The association between neuraxial anesthesia and the development of childhood asthma – a secondary analysis of the Newborn Epigenetics Study Cohort

Abstract

Objectives:

Childhood asthma is a common chronic illness that has been associated with mode of delivery. However, the effect of cesarean delivery alone does not fully account for the increased prevalence of childhood asthma. We tested the hypothesis that neuraxial anesthesia used for labor analgesia and cesarean delivery alters the risk of developing childhood asthma.

Methods:

Within the Newborn Epigenetics Study birth cohort, 196 mother and child pairs with entries in the electronic anesthesia records were included. From these records, data on maternal anesthesia type, duration of exposure, and drugs administered peripartum were abstracted and combined with questionnaire-derived prenatal risk factors and medical records and questionnaire-derived asthma diagnosis data in children. Logistic regression models were used to evaluate associations between type of anesthesia, duration of anesthesia, and the development of asthma in males and females.

Results:

We found that longer duration of epidural anesthesia was associated with a lower risk of asthma in male children (OR= 0.80; 95% CI = 0.66 to 0.95) for each hour of epidural exposure. Additionally, a unit increase in the composite dose of local anesthetics and opioid analgesics administered via the spinal route was associated with a lower risk of asthma in both male (OR= 0.59, 95% CI = 0.36 to 0.96) and female children (OR 0.26, 95% CI 0.09 to 0.82).

Conclusion:

Our data suggest that peripartum exposure to neuraxial anesthesia may reduce the risk of childhood asthma primarily in males. Larger human studies and model systems with longer follow-up are required to elucidate these findings.

Introduction

Neuraxial anesthesia is widely used to provide analgesia for labor and anesthesia for cesarean section. The most recent estimates demonstrate that 94% of all cesarean deliveries in the United States are performed using neuraxial anesthetic techniques.¹ Furthermore both the use and availability of epidural analgesia for labor has been steadily increasing in the United States.² Despite the widespread use of neuraxial anesthesia for delivery, there is a paucity of information on the effects of anesthetic techniques on long term childhood outcomes. Most prior anesthesia studies have been limited to investigating the effect of both neuraxial anesthesia and general anesthesia on neonatal acid base status, Apgar scores and neurologic and adaptive capacity scoring systems (NACS) as short term surrogate markers of fetal well-being.^3-8 Both lower early NACS and Apgar scores at one minute have been observed in neonates delivered with general anesthesia when compared with those delivered by neuraxial anesthesia.^3,7 However the potential adverse short term neonatal effects of general anesthesia maybe related to the urgency of the cesarean delivery and the choice of muscle relaxants used.^7,8 In contrast to general anesthesia, emerging evidence has demonstrated that neuraxial anesthesia may have immunomodulatory effects, modulate the stress response and oxidative stress response to labor and delivery that may affect fetal adaptation postnatally.^9-15 As a result neuraxial anesthesia may have longer lasting effects that may influence the development of common childhood diseases.

Chronic childhood diseases such as asthma remain a major public health problem. With an increasing prevalence globally asthma remains one of the leading causes of hospitalizations, emergency room visits and missed school days in children.^16,17 Gender differences also exist in the prevalence of asthma in childhood, with studies reporting an increased risk of wheezing and asthma in boys when compared with girls. The etiology of childhood asthma while multifactorial, may result from a combination of inherited genetic predisposition, stress response at delivery, immune sensitization, exogenous environmental exposure and effects on the neonatal microbiome.¹⁶ The role of peripartum factors remains unclear but likely play a role, as an association between mode of delivery and the development of childhood asthma has been demonstrated.^18-20 A recent meta-analysis demonstrated an increased risk of childhood asthma in children delivered by cesarean delivery when compared with those delivered by vaginal delivery.¹⁹ While these studies account for most known risk factors and environmental exposure, they do not account for the type of anesthesia used for delivery. This omission is despite different anesthetic techniques being used for cesarean delivery and vaginal delivery with different degrees of exposure of the fetus to these anesthetic agents.

Our working hypothesis is that neuraxial anesthesia used for labor analgesia and cesarean delivery alters the risk of developing childhood asthma in a sex-specific manner. For this study we leveraged resources of a cohort of children, whose mothers received neuraxial anesthesia for vaginal delivery or cesarean section and have been followed longitudinally from birth until age seven years.

