Cumulative maternal lifetime stress & child asthma: effect modification by BMI

Abstract

Background:

Investigations of maternal psychosocial stress and child asthma have produced mixed findings, which may reflect inconsistent consideration of modifying factors.

Objective:

To examine associations between maternal lifetime stress and child asthma, and to assess effect modification by maternal pre-pregnancy body mass index and race/ethnicity in a prenatal cohort of mother-child dyads.

Methods:

Maternal lifetime stress was assessed using the Life Stressor Checklist-Revised, administered during pregnancy and child asthma was ascertained by parent-report in study follow-up visits. In the overall group and stratified by race/ethnicity, we used multivariable logistic regression and varying coefficient modeling to investigate the association between maternal stress and child asthma, assessing for effect modification by pre-pregnancy body mass index.

Results:

Women were predominately Black (Black/Hispanic-Black 44.5%) or non-Black Hispanic (37.6%), with elevated pre-pregnancy body mass index (25.1% overweight, 29.8% obese); 17% of children had asthma. Higher maternal stress was associated with increased relative odds of child asthma only in dyads with women in the obese (≥30 kilograms/meters squared) category (odds ratio 1.84, 95% confidence interval 1.27–2.67). Varying coefficient models demonstrated stronger positive associations between increased maternal lifetime stress and child asthma in women with higher pre-pregnancy body mass index; the strongest association was observed in the Black group.

Conclusion:

Maternal pre-pregnancy body mass index modified the association between maternal lifetime stress and child asthma. These findings underscore the need to consider complex interactions to fully elucidate intergenerational stress effects on early childhood asthma.

INTRODUCTION

Asthma is a leading childhood disease (1) associated with significant morbidity, healthcare utilization and reduced quality of life (2). Approximately 8% of children in the United States (U.S.) have asthma, with highest rates in historically minoritized racial/ethnic groups and those in under-resourced backgrounds.

Programming of child respiratory health begins in utero with disease risk influenced by a range of environmental factors, including maternal psychosocial stress (3). Stress-related neuropsychobiologic changes can alter immune regulation and oxidative stress systems which serve as mechanistic links between maternal lifetime stress and child asthma (3). Stress research has evolved towards a life course framework that considers psychosocial stressors experienced over a woman’s lifetime with evidence that traumatic stressors can be particularly robust potentiators of inflammatory processes and oxidative stress carried into pregnancy to impact fetal programming (4–6). . Lifetime exposure to traumatic stressors, even when remote in time, impact stress-related programming of respiratory disease starting prenatally (5, 7).

Excess maternal pre-pregnancy weight and consequent persistent low-grade inflammation, oxidative stress, and metabolic dysregulation (8) also alters fetal development (9, 10). Studies link pre-pregnancy body mass index (BMI) with child asthma, rhinitis and dermatitis (11), even after adjusting for child BMI. While being overweight or obese are related to maternal history of traumatic experiences and to chronic stress more generally (2,9–12)(12–14), there are a range of demographic and environmental influences at both the community- and individual-level that contribute to obesity across a woman’s lifespan (15–18). Due to factors such as structural and interpersonal racism, lower-income, BILPOC (Black, Indigenous, LatinX, People of Color) women may disproportionately experience traumatic events(19, 20) or live in historically segregated communities shaped by discriminatory practices, such as redlining and disinvestment, that have contributed to historical and contemporary stress exposures, as well as other obesogenic factors (21–23), and therefore are particularly at risk for obesity (24). Acknowledging that race is a sociopolitical construct that serves as a proxy for systemic factors that contribute to disparate experiences and exposures that may have intergenerational effects(7, 25), studies are needed to disentangle the complex relationships among chronic maternal stress, pre-pregnancy obesity, and race/ethnicity in relation to asthma risk in the next generation.

