Influence of childhood growth on asthma and lung function in adolescence

Background Low birth weight and rapid infant growth in early infancy are associated with increased risk of childhood asthma, but little is known about the role of postinfancy growth in asthmatic children. Objectives We sought to examine the associations of children's growth patterns with asthma, bronchial responsiveness, and lung function until adolescence. Methods Individual growth trajectories from birth until 10 years of age were estimated by using linear spline multilevel models for 9723 children participating in a population-based prospective cohort study. Current asthma at 8, 14, and 17 years of age was based on questionnaires. Lung function and bronchial responsiveness or reversibility were measured during clinic visits at 8 and 15 years of age. Results Rapid weight growth between 0 and 3 months of age was most consistently associated with increased risks of current asthma at the ages of 8 and 17 years, bronchial responsiveness at age 8 years, and bronchial reversibility at age 15 years. Rapid weight growth was associated with lung function values, with the strongest associations for weight gain between 3 and 7 years of age and higher forced vital capacity (FVC) and FEV1 values at age 15 years (0.12 [95% CI, 0.08 to 0.17] and 0.11 [95% CI, 0.07 to 0.15], z score per SD, respectively) and weight growth between 0 and 3 months of age and lower FEV1/FVC ratios at age 8 and 15 years (−0.13 [95% CI, −0.16 to −0.10] and −0.04 [95% CI, −0.07 to −0.01], z score per SD, respectively). Rapid length growth was associated with lower FVC and FVC1 values at age 15 years. Conclusion Faster weight growth in early childhood is associated with asthma and bronchial hyperresponsiveness, and faster weight growth across childhood is associated with higher FVC and FEV1 values.

Asthma is the most prevalent chronic respiratory disease in children worldwide. 1,2 Many factors have been associated with increased risk of asthma or lower lung function, such as gestational age, tobacco smoke exposure, breast-feeding habits, and a family history of asthma or allergy. [3][4][5][6][7] Respiratory morbidity might also be the result of abnormal growth. Fetal growth [8][9][10] and low birth weight [10][11][12][13][14][15][16] have been associated with asthma. Several studies have explored the associations of infant or childhood growth with the risk of asthma or lung function in later life. 10,[17][18][19][20][21][22][23][24][25][26] They reported an increased risk of asthma symptoms in preschool children with accelerated growth in early infancy, 10,23 an increased incidence of asthma at 6 years after a rapid increase in body mass index in early childhood, 22 a lower FEV 0.4 value in the first months of life in children with greater postnatal weight gain, 21 and a negative association of growth with lung function during the first year of life. 27 However, other studies observed no evidence for increased risk of asthma caused by rapid growth 24 or observed that weight gain during the first year was positively associated with lung function. 19 These inconsistencies could be explained in part by methodological issues, including differences in the definitions of growth or asthma outcomes and adjustment for potential confounders.
It was suggested in a published study that growth in early infancy, especially from birth to 3 months, 10 might be an Abbreviations used ALSPAC: Avon Longitudinal Study of Parents and Children FEF  : Forced expiratory flow between 25% and 75% FVC: Forced vital capacity OR: Odds ratio important influence on asthma risk. However, it is unknown whether this association persists until adolescence or influences lung function, although tracking of lung function suggests that its trajectory is established by midchildhood. 20,28 It is also not known whether the first 3 months after birth is the only important time period or whether any specific period after the first year of age might play a role as well.
The underlying mechanism of the associations between growth and respiratory morbidity might include abnormal growth and development of the lungs or immunologic or inflammatory effects, such as adiposity-related systemic and tissue-specific inflammation. [29][30][31][32][33] To test our hypothesis that rapid early growth is negatively associated with respiratory health, we examined the association of children's growth trajectories from birth until age 10 years with current asthma, bronchial responsiveness or reversibility, and lung function in adolescence in a populationbased prospective birth cohort study among 9723 children.

METHODS Design and setting
Subjects were participants in the Avon Longitudinal Study of Parents and Children (ALSPAC) in the United Kingdom, which has been described previously 34 and on the study's Web site (www.bristol.ac.uk/alspac). In brief, 15,247 pregnant women residing in one of 3 Bristol-based health districts with an expected delivery date of between April 1, 1991, andDecember 31, 1992, were recruited and gave birth to 14,316 singleton children who were alive at the age of 1 year. Children with no information on either growth trajectories (n 5 701) or any asthma outcome (n 5 3,892) were excluded, leaving a total of 9,723 children included in the current analyses (see Fig E1 in this article's Online Repository at www.jacionline.org). Ethical approval for the study was obtained from the ALSPAC Law and Ethics Committee and local research ethics committees. Witten informed consent was obtained from all participants and their parents or guardians.

