Abstract
Background Harmful effects of prenatal tobacco exposure and possible interaction with 17q12-21 genetic variants have been shown for some asthma outcomes in childhood, whereas findings related to allergy outcomes are more inconsistent. This study aimed to examine the effect of prenatal tobacco exposure and relation to 17q12-21 genotype on a wide array of asthma and allergy-related outcomes in early childhood.
Methods Prenatal tobacco exposure was determined by maternal smoking during the third trimester (yes/no) in 411 children from the phenotyped Copenhagen Prospective Studies on Asthma in Childhood 2000 (COPSAC2000) birth cohort with clinical follow-up to age 7 years. The rs7216389 single nucleotide polymorphism was used as main representative of the 17q12-21 locus. Asthma end-points included asthma diagnosis, exacerbations, episodes with troublesome lung symptoms and lower respiratory tract infections, spirometry, plethysmography, bronchial responsiveness to methacholine, exercise and cold dry air. Allergy-related endpoints included aeroallergen sensitisation, allergic rhinitis, fractional exhaled nitric oxide, blood eosinophil count and urine eosinophil protein X levels. Statistical analyses were done using Cox regression, linear regression, logistic regression and quasi-Poisson regression.
Results Prenatal tobacco exposure increased the risk of asthma (adjusted hazard ratio (aHR) 2.05, 95% CI 1.13–3.73; p=0.02), exacerbations (aHR 3.76, 95% CI 2.05–6.91; p<0.001), number of LRTIs (adjusted incidence rate ratio 1.87, 95% CI 1.34–2.55; p<0.001), and was associated with decreased spirometry indices (forced expiratory volume in 1 s (FEV1) adjusted mean difference (aMD) −0.07 L, 95% CI −0.13– −0.005 L, p=0.03; maximal mid-expiratory flow aMD −0.19 L·s−1, −0.34– −0.04 L·s−1, p=0.01) and increased bronchial responsiveness to methacholine (provocative dose of methacholine causing a 20% drop in FEV1 adjusted geometric mean ratio 0.55, 95% CI 0.31–0.96; p=0.04). In contrast, there was no association with any allergy-related end-points. The effect on asthma depended on 17q12-21 genotype with an increased risk only among children without risk alleles.
Conclusion Prenatal tobacco exposure was associated with asthma dependent on 17q12-21 genotype and with exacerbations, lung function and bronchial responsiveness, but not with any allergy-related outcomes. This suggests that tobacco exposure in utero leads to adverse lung developmental/structural effects rather than susceptibility to develop allergy and type 2 inflammation.
Abstract
Prenatal tobacco exposure was associated with asthma, exacerbations and impaired lung function, but not allergy. The effect on asthma depended on 17q12-21 genotype with an increased risk only among children without risk alleles. https://bit.ly/3y49BKK
Introduction
In most previous studies, prenatal tobacco exposure has been associated with an increased risk of developing childhood asthma [1–5], asthma exacerbations [3, 6] and decreased lung function at school age [1, 4–6], whereas findings related to the risk of developing allergy-related outcomes [2, 7–11] and atopic dermatitis [10, 12–14] are more inconsistent. There is a lack of longitudinal studies investigating the effect of prenatal tobacco exposure on multiple asthma and allergy-related end-points in the same population, which may shed light on whether the harmful in utero effects are associated with adverse lung developmental/structural effects and or/and susceptibility to develop allergy and type 2 inflammation. Furthermore, a gene–environment interaction between the 17q12-21 genotype and tobacco exposure has been suggested to increase the risk of asthma in one study investigating prenatal tobacco exposure [15] and in two studies investigating tobacco exposure in early childhood [16, 17]. However, whether such interaction is present for intermediary asthma traits such as forced flows, airway resistance and bronchial hyperresponsiveness and for allergy-related traits and type 2 inflammation remains to be elucidated.
In this study, we utilise the detailed clinically phenotyped Copenhagen Prospective Studies on Asthma in Childhood 2000 (COPSAC2000) birth cohort [18] to investigate the effects of prenatal tobacco exposure and relation to 17q12-21 genetic variants on development of multiple asthma and allergy-related traits in the first 7 years of life. The combination of objective assessments of tobacco exposure in utero, 17q-12-21 genotype data and a broad array of longitudinally collected asthma and allergy outcomes from birth in the same population makes this study a unique contribution to the literature on harmful effects of prenatal tobacco exposure on respiratory outcomes in childhood.