In a racially and ethnically diverse prenatal cohort, we investigated the association between maternal lifetime stress and child asthma and assessed whether associations were modified by maternal pre-pregnancy BMI and race/ethnicity. As both chronic stress and obesity contribute independently to an altered inflammatory and oxidative state in pregnancy that may result from similar, but not completely overlapping pathways, we hypothesized that maternal pre-pregnancy obesity would modify the effects of lifetime maternal stress on child. In addition, given overlapping evidence showing that obesity-related inflammation during pregnancy impacts children’s outcomes in an ethnic-specific manner (26), we hypothesized that race/ethnicity would further modify these associations. Some of these analyses have been reported in abstract form (27).

METHODS

Study Participants

Analyses included mother-child dyads enrolled in the PRogramming of Intergenerational Stress Mechanisms (PRISM) pregnancy cohort designed to prospectively examine associations of maternal psychosocial stress and other environmental exposures with child health and development. PRISM recruited 1,003 women receiving prenatal care from the Beth Israel Deaconess Medical Center and East Boston Neighborhood Health Center in Boston, MA (March 2011 – December 2013) and Mount Sinai Hospital in New York City, NY (April 2013 – July 2020) at 23.2±9.1 weeks gestation. Eligible women were English- or Spanish-speaking, ≥18 years of age, and had a singleton pregnancy. Exclusions included report of ≥7 alcoholic drinks/week prior to pregnancy recognition or any alcohol or drug use after pregnancy recognition, human immunodeficiency virus positive status, or a major congenital or genetic abnormality that could impact participation. Among these, 793 had complete data on maternal stress, pre-pregnancy BMI, race/ethnicity, and child asthma. Participants with self-reported “other” race/ethnicity (n=33) were excluded from the analysis due to limited sample size to examine race-specific effects, yielding a final analytic sample of N=760. Characteristics of those enrolled and those included in analyses were similar (p-values>0.20; Online Supplement, Table E1). Procedures were approved by the relevant institutions’ human studies committees and written informed consent was obtained from women in their preferred language.

Measures

Maternal Lifetime Stress Exposure

Research staff administered the 30-item Life Stressor Checklist-Revised (LSC-R) (28) assessing experienced or witnessed stressful life events (e.g. serious disasters, accidents, financial stressors, violence/robbery, sexual harassment or abuse, physical or mental health conditions, incarceration, emotional or physical abuse or neglect, divorce or separation) over the life course. To consider impact during pregnancy, we used the LSC-R weighted score of events, which considered the negative impact of each endorsed item by incorporating responses to the prompt, “How much has this affected your life in the past year?”. The weighted score of endorsed events was derived by summing the severity rating (1 [“not at all”] to 5[“extremely”]) from all endorsed events (range of potential scores, 0 to 150). The LSC-R has established test-retest reliability and has been validated in diverse populations (29, 30).

Childhood Asthma Outcome

Women were asked about their child’s respiratory health at approximately 4-month intervals through telephone interviews and face-to-face follow-up for the first 30 months of life, then annually. Child asthma status was based on affirmative maternal report of any of the following: “Has your child ever had asthma (since birth)?”; “Has a doctor or other healthcare provider ever said that your baby/child had asthma?”; “Does your baby/child currently take medicine for his/her asthma?”; and/or “Has your baby/child been hospitalized overnight for the asthma (since birth)?”.

