Using a Mixture-of-Bivariate-Regressions Model to Explore Heterogeneity of Effects of the Use of Inhaled Corticosteroids on Gestational Age and Birth Weight Among Pregnant Women With Asthma

Using a Mixture-of-Bivariate-Regressions Model to Explore Heterogeneity of Effects of the Use of... Abstract Asthma is a heterogeneous disease, and responses to asthma medications vary noticeably among patients. A substantively oriented objective of this study was to explore the potentially heterogeneous effects of exposure to maternal inhaled corticosteroids (ICS) on gestational age (GA) at delivery and birth weight (BW) using a cohort of 6,197 pregnancies among women with asthma (Quebec, Canada, 1998–2008). A methodologically oriented objective was to comprehensively describe the application of a Bayesian 2-component mixture-of-bivariate-regressions model to address this issue and estimate the effects of ICS on GA and BW jointly. Based on the proposed model, no association between ICS and GA/BW was found for a large proportion of asthmatic pregnancies. However, a positive association between ICS exposure and GA/BW was revealed in a small subset of pregnancies comprising mainly preterm and low–birth-weight infants. A novel application of this model was also subsequently performed using BW z score instead of BW as the outcome variable. In conclusion, the studied mixture-of-bivariate-regressions model was useful for detecting heterogeneity in the effect of ICS on GA and BW in our population of women with asthma. These analyses pave the way for analogous uses of this model for general assessment of exposure effect heterogeneity for these perinatal outcomes. asthma, birth weight, gestational age, heterogeneity, inhaled corticosteroids, mixture of regressions Asthma is a widespread and complex disease whose etiology involves both genetic and environmental factors (1–3). Worldwide, inhaled corticosteroids (ICS) are the first-line controller therapy for asthma at all ages (4–7). Common add-on therapies to ICS include leukotriene receptor antagonists and long-acting β2-agonists (LABA) (4, 5). Despite well-accepted guidelines, it is recognized that responses to asthma medications vary considerably among patients (8–10). While adherence to treatment, therapy combinations, and smoking explain some heterogeneity in response to asthma medication (11–16), it has been suggested to have a strong genetic foundation (8, 17, 18). Asthma is one of the most frequent chronic medical conditions to be reported during pregnancy (19). The prevalence of asthma among pregnant women has been observed to increase in the past several decades (20, 21), with recent reported numbers ranging from 7.8% to 13% (22–24). During pregnancy, current guidelines recommend continuing pharmacological therapy, as uncontrolled asthma puts the pregnant woman and the fetus at risk of adverse outcomes (5, 7, 25). Indeed, data from well-regarded studies have shown that optimally treated asthmatic women have fewer adverse perinatal and maternal outcomes than those without therapy (26–28). Moreover, difficult-to-control asthma requiring systemic corticosteroids during pregnancy has been associated with increased risk of maternal preeclampsia, perinatal mortality, hyperbilirubinemia, preterm birth (PTB; defined as gestational age (GA) at delivery <37 weeks), and low birth weight (LBW; defined as birth weight (BW) <2,500 g) (23, 29–34). The use of ICS at low-to-moderate doses by asthmatic pregnant women is generally regarded as safe with respect to adverse perinatal outcomes, including GA- and BW-related variables (22, 35). However, doubts remain concerning the effect of higher doses of ICS on GA and BW (22, 36, 37). While the dose-response relationship between ICS and GA and BW is one important pharmacoepidemiologic research direction (22), exploration of heterogeneity in the effects of ICS on BW and GA has not received due attention, being circumscribed to analyses stratified on asthma severity (38) or fetal sex (38, 39). Because recent data suggest that ICS treatment may lead to greater benefit for the control of asthma in certain subpopulations (40–44), it is reasonable to conjecture that ICS could exhibit heterogeneous effects on GA and BW due to patients’ or pregnancies’ characteristics. Finite mixture models are frequently used for the purpose of clustering (45–47). From this point of view, one desires to find homogeneous subpopulations among a hypothesized heterogeneous population (48). Specifically, this approach assumes the existence of K latent (unknown) subpopulations and aims to recover them by fitting a K-component mixture distribution, where the component-specific distributions are often taken to be from the same parametric family (e.g., such as normal distributions). Finite mixtures have been regularly proposed for modeling BW distributions (49–54). For this outcome, it has been suggested that the mixture’s components are the result of different types of births: The main component represents the subpopulation of normal births, while the other, minor components (typically 1 or 2 additional components) represent subpopulations of abnormal births (49, 55, 56). Recently, finite mixture models have also been applied in asthma pharmacogenetics to handle individual heterogeneity in response to the drug montelukast, a leukotriene receptor antagonist (57). In this paper, we use a finite mixture-of-regressions model to explore possible heterogeneity in the effects of ICS on GA and BW in a population of asthmatic pregnant women from the province of Quebec, Canada. In a mixture-of-regressions model, regression coefficients are allowed to vary from component to component, thus allowing for a given covariate, say ICS use, to have different effects on the outcome across the K subpopulations underlying the study population. Herein, rather than specifying a mixture of regressions for GA and BW separately, we use the Bayesian bivariate mixture-of-regressions model proposed by Schwartz et al. (56), which accounts for the strong correlation between these 2 perinatal outcomes. From a methodological perspective, our analyses serve as a worked example for illustrating the usefulness of this model for general heterogeneity of effect assessment. METHODS Bayesian finite mixture-of-bivariate-regressions model Prior to presenting our mixture-of-bivariate-regressions model, we refer the reader to Web Appendix 1 (available at https://academic.oup.com/aje) for a review of the univariate mixture-of-regressions model. Therein, we illustrate a key feature of mixtures of regressions, namely, how they model outcome heterogeneity and heterogeneity of effect due to unmeasured categorical predictors. Schwartz et al. (56) introduced a Bayesian finite mixture of bivariate regressions to jointly model GA and BW for effect estimation. They used a mixture specified by bivariate normal distributions, where each component-specific bivariate normal distribution is decomposed as a marginal form times a conditional form. Let G and B be the joint response variables corresponding to GA and BW, respectively. Let A be the exposure of interest, maternal ICS use during pregnancy, and let X be a set of p adjustment variables. Let (gi,bi) be the response data for the ith pregnancy (i=1,…,n), along with ai and xi′=(x1i,…,xpi), the corresponding observed exposure and covariates. Following Schwartz et al. (56), we specify the joint density function for (Gi,Bi) as f(gi,bi|xi,θ)=∑k=1Kπkfk(gi,bi)=∑k=1KπkN(gi|μg,k+aiτg,k+xi′βg,k,σg,k2)×N(bi|μb,k+aiτb,k+xi′βb,k+(gi−(μg,k+aiτg,k+xi′βg,k))β∗k,σb|g,k2), (1) where θ represents the set of unknown parameters of the model. In equation 1, τg,k and τb,k are the kth-component regression coefficients for the exposure (ICS use) associated with GA and BW, respectively. Moreover, the vectors of regression coefficients βg,k and βb,k are the kth-component regression coefficients associated with covariates X. Interestingly, the component-specific bivariate decomposition allows for the interpretation of BW conditional on GA within each component of the mixture: In component k, the coefficient β∗k in the conditional model for B given G encodes the mean change in B for a 1-unit increase in G after the contributions of A and X to explain G have been removed. In other words, it represents the residual association between GA and BW in each subpopulation. The residual variance associated with B(G) within component k is given by σb|g,k2(σg,k2). As such, model 1 (equation 1) allows for a component-specific residual variance in the marginal model for G and in the conditional model for B given G, thus permitting some heteroscedasticity in these residuals overall. To the notable exception of K, which is considered fixed, all parameters in model 1 are unknown and need to be estimated. Assuming independent units, the joint density function of B=(B1,…,Bn) and G=(G1,…,Gn) given x=(x1,…,xn) is f(g,b|x,θ)=Πi=1nf(gi,bi|xi,θ). Inference is based on the posterior distribution of θ, which is proportional to the likelihood times the prior distribution of θ. As in Schwartz et al. (56), this prior distribution is the product of standard prior distributions for the model parameters: a Dirichlet distribution for the mixing proportions (πk), multivariate normal distributions for the regression coefficients, and gamma distributions for the inverse of the residual variances. Data Data on medication prescriptions filled in community pharmacies, outpatient medical visits, emergency-department visits, medical procedures, and hospitalizations were retrieved from 2 administrative databases in Quebec: the Régie de l’assurance maladie du Québec and the Maintenance et exploitation des données pour l’étude de la clientèle hospitalière databases. These data were first used by Cossette et al. (36), who investigated the dose-response relationship between maternal use of ICS and LABA during pregnancy and the adverse perinatal outcomes LBW, PTB, and small size for GA (defined as the <10% BW quantile for a given GA in a population, by sex). Study design Our cohort was formed by applying the same inclusion and exclusion criteria as those used by Cossette et al. (36) to define their cohort. The inclusion criteria were: 1) singleton delivery (live birth or stillbirth) between 1998 and 2008; 2) maternal age ≤45 years at the beginning of pregnancy; 3) ≥1 diagnosis of asthma and ≥1 prescription for an asthma medication filled in the year before or during pregnancy; and 4) drug insurance coverage by the Régie de l’assurance maladie du Québec plan for at least 1 year before and throughout pregnancy. Exclusion criteria were: use of theophylline, cromoglycate, nedocromil, ketotifen, and LABA without ICS during pregnancy. Because our mixture-of-regressions model was defined for a set of independent observations, we kept only the most recent pregnancy if a woman had several pregnancies during follow-up (1,177 pregnancies were excluded). Our choice to select the last recorded pregnancy in the follow-up period rather than the first was made on the grounds that recent pregnancies are more likely to reflect current medical practice regarding the management of asthma during pregnancy. Our cohort comprised a total of 6,197 pregnancies, among which 5,021 (81%) were such that they corresponded to a woman’s unique pregnancy between 1998 and 2008, while the other 1,176 pregnancies (19%) corresponded to the last pregnancy for a woman who had more than 1 pregnancy during follow-up. Outcomes, exposure, and covariates The outcomes of interest were GA at delivery, measured in completed weeks, and BW, measured in grams. Maternal ICS exposure during pregnancy was measured with an algorithm based on prescription renewals that was used in prior studies (58, 59). ICS exposure was classified in 2 categories: use (>0 μg/day) and no use (0 μg/day). Our list of adjustment variables contained 26 identified risk factors for LBW, PTB, or small size for GA (36). These variables, some of which may be potentially confounding, can be divided into the following 4 categories: 1) mother’s and baby’s characteristics: maternal age, baby’s sex, receipt of social assistance, and rural residency; 2) maternal chronic conditions in the year before or during pregnancy: antiphospholipid syndrome, chronic hypertension, diabetes mellitus, and uterine defects; 3) pregnancy-related variables: gestational diabetes, eclampsia/preeclampsia, anemia, placental conditions, placental abruption, vaginal bleeding, maternal infections, fetal-maternal hemorrhage, pregnancy-induced hypertension, and use of β-blockers; and 4) asthma-related variables during pregnancy: leukotriene-receptor antagonists, short-acting β2-agonists, oral