Patients and Methods

Study participants

This retrospective cohort study was approved by the Duke University Medical Center Institutional Review Board and reported in line with recommendations from the STROBE statement. We included pregnant women and their offspring initially enrolled as part of the Newborn Epigenetics STudy (NEST), a pre-birth cohort study which recruited pregnant women between 2005 and 2011. These pregnant women received prenatal care at six prenatal clinics in Durham Country, North Carolina and were enrolled using a recruitment protocol whose details have been described previously.^21,22 Briefly we included all women who were > 18 years of age, english or spanish speaking, intending to use Duke University Medical Center or Duke Regional Hospital for obstetric care for the index pregnancy. Excluded were those intending to relinquish custody of the index child, or those with confirmed human immunodeficiency virus (HIV) infection. A total of 3,690 eligible women were approached and 2,595 agreed to participate in the initial NEST. Of those patients 1359 were excluded due to missing data and these patients were similar to those with complete data except for differences in birth weight, gestational age and maternal race. This analysis is however limited to the 196 mothers and offspring (104 mother-boy pairs and 92 mother-girls pairs) who had complete data on anesthesia exposure obtained from the Duke University Hospital Electronic Innovian®` Anesthesia record, after excluding patients who needed general anesthesia (sample size n=8) and delivered extremely preterm (defined as less than 28 weeks of gestational age, n=28) (Figure. 1).

Figure 1.

Flow chart of participants included in this study

Data collection

Anesthesia exposure

The duration of anesthesia exposure was characterized from the time of placement of neuraxial anesthesia (spinal, epidural or combined spinal epidural) to the time of delivery of the baby obtained from trigger notes entered in the electronic anesthesia record. We specifically extracted data on doses of local anesthetics, opioids, antiemetics, antibiotics and vasopressors administered to the mother and their routes of administration. We included antiemetics and vasopressors since these are commonly used to treat nausea and vomiting and hypotension which are common adverse effects associated with neuraxial anesthesia.^23,24 Data on specific drugs administered were only available for those patients who had an anesthetic intervention during labor and delivery.

Assessment of asthma and confounders

Anesthesia data were merged with data collected from the initial NEST cohort. This included maternal sociodemographic and obstetric data such as age, race/ethnicity, prepreg Nancy body mass index (BMI), educational status, cigarette smoking status, asthma status, gravidity, parity and mode of delivery. For the offspring we collected data on sex, birthweight, duration of breastfeeding and childhood asthma diagnosis primarily obtained from questionnaires. To determine the diagnosis of childhood asthma, we developed an algorithm using a combination of billing and prescription records, and two questions from the follow-up questionnaire., i.e., 1) “What was the outcome of your child’s doctor visits? Normal/Concerns, and if there are concerns, specify?” and 2) “Was the child diagnosed with any condition by his/her doctor? Yes/No.”For those who reported a current asthma diagnosis we then verified this diagnosis with an existing prescription for asthma medications. Supporting the validity of this algorithm, the accuracy compared with a full medical records review was >99%. Furthermore, the prevalence of asthma in our cohort was consistent with the known prevalence of asthma in children aged 5-9 years.²⁵

Childhood asthma was analyzed as a dichotomous variable. In the analysis, confounders included infant sex, maternal smoking during pregnancy (yes/no), pre-pregnancy BMI, maternal weight gain during pregnancy (inadequate/adequate/excessive), maternal educational status (graduated high school, college or college graduate), maternal ethnicity (African American, White or Hispanic), birth weight (≤4kg or >4kg), breastfeeding more than 3 months, a history of maternal asthma, gestational age and mode of delivery (cesarean delivery or vaginal delivery).

Statistical analysis

We performed a univariate logistic regression to evaluate the association between childhood asthma and each maternal characteristic. Multivariable logistic regression models were used to evaluate associations between childhood asthma and anesthesia exposure in males and in females separately, adjusting for confounders. The variables of interest included the dosage and the duration of administration. Drugs were grouped into six categories (Table 2) based on the types of drugs and route of administration, and we considered composite measures to reduce dimensionality and to assess their effects in aggregate. The six categories included: 1) epidural-administered local anesthetics and opioid analgesics, 2) spinal-administered local anesthetics and opioid analgesics, 3) intravenously administered fentanyl, 4) antiemetics (ondansetron and metoclopramide), 5) antibiotics (cefazolin and clindamycin), and 6) phenylephrine. The formation of these categories is described below.