Methods
Study population
COPSAC2000 is an ongoing, prospective birth cohort of 411 children born to mothers with a history of asthma. Enrolment took place during the first month of life, excluding children with severe congenital abnormality, gestational age <36 weeks or lung symptoms prior to enrolment. The cohort has been described in detail previously [18, 19].
Ethics
The study was conducted in accordance with the guiding principles of the Declaration of Helsinki and was approved by the local ethics committee (KF 01-289/96) and the Danish Data Protection Agency (2008-41-1754). Both parents gave oral and written informed consent before enrolment.
Prenatal tobacco exposure
Prenatal tobacco exposure was defined as maternal smoking during the third trimester (yes/no) and was determined by parental interviews at enrolment of the child at the COPSAC clinic [20]. In addition, we measured dry blood spot cotinine level by age 1–12 days and postnatal tobacco exposure by hair nicotine level by age 1 year [21].
Asthma-related outcomes
Asthma
Asthma was solely diagnosed by the COPSAC paediatricians according to international guidelines such as those from the Global Initiative for Asthma, as detailed previously [22], based on a pre-defined, validated algorithm using repeated symptoms and response to treatment: 1) recurrent episodes of troublesome lung symptoms; 2) symptoms typical of asthma; 3) response to intermittent rescue use of inhaled β2-agonist; and 4) response to a 3-month course of inhaled corticosteroids (ICS) (budesonide, 400 μg·day−1) and relapse when stopping treatment (supplementary material) [22, 23]. Asthma at age 0–7 years was used as the main outcome. Asthma at age 7 years was defined as having ICS-dependent asthma between the sixth and eighth birthday.
As secondary asthma end-points, we examined persistent and transient asthma, allergic and nonallergic asthma, and T2-high and T2-low asthma (supplementary material).
Exacerbations
Exacerbations at age 0–7 years were defined from the need for oral prednisolone (1 mg·kg−1 for 3 days) or high-dose ICS (budesonide 1600 μg·day−1 for 14 days) prescribed by the COPSAC paediatricians, or need for hospitalisation [23].
Troublesome lung symptoms
Troublesome lung symptoms were explained to the parents as wheeze or whistling sounds, breathlessness or recurrent troublesome cough severely affecting the wellbeing of the child, and were registered in a daily diary from birth, as described by Bisgaard et al. [22]. The diaries were reviewed with the parents at the visits [19, 22]. An episode was defined as three consecutive days with symptoms [22]. Only children with >90% of days recorded in the diary were included.
Lower respiratory tract infections
Lower respiratory tract infections (LRTIs) at age 0–3 years included a clinical diagnosis of bronchiolitis or pneumonia [24].
Lung function and bronchial responsiveness
Forced expiratory volume in 1 s (FEV1), forced vital capacity (FVC) and maximal mid-expiratory flow (MMEF) were measured by spirometry and specific airway resistance (sRaw) was measured using whole-body plethysmography. Bronchial responsiveness to cold dry air, defined as the percentage change from baseline in sRaw and FEV1, was measured at 6 years. The provocative dose of methacholine causing a 20% drop in FEV1 (PD20) was estimated from the dose–response curves at age 6.5 years. Bronchial responsiveness to exercise was defined as the maximum percentage decline in FEV1 from baseline within 10 min after exercise at age 7 years [25]. Quality control of the spirometry measurements are described in the supplementary material.
Neonatal lung function (forced expiratory volume in the first 0.5 s (FEV0.5) and forced expiratory flow at 50% of forced vital capacity (FEF50)) and bronchial reactivity to methacholine (PD15) were measured at 1 month of age [20, 25].
Allergy-related outcomes
Allergic sensitisation
Allergic sensitisation to common aeroallergens was determined at age 6 years by skin-prick test (ALK-Abelló, Copenhagen, Denmark), defining a positive test as a mean wheal diameter of 3 mm or larger than the negative control, and/or by specific-IgE measurements (ImmunoCAP; Phadia AB, Uppsala, Sweden) using a cut-off at ≥0.35 kU·L−1 [26] (supplementary material).
