Prenatal lead exposure and childhood lung function: Influence of maternal cortisol and child sex

Abstract

Introduction:

Maternal hypothalamic-pituitary-adrenal (HPA) axis disruption in pregnancy may contribute to the programming of childhood respiratory disease and may modify the effect of chemical toxins, like lead (Pb), on lung development. Child sex may further modify these effects. We sought to prospectively examine associations between maternal HPA axis disruption, prenatal Pb and childhood lung function and explore potential effect modification by maternal cortisol and child sex on the association between prenatal Pb and lung function outcomes.

Materials and Methods:

Analyses included 222 mothers and children enrolled in a longitudinal birth cohort study in Mexico City. Maternal diurnal salivary cortisol was assessed in pregnancy; cortisol awakening response (CAR) and diurnal slope were calculated. Blood Pb was measured during the second trimester of pregnancy. Post-bronchodilator lung function was tested at ages 8–11 years. Associations were modeled using generalized linear models with interaction terms, adjusting for covariates.

Results:

A higher (flatter) diurnal slope was associated with lower FEV₁/FVC ratio (β: −0.433, 95%CI [−0.766, −0.101]). We did not find any main effect associations between prenatal Pb and lung function outcomes. We report an interaction between Pb and cortisol in relation to FEV₁/FVC and FEF_25–75% (p_interaction<0.05 for all). Higher prenatal Pb was associated with reduced FEV₁/FVC only in children whose mothers had a high CAR. Higher prenatal Pb was also associated with reduced FEV₁/FVC and FEF_25–75% in mothers with a flatter diurnal slope. A 3-way interaction between prenatal Pb, CAR and sex on FEV₁/FVC, indicated that boys born to women with high CAR and higher prenatal Pb levels had lower FEV₁/FVC ratios (p_interaction=0.067).

Conclusions:

Associations between prenatal Pb and childhood lung function were modified by disrupted maternal cortisol in pregnancy and child sex. These findings underscore the need to consider complex interactions to fully elucidate effects of prenatal Pb exposure on childhood lung function.

1. Introduction

Lung function growth patterns are established as early as 7 years of age and are an important determinant of peak lung function and subsequent rate of decline in adulthood (Lange et al., 2015; Stern et al., 2007). Therefore, identifying potentially modifiable environmental factors contributing to poor lung function growth patterns in early life has implications for long-term respiratory health (Schultz et al., 2018; Stern et al., 2007). Lung development begins in utero through a carefully orchestrated sequence of events (Pinkerton and Joad, 2006) involving the coordinated functioning of interactive networks, including neuroendocrine, autonomic, and immune function, influencing lung growth and development. Beginning in utero, environmental factors, can disrupt these signaling pathways in adverse ways.

The maternal-fetal hypothalamic-pituitary-adrenal (HPA) axis plays a significant role in fetal development (Sheng et al., 2020; Wright, 2010). HPA axis disruption in pregnancy can contribute to the programming of respiratory disease through direct trans-placental passage of maternal hormones like cortisol and/or by indirectly impacting maternal-fetal immune function (Wright, 2010; Wright, 2012). Cortisol secretion typically follows a homeostatic diurnal pattern with low levels at awakening that rise quickly that then fall during the course of the day. Altered prenatal cortisol production, evidenced by a flatter diurnal slope, is linked to infant respiratory illnesses broadly (Beijers et al., 2010) including increased risk of repeated wheezing in infancy (Wright et al., 2013). Additionally, overlapping animal (Cory-Slechta et al., 2008) and human (Campbell et al., 2019; Enlow et al., 2017; Flom et al., 2018) studies demonstrate that disrupted prenatal HPA axis functioning can moderate the effects of other environmental toxins, on child developmental outcomes.