Table 2.

List of the median (1^st quartile, 3^rd quartile) doses of the drugs administered and the duration of exposure to epidural and spinal anesthetics for delivery for male and female children.

		Median dosage [interquartile range]
Indices	Anesthetics	Male (N=104)				Female (N=92)
		N1	Normal (N=72)	N	Asthma (N=32)	N	Normal (N=76)	N	Asthma (N=16)
Epidural administration	Bupivacaine (mg)	2	13.75 [12.50, 15.00]	2	32.00 [20.00, 50.00]	7	18.75 [18.00, 75.00]	0	NA [NA, NA]
	Fentanyl (mg)	16	0.10 [0.04, 0.15]	5	0.10 [0.05, 0.25]	14	0.05 [0.03, 0.25]	1	0.05 [0.05, 0.05]
	Lidocaine (mg)	13	200.00 [100.00, 600.00]	4	300.00 [260.00, 400.00]	6	272.5 [100.00, 500.00]	4	300.00 [100.00, 400.00]
	Local anesthesia infusion (ml)2	55	95 [8, 528]	19	79 [10, 375]	52	67 [20, 530]	9	71 [15, 294]
	Morphine (mg)	3	3 [3, 3]	1	3.00 [3.00, 3.00]	1	3 [3, 3]	1	3 [3, 3]
Spinal administration	Bupivacaine (mg)	13	12.00 [10.00, 13.50]	9	12.00 [10.50, 12.00]	16	12.00 [9.00, 12.00]	3	12.00 [11.25, 12.00]
	Fentanyl (mg)	15	0.02 [0.01, 0.03)	6	0.02 [0.01, 0.02]	15	0.02 [0.01, 0.02]	3	0.02 [0.01, 0.02]
	Morphine (mg)	10	0.15 [0.10, 0.20]	6	0.15 [0.10, 0.15]	14	0.15 [0.10, 0.15]	1	0.15 [0.15, 0.15]
Intravenous administered	Fentanyl (mg)	11	0.05 [0.02, 0.24]	5	0.05 [0.05, 0.25]	8	0.05 [0.02, 0.15]	3	0.02 [0.02, 0.20]
Fentanyl
Anti-emetics	Metoclopramide (mg)	6	10 [10, 10]	2	10 [10, 10]	2	10 [10, 10]	0	NA [NA, NA]
	Ondansetron (mg)	15	4 [4, 8]	9	4 [4, 4]	19	4 [4, 4]	4	4 [4, 4]
Antibiotics	Cefazolin (mg)	22	2,000 [1,000, 2,000]	12	2000 [1000, 2000]	22	2,000 [1,000, 2,000]	5	2,000 [2,000, 2,000]
	Clindamycin (mg)	1	900 [900, 900]	0	NA [NA, NA]	3	900 [900, 1,700]	1	600 [600, 600]
Phenylephrine	Phenylephrine (mg)	21	0.80 [0.05, 2.50]	13	0.60 [0.10, 3.05]	20	1.26 [0.05, 8.00]	7	0.40 [0.10, 3.05]
Epidural duration	Epidural duration (hrs)	58	8.62 [0.37, 33.10]	22	5.35 [0.23, 16.23]	59	4.9 [0.33, 27.92]	13	3.18 [0.58, 17.05]
Spinal duration	Spinal duration (hrs)	21	0.38 [0.18, 5.42]	14	0.49 [0.37, 5.15]	24	0.57 [0.33, 10.13]	6	0.57 [0.27, 0.85]

^1:N is the sample size.