Other Covariates

Women reported their birthdate, race, ethnicity, highest level of educational attainment and pre-pregnancy height and weight at enrollment. Child sex was obtained from the medical record at birth. Maternal education was categorized as low (high school or less; ≤12 years) vs. high (some college or above; >12 years). Maternal race and ethnicity were categorized as Black (Black and Black-Hispanic), Hispanic (non-Black Hispanic), and White (non-Hispanic White). Pre-pregnancy BMI was calculated as weight in kilograms (kg) divided by height in meters (m) squared. For descriptive purposes we categorized pre-pregnancy BMI into healthy/underweight (<25 kg/m²), Overweight (25 to <30 kg/m²), Class I obesity (30 to <35 kg/m²), Class 2 obesity (35 to <40 kg/m²) and Class 3 obesity (“severe” obesity; ≥40 kg/m²) based on U.S. Centers for Disease Control and Prevention guidelines (31). Class 1, Class 2, and Class 3 obesity were collapsed into one obesity group for statistical modeling due to sample size. Women who reported a history of atopy (asthma, hay fever, or eczema) were characterized as having active atopy during pregnancy if they 1) answered yes to “Do you still have asthma?” or “Do you still have hay fever?” or “Do you still have eczema?” at enrollment, or 2) reported at least one of the following: emergency visit or hospitalization for asthma or asthma symptoms (wheezing, shortness of breath, cough, chest tightness, nighttime awakenings due to asthma) within the past year and/or 3) reported use of medications for asthma, hay fever, or eczema during pregnancy or within one year prior to enrollment.

Data Analysis

We derived the distribution of LSC-R score, child asthma, and covariates in the sample overall and by maternal race/ethnicity. We used Wilcoxon rank sum test and χ² test as appropriate to compare the difference in covariates by maternal race/ethnicity. We also examined Spearman correlations between LSC-R score and maternal pre-pregnancy BMI (hereafter maternal BMI). For primary analyses, we examined the association between maternal lifetime stress (continuous, weighted LSC-R) and child asthma using multivariable logistic regression adjusting for child sex, maternal age, maternal race/ethnicity (overall sample only), maternal active atopy, and maternal education in the overall sample and subsequently stratified by maternal BMI and race/ethnicity. In addition, we fit the logistic regression models with an interaction term between LSC-R score and maternal BMI categories, both in the overall sample and when stratified by maternal race/ethnicity. To ascertain the distinct impacts across BMI groups, we further employed a multiple degrees of freedom chi-square test for each interaction model. This method is pivotal for evaluating whether the relationship between LSC-R and child asthma varied significantly among the different BMI categories. Next, we used varying coefficient models to assess whether maternal BMI (treated as a continuous variable) modified the association between maternal lifetime stress and child asthma, as participants with low or high BMI could represent susceptible populations, and interactions might be non-linear. Varying coefficient models are similar to logistic regression models but do not assume a constant effect of the exposure (LSC-R) across different maternal BMI levels. Thus, maternal BMI was entered into the model as an unspecified, potentially non-linear, smooth function estimated from the data both when assessing the main effect of maternal BMI and the maternal BMI x LSC-R interaction. The resulting model allowed the association of maternal lifetime stress and child asthma to vary as a potentially nonlinear function of maternal BMI. The varying coefficient model takes the form ${logit(ChildAsthma}_i)={\beta }_0+{\beta }_1{LSCR}_i+{\beta }_2({MaternalBMI}_i)+{\beta }_3({MaternalBMI}_i){LSCR}_i+{\beta }_4{x}_i$, where ${\beta }_2\left({MaternalBMI}_i\right)$ is the main, potentially nonlinear, smooth log odds ratio between child asthma and maternal BMI when ${LSCR}_i=0,{{\beta }_1+\beta }_3\left({MaternalBMI}_i\right)$ is the log odds ratio between maternal LSC-R and child asthma at a given value of maternal BMI, and $x_i$ is a vector containing the additional covariates. Figures were created to present estimates and corresponding confidence intervals for the LSC-R – asthma odds ratio, characterized by $\mathrm{e}\mathrm{x}\mathrm{p}[{\beta }_1+{\beta }_3\left(MaternalBMI\right)]$, as a function of maternal BMI. To additionally examine effect modification by race/ethnicity, we next performed the varying coefficient models stratifying by maternal race/ethnicity. Analyses were conducted using the mgcv package in R (v4.0.2, Vienna, Austria), as well as SAS statistical software (v9.4, SAS Institute Inc., Cary, NC).

