Maternal vitamin D status modifies the effects of early life tobacco exposure on child lung function

Abstract

Background:

Prior studies suggest that vitamin D may modify the effects of environmental exposures, however, none have investigated gestational vitamin D and cumulative tobacco smoke exposure (TSE) throughout pregnancy and early life.

Objective:

We investigated the effects of early life TSE on child lung function and the modulatory effects of gestational vitamin D on this association.

Methods:

The Vitamin D Antenatal Asthma Reduction Trial recruited nonsmoking pregnant women and followed the mother-child pairs to age 6 years. TSE was assessed with questionnaires and plasma cotinine measurements in the mothers (10-18 and 32-38 gestational weeks) and children (1, 3, and 6 years). Cumulative TSE was calculated from the repeated cotinine measurements. 25-hydroxyvitamin D (25[OH]D) levels were measured at 10-18 and 32-38 gestational weeks. Lung function was assessed at 6 years with spirometry and impulse oscillometry.

Results:

Of the 476 mother-child pairs, 205 (43%) had increased cotinine levels at ≥1 time point. Cumulative TSE was associated with decreased FEV₁ (β −0.043 L, P=0.018) and increased respiratory resistance (R5; β 0.060 kPa/L/s, P=0.002). This association persisted in subjects with insufficient (<30 ng/ml) 25(OH)D levels throughout pregnancy (β 0.077 kPa/L/s, P=0.016 for R5) but not among those with sufficient levels throughout pregnancy.

Conclusion:

Cumulative TSE from pregnancy to childhood is associated with dose- and duration-dependent decreases in child lung function at 6 years even in the absence of reported maternal smoking. Gestational vitamin D may modulate this effect and have therapeutic potential for minimizing the adverse effect of TSE on lung throughout early life.

Graphical Abstract

Capsule summary:

This study demonstrates dose- and duration-dependent decreases in child lung function with exposure to secondhand smoking and provides a novel potential effect modifier with therapeutic potential: gestational vitamin D status.

INTRODUCTION

With an estimated 1.14 billion tobacco smokers and 7.69 million smoking-attributed deaths in 2019, tobacco smoking is one of the biggest preventable global health crises.¹ In addition to the detrimental effects of active smoking, secondhand tobacco smoke exposure (TSE) poses a health hazard especially for children whose developing lungs might be particularly susceptible to environmental insults. In 2016, 7.2% of women smoked during pregnancy, with the prevalence as high as 16.7% in certain subgroups. ² Even among women who do not smoke, exposure to secondhand smoke poses a problem. One study found that 38.5% of pregnant women attending an antenatal clinic reported being exposed to tobacco smoke in the past week. ³ Furthermore, TSE affects up to approximately 40% of children globally ⁴ and may lead to harmful health outcomes, such as increased respiratory infection susceptibility, asthma development, and other morbidities.^5-7 As nicotine and other toxic substances in tobacco smoke cross the placenta and their concentrations in the fetus may even exceed the maternal concentrations,⁸ the effects of tobacco smoke can start already in utero. Gestational TSE has been linked to different adverse events including low birth weight, preterm birth, physician-diagnosed asthma, and reduced childhood lung function.^9-14 However, most studies investigating the effects tobacco smoke are questionnaire-based and have measured the exposure at a single time point.

Self-reported measures of TSE are susceptible to the effects of social pressure and recall bias, and often fail to identify secondhand exposures.¹⁵ Cotinine is the predominant nicotine metabolite and can be used as an objective marker of both primary and secondary tobacco exposure from any nicotine delivery device.¹⁶ It is detectable in plasma for approximately 3 to 5 days after exposure.¹⁷ Previous studies have reported significant differences between self-reported TSE and cotinine measurements,^15,18 indicating that self-reported measures vastly underreport the exposure to nicotine products.