^2:0.1-0.125% Bupivacaine with 2 μg/ml fentanyl

First, based on the drug functions, we considered major categories as local anesthetics and opioid analgesics, antiemetics (ondansetron and metoclopramide), antibiotics (cefazolin and clindamycin), and phenylephrine. Several other drugs not considered in this analysis were administered to a few subjects (i.e., 14 individuals in total), including diphenhydramine (2 individuals), ephedrine (1 individual), epinephrine (2 individuals), ketamine (3 individuals), and midazolam (8 individuals). After excluding the 14 individuals, we had 196 individuals who did not receive the above drugs and only had those drugs listed in Table 2 during delivery. Since local anesthetics and opioid analgesics maybe administered via different routes, and the impact of this administration may be fundamentally different, we further considered separate groups for local anesthetics and opioid analgesics: 1) epidural-administrated local anesthetics and opioid analgesics, 2) spinal-administrated local anesthetics and opioid analgesics, and 3) intravenously-administered fentanyl (listed in Table 2).

Finally, to empirically assess the appropriateness of the six drug categories, we conducted factor analysis on the 14 drugs listed in Table 2 using the principal components method with oblique promax rotation. Factor analysis uses eigenvalue decomposition to transform the drugs of similar types into factors, which can then facilitate the understanding of how the drugs are similar or different. As shown in Figure S1, the results of factor analysis are generally in agreement with the six categories we defined: epidural-administered local anesthetics and opioid analgesics (i.e., Factors 2 and 5 in Figure S1); spinal-administered local anesthetics and opioid analgesics (i.e., Factor 1); intravenously-administered fentanyl (i.e., Factor 3); antiemetics (Factor 6); antibiotics (Factor 4), and phenylephrine (Factor 7).

For each of the six drug categories that contained more than one drug, we computed the composite dosage by summing the standardized values of the listed drugs in the category. Standardization before summing facilitates pooling all drugs on a comparable scale. We included the drug dosages via the composite measures; we also included the duration of epidural administration and the duration of spinal administration as explanatory variables in the regression model. We performed no a priori sample size calculation, however for our final sample size of 196 mother and child pairs a posthoc power analysis was performed to ensure we had adequate power (>80%) for the majority of variables of interest in our model. All analyses were conducted using SAS (Version 9.4, The SAS institute, Cary, NC).

Results

Table 1 shows the univariate analysis for the maternal characteristics for the mother and child pairs, stratified by sex for the development of childhood asthma. Overall males comprised 53.06% of the sample, and African Americans comprised 50.51%. Maternal asthma was reported by 14.29% of participants, smoking by 28.57%, maternal obesity before pregnancy by 31.63% and 37.24% of offspring were delivered by Cesarean Section. These factors were more frequently reported by women with asthmatic children and were accounted for in the model.

Table 1.

Univariate analysis of maternal characteristics of the study participants.

Category		Male (N=104)			Female (N=92)
		Normal (N=72) (%)	Asthma (N=32) (%)	p- value	Normal (N=76) (%)	Asthma (N=16) (%)	p- value
Maternal age at delivery (years)		29 [24, 33]	27 [23,34]	0.579	29 [23, 33]	31 [22, 36]	0.609
Gestational	Less than adequate	15.28	31.25	0.015	19.74	25.00	0.508
weight gain	Adequate	33.33	12.50		18.42	12.50
	Excessive	51.39	56.25	0.080	61.84	62.50	0.632
Gestational age (weeks)		39 [37, 40]	39 [38,40]	0.842	39 [39, 40]	38 [37, 39]	0.010
Parity	0	79.17	78.12		72.37	81.25
	1	9.72	3.12	0.306	14.47	12.50	0.751
	2	6.94	6.25	0.916	1.32	0	0.997
	3	2.78	0	0.993	5.26	6.25	0.961
	4	1.39	3.12	0.566	2.63	0	0.995
	5	0	0	NA	2.63	0	0.995
	>5	0	9.38	0.990	1.32	0	0.997
Education	Less than high school	23.61	31.25	0.414	22.37	12.50	0.383
	High school graduate/GED/College	76.39	68.75		77.63	87.50
Smoking status	Non-smoker	77.78	65.62		67.11	75.00
	Smoker	22.22	34.38	0.195	32.89	25.00	0.538
Maternal asthma	Yes	15.28	12.50	0.999	13.16	18.75	1.000
	No	84.72	87.50		86.84	81.25
Maternal obesity1	Yes	26.39	37.50	0.255	31.58	43.75	0.352
	No	73.61	62.50		68.42	56.25
Race/Ethnicity	African American	37.50	56.25		56.58	68.75
	Hispanic	25.00	18.75	0.217	13.16	6.25	0.394
	White	37.50	25.00	0.108	30.26	25.00	0.546
Breastfeeding	Yes	59.72	53.12	0.530	56.58	56.25	0.981
	No	40.28	46.88		43.42	43.75
Delivery route	C-section	34.72	46.88	0.242	32.89	50.00	0.200
	Vaginal delivery	65.28	53.12		67.11	50.00
Birth weight1	<4 Kg	86.11	93.75		93.42	93.75
	≥4 Kg	13.89	6.25	0.829	6.58	6.25	0.029