RESULTS

Table 1 delineates participant characteristics for the sample overall and by maternal race/ethnicity. Women were primarily from historically minoritized racial/ethnic groups (44.5% Black, 37.6% Hispanic), 58.9% reported educational attainment of high school or less, and 43% reported atopy during pregnancy. In the overall cohort, 25.5% of women were overweight, and 29.8% were obese. Black (29% for Class 1/Class 2, 7.4% for Class 3) and Hispanic (27% for Class 1/Class2, 3.9% for Class 3) women had a higher prevalence of obesity and extreme obesity compared to White women (10.3% for Class 1/Class 2, 0.7% for Class 3) (p<0.001). Overall, the median (interquartile range) of LSC-R scores was 4 (17) with a range of 0–55. On average, White women had lower LSC-R scores, higher educational attainment status, and were older at enrollment when compared to Black and Hispanic women (all p<0.01). Hispanic women were less likely to have active atopy during pregnancy compared to Black and White women (p=0.01). Child sex was similar for all race/ethnicity groups (p=0.96). Overall, 17.1% of children had asthma with a mean age (standard deviation) of 3.1 (±2.0) years at initial report. Asthma incidence was similar in Black (18.1%) and Hispanic (19.6%) children, but significantly lower in White (9.6%) children (p=0.03).

Table 1.

Participant characteristics of mother-child dyads enrolled in the PRISM Study, by maternal race/ethnicity

	All (N=760)		Black (n=338)		Hispanic (n=286)		White (n=136)
Maternal age at delivery
years (median, IQR*)	29.4	(24.9–33.9)	27.7	(24–32.5)	28.2	(23.9–33.2)	33.3	(30.9–36.0)
Maternal LSC-R weighted score
LSC-R weighted (median, IQR*)	9	(4–17)	11	(6–20)	8	(4–16)	5	(3–9)
Maternal education (n, %)
>12 years	448	59.0	188	55.6	128	44.8	132	97.1
≤12 years	312	41.1	150	44.4	158	55.2	4	2.9
Maternal BMI (kg/m2) categories (n, %) *
Normal/underweight (<25)	343	45.1	122	36.1	121	42.3	100	73.5
Overweight (25 to <30)	191	25.1	93	27.5	77	26.9	21	15.4
Class 1 and Class 2 Obesity (30 to <40)	189	24.9	98	29.0	77	26.9	14	10.3
Class 3 obesity (≥40)	37	4.9	25	7.4	11	3.9	1	0.7
Maternal active atopy (n, %)
No	433	57.0	176	52.1	184	64.3	73	53.7
Yes	327	43.0	162	47.9	102	35.7	63	46.3
Child sex (n, %)
Girls	357	47.0	159	47.0	133	46.5	65	47.8
Boys	403	53.0	179	53.0	153	53.5	71	52.2
Child ever asthma (n, %)
No	630	82.9	277	82.0	230	80.4	123	90.4
Yes	130	17.1	61	18.1	56	19.6	13	9.6

Maternal race and ethnicity: Black (non-Hispanic Black and Black-Hispanic), Hispanic (Hispanic/non-Black), and White (non-Hispanic White).

^*IQR = interquartile range (25th percentile–75th percentile).

LSC-R = Life Stressor checklist - Revised

BMI = body mass index

Figure 1 and Figure 2 present boxplots of maternal LSC-R weighted scores and pre-pregnancy BMI, respectively, stratified by maternal race/ethnicity. Both LSC-R weighted scores and BMI were highest among Black, followed by Hispanic, then White women. In addition, the ranges of variation for maternal pre-pregnancy BMI and LSC-R weighted scores were most narrow among White, compared to Black and Hispanic (see Online Supplement, Figure E1) women. For example, the 75^th percentile of LSC-R in Black women (LSC-R=20) was similar to the 95^th percentile in White women (LSC-R=21); the highest pre-pregnancy BMI and LSC-R score were 60 kg/m² and 55, respectively in Black women compared to 44.6 kg/m² and 31 in White women. Correlations between LSC-R weighted scores and maternal pre-pregnancy BMI were weak (Spearman correlation, r=0.08 (p=0.04,) among all participants; correlations were r=−0.008, p=0.88, r=−0.036 p=0.54, and r=0.18, p=0.04, in Black, Hispanic, and White sub-groups respectively).