Vitamin D has immunomodulatory and anti-inflammatory effects¹⁹ and its deficiency has been associated with decreased lung function in adult smokers.²⁰ Furthermore, vitamin D has been linked to fetal lung and immune system development,^21-23 and might therefore be protective against the effects of tobacco smoke on the developing lung. However, vitamin D insufficiency is extremely common during pregnancy worldwide.²⁴ The Vitamin D Antenatal Asthma Reduction Trial (VDAART) randomized non-smoking pregnant women to high-dose vitamin D₃ supplementation or placebo.²⁵ Comprehensive follow-up data of the cohort included questionnaires on TSE, plasma cotinine levels, and plasma 25-hydroxyvitamin D (25-[OH]D) levels at multiple time points during pregnancy and childhood, and lung function measurement at age 6 years using spirometry and impulse oscillometry (IOS). We have previously shown that gestational vitamin D status is positively associated with childhood lung function.²⁶

In the present study, we hypothesized that in utero vitamin D deficiency leads to altered lung development and as a result, increased susceptibility to pre- and postnatal TSE. We first assessed the effect of cumulative TSE on child lung function at age 6 years using data on serum cotinine measurements at multiple time-points during pregnancy and childhood in the well-characterized VDAART cohort. We then examined the potential modulatory effects of gestational vitamin D status on the association between cumulative TSE during early life and childhood lung function.

METHODS

Study populations

The VDAART (clinicaltrials.gov identifier: NCT00920621) is a double-blinded multi-center trial that randomized non-smoking pregnant women at 10-18 gestational weeks (GW) for high-dose vitamin D₃ supplementation (4000 IU/d) or placebo.²⁵ All enrolled women and/or the biological fathers of the children had a history of asthma, eczema, or allergic rhinitis. Written informed consent was obtained from the mothers and the study was approved by the institutional review boards at the three clinical centers (Boston, St. Louis, and San Diego) and Brigham and Women’s Hospital, which served as the Data Coordinating Center.

Tobacco smoke exposure

Active smoking of the mothers and household smoking information in VDAART was collected through questionnaires administered at study enrollment, monthly during pregnancy, and quarterly after delivery (please see the Online Data Supplement). Only women who were self-reported nonsmokers at enrollment were included in the trial.

A detailed description of the metabolomic profiling method is provided in the Online Data Supplement. Global untargeted metabolomic data were measured by Metabolon Inc.²⁷ from the plasma samples of the mothers at 10-18 GW and at 32-38 GW and children at ages 1, 3, and 6 years. From the relatively quantified metabolomic data, we identified the tobacco metabolite cotinine as our primary metabolite of interest in objectively quantifying TSE. There were high missing proportions across all time points, resulting from either non-detection of cotinine or levels below the limit of detection. We created dichotomized variables for cotinine where missing was considered absent, and having a measured value was considered present at each time point. The group with detectable cotinine levels was further divided into “low” (lower 50%) or “high” (upper 50%) cotinine level by median value of each time point, creating a three-category variable (0/1/2 representing absent cotinine/lower 50%/upper 50%). Cumulative TSE from pregnancy to age 6 was quantified by adding up numeric values of the three-category cotinine variables from all five time points and normalizing this by the number of time points where metabolomic data were available, creating a scale from 0 to 2. Subjects were then divided into three groups based on their cumulative cotinine level indicator for group-wise comparisons: no cotinine exposure (no cotinine detected at any time point, i.e., cumulative cotinine indicator of 0), medium cotinine exposure (cumulative cotinine indicator of 0.5-1.49), and high cotinine exposure (cumulative cotinine indicator of ≥1.5).

Vitamin D measurements

Plasma 25[OH]D level was measured in the mothers at 10-18 and 32-38 GW and in the children at age 6 years using the DiaSorin Liaison chemiluminescence immunoassay.²⁵ Vitamin D sufficiency was defined plasma 25(OH)D level of ≥30 ng/ml.²⁶

Lung function measurements

Methods for quality control of the lung function measurements have been previously published.²⁸ Spirometry was performed at age 6 years according to the American Thoracic Society and European Respiratory Society guidelines²⁹ using the Jaeger MasterScreen PFT System (VyaireTM Medical, Inc.) with the child standing and using a nose clip, and regarded as successful if ≥2 acceptable and repeatable recordings were acquired. The mean value of each acceptable spirometry index was used in the analyses.