Growth trajectories
Height and weight measurements were available from birth up to age 10 years from a variety of sources (see the Methods section and Table E1 in this article's Online Repository at www.jacionline.org for full details). Linear spline multilevel models were used to estimate trajectories of height and weight. The models estimate mean and person-specific birth weight or length and mean and person-specific rates of weight or height growth between 0 and 3 months, 3 months and 1 year, 1 and 3 years, 3 and 7 years, and 7 and 10 years of age and are described in full elsewhere. 35 Early growth was defined as growth between birth and the age of 1 year, midchildhood growth as growth between the ages of 1 and 7 years, and late childhood growth as growth between 7 and 10 years of age. We generated SD scores (z scores) for birth weight and length and rate of weight/height growth in each period of childhood by subtracting the mean from the person-specific value and dividing by the SD. These SD scores for birth weight/length and rates of growth are used as exposures in our analyses.

Asthma and lung function
Current asthma status was obtained at the ages of 8, 14, and 17 years. Current asthma was defined as a reported doctor's diagnosis of asthma ever and reported wheezing, asthma, or use of asthma medication in the previous 12 months. Skin prick test reactivity was determined at the age of 7 years. A child was deemed to react to an allergen (grass, house dust, or cat) if their wheal and/or flare responses were 2 mm or greater and they had no reaction to the negative control. Bronchial hyperresponsiveness, unselected for asthma or wheezing, was measured at the ages of 8 and 15 years. 36 At age 8 years, we tested the provoking dose of methacholine causing a decrease in FEV 1 from baseline. The dose-response slope was calculated by fitting a linear function to the plot of percentage decrease from baseline. We dichotomized bronchial responsiveness using the highest tertile as responders and the rest as nonresponders. At age 15 years, we defined bronchial reversibility as a change of equal to or greater than 12% between FEV 1 before and after inhalation of a standard dose (400 mg) of salbutamol. 37 Spirometry (Vitalograph 2120; Vitalograph, Maids Moreton, United Kingdom) was performed at 8 and 15 years of age according to American Thoracic Society standards. 38 Lung function measurements (FEV 1 , forced vital capacity [FVC], forced expiratory flow between 25% and 75% [FEF  ], FEV 1 /FVC ratio, and FEF 25-75 /FVC ratio) were converted into sex-, age-, and height-adjusted z scores (see Table E1 for time points of outcomes). 39

Covariates
Maternal age, highest qualification, body mass index, parity, and a history of asthma or atopy were reported in questionnaires at 12 weeks of gestation, and smoking during pregnancy was assessed at 18 weeks of gestation by using self-completion questionnaires sent to the mothers. Maternal anxiety during pregnancy was measured at 32 weeks of pregnancy and was defined as the highest quartile of the Crown-Crisp Experiential Index. 40 Children's gestational age and sex were obtained from birth records. Breast-feeding status at age 8 months was obtained from maternal self-completion questionnaires.

Statistical analysis
We used logistic regression models to assess associations between growth trajectories and current asthma, atopy, and bronchial responsiveness or reversibility. Linear regression models were used to assess associations of growth trajectories with lung function measurements. Analyses were adjusted for potential confounders, including maternal age, body mass index, anxiety, education, history of asthma or atopy, smoking habits, parity, and the child's sex, gestational age at birth, and breast-feeding status. Models of weight gain were additionally adjusted for birth weight, and preceding rates of height-adjusted weight growth trajectories and models of height gain were additionally adjusted for preceding rates of height growth trajectories and birth weight. Models for current asthma or lung function were additionally adjusted for previous current asthma or lung function measurements. In addition, body mass index at the age of outcome assessment was added as an interaction to explore potential effect modification on the associations of childhood growth with asthma and lung function.
Missing data in confounders were imputed by using multiple imputations. Percentages of missing values within the population for analysis were lower than or near 10%, except for maternal body mass index (13.1%), anxiety (13.6%), and child's breast-feeding duration (11.5%). Ten new data sets were created by means of imputation based on all covariates, determinants, and outcomes in the model. 41 All data sets were analyzed separately, after which results were combined. No differences in results were observed between analyses with imputed missing data or complete cases only. Therefore we present only results based on imputed data sets. Statistical analyses were performed with the Statistical Package of Social Sciences version 19.0 for Windows (SPSS, Chicago, Ill). Table I. Children were born at a median gestational age of 40 weeks (95% range, 35-42 weeks), with an average birth weight of 3436 grams (SD, 524 grams). Current asthma was reported in 13.9%, 13.2%, and 15.3% of the children at the age of 8, 14, and 17 years. All covariates differed between those included and those excluded from this study, apart from maternal history of asthma (see Table E2 in this article's Online Repository at www.jacionline.org).