corticosteroids, intranasal corticosteroids, ≥1 emergency department visit for asthma, LABA, ≥1 hospitalization for asthma, and severity of asthma (in the year before conception). Statistical analyses Descriptive statistics were used to determine the characteristics of the pregnancies as a function of ICS use. Simple linear models regressing each outcome (GA, BW) as a function of each variable (exposure, adjustment variables) were fitted on our cohort of pregnancies. The associations of ICS with GA and BW were then estimated separately for each outcome using a standard adjusted linear regression model. Because GA is a potential intermediate variable in the pathway between ICS and BW, the BW-adjusted regression model did not include GA. Finally, the mixture model (model 1) with K=1 and K=2 was used to explore heterogeneity of the effects of ICS on GA/BW. The prior distribution on the parameters of model 1 is presented in Table 1. Table 1. Prior and Hyperparameter Values for the Parameters of Mixture-of-Bivariate-Regressions Models of Inhaled Corticosteroid Exposure During Pregnancy Among Asthmatic Women, Quebec, Canada, 1998–2008a Hyperparameter Mixture Model 1-Component Model (K = 1) 2-Component Model (K = 2) k = 1 k = 2 αk 1 1 1 μg,k0 37 37 32 τg,k0 0 0 0 βg,k0 0→ 0→ 0→ Σg,k 100,000I b 100,000I 100,000I μb,k0 3,500 3,100 1,900 τb,k0 0 0 0 βb,k0 0→ 0→ 0→ β∗k0 0 0 0 Σb,k 100,000I 100,000I 100,000I hg,k=hb,k 1 1 1 rg,k=rb,k 1 1 1 Hyperparameter Mixture Model 1-Component Model (K = 1) 2-Component Model (K = 2) k = 1 k = 2 αk 1 1 1 μg,k0 37 37 32 τg,k0 0 0 0 βg,k0 0→ 0→ 0→ Σg,k 100,000I b 100,000I 100,000I μb,k0 3,500 3,100 1,900 τb,k0 0 0 0 βb,k0 0→ 0→ 0→ β∗k0 0 0 0 Σb,k 100,000I 100,000I 100,000I hg,k=hb,k 1 1 1 rg,k=rb,k 1 1 1 a A Dirichlet distribution is specified for the mixing proportions: (π1,…,πK)~Dirichlet(α1,…,αK). Multivariate normal (N) distributions are specified for the gestational age and birth weight regression coefficients: (μg,k,τg,k,βg,k)~N((μg,k0,τg,k0,βg,k0),Σg,k), (μb,k,τb,k,βb,k,β∗k)~N((μb,k0,τb,k0,βb,k0,β∗k0),Σb,k). Gamma distributions are specified for the gestational age and birth weight residual precisions: (σg,k2)−1~Gamma(hg,k,rg,k);(σb|g,k2)−1~Gamma(hb,k,rb,k). bI = identity matrix. Table 1. Prior and Hyperparameter Values for the Parameters of Mixture-of-Bivariate-Regressions Models of Inhaled Corticosteroid Exposure During Pregnancy Among Asthmatic Women, Quebec, Canada, 1998–2008a Hyperparameter Mixture Model 1-Component Model (K = 1) 2-Component Model (K = 2) k = 1 k = 2 αk 1 1 1 μg,k0 37 37 32 τg,k0 0 0 0 βg,k0 0→ 0→ 0→ Σg,k 100,000I b 100,000I 100,000I μb,k0 3,500 3,100 1,900 τb,k0 0 0 0 βb,k0 0→ 0→ 0→ β∗k0 0 0 0 Σb,k 100,000I 100,000I 100,000I hg,k=hb,k 1 1 1 rg,k=rb,k 1 1 1 Hyperparameter Mixture Model 1-Component Model (K = 1) 2-Component Model (K = 2) k = 1 k = 2 αk 1 1 1 μg,k0 37 37 32 τg,k0 0 0 0 βg,k0 0→ 0→ 0→ Σg,k 100,000I b 100,000I 100,000I μb,k0 3,500 3,100 1,900 τb,k0 0 0 0 βb,k0 0→ 0→ 0→ β∗k0 0 0 0 Σb,k 100,000I 100,000I 100,000I hg,k=hb,k 1 1 1 rg,k=rb,k 1 1 1 a A Dirichlet distribution is specified for the mixing proportions: (π1,…,πK)~Dirichlet(α1,…,αK). Multivariate normal (N) distributions are specified for the gestational age and birth weight regression coefficients: (μg,k,τg,k,βg,k)~N((μg,k0,τg,k0,βg,k0),Σg,k), (μb,k,τb,k,βb,k,β∗k)~N((μb,k0,τb,k0,βb,k0,β∗k0),Σb,k). Gamma distributions are specified for the gestational age and birth weight residual precisions: (σg,k2)−1~Gamma(hg,k,rg,k);(σb|g,k2)−1~Gamma(hb,k,rb,k). bI = identity matrix. We implemented our mixture-of-bivariate-regressions model (equation 1) using a Gibbs sampling algorithm for augmented mixture (60) programmed with the R (R Foundation for Statistical Computing, Vienna, Austria) and C (61) languages. For each of the 2 mixture models considered, draws from the posterior distribution of the model parameters were obtained using the Gibbs sampler. A total of 100,000 iterations (burn-in = 100,000 iterations) were retained for each model and used to calculate posterior means and standard deviations and 95% credible intervals. Component allocation was monitored for the mixture model with K = 2. More precisely, while component membership was unknown for each pregnancy in our cohort, the proportion of times that each pregnancy is attributed to a given component over the Gibbs sampling iterations is an estimation of the posterior probability that the pregnancy belongs to that component. Hard clustering was done by assigning a pregnancy to the component with the highest posterior probability. An indicator variable Z was created and set to Zi=1 whenever pregnancy i was allocated more than 50% of times to the second component ( Zi=0 otherwise) over the Gibbs sampling iterations. We then used standard multiple logistic regression with outcome Z to investigate whether any variable was associated with this indicator. Sensitivity analyses regarding the choice of priors and adjustment covariates were performed subsequently; results are presented in Web Appendix 2. We also conducted unplanned analyses to obtain further insights on the problem and results: We performed 3 standard adjusted logistic regression analyses using LBW, PTB, and very PTB (GA <32 weeks) as outcome variables. Finally, because many perinatal studies rely on a GA-normalized BW variable as the primary outcome instead of or in addition to BW (62, 63), a novel application of the mixture of regressions with K = 2 was done to model ICS exposure effect heterogeneity using GA and BW z scores as response variables (with the priors for the BW z score intercepts centered at 0). The BW z scores were computed using a Canadian chart described by Kramer et al. (64). Approval from the Commission d’accès à l’information du Québec was obtained prior to requesting and linking the information from the Maintenance et exploitation des données pour l’étude de la clientèle hospitalière and Régie de l’assurance maladie du Québec databases. These analyses were approved by the ethics committee of the Hôpital du Sacré-Coeur de Montréal. RESULTS In our cohort, 3,461 babies out of 6,197 (55.8%) were exposed to ICS during their mother’s pregnancy. Among pregnancies exposed to ICS, 86.3% were exposed to a dose of less than 250 μg/day (fluticasone propionate equivalent) on average. Descriptive statistics are presented in Table 2 according to binary exposure status. A larger proportion of women treated with ICS received social assistance. Overall, asthma-related variables featured the greatest imbalance between pregnancies exposed to ICS and those unexposed. Notably, pregnancies exposed to ICS were associated with more severe asthma, were associated with more visits to emergency departments because of asthma, and were more exposed to leukotriene-receptor antagonists, short-acting β2-agonists, oral corticosteroids, and LABA (by design, no pregnancies unexposed to ICS were exposed to LABA in our cohort). Table 2. Distribution of Potentially Confounding Covariates According to Inhaled Corticosteroid Exposure During Pregnancy Among Asthmatic Women, Quebec, Canada, 1998–2008 Characteristic ICS Exposure Status Not Exposed Exposed No. % No. % No. of pregnancies 2,736 44.15 3,461 55.85 Mother’s and baby’s characteristics  Maternal age, years   <18 39 1.43 50 1.44   18–34 2,349 85.86 2,877 83.13   >34 348 12.72 534 15.43  Baby’s sex (female vs. male) 1,362 49.78 1,668 48.19  Receipt of social assistance 1,370 50.07 2,015 58.22  Rural residency 517 18.90 685 19.79 Maternal chronic conditions  Antiphospholipid syndrome 16 0.58 21 0.61  Chronic hypertension 75 2.74 118 3.41  Diabetes mellitus 87 3.18 136 3.93  Uterine defects 363 13.27 495 14.30 Pregnancy-related variables  Gestational diabetes 259 9.47 376 10.86  Eclampsia/preeclampsia 78 2.85 116 3.35  Anemia 385 14.07 509 14.71  Placental conditions 107 3.91 137 3.96  Placental abruption 95 3.47 125 3.61  Vaginal bleeding 381 13.93 448 12.94  Maternal infections 418 15.28 490 14.16  Fetal-maternal hemorrhage 5 0.18 10 0.29  Pregnancy-induced hypertension 167 6.10 241 6.96  Use of β-blockers 23 0.84 31 0.90 Asthma-related variables  Leukotriene-receptor antagonists 9 0.33 90 2.60  No. of SABA doses per week   0 1,549 56.62 434 12.54   0.01–3 762 27.85 1,346 38.89   >3 425 15.53 1,681 48.57  Oral corticosteroids 104 3.80 589 17.02  Intranasal corticosteroids 196 7.16 596 17.22  ≥1 ED visit for asthma 170 6.21 661 19.10  LABA 0 0 553 15.98  ≥1 hospitalization for asthma 6 0.22 63 1.82  Severity of asthma prior to pregnancy   Mild 2,479 90.61 2,499 72.20   Moderate 226 8.26 623 18.00   Severe 31 1.13 339 9.79 Characteristic ICS Exposure Status Not Exposed Exposed No. % No. % No. of pregnancies 2,736 44.15 3,461 55.85 Mother’s and baby’s characteristics  Maternal age, years   <18 39 1.43 50 1.44   18–34 2,349 85.86 2,877 83.13   >34 348 12.72 534 15.43  Baby’s sex (female vs. male) 1,362 49.78 1,668 48.19  Receipt of social assistance 1,370 50.07 2,015 58.22  Rural residency 517 18.90 685 19.79 Maternal chronic conditions  Antiphospholipid syndrome 16 0.58 21 0.61  Chronic hypertension 75 2.74 118 3.41  Diabetes mellitus 87 3.18 136 3.93  Uterine defects 363 13.27 495 14.30 Pregnancy-related variables  Gestational diabetes 259 9.47 376 10.86  Eclampsia/preeclampsia 78 2.85 116 3.35  Anemia 385 14.07 509 14.71  Placental conditions 107 3.91 137 3.96  Placental abruption 95 3.47 125 3.61  Vaginal bleeding 381 13.93 448 12.94  Maternal infections 418 15.28 490 14.16  Fetal-maternal hemorrhage 5 0.18 10 0.29  Pregnancy-induced hypertension 167 6.10 241 6.96  Use of β-blockers 23 0.84 31 0.90 Asthma-related variables  Leukotriene-receptor antagonists 9 0.33 90 2.60  No. of SABA doses per week   0 1,549 56.62 434 12.54   0.01–3 762 27.85 1,346 38.89   >3 425 15.53 1,681 48.57  Oral corticosteroids 104 3.80 589 17.02  Intranasal corticosteroids 196 7.16 596 17.22  ≥1 ED visit for asthma 170 6.21 661 19.10  LABA 0 0 553 15.98  ≥1 hospitalization for asthma 6 0.22 63 1.82  Severity of asthma prior to pregnancy   Mild 2,479 90.61 2,499 72.20   Moderate 226 8.26 623 18.00   Severe 31 1.13 339 9.79 Abbreviations: ED, emergency department; ICS, inhaled corticosteroids; LABA, long-acting β2-agonists; SABA, short-acting β2-agonists. Table 2. Distribution of Potentially Confounding Covariates According to Inhaled Corticosteroid Exposure During Pregnancy Among Asthmatic Women, Quebec, Canada, 1998–2008 Characteristic ICS Exposure Status Not Exposed Exposed No. % No. % No. of pregnancies 2,736 44.15 3,461 55.85 Mother’s and baby’s characteristics  Maternal age, years   <18 39 1.43 50 1.44   18–34 2,349 85.86 2,877 83.13   >34 348 12.72 534 15.43  Baby’s sex (female vs. male) 1,362 49.78 1,668 48.19  Receipt of social assistance 1,370 50.07 2,015 58.22  Rural residency 517 18.90 685 19.79 Maternal chronic conditions  Antiphospholipid syndrome 16 0.58 21 0.61  Chronic hypertension 75 2.74 118 3.41  Diabetes mellitus 87 3.18 136 3.93  Uterine defects 363 13.27 495 14.30 Pregnancy-related variables  Gestational diabetes 259 9.47 376 10.86  Eclampsia/preeclampsia 78 2.85 116 3.35  Anemia 385 14.07 509 14.71  Placental conditions 107 3.91 137 3.96  Placental abruption 95 3.47 125 3.61  Vaginal bleeding 381 13.93 448 12.94  Maternal infections 418 15.28 490 14.16  Fetal-maternal hemorrhage 5 0.18 10 0.29  Pregnancy-induced hypertension 167 6.10 241 6.96  Use of β-blockers 23 0.84 31 0.90 Asthma-related variables  Leukotriene-receptor antagonists 9 0.33 90 2.60  No. of SABA doses per week   0 1,549 56.62 434 12.54   0.01–3 762 27.85 1,346 38.89   >3 425 15.53 1,681 48.57  Oral corticosteroids 104 3.80 589 17.02  Intranasal corticosteroids 196 7.16 596 17.22  ≥1 ED visit for asthma 170 6.21 661 19.10  LABA 0 0 553 15.98  ≥1 hospitalization for asthma 6 0.22 63 1.82  Severity of asthma prior to pregnancy   Mild 2,479 90.61 2,499 72.20   Moderate 226 8.26 623 18.00   Severe 31 1.13 339 9.79 Characteristic ICS Exposure Status Not Exposed Exposed No. % No. % No. of pregnancies 2,736 44.15 3,461 55.85 Mother’s and baby’s characteristics  Maternal age, years   <18 39 1.43 50 1.44   18–34 2,349 85.86 2,877 83.13   >34 348 12.72 534 15.43  Baby’s sex (female vs. male) 1,362 49.78 1,668 48.19  Receipt of social assistance 1,370 50.07 2,015 58.22  Rural residency 517 18.90 685 19.79 Maternal chronic conditions  Antiphospholipid syndrome 16 0.58 21 0.61  Chronic hypertension 75 2.74 118 3.41  Diabetes mellitus 87 3.18 136 3.93  Uterine defects 363 13.27 495 14.30 Pregnancy-related variables  Gestational diabetes 259 9.47 376 10.86  Eclampsia/preeclampsia 78 2.85 116 3.35  Anemia 385 14.07 509 14.71  Placental conditions 107 3.91 137 3.96  Placental abruption 95 3.47 125 3.61  Vaginal bleeding 381 13.93 448 12.94  Maternal