Figure 1. Maternal Life Stressor-Checklist-Revised (LSC-R)-weighted scores by maternal race/ethnicity.

Boxplots of maternal LSC-R weighted scores stratified by maternal race/ethnicity (Black: Black/Black-Hispanic; Hispanic: non-Black Hispanic; White: non-Hispanic White). Bottom and top of boxes denote 25^th percentile (Q1) and 75^th percentile (Q3) of distribution; thick line within boxes denotes median. Whiskers denote [Q1 – 1.5*IQR] and [Q3 + 1.5*IQR]; dark black dots represent potential extreme values. IQR: interquartile range.

Figure 2. Maternal pre-pregnancy body mass index (BMI) (kg/m²) by maternal race/ethnicity.

Boxplots of maternal pre-pregnancy BMI (kg/m²) were derived, stratified by maternal race/ethnicity (Black: Black/Black-Hispanic; Hispanic: non-Black Hispanic; White: non-Hispanic White). Bottom and top of boxes denote 25^th percentile (Q1) and 75^th percentile (Q3) of the distribution; thick line within boxes denote median. Whiskers of denote [Q1 – 1.5*IQR] and [Q3 + 1.5*IQR]; dark black dots represent potential extreme values. IQR: interquartile range.

Multivariable Logistic Regression Models

Table 2 summarizes results from the logistic regression models of LSC-R predicting child asthma, in the overall sample as well as stratified by maternal race/ethnicity and pre-pregnancy BMI. Results from the overall sample demonstrated that a one-standard deviation increase in LSC-R weighted score was associated with increased relative odds of child asthma in women in the obese category defined as ≥30 kg/m² (OR=1.84, 95% CI: 1.27–2.67), whereas this association was not detected among women in the normal/underweight (<25 kg/m²) or overweight groups (≥25 and <30 kg/m²). When stratified by maternal race/ethnicity, the association between higher maternal LSC-R and child asthma was observed in women in the obese category (≥30 kg/m²) in Black (OR=1.66, 95% CI: 1.06–2.60) and Hispanic (OR=2.33, 95% CI: 1.10–4.96) women. Small cell size in the non-Hispanic White group prohibited evaluation of the association between maternal LSC-R and child asthma in this group (n=2 with asthma among n=15 in obese group). Further, multiple degree of freedom tests of the interaction models confirmed a significant interaction between maternal LSC-R scores and BMI categories for the overall sample (p=0.03) and Black group (p=0.01). These findings highlight the varying impact of LSC-R on child asthma across different BMI categories and racial/ethnic groups.

Table 2.

The association between maternal lifetime stress and child asthma, stratified by maternal race/ethnicity and maternal pre-pregnancy BMI categories

Odds ratios and 95% confidence intervals of maternal Life Stressor Checklist-Revised (LSC-R) weighted score
	All				Black				Hispanic				White
	OR	95%	CI		OR	95%	CI		OR	95%	CI		OR	95%	CI
Overall	1.07	0.88	1.30		0.94	0.72	1.24		1.21	0.89	1.65		1.50	0.60	3.72
By BMI (kg/m 2 ) categories
Normal/underweight (<25)	0.81	0.57	1.16		0.46	0.23	0.94		1.11	0.68	1.81		1.74	0.51	6.01
Overweight (25 to <30)	0.93	0.62	1.42		0.93	0.55	1.56		0.97	0.46	2.08		n/a	n/a	n/a
Obese (≥30)	1.84	1.27	2.67	*	1.66	1.06	2.60	*	2.33	1.10	4.96	*	n/a	n/a	n/a

Models adjusted for maternal age at birth, maternal education, maternal active atopy, child sex (and maternal race/ethnicity for the overall sample model).