Childhood growth with asthma
We observed no evidence of an association between higher birth length or weight and current asthma (Table II). Height growth in midchildhood tended to be negatively associated with current asthma at age 8 years, with the strongest evidence of association for height gain between 3 and 7 years of age and asthma at 8 years of age (odds ratio [OR], 0.75 [95% CI, 0.66-0.86] per SD increase). More rapid weight gain during early childhood tended to be positively associated with current asthma, with the most consistent associations observed for weight gain between 0 and 3 months of age and asthma at 8 Table II). We did not find strong evidence that current body mass index modified the association of childhood growth with asthma. P values for the crossproduct of growth measurements with body mass index were greater than .05. No associations were observed between any growth measures and skin prick test reactivity (see Table E3 in this article's Online Repository at www.jacionline.org).

Childhood growth with bronchial responsiveness
We observed no evidence of an association between higher birth length or weight and bronchial hyperresponsiveness (Table III). Also, no evidence was found for associations between height gain in early childhood, midchildhood, or late childhood and bronchial responsiveness or reversibility at 8 and 15 years of age, respectively. Higher weight gain in early childhood (between 0 and 3 and 3 and 12 months of age only) was associated with an increased risk of bronchial responsiveness to methacholine at 8 years (OR, 1.11 [95% CI, 1.03-1.20] and 1.09 [95% CI, 1.00-1.19], respectively, per SD increase) and bronchial responsiveness to salbutamol at 15 years (OR, 1.14 [95% CI, 1.00-1.31] and 1.24 [95% CI, 1.07-1.42], respectively, per SD increase). No strong evidence was observed for effect modification of childhood growth with current body mass index on bronchial responsiveness or reversibility (P for interaction > .05). After birth, more rapid height gain in early childhood and midchildhood was most consistently associated with a lower FVC and FEV 1 values at age 15 years but not with other lung function variables or ratios or with lung function at age 8 years (Fig 1, A-I).

Childhood growth with lung function
Higher birth weight was most strongly associated with higher FVC, FEV 1 , and FEF 25- ; Fig 1, B, D, and F). Also, higher birth weight was associated with reduced FEV 1 /FVC and FEF 25-75 /FVC ratios at age 15 years (Fig 1, H and J). After birth, more rapid weight growth throughout childhood was associated with higher FVC and FEV 1 values, with the greatest effect estimates for weight gain in midchildhood and FVC and FEV 1 values at age 15 years (0.12 [95% CI, 0.08-0.17] and 0.11 [95% CI, 0.07-0.15], z score per SD, respectively; Fig 1, B and D). For the other lung function variables, more rapid weight gain in early childhood was associated with a decreased FEF 25-75 value at 8 years of age only (Fig 1, F). We observed lower FEV 1 /FVC and FEF  /FVC ratios at the ages of 8 and 15 years for early rapid weight gain, followed by normal ratios for midchildhood weight gain but lower ratios for late rapid weight gain (Fig 1, H and J). We observed effect modification of childhood weight growth by current body mass index on lung function (P for interaction < .05) but not of childhood height growth (P for interaction > .05). Stratified analyses for body mass index showed that the effect estimates of childhood weight growth for FVC and FEV 1 were larger in the group of children with a normal body mass index compared with the overweight children (see Table E5 in this article's Online Repository at www.jacionline.org).

DISCUSSION
Our results suggest positive associations of rapid weight growth during early childhood and midchildhood with current asthma, higher weight growth during early childhood with increased bronchial responsiveness or reversibility, and higher weight growth in childhood with higher overall lung volumes but increased measures of obstruction (FEV 1 /FVC and FEF 25-75 /FVC ratios) in childhood. Higher length at birth and height growth in childhood were associated with lower lung volumes but less consistently associated with the other respiratory outcomes.