Toxic metals, particularly exposure to lead (Pb), have been cross-sectionally associated with lung function deficits in children (Little et al., 2017; Madrigal et al., 2018; Zheng et al., 2013) and adults (Leem et al., 2015; Pak et al., 2012; Wei et al., 2020; Yang et al., 2019). Prenatal Pb exposure has been linked with many adverse developmental outcomes in children, including diagnoses linked with poor lung growth and development such as preterm birth, restricted fetal growth (Rodosthenous et al., 2017; Vigeh et al., 2011), allergic disorders and infantile wheezing (Kim et al., 2019; Pesce et al., 2021; Shaheen et al., 2004). Associations between prenatal Pb exposure and childhood lung function have not been examined. Because Pb exposure and disrupted cortisol are both associated with immune modulation and inflammation, they may act synergistically (Kasten-Jolly and Lawrence, 2014). Associations between Pb and cortisol exposures in utero have not been examined in relationship to prenatal programming of childhood lung function. Finally, susceptibility to prenatal Pb and disrupted maternal cortisol varies by fetal sex (Cowell and Wright, 2017; Kasten-Jolly and Lawrence, 2017). While animal and human studies also demonstrate sex differences in lung development (Ishak et al., 2014a; Torday and Nielsen, 1987), the environmental factors that may underlie these differences are not well understood.

Leveraging a prospective pregnancy cohort study, we first examined independent effects of maternal-fetal HPA axis functioning assessed via diurnal salivary cortisol indices, prenatal blood Pb and lung function ascertained in children at ages 8–11 years. We hypothesized that higher prenatal Pb exposure and disrupted maternal cortisol (a higher cortisol awakening response [CAR] and flatter diurnal slope) would be associated with lower lung function in childhood. We next examined interactions between prenatal Pb and maternal cortisol. Finally, we explored 3-way interactions between Pb, cortisol and child sex.

2. Materials and Methods

2.1. Study population

Between July 2007 and February 2011, pregnant women receiving prenatal care through the Mexican Social Security System (Instituto Mexicano del Seguro Social –IMSS) were recruited into the PROGRESS study. Women were eligible to participate if they met the following criteria:< 20 weeks gestation, at least 18 years of age, had completed primary education, planned to stay in Mexico City for the next 3 years, had access to a telephone, had no medical history of heart or kidney disease, did not consume alcohol daily, and did not use any steroid or anti-epilepsy medications (Burris et al., 2013). Procedures were approved by institutional review boards at the Harvard School of Public Health, Icahn School of Medicine at Mount Sinai, and the Mexican National Institute of Public Health. Women provided written informed consent and children provided assent once they reached 7 years of age.

2.2. Salivary cortisol

Between 19.1 ± 2.3 weeks gestation (mean ± standard deviation [SD]), participants collected diurnal saliva samples at home for the assessment of salivary cortisol as described previously (Braun et al., 2014). Briefly, women provided five saliva samples each day over two consecutive days using the passive drool technique (Strazdins et al., 2005). Women were instructed to provide samples into Salicaps (IBL International, Hamburg, Germany) upon awakening (“when you open your eyes”), 45 minutes after waking, 4 hours after waking, 10 hours after waking, and at bedtime (“right before getting into bed”). Participants were instructed not to eat, brush their teeth, or drink liquids for at least 15 minutes before providing a sample and not drink beverages containing caffeine before collecting the first two samples. Women recorded the collection time of each sample. They also answered questions about mood states, exercise, sleep and medication use on collection days, and this information was considered if atypical levels were observed. Saliva samples were assayed in the same batch in duplicate for cortisol using a commercially available chemi-luminescence-assay with sensitivity of ~0.16 ng/ml (IBL; Hamburg, Germany, Clemens Kirschbaum). Control sera covering at least three levels of cortisol were run during each 24-hour time period and intra- and interassay coefficients of variation were less than 8%.

