of autism spectrum disorder in offspring.⁴ However, residual confounding likely explains these findings, as a subsequent population-based study of 123 175 live births in Manitoba, Canada, did not confirm this association after accounting for a larger set of confounders.⁵ Therefore, there is a need for more research that addresses selection and confounding bias to clarify the presence and extent of an association. Because assigning women to receive epidural pain relief or placebo in a randomized trial is unethical and immoral, high-quality longitudinal studies remain the only viable option for evaluating potential links between epidural pain relief and adverse neurodevelopmental outcomes in human offspring.

A new population-based study⁶ from Scotland published in JAMA Network Open aims to shed more light on this topic. In this study, Kearns et al⁶ used data from 435 281 mother-offspring pairs in Scotland to explore whether associations exist between epidural pain relief and a number of childhood development indicators and adverse neonatal outcomes. In the regression analyses of childhood development indicators at age 2 years, null findings were reported across all models. However, findings from the associations between epidural pain relief and neonatal outcomes were less clear. In confounder-adjusted analyses, epidural pain relief was associated with a small increased risk of neonatal resuscitation (adjusted relative risk, 1.07; 95% CI, 1.03-1.11) and neonatal intensive care unit admission (adjusted relative risk, 1.14; 95% CI, 1.11-1.17) but was not associated with an increased risk of an offspring Apgar score less than 7 or 4 at 5 minutes. In additional mediation analysis, the researchers determined that delivery mode was a key mediating factor for the positive associations between epidural pain relief with neonatal resuscitation and neonatal intensive care admission.

Overall, these findings add further evidence to suggest that epidural pain relief does not influence the risk of adverse neurodevelopmental outcomes in offspring. However, this study does have some important limitations. First, 12 heterogeneous neonatal and neurodevelopmental outcomes were included in the analyses. As the authors did not correct for multiple group comparisons, statistically significant exposure-outcome associations may be due to chance (type 1 error). Second, residual confounding may explain the positive associations observed between epidural pain relief with neonatal resuscitation and neonatal intensive care admission, especially as provider-level factors, such as availability of trained neonatologists, and hospital-level factors, such as level of neonatal care, were not accounted for in the analysis. Furthermore, in the post hoc propensity score analyses examining confounding by indication, there was no information for the clinical indications for epidural pain relief. Third, there is substantial evidence disproving a link between epidural pain relief and an increased risk of cesarean delivery³; therefore, the appropriateness of the mediation analysis is questionable. Fourth, the validity, accuracy, and reliability of the measures used to classify the neurodevelopmental outcomes at age 2 years are not known.

So where do we go from here? Limiting the effect of bias is crucial when evaluating the presence, magnitude, and direction of an exposure-outcome relationship. Therefore, future observational studies examining epidural pain relief as the exposure and neonatal or neurodevelopmental measures as outcomes must ensure that appropriate patient-level, family-level, and hospital-level residual confounders are included and justified a priori. To aid researchers in this process, direct acyclic graphs can be useful for diagnosing and illustrating sources of bias, allowing for the selection of appropriate confounders for estimating risk estimates or causal effects from the observed data and providing a rationale for covariate selection.

For pregnant women and their maternal health care professionals, the effect of prior studies purporting a link between epidural pain relief and adverse neurodevelopment outcomes should not be underestimated. The media attention given to these studies, especially when published in high-impact journals, may cause apprehension among pregnant women and change how maternal health care professionals counsel pregnant women about epidural pain relief. It also diverts attention away from key maternal benefits afforded by epidural pain relief. Women who receive epidural pain relief have lower pain scores, are more satisfied, and are less likely to request additional pain relief compared with women who receive systemic opioid analgesia.¹ In addition, the presence of an indwelling epidural catheter enables the use of epidural anesthesia for an unplanned intrapartum cesarean delivery, thereby allowing women to be awake and comfortable during delivery. If an epidural catheter is not available, the likelihood of using general anesthesia may increase, which is associated with an increased risk of maternal and perinatal morbidity, such as failed intubation; intraoperative hemorrhage; postoperative pain, nausea, and vomiting; and fetal exposure to general anesthesia.

Despite its limitations, the study by Kearns et al⁶ provides further evidence that there is no meaningful association between epidural pain relief and adverse neurodevelopmental outcomes in the offspring. Although we may not yet be able to end research on this purported link, these findings provide further reassurance to pregnant women considering epidural pain relief during labor and their maternal and anesthesia health care professionals.

Published: October 28, 2021. doi:10.1001/jamanetworkopen.2021.31722

Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2021 Butwick AJ et al. JAMA Network Open.

Corresponding Author: Alexander J. Butwick, MS, Department of Anesthesiology, Perioperative, and Pain Medicine, Stanford University School of Medicine, 300 Pasteur Dr, Stanford, CA 94305 (ajbut@stanford.edu).

Conflict of Interest Disclosures: None reported.