Comparison with previous studies
Previous studies of the association of childhood growth with asthma have reported an increased risk of asthma symptoms in preschool children with accelerated growth in early infancy. 10,23 A previous study that measured asthma at an older age (6 years) showed no evidence for increased risks caused by changes in growth using similarly defined growth trajectories as in our study. However, those authors did report increased risks of ever wheezing in those with higher weight growth in early childhood. 24 Differences in results with our study might be explained Values are ORs (95% CIs). Models are adjusted for maternal age, education level, history of asthma, body mass index, parity, smoking during pregnancy, anxiety, and the child's sex, gestational age, breast-feeding duration, and previous height or weight gain. Models of weight were additionally adjusted for preceding height and weight growth trajectories, and models of height were additionally adjusted for preceding height growth trajectories and birth weight. Also, models were additionally adjusted for previous current asthma. Values are ORs (95% CIs) of bronchial responsiveness or reversibility. Models are adjusted for maternal age, education level, history of asthma, body mass index, parity, smoking during pregnancy, anxiety, and the child's sex, gestational age, breast-feeding duration, and previous height or weight gain. Models of weight were additionally adjusted for preceding height and weight growth trajectories, and models of height were additionally adjusted for preceding height growth trajectories and birth weight. Methacholine responsive, Highest tertile versus lower tertiles; Salbutamol responsive, greater than 12% change in FEV 1 vs less than 12% change. Values are differences in z score lung function (95% CIs). Z scores were calculated for sex, age, and height at the time of measurement. FEV 1 /FVC and FEF 25-75 /FVC sex-adjusted z scores were additionally adjusted for age and height of measurement. Models are adjusted for maternal age, education level, body mass index, parity, smoking during pregnancy, anxiety, history of asthma, and the child's sex, gestational age, and breast-feeding duration. Models of weight were additionally adjusted for preceding height and weight growth trajectories, and models of height were additionally adjusted for preceding height growth trajectories and birth weight. Models for lung function at 15 years of age were additionally adjusted for lung function measures at age 8 years.
by differences in the study populations (general population, term-born children only) and the age at which asthma was measured (early childhood, midchildhood, or late childhood). A meta-analysis on body mass index gain in early childhood and midchildhood suggested that more rapid body mass index gain in early childhood but not thereafter was associated with an increased incidence of asthma at age 6 years, 22 which is consistent with our findings about asthma at age 8 years.
To the best of our knowledge, no previous studies have examined the relationship between childhood growth and bronchial responsiveness or reversibility. However, because asthma is associated with bronchial hyperresponsiveness, 42 the association between early childhood weight gain and the objective measure of bronchial responsiveness is in line with previous studies on growth and asthma outcomes, 9,10,[22][23][24]27 and this strengthens our conclusions about the association with asthma by using both objective and self-reported outcome measures. Previous studies that measured lung function during early childhood reported lower FEV 0.4 values in the first months of life in term-born children with greater postnatal weight gain. 21 Turner et al 27 showed a negative association of growth between 1 and 12 months of age and lung function change (V9max FRC [maximal flow at functional residual capacity]) during the same period. Only a tendency toward an association of growth with lower FEF 25-75 values at 11 years of age was observed. 27 Our findings were in line with these results. In contrast, Canoy et al 19 showed in adults that weight gain during the first year was positively associated with adult lung function independent of birth weight. Additionally, we showed in a large number of subjects that weight gain in midchildhood and late childhood was associated with lung function independent of birth weight and weight gain in early childhood.