2.3. Prenatal maternal blood lead

Venous blood from women was drawn into trace metal vacutainer (Becton-Dickinson and Company, Franklin Lakes, New Jersey) tubes containing EDTA during the second trimester (2T) between 16 and 20 weeks of gestation. Samples were refrigerated at 4 °C until shipment to the laboratory, where they were kept at −20°C until analyzed. Blood Pb was analyzed as previously described (Kupsco et al., 2019; Renzetti et al., 2017). Briefly, after digestion in concentrated HNO₃ and 30% H₂O₂ blood was analyzed on an Agilent 8800 ICP Triple Quad (ICP-QQQ) instrument (Agilent Technologies, Inc., Santa Clara, CA) in MS/MS mode. Quality control measures were as previously described (Renzetti et al., 2017).

2.4. Child lung function

Subsequent funding was obtained to assess lung function in children aged 8–11 years, of which 277 children completed testing between October 2018 and March 2020. Children underwent pre- and post-bronchodilator spirometric testing in their homes by a trained field physician and nurse. Prior to testing, height was measured with a fixed stadiometer to the nearest 0.1 cm, and an electronic scale was used to measure weight to the nearest 0.1 kg. Spirometry was conducted in accordance to American Thoracic Society (ATS) guidelines with a portable MedGraphics^™ PC based USB spirometer which displays real-time flow-volume plots to facilitate testing(Miller et al., 2005). Flow was measured with a heated screen pneumotachograph (flow range 0±20 l/s, accuracy 0.2 to12 l/s ±2 %) and volume was measured by digital integration. Calibration preceded each session using a standard 3L syringe accounting for ambient temperature, air pressure and humidity. Participants were included if they reported no acute respiratory symptoms for ≥3 weeks. Short-acting beta-agonists, anticholinergic and theophylline preparations were withheld 4 hours prior to testing; long-acting beta-agonists were withheld for 12 hours and long-acting theophylline preparations for 24 hours. Parameters recorded from a minimum of 3 (and no more than 8) maneuvers included: forced vital capacity (FVC, liters), forced expiratory volume at 1 s (FEV₁, liters), FEV₁/FVC and forced expiratory flow at 25–75% of the pulmonary volume (FEF_25–75%, liters). Participants received 2 (200 mcg total) puffs of salbutamol through a spacer with mask and spirometry was repeated 15 minutes later. All tests were overread for acceptability and reproducibility by a pulmonologist. 32 children who did not achieve acceptable tests had repeated testing with 23/32 achieving acceptable and reproducible results. Pre-bronchodilator, 245 (88%) tests met the criteria for acceptability and reproducibility and 236 (85%) tests met criteria post-bronchodilator. For analyses, PFT parameters were regressed on age, sex, height, then calculated residual values were divided by the standard deviation (SD) of the residuals, to convert to z-scores with mean of 0 and SD of 1. Prenatal Pb data was missing in 14 of the 236 children with reproducible post-bronchodilator spirometry (n for analysis = 222). A detailed flow diagram of participants included in analysis is shown in Figure S1. We did not find any significant differences between participants invited into the study and the remaining age-eligible participants (Table S1)

2.5. Covariates

We considered covariates previously linked to cortisol and childhood lung function but not on the causal pathway (eg. gestational age or birthweight) and confirmed covariates based on formulation of a Directed Acyclic Graph (DAG; Online Supplement, Figure S2). Models were adjusted for the minimal sufficient adjustment sets for estimating the total effect of prenatal cortisol on childhood lung function including maternal age (continuous in years), educational attainment at enrollment (<high school, some high school or high school graduate, >high school) and environmental tobacco smoke (ETS) exposure. Smoking data were obtained via self-reported questionnaire. In our sample, only 2 mothers reported smoking in pregnancy, therefore only prenatal exposure to ETS was included in the models and was defined as report of anyone smoking inside the home during the second or third trimester of pregnancy.