Rhinitis
Rhinitis was diagnosed by the COPSAC paediatricians based on parental interviews on symptoms in the child's seventh year of life, defined by troublesome sneezing or blocked or runny nose, severely affecting the wellbeing of the child in periods without common cold or flu. Allergic rhinitis was diagnosed in children with sensitisation to aeroallergens clearly related to the symptomatic periods [27].
Fractional exhaled nitric oxide, blood eosinophil count and eosinophil protein X
Fractional exhaled nitric oxide (FeNO) was measured at age 7 years using an online technique and blood samples were analysed for eosinophil count at 6 months and 7 years of age [27]. Eosinophil protein X (EPX) was measured in urine samples collected at ages 1 and 6 months using a double-antibody immunoassay [28].
Atopic dermatitis
Atopic dermatitis at age 0–7 years was diagnosed by COPSAC paediatricians with clinical assessment according to Hanifin and Rajka's diagnostic criteria based on the presence of three out of four major criteria and at least three out of 23 minor criteria [29].
Genotyping
The rs7216389 single nucleotide polymorphism (SNP) located on chromosome 17q12-21 was genotyped by allelic discrimination [23]. We chose rs7216389, with T being the risk allele, as main representative of the 17q12-21 locus to adhere with previous COPSAC studies [23, 30] and because it has been strongly associated with early-onset asthma [31]. We also included two additional SNPs in the 17q12-21 region, rs12936231 and rs2305480, with C and G being the risk alleles, respectively (supplementary material).
Covariates
Exposures known to be associated with tobacco smoking, risk of asthma or both, were included as covariates. These included birth by caesarean section (yes/no), older children living in the child's home at birth (yes/no), social circumstances, duration of exclusive breastfeeding, age at start of daycare, household cat ownership at birth (yes/no), exposure to passive smoke during childhood (yes/no), mother's sensitisation, fine particulate matter (PM2.5), blackness of the particulate matter (“black smoke”), neonatal lung function (FEV0.5, FEF50) and neonatal bronchial responsiveness (PD15) (supplementary material).
Statistical analysis
Age of onset of asthma, exacerbations and atopic dermatitis were analysed using Cox proportional hazard regression models computing hazard ratios (HR). The proportionality of the hazards across time was checked using Kaplan–Meier curves. Maternal smoking in the third trimester was used as dichotomous exposure, whereas cotinine and nicotine were used as a continuous exposure. In the Cox proportional hazard regression model examining the risk of exacerbations, the effect modification by history of LRTI was investigated in an interaction model by adding cross-products to the model. Lung function measures were calibrated for height, sex and age. The effect of prenatal tobacco exposure on lung function, FeNO, blood eosinophil count and EPX were analysed using linear regression computing estimates of mean difference (MD) and geometric mean ratios (GMR). Asthma, allergic rhinitis and allergic sensitisation were analysed by logistic regression computing odds ratios (OR). Number of episodes with troublesome lung symptoms, exacerbations and LRTIs were analysed by quasi-Poisson regression computing incidence rate ratios (IRR). The quasi-Poisson model was used to account for overdispersion in the data (supplementary material).
Effect modification by 17q12-21 genotype was investigated in genotype-stratified analyses and interaction models by adding cross-products to the models alongside main effects of each term.
Missing data were not included in the analyses and we did not use imputation for missing values. All analyses were adjusted for the confounders caesarean section, older children living in the child's home at birth, social circumstances, duration of exclusive breastfeeding, age at start of daycare and household cat ownership at birth. Analyses of urine EPX were further adjusted for creatinine excretion [28]. As we showed previously that prenatal tobacco exposure is associated with decreased lung function at age 1 month [20], in addition we adjusted FEV1, MMEF and PD20 for neonatal lung function. We also previously showed that cat exposure in early life decreases asthma risk from the 17q12-21 high-risk variant [32], and therefore adjusted our analysis for risk of asthma and exacerbations for household cat at birth.
Statistical analyses were performed in R project 4.0.0 (www.R-project.org) using a significance level of p<0.05.
Results
Baseline characteristics
Information on maternal smoking during the third trimester was available for all 411 children, showing that 63 (15.3%) were exposed. Dry blood spot cotinine level was analysed in 388 (94.4%) and hair nicotine level in 368 (89.5%) children. All three measures of tobacco exposure were highly correlated (supplementary figure E1).
Children exposed versus not exposed to maternal smoking during pregnancy had lower birth length and weight, were exclusively breastfed for a shorter time, had lower social circumstances, were more often exposed to passive smoke during childhood and more often had household cat ownership at birth (supplementary table E1).
Prenatal tobacco exposure and asthma-related outcomes
Asthma
The risk of developing asthma at age 0–7 years was higher in exposed children versus those who were not exposed (adjusted (a)HR 2.05, 95% CI 1.13–3.73; p=0.02) (table 1, figure 1), which was also significant after adjusting for passive smoke exposure in childhood, PM2.5, black smoke and mother's sensitisation (supplementary table E2). Higher dry blood spot cotinine level at birth was also associated with an increased risk of asthma: HR per doubling of cotinine level (1.18, 95% CI 1.01–1.37; p=0.03) (supplementary figure E2), whereas hair nicotine level at age 1 year was not (HR per doubling nicotine level 1.06, 95% CI 0.95–1.18; p=0.31) (supplementary table E3).
Prenatal tobacco exposure and asthma end-points
Kaplan–Meier curve showing the risk of asthma during the first 7 years of life for children with and without prenatal tobacco exposure.
Current asthma at age 7 years was diagnosed in 26.5% of the exposed children, compared to 15.6% of children who were not exposed (aOR 2.53, 95% CI 0.98–6.21; p=0.047) (table 1).
We found no significant difference on the effect of prenatal tobacco exposure on persistent versus transient asthma, allergic versus non-allergic asthma or T2-high versus T2-low asthma (supplementary table E4).
Exacerbations
The risk of exacerbations at age 0–7 years was increased in children exposed versus not exposed (aHR 3.76, 95% CI 2.05–6.91; p<0.001) (table 1, figure 2), which was also significant after adjusting for passive smoke in childhood, PM2.5, black smoke and mother's sensitisation (supplementary table E2). Higher cotinine level was associated with an increased risk of exacerbations (HR 1.25, 95% CI 1.07–1.47; p<0.01) (supplementary figure E3), which was also the case for hair nicotine level (HR 1.18, 95% CI 1.05–1.32; p<0.01) (supplementary table E3).
Kaplan–Meier curve showing the risk of exacerbations during the first 7 years of life for children with and without prenatal tobacco exposure.
The number of exacerbations was higher in children exposed versus not exposed (median 0 (IQR 0–1) versus 0 (0–0), aIRR 2.41, 95% CI 0.99–5.29; p=0.04) (table 1).
Restricting the analyses to children with a diagnosis of asthma (n=72) showed a trend of increased risk of exacerbations in exposed children (HR 1.79, 95% CI 0.92–3.50; p=0.09). The effect of prenatal tobacco smoke on exacerbations was not significantly dependent on history of LRTI (p-interaction=0.24).
Episodes with lung symptoms
The number of LRTIs was higher in children exposed versus not exposed (median 2 (IQR 0.5–3.0) versus 1 (0–2), aIRR 1.87, 95% CI 1.34–2.55; p<0.001), but there was no significant difference for episodes with troublesome lung symptoms (median 10 (IQR 5.25–13.50) versus 6 (2–11), aIRR 1.13, 95% CI 0.83–1.52; p=0.43) (table 1).
Lung function and bronchial responsiveness
Prenatal tobacco exposure was associated with decreased FEV1 (aMD −0.07 L, 95% CI −0.13– −0.005; p=0.03), decreased MMEF (aMD −0.19 L·s−1, 95% CI −0.34– −0.04; p=0.01) and increased bronchial responsiveness to methacholine, i.e. lower PD20 (aGMR 0.55, 95% CI 0.31–0.96; p=0.04). Adjusting FEV1, MMEF and PD20 for neonatal lung function did not change the results (supplementary table E5). There was no association with FVC, sRaw or responsiveness to exercise or cold air (table 1).
Prenatal tobacco exposure and allergy-related outcomes
Aeroallergen sensitisation and allergic rhinitis were diagnosed in 33% and 13% of children exposed to prenatal tobacco smoke versus 29% and 18% of children not exposed, respectively, showing no significant differences. Furthermore, there was no association with FeNO level, blood eosinophil count or urine EPX levels (table 2).
Prenatal tobacco exposure and allergy-related and skin end-points
Prenatal tobacco exposure and atopic dermatitis
Atopic dermatitis was diagnosed in 32% of children exposed to prenatal tobacco smoke versus 45% of children not exposed, showing no significant differences (table 2).
Effect modification by 17q12-21 genotype
The SNP rs7216389 was genotyped in 388 (94.4%) of the 411 children with a genotype distribution of TT 29.4%, CT 47.9% and CC 22.7% (asthma risk allele: T). We investigated interaction with 17q12-21 genotype for the end-points, where a main effect of prenatal tobacco exposure was observed. In children with rs7216389 wild-type CC genotype, prenatal tobacco exposure increased the risk of asthma (HR 4.25, 95% CI 1.42–12.73; p<0.01), whereas there was no increased risk among children with high-risk TT genotype (HR 1.00, 95% CI 0.38–2.63; p=1.00, p-interaction=0.048) (figure 3). There was no interaction for other asthma-related end-points, including exacerbations, lung function and bronchial responsiveness (table 3). Using dry blood spot cotinine or hair nicotine as tobacco exposure measures showed similar results (supplementary table E6). Using rs12936231 and rs2305480 genotypes also showed similar results (supplementary table E7).
Kaplan–Meier curves showing the risk of asthma during the first 7 years of life for children with and without prenatal tobacco exposure, stratified by 17q12-21 genotype (rs7216389).
Prenatal tobacco exposure and risk of asthma end-points by 17q12-21 (rs7216389) genotype
Adjusting the Cox regression analysing the risk of asthma and exacerbations for household cat at birth did not change the results (supplementary table E8).
Discussion
Primary findings
In our Danish COPSAC2000 birth cohort, children with prenatal tobacco exposure had a doubled risk of developing asthma, tripled risk of exacerbations, a doubled risk of LRTIs, reduced lung function and increased bronchial responsiveness by 7 years compared to children not exposed. In contrast, there was no association between prenatal tobacco exposure and risk of allergy, type 2 inflammation markers or atopic dermatitis. The effect on asthma depended on 17q12-21 genotype with an increased risk only among children without risk alleles, whereas there was no interaction for the other asthma or allergy-related end-points.
Strength and limitations
The major strength of the study is the prospective monitoring of asthma, exacerbations, LRTIs, lung function, bronchial responsiveness, allergy, type 2 inflammation markers and atopic dermatitis in the first 7 years of life [18]. The children were seen for clinical evaluation and lung function assessment at 1 month of age, with repeated scheduled half-yearly assessments until age 7 years (16 visits) in conjunction with acute care visits upon onset of respiratory symptoms. Furthermore, troublesome lung symptoms were registered prospectively by the parents in daily diaries from birth, thereby diminishing recall bias [18, 33].
Children with asthma were exclusively diagnosed, treated and followed as outpatients by the COPSAC paediatricians according to a pre-determined validated quantitative symptom algorithm, response to ICS treatment and relapse after discontinuation [34]. This approach makes the COPSAC cohort unique compared to other studies using questionnaire-based diagnoses or diagnoses made by local doctors, which are much more heterogeneous, although there is still a risk for over- and under-diagnosing asthma in children [35, 36]. Having all these asthma and allergy-related traits in the same population combined with objective measures of prenatal tobacco exposure and genotyping data made it possible to explore the effects of prenatal tobacco exposure and relation to 17q12-21 genotype on a wide array of different, longitudinally collected end-points, which is an important contribution to the literature on harmful effects of tobacco exposure on respiratory outcomes in childhood.
Prenatal tobacco smoke exposure was determined by parental interview data, but as this may be subject to recall bias and under-reporting, we included objective measures of cotinine and nicotine [20, 21] and parental information about passive smoke exposure during childhood. Measurements of nicotine in hair at age 1 year is a measure of postnatal tobacco exposure, but were highly correlated to maternal smoking during the third trimester of pregnancy. Information on a wide range of other environmental exposures known to be associated with tobacco smoking, risk of asthma or both is another advantage. This allowed robust confounder adjustments, which did not change the findings. The wide range of different lung function assessments is one of the strengths of our study, as it makes it possible to address how prenatal tobacco expose is associated with lung volumes, airway resistance and response to different provocation tests. However, it is an inherent limitation of studies of young children that the asthma diagnosis cannot be measured objectively by lung function tests, and therefore must rely on symptoms and response to treatment. Current European Respiratory Society guidelines recommend not diagnosing asthma in children aged 5–16 years based on clinical history alone or a single abnormal objective test. The guidelines recommend using the objective measurements of spirometry, bronchodilator reversibility testing and FeNO as first-line tests when diagnosing asthma in children at this age [37]. As our cohort includes children up to 7 years of age, it may be a limitation that the diagnosis of asthma in children aged >5 years, i.e. 5–7 years, was based on clinical history and response to treatment instead of objective tests.
Compared to other ongoing birth cohorts, our cohort is small with 411 children, which may be a limitation of this study. Hence, power issues might be a challenge where weak, but true, estimates are at play, e.g. for the allergy-related outcomes. The COPSAC2000 cohort of children with high risk of asthma and the limited ethnic variation diminish the generalisability of our findings. Loss to follow-up might be influenced by the child's respiratory health, i.e. fewer losses to follow-up with asthma and allergy, and our results might be biased due to this.
Finally, studying gene–environment interactions is very complex and may in our case include genes outside the 17q12-21 locus. Therefore, unmeasured gene–environment interactions may be a limitation of our study. Additionally, there might be unmeasured confounders that should have been adjusted for.
Interpretation
We found that prenatal tobacco exposure was associated with an increased risk of asthma, exacerbations, LRTIs, reduced lung function and increased bronchial responsiveness in the first 7 years of life. These findings align with previous studies showing that prenatal tobacco exposure is associated with either an increased risk of childhood asthma [2–4], exacerbation risk [3, 6] or decreased lung function at school age [4, 6], but our study is the first to analyse the effect on a wide array of asthma outcomes, i.e. symptom load, respiratory infections, severe exacerbations, forced flows, airway resistance and bronchial responsiveness in the same population.
Interestingly, prenatal tobacco exposure was associated with bronchial responsiveness to methacholine, but not with responsiveness to exercise or cold-air challenge, which may be due to the differences between the challenges, with the former being direct and the latter two being indirect provocation tests. It has been suggested that prenatal tobacco exposure impairs airway development and/or lung elastic properties in utero [38], leading to lung function deficits and an increased risk of asthma. The possible mechanisms have been studied in guinea pigs [39] and mice [40, 41], showing that tobacco exposure during pregnancy altered gene expression involved in lung development, leading to airway remodelling and increased bronchial responsiveness to methacholine in the offspring. Additionally, a review [1] including mice, rats, sheep and monkey studies proposed that airway remodelling and airway hyperresponsiveness are probably caused by prenatal tobacco exposure and further suggested that prenatal exposure to nicotine stimulated lung branching and dysanaptic lung growth, i.e. incongruence between growth of the lungs and the airways, leading to an increased number of small-diameter airways. Dysanaptic airway growth may lead to decreased forced expiratory flows, increased airway responsiveness and predisposition to airway disease. Children exposed to prenatal tobacco may suffer from dysanaptic airways rather than asthma as we found reduced forced flows and increased bronchial responsiveness to methacholine (direct test), but not increased responsiveness to exercise or cold air challenge (indirect tests) or elevated type 2 markers. Dysanaptic lung growth might occur in the absence of asthma, but it is earlier shown that dysanapsis is associated with worse disease severity in children suffering from asthma [42] and it is not unlikely that prenatal tobacco exposure is associated with both dysanaptic airways and asthma as suggested by our data. The lack of an association between prenatal tobacco exposure and responsiveness to exercise or cold-air challenge in our study could also be explained by these indirect tests' lower sensitivity compared to the direct methacholine challenge test and/or lack of power in these analyses.
In our study, there was no association between prenatal tobacco exposure and risk of developing aeroallergen sensitisation, allergic rhinitis or elevated type 2 inflammation markers, including FeNO, blood eosinophil count and urine EPX level. This suggests that prenatal tobacco exposure has adverse lung developmental/structural effects in utero, which may lead to diminished lung function and increased risk of asthma, whereas there are no adverse effects on allergy outcomes or type 2 inflammation in early childhood.
A systematic review of the effects of prenatal and childhood tobacco exposure identified studies showing both increased and reduced risk of allergic sensitisation, but only three studies examined prenatal tobacco exposure and showed a trend of reduced risk of sensitisation [8]. A cross-sectional study in 1714 Italian children aged 7–16 years, based on parental questionnaire and skin-prick tests, showed that maternal smoking during pregnancy was a risk factor for allergic sensitisation [9]. In contrast, a study of 3316 children from the BAMSE (Children, Allergy, Milieu, Stockholm, Epidemiology) study [11] showed no association between maternal smoking during pregnancy and sensitisation up to 16 years of age. The German Multicentre Allergy Study of 1314 children [7] and a study of 10 860 children from the Mechanisms of the Development of Allergy (MeDALL) birth cohort consortium [2] both showed that any maternal smoking during pregnancy was not associated with risk of allergic rhinitis, while the latter showed that maternal smoking of >10 cigarettes per day during the whole pregnancy period increased the risk. Lee et al. [10] found that prenatal exposure to maternal passive smoking, but not prenatal exposure to maternal active smoking, was associated with allergic rhinitis in Chinese children. One study has shown that prenatal tobacco exposure was not associated with FeNO at age 6 years [7], whereas no previous studies have examined the effect of prenatal exposure on blood eosinophil count or urine EPX levels in childhood.
We found no association between prenatal tobacco exposure and risk of developing atopic dermatitis in the first 7 years of life, which adds to the interpretation that prenatal tobacco exposure increases the risk of asthma, but not atopy-related end-points. This is in line with studies of Chinese children [10] and Swedish children in the BAMSE study [12]. However, Lee et al. [10] found an increased risk of atopic dermatitis among children exposed to passive smoking during childhood and speculated that the lack of association between active smoking during pregnancy and atopic dermatitis in offspring is due to a small number of smoking pregnant mothers. Two studies showed that prenatal tobacco exposure increased the risk of atopic dermatitis [13, 14], but these studies were performed in infants.
We observed that the harmful effect of prenatal tobacco exposure on risk of asthma development interacted with 17q12-21 genotype where only children with wild-type genotype had an increased risk. Interestingly, there was no such interaction for exacerbations, LRTIs, lung function or bronchial responsiveness. Prenatal tobacco exposure may increase the risk of asthma dependent on 17q12-21 genotype by altering the expression of genes regulating immune responses [31, 43, 44], whereas lung developmental/structural effects of tobacco exposure affecting lung function in childhood seems independent of 17q12-21 genotype, which fits well with previous studies showing no association between 17q12-21 genotype and lung function [23, 45].
Our 17q12-21 interaction finding for asthma is opposite to previous studies, which have shown the highest risk of asthma in children with high-risk genotype exposed to tobacco smoke [15–17]. However, two of the studies examined the effect of early-life tobacco exposure, not prenatal exposure [16, 17], which may account for the opposite findings, as tobacco smoke has irritant effects in childhood. The third study examined the effect of prenatal tobacco exposure in the Generation R Study (2438 mothers and children) and in the Prevention and Incidence of Asthma and Mite Allergy (PIAMA) cohort (2023 mothers and children) and only observed significant interaction with 17q12-21 genotype in a combined analyses of the cohorts [15].
The at-risk nature of the COPSAC2000 birth cohort might be another explanation for the contrasting findings, as mothers with a history of asthma might have different behaviour, environment, awareness of symptoms and use of medication. Furthermore, children of mothers who smoked in pregnancy were more likely to have cat. This may be important, as we previously showed that cat exposure in early life decreases asthma risk from the 17q12-21 high-risk variant [32], but adjusting the analysis for household cat did not change the results, supporting that this cannot explain our findings. Finally, our opposite findings could also be spurious due to the low numbers in our cohort.
Conclusion
Prenatal tobacco exposure was associated with an increased risk of asthma dependent on 17q12-21 variants, exacerbations, LRTIs, decreased lung function and increased bronchial responsiveness, but showed no association with risk of allergy outcomes or elevated type 2 inflammation markers during early childhood. These findings suggest that prenatal tobacco exposure predominantly has adverse lung developmental/structural effects in utero, leading to diminished lung function and increased risk of non-atopic asthma.
Supplementary material
Supplementary Material
Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.
Supplementary material ERJ-00453-2021.Supplement
Figure E1. Plots showing the correlation between tobacco exposure in third trimester and cotinine in dry blood spot (a), nicotine in hair at age 1 year (b) and correlation between nicotine in hair at age 1 year and cotinine in dry blood spot (c). ERJ-00453-2021.Figure_E1
Figure E2. Kaplan-Meier curve showing risk of asthma during the first 7 years of life for children with high versus low cotinine in dry blood spot. ERJ-00453-2021.Figure_E2
Figure E3. Kaplan-Meier curve showing risk of exacerbations during the first 7 years of life for children with high versus low cotinine in dry blood spot. ERJ-00453-2021.Figure_E3
Figure E4. Kaplan-Meier curve showing risk of asthma during the first 7 years of life for children with different 17q12-21 (rs7216389) genotypes. ERJ-00453-2021.Figure_E4
Figure E5. Kaplan-Meier curve showing risk of exacerbations during the first 7 years of life for children with different 17q12-21 (rs7216389) genotypes. ERJ-00453-2021.Figure_E5
Shareable PDF
Supplementary Material
This one-page PDF can be shared freely online.
Shareable PDF ERJ-00453-2021.Shareable
Acknowledgements
We express our deepest gratitude to the children and families of the COPSAC2000 cohort study for all their support and commitment. We acknowledge and appreciate the unique efforts of the COPSAC research team.
Footnotes
This article has supplementary material available from erj.ersjournals.com
Governance: We are aware of and comply with recognised codes of good research practice, including the Danish Code of Conduct for Research Integrity. We comply with national and international rules on the safety and rights of patients and healthy subjects, including good clinical practice as defined in the EU's Directive on Good Clinical Practice, the International Conference on Harmonisation's good clinical practice guidelines and the Helsinki Declaration. Privacy is important to us which is why we follow national and international legislation on General Data Protection Regulation, the Danish Act on Processing of Personal Data and the practice of the Danish Data Inspectorate.
Author contributions: The guarantor of the study is H. Bisgaard, from conception and design to conduct of the study and acquisition of data, data analysis, and interpretation of data. H. Bisgaard had full access to all the data in the study and had final responsibility for the decision to submit for publication. R.B. Sunde wrote the first draft of the manuscript. R.B. Sunde, B. Chawes, J. Thorsen and C-E.T. Pedersen were responsible for data analysis, interpretation and writing the manuscript. K. Bønnelykke and J. Stokholm were responsible for interpretation and writing the manuscript. All co-authors have provided important intellectual input and contributed considerably to the analyses and interpretation of the data. All authors guarantee that the accuracy and integrity of any part of the work have been appropriately investigated and resolved and all have approved the final version of the manuscript. No honorarium, grant, or other form of payment was given to any of the authors to produce this manuscript.
Conflict of interest: R.B. Sunde has nothing to disclose.
Conflict of interest: J. Thorsen has nothing to disclose.
Conflict of interest: C-E.T. Pedersen has nothing to disclose.
Conflict of interest: J. Stokholm has nothing to disclose.
Conflict of interest: K. Bønnelykke has nothing to disclose.
Conflict of interest: B. Chawes has nothing to disclose.
Conflict of interest: H. Bisgaard has nothing to disclose.
Support statement: All funding received by COPSAC is listed at www.copsac.com. The Lundbeck Foundation (grant number R16-A1694); the Ministry of Health (grant number 903516); Danish Council for Strategic Research (grant number 0603-00280B) and the Capital Region Research Foundation have provided core support to the COPSAC Research Center. Funding information for this article has been deposited with the Crossref Funder Registry.
- Received February 12, 2021.
- Accepted June 21, 2021.
- Copyright ©The authors 2022. For reproduction rights and permissions contact permissions{at}ersnet.org
References