Interpretation of results
The most prominent and novel findings in this study are the positive associations of weight gain in early childhood, specifically weight gain in the first 3 months of life, and lung function changes at 8 and 15 years of age. This early postnatal period has been observed previously to be important for the development of asthma symptoms and decreased lung function up to preschool age. 10,21,23,24 Our results suggest that the effects of rapid weight gain in the first 3 months of life on asthma and bronchial hyperresponsiveness persist until adolescence. Weight gain between 0 and 3 months of age was associated with asthma at both 8 and 17 years of age, whereas weight gain between 3 and 12 months of age was only associated with asthma at age 8 years. This might be due to a different underlying mechanism between these intervals, such as early developmental influences resulting in persistent changes in airway or immune development after more rapid growth in the first months after birth. Whereas, more rapid growth and multiple other exposures of influence after these first months leads to modifiable changes. However, this also might be a chance finding or caused by the smaller number of children in the older age group. Therefore these associations need to be replicated in other studies. Additionally, rapid weight growth in midchildhood and late childhood was associated with changes in lung function variables. The underlying mechanisms of rapid weight gain in childhood on asthma and lung function outcomes are unclear and should be assessed in future studies. We speculate that abnormal growth and development of the lungs, possibly with mismatch between airway and alveolar growth or immunologic and inflammatory effects with lung and airway remodeling, might play a role. 29,32,43 Both airways and alveoli continue to develop, at least until adolescence. 44 Therefore exposures, including growth, during childhood are likely to influence this development. Our results suggest that growth in the first months after birth is the most important in the association with asthma and lung function because most rapid developmental and growth changes occur during this time. Growth after this early period also might influence lung function but might not completely reverse the early effects. Also, higher leptin levels have been associated with lower lung function in childhood, suggesting a possible mechanism mediated by adiposity-related hormones that might influence pulmonary growth factors, such as vascular endothelial growth factor, transforming growth factor, and insulin-like growth factor. 45 Also, unmeasured factors determining early growth might influence respiratory health in adolescence. The FEV 1 / FVC ratio is a measure of obstruction, and decreased values are a feature of asthma. We observed increases in FVC and FEV 1 values in association with rapid early weight gain but a lower FEV 1 /FVC ratio, which would be consistent with greater influence of early rapid weight gain on lung volume than airway growth. Because weight gain in infancy is proportionally greater than in subsequent years, effects of rapid weight gain on an imbalance between FEV and FVC values might be most influenced during this specific period. The FEF 25-75 /FVC ratio has also been suggested as a measure of dysanapsis in which airways are small in relation to total lung capacity, 46 and therefore our finding of rapid weight gain associations with lower FEF 25-75 /FVC ratios would be consistent with this explanation. The associations of early childhood and midchildhood height growth and decreased FVC and FEV 1 values at the age of 15 years could also point to dysanapsis; however, we did not observe an association of height growth with the FEV 1 /FVC ratio. Therefore height growth is less likely to be associated with a mismatch between airway and alveolar growth.
Although we did not observe an association of growth with skin prick test reactivity, another possible explanation for effects of rapid weight gain on lung function is through influence of adipose tissue on the developing immune system through secretion of immunologically active factors, including adipokines and chemokines, which stimulate release of TNF-a and interleukins. 47 In mice leptin has been shown to enhance airway responsiveness, suggesting an immunomodulatory role, 48 and the effect has also been reported in human subjects, although results are inconsistent. [49][50][51] If this potential underlying mechanism is a factor in the association between growth and asthma, we suggest that not only obesity but also weight gain in normal and overweight children leads to increased leptin levels. We observed no evidence that body mass index modified associations of childhood growth with asthma and bronchial responsiveness or reversibility or on the association of childhood height growth with lung function. We observed effect modification of childhood weight growth by current body mass index on lung function (P for interaction < .05). Stratified analyses for body mass index showed that the effect estimates of childhood weight growth for FVC and FEV 1 values were larger in the group of children with a normal body mass index compared with overweight children (see Table E2). This suggests that the development of lung and airway volumes of overweight subjects are less influenced by weight gain in childhood than lung and airway volumes of normal-weight children. The underlying mechanisms should be studied in future research. Finally, a common unknown factor that increases weight gain and is also responsible for a higher risk of respiratory morbidity, such as shared genetic risk, might be involved. 52

Strengths and limitations
This study was embedded in a population-based prospective cohort study, with a large number of subjects being studied from pregnancy onward and detailed and prospectively acquired information about growth and respiratory morbidity. Modeled growth trajectories for this population enabled us to take account of different timings and numbers of measurements between children. Additionally, previous changes in weight and height were considered in the models, and therefore changes in the time intervals reflect growth during that specific interval independent of earlier growth and not simply catch-up or down growth after early aberrant growth. Because puberty can influence pulmonary physiology, an important limitation of our growth models is that they do not continue during puberty. Growth and puberty have a bidirectional relationship, which makes disentangling their effects complex. Further research is needed to explore the effects of growth during and after puberty with respiratory health. Lung function measurements were made by using the same methods at 2 time points, and methacholine challenge or bronchodilator reversibility were used to evaluate bronchial responsiveness, producing objective respiratory outcomes. After Bonferroni corrections for multiple testing (P <.00625), only the associations of growth with asthma did not hold. However, statistical corrections for multiple comparisons seem conservative because all exposures and outcomes are correlated. We adjusted for a large number of confounders.
Loss to follow-up data in the ALSPAC cohort is associated with social deprivation but not with atopic predisposition. 53 Nonresponse and loss to follow-up would lead to biased effect estimates if associations of childhood growth with respiratory outcomes differed between those included and not included in the analyses. This is unlikely but difficult to study. Therefore we believe that the differences in covariates between those included and excluded did not influence the effect estimates but might have affected the generalizability of the observed effects.
Also, we were unable to take fetal growth into account. Growth in childhood might be the result of various fetal growth patterns that could underlie associations of growth in childhood with asthma and lung function. However, previous studies showed inconsistent effects of fetal growth with respiratory outcomes. 8,10 In conclusion, our results suggest that rapid weight growth during specific intervals in childhood is associated with current asthma, increased bronchial responsiveness and reversibility, and higher lung volumes and measures of obstruction. Rapid length growth was only associated with lower overall lung volume. Therefore changes in weight, especially early weight growth, appear to be important in lung development. Further studies are needed to replicate these findings and to explore the underlying mechanisms of the effect of growth in specific periods on respiratory health and differential lung growth.
We thank all the families who took part in this study, the midwives for their help in recruiting them, and the whole ALSPAC team, which includes interviewers, computer and laboratory technicians, clerical workers, research scientists, volunteers, managers, receptionists, and nurses.
Key message d Faster weight growth in early childhood is positively associated with asthma and bronchial hyperresponsiveness during adolescence.

Growth trajectories
Length and height data for the children were available from several sources. Birth length (crown-heel) was measured with a Harpenden Neonatometer (Holtain, Crosswell, Crymych, United Kingdom) by trained research staff who visited newborns soon after birth (median, 1 day; range, 1 to 14 days). From birth to age 5 years, measurements were available from health visitor records, which form part of standard childcare in the United Kingdom, for the majority of the cohort. On average, up to 4 measurements were taken at 2, 10, 21, and 48 months of age, which has been demonstrated previously to have good accuracy. E1 For a random 10% of the cohort, direct measurements from a series of research clinics held between the ages of 4 months and 5 years were also available. At these clinics, crown-heel length for children aged 4 to 25 months was measured with a Harpenden Neonatometer (Holtain), and from 25 months onward, standing height was measured with a Leicester height measure (Seca, Hamburg, Germany). Weight was measured by using a Seca scale. From age 7 years upward, all children were invited to annual clinics, at which time standing height was measured (without shoes) to the last complete millimeter by using the Harpenden stadiometer (Holtain), and Tanita Body Fat Analyses (Model TBF 305) were used for weight measurement. Across all ages, parent-reported child heights and weights were also available from questionnaires. These measurements were comparable with routinely collected child health record height and weight data with no systematic bias. E1 Therefore all the above-described height and weight measurements were used to generate growth trajectories for height and weight from children with at least 2 observed measurements.
We used fractional polynomials to find the best-fitting average trajectory and used this to derive approximate knot points for a linear spline model. E2 A separate model was created for female and male subjects. We simplified the models with the aim of having the same knot points in female and male subjects for both weight and height and having knot points at round ages in months. Model fit with these models was not appreciably lower than in the optimal model. The defined knot points are 3 months and 1, 3, and 7 years, thereby creating the following growth rate trajectories: 0 to 3 months, 3 months to 1 year, 1 to 3 years, 3 to 7 years, and 7 to 10 years. Growth rates are presented as changes in SD, which were generated by subtracting the mean from the person-specific value and dividing by the SD.
Please note that the study Web site contains details of all the data, which are available through a fully searchable data dictionary at the following Web page: www.bris.ac.uk/alspac/researchers/data-access/data-dictionary. Growth model 3 Growth model 4 Growth model 5 Growth model 6 Growth model 7 Growth model 8 Growth model Asthma Bronchial responsiveness (methacholine) Spirometry 9 Growth model 10 Growth model 11 12 13 14 Asthma 15 Bronchial reversibility (salbutamol) Spirometry 16 17 Asthma    Values are differences in z score lung function (95% CIs). Z scores were calculated for sex, age, and height at the time of measurement. Models are adjusted for maternal age, education level, body mass index, parity, smoking during pregnancy, anxiety, history of asthma, and the child's sex, gestational age, and breast-feeding duration. Models of weight were additionally adjusted for preceding height and weight growth trajectories, and models of height were additionally adjusted for preceding height growth trajectories and birth weight. Models for lung function at age 15 years were additionally adjusted for lung function measures at age 8 years.