IOS was performed at age 6 years using the Jaeger MasterScreen IOS System (VyaireTM Medical, Inc.)²⁸ with the child sitting and using a nose clip, and regarded as successful if ≥2 acceptable and repeatable recordings with a regular tidal breathing pattern of ≥20 seconds were acquired. Respiratory resistance at 5 Hz (R5), which is considered to reflect the resistive properties of both large and small airways,³⁰ was selected a priori as the main outcome variable of IOS. The mean value of each IOS index was used in the analyses.

Statistical analyses

Student T test and Chi-squared tests were used to compare demographic data between groups, and Wilcoxon test to compare cotinine levels between groups. Due to the lack of validated multiethnic reference values for IOS, the lung function data were analyzed using height, sex, and race adjusted models. The GLI percent predicted spirometry values³¹ were used in Figure 4. The association between cumulative cotinine level (scale 0-2) and child lung function was evaluated using linear regression models adjusted for child height at age 6 years, sex (male or female), race (White, Black, or other), maternal education (<college or college/graduate school), study site (Boston, San Diego, or St. Louis), and trial intervention (high-dose vitamin D₃ or placebo). R Version 4.0.3 was used for the statistical analyses and a confidence level of 95% was applied.

RESULTS

Baseline characteristics

Of the 806 mother-child pairs in VDAART, 805 subjects had data on at least one cotinine measurement and of these subjects, 476 had successful lung function data at age 6 years and were included in the present post-hoc study (Figure 1). No differences were observed between the baseline characteristics of included and excluded VDAART mother-child pairs apart from study site distribution (P=0.019) (Table E1). The number of subjects with plasma cotinine levels at each time point was: 462 for 10-18 gestational weeks (GW), 462 for 32-38 GW, 305 for age 1 year, 300 for age 3 years, and 391 for age 6 years. Table 1 presents the baseline characteristics of the study subjects (N=476) stratified by TSE. Subjects with TSE, reflected by increased plasma cotinine level on at least 1 time point, had lower maternal education level (P<0.001), household income (P<0.001) and 25(OH)D levels (P<0.001), and were more often Black (P<0.001) and from the St. Louis study center (63% vs. 20%) than those without TSE. Children with TSE were also more often preterm (P=0.024) and had asthma (P=0.017) and allergic sensitization (P=0.035) at age 6 years than their peers, but these differences did not remain significant in logistic regression models adjusted for potential confounders (β 0.024, 95% confidence interval [95% CI] −0.022 to 0.071, P=0.306 for prematurity; β 0.031, 95% CI −0.057 to 0.119, P=0.488 for asthma; and β 0.005, 95% CI −0.114 to 0.125, P=0.933 for allergic sensitization).

Figure 1.

Flowchart of the study.

Table 1.

Baseline characteristics of the study subjects (n=476) stratified by detection of increased plasma cotinine levels^†.