Maternal race/ethnicity: Black (non-Hispanic Black and Black-Hispanic), Hispanic (Hispanic/non-Black), and White (non-Hispanic White).

OR: odds ratio corresponding to per Standard deviation (SD) increase in LSC-R weighted score (SD=10.6)

^*Significant at α=0.05 level.

n/a: Model unstable due to small cell size

BMI = body mass index

Varying Coefficient Models

We used varying coefficient modeling to further examine the association between maternal lifetime stress and child asthma and assess for the interaction along a continuum of maternal BMI. Figure 3 presents the odds ratio characterizing the association between LSC-R and childhood asthma, together with pointwise 95% confidence intervals, as a continuous function of centered and scaled (by one SD) BMI. In the cohort overall, the varying coefficient model demonstrated significant association between higher LSC-R and child asthma at higher BMI, specifically ≥36.1 kg/m² (Figure 3A). In models stratified by race/ethnicity, the association was significant in Black women at BMI ≥43 kg/m² (Figure 3B). Although varying coefficient models allow detection of non-linear interactions between LSC-R and BMI -- that is, it allows the LSC-R – child asthma odds ratio to vary non-linearly as a function of BMI -- our results suggest that this LSC-R and child asthma odds ratio is linear as a function of BMI. Among Hispanic women, the overall ORs were >1 along the continuum of maternal BMI when it was above the mean (SD=0 and above); however, the association did not reach statistical significance (Figure 3C).

Figure 3. Odds ratio of child asthma in relation to maternal lifetime stress exposure: Effect modification by pre-pregnancy body mass index (BMI).

Varying-coefficient models for (A) Overall, (B) Black, (C) Hispanic/non-Black. The figure demonstrates the change in OR on the association between maternal LSC-R (per SD increase) and child asthma across different levels of pre-pregnancy BMI (centered and scaled by one SD). The OR was estimated as a smooth function of pre-pregnancy BMI (solid line) and 95% CIs (dashed line). Models adjusted for maternal age, education, atopy during pregnancy, and child sex (and race/ethnicity for the overall sample model). These results suggest that the interaction between LSC-R and pre-pregnancy BMI on child asthma is likely linear. SD=11.6 for LSC-R; SD=6.6 for BMI.

DISCUSSION

In this ethnically mixed sample, among children born to women with a pre-pregnancy BMI categorized as obese (≥30kg/m2), higher maternal lifetime stress was associated with increased odds of asthma (OR=1.84, 95% CI: 1.27–2.67); this association was not significant across other BMI categories. Varying coefficient models assessing the association of maternal lifetime stress and child asthma across the continuum of maternal pre-pregnancy BMI confirmed statistically significant associations at the highest levels of BMI, with race/ethnicity-stratified models detecting the strongest association in children born to Black women with Class 3 obesity (“severe” obesity; ≥40 kg/m²). Although a similar pattern was seen in Hispanic dyads, associations were not significant in the varying coefficient models.

Our approach to these analyses was grounded in life course theory of racial/ethnic health disparities(32) and the recognized importance of structural racism and its consequent differential impact on the lived experiences of women from different racial/ethnic backgrounds that can become biologically embedded and influence asthma risk in the next generation (33, 34), A number of lines of evidence support such a life course environmental justice perspective. First, there is growing research linking poorer pregnancy and birth outcomes with increased exposure to adverse childhood experiences (ACEs), which include harmful events and conditions assessed on the LSC-R, including child abuse and neglect, domestic violence, psychopathology in a primary caregiver, and incarceration of a parent or caregiver (35, 36). While much of the stress and asthma literature considers contemporary life events experienced during the index pregnancy (3), our group has demonstrated associations between maternal early life exposure to adverse experiences (5, 37) with adverse respiratory outcomes in offspring, underscoring the role for even remote experiences.

Second, both ACEs (38) and adverse adult experiences (39) have been documented to be more prevalent among U.S. non-Hispanic Black than non-Hispanic White individuals. In our sample, Black and Hispanic women experienced higher levels of cumulative lifetime traumatic and non-traumatic stressors compared to non-Hispanic White women. Women of color were also nearly twice as likely to be overweight and three times as likely to be obese prior to pregnancy when compared to White women consistent with prior research demonstrating racial/ethnic disparities in obesity in adults and particularly in women (24). Women meeting criteria for severe obesity (BMI ≥40) were primarily Black women in our sample who also experienced greater lifetime stress or adversity. Recent work by Sandsaeter and colleagues linked childhood adversities with increased pre-pregnancy BMI, demonstrating that the strength of the positive associations between childhood adversities and pre-pregnancy obesity increased with increasing obesity level (40).

Third, both lifetime maternal stress and obesity have been shown to increase oxidative stress levels in pregnancy, with increasing attention on the role of oxidative stress in asthma development (41). In prior analyses in the cohort leveraged herein, our group examined the association between increased maternal lifetime stress as assessed by the LSC-R and placental mitochondrial DNA (mtDNA) mutational load, conceptualized as a dosimeter of cumulative oxidative stress carried into pregnancy. Cumulative maternal stress was associated with greater mitochondrial mutational load, particularly among Black women when compared to White and Hispanic women. Whether racial/ethnic differences in mutational load on placental functioning mediates in part, the impact of lifetime maternal adversity on asthma development in offspring, warrants further study. Other potential mechanistic links between chronic stress exposure and asthma risk include immune modulation, epigenetic modifications, and microbiome alterations (42–46). Maternal stress is associated with an increase in endogenous and exogenous glucocorticoids which leads to increased inactivating enzymes in the placenta and fetus (11BetaHSD2), potentially causing altered immunity (cytokine milieu perturbations and increased allergic sensitization), increased immune tolerance, hematopoietic stem cells, corticotrophin releasing hormone, and T cell differentiation/selection (47). Placental epigenetic alterations in the oxidative stress pathway have been noted in association with maternal asthma which may be another link to childhood asthma that remains to be determined (43). Airway microbiota in children has also been noted to be altered in association with prenatal stress and the development of asthma(45, 48).

Finally, we assessed for effect modification by race/ethnicity although our ability to make definitive conclusions was limited due to our sample size across different race and ethnicities and BMI categories. Differences by race and ethnicity may reflect a surrogate marker for the impact of structural racism. Race is a sociopolitical construct and not biologic (49). Structural racism includes a system in which institutional and public policies and practices create and fuel racial inequities that impact opportunities, resources, and wellbeing often mediated through social determinants of health including where individuals are born, grow, live, learn, play, work, and age (50). Structural racism normalizes historical cultural and political practices that disadvantage racial/ethnic minoritized groups and may increase cumulative lifetime stress(50). Chronic stress has been associated with the development or severity of a number of diseases and biologic processes including hypertensive disorders, cardiovascular disease, asthma, rheumatoid arthritis, anxiety disorders, depression, chronic pain, human immunodeficiency virus/AIDS, stroke, certain types of cancer, accelerated biological aging and premature mortality (64, 65, and 66). Cumulative lifetime exposures stressors, even when remote, may have particular influence on both maternal obesity and stress-related programming of respiratory disease in offspring. This may be particularly relevant for women from BILPOC or other minoritized groups who as result of structural racism may disproportionately struggle with multiple adverse exposures, including environmental toxicants, exacerbations of chronic diseases such as asthma, lower socioeconomic resources, food insecurity, higher pre-pregnancy BMI, psychosocial stress and other adverse factors that may cluster together and have synergistic effects that impact the child.