2.6. Statistical Analysis

Cortisol values were log-transformed to reduce skewness. We examined two features of women’s diurnal cortisol patterns. The rapid increase in cortisol concentrations in the morning, known as the cortisol awakening response (CAR) was estimated from the change in cortisol concentrations between the first and second saliva samples of each day separately. The decline in concentrations over the course of the day is known as diurnal slope and was estimated for each individual collection day using the change in salivary cortisol concentrations between the 1^st and 5^th samples of each day, excluding the 2^nd sample. Steeper slopes indicate a more rapidly declining cortisol output throughout the course of the day, flatter slopes indicate a slower decline in cortisol output over the course of the day. In addition, cortisol values were averaged across both collection days to increase stability. Blood Pb was right skewed and was natural log (ln)-transformed.

All associations were tested using generalized linear models. For main analyses, cortisol measures and blood Pb were analyzed as continuous variablesCortisol measures were dichotomized at the median for interaction analyses. We first tested the main effects of cortisol measures and Pb exposure on each lung function outcome separately, adjusting for covariates. We then examined effect modification by CAR (≤median/>median), diurnal slope (≤median/>median), and child sex (male/female) for lung function outcomes by including the relevant interaction terms. Finally, we considered 3-way interactions between these variables (Blood Pb × sex × CAR; Blood Pb × sex × slope). Analyses were performed in R Version 3.5.1 (Vienna, Austria) and SPSS version 24 (Chicago, IL) with statistical significance set at 0.05.

3. Results

Participants’ demographic characteristics are shown in Table 1. The majority of mothers included in this study had 12 or fewer years of schooling (79.4%). Over a third of women (36.6%) reported exposure to a smoker in the home during pregnancy. Our sample included slightly more male than female children. The mean age of children at testing was 9.61 years. The median log transformed blood Pb was 1.11μg/dL, median CAR −0.24 and diurnal slope −0.92. Blood Pb and cortisol measures were not significantly correlated (Pearson correlation values <0.1 and p-values >0.38 for both).

Table 1.

PROGRESS participant characteristics N=222

Categorical variables	n (%)
Child’s sex
Male	121 (54.5)
Female	101 (45.5)
Report ETS in the home in pregnancy*	82 (36.9)
Maternal education at enrollment
< than high school	91 (41.0)
Some high school or high school graduate	88 (39.6)
> than high school	43 (19.4)
Continuous variables
Cortisol Awakening Response (μg/dl; median, IQR) Ω	−0.24 (−4.79, 5.90)
Diurnal slope (median, IQR) Ω	−0.92 (−1.19, −0.66)
Log 2nd trimester blood Pb, (μg/dl; median, IQR)	1.11 (0.67, 1.51)
FEV1 (L; mean, SD)	2.11 (0.39)
FVC (L; mean, SD)	2.38 (0.45)
FEF25–75% (L/s; mean, SD)	2.70 (0.65)
FEV1/FVC ratio (mean, SD)	88.7 (5.69)
z-score of FEV1 (median, IQR) †	−0.02 (−0.61, 0.73)
z-score of FVC (median, IQR) †	0.03 (−0.64, 0.61)
z-score of FEF25–75% (median, IQR) †	0.02 (−0.58, 0.65)
z-score of FEV1/FVC ratio (median, IQR) †	0.10 (−0.57, 0.68)
Child age at spirometry, median (IQR)	9.57 (9.23–10.2)
Child height at spirometry (cm), median (IQR)	135.5 (130.5–141.6)
Maternal age at enrollment years, median (IQR)	27.1 (23.8–31.5)

^*maternal report of smokers inside the home at second or third trimester

^ΩData available for n=206 participants

^†Adjusted for age, sex, height.

3.1. Main Effect Models

Table 2 shows the three separate multivariable-adjusted regression models between cortisol measures, 2^nd trimester Pb, and lung function z-scores. We found that a more positive diurnal slope (flatter) was significantly associated with lower FEV₁/FVC ratio (β: −0.433, 95%CI [−0.766, −0.101], p=0.011) and had a suggestive association with lower FEF_25–75% (β: −0.308, 95%CI [−0.627, 0.011], p=0.06). While 2^nd trimester blood Pb levels were negatively associated with lung function outcomes, none of these associations were statistically significant. Results did not substantially vary if both Pb and each cortisol measure were included in the same model (Supplemental Table S2).

Table 2.

Multivariable-adjusted regression models between 2^nd trimester Pb, cortisol measures and lung function outcomes.

	Post-bronchodilator z-score
	FEV1	FVC	FEV1/FVC	FEF25-75%
CAR	0.002 (−0.014, 0.019) p=0.768	0.009 (−0.008, 0.025) p=0.303	−0.011 (−0.027, 0.006) p=0.202	−0.008 (−0.024, 0.007) p=0.301
Diurnal slope	−0.224 (−0.560, 0.112) p=0.193	0.005 (−0.326, 0.342) p=0.977	−0.433 (−0.766, −0.101) p=0.011	−0.308 (−0.627, 0.011) p=0.060
Pb	−0.171 (−0.392, 0.051) p=0.133	−0.142 (−0.362, 0.078) p=0.208	0.030 (−0.253, 0.193) p=0.794	−0.146 (−0.358, 0.066) p=0.178

Adjusted for maternal age and education at enrollment and report of environmental tobacco smoke in pregnancy.

3.2. Interaction models

In two-way interaction models, we did not find evidence for effect modification by sex (all interaction p-values>0.25). We did find evidence for effect modification by CAR (p_interaction=0.009) on the association between prenatal blood Pb and FEF_25–75% (Figure 1). In stratified models, increasing blood Pb was significantly associated with lower FEF_25–75% (β−: 0.357, [95%CI: −0.677, −0.037]) among children born to women with a high (>median) CAR but not in those children whose mothers had low (≤median) CAR (β: 0.209, [95%CI: −0.098, 0.515]). We also found evidence for effect modification by diurnal slope (p_interaction=0.048 and (p_interaction=0.028) on associations between prenatal blood Pb and FEV₁/FVC and FEF_25–75% (Figure 2). Similarly, we found that increasing prenatal blood Pb was associated with lower lung function outcomes among children born to mothers with a more positive (flattened) diurnal slope.

Figure 1.

Multivariable analyses characterizing the association between continuous blood Pb at second trimester and post-BD FEF z-score (p_interaction=0.009) by low (≤median) and high (>median) maternal 2^nd CAR.

Figure 2.

Multivariable analyses characterizing the association between continuous blood Pb at second trimester and (A) post-BD FEV1/FVC (p_interaction=0.048) and (B) post-BD FEF z-score (p_interaction=0.028) by low (≤median) and high (>median) maternal diurnal cortisol slope.

3.3. Exploratory 3-way interactions

We also found evidence of a three-way interaction between child sex, CAR and blood Pb. Figure 3 depicts the 3-way interaction after stratifying by the relevant covariates. Male children born to women with a high CAR had higher blood Pb and lower FEV₁/FVC ratio (p_interaction=0.067).

Figure 3.

Multivariable analyses characterizing the association between continuous blood Pb at second trimester and post-BD FEV₁/FVC z-score (p_interaction=0.067) by child sex and low (≤median) and high (>median) maternal 2^nd CAR. Male children/high cortisol n=58, male children/low cortisol n=53, female children/high cortisol n=45, female children/low cortisol n=50.