Characteristic	No cotinine detected (n=271)	Increased cotinine at ≥1 time point (n=205)	P	Missing data (%)
Maternal education, N (%)			<0.001	0
< High school	13 (5)	48 (23)
High school or technical school	47 (17)	93 (45)
College or graduate school	211 (78)	64 (31)
Household income, N (%)			<0.001	0
< $50,000	83 (31)	114 (56)
≥ $50,000	142 (52)	24 (12)
Refused to answer or does not know	46 (17)	67 (33)
Medical history of the mother, N (%)
Asthma	109 (40)	92 (45)	0.355	0
Hay fever	188 (69)	129 (63)	0.168	0
Eczema	85 (31)	68 (33)	0.750	0
Vitamin D3 intervention, N (%)	138 (51)	103 (50)	0.957	0
25-hydroxyvitamin D level (ng/ml, mean (SD))
Mother at 10-18 gestational weeks	25.45 (9.41)	18.83 (8.94)	<0.001	1
Mother at 32-38 gestational weeks	35.49 (13.31)	29.82 (15.13)	<0.001	3
Child at age 6 years	31.22 (9.99)	24.72 (8.08)	<0.001	17
Child sex (male, N (%))	139 (51)	108 (53)	0.835	0
Child race, N (%)			<0.001	0
Black	67 (25)	165 (81)
White	128 (47)	26 (13)
Other	76 (28)	14 (7)
Gestational age <37 weeks at delivery, N (%)	6 (2)	14 (7)	0.024	0
Child allergic sensitization at age 6 years ‡ , N (%)	117 (55)	114 (66)	0.035	19
Child asthma at age 6 yrs. §, N (%)	37 (14)	46 (22)	0.017	0
Study site (%)			<0.001	0
Boston, MA	70 (26)	54 (26)
San Diego, CA	146 (54)	22 (11)
St. Louis, MO	55 (20)	129 (63)

^†Increased plasma cotinine level at any of the investigated time points: 10-18 gestational weeks, 32-38 gestational weeks, or at 1, 3, or 6 years after delivery.

^‡Serum specific IgE positive (≥35 kU/L) to ≥1 food or aeroallergen and/or total serum IgE level of ≥60 kU/1 at the age of 3 and/or 6 years.

^§Defined as physician-diagnosed asthma at any time from birth to age 6 years and a report of asthma medication use or wheeze after age 5 years.

Questionnaires vs. cotinine measurements

Table E2 illustrates the number of subjects with TSE based on the questionnaires and cotinine levels at each time point. Reported household smoking was associated with significantly increased cotinine levels at all time points, with the most prominent effects seen among the youngest children (ages 1 and 3 years) (Figure 2). However, high cotinine levels were also detected in subjects without reported household smoking, indicating possible unrecognized sources or inaccurate reporting of TSE. Active maternal smoking at 32-38 GW was also associated with significantly increased cotinine levels in the mother (Figure E1).

Figure 2.

Association between reported secondhand tobacco smoke exposure and plasma cotinine level (in standard deviations [SD]) in A) mothers at 10-18 gestational weeks and in the children at ages B) 1, C) 3, and D) 6 years. Analyses were performed using Wilcoxon Rank Sum Test.

Smoke exposure and childhood lung function

Figure 3 illustrates the association between cumulative TSE, reflected by the repeated plasma cotinine measurements, and child lung function at age 6 years. Cumulative cotinine level was negatively associated with forced expiratory volume in 1 second (FEV₁) and forced vital capacity (FVC) in spirometry and positively associated with R5 in IOS – all indicating that TSE from pregnancy to childhood is associated with dose- and duration-dependent decreases in child lung function. All associations between cumulative cotinine level and child lung function remained significant when adjusting for child asthma status, allergic sensitization, and prematurity (Table E3). Race-stratified analyses revealed similar effect directions indicating decreasing child lung function with increasing cumulative plasma cotinine level in both Black and White subjects (Figure E2).

Figure 3.

Association between cumulative tobacco smoke exposure from pregnancy (10-18 and 32-38 gestational weeks) to childhood (ages 1, 3, and 6 years) and child lung function at age 6 years assessed with spirometry and impulse oscillometry. †Linear regression model adjusted for child height, sex, race, maternal education level, study site, and trial intervention. Abbreviations: FEV₁, forced expiratory volume in 1 second; FVC, forced vital capacity; and R5, respiratory resistance at 5 Hz.

Figure 4 demonstrates the differences in lung function between subjects with no cotinine detected, medium cotinine level, and high cotinine level throughout the follow-up (cumulative cotinine level of 0 for no cotinine, 0.5-1.49 for medium level, and ≥1.5 for high level on a scale from 0 to 2). We observed decreasing lung function with increasing cotinine level group with those in the high cotinine level group having 19 ml lower FEV₁ (16%, 1.04 L vs. 1.23 L, adjusted P=0.028), 22 ml lower FVC (16%, 1.17 L vs. 1.39 L, adjusted P=0.010), and 0.17 kPa/L/s higher R5 (20%, 1.03 kPa/L/s vs. 0.86 kPa/L/s, adjusted P<0.001) than those without increased cotinine levels during follow-up. Baseline characteristics for these cotinine level groups are presented in Table E4.

Figure 4.

Association between cumulative tobacco smoke exposure from pregnancy (10-18 and 32-38 gestational weeks) to childhood (ages 1, 3, and 6 years) and child lung function at age 6 years. A) Forced expiratory volume in 1 second (FEV1), B) forced vital capacity (FVC), and C) respiratory resistance at 5 Hz (R5). †Subjects were divided into three groups based on the cumulative cotinine level (scale 0-2): none (0/2), medium (0.5-1.49/2), or high (≥1.5/2). P-values indicate association between the 3-level TSE group and child lung function in a linear regression model adjusted for covariates.

When the study subjects were stratified based on TSE timing to only prenatal (n=23), only postnatal (n=69), and both pre- and postnatal (n=111) exposure, all groups showed a trend towards decreased lung function with TSE, but statistically significant association was only observed between pre- and postnatal TSE and FEV₁ (Figure E3).

Vitamin D and tobacco smoke exposure

In total 189 study subjects had insufficient (<30 ng/ml) gestational 25(OH)D levels at both time points (10-18 and 32-38 GW), 76 had sufficient (≥30 ng/ml) levels at both time points, and 194 had sufficient level at one time point and insufficient at the other. No association between cumulative cotinine level and child lung function was observed among subjects with sufficient 25(OH)D level throughout pregnancy, whereas significant associations demonstrating decreasing lung function with increasing cotinine level were observed among those with insufficient 25(OH)D levels at both time points, and nonsignificant trends were seen among subjects with sufficient and insufficient levels (Figure 5). A significant interaction between cumulative cotinine level and gestational 25(OH)D status was observed with R5 but not with the spirometry indices (Table E5). Interaction between continuous 25(OH)D level at different time points during pregnancy and cumulative cotinine level was only significant for baseline 25(OH)D level and R5 (Table E6).

Figure 5.

Association between cumulative tobacco smoke exposure from pregnancy to childhood and child lung function at age 6 years among subjects with A) high (≥30 ng/ml), B) high (≥30 ng/ml) and low (<30 ng/ml), and B) low (<30 ng/ml) maternal plasma 25-hydroxyvitamin D (25[OH]D) levels during pregnancy. Cotinine level was measured in mothers at 10-18 and 32-38 gestational weeks, and in children at ages 1, 3, and 6 years. †Linear regression model adjusted for child height, sex, race, maternal education level, study site, and child 25(OH)D level at age 6 years. Abbreviations: FEV₁, forced expiratory volume in 1 second; FVC, forced vital capacity; and R5, respiratory resistance at 5 Hz.

When the study subjects were stratified by trial intervention, a statistically significant association between cumulative cotinine level and all lung function indices was observed among the placebo arm, whereas the high-dose prenatal vitamin D₃ supplementation arm only showed significant associations with R5 (Figure E4). However, interaction between the trial intervention and cumulative cotinine level did not reach statistical significance (Table E7).

DISCUSSION

Our study indicates that cumulative TSE, reflected by increased plasma cotinine levels at multiple time points during pregnancy and childhood, is associated with dose- and duration-dependent decreases in child lung function at age 6 years. Recent global health reports reiterate the enormous public health burden of tobacco use,¹ and our study demonstrates the negative implications of TSE in a vulnerable population. The majority of reported TSE in this study were secondhand, indicating that exposure to secondhand sources can have detrimental effects on lung function outcomes in early childhood. Additionally, our findings suggest a novel potential modulator of the effects with therapeutic potential: gestational vitamin D status, and that adequate gestational vitamin D status may confer resilience against early life TSE.