For categorical variables levels with empty p-values are set as reference in univariate analysis. Data are median ( 1st quartile, 3^rd quartile) or %.

¹Maternal BMI>30 kg/m²

Table 2 shows the quartile dosages of the drugs, duration of spinal and epidural anesthesia grouped by categories, administered during labor and delivery and stratified by asthma status. These tables provide the distributions of potential covariates in male and female children.

Table 3 summarizes the association between anesthesia drugs, type of anesthesia and duration of anesthesia exposure and the development of childhood asthma adjusted for confounders as odd ratios (ORs) and 95% confidence intervals (CI). Interestingly we found that epidural duration was associated with a lower risk of asthma in male children and for each hour of epidural analgesia exposure, the OR of developing asthma in male children was 0.80 (95% CI = 0.66 to 0.95). A unit increase in the composite dose of local anesthetics and opioid analgesics administered via the spinal route were also associated with an even lower risk of asthma in male (OR= 0.59, 95% CI = 0.36 to 0.96) and in female children (OR 0.26, 95% CI 0.09 to 0.82). Antiemetic use was also associated with a lower risk of asthma in female children OR=0.26, 95%CI=0.06 to 0.63). However, an increasing dose of phenylephrine administration was only marginally associated with an increased risk of asthma in male children only (OR 1.15, 95% CI= 1.00 to 1.33).

Table 3.

Odds ratios for the associations between duration of exposure and drugs administered during delivery and the diagnosis of childhood asthma in males and female children. Model adjusted for maternal asthma, maternal smoking status, maternal obesity, maternal weight gain during pregnancy, ethnicity, maternal educational attainment, birth weight, breastfeeding, mode of delivery and gestational age.

Characteristic	Childhood asthma
	Male (N=104)			Female (N=92)
	Odds Ratio2	95% CI	P- value	Odds Ratio	95% CI	P- value
Epidural-administered anesthetics and opioid analgesics1	1.17	[0.87, 1.57]	0.292	0.80	[0.49, 1.31]	0.368
Spinal-administered anesthetics and opioid analgesics1	0.59	[0.36, 0.96]	0.035	0.26	[0.09, 0.82]	0.021
Anti-emetics	1.28	[0.85, 1.91]	0.236	0.19	[0.06, 0.63]	0.007
Antibiotics	0.75	[0.30, 1.84]	0.523	0.75	[0.26, 2.11]	0.580
Intravenous-administered fentanyl2	0.71	[0.17, 2.97]	0.643	9.92	[0.95, 103.94]	0.056
Phenylephrine2	1.15	[1.00, 1.33]	0.051	1.01	[0.90, 1.15]	0.818
Epidural duration	0.80	[0.66, 0.95]	0.013	1.00	[0.83, 1.20]	0.974
Spinal duration	1.43	[0.75, 2.73]	0.285	0.92	[0.20, 4.27]	0.919

^1:Epidural and spinal anesthetics obtained by the sum of the standardized values of their members.

^2:Unit of measurement is 0.1mg

Discussion

In this study we characterized anesthesia administration during labor and delivery and evaluated exposure in relation to the development of childhood asthma adjusted for confounders. A key finding of this study was that whereas epidural anesthesia itself was not associated with a lower risk of asthma; the longer duration was associated with a significant reduction the incidence of asthma in male children. Even though increasing doses of spinal local anesthetics and opioids were significantly associated with a lower risk of asthma in both male and female children, the actual difference in the doses of local anesthetics and opioids used for spinal anesthesia for delivery do not vary significantly in clinical practice. Additionally, we demonstrated that 2 common drugs used in patients with neuraxial anesthesia may have opposing effects with antiemetics associated with a reduced risk of asthma in female children only and phenylephrine administration marginally associated with an increase in the risk of asthma in male children only.

Neuraxial anesthesia is widely used for labor analgesia and cesarean delivery in the developed world with epidural analgesia used primarily for labor analgesia and for cesarean section in laboring patients. Spinal anesthesia in contrast is used primarily as the anesthetic technique for cesarean delivery for both scheduled and unscheduled cases.² Our findings suggests that both types of neuraxial anesthetic techniques may in fact have long-term benefits on the development of childhood asthma. Labor and delivery represent a period of significant physical and psychological stress resulting from pain and fear in the perinatal period.^26-28 Several epidemiological studies have demonstrated a link between maternal stress and the development of asthma phenotypes in offspring.^29-32 The observed reduction in the risk of childhood asthma with increased duration of exposure to epidural anesthesia to a lesser extent, increasing doses of spinal anesthesia and opioids, maybe due to an attenuation of maternal or fetal stress response to labor and delivery by mechanisms that still remain unclear. It also remains unclear whether these benefits are maintained throughout the lifecourse.

The perinatal psycho-social stress induced by labor pain or the anticipation of labor pains and delivery leads to the activation of the hypothalamo-pituitary axis (HPA) which is involved in the inflammatory response and can have effects on innate and adaptive immunity in the offspring.^33-35 The resulting hormones and neuropeptides released as a result of this stress response, in turn are involved in airway responses and inflammatory changes that modulate risk for the development of reactive airways disease. ^33-35 Furthermore, genetic and epigenetic regulatory mechanisms triggered by peripartum stress may affect glucocorticoid receptor responsiveness and lead to changes in neuroendocrine, oxidative stress and immunomodulatory mechanisms in utero, that may increase susceptibility to asthma in childhood.^33-35 Together, modulation of parturition-related physiological changes related to maternal stress may mediate the inverse association observed between neuraxial anesthesia administration and asthma.

Evidence exists that maternal stress response may be partially attenuated by neuraxial anesthesia. Neuraxial anesthesia reduces maternal beta-endorphin ACTH and cortisol surge that usually accompanies labor.^12,14,36 Epidural analgesia also reduces maternal epinephrine levels with less of an effect observed on norepinephrine levels.¹³ However, the effect of neuraxial anesthesia on the fetal stress response is less clear. While epidural analgesia appears not to affect fetal catecholamine, ACTH and cortisol levels, maternal and fetal cortisol levels show a strong correlation, suggesting that maternal cortisol may be transferred transplacentally to the fetus, such that maternal stress levels on the fetus may be indirect.^14,36 The studies investigating the effects of neuraxial anesthesia on both maternal and fetal stress response are however limited by small sample sizes and because most of these studies are older, they are based on higher concentrations of local anesthetic which do not reflect current obstetric anesthesia practice where low concentration epidurals administered with opioids are used.

An immunomodulatory mechanism may also partly explain the association between a reduction in the development of childhood asthma and the use of neuraxial anesthesia. It has previously been demonstrated that the use of epidural anesthesia for labor and delivery was associated with lower levels of allergen specific IgE and total IgE levels in children at aged two years when compared with mothers who delivered without epidural analgesia.¹⁰ It is unclear if these effects are due primarily to effects on the stress response or direct immunomodulatory effects of local anesthetics since epidurally-administered local anesthetics may be transferred to the fetus transplacentally and and have been detected at low levels in breast milk.³⁷-³⁹ Local anesthetics both in vivo and in vitro and are also known to suppress cytokine production and may inhibit dendritic cell-mediated differentiation of Th1 cells with little effect on the differentiation of Th2 and regulatory T cells. ^40,41

Neuraxial anesthesia may also reduce markers of oxidative stress in the placenta and so may have other long term outcomes on the offspring which may have not been evident from studies investigating maternal and fetal catecholamines and steroid hormone levels.¹¹ Reasons for male-specific protective effects of epidural anesthesia on asthma was surprising. This sex specific response also plays a role in the development of asthma as highlighted by another study that demonstrated an association between prenatal maternal stress and the development of wheezing in male offspring in a Mexican cohort.³² This vulnerability of male offspring to stress suggests increased vulnerability to prenatal stress mediated in part by placental and placental glucocorticoid response and increased susceptibility to stress induced oxidation in utero making them more responsive to the beneficial effects of neuraxial anesthesia.^42,43 Nonetheless, with the growing use of neuraxial anesthesia for delivery, and asthma disproportionately affecting male children, but not female children, these associations, will require reconciliation in larger studies and in model systems to elucidate the possible mechanisms involved.

The effect of antiemetics use by mothers on asthma development in female offspring was a surprising finding. Interestingly the most common antiemetic administered in our cohort was ondansetron – a 5 hydroxytryptamine (5HT₃) or serotonin receptor antagonist. Serotonin plays a role in the development of airway hyperresponsiveness caused by early life exposure to allergens through cholinergic-mediated airway smooth muscle contraction.^44,45 Higher levels of serotonin are observed in patients with asthma and are negatively correlated with lung function.⁴⁶ Similarly pharmacological interventions that reduce serotonin levels may improve asthma severity and improve lung function.⁴⁷ Given the widespread use of antiemetics in pregnancy this potential association warrants further investigation.

The marginal effect of phenylephrine on asthma development in male children also further highlights the susceptibility of the male children to the development of asthma. The biological plausibility for this effect is not entirely clear but could represent an indirect effect where an increasing dose of phenylephrine is a surrogate marker for women who experienced more severe hypotension associated with spinal anesthesia. Hypotension induced by spinal anesthesia has been associated with low Apgar scores and fetal acidosis but longer term effects are less clear.⁴⁸-⁵⁰ Phenylephrine, an α1 agonist, also undergoes placental transfer and may have direct effects on bronchial smooth muscle tone.^51,52 Larger studies will be needed to determine effect of phenylephrine on asthma development especially in light of the fact that a recently released international consensus statement recommended α-agonists as the most appropriate agent for both the treatment and prevention of hypotension associated with spinal anesthesia.⁵⁰

A major strength of our study is the longitudinal design and a rigorous ascertainment of neuraxial anesthesia, associated drugs and asthma status. However, our findings should also be interpreted in the context of methodological limitations. We did not include patients who received no neuraxial anesthesia in our study population. This is due to the high utilization of neuraxial anesthesia for labor analgesia and delivery in the patient population recruited for the initial study (80% epidural rate for labor analgesia and 95% neuraxial anesthesia rate for anesthesia for cesarean delivery). Furthermore, patients who do not receive neuraxial anesthesia particularly for labor are not included in the electronic anesthesia database making collection of data on types of drugs and duration of exposure during the peripartum period less reliable. This study represents a secondary analysis of the primary NEST study which was conducted with very different aims and anesthetic information was not prospectively collected. A large proportion of the initial NEST cohort was also excluded due to missing breastfeeding data leading to differences in gestational age, birth weight and race between the cohort with and without missing data. Including these factors in our final model to reduce the risk of bias did not alter the findings of our study. Additionally, we did not collect data on maternal anxiety, fear or other psycho-social stressors during the peripartum period and the effects of these factors could not be assessed in final model the development of asthma in offspring. Furthermore, although we were adequately powered to examine sex-specific effects of neuraxial anesthesia on common childhood outcomes, we were under-powered to evaluate interactions within specific sex strata. Our posthoc power analysis also demonstrated that our study was specifically underpowered to determine the association between spinal anesthesia duration, epidural anesthesia duration and phenylephrine administration with the development of childhood asthma in female children only. Due to the small sample size we limited the variables and confounders included in the model and additional variables such as the urgency of the cesarean section which might determine the mode of neuraxial anesthesia used were not included. We however plan to explore the role of urgency of the cesarean delivery on our primary outcome in future larger studies.

Conclusions

This study present early evidence for associations between neuraxial anesthesia and a lower risk of asthma - these associations may be stronger in males. Clinically, early placement of neuraxial anesthesia in women in labor who experience elevated levels of peripartum psychosocial stress could be a useful intervention for reducing the risk of childhood asthma. If replicated in larger studies with longer follow-up, investigating mechanisms could inform intervention strategies.