In utero smoking exposure induces changes to lung clearance index and modifies risk of wheeze in infants

Fetal exposure to tobacco smoking throughout pregnancy is associated with wheezing in infancy. We investigated the influence of in utero smoking exposure on lung ventilation homogeneity and the relationship between lung ventilation inhomogeneity at 7 weeks of age and wheezing in the first year of life.


| INTRODUCTION
2][3] A meta-analysis of 10 cohort studies found that smoking during pregnancy was associated with wheezing in infancy (OR: 1.50, 95% CI: 1.27-1.77,p = <.01). 4Mechanistically adverse respiratory outcomes are thought to be the consequence of the direct effects of nicotine, carbon monoxide, and other harmful chemicals on oxygen and nutrient delivery, oxidative stress responses and epigenome reprogramming in the developing fetus via promoting airways remodeling, reactivity and inflammation. 2spite this evidence and intensive health education, at least 15%-20% of mothers do not quit smoking during pregnancy. 5ere is also strong evidence to demonstrate that maternal smoking during pregnancy has a negative effect on childhood respiratory health.An association between reduced infant lung function which tracks throughout life, 6 and the development of wheezing and asthma in childhood has been demonstrated. 7Even though it is highly plausible, it remains to be determined whether changes in infant lung function play a significant role in the relationship betweenin utero tobacco smoke exposure and later wheeze.This is relevant because infant lung function may be employable to more precisely predict the risk of future respiratory symptom development in babies at risk as a consequence of adverse environmental exposures.Furthermore, the effect of in utero tobacco exposure on the baby's gas mixing within the lung during the first weeks of life has not been determined.

The lung clearance index (LCI) is a parameter derived from
Multiple Breath Washout (MBW), which evaluates the unevenness of gas mixing within the lung (termed ventilation inhomogeneity) and primarily reflects peripheral airway function. 8A higher LCI indicates greater ventilation inhomogeneities.One study reported that LCI was elevated in African infants at 7 weeks of age exposed to household tobacco smoke compared to infants that were unexposed. 9ferring a relationship of prenatal tobacco smoke exposure with unevenness of gas mixing within the lung as well as with the development of wheeze in infancy we hypothesized that higher LCI levels during the first weeks after birth may be found in infants with wheeze in the first year of life.We also aimed to investigate an interaction, where LCI is an "effect modifier" (or moderator) and an increased LCI is one of the preconditions under which the relationship between intrauterine tobacco smoke exposure and infant wheeze can be observed.

| METHODS
Data were derived from a prospective cohort study of infants born to mothers with asthma who participated during pregnancy in the multicenter (Newcastle, Sydney, Brisbane, and Canberra in Australia) Breathing for Life randomized controlled trial (RCT) comparing FeNO-guided asthma management to usual care. 101200 women were enrolled in the RCT, 599 were randomized to the usual care group, and 601 to the FeNO group.In the intervention group, a modified FeNO algorithm was applied every 6-12 weeks (FeNO was used to adjust the ICS dose, and a long-acting beta-agonist was added when symptoms were uncontrolled).This study was approved by the Human Research Ethics Committee (Reference Number 12/ 10/17/3.04),and participation was based on written informed consent.
At enrollment in pregnancy, between 12 and 23 weeks gestation, information was collected about sociodemographic characteristics and smoking status, and an exhaled (e)CO measurement was performed (piCO Smokerlyzer Breath CO Monitor).Participants were asked in relation to any self-reported cigarette use in the past week.
They were considered current smokers if they self-reported smoking, or eCO ≥6 ppm. 11Women <18 years of age or with drug/alcohol dependence were excluded.Residential postcodes were used for socioeconomic status estimates based on the Socioeconomic Index for Areas (SEIFA) numbers.SEIFA scores residential areas according to their relative socioeconomic advantage and disadvantage using Census data.A lower score indicates that the area of residence is relatively disadvantaged compared to an area with a higher score.For those randomized to usual care, one study visit was provided during pregnancy.Participants randomized to the intervention group had study visits every 3-6 weeks until delivery.Infants were prospectively followed up in Newcastle and Sydney only with measurements performed including lung function at 5 to 9 weeks of age (corrected for prematurity [<37 weeks' gestation]), clinical assessment, and parent-reported respiratory questionnaire data at 6 and 12 months of age.Infants were seen between May 2014 and December 2019. 12e age at test shown in Table 1 is corrected for prematurity.

| Infant lung function
Lung function was performed in unsedated infants during behaviorally defined quiet natural sleep, following a feed.Testing was performed with the infant lying supine, using an appropriately-sized infant mask to maintain a tight face seal (sizes 0, 0/1, and 1; Homedica AG), according to the ERS/ATS standards of infant lung function testing. 13,14The additional equipment related dead space volume of the mask was corrected during analysis (size 0-0.005L; size 0/1-0.0085L; and size 1-0.012L).Tidal breathing assessment was performed for 90 s to obtain at least 30 good-quality breaths, 15 with flow measured using an ultrasonic flow meter (Spiroson ® ; EcoMedics AG).
During the wash-in phase, the infant breathed an inert gas mixture containing 4% SF 6 , 21% O 2 and balance N 2 until SF 6 reached an equilibrium of 4% within the lungs.At this point washout commenced, breathing medical-grade air until the end-tidal SF 6 level reached 1/40th of its starting concentration. 16Trials were performed in triplicate to obtain a minimum of two technically acceptable trials.For signal collection, a software package (Spiroware 2, EcoMedics AG) was used for tidal breathing and SF 6 MBW.After adjusting for environmental settings, temperature model parameters, and mask dead space volume used during testing, analysis was performed with Wbreath (V.3.28.0;Ndd Medizintechnik).Average LCI was reported from a minimum of two technically acceptable trials.Tidal breathing flow-volume loop (TBFVL) data was reported from the average of at least 30 consecutive good-quality breaths and the best attempt was selected.Tidal breathing flow-volume loop (TBFVL) and MBW analysis were only performed on data with a volume drift of <3 mL•s −1 , as trials with values above the drift threshold were assumed to contain a leak.

| Statistical analysis
Descriptive statistics and regression analysis were performed using Stata SE v15 for Windows (Stata CorporationA).Differences between groups were assessed using Chi-square tests, Independent-Samples t-test or Mann-Whitney U-tests, as collinearity between self-reported smoking and eCO is shown in Supporting Information S1: e-Table 1).Three-hundred-and-ninetyfour infants had acceptable tidal breathing data and a maternal eCO measurement and 257 infants had acceptable MBW data and maternal eCO available (Table 2).

| Effects of intrauterine tobacco smoke exposure on perinatal outcomes and wheeze
Smoke-exposed infants identified by self-reported smoking had significantly lower socioeconomic backgrounds percentile suggesting they live in more disadvantaged areas (SEIFA decile: 5.0 vs. 6.0, p = .029),lower birth weight (3120 vs. 3420 g, p = .028),and lower birth length (50.0 vs. 51.0cm, p = .039).Smoke-exposed infants identified by self-reported and/or by higher eCO levels were less likely to have been breastfed exclusively up until the day of infant lung function testing (on average first 6 weeks of life, Table 1) than nonexposed infants (28.5% vs. 55.6%,p = .001,Table 1).Comparison of self-reported smoking status and eCO level >6ppm assessments produces a kappa value of 0.70 which suggests a moderate strength of agreement (Supporting Information S1: e-Table 1).
Wheezing in the first year of life was significantly more frequent in smoke-exposed infants identified by self-reported | 1689 smoking compared to nonexposed (77% vs. 50%, p = .005)and also in smoke-exposed infants identified by eCO>6ppm (73.1% vs. 50.3%, p = .025,Table 1).The percentage of infants born to mothers from the FeNO intervention group was slightly higher than that from the usual care group.This is Iikely a direct consequence of the pragmatic RCT design with fewer study visits, and thus less engagement, during pregnancy in the usual care group. 10

| Effects of intrauterine tobacco smoke exposure on infant lung function
A comparative analysis of lung function between infants born to selfreported smokers versus nonsmoking mothers and between infants born to mothers with eCO > 6ppm versus ≤ 6ppm revealed the following differences (Table 2).Changes in TBFVL parameters due to smoking exposure were described, however, no statistically significant differences were found.Regardless of the definition used for maternal smoking during pregnancy, functional residual capacity (FRC) was lower (0.073 vs. 0.081, p = .006,Table 2).In addition, LCI was significantly higher in infants born to tobacco-smoking mothers (7.90 vs. 7.64, p = .030for self-reported smoking vs. no smoking, n = 265; 7.94 vs. 7.65, p = .011for eCO >6 ppm vs. eCO <=6 ppm, n = 257, Table 2).When using a cut-off for LCI based on the population LCI median, infants born to a mother who smoked during pregnancy were significantly more likely to have an LCI above the median value (Table 2).

| Association between infant lung function and wheeze
Next, we investigated risk factors for wheezing in the first 12 months of life and associations with infant lung function.There were no significant differences in birth weight, birth length, age at the test, and sex between the groups (Table 3).Infants with wheeze had higher LCI levels (7.75 vs. 7.61, p = .044,Table 3) and were significantly more commonly exposed to tobacco smoke (12% vs. 4%, p = .005),born prematurely (10.9% vs. 5.1%, p = .043),and had a lower length at 6-7 weeks of age (56.0 vs. 56.9cm, p = .005).
Twenty-two per cent of infants (n = 68 out of 312) had an Emergency Department presentation for any respiratory symptoms in the first year of life, and 94% of those (n = 64 out of 68) presented for wheezy symptoms (Table 3).

| Interactions between LCI and intrauterine tobacco smoke exposure for the risk of wheeze
Employing a regression analysis using an interaction term and adjusting for possible confounders, we found a significant statistical interaction between self-reported smoking in pregnancy and LCI 1.000-1.001,p = .040,n = 186) were significantly associated with wheeze in the first year of life for smoke-exposed infants identified by self-reported smoking (Supporting Information S1: e-Table 2).

| DISCUSSION
To the best of our knowledge, this is the first study to describe ventilation inhomogeneity in predominantly Caucasian infants born to asthmatic mothers and exposed to smoking in utero using both self-reported smoking status and biochemical validation using eCO.
We demonstrate an interaction between smoking in pregnancy and LCI for the development of wheeze in the first year of life.
Furthermore, it highlights the potential for further investigations that could employ infant lung function, and specifically LCI, to more precisely predict future respiratory symptom development in babies at risk as a result of, for instance, in-utero tobacco smoke exposure.
A complex mixture of more than 7,000 chemicals is found in tobacco fumes and probably several of those impact lung development. 17However,nicotine is of particular interest as it crosses the placental barrier and fetal levels in blood and lungs closely mimic those measured in the mother's blood.Nicotinic acetylcholine receptors (nAChR) are expressed by airway epithelial cells and fibroblasts in the lungs. 18nAChR expression in fetal lungs was further upregulated by nicotine when pregnant rhesus monkeys were treated with subcutaneous nicotine to maintain a nicotine level in amnionic fluid that is consistent with the level in pregnant human smokers. 19cotine exposure was associated with increased lung collagen deposition and a decrease in expiratory flows similar to those seen in infants. 18,202][23] The critical period for prenatal nicotine exposure to affect expiratory flow parameters may correspond to the canalicular and saccular period of lung development. 21creased ventilation inhomogeneity has been demonstrated in lung diseases including cystic fibrosis (CF), primary ciliary dyskinesia (PCD), 24 and severe asthma. 25,26Higher LCI and lower FRC levels shortly after birth were found in infants with CF compared to healthy controls (7.08 vs. 6.78;p = .034), 27demonstrating that LCI measurement early in life serves as a highly sensitive noninvasive marker to detect and monitor the presence of obstructive lung disease in CF.In our study, we used MBW and TBFVL to compare infants exposed and nonexposed to smokingin utero.The increase in LCI associated with smoke exposure was comparable to the magnitude of effect at this age described in other important respiratory diseases such as CF (LCI in our cohort: 7.90 vs. 7.64; p = .030). 27The increased LCI values reported here probably reflect changes in peripheral airway caliber, rather than the early suppurative lung disease changes described in CF.However, this requires further studies using high-resolution imaging.Furthermore, despite a statistically significant change in MBW parameters being seen due to in-utero smoking exposure, we T A B L E 3 Baseline characteristics, multiple breath washout and genetic risk variants in infants with wheeze in the first year of life.did not observe similar changes in TBFVL.Longer and more tortuous airways due to lung branching have been demonstrated to be nicotine dose-dependent. 23Possibly, the absence of an effect on the TBFVL in this study may suggest that the in-utero smoking exposure led to structural airway changes rather than changes to the respiratory system mechanics, suggesting that small airway development has been affected.Alternatively, detectable changes in TBFVL are masked by the effect of asthma in pregnancy on tidal breathing lung function parameters in our cohort of infants. 12wever, it is also possible that, as compared to a previous study conducted 25 years ago, 28 the cumulative dose of intrauterine tobacco smoke exposure has become less as a result of legislative smoking bans for reducing harms from secondhand smoke exposure precluding us to identify effects on TBVL parameters. 29In any case, LCI may be a more sensitive marker for structural, or functional, airway changes.It has yet to be determined whether lung ventilation inhomogeneities persist in early childhood and beyond in tobacco smoke-exposed babies.
Our study demonstrated, as expected, that infants with a history of in-utero exposure to cigarette smoke were more likely to have later wheeze.In this same cohort, we have previously shown that maternal asthma in pregnancy is a risk factor for impaired lung function in young male infants. 12The literature suggests that early life exposures may limit lung development in utero and early infancy, preventing individuals from ever achieving their maximal lung function level, 30 and leading to an increased risk of wheezing.
Our findings indicate that prematurity, birth weight and length were associated with later wheeze.Although the association was not statistically consistent for both self-reported and eCO >6 ppm, the effect of prematurity and birth weight on later wheeze was high when smoke-exposed infants were identified by self-reported smoking.On the other hand, only birth length was associated with wheeze in the first year of life for infants born to mothers with eCO >6 ppm.The results could suggest a mechanism by which in-utero tobacco smoke exposure is more sensitive to a minimum dose-related response, which is consistent with previous studies. 31,32limitation of our study is that all infants were born to mothers with asthma who have previously been shown to have lower tidal breathing lung function parameters in early life 12 and increased risk of developing wheeze and asthma.It is conceivable that in asthmatic mothers the association between tobacco smoking during pregnancy and adverse pregnancy outcomes is mediated, in part, through suboptimal asthma control, or greater incidence of significant asthma exacerbations.This would be an important topic for future in-depth analysis.As infants born to women with asthma in pregnancy have reduced lung function, 12 our results cannot be extrapolated to the general population without further research on infants born to nonasthmatic women.However, our results are relevant to a large proportion of children at risk of developing respiratory symptoms as up to 40% of asthmatic children have a mother with asthma. 33other limitation was that eCO measurements were not performed throughout pregnancy.The eCO measurement depended on the randomization of the control or the intervention arm, which dictated a different study visit regimen.Consequently, we used only the baseline eCO measurement for both arms, it being unfeasible to use the data from attendance at other visits.Additionally, the respiratory questionnaire applied at 12 months of age to assess the wheezing occurrence was parent self-reported, which might have introduced a response bias.The strengths of our study are a multicenter study design, validation of active smoking status in pregnancy using eCO, and use of a sensitive tool to assess peripheral airway function in infancy.
Therefore, interactions between airway function and environmental exposure modulate the risk of developing infant wheeze.We suggest that interventions to prevent tobacco smoke effects on the offspring's airway function such as antioxidant supplementation during pregnancy 34 should be prioritized to reduce the respiratory disease burden in later life.LCI could be explored in future studies as a useful marker for identifying those infants at higher risk of developing respiratory symptoms, along with characterization of genetic variants and environmental exposures, in a precision medicine approach.
Note: SEIFA index is presented as a decile.Asthma exacerbation during pregnancy: hospitalization; hospitalization/emergency department presentation; and exacerbations requiring medical intervention including hospitalization, emergency department presentation, oral corticosteroid use, and unscheduled doctor visits.Abbreviations: IQR, interquartile range; SD, standard deviation; SEIFA, socioeconomic index for areas.appropriate.Interactions were tested using a modified Poisson regression model with a robust error variance used for binary data including an interaction term for LCI and self-reported smoking or LCI and eCO >6 ppm.All models were adjusted for sex, prematurity, environmental tobacco smoke, birth weight, birth length, length at test, SEIFA and asthma management group during pregnancy.Consent to participate in the follow-up birth cohort was obtained for 64% (n = 662/1035) of infants, of whom 90% (n = 594/662) attended infant lung function at age 7 weeks.Technically acceptable lung function data were obtained for 71% (n = 423/594) infants.Acceptable tidal breathing data was obtained in 411 infants and acceptable MBW data in 265 infants (Figure1).Of those with acceptable data, 10% (n = 42/423) of all mothers self-reported tobacco smoking during pregnancy (mean age 28.4 ± 6) and 381 self-reported not having smoked during pregnancy (mean age 30.2 ± 5, Table1).Four hundred and five women had eCO levels measured during pregnancy.12%(48/405) of mothers had an eCO level >6ppm (mean age 27 ± 6.2) while 357 mothers had an eCO lower than 6 ppm indicating abstinence on the day of measurement (mean age 30.2 ± 5, Table1;