The effects of TSE on lung development have been well-studied,^32-35 and prenatal TSE, specifically, has been shown compromise fetal lung development and dysregulate genes involved in airway remodeling.³⁶ The exact mechanisms are not fully understood and appear to vary based on genetic and environmental factors. Active maternal smoking increases the risk of offspring comorbidities,³⁷ but even passive maternal TSE can affect fetal development and outcomes, although generally observed to a lesser degree than active maternal smoking.³⁸ Delineating the specific impacts of postnatal exposure to tobacco smoke is challenging because children are often exposed both prenatally and postnatally. However, early childhood TSE has been associated with poor pulmonary outcomes, independent of prenatal exposure,³⁹ which highlights the continued risk to respiratory health even if parents abstain from smoking during pregnancy. Postnatal TSE can show effects on airway inflammation and remodeling similar to prenatal exposures,⁴⁰ and the systemic inflammatory effects produced by secondhand smoke can persist >3 times the duration of exposure resulting in chronic low-grade systemic inflammation in children exposed to tobacco smoke daily.⁴¹ We focused on cumulative TSE to better understand the effects of longitudinal exposure to mainly secondhand smoking and observed a significant association between objectively measured TSE from pregnancy to childhood and child lung function at age 6 years, supporting that both pre- and postnatal exposure to tobacco smoke contribute to the detrimental effects on child respiratory health. In line with this finding, our analyses stratified by TSE timing showed nonsignificant trends towards decreased lung function among subjects with either pre- or postnatal TSE and a statistically significant association among subjects with both pre- and postnatal TSE. However, this analysis was limited by low number of subjects with only pre- or postnatal TSE.

Self-reported measures of tobacco smoke are susceptible to biases and can vastly underestimate actual exposures to tobacco smoke.¹⁵ This is in line with our findings on the association between reported household smoking and concurrent plasma cotinine levels: even though as a group, subjects who reported TSE had higher plasma cotinine levels, high cotinine levels were also seen among subjects without self-reported exposures. Therefore, we used cotinine, which can be considered as an objective measure of both primary and secondary exposures from any nicotine delivery,¹⁶ in our main analyses. Doing so, we found that many of the mothers had levels consistent with active smoking rather than secondhand exposures. Using data from cotinine measurements at multiple time points throughout pregnancy and childhood and accounting for the level of exposure, we were able to detect significant duration- and dose-dependent effects of TSE on child lung function.

The most novel finding of our study is the potential modulatory effect of gestational vitamin D status on lung susceptibility to TSE. Vitamin D deficiency has been associated with decreased lung function and a more rapid lung function decline in adult smokers.²⁰ Coupled with the findings of our present study, this suggests that maintaining sufficient vitamin D levels throughout life could protect against continued exposures during the lifespan. The modifying effects of vitamin D to other detrimental environmental exposures have also been reported in both animal and human studies.^42-44 The mechanisms behind this effect modification are unclear but may be related to the immunomodulatory, anti-inflammatory, and antioxidant effects of vitamin D.^19,20 Tobacco smoke also decreases the production of 1,25-dihydroxyvitamin D in lung epithelial cells,⁴⁵ and higher serum levels of vitamin D could potentially overcome this effect. Furthermore, vitamin D influences fetal lung growth and development.^21-23