4. Discussion

To our knowledge, this is the first prospective study examining associations among maternal diurnal cortisol profiles in pregnancy, prenatal Pb and lung function in childhood. We did find a main effect of cortisol in that a higher (flatter) maternal diurnal cortisol slope in pregnancy was associated with lower measures of FEV₁/FVC and FEF_25–75% in children aged 8–11 years. While the associations were in the expected directions, we did not find statistically significant associations between prenatal Pb and lung function outcomes in our main effect models. Assessment of effect modification more fully elucidated the effect of Pb on child lung function outcomes. The effect of prenatal Pb was modified by maternal cortisol levels, with its effects being most pronounced in children born to women with a high CAR and a flattened diurnal slope compared to those born to women with a lower CAR and a more typical diurnal slope (steeper decline over the day). In exploratory analyses, we found evidence of a three-way interaction between prenatal Pb, CAR and child sex, which indicated that male children born to women with a high CAR and higher blood Pb had the lowest FEF_25–75%.

Only two other studies have reported associations between altered diurnal cortisol rhythms and respiratory outcomes. Beijers and colleagues found that profiles including a flattened slope were associated with a nonspecific composite score of childhood respiratory outcomes (Beijers et al., 2010). Our group also previously reported on the association between a flattened diurnal slope and repeated wheeze in infancy (Wright et al., 2013). Similar to our results presented here, both studies linked a blunted HPA axis (characterized by flattening of the diurnal slope) with higher risk of adverse respiratory outcomes. We extend these findings by considering objective measures of lung function in older children.

Previous cross- sectional studies have reported associations between Pb and lung function, mostly limited to reported reductions in FEV₁ and FVC (Little et al., 2017; Madrigal et al., 2018; Zeng et al., 2017). Urinary Pb was inversely associated with FVC in children aged 6–17 years participating in the United States National Health and Nutrition Examination Survey (Madrigal et al., 2018). Concurrent blood lead was associated with lower FVC and VC in a study of 373 Polish children aged 10–15 years. Blood Pb was associated with lower FEV₁ but not lower FVC in 206 pre-school aged children in China (Zeng et al., 2017). In a small study of 107 elementary school-aged children in Mongolia, higher airborne Pb exposure was associated with reduced peak expiratory flow rate (Madaniyazi et al., 2013). Studies that have analyzed prenatal exposure to Pb and respiratory outcomes have reported mostly null results. In an analysis of data from the EDEN study in France, the authors did not find any significant associations between maternal serum Pb or cord blood Pb and higher risk of asthma or atopic diseases (Pesce et al., 2021). Higher cord blood Pb was associated with longer duration of atopic dermatitis but not development or severity in pre-school aged children in Korea (Kim et al., 2019). The Avon Longitudinal Study of Parents and Children did not report any associations between cord blood Pb and risk of early childhood wheezing and eczema (Shaheen et al., 2004).

Our data also supported an interaction between prenatal Pb exposure and maternal cortisol production in utero in relation to lower lung function parameters in childhood (i.e., children born to mothers with higher prenatal Pb exposure). Children born to women with adverse cortisol profiles were more likely to have lower lung function outcomes. Both Pb and disrupted cortisol may independently enhance a proinflammatory immune milieu or an oxidative state that influences the programming of lung development (Al-Hussainy and Mohammed, 2021; Cowell et al., 2020; Osborne et al., 2018; Zhou et al., 2019). Blood Pb has also been associated with alterations in immune function including increased levels of inflammatory markers (Kim et al., 2019; Metryka et al., 2018), total IgE (Wang et al., 2017; Wells et al., 2014) and eosinophils (Wells et al., 2014). Similarly, disruption of the HPA axis has been associated with increased levels of pro-inflammatory cytokines and affecting Th1/Th2 balance (Wright, 2010). Thus, when both exposures co-occur, one may potentiate the effects of the other.

While we did not find sex differences in the association between Pb and cortisol main effects on lung function outcomes, in exploratory analyses we did find evidence of effect modification by both child sex and maternal CAR, for the association between prenatal Pb and FEF_25–75%. The association between prenatal Pb and FEF_25–75% was strongest for male children born to women with a high CAR. Differential maturation in lung development of males relative to females may predispose male infants to childhood respiratory diseases (Demissie et al., 1998; Ishak et al., 2014b; Liptzin et al., 2015) and they may be more susceptible to the pro-inflammatory effects of prenatal exposure to both Pb and cortisol. Interestingly, some animal studies examining the effects of prenatal Pb exposure on adult neurobehavior have only found effects of Pb when considered along with stress/stress correlates such as cortisol, often in a sex-specific manner. For example, Corey-Slechta and colleagues (Cory-Slechta et al., 2004) observed interactions between maternal stress and maternal Pb exposure on fixed-interval schedule controlled performance more frequently in female than in male offspring. Sex-specific synergistic effects of maternal Pb and stress/stress correlates have also been reported for other outcomes including stress responsivity (Virgolini et al., 2006), behavioral outcomes (Weston et al., 2014) and epigenetic changes (Sobolewski et al., 2018) underscoring the need to examine these higher order interactions in prenatal programming effects.

Strengths of this study include the prospective design, availability of high quality post-bronchodilator lung function measurements and consideration of important confounders. This is the first study to examine effects of HPA axis disruption in pregnant women on lung function assessing cortisol rhythms through repeated measures over multiple days. Potential limitations should also be considered. We only examined cortisol rhythms and Pb at a single time-point in pregnancy. We measured salivary diurnal cortisol over the course of two days, which captures more subacute functioning (over days). We also examined our data for improbable patterns and cross-referenced it with diary information on day of collection (e.g., sample timing, eating, sickness, sleep), although we acknowledge that salivary cortisol is susceptible to other situational influences that we may not have captured. While our focus was the prenatal period, it will also be important to elucidate if postnatal Pb/maternal psychological functioning impact the child’s own developing stress response systems (i.e., HPA axis, autonomic functioning, immune function) and ultimately examine whether the associations between maternal cortisol disruption, Pb exposure in utero and lung function is, in part, mediated through effects on child stress physiology. These first-time findings should stimulate further research in this area including consideration of postnatal factors as well using life course longitudinal designs. Given our current sample size, this was beyond the scope of our study. We only invited a third of the original cohort to participate and the sample size may have also impacted our power to fully examine interactions and these findings warrant replication in future studies with increased sample size. As with any observational study, we cannot rule out residual confounding due to unmeasured factors that may influence lung function in childhood. We acknowledge that there might be other covarying exposures not considered in our analyses that can impact both HPA axis functioning and lung growth and development starting prenatally, including air pollution, other toxic metals and additional socioeconomic factors. PROGRESS is composed of low-income families in an urban setting and our results may translate to other disadvantaged populations who face similar exposures to metals, stressors and poverty. Similar studies conducted in other regions may help elucidate whether these results generalize to other populations.

5. Conclusions

These data highlight the need to consider effect modifiers in order to more fully elucidate the role that Pb toxicity may play in programming lung structure and function. Understanding mechanisms underlying modifying effects of cortisol as well as child sex on the link between prenatal Pb exposure and childhood lung function will further elucidate how respiratory disease is programmed starting prenatally. Furthermore, exposure to stress correlates and Pb are particularly important exposures in Mexico impacting pediatric morbidity (Caravanos et al., 2014; Forno et al., 2015). Thus, continued exploration of the links between environmental exposures and HPA axis disruption in pregnancy in relation to lung function may be particularly relevant for urban populations of lower socioeconomic status who are particularly burdened by these exposures. Enhanced knowledge of those who may most likely benefit from prevention and intervention strategies may serve to develop more efficacious protocols to promote optimal lung development and potentially reduce the burden of childhood respiratory disease. Given known associations between child and adult lung function, these interventions may have lasting implications.

Abbreviations:

CAR

Cortisol awakening response

CI

confidence interval

FEF_25–75%

forced expiratory flow at 25–75% of the pulmonary volume

FEV₁

forced expiratory volume in one second

FVC

forced vital capacity

HPA

Hypothalamic-pituitary-adrenal