We note several strengths in our analyses including the reasonably large ethnically and socioeconomically diverse, urban, higher-risk sample with well-characterized lifetime cumulative traumatic and non-traumatic stress and a sample size that allowed consideration of complex interactions with pre-pregnancy BMI and maternal race/ethnicity. There are also some limitations to consider. Maternal lifetime stress and trauma exposure were obtained using maternal report of LSCR, however the LSC-R is a measure of cumulative stress that has established test-retest reliability and has been validated in diverse populations(28) and is associated with biologic responses implicated in asthma(51–55). Although this is a widely used measure with documented validity and reliability, there are data suggesting possible under-reporting of trauma exposure which would lead to an underestimate of associations(56). Moreover, the LSC-R was collected prior to birth and thus unlikely to have been differentially reported based on whether the child developed asthma. We use BMI calculated from self-reported data versus measured. Although several studies have reported a degree of under-estimation of pre-pregnancy weight when comparing self-reported weight to weight measured in the 1^st trimester or from pre-pregnancy electronic medical record data, findings included a high degree of concordance in classification of BMI category(57) and similar results when assessing associations of perinatal outcomes with self-reported or measured BMI(58), which may extend to childhood findings. Notably if underestimation of weight is higher in individuals with higher BMIs(58), associations with LSCR and child asthma, by pre-pregnancy BMI may be underestimated particularly in the subgroup of Black or Hispanic women which on average had higher BMIs (56).

Women enrolled in PRisM are predominately Black and Latina from an urban environment, which are groups that have been historically underrepresented in pregnancy cohort investigations, despite increased prevalence of chronic conditions such as asthma in these same groups. This may limit generalizability to samples with other racial, ethnic and sociodemographic makeup. The small sample size in some subgroups (e.g., non-Hispanic Whites) may also impact the stability of estimates and limited our ability to draw inferences from this subgroup(59). Lastly, child asthma was based on maternal report of a diagnosis by a healthcare provider; parental/caregiver report is the most commonly used measure of a clinical diagnosis at this age (60). However, there may be misclassification of parent-report of child asthma outcomes, which would likely be non-differential and lead to underestimation of associations. Clinical assessments of asthma may decrease misclassification in future investigations.

In summary, these analyses demonstrate complex associations among maternal lifetime stress, pre-pregnancy obesity, race/ethnicity and childhood asthma risk, identifying those at particular risk. Detailed life course epidemiologic studies are needed to better understand these associations and the upstream factors related to differential lived experiences that place women on a suboptimal health trajectory that can influence programming of adverse respiratory health outcomes in the next generation. Future studies should consider other stressful conditions or experiences such as discrimination or the intersectionality of racism and sexism in relation to intergenerational transmission of asthma risk (61). It will be important for future work in larger studies to consider how prenatal and postnatal environmental exposures, such as ambient air pollutants that influence inflammatory, oxidative stress or other biologic pathways (41, 42, 44), may modify the association between maternal stress and child asthma, to further characterize children at highest risk for asthma or contribute to understanding of the observed associations with maternal factors in this current study(59). Studies are also needed to assess underlying mechanisms operating in pregnancy that differentially program asthma risk across historically minoritized groups compared to more advantaged populations. Once replicated in other cohorts, these findings have the potential to impact public health policy for expectant mothers or preconception; while longer-term multisector solutions are developed and implemented to address the structural barriers that impact health, in the short term it may be important to consider whether interventions targeting prenatal stress (mindfulness based stress reduction(62)) and gestational weight gain (structured exercise(63)) have beneficial effects on child respiratory health.

Abbreviations:

AIDS

Acquired immunodeficiency virus

BILPOC

Black Indigenous Latino Persons of Color

BMI

Body Mass index

LSC-R

Lifetime Stressor Checklist – Revised

MtDNA

Mitochondrial DNA

OR

Odds Ratio

PRISM

PRogramming of Intergenerational Stress Mechanisms