We previously observed that gestational vitamin D levels were significantly associated with offspring lung function in VDAART and a level of ≥30 ng/ml appeared optimal for offspring respiratory health.^26,46,47 In the present study, we hypothesized that in utero vitamin D deficiency leads to altered lung development and as a result, increased susceptibility to TSE both pre- and postnatally. Supporting this hypothesis, we observed no significant association between TSE and child lung function in mothers with sufficient vitamin D levels throughout pregnancy whereas a significant effect was seen among those with insufficient levels. However, the interaction between TSE and gestational vitamin D insufficiency reached statistical significance only with R5, suggesting that IOS might be more sensitive than spirometry in detecting early lung function impairment related to TSE. Small airways are thought to be the primary site of pathological response to tobacco smoke⁴⁸ and therefore, the superior sensitivity of R5 might be related to its potential to detect dysfunction of both large and small airways as compared to FEV₁ and FVC which mainly reflect large and medium-sized airways.³⁰ Furthermore, although the associations between TSE and lung function were stronger among the placebo arm than in the vitamin D₃ supplementation arm in the stratified analyses, the interaction between trial intervention and TSE was not significant. Coupled with the finding that maternal baseline but not third trimester vitamin D level showed a significant interaction with TSE for R5, our findings suggest that vitamin D sufficiency from early pregnancy (at or before 10-18 gestational week) may be required to reach the optimal protective effects of vitamin D against TSE. Because VDAART only supplemented the mothers during pregnancy, it would be of interest to see whether additional vitamin D supplementation in childhood would further protect against the effects of secondhand TSE on childhood lung function.

This study has several limitations. First, the nature of relative quantification of global metabolomics prevented us from establishing an exact level of plasma cotinine for clinical translation. Thus, a study that measures specific cotinine levels is warranted. Furthermore, the half-life of cotinine in plasma is approximately 15-19 hours and thus, our cotinine measures may be representative of shorter-term exposure to smoking. Nevertheless, our comprehensive analyses of cumulative TSE during pregnancy and/or early childhood still showed significant impacts on child lung function and airway resistance. The multiple cotinine measurements throughout the study period likely attenuated the limitation of short-term representation of these measures. Furthermore, in line with previous studies,¹⁵ only approximately half of the study subjects with increased plasma cotinine levels reported maternal or other household smoking in the questionnaires. Failure to identify TSE by the questionnaires is likely multifactorial including exposure sources not specified by the questionnaires such as workplace exposure, neighbor/friends, or urban environment as well as recall bias and social pressure. Second, the analyses stratified by gestational vitamin D status and the interaction models were limited by sample size and therefore, further studies are needed to verify the effect modulatory potential of gestational vitamin D and to determine the exact mechanism behind this effect. Nevertheless, despite the safe and inexpensive nature of gestational vitamin D supplementation, vitamin D insufficiency is extremely common among pregnant women. Our findings support the importance of ensuring vitamin D sufficiency throughout pregnancy, and coupled with our prior findings in adults, throughout the lifespan. Third, we did not have measures of infant respiratory function. However, since lung function tracks from birth throughout life,^49-51 we would likely see effects also with those measures. Future studies should consider infant respiratory function measurements. Lastly, loss to follow-up and missing data should be acknowledged as potential sources of bias in longitudinal studies even though we did not observe significant differences in child and parental baseline characteristics between the study subjects and those excluded except for study site distribution. Furthermore, although our analyses were adjusted for different baseline characteristics, the possibility of residual confounding should be acknowledged.

In conclusion, with comprehensive multiple time-point follow-up data including detailed questionnaires and objective blood marker of TSE, cotinine, we identified a dose- and duration-dependent association between TSE from pregnancy to childhood and child lung function at age 6 years even in the absence of reported maternal smoking during pregnancy. Our findings also indicate a novel potential effect modulator with therapeutic potential: gestational vitamin D status, and suggests that adequate vitamin D status may confer resilience against tobacco smoke exposure. These finding highlight the importance of preventing exposure to even small amounts of tobacco smoke during early life and ensuring adequate vitamin D levels throughout pregnancy to prevent early lung function impairment.

Clinical Implications:

Early life exposure to secondhand smoking is associated with dose- and duration-dependent decreases in child lung function which appear to be modulated by gestational vitamin D status.

Abbreviations:

25(OH)D

25-hydroxyvitamin D

CI

Confidence interval

FEV₁

Forced expiratory volume in 1 second

FVC

Forced vital capacity

GW

Gestational week

IOS

Impulse oscillometry

R5

Respiratory resistance at 5 Hz

TSE

Tobacco smoke exposure

VDAART

Vitamin D Antenatal Asthma Reduction Trial