infections 418 15.28 490 14.16  Fetal-maternal hemorrhage 5 0.18 10 0.29  Pregnancy-induced hypertension 167 6.10 241 6.96  Use of β-blockers 23 0.84 31 0.90 Asthma-related variables  Leukotriene-receptor antagonists 9 0.33 90 2.60  No. of SABA doses per week   0 1,549 56.62 434 12.54   0.01–3 762 27.85 1,346 38.89   >3 425 15.53 1,681 48.57  Oral corticosteroids 104 3.80 589 17.02  Intranasal corticosteroids 196 7.16 596 17.22  ≥1 ED visit for asthma 170 6.21 661 19.10  LABA 0 0 553 15.98  ≥1 hospitalization for asthma 6 0.22 63 1.82  Severity of asthma prior to pregnancy   Mild 2,479 90.61 2,499 72.20   Moderate 226 8.26 623 18.00   Severe 31 1.13 339 9.79 Abbreviations: ED, emergency department; ICS, inhaled corticosteroids; LABA, long-acting β2-agonists; SABA, short-acting β2-agonists. The crude results derived using standard univariate linear regression models are presented in Table 3. We observed that pregnancies exposed and unexposed to ICS were similar in terms of mean GA and BW. The standard univariate-adjusted analysis for GA (second and third columns in Table 4) revealed a positive association between ICS and GA (for GA associated with ICS, mean difference = 0.17 weeks, 95% confidence interval (CI): 0.05, 0.29). Although GA is arguably the strongest predictor of BW, the small positive association between ICS and GA did not translate to a significant association between ICS and BW (mean difference = 29.16 g, 95% CI: −4.81, 63.12). The regression coefficients for GA and BW obtained from the mixture model with K = 1 (last 2 columns in Table 4) were similar to those obtained from the standard univariate-adjusted regression models. Table 3. Associations Between Selected Characteristics of Asthmatic Pregnant Women and Their Pregnancy Outcomes, Quebec, Canada, 1998–2008a Characteristicb Pregnancy Outcome Gestational Age, weeks Birth Weight, g RC 95% CI RC 95% CI Gestational age, weeks N/A 185.97 180.52, 191.43 ICS exposure 0.08 −0.03, 0.18 1.54 −27.79, 30.87 Mother’s and baby’s characteristics  Maternal age, years   <18 0 Referent 0 Referent   18–34 −0.05 −0.47, 0.38 147.08 24.59, 269.57   >34 −0.38 −0.82, 0.06 111.91 −15.53, 239.36  Baby’s sex (female vs. male) −0.02 −0.12, 0.08 −147.41 −176.31, −118.51  Receipt of social assistance −0.29 −0.39, −0.19 −112.44 −141.56, −83.32  Rural residency −0.16 −0.29, −0.03 −40.03 −76.85, −3.21 Maternal chronic conditions  Antiphospholipid syndrome −1.09 −1.75, −0.43 −322.57 −511.45, −133.69  Chronic hypertension −0.54 −0.83, −0.25 −55.23 −139.06, 28.60  Diabetes mellitus −0.66 −0.93, −0.39 33.10 −45.09, 111.29  Uterine defects −0.59 −0.74, −0.44 −35.87 −78.03, 6.29 Pregnancy-related variables  Gestational diabetes −0.29 −0.46, −0.12 63.98 15.98, 111.98  Eclampsia/preeclampsia −1.33 −1.62, −1.04 −272.27 −355.63, −188.91  Anemia −0.12 −0.27, 0.02 58.25 16.83, 99.68  Placental conditions −0.80 −1.06, −0.54 −239.88 −314.53, −165.23  Placental abruption −1.85 −2.12, −1.58 −466.03 −543.88, −388.18  Vaginal bleeding −0.88 −1.02, −0.73 −207.58 −250.05, −165.10  Maternal infections −0.31 −0.46, −0.17 −33.40 −74.58, 7.78  Fetal-maternal hemorrhage −2.84 −3.87, −1.81 −492.31 −788.45, −196.18  Pregnancy-induced hypertension −0.30 −0.51, −0.10 −22.43 −81.15, 36.30  Use of β-blockers −0.48 −1.03, 0.07 −94.24 −250.93, 62.44 Asthma-related variables  Leukotriene-receptor antagonists −0.13 −0.53, 0.28 −76.21 −192.35, 39.94  No. of SABA doses per week   0 0 Referent 0 Referent   0.01–3 0.12 −0.01, 0.24 7.92 −27.93, 43.76   >3 −0.08 −0.20, 0.05 −41.16 −77.02, −5.31  Oral corticosteroids −0.07 −0.23, 0.09 −29.43 −75.63, 16.78  Intranasal corticosteroids −0.05 −0.20, 0.11 21.80 −21.82, 65.42  ≥1 ED visit for asthma 0.10 −0.05, 0.25 9.53 −33.21, 52.27  LABA −0.12 −0.30, 0.06 −20.40 −71.49, 30.68  ≥1 hospitalization for asthma −0.16 −0.65, 0.32 −101.28 −240.06, 37.49  Severity of asthma prior to pregnancy   Mild 0 Referent 0 Referent   Moderate 0.10 −0.05, 0.25 4.44 −38.10, 46.98   Severe −0.08 −0.30, 0.13 −98.31 −160.05, −36.58 Characteristicb Pregnancy Outcome Gestational Age, weeks Birth Weight, g RC 95% CI RC 95% CI Gestational age, weeks N/A 185.97 180.52, 191.43 ICS exposure 0.08 −0.03, 0.18 1.54 −27.79, 30.87 Mother’s and baby’s characteristics  Maternal age, years   <18 0 Referent 0 Referent   18–34 −0.05 −0.47, 0.38 147.08 24.59, 269.57   >34 −0.38 −0.82, 0.06 111.91 −15.53, 239.36  Baby’s sex (female vs. male) −0.02 −0.12, 0.08 −147.41 −176.31, −118.51  Receipt of social assistance −0.29 −0.39, −0.19 −112.44 −141.56, −83.32  Rural residency −0.16 −0.29, −0.03 −40.03 −76.85, −3.21 Maternal chronic conditions  Antiphospholipid syndrome −1.09 −1.75, −0.43 −322.57 −511.45, −133.69  Chronic hypertension −0.54 −0.83, −0.25 −55.23 −139.06, 28.60  Diabetes mellitus −0.66 −0.93, −0.39 33.10 −45.09, 111.29  Uterine defects −0.59 −0.74, −0.44 −35.87 −78.03, 6.29 Pregnancy-related variables  Gestational diabetes −0.29 −0.46, −0.12 63.98 15.98, 111.98  Eclampsia/preeclampsia −1.33 −1.62, −1.04 −272.27 −355.63, −188.91  Anemia −0.12 −0.27, 0.02 58.25 16.83, 99.68  Placental conditions −0.80 −1.06, −0.54 −239.88 −314.53, −165.23  Placental abruption −1.85 −2.12, −1.58 −466.03 −543.88, −388.18  Vaginal bleeding −0.88 −1.02, −0.73 −207.58 −250.05, −165.10  Maternal infections −0.31 −0.46, −0.17 −33.40 −74.58, 7.78  Fetal-maternal hemorrhage −2.84 −3.87, −1.81 −492.31 −788.45, −196.18  Pregnancy-induced hypertension −0.30 −0.51, −0.10 −22.43 −81.15, 36.30  Use of β-blockers −0.48 −1.03, 0.07 −94.24 −250.93, 62.44 Asthma-related variables  Leukotriene-receptor antagonists −0.13 −0.53, 0.28 −76.21 −192.35, 39.94  No. of SABA doses per week   0 0 Referent 0 Referent   0.01–3 0.12 −0.01, 0.24 7.92 −27.93, 43.76   >3 −0.08 −0.20, 0.05 −41.16 −77.02, −5.31  Oral corticosteroids −0.07 −0.23, 0.09 −29.43 −75.63, 16.78  Intranasal corticosteroids −0.05 −0.20, 0.11 21.80 −21.82, 65.42  ≥1 ED visit for asthma 0.10 −0.05, 0.25 9.53 −33.21, 52.27  LABA −0.12 −0.30, 0.06 −20.40 −71.49, 30.68  ≥1 hospitalization for asthma −0.16 −0.65, 0.32 −101.28 −240.06, 37.49  Severity of asthma prior to pregnancy   Mild 0 Referent 0 Referent   Moderate 0.10 −0.05, 0.25 4.44 −38.10, 46.98   Severe −0.08 −0.30, 0.13 −98.31 −160.05, −36.58 Abbreviations: CI, confidence interval; ED, emergency department; ICS, inhaled corticosteroids; LABA, long-acting β2-agonists; N/A, not applicable; RC, regression coefficient; SABA, short-acting β2-agonists. a Regression coefficients with 95% CIs were obtained from simple univariate linear models. b All variables are binary (yes = 1/no = 0) unless otherwise indicated. Table 3. Associations Between Selected Characteristics of Asthmatic Pregnant Women and Their Pregnancy Outcomes, Quebec, Canada, 1998–2008a Characteristicb Pregnancy Outcome Gestational Age, weeks Birth Weight, g RC 95% CI RC 95% CI Gestational age, weeks N/A 185.97 180.52, 191.43 ICS exposure 0.08 −0.03, 0.18 1.54 −27.79, 30.87 Mother’s and baby’s characteristics  Maternal age, years   <18 0 Referent 0 Referent   18–34 −0.05 −0.47, 0.38 147.08 24.59, 269.57   >34 −0.38 −0.82, 0.06 111.91 −15.53, 239.36  Baby’s sex (female vs. male) −0.02 −0.12, 0.08 −147.41 −176.31, −118.51  Receipt of social assistance −0.29 −0.39, −0.19 −112.44 −141.56, −83.32  Rural residency −0.16 −0.29, −0.03 −40.03 −76.85, −3.21 Maternal chronic conditions  Antiphospholipid syndrome −1.09 −1.75, −0.43 −322.57 −511.45, −133.69  Chronic hypertension −0.54 −0.83, −0.25 −55.23 −139.06, 28.60  Diabetes mellitus −0.66 −0.93, −0.39 33.10 −45.09, 111.29  Uterine defects −0.59 −0.74, −0.44 −35.87 −78.03, 6.29 Pregnancy-related variables  Gestational diabetes −0.29 −0.46, −0.12 63.98 15.98, 111.98  Eclampsia/preeclampsia −1.33 −1.62, −1.04 −272.27 −355.63, −188.91  Anemia −0.12 −0.27, 0.02 58.25 16.83, 99.68  Placental conditions −0.80 −1.06, −0.54 −239.88 −314.53, −165.23  Placental abruption −1.85 −2.12, −1.58 −466.03 −543.88, −388.18  Vaginal bleeding −0.88 −1.02, −0.73 −207.58 −250.05, −165.10  Maternal infections −0.31 −0.46, −0.17 −33.40 −74.58, 7.78  Fetal-maternal hemorrhage −2.84 −3.87, −1.81 −492.31 −788.45, −196.18  Pregnancy-induced hypertension −0.30 −0.51, −0.10 −22.43 −81.15, 36.30  Use of β-blockers −0.48 −1.03, 0.07 −94.24 −250.93, 62.44 Asthma-related variables  Leukotriene-receptor antagonists −0.13 −0.53, 0.28 −76.21 −192.35, 39.94  No. of SABA doses per week   0 0 Referent 0 Referent   0.01–3 0.12 −0.01, 0.24 7.92 −27.93, 43.76   >3 −0.08 −0.20, 0.05 −41.16 −77.02, −5.31  Oral corticosteroids −0.07 −0.23, 0.09 −29.43 −75.63, 16.78  Intranasal corticosteroids −0.05 −0.20, 0.11 21.80 −21.82, 65.42  ≥1 ED visit for asthma 0.10 −0.05, 0.25 9.53 −33.21, 52.27  LABA −0.12 −0.30, 0.06 −20.40 −71.49, 30.68  ≥1 hospitalization for asthma −0.16 −0.65, 0.32 −101.28 −240.06, 37.49  Severity of asthma prior to pregnancy   Mild 0 Referent 0 Referent   Moderate 0.10 −0.05, 0.25 4.44 −38.10, 46.98   Severe −0.08 −0.30, 0.13 −98.31 −160.05, −36.58 Characteristicb Pregnancy Outcome Gestational Age, weeks Birth Weight, g RC 95% CI RC 95% CI Gestational age, weeks N/A 185.97 180.52, 191.43 ICS exposure 0.08 −0.03, 0.18 1.54 −27.79, 30.87 Mother’s and baby’s characteristics  Maternal age, years   <18 0 Referent 0 Referent   18–34 −0.05 −0.47, 0.38 147.08 24.59, 269.57   >34 −0.38 −0.82, 0.06 111.91 −15.53, 239.36  Baby’s sex (female vs. male) −0.02 −0.12, 0.08 −147.41 −176.31, −118.51  Receipt of social assistance −0.29 −0.39, −0.19 −112.44 −141.56, −83.32  Rural residency −0.16 −0.29, −0.03 −40.03 −76.85, −3.21 Maternal chronic conditions  Antiphospholipid syndrome −1.09 −1.75, −0.43 −322.57 −511.45, −133.69  Chronic hypertension −0.54 −0.83, −0.25 −55.23 −139.06, 28.60  Diabetes mellitus −0.66 −0.93, −0.39 33.10 −45.09, 111.29  Uterine defects −0.59 −0.74, −0.44 −35.87 −78.03, 6.29 Pregnancy-related variables  Gestational diabetes −0.29 −0.46, −0.12 63.98 15.98, 111.98  Eclampsia/preeclampsia −1.33 −1.62, −1.04 −272.27 −355.63, −188.91  Anemia −0.12 −0.27, 0.02 58.25 16.83, 99.68  Placental conditions −0.80 −1.06, −0.54 −239.88 −314.53, −165.23  Placental abruption −1.85 −2.12, −1.58 −466.03 −543.88, −388.18  Vaginal bleeding −0.88 −1.02, −0.73 −207.58 −250.05, −165.10  Maternal infections −0.31 −0.46, −0.17 −33.40 −74.58, 7.78  Fetal-maternal hemorrhage −2.84 −3.87, −1.81 −492.31 −788.45, −196.18  Pregnancy-induced hypertension −0.30 −0.51, −0.10 −22.43 −81.15, 36.30  Use of β-blockers −0.48 −1.03, 0.07 −94.24 −250.93, 62.44 Asthma-related variables  Leukotriene-receptor antagonists −0.13 −0.53, 0.28 −76.21 −192.35, 39.94  No. of SABA doses per week   0 0 Referent 0 Referent   0.01–3 0.12 −0.01, 0.24 7.92 −27.93, 43.76   >3 −0.08 −0.20, 0.05 −41.16 −77.02, −5.31  Oral corticosteroids −0.07 −0.23, 0.09 −29.43 −75.63, 16.78  Intranasal corticosteroids −0.05 −0.20, 0.11 21.80 −21.82, 65.42  ≥1 ED visit for asthma 0.10 −0.05, 0.25 9.53 −33.21, 52.27  LABA −0.12 −0.30, 0.06 −20.40 −71.49, 30.68  ≥1 hospitalization for asthma −0.16 −0.65, 0.32 −101.28 −240.06, 37.49  Severity of asthma prior to pregnancy   Mild 0 Referent 0 Referent   Moderate 0.10 −0.05, 0.25 4.44 −38.10, 46.98   Severe −0.08 −0.30, 0.13 −98.31 −160.05, −36.58 Abbreviations: CI, confidence interval; ED, emergency department; ICS, inhaled corticosteroids; LABA, long-acting β2-agonists; N/A, not applicable; RC, regression coefficient; SABA, short-acting β2-agonists. a Regression coefficients with 95% CIs were obtained from simple univariate linear models. b All variables are binary (yes = 1/no = 0) unless otherwise indicated. Table 4. Associations of Inhaled Corticosteroid Exposure With Gestational Age and Birth Weight (Adjusted Models) Among Asthmatic Pregnant Women, Quebec, Canada, 1998–2008 Characteristica Univariate-Adjusted Model 1-Component Adjusted Mixture Model GA, weeks BW, g GA, weeks BW, g RC 95% CI RC 95% CI RC 95% CrI RC 95% CrI Intercept 38.98 38.55, 39.40 3,256.19 3,133.98, 3,378.40 39.02 38.60, 39.44 3,274.01 3,157.17, 3,390.80 ICS exposure 0.17 0.05, 0.29 29.16 −4.81, 63.12 0.17 0.05, 0.29 28.78 −5.14, 62.51 Mother’s and baby’s characteristics  Maternal age, years   <18 0 Referent 0 Referent 0 Referent 0 Referent   18–34 0.05 −0.36, 0.47 157.72 38.89, 276.55 0.01 −0.39, 0.42 140.05 26.61, 252.76   >34 −0.19 −0.62, 0.24 113.81 −10.33, 237.94 −0.23 −0.65, 0.19 95.80 −22.97, 213.66  Baby’s sex (female vs. male) 0.01 −0.09, 0.11 −142.11 −170.32, −113.91 0.01 −0.09, 0.11 −142.37 −170.51, −114.27  Receipt of social assistance −0.30 −0.40, −0.20 −118.13 −146.87, −89.39 −0.30 −0.40, −0.20 −118.02 −146.69, −89.44  Rural residency −0.21 −0.33, −0.08 −50.49 −86.44, −14.54 −0.21 −0.33, −0.08 −50.49 −86.41, −14.73 Maternal chronic conditions  Antiphospholipid syndrome −0.92 −1.55, −0.28 −341.40 −524.57, −158.22 −0.86 −1.48, −0.23 −313.12 −488.25, −138.52  Chronic hypertension −0.23 −0.52, 0.06 −32.30 −116.42, 51.82 −0.23 −0.52, 0.06 −32.02 −114.92, 50.86  Diabetes mellitus −0.38 −0.65, −0.12 105.14 28.58, 181.70 −0.39 −0.65, −0.12 102.76 26.36, 178.47  Uterine defects −0.47 −0.61, −0.32 −28.85 −70.22, 12.53 −0.47 −0.61, −0.32 −28.71 −69.94, 12.25 Pregnancy-related variables  Gestational diabetes −0.20 −0.36, −0.03 81.78 34.46, 129.11 −0.20 −0.36, −0.03 81.00 33.65, 128.08  Eclampsia/preeclampsia −1.22 −1.51, −0.92 −270.91 −356.08, −185.74 −1.21 −1.50, −0.91 −266.56 −350.73, −182.23  Anemia −0.04 −0.18, 0.10 74.53 34.18, 114.89 −0.04 −0.18, 0.10 73.70 33.66, 113.68  Placental conditions −0.33 −0.59, −0.07 −144.19 −218.69, −69.69 −0.33 −0.58, −0.07 −143.47 −216.75, −70.32  Placental abruption −1.34 −1.64, −1.03 −356.21 −443.47, −268.96 −1.32 −1.62, −1.02 −350.66 −436.76, −264.08  Vaginal bleeding −0.42 −0.59, −0.25 −92.93 −141.19, −44.68 −0.42 −0.59, −0.26 −94.39 −142.35, −46.30  Maternal infections −0.17 −0.31, −0.03 −9.77 −49.99, 30.45 −0.17 −0.31, −0.03 −9.91 −50.15, 30.26  Fetal-maternal hemorrhage −2.40 −3.40, −1.40 −335.49 −623.96, −47.02 −2.28 −3.23, −1.31 −277.66 −538.28, −16.82  Pregnancy-induced hypertension −0.01 −0.22, 0.20 17.90 −42.12, 77.92 −0.01 −0.22, 0.19 16.95 −42.28, 76.64  Use of β-blockers −0.12 −0.66, 0.42 −64.52 −219.89, 90.84 −0.11 −0.64, 0.42 −61.00 −210.51, 88.67 Asthma-related variables  Leukotriene-receptor antagonists 0.03 −0.38, 0.43 −30.25 −146.78, 86.27 0.03 −0.37, 0.43 −29.39 −144.01, 85.32  No. of SABA doses per week   0 0 Referent 0 Referent 0 Referent 0 Referent   0.01–3 0.06 −0.08, 0.19 −1.93 −39.56, 35.70 0.05 −0.08, 0.19 −2.15 −39.68, 35.51   >3 −0.18 −0.33, −0.04 −44.09 −86.45, −1.72 −0.18 −0.33, −0.04 −44.34 −86.88, −1.57  Oral corticosteroids −0.08 −0.26, 0.10 −29.51 −80.98, 21.96 −0.08 −0.26, 0.10 −28.86 −80.08, 22.19  Intranasal corticosteroids −0.01 −0.15, 0.14 23.75 −19.35, 66.85 −0.01 −0.15, 0.14 23.54 −19.33, 66.36  ≥1 ED visit for asthma 0.13 −0.03, 0.30 30.54 −16.32, 77.40 0.13 −0.03, 0.29 30.19 −16.68, 76.65  LABA −0.11 −0.30, 0.08 1.39 −53.35, 56.13 −0.11 −0.30, 0.08 1.60 −52.88, 56.02  ≥1 hospitalization for asthma −0.08 −0.56, 0.41 −81.84 −220.82, 57.14 −0.07 −0.54, 0.41 −78.29 −214.34, 57.85  Severity of asthma prior to pregnancy   Mild 0 Referent 0 Referent 0 Referent 0 Referent   Moderate 0.17 0.02, 0.33 11.06 −33.85, 55.97 0.17 0.02, 0.33 10.92 −34.01, 55.62   Severe 0.06 −0.17, 0.29 −71.50 −138.28, −4.72 0.06 −0.17, 0.29 −71.36 −137.81, −4.57 GA residual 184.83 179.31, 190.37 Characteristica Univariate-Adjusted Model 1-Component Adjusted Mixture Model GA, weeks BW, g GA, weeks BW, g RC 95% CI RC 95% CI RC 95% CrI RC 95% CrI Intercept 38.98 38.55, 39.40 3,256.19 3,133.98, 3,378.40 39.02 38.60, 39.44 3,274.01 3,157.17, 3,390.80 ICS exposure 0.17 0.05, 0.29 29.16 −4.81, 63.12 0.17 0.05, 0.29 28.78 −5.14, 62.51 Mother’s and baby’s characteristics  Maternal age, years   <18 0 Referent 0 Referent 0 Referent 0 Referent   18–34 0.05 −0.36, 0.47 157.72 38.89, 276.55 0.01 −0.39, 0.42 140.05 26.61, 252.76   >34 −0.19 −0.62, 0.24 113.81 −10.33, 237.94 −0.23 −0.65, 0.19 95.80 −22.97, 213.66  Baby’s sex (female vs. male) 0.01 −0.09, 0.11 −142.11 −170.32, −113.91 0.01 −0.09, 0.11 −142.37 −170.51, −114.27  Receipt of social assistance −0.30 −0.40, −0.20 −118.13 −146.87, −89.39 −0.30 −0.40, −0.20 −118.02 −146.69, −89.44  Rural residency −0.21 −0.33, −0.08 −50.49 −86.44, −14.54 −0.21 −0.33, −0.08 −50.49 −86.41, −14.73 Maternal chronic conditions  Antiphospholipid syndrome −0.92 −1.55, −0.28 −341.40 −524.57, −158.22 −0.86 −1.48, −0.23 −313.12 −488.25, −138.52  Chronic hypertension −0.23 −0.52, 0.06 −32.30 −116.42, 51.82 −0.23 −0.52, 0.06 −32.02 −114.92, 50.86  Diabetes mellitus −0.38 −0.65, −0.12 105.14 28.58, 181.70 −0.39 −0.65, −0.12 102.76 26.36, 178.47  Uterine defects −0.47 −0.61, −0.32 −28.85 −70.22, 12.53 −0.47 −0.61, −0.32 −28.71 −69.94, 12.25 Pregnancy-related variables  Gestational diabetes −0.20 −0.36, −0.03 81.78 34.46, 129.11 −0.20 −0.36, −0.03 81.00 33.65, 128.08  Eclampsia/preeclampsia −1.22 −1.51, −0.92 −270.91 −356.08, −185.74 −1.21 −1.50, −0.91 −266.56 −350.73, −182.23  Anemia −0.04 −0.18, 0.10 74.53 34.18, 114.89 −0.04 −0.18, 0.10 73.70 33.66, 113.68  Placental conditions −0.33 −0.59, −0.07 −144.19 −218.69, −69.69 −0.33 −0.58, −0.07 −143.47 −216.75, −70.32  Placental abruption −1.34 −1.64, −1.03 −356.21 −443.47, −268.96 −1.32 −1.62, −1.02 −350.66 −436.76, −264.08  Vaginal bleeding −0.42 −0.59, −0.25 −92.93 −141.19, −44.68 −0.42 −0.59, −0.26 −94.39 −142.35, −46.30  Maternal infections −0.17 −0.31, −0.03 −9.77 −49.99, 30.45 −0.17 −0.31, −0.03 −9.91 −50.15, 30.26  Fetal-maternal hemorrhage −2.40 −3.40, −1.40 −335.49 −623.96, −47.02 −2.28 −3.23, −1.31 −277.66 −538.28, −16.82  Pregnancy-induced hypertension −0.01 −0.22, 0.20 17.90 −42.12, 77.92 −0.01 −0.22, 0.19 16.95 −42.28, 76.64  Use of β-blockers −0.12 −0.66, 0.42 −64.52 −219.89, 90.84 −0.11 −0.64, 0.42 −61.00 −210.51, 88.67 Asthma-related variables  Leukotriene-receptor antagonists 0.03 −0.38, 0.43 −30.25 −146.78, 86.27 0.03 −0.37, 0.43 −29.39 −144.01, 85.32  No. of SABA doses per week   0 0 Referent 0 Referent 0 Referent 0 Referent   0.01–3 0.06 −0.08, 0.19 −1.93 −39.56, 35.70 0.05 −0.08, 0.19 −2.15 −39.68, 35.51   >3 −0.18 −0.33, −0.04 −44.09 −86.45, −1.72 −0.18 −0.33, −0.04 −44.34 −86.88, −1.57  Oral corticosteroids −0.08 −0.26, 0.10 −29.51 −80.98, 21.96 −0.08 −0.26, 0.10 −28.86 −80.08, 22.19  Intranasal corticosteroids −0.01 −0.15, 0.14 23.75 −19.35, 66.85 −0.01 −0.15, 0.14 23.54 −19.33, 66.36  ≥1 ED visit for asthma 0.13 −0.03, 0.30 30.54 −16.32, 77.40 0.13 −0.03, 0.29 30.19 −16.68, 76.65  LABA −0.11 −0.30, 0.08 1.39 −53.35, 56.13 −0.11 −0.30, 0.08 1.60 −52.88, 56.02  ≥1 hospitalization for asthma −0.08 −0.56, 0.41 −81.84 −220.82, 57.14 −0.07 −0.54, 0.41 −78.29 −214.34, 57.85  Severity of asthma prior to pregnancy   Mild 0 Referent 0 Referent 0 Referent 0 Referent   Moderate 0.17 0.02, 0.33 11.06 −33.85, 55.97 0.17 0.02, 0.33 10.92 −34.01, 55.62   Severe 0.06 −0.17, 0.29 −71.50 −138.28, −4.72 0.06 −0.17, 0.29 −71.36 −137.81, −4.57 GA residual 184.83 179.31, 190.37 Abbreviations: BW, birth weight; CI, confidence interval; CrI, credible interval; ED, emergency department; GA, gestational age; ICS, inhaled corticosteroids; LABA, long-acting β2-agonists; RC, regression coefficient; SABA, short-acting β2-agonists. a All adjustment variables are binary (yes = 1/no = 0) unless otherwise indicated. Table 4. Associations of Inhaled Corticosteroid Exposure With Gestational Age and Birth Weight (Adjusted Models) Among Asthmatic Pregnant Women, Quebec, Canada, 1998–2008 Characteristica Univariate-Adjusted Model 1-Component Adjusted Mixture Model GA, weeks BW, g GA, weeks BW, g RC 95% CI RC 95% CI RC 95% CrI RC 95% CrI Intercept 38.98 38.55, 39.40 3,256.19 3,133.98, 3,378.40 39.02 38.60, 39.44 3,274.01 3,157.17, 3,390.80 ICS exposure 0.17 0.05, 0.29 29.16 −4.81, 63.12 0.17 0.05, 0.29 28.78 −5.14, 62.51 Mother’s and baby’s characteristics  Maternal age, years   <18 0 Referent 0 Referent 0 Referent 0 Referent   18–34 0.05 −0.36, 0.47 157.72 38.89, 276.55 0.01 −0.39, 0.42 140.05 26.61, 252.76   >34 −0.19 −0.62, 0.24 113.81 −10.33, 237.94 −0.23 −0.65, 0.19 95.80 −22.97, 213.66  Baby’s sex (female vs. male) 0.01 −0.09, 0.11 −142.11 −170.32, −113.91 0.01 −0.09, 0.11 −142.37 −170.51, −114.27  Receipt of social assistance −0.30 −0.40, −0.20 −118.13 −146.87, −89.39 −0.30 −0.40, −0.20 −118.02 −146.69, −89.44  Rural residency −0.21 −0.33, −0.08 −50.49 −86.44, −14.54 −0.21 −0.33, −0.08 −50.49 −86.41, −14.73 Maternal chronic conditions  Antiphospholipid syndrome −0.92 −1.55, −0.28 −341.40 −524.57, −158.22 −0.86 −1.48, −0.23 −313.12 −488.25, −138.52  Chronic hypertension −0.23 −0.52, 0.06 −32.30 −116.42, 51.82 −0.23 −0.52, 0.06 −32.02 −114.92, 50.86  Diabetes mellitus −0.38 −0.65, −0.12 105.14 28.58, 181.70 −0.39 −0.65, −0.12 102.76 26.36, 178.47  Uterine defects −0.47 −0.61, −0.32 −28.85 −70.22, 12.53 −0.47 −0.61, −0.32 −28.71 −69.94, 12.25 Pregnancy-related variables  Gestational diabetes −0.20 −0.36, −0.03 81.78 34.46, 129.11 −0.20 −0.36, −0.03 81.00 33.65, 128.08  Eclampsia/preeclampsia −1.22 −1.51, −0.92 −270.91 −356.08, −185.74 −1.21 −1.50, −0.91 −266.56 −350.73, −182.23  Anemia −0.04 −0.18, 0.10 74.53 34.18, 114.89 −0.04 −0.18, 0.10 73.70 33.66, 113.68  Placental conditions −0.33 −0.59, −0.07 −144.19 −218.69, −69.69 −0.33 −0.58, −0.07 −143.47 −216.75, −70.32  Placental abruption −1.34 −1.64, −1.03 −356.21 −443.47, −268.96 −1.32 −1.62, −1.02 −350.66 −436.76, −264.08  Vaginal bleeding −0.42 −0.59, −0.25 −92.93 −141.19, −44.68 −0.42 −0.59, −0.26 −94.39 −142.35, −46.30  Maternal infections −0.17 −0.31, −0.03 −9.77 −49.99, 30.45 −0.17 −0.31, −0.03 −9.91 −50.15, 30.26  Fetal-maternal hemorrhage −2.40 −3.40, −1.40 −335.49 −623.96, −47.02 −2.28 −3.23, −1.31 −277.66 −538.28, −16.82  Pregnancy-induced hypertension −0.01 −0.22, 0.20 17.90 −42.12, 77.92 −0.01 −0.22, 0.19 16.95 −42.28, 76.64  Use of β-blockers −0.12 −0.66, 0.42 −64.52 −219.89, 90.84 −0.11 −0.64, 0.42 −61.00 −210.51, 88.67 Asthma-related variables  Leukotriene-receptor antagonists 0.03 −0.38, 0.43 −30.25 −146.78, 86.27 0.03 −0.37, 0.43 −29.39 −144.01, 85.32  No. of SABA doses per week   0 0 Referent 0 Referent 0 Referent 0 Referent   0.01–3 0.06 −0.08, 0.19 −1.93 −39.56, 35.70 0.05 −0.08, 0.19 −2.15 −39.68, 35.51   >3 −0.18 −0.33, −0.04 −44.09 −86.45, −1.72 −0.18 −0.33, −0.04 −44.34 −86.88, −1.57  Oral corticosteroids −0.08 −0.26, 0.10 −29.51 −80.98, 21.96 −0.08 −0.26, 0.10 −28.86 −80.08, 22.19  Intranasal corticosteroids −0.01 −0.15, 0.14 23.75 −19.35, 66.85 −0.01 −0.15, 0.14 23.54 −19.33, 66.36  ≥1 ED visit for asthma 0.13 −0.03, 0.30 30.54 −16.32, 77.40 0.13 −0.03, 0.29 30.19 −16.68, 76.65  LABA −0.11 −0.30, 0.08 1.39 −53.35, 56.13 −0.11 −0.30, 0.08 1.60 −52.88, 56.02  ≥1 hospitalization for asthma −0.08 −0.56, 0.41 −81.84 −220.82, 57.14 −0.07 −0.54, 0.41 −78.29 −214.34, 57.85  Severity of asthma prior to pregnancy   Mild 0 Referent 0 Referent 0 Referent 0 Referent   Moderate 0.17 0.02, 0.33 11.06 −33.85, 55.97 0.17 0.02, 0.33 10.92 −34.01, 55.62   Severe 0.06 −0.17, 0.29 −71.50 −138.28, −4.72 0.06 −0.17, 0.29 −71.36 −137.81, −4.57 GA residual 184.83 179.31, 190.37 Characteristica Univariate-Adjusted Model 1-Component Adjusted Mixture Model GA, weeks BW, g GA, weeks BW, g RC 95% CI RC 95% CI RC 95% CrI RC 95% CrI Intercept 38.98 38.55, 39.40 3,256.19 3,133.98, 3,378.40 39.02 38.60, 39.44 3,274.01 3,157.17, 3,390.80 ICS exposure 0.17 0.05, 0.29 29.16 −4.81, 63.12 0.17 0.05, 0.29 28.78 −5.14, 62.51 Mother’s and baby’s characteristics  Maternal age, years   <18 0 Referent 0 Referent 0 Referent 0 Referent   18–34 0.05 −0.36, 0.47 157.72 38.89, 276.55 0.01 −0.39, 0.42 140.05 26.61, 252.76   >34 −0.19 −0.62, 0.24 113.81 −10.33, 237.94 −0.23 −0.65, 0.19 95.80 −22.97, 213.66  Baby’s sex (female vs. male) 0.01 −0.09, 0.11 −142.11 −170.32, −113.91 0.01 −0.09, 0.11 −142.37 −170.51, −114.27  Receipt of social assistance −0.30 −0.40, −0.20 −118.13 −146.87, −89.39 −0.30 −0.40, −0.20 −118.02 −146.69, −89.44  Rural residency −0.21 −0.33, −0.08 −50.49 −86.44, −14.54 −0.21 −0.33, −0.08 −50.49 −86.41, −14.73 Maternal chronic conditions  Antiphospholipid syndrome −0.92 −1.55, −0.28 −341.40 −524.57, −158.22 −0.86 −1.48, −0.23 −313.12 −488.25, −138.52  Chronic hypertension −0.23 −0.52, 0.06 −32.30 −116.42, 51.82 −0.23 −0.52, 0.06 −32.02 −114.92, 50.86  Diabetes mellitus −0.38 −0.65, −0.12 105.14 28.58, 181.70 −0.39 −0.65, −0.12 102.76 26.36, 178.47  Uterine defects −0.47 −0.61, −0.32 −28.85 −70.22, 12.53 −0.47 −0.61, −0.32 −28.71 −69.94, 12.25 Pregnancy-related variables  Gestational diabetes −0.20 −0.36, −0.03 81.78 34.46, 129.11 −0.20 −0.36, −0.03 81.00 33.65, 128.08  Eclampsia/preeclampsia −1.22 −1.51, −0.92 −270.91 −356.08, −185.74 −1.21 −1.50, −0.91 −266.56 −350.73, −182.23  Anemia −0.04 −0.18, 0.10 74.53 34.18, 114.89 −0.04 −0.18, 0.10 73.70 33.66, 113.68  Placental conditions −0.33 −0.59, −0.07 −144.19 −218.69, −69.69 −0.33 −0.58, −0.07 −143.47 −216.75, −70.32  Placental abruption −1.34 −1.64, −1.03 −356.21 −443.47, −268.96 −1.32 −1.62, −1.02 −350.66 −436.76, −264.08  Vaginal bleeding −0.42 −0.59, −0.25 −92.93 −141.19, −44.68 −0.42 −0.59, −0.26 −94.39 −142.35, −46.30  Maternal infections −0.17 −0.31, −0.03 −9.77 −49.99, 30.45 −0.17 −0.31, −0.03 −9.91 −50.15, 30.26  Fetal-maternal hemorrhage −2.40 −3.40, −1.40 −335.49 −623.96, −47.02 −2.28 −3.23, −1.31 −277.66 −538.28, −16.82  Pregnancy-induced hypertension −0.01 −0.22, 0.20 17.90 −42.12, 77.92 −0.01 −0.22, 0.19 16.95 −42.28, 76.64  Use of β-blockers −0.12 −0.66, 0.42 −64.52 −219.89, 90.84 −0.11 −0.64, 0.42 −61.00 −210.51, 88.67 Asthma-related variables  Leukotriene-receptor antagonists 0.03 −0.38, 0.43 −30.25 −146.78, 86.27 0.03 −0.37, 0.43 −29.39 −144.01, 85.32  No. of SABA doses per week   0 0 Referent 0 Referent 0 Referent 0 Referent   0.01–3 0.06 −0.08, 0.19 −1.93 −39.56, 35.70 0.05 −0.08, 0.19 −2.15 −39.68, 35.51   >3 −0.18 −0.33, −0.04 −44.09 −86.45, −1.72 −0.18 −0.33, −0.04 −44.34 −86.88, −1.57  Oral corticosteroids −0.08 −0.26, 0.10 −29.51 −80.98, 21.96 −0.08 −0.26, 0.10 −28.86 −80.08, 22.19  Intranasal corticosteroids −0.01 −0.15, 0.14 23.75 −19.35, 66.85 −0.01 −0.15, 0.14 23.54 −19.33, 66.36  ≥1 ED visit for asthma 0.13 −0.03, 0.30 30.54 −16.32, 77.40 0.13 −0.03, 0.29 30.19 −16.68, 76.65  LABA −0.11 −0.30, 0.08 1.39 −53.35, 56.13 −0.11 −0.30, 0.08 1.60 −52.88, 56.02  ≥1 hospitalization for asthma −0.08 −0.56, 0.41 −81.84 −220.82, 57.14 −0.07 −0.54, 0.41 −78.29 −214.34, 57.85  Severity of asthma prior to pregnancy   Mild 0 Referent 0 Referent 0 Referent 0 Referent   Moderate 0.17 0.02, 0.33 11.06 −33.85, 55.97 0.17 0.02, 0.33 10.92 −34.01, 55.62   Severe 0.06 −0.17, 0.29 −71.50 −138.28, −4.72 0.06 −0.17, 0.29 −71.36 −137.81, −4.57 GA residual 184.83 179.31, 190.37 Abbreviations: BW, birth weight; CI, confidence interval; CrI, credible interval; ED, emergency department; GA, gestational age; ICS, inhaled corticosteroids; LABA, long-acting β2-agonists; RC, regression coefficient; SABA, short-acting β2-agonists. a All adjustment variables are binary (yes = 1/no = 0) unless otherwise indicated. The results for the mixture model with K = 2 are presented in Table 5. The mixing proportions π1 and π2, which represent the relative size of each component (subpopulation) in the population, were estimated to be 0.92 (95% credible interval (CrI): 0.90, 0.93) and 0.08 (95% CrI: 0.07, 0.10), respectively. We observed that the second component was associated with pregnancies with smaller GA and BW. The first component was centered at 39.51 weeks (95% CrI: 39.20, 39.82) and 3,358 g (95% CrI: 3,252, 3,464). The second component was centered at 33.24 weeks (95% CrI: 30.82, 35.44) and 2,182 g (95% CrI: 1,848, 2,511). In the first component, the mean difference associated with ICS for GA was −0.01 week (95% CrI: −0.10, 0.08), while the mean difference for BW was 1.27 g (95% CrI: −29.98, 32.15). Corresponding figures in the second component were 1.55 weeks (95% CrI: 0.67, 2.42) for GA and 200.3 g (95% CrI: 34.10, 366.0) for BW. In general, we observed large estimation uncertainty for many regression coefficients in the second component as a result of the relatively small representation of this component in our cohort of size 6,197 (equivalent of n times the estimated π2 value, i.e., 496). As reported in Web Appendix 2, our additional analyses showed robustness of the results to the choice of the prior distribution. Table 5. Associations of Inhaled Corticosteroid Exposure With Gestational Age and Birth Weight Among Asthmatic Pregnant Women in Each Component of an Adjusted 2-Component Adjusted Mixture-of-Bivariate-Regressions Model, Quebec, Canada, 1998–2008 Characteristica Component 1b Component 2c GA, weeks BW, g GA, weeks BW, g RC 95% CrI RC 95% CrI RC 95% CrI RC 95% CrI Intercept 39.51 39.20, 39.82 3,357.68 3,251.77, 3,463.80 33.24 30.82, 35.44 2,182.29 1,848.04, 2,511.27 ICS exposure −0.01 −0.10, 0.08 1.27 −29.98, 32.15 1.55 0.67, 2.42 200.33 34.10, 365.97 Mother’s and baby’s characteristics  Maternal age, years   <18 0 Referent 0 Referent 0 Referent 0 Referent   18–34 −0.27 −0.57, 0.02 110.16 7.03, 212.89 2.52 0.42, 4.85 396.48 84.92, 712.79   >34 −0.34 −0.65, −0.03 108.30 0.08, 215.71 1.02 −1.22, 3.43 33.24 −303.02, 369.88  Baby’s sex (female vs. male) 0.04 −0.03, 0.11 −143.34 −169.07,−117.62 −0.32 −1.08, 0.43 −142.10 −286.54, 1.54  Receipt of social assistance −0.20 −0.28, −0.13 −98.18 −124.44, −71.92 −0.70 −1.46, 0.07 −169.27 −314.51, −22.65  Rural residency −0.15 −0.24, −0.06 −38.53 −71.32, −5.54 −0.50 −1.45, 0.42 −90.59 −269.02, 81.75 Maternal chronic conditions  Antiphospholipid syndrome −0.53 −1.00, −0.07 −244.61 −410.75, −77.66 −0.54 −3.85, 2.51 −213.44 −694.67, 256.87  Chronic hypertension −0.33 −0.54, −0.13 −56.37 −132.13, 19.98 0.14 −1.94, 2.10 −8.04 −385.33, 353.93  Diabetes mellitus −0.51 −0.71, −0.32 63.96 −5.02, 133.07 −0.12 −2.24, 1.89 257.60 −89.35, 591.89  Uterine defects −0.52 −0.62, −0.42 −31.51 −69.19, 6.01 −0.73 −1.94, 0.44 −164.26 −389.89, 54.18 Pregnancy-related variables  Gestational diabetes −0.38 −0.50, −0.26 54.80 11.82, 98.01 1.22 −0.05, 2.44 207.57 −39.53, 443.16  Eclampsia/preeclampsia −0.52 −0.75, −0.28 −79.58 −162.51, 4.65 −1.87 −3.28, −0.47 −539.05 −804.27, −278.65  Anemia 0.10 0.00, 0.20 107.97 71.27, 144.83 −0.50 −1.53, 0.51 −80.34 −275.09, 109.39  Placental conditions −0.30 −0.50, −0.10 −151.93 −220.79, −82.81 −0.16 −1.77, 1.39 50.95 −240.65, 334.24  Placental abruption −0.62 −0.86, −0.38 −228.92 −311.94, −145.29 −2.10 −3.67, −0.51 −394.53 −675.78, −110.45  Vaginal bleeding −0.18 −0.30, −0.05 −46.73 −92.54, −0.82 −1.59 −2.70, −0.52 −261.21 −463.43, −62.13  Maternal infections −0.11 −0.21, 0.00 4.89 −32.26, 41.89 −0.49 −1.52, 0.52 −59.10 −252.09, 128.67  Fetal-maternal hemorrhage −0.24 −1.13, 0.71 14.26 −250.99, 285.98 −3.18 −6.34, −0.10 −275.13 −770.42, 207.84  Pregnancy-induced hypertension −0.16 −0.31, −0.01 2.59 −52.34, 57.77 0.22 −1.30, 1.67 −145.82 −435.69, 133.06  Use of β-blockers −0.09 −0.47, 0.29 −39.42 −180.71, 101.34 0.43 −2.52, 3.31 −210.51 −692.39, 281.47 Asthma-related variables  Leukotriene-receptor antagonists 0.03 −0.26, 0.32 −37.40 −141.50, 66.76 −0.73 −3.51, 1.88 −35.11 −490.04, 402.69  No. of SABA doses per week   0 0 Referent 0 Referent 0 Referent 0 Referent   0.01–3 0.04 −0.06, 0.13 −2.30 −36.50, 32.14 0.39 −0.57, 1.37 30.58 −152.98, 215.72   >3 −0.03 −0.14, 0.08 −21.05 −59.77, 17.69 −0.69 −1.74, 0.36 −40.12 −235.79, 154.51  Oral corticosteroids 0.00 −0.13, 0.13 −28.34 −75.45, 18.65 −0.11 −1.36, 1.12 149.64 −80.78, 375.93  Intranasal corticosteroids 0.00 −0.10, 0.11 25.19 −14.07, 64.54 −0.43 −1.59, 0.67 −61.50 −277.99, 147.23  ≥1 ED visit for asthma 0.06 −0.06, 0.18 25.28 −17.13, 68.02 0.24 −1.03, 1.50 −77.18 −312.31, 153.70  LABA −0.07 −0.21, 0.06 11.48 −38.86, 61.59 −0.15 −1.55, 1.22 −52.13 −322.39, 213.72  ≥1 hospitalization for asthma 0.07 −0.29, 0.43 −65.35 −190.57, 61.27 −0.35 −3.11, 2.18 3.94 −438.72, 426.41  Severity of asthma prior to pregnancy   Mild 0 Referent 0 Referent 0 Referent 0 Referent   Moderate 0.09 −0.02, 0.20 −6.30 −47.13, 34.96 0.68 −0.46, 1.81 133.34 −82.30, 348.92   Severe 0.04 −0.13, 0.21 −84.24 −144.89, −23.41 −1.18 −3.01, 0.58 −205.41 −523.13, 104.07 GA residual 156.44 146.77, 166.14 160.71 148.69, 172.67 Characteristica Component 1b Component 2c GA, weeks BW, g GA, weeks BW, g RC 95% CrI RC 95% CrI RC 95% CrI RC 95% CrI Intercept 39.51 39.20, 39.82 3,357.68 3,251.77, 3,463.80 33.24 30.82, 35.44 2,182.29 1,848.04, 2,511.27 ICS exposure −0.01 −0.10, 0.08 1.27 −29.98, 32.15 1.55 0.67, 2.42 200.33 34.10, 365.97 Mother’s and baby’s characteristics  Maternal age, years   <18 0 Referent 0 Referent 0 Referent 0 Referent   18–34 −0.27 −0.57, 0.02 110.16 7.03, 212.89 2.52 0.42, 4.85 396.48 84.92, 712.79   >34 −0.34 −0.65, −0.03 108.30 0.08, 215.71 1.02 −1.22, 3.43 33.24 −303.02, 369.88  Baby’s sex (female vs. male) 0.04 −0.03, 0.11 −143.34 −169.07,−117.62 −0.32 −1.08, 0.43 −142.10 −286.54, 1.54  Receipt of social assistance −0.20 −0.28, −0.13 −98.18 −124.44, −71.92 −0.70 −1.46, 0.07 −169.27 −314.51, −22.65  Rural residency −0.15 −0.24, −0.06 −38.53 −71.32, −5.54 −0.50 −1.45, 0.42 −90.59 −269.02, 81.75 Maternal chronic conditions  Antiphospholipid syndrome −0.53 −1.00, −0.07 −244.61 −410.75, −77.66 −0.54 −3.85, 2.51 −213.44 −694.67, 256.87  Chronic hypertension −0.33 −0.54, −0.13 −56.37 −132.13, 19.98 0.14 −1.94, 2.10 −8.04 −385.33, 353.93  Diabetes mellitus −0.51 −0.71, −0.32 63.96 −5.02, 133.07 −0.12 −2.24, 1.89 257.60 −89.35, 591.89  Uterine defects −0.52 −0.62, −0.42 −31.51 −69.19, 6.01 −0.73 −1.94, 0.44 −164.26 −389.89, 54.18 Pregnancy-related variables  Gestational diabetes −0.38 −0.50, −0.26 54.80 11.82, 98.01 1.22 −0.05, 2.44 207.57 −39.53, 443.16  Eclampsia/preeclampsia −0.52 −0.75, −0.28 −79.58 −162.51, 4.65 −1.87 −3.28, −0.47 −539.05 −804.27, −278.65  Anemia 0.10 0.00, 0.20 107.97 71.27, 144.83 −0.50 −1.53, 0.51 −80.34 −275.09, 109.39  Placental conditions −0.30 −0.50, −0.10 −151.93 −220.79, −82.81 −0.16 −1.77, 1.39 50.95 −240.65, 334.24  Placental abruption −0.62 −0.86, −0.38 −228.92 −311.94, −145.29 −2.10 −3.67, −0.51 −394.53 −675.78, −110.45  Vaginal bleeding −0.18 −0.30, −0.05 −46.73 −92.54, −0.82 −1.59 −2.70, −0.52 −261.21 −463.43, −62.13  Maternal infections −0.11 −0.21, 0.00 4.89 −32.26, 41.89 −0.49 −1.52, 0.52 −59.10 −252.09, 128.67  Fetal-maternal hemorrhage −0.24 −1.13, 0.71 14.26 −250.99, 285.98 −3.18 −6.34, −0.10 −275.13 −770.42, 207.84  Pregnancy-induced hypertension −0.16 −0.31, −0.01 2.59 −52.34, 57.77 0.22 −1.30, 1.67 −145.82 −435.69, 133.06  Use of β-blockers −0.09 −0.47, 0.29 −39.42 −180.71, 101.34 0.43 −2.52, 3.31 −210.51 −692.39, 281.47 Asthma-related variables  Leukotriene-receptor antagonists 0.03 −0.26, 0.32 −37.40 −141.50, 66.76 −0.73 −3.51, 1.88 −35.11 −490.04, 402.69  No. of SABA doses per week   0 0 Referent 0 Referent 0 Referent 0 Referent   0.01–3 0.04 −0.06, 0.13 −2.30 −36.50, 32.14 0.39 −0.57, 1.37 30.58 −152.98, 215.72   >3 −0.03 −0.14, 0.08 −21.05 −59.77, 17.69 −0.69 −1.74, 0.36 −40.12 −235.79, 154.51  Oral corticosteroids 0.00 −0.13, 0.13 −28.34 −75.45, 18.65 −0.11 −1.36, 1.12 149.64 −80.78, 375.93  Intranasal corticosteroids 0.00 −0.10, 0.11 25.19 −14.07, 64.54 −0.43 −1.59, 0.67 −61.50 −277.99, 147.23  ≥1 ED visit for asthma 0.06 −0.06, 0.18 25.28 −17.13, 68.02 0.24 −1.03, 1.50 −77.18 −312.31, 153.70  LABA −0.07 −0.21, 0.06 11.48 −38.86, 61.59 −0.15 −1.55, 1.22 −52.13 −322.39, 213.72  ≥1 hospitalization for asthma 0.07 −0.29, 0.43 −65.35 −190.57, 61.27 −0.35 −3.11, 2.18 3.94 −438.72, 426.41  Severity of asthma prior to pregnancy   Mild 0 Referent 0 Referent 0 Referent 0 Referent   Moderate 0.09 −0.02, 0.20 −6.30 −47.13, 34.96 0.68 −0.46, 1.81 133.34 −82.30, 348.92   Severe 0.04 −0.13, 0.21 −84.24 −144.89, −23.41 −1.18 −3.01, 0.58 −205.41 −523.13, 104.07 GA residual 156.44 146.77, 166.14 160.71 148.69, 172.67 Abbreviations: BW, birth weight; CrI, credible interval; ED, emergency department; GA, gestational age; ICS, inhaled corticosteroids; LABA, long-acting β2-agonists; RC, regression coefficient; SABA, short-acting β2-agonists. a All adjustment variables are binary (yes = 1/no = 0) unless otherwise indicated. b Mixing proportion π1: 0.92, 95% CrI: 0.90, 0.93. c Mixing proportion π2: 0.08, 95% CrI: 0.07, 0.10. Table 5. Associations of Inhaled Corticosteroid Exposure With Gestational Age and Birth Weight Among Asthmatic Pregnant Women in Each Component of an Adjusted 2-Component Adjusted Mixture-of-Bivariate-Regressions Model, Quebec, Canada, 1998–2008 Characteristica Component 1b Component 2c GA, weeks BW, g GA, weeks BW, g RC 95% CrI RC 95% CrI RC 95% CrI RC 95% CrI Intercept 39.51 39.20, 39.82 3,357.68 3,251.77, 3,463.80 33.24 30.82, 35.44 2,182.29 1,848.04, 2,511.27 ICS exposure −0.01 −0.10, 0.08 1.27 −29.98, 32.15 1.55 0.67, 2.42 200.33 34.10, 365.97 Mother’s and baby’s characteristics  Maternal age, years   <18 0 Referent 0 Referent 0 Referent 0 Referent   18–34 −0.27 −0.57, 0.02 110.16 7.03, 212.89 2.52 0.42, 4.85 396.48 84.92, 712.79   >34 −0.34 −0.65, −0.03 108.30 0.08, 215.71 1.02 −1.22, 3.43 33.24 −303.02, 369.88  Baby’s sex (female vs. male) 0.04 −0.03, 0.11 −143.34 −169.07,−117.62 −0.32 −1.08, 0.43 −142.10 −286.54, 1.54  Receipt of social assistance −0.20 −0.28, −0.13 −98.18 −124.44, −71.92 −0.70 −1.46, 0.07 −169.27 −314.51, −22.65  Rural residency −0.15 −0.24, −0.06 −38.53 −71.32, −5.54 −0.50 −1.45, 0.42 −90.59 −269.02, 81.75 Maternal chronic conditions  Antiphospholipid syndrome −0.53 −1.00, −0.07 −244.61 −410.75, −77.66 −0.54 −3.85, 2.51 −213.44 −694.67, 256.87  Chronic hypertension −0.33 −0.54, −0.13 −56.37 −132.13, 19.98 0.14 −1.94, 2.10 −8.04 −385.33, 353.93  Diabetes mellitus −0.51 −0.71, −0.32 63.96 −5.02, 133.07 −0.12 −2.24, 1.89 257.60 −89.35, 591.89  Uterine defects −0.52 −0.62, −0.42 −31.51 −69.19, 6.01 −0.73 −1.94, 0.44 −164.26 −389.89, 54.18 Pregnancy-related variables  Gestational diabetes −0.38 −0.50, −0.26 54.80 11.82, 98.01 1.22 −0.05, 2.44 207.57 −39.53, 443.16  Eclampsia/preeclampsia −0.52 −0.75, −0.28 −79.58 −162.51, 4.65 −1.87 −3.28, −0.47 −539.05 −804.27, −278.65  Anemia 0.10 0.00, 0.20 107.97 71.27, 144.83 −0.50 −1.53, 0.51 −80.34 −275.09, 109.39  Placental conditions −0.30 −0.50, −0.10 −151.93 −220.79, −82.81 −0.16 −1.77, 1.39 50.95 −240.65, 334.24  Placental abruption −0.62 −0.86, −0.38 −228.92 −311.94, −145.29 −2.10 −3.67, −0.51 −394.53 −675.78, −110.45  Vaginal bleeding −0.18 −0.30, −0.05 −46.73 −92.54, −0.82 −1.59 −2.70, −0.52 −261.21 −463.43, −62.13  Maternal infections −0.11 −0.21, 0.00 4.89 −32.26, 41.89 −0.49 −1.52, 0.52 −59.10 −252.09, 128.67  Fetal-maternal hemorrhage −0.24 −1.13, 0.71 14.26 −250.99, 285.98 −3.18 −6.34, −0.10 −275.13 −770.42, 207.84  Pregnancy-induced hypertension −0.16 −0.31, −0.01 2.59 −52.34, 57.77 0.22 −1.30, 1.67 −145.82 −435.69, 133.06  Use of β-blockers −0.09 −0.47, 0.29 −39.42 −180.71, 101.34 0.43 −2.52, 3.31 −210.51 −692.39, 281.47 Asthma-related variables  Leukotriene-receptor antagonists 0.03 −0.26, 0.32 −37.40 −141.50, 66.76 −0.73 −3.51, 1.88 −35.11 −490.04, 402.69  No. of SABA doses per week   0 0 Referent 0 Referent 0 Referent 0 Referent   0.01–3 0.04 −0.06, 0.13 −2.30 −36.50, 32.14 0.39 −0.57, 1.37 30.58 −152.98, 215.72   >3 −0.03 −0.14, 0.08 −21.05 −59.77, 17.69 −0.69 −1.74, 0.36 −40.12 −235.79, 154.51  Oral corticosteroids 0.00 −0.13, 0.13 −28.34 −75.45, 18.65 −0.11 −1.36, 1.12 149.64 −80.78, 375.93  Intranasal corticosteroids 0.00 −0.10, 0.11 25.19 −14.07, 64.54 −0.43 −1.59, 0.67 −61.50 −277.99, 147.23  ≥1 ED visit for asthma 0.06 −0.06, 0.18 25.28 −17.13, 68.02 0.24 −1.03, 1.50 −77.18 −312.31, 153.70  LABA −0.07 −0.21, 0.06 11.48 −38.86, 61.59 −0.15 −1.55, 1.22 −52.13 −322.39, 213.72  ≥1 hospitalization for asthma 0.07 −0.29, 0.43 −65.35 −190.57, 61.27 −0.35 −3.11, 2.18 3.94 −438.72, 426.41  Severity of asthma prior to pregnancy   Mild 0 Referent 0 Referent 0 Referent 0 Referent   Moderate 0.09 −0.02, 0.20 −6.30 −47.13, 34.96 0.68 −0.46, 1.81 133.34 −82.30, 348.92   Severe 0.04 −0.13, 0.21 −84.24 −144.89, −23.41 −1.18 −3.01, 0.58 −205.41 −523.13, 104.07 GA residual 156.44 146.77, 166.14 160.71 148.69, 172.67 Characteristica Component 1b Component 2c GA, weeks BW, g GA, weeks BW, g RC 95% CrI RC 95% CrI RC 95% CrI RC 95% CrI Intercept 39.51 39.20, 39.82 3,357.68 3,251.77, 3,463.80 33.24 30.82, 35.44 2,182.29 1,848.04, 2,511.27 ICS exposure −0.01 −0.10, 0.08 1.27 −29.98, 32.15 1.55 0.67, 2.42 200.33 34.10, 365.97 Mother’s and baby’s characteristics  Maternal age, years   <18 0 Referent 0 Referent 0 Referent 0 Referent   18–34 −0.27 −0.57, 0.02 110.16 7.03, 212.89 2.52 0.42, 4.85 396.48 84.92, 712.79   >34 −0.34 −0.65, −0.03 108.30 0.08, 215.71 1.02 −1.22, 3.43 33.24 −303.02, 369.88  Baby’s sex (female vs. male) 0.04 −0.03, 0.11 −143.34 −169.07,−117.62 −0.32 −1.08, 0.43 −142.10 −286.54, 1.54  Receipt of social assistance −0.20 −0.28, −0.13 −98.18 −124.44, −71.92 −0.70 −1.46, 0.07 −169.27 −314.51, −22.65  Rural residency −0.15 −0.24, −0.06 −38.53 −71.32, −5.54 −0.50 −1.45, 0.42 −90.59 −269.02, 81.75 Maternal chronic conditions  Antiphospholipid syndrome −0.53 −1.00, −0.07 −244.61 −410.75, −77.66 −0.54 −3.85, 2.51 −213.44 −694.67, 256.87  Chronic hypertension −0.33 −0.54, −0.13 −56.37 −132.13, 19.98 0.14 −1.94, 2.10 −8.04 −385.33, 353.93  Diabetes mellitus −0.51 −0.71, −0.32 63.96 −5.02, 133.07 −0.12 −2.24, 1.89 257.60 −89.35, 591.89  Uterine defects −0.52 −0.62, −0.42 −31.51 −69.19, 6.01 −0.73 −1.94, 0.44 −164.26 −389.89, 54.18 Pregnancy-related variables  Gestational diabetes −0.38 −0.50, −0.26 54.80 11.82, 98.01 1.22 −0.05, 2.44 207.57 −39.53, 443.16  Eclampsia/preeclampsia −0.52 −0.75, −0.28 −79.58 −162.51, 4.65 −1.87 −3.28, −0.47 −539.05 −804.27, −278.65  Anemia 0.10 0.00, 0.20 107.97 71.27, 144.83 −0.50 −1.53, 0.51 −80.34 −275.09, 109.39  Placental conditions −0.30 −0.50, −0.10 −151.93 −220.79, −82.81 −0.16 −1.77, 1.39 50.95 −240.65, 334.24  Placental abruption −0.62 −0.86, −0.38 −228.92 −311.94, −145.29 −2.10 −3.67, −0.51 −394.53 −675.78, −110.45  Vaginal bleeding −0.18 −0.30, −0.05 −46.73 −92.54, −0.82 −1.59 −2.70, −0.52 −261.21 −463.43, −62.13  Maternal infections −0.11 −0.21, 0.00 4.89 −32.26, 41.89 −0.49 −1.52, 0.52 −59.10 −252.09, 128.67  Fetal-maternal hemorrhage −0.24 −1.13, 0.71 14.26 −250.99, 285.98 −3.18 −6.34, −0.10 −275.13 −770.42, 207.84  Pregnancy-induced hypertension −0.16 −0.31, −0.01 2.59 −52.34, 57.77 0.22 −1.30, 1.67 −145.82 −435.69, 133.06  Use of β-blockers −0.09 −0.47, 0.29 −39.42 −180.71, 101.34 0.43 −2.52, 3.31 −210.51 −692.39, 281.47 Asthma-related variables  Leukotriene-receptor antagonists 0.03 −0.26, 0.32 −37.40 −141.50, 66.76 −0.73 −3.51, 1.88 −35.11 −490.04, 402.69  No. of SABA doses per week   0 0 Referent 0 Referent 0 Referent 0 Referent   0.01–3 0.04 −0.06, 0.13 −2.30 −36.50, 32.14 0.39 −0.57, 1.37 30.58 −152.98, 215.72   >3 −0.03 −0.14, 0.08 −21.05 −59.77, 17.69 −0.69 −1.74, 0.36 −40.12 −235.79, 154.51  Oral corticosteroids 0.00 −0.13, 0.13 −28.34 −75.45, 18.65 −0.11 −1.36, 1.12 149.64 −80.78, 375.93  Intranasal corticosteroids 0.00 −0.10, 0.11 25.19 −14.07, 64.54 −0.43 −1.59, 0.67 −61.50 −277.99, 147.23  ≥1 ED visit for asthma 0.06 −0.06, 0.18 25.28 −17.13, 68.02 0.24 −1.03, 1.50 −77.18 −312.31, 153.70  LABA −0.07 −0.21, 0.06 11.48 −38.86, 61.59 −0.15 −1.55, 1.22 −52.13 −322.39, 213.72  ≥1 hospitalization for asthma 0.07 −0.29, 0.43 −65.35 −190.57, 61.27 −0.35 −3.11, 2.18 3.94 −438.72, 426.41  Severity of asthma prior to pregnancy   Mild 0 Referent 0 Referent 0 Referent 0 Referent   Moderate 0.09 −0.02, 0.20 −6.30 −47.13, 34.96 0.68 −0.46, 1.81 133.34 −82.30, 348.92   Severe 0.04 −0.13, 0.21 −84.24 −144.89, −23.41 −1.18 −3.01, 0.58 −205.41 −523.13, 104.07 GA residual 156.44 146.77, 166.14 160.71 148.69, 172.67 Abbreviations: BW, birth weight; CrI, credible interval; ED, emergency department; GA, gestational age; ICS, inhaled corticosteroids; LABA, long-acting β2-agonists; RC, regression coefficient; SABA, short-acting β2-agonists. a All adjustment variables are binary (yes = 1/no = 0) unless otherwise indicated. b Mixing proportion π1: 0.92, 95% CrI: 0.90, 0.93. c Mixing proportion π2: 0.08, 95% CrI: 0.07, 0.10. Figure 1 presents the clustering of the 6,197 pregnancies on the GA-BW plane. All but 1 of the pregnancies allocated to the second component were preterm, while the vast majority (82%) corresponded to LBW infants. In parallel, most PTB pregnancies allocated to the first component did not correspond to LBW infants. At the 5% level, a high probability of membership in the second component was found to be negatively associated with exposure to ICS and gestational diabetes and positively associated with receipt of social assistance, fetal-maternal hemorrhage, placental abruption, vaginal bleeding, anemia, eclampsia/preeclampsia, and use of more than 3 doses of short-acting β2-agonists per week (see Table 6). Figure 1. View largeDownload slide Allocation of pregnancies in the first (gray dots) and second (black dots) components of a 2-component mixture-of-bivariate-regressions model of inhaled corticosteroid exposure among asthmatic pregnant women, Quebec, Canada, 1998–2008. Figure 1. View largeDownload slide Allocation of pregnancies in the first (gray dots) and second (black dots) components of a 2-component mixture-of-bivariate-regressions model of inhaled corticosteroid exposure among asthmatic pregnant women, Quebec, Canada, 1998–2008. Table 6. Associations Between Selected Characteristics of Asthmatic Pregnant Women and Membership in the Second Component of a Mixture-of-Bivariate-Regressions Model of Inhaled Corticosteroid Exposure (Multiple Logistic Regression Analysis), Quebec, Canada, 1998–2008 Characteristica OR 95% CI P Value ICS exposure 0.62 0.47, 0.82 <0.001 Mother’s and baby’s characteristics  Maternal age, years   <18 1.00 Referent   18–34 0.49 0.23, 1.04 0.06   >34 0.75 0.34, 1.66 0.48  Baby’s sex (female vs. male) 0.94 0.75, 1.18 0.59  Receipt of social assistance 1.43 1.13, 1.82 <0.01  Rural residency 1.18 0.89, 1.58 0.25 Maternal chronic conditions  Antiphospholipid syndrome 1.57 0.44, 5.57 0.49  Chronic hypertension 0.75 0.36, 1.55 0.44  Diabetes mellitus 0.69 0.36, 1.31 0.25  Uterine defects 0.81 0.57, 1.14 0.23 Pregnancy-related variables  Gestational diabetes 0.60 0.38, 0.94 0.02  Eclampsia/preeclampsia 3.70 2.29, 5.98 <0.0001  Anemia 1.35 1.01, 1.82 0.04  Placental conditions 1.24 0.77, 2.01 0.37  Placental abruption 2.03 1.27, 3.26 <0.01  Vaginal bleeding 1.91 1.38, 2.60 <0.001  Maternal infections 1.16 0.86, 1.57 0.34  Fetal-maternal hemorrhage 5.18 1.52, 17.6 <0.01  Pregnancy-induced hypertension 0.74 0.45, 1.23 0.24  Use of β-blockers 1.60 0.54, 4.71 0.40 Asthma-related variables  Leukotriene-receptor antagonists 0.82 0.31, 2.14 0.68  No. of SABA doses per week   0 1.00 Referent   0.01–3 1.06 0.78, 1.45 0.71   >3 1.56 1.11, 2.18 <0.01  Oral corticosteroids 1.35 0.91, 2.00 0.14  Intranasal corticosteroids 1.00 0.71, 1.42 0.99  ≥1 ED visit for asthma 0.77 0.52, 1.15 0.20  LABA 1.27 0.83, 1.96 0.28  ≥1 hospitalization for asthma 1.49 0.60, 3.72 0.39  Severity of asthma prior to pregnancy   Mild 1.00 Referent   Moderate 0.95 0.67, 1.37 0.80   Severe 0.83 0.49, 1.41 0.50 Characteristica OR 95% CI P Value ICS exposure 0.62 0.47, 0.82 <0.001 Mother’s and baby’s characteristics  Maternal age, years   <18 1.00 Referent   18–34 0.49 0.23, 1.04 0.06   >34 0.75 0.34, 1.66 0.48  Baby’s sex (female vs. male) 0.94 0.75, 1.18 0.59  Receipt of social assistance 1.43 1.13, 1.82 <0.01  Rural residency 1.18 0.89, 1.58 0.25 Maternal chronic conditions  Antiphospholipid syndrome 1.57 0.44, 5.57 0.49  Chronic hypertension 0.75 0.36, 1.55 0.44  Diabetes mellitus 0.69 0.36, 1.31 0.25  Uterine defects 0.81 0.57, 1.14 0.23 Pregnancy-related variables  Gestational diabetes 0.60 0.38, 0.94 0.02  Eclampsia/preeclampsia 3.70 2.29, 5.98 <0.0001  Anemia 1.35 1.01, 1.82 0.04  Placental conditions 1.24 0.77, 2.01 0.37  Placental abruption 2.03 1.27, 3.26 <0.01  Vaginal bleeding 1.91 1.38, 2.60 <0.001  Maternal infections 1.16 0.86, 1.57 0.34  Fetal-maternal hemorrhage 5.18 1.52, 17.6 <0.01  Pregnancy-induced hypertension 0.74 0.45, 1.23 0.24  Use of β-blockers 1.60 0.54, 4.71 0.40 Asthma-related variables  Leukotriene-receptor antagonists 0.82 0.31, 2.14 0.68  No. of SABA doses per week   0 1.00 Referent   0.01–3 1.06 0.78, 1.45 0.71   >3 1.56 1.11, 2.18 <0.01  Oral corticosteroids 1.35 0.91, 2.00 0.14  Intranasal corticosteroids 1.00 0.71, 1.42 0.99  ≥1 ED visit for asthma 0.77 0.52, 1.15 0.20  LABA 1.27 0.83, 1.96 0.28  ≥1 hospitalization for asthma 1.49 0.60, 3.72 0.39  Severity of asthma prior to pregnancy   Mild 1.00 Referent   Moderate 0.95 0.67, 1.37 0.80   Severe 0.83 0.49, 1.41 0.50 Abbreviations: CI, confidence interval; ED, emergency department; ICS, inhaled corticosteroids; LABA, long-acting β2-agonists; OR, odds ratio; SABA, short-acting β2-agonists. a All variables are binary (yes = 1/no = 0) unless otherwise indicated. View Large Table 6. Associations Between Selected Characteristics of Asthmatic Pregnant Women and Membership in the Second Component of a Mixture-of-Bivariate-Regressions Model of Inhaled Corticosteroid Exposure (Multiple Logistic Regression Analysis), Quebec, Canada, 1998–2008 Characteristica OR 95% CI P Value ICS exposure 0.62 0.47, 0.82 <0.001 Mother’s and baby’s characteristics  Maternal age, years   <18 1.00 Referent   18–34 0.49 0.23, 1.04 0.06   >34 0.75 0.34, 1.66 0.48  Baby’s sex (female vs. male) 0.94 0.75, 1.18 0.59  Receipt of social assistance 1.43 1.13, 1.82 <0.01  Rural residency 1.18 0.89, 1.58 0.25 Maternal chronic conditions  Antiphospholipid syndrome 1.57 0.44, 5.57 0.49  Chronic hypertension 0.75 0.36, 1.55 0.44  Diabetes mellitus 0.69 0.36, 1.31 0.25  Uterine defects 0.81 0.57, 1.14 0.23 Pregnancy-related variables  Gestational diabetes 0.60 0.38, 0.94 0.02  Eclampsia/preeclampsia 3.70 2.29, 5.98 <0.0001  Anemia 1.35 1.01, 1.82 0.04  Placental conditions 1.24 0.77, 2.01 0.37  Placental abruption 2.03 1.27, 3.26 <0.01  Vaginal bleeding 1.91 1.38, 2.60 <0.001  Maternal infections 1.16 0.86, 1.57 0.34  Fetal-maternal hemorrhage 5.18 1.52, 17.6 <0.01  Pregnancy-induced hypertension 0.74 0.45, 1.23 0.24  Use of β-blockers 1.60 0.54, 4.71 0.40 Asthma-related variables  Leukotriene-receptor antagonists 0.82 0.31, 2.14 0.68  No. of SABA doses per week   0 1.00 Referent   0.01–3 1.06 0.78, 1.45 0.71   >3 1.56 1.11, 2.18 <0.01  Oral corticosteroids 1.35 0.91, 2.00 0.14  Intranasal corticosteroids 1.00 0.71, 1.42 0.99  ≥1 ED visit for asthma 0.77 0.52, 1.15 0.20  LABA 1.27 0.83, 1.96 0.28  ≥1 hospitalization for asthma 1.49 0.60, 3.72 0.39  Severity of asthma prior to pregnancy   Mild 1.00 Referent   Moderate 0.95 0.67, 1.37 0.80   Severe 0.83 0.49, 1.41 0.50 Characteristica OR 95% CI P Value ICS exposure 0.62 0.47, 0.82 <0.001 Mother’s and baby’s characteristics  Maternal age, years   <18 1.00 Referent   18–34 0.49 0.23, 1.04 0.06   >34 0.75 0.34, 1.66 0.48  Baby’s sex (female vs. male) 0.94 0.75, 1.18 0.59  Receipt of social assistance 1.43 1.13, 1.82 <0.01  Rural residency 1.18 0.89, 1.58 0.25 Maternal chronic conditions  Antiphospholipid syndrome 1.57 0.44, 5.57 0.49  Chronic hypertension 0.75 0.36, 1.55 0.44  Diabetes mellitus 0.69 0.36, 1.31 0.25  Uterine defects 0.81 0.57, 1.14 0.23 Pregnancy-related variables  Gestational diabetes 0.60 0.38, 0.94 0.02  Eclampsia/preeclampsia 3.70 2.29, 5.98 <0.0001  Anemia 1.35 1.01, 1.82 0.04  Placental conditions 1.24 0.77, 2.01 0.37  Placental abruption 2.03 1.27, 3.26 <0.01  Vaginal bleeding 1.91 1.38, 2.60 <0.001  Maternal infections 1.16 0.86, 1.57 0.34  Fetal-maternal hemorrhage 5.18 1.52, 17.6 <0.01  Pregnancy-induced hypertension 0.74 0.45, 1.23 0.24  Use of β-blockers 1.60 0.54, 4.71 0.40 Asthma-related variables  Leukotriene-receptor antagonists 0.82 0.31, 2.14 0.68  No. of SABA doses per week   0 1.00 Referent   0.01–3 1.06 0.78, 1.45 0.71   >3 1.56 1.11, 2.18 <0.01  Oral corticosteroids 1.35 0.91, 2.00 0.14  Intranasal corticosteroids 1.00 0.71, 1.42 0.99  ≥1 ED visit for asthma 0.77 0.52, 1.15 0.20  LABA 1.27 0.83, 1.96 0.28  ≥1 hospitalization for asthma 1.49 0.60, 3.72 0.39  Severity of asthma prior to pregnancy   Mild 1.00 Referent   Moderate 0.95 0.67, 1.37 0.80   Severe 0.83 0.49, 1.41 0.50 Abbreviations: CI, confidence interval; ED, emergency department; ICS, inhaled corticosteroids; LABA, long-acting β2-agonists; OR, odds ratio; SABA, short-acting β2-agonists. a All variables are binary (yes = 1/no = 0) unless otherwise indicated. View Large Lastly, when reanalyzing our cohort using LBW as opposed to BW in a standard adjusted logistic regression model, we obtained an odds ratio associated with ICS of 0.79 (95% CI: 0.63, 1.00). For the outcomes PTB and very PTB, the adjusted ICS odds ratios were 0.84 (95% CI: 0.68, 1.04) and 0.28 (95% CI: 0.15, 0.51), respectively. The results for the 2-component mixture secondary analysis using GA and BW z scores are presented in Web Appendix 3 (Web Table 1). The relative sizes of the 2 subpopulations were similar to those found in the corresponding analysis using GA and BW. The association between ICS and GA also remained similar in both components; however, ICS was not found to be associated with BW z score in either component. In the smallest component, only diabetes mellitus and eclampsia/preeclampsia were found to be positively and negatively associated with BW z score, respectively. Unlike in the main analyses, the within-component residual association between GA and BW z score (β∗k) did not significantly differ from zero. DISCUSSION Adjusting for a large set of covariates, including many variables directly related to asthma severity and control, the credible intervals for exposure to ICS associated with GA and BW in the first, major component of our K = 2 mixture model were found to be centered approximately at zero and relatively narrow. These results, which concern a large proportion of pregnancies, are consistent with those of many other studies that have failed to associate ICS exposure during pregnancy with adverse perinatal outcomes, including GA- and BW-related variables (35). Interestingly, our 2-component mixture model revealed that there is a small group of asthmatic women, representing about 8% of the population, in whom ICS is positively associated with GA and BW, with a magnitude which could be qualified as clinically significant. In this subpopulation, pregnancies exposed to ICS were, on average, approximately 11 days longer than those that were not (95% CrI: 4.70, 16.0). This increase in GA reverberated on BW, as was seen from the estimated average increase of about 200 g in BW when pregnancies were exposed to ICS (95% CrI: 34.10, 366.0). However, our bivariate model makes it possible to note that this mean difference of about 200 g is somewhat smaller than what would be expected from an increase of 1.5 weeks in GA, since, in this component, the residual mean change for the effect of 1 additional week in GA on BW was estimated to be 160.7 g (95% CrI: 148.7, 172.7). This observation is consistent with the BW z score analysis, where the point estimate for the association between ICS and BW z score in the second component was found to be small and negative (estimate = −0.07, 95% CrI: −0.35, 0.21). Differences in conclusions between the standard linear regressions (or K = 1 mixture model) for GA and BW and the standard logistic regressions which instead parameterized PTB and LBW may suggest that the effect of exposure to ICS takes place at low quantiles of the joint GA/BW distribution. This is more strongly supported by the results from the main 2-component mixture analysis, for which a positive association between ICS and GA/BW was seen in the most problematic asthmatic pregnancies. Indeed, nearly all hard-clustered pregnancies in the smaller component were preterm, which would explain why inconclusive results were obtained in our standard adjusted logistic regression analysis using PTB as the outcome (odds ratio = 0.84, 95% CI: 0.68, 1.04). The significant results obtained for very PTB (odds ratio = 0.28, 95% CI: 0.15, 0.51) could also be anticipated from the mixture model’s results, since 32 weeks is a central point for GA in the second component (recall Figure 1). To our knowledge, only 1 previous study has examined the association between maternal exposure to ICS and very PTB using ICS-unexposed asthmatic women as the reference group: Using very PTB as the outcome variable, Schatz et al. (65) found a nonsignificant unadjusted odds ratio for ICS of 0.94 (P > 0.05) based on a sample of 2,123 asthmatic women. Our study was a reanalysis of Cossette et al.’s cohort of pregnancies among women with asthma (36); as such, the same limitations regarding the data pertain to the present work. Generalization of the results to a general obstetrical population is questionable, since the Régie de l’assurance maladie du Québec prescription insurance plan is a compulsory insurance plan for Quebec residents who are not eligible for a private one. Overall, people enrolled in this plan have a less favorable socioeconomic status, which may influence ICS adherence. Indeed, recall that exposure to ICS was based on dispensed prescriptions, which might not reflect actual use. Moreover, our list of measured covariates omitted important potential confounders or risk factors, such as maternal cigarette smoking during pregnancy, prepregnancy body mass index, ethnicity, and parity. We also lacked genetic information, which could be related to global responsiveness to ICS and the occurrence of adverse perinatal outcomes such as PTB and LBW. However, additional analyses which used the number of pregnancies in the follow-up period to define a proxy for parity yielded essentially the same results (not reported). Moreover, Cossette et al. found, in sensitivity analyses, that their estimates of the effect of ICS on the perinatal outcomes LBW, PTB, and small size for GA were relatively robust to unmeasured smoking status (36). A similar conclusion was reached for maternal obesity status (see Web Appendix 4, including Web Table 2, for results and discussion). Because one can define the mixture-of-regression model to represent homogenous subpopulations indexed by common values for unmeasured binary or categorical covariates, it is, however, possible that these missing covariates were partially accounted for by our model. Finally, a limitation specific to this study is the possibility that the ever/never exposure classification may have attenuated true risks associated with the effect of ICS use on GA/BW, or even created a protective effect. This bias would arise from increased opportunity for exposure among women remaining pregnant longer and has been shown to explain an apparent protective effect of maternal influenza immunization in pregnancy on risk of PTB (66). Many of our study’s strengths directly follow from those of Cossette et al. (36)—notably the large number of pregnancies, medication-exposure assessment from data prospectively collected independently of the outcomes and recorded in pharmacy records, the inclusion of multiple potential confounders, and the use of previously validated algorithms to measure ICS and short-acting β2-agonist exposure and the control and severity of asthma. While the size of our cohort was deemed sufficiently large to apply a mixture-of-regressions model, a larger number of pregnancies would have allowed a more thorough assessment of the heterogeneity of the associations between exposure to ICS and GA and BW. Indeed, because the regression coefficients’ CrIs were generally wide for the smallest subpopulation, we chose not to study the effect of ICS on GA and BW as a function of dosage. Similarly, we did not consider adding a third mixture component, which, for a larger cohort, may have revealed more specific insights on the role of ICS in the most problematic asthmatic pregnancies. Indeed, because it is well-known that prematurity is a multifactorial syndrome (67, 68), it is likely that pregnancies in the smallest component still exhibit significant individual heterogeneity with regard to the ICS effect. As it stands, we nonetheless believe that our mixture-of-regressions model enabled us to carry out a unique and highly valuable exploration of our cohort data, which will likely serve to refine or generate new hypotheses as to how treatment of asthma with ICS during pregnancy affects fetal growth. In conclusion, we have introduced a finite mixture-of-bivariate-regressions model to explore the heterogeneity of ICS exposure effects on BW (or BW z score) and GA in a large cohort of asthmatic pregnant women, but this model can certainly be used for risk assessment for other exposures. When using this model with BW z score instead of BW, it is expected that the coefficients β∗k(k=1,2,…,K) will be zero if the GA-dependent normalization scale for BW is appropriate. The proposed mixture is thus overly flexible for exploring effect heterogeneity with a BW z score outcome, since it models the theoretical null dependence between GA and BW z score. While this could be viewed as undesirable, one may find it useful to be able to check for practical deviation of β∗k from zero, indicating a residual correlation between GA and BW z score and thus suggesting an external reference population inadequate for the study population. With a BW z score outcome, a straightforward modification of the model to strictly enforce the β∗k coefficient values to zero could alternatively be done, thereby reducing by K the number of parameters to estimate. ACKNOWLEDGMENTS Author affiliations: Département de mathématiques, Faculté des sciences, Université du Québec à Montréal, Montréal, Québec, Canada (Mariia Samoilenko, Geneviève Lefebvre); Faculté de pharmacie, Université de Montréal, Montréal, Québec, Canada (Lucie Blais, Geneviève Lefebvre); Centre de recherche, Hôpital du Sacré-Coeur de Montréal, Montréal, Québec, Canada (Lucie Blais); Centre de recherche Clinique Étienne-Le Bel, Centre hospitalier universitaire de Sherbrooke, Sherbrooke, Québec, Canada (Lucie Blais); and Département d’obstétrique-gynécologie, Centre hospitalier universitaire de Sainte-Justine, Université de Montréal, Montréal, Québec, Canada (Isabelle Boucoiran). This work was supported by grants from the Fonds de recherche Québec–Santé (FRQ-S) and the Natural Sciences and Engineering Research Council of Canada. G.L. is an FRQ-S Research Scholar. G.L. thanks Calcul Québec (Montréal, Québec, Canada) and Dr. Daniel Stubbs (Calcul Québec) for programming support. Conflict of interest: none declared. Abbreviations BW birth weight CI confidence interval CrI credible interval GA gestational age ICS inhaled corticosteroids LABA long-acting β2-agonists LBW low birth weight PTB preterm birth REFERENCES 1 Ober C , Yao TC . The genetics of asthma and allergic disease: a 21st century perspective . Immunol Rev . 2011 ; 242 ( 1 ): 10 – 30 . Google Scholar Crossref Search ADS PubMed 2 Noval Rivas M , Crother TR , Arditi M . The microbiome in asthma . Curr Opin Pediatr . 2016 ; 28 ( 6 ): 764 – 771 . Google Scholar Crossref Search ADS PubMed 3 Borish L , Culp JA . Asthma: a syndrome composed of heterogeneous diseases . Ann Allergy Asthma Immunol . 2008 ; 101 ( 1 ): 1 – 8 ; quiz 8–11, 50. Google Scholar Crossref Search ADS PubMed 4 Lougheed MD , Lemiere C , Ducharme FM , et al. . Canadian Thoracic Society 2012 guideline update: diagnosis and management of asthma in preschoolers, children and adults . Can Respir J . 2012 ; 19 ( 2 ): 127 – 164 . Google Scholar Crossref Search ADS PubMed 5 Global Initiative for Asthma . Global Strategy for Asthma Management and Prevention. Updated 2016. Bethesda, MD : National Heart, Lung, and Blood Institute ; 2016 . https://ginasthma.org/wp-content/uploads/2016/04/wms-GINA-2016-main-report-final.pdf. Accessed September 7, 2016. 6 National Asthma Education and Prevention Program . Expert Panel Report 3 (EPR-3): guidelines for the diagnosis and management of asthma—summary report 2007 . J Allergy Clin Immunol . 2007 ; 120 ( 5 suppl ): S94 – S138 . PubMed 7 Healthcare Improvement Scotland; Scottish Intercollegiate Guidelines Network (SIGN); British Thoracic Society . British Guideline on the Management of Asthma: a National Clinical Guideline. (SIGN 153). Edinburgh, United Kingdom : Healthcare Improvement Scotland ; 2016 . https://www.brit-thoracic.org.uk/document-library/clinical-information/asthma/btssign-asthma-guideline-2016/. Accessed October 11, 2016. 8 Lima JJ . Treatment heterogeneity in asthma: genetics of response to leukotriene modifiers . Mol Diagn Ther . 2007 ; 11 ( 2 ): 97 – 104 . Google Scholar Crossref Search ADS PubMed 9 Dahlin A , Tantisira KG . Integrative systems biology approaches in asthma pharmacogenomics . Pharmacogenomics . 2012 ; 13 ( 12 ): 1387 – 1404 . Google Scholar Crossref Search ADS PubMed 10 Farzan N , Vijverberg SJ , Arets HG , et al. . Pharmacogenomics of inhaled corticosteroids and leukotriene modifiers: a systematic review . Clin Exp Allergy . 2017 ; 47 ( 2 ): 271 – 293 . Google Scholar Crossref Search ADS PubMed 11 Simpson KR , Creehan PA . AWHONN’s Perinatal Nursing . 4th ed . Philadelphia, PA : Wolters Kluwer/Lippincott Williams & Wilkins Health ; 2013 . 12 Barnes PJ . Glucocorticosteroids: current and future directions . Br J Pharmacol . 2011 ; 163 ( 1 ): 29 – 43 . Google Scholar Crossref Search ADS PubMed 13 Barnes PJ . Corticosteroid resistance in patients with asthma and chronic obstructive pulmonary disease . J Allergy Clin Immunol . 2013 ; 131 ( 3 ): 636 – 645 . Google Scholar Crossref Search ADS PubMed 14 Matsumura Y . Heterogeneity of glucocorticoid resistance in patients with bronchial asthma . Int J Biomed Sci . 2010 ; 6 ( 3 ): 158 – 166 . Google Scholar PubMed 15 Lazarus SC , Chinchilli VM , Rollings NJ , et al. . Smoking affects response to inhaled corticosteroids or leukotriene receptor antagonists in asthma . Am J Respir Crit Care Med . 2007 ; 175 ( 8 ): 783 – 790 . Google Scholar Crossref Search ADS PubMed 16 Bourdin A , Halimi L , Vachier I , et al. . Adherence in severe asthma . Clin Exp Allergy . 2012 ; 42 ( 11 ): 1566 – 1574 . Google Scholar Crossref Search ADS PubMed 17 Silverman ES , Liggett SB , Gelfand EW , et al. . The pharmacogenetics of asthma: a candidate gene approach . Pharmacogenomics J . 2001 ; 1 : 27 – 37 . Google Scholar Crossref Search ADS PubMed 18 Davis JS , Weiss ST , Tantisira KG . Asthma pharmacogenomics: 2015 update . Curr Allergy Asthma Rep . 2015 ; 15 ( 7 ): 42 . Google Scholar Crossref Search ADS PubMed 19 Murphy VE . Managing asthma in pregnancy . Breathe (Sheff) . 2015 ; 11 ( 4 ): 258 – 267 . Google Scholar Crossref Search ADS PubMed 20 Katz O , Sheiner E . Asthma and pregnancy: a review of two decades . Expert Rev Respir Med . 2008 ; 2 ( 1 ): 97 – 107 . Google Scholar Crossref Search ADS PubMed 21 Murphy VE , Clifton VL , Gibson PG . Asthma exacerbations during pregnancy: incidence and association with adverse pregnancy outcomes . Thorax . 2006 ; 61 ( 2 ): 169 – 176 . Google Scholar Crossref Search ADS PubMed 22 Gregersen TL , Ulrik CS . Safety of bronchodilators and corticosteroids for asthma during pregnancy: what we know and what we need to do better . J Asthma Allergy . 2013 ; 6 : 117 – 125 . Google Scholar PubMed 23 Namazy JA , Murphy VE , Powell H , et al. . Effects of asthma severity, exacerbations and oral corticosteroids on perinatal outcomes . Eur Respir J . 2013 ; 41 ( 5 ): 1082 – 1090 . Google Scholar Crossref Search ADS PubMed 24 Grzeskowiak LE , Clifton VL . Asthma management during pregnancy: how long before we can all breathe a little easier? J Asthma . 2015 ; 52 ( 10 ): 1020 – 1022 . Google Scholar Crossref Search ADS PubMed 25 National Heart, Lung, and Blood Institute; National Asthma Education and Prevention Program Asthma and Pregnancy Working Group . NAEPP Expert Panel Report. Managing asthma during pregnancy: recommendations for pharmacologic treatment—2004 update . J Allergy Clin Immunol . 2005 ; 115 ( 1 ): 34 – 46 . Crossref Search ADS PubMed 26 Schatz M , Zeiger RS , Hoffman CP , et al. . Perinatal outcomes in the pregnancies of asthmatic women: a prospective controlled analysis . Am J Respir Crit Care Med . 1995 ; 151 ( 4 ): 1170 – 1174 . Google Scholar PubMed 27 Schatz M , Zeiger RS , Harden KM , et al. . The safety of inhaled beta-agonist bronchodilators during pregnancy . J Allergy Clin Immunol . 1988 ; 82 ( 4 ): 686 – 695 . Google Scholar Crossref Search ADS PubMed 28 Perlow JH , Montgomery D , Morgan MA , et al. . Severity of asthma and perinatal outcome . Am J Obstet Gynecol . 1992 ; 167 ( 4 ): 963 – 967 . Google Scholar Crossref Search ADS PubMed 29 Schatz M . The safety of asthma and allergy medications during pregnancy . Can J Allergy Clin Immunol . 1998 ; 3 ( 5 ): 242 – 254 . 30 Schatz M , Zeiger RS , Harden K , et al. . The safety of asthma and allergy medications during pregnancy . J Allergy Clin Immunol . 1997 ; 100 ( 3 ): 301 – 306 . Google Scholar Crossref Search ADS PubMed 31 Dombrowski MP , Schatz M , Wise R , et al. . Randomized trial of inhaled beclomethasone dipropionate versus theophylline for moderate asthma during pregnancy . Am J Obstet Gynecol . 2004 ; 190 ( 3 ): 737 – 744 . Google Scholar Crossref Search ADS PubMed 32 Murphy VE , Namazy JA , Powell H , et al. . A meta-analysis of adverse perinatal outcomes in women with asthma . BJOG . 2011 ; 118 ( 11 ): 1314 – 1323 . Google Scholar Crossref Search ADS PubMed 33 Rejnö G , Lundholm C , Gong T , et al. . Asthma during pregnancy in a population-based study—pregnancy complications and adverse perinatal outcomes . PLoS One . 2014 ; 9 ( 8 ): e104755 . Google Scholar Crossref Search ADS PubMed 34 Namazy JA , Chambers C , Schatz M . Safety of therapeutic options for treating asthma in pregnancy . Expert Opin Drug Saf . 2014 ; 13 ( 12 ): 1613 – 1621 . Google Scholar Crossref Search ADS PubMed 35 Namazy JA , Schatz M . Pharmacotherapy options to treat asthma during pregnancy . Expert Opin Pharmacother . 2015 ; 16 ( 12 ): 1783 – 1791 . Google Scholar Crossref Search ADS PubMed 36 Cossette B , Forget A , Beauchesne MF , et al. . Impact of maternal use of asthma-controller therapy on perinatal outcomes . Thorax . 2013 ; 68 ( 8 ): 724 – 730 . Google Scholar Crossref Search ADS PubMed 37 Samoilenko M , Blais L , Cossette B , et al. . Assessing the dose-response relationship between maternal use of inhaled corticosteroids therapy and birth weight: a generalized propensity score approach . Obs Stud . 2016 ; 2016 ( Article 8 ): 90 – 118 . 38 Murphy VE , Gibson PG , Giles WB , et al. . Maternal asthma is associated with reduced female fetal growth . Am J Respir Crit Care Med . 2003 ; 168 ( 11 ): 1317 – 1323 . Google Scholar Crossref Search ADS PubMed 39 Hodyl NA , Stark MJ , Osei-Kumah A , et al. . Fetal glucocorticoid-regulated pathways are not affected by inhaled corticosteroid use for asthma during pregnancy . Am J Respir Crit Care Med . 2011 ; 183 ( 6 ): 716 – 722 . Google Scholar Crossref Search ADS PubMed 40 Huffaker MF , Phipatanakul W . Pediatric asthma: guidelines-based care, omalizumab, and other potential biologic agents . Immunol Allergy Clin North Am . 2015 ; 35 ( 1 ): 129 – 144 . Google Scholar Crossref Search ADS PubMed 41 Langmack EL , Martin RJ . Heterogeneity of response to asthma controller therapy: clinical implications . Curr Opin Pulm Med . 2010 ; 16 ( 1 ): 13 – 18 . Google Scholar Crossref Search ADS PubMed 42 Drazen JM . A step toward personalized asthma treatment . N Engl J Med . 2011 ; 365 ( 13 ): 1245 – 1246 . Google Scholar Crossref Search ADS PubMed 43 Blake K , Raissy H . Pharmacogenomic testing in the asthma clinic: will inhaled corticosteroids lead the way? Pediatr Allergy Immunol Pulmonol . 2012 ; 25 ( 1 ): 44 – 47 . Google Scholar Crossref Search ADS 44 Tantisira KG , Lasky-Su J , Harada M , et al. . Genomewide association between GLCCI1 and response to glucocorticoid therapy in asthma . N Engl J Med . 2011 ; 365 ( 13 ): 1173 – 1183 . Google Scholar Crossref Search ADS PubMed 45 McLachlan GJ , Basford KE . Mixture Models: Inference and Applications to Clustering . New York, NY : Marcel Dekker, Inc. ; 1988 . 46 Melnykov V , Maitra R . Finite mixture models and model-based clustering . Stat Surv . 2010 ; 4 : 80 – 116 . Google Scholar Crossref Search ADS 47 Andrews JL , McNicholas PD . Model-based clustering, classification, and discriminant analysis via mixtures of multivariate t-distributions . Stat Comput . 2012 ; 22 ( 5 ): 1021 – 1029 . Google Scholar Crossref Search ADS 48 Frühwirth-Schnatter S . Finite Mixture and Markov Switching Models . New York, NY : Springer Science+Business Media LLC ; 2006 . 49 Charnigo R , Chesnut LW , Lobianco T , et al. . Thinking outside the curve, part I: modeling birthweight distribution . BMC Pregnancy Childbirth . 2010 ; 10 : 37 . Google Scholar Crossref Search ADS PubMed 50 Gage TB , Therriault G . Variability of birth-weight distributions by sex and ethnicity: analysis using mixture models . Hum Biol . 1998 ; 70 ( 3 ): 517 – 534 . Google Scholar PubMed 51 Deng W , Charnigo R , Dai H , et al. . Characterizing components in a mixture model for birthweight distribution . J Biom Biostat . 2011 ; 2 ( 4 ): 118 . Google Scholar Crossref Search ADS 52 Tentoni S , Astolfi P , De Pasquale A , et al. . Birthweight by gestational age in preterm babies according to a Gaussian mixture model . BJOG . 2004 ; 111 ( 1 ): 31 – 37 . Google Scholar Crossref Search ADS PubMed 53 Oja H , Koiranen M , Rantakallio P . Fitting mixture models to birth weight data: a case study . Biometrics . 1991 ; 47 ( 3 ): 883 – 897 . Google Scholar Crossref Search ADS PubMed 54 Platt RW , Abrahamowicz M , Kramer MS , et al. . Detecting and eliminating erroneous gestational ages: a normal mixture model . Stat Med . 2001 ; 20 ( 23 ): 3491 – 3503 . Google Scholar Crossref Search ADS PubMed 55 Whitmore GA , Su Y . Modeling low birth weights using threshold regression: results for US birth data . Lifetime Data Anal . 2007 ; 13 ( 2 ): 161 – 190 . Google Scholar Crossref Search ADS PubMed 56 Schwartz SL , Gelfand AE , Miranda ML . Joint Bayesian analysis of birthweight and censored gestational age using finite mixture models . Stat Med . 2010 ; 29 ( 16 ): 1710 – 1723 . Google Scholar Crossref Search ADS PubMed 57 York TP , Vargas-Irwin C , Anderson WH , et al. . Asthma pharmacogenetic study using finite mixture models to handle drug-response heterogeneity . Pharmacogenomics . 2009 ; 10 ( 5 ): 753 – 767 . Google Scholar Crossref Search ADS PubMed 58 Blais L , Beauchesne MF , Rey E , et al. . Use of inhaled corticosteroids during the first trimester of pregnancy and the risk of congenital malformations among women with asthma . Thorax . 2007 ; 62 ( 4 ): 320 – 328 . Google Scholar Crossref Search ADS PubMed 59 Martel MJ , Rey E , Beauchesne MF , et al. . Use of inhaled corticosteroids during pregnancy and risk of pregnancy induced hypertension: nested case-control study . BMJ . 2005 ; 330 ( 7485 ): 230 . Google Scholar Crossref Search ADS PubMed 60 Diebolt J , Robert CP . Estimation of finite mixture distributions through Bayesian sampling . J R Stat Soc Series B Stat Methodol . 1994 ; 56 ( 2 ): 363 – 375 . 61 Loosemore S , McGrath R , Oram A , et al. . The GNU C Reference Manual . Boston, MA : Free Software Foundation, Inc. ; 2001 . https://www.gnu.org/software/gnu-c-manual/). Accessed July 2, 2018. 62 Arnold CC , Kramer MS , Hobbs CA , et al. . Very low birth weight: a problematic cohort for epidemiologic studies of very small or immature neonates . Am J Epidemiol . 1991 ; 134 ( 6 ): 604 – 613 . Google Scholar Crossref Search ADS PubMed 63 Hutcheon JA , Platt RW . The missing data problem in birth weight percentiles and thresholds for “small-for-gestational-age.” Am J Epidemiol . 2008 ; 167 ( 7 ): 786 – 792 . Google Scholar Crossref Search ADS PubMed 64 Kramer MS , Platt RW , Wen SW , et al. . A new and improved population-based Canadian reference for birth weight for gestational age . Pediatrics . 2001 ; 108 ( 2 ): E35 . Google Scholar Crossref Search ADS PubMed 65 Schatz M , Dombrowski MP , Wise R , et al. . The relationship of asthma medication use to perinatal outcomes . J Allergy Clin Immunol . 2004 ; 113 ( 6 ): 1040 – 1045 . Google Scholar Crossref Search ADS PubMed 66 Vazquez-Benitez G , Kharbanda EO , Naleway AL , et al. . Risk of preterm or small-for-gestational-age birth after influenza vaccination during pregnancy: caveats when conducting retrospective observational studies . Am J Epidemiol . 2016 ; 184 ( 3 ): 176 – 186 . Google Scholar Crossref Search ADS PubMed 67 Watson LF , Rayner JA , Forster D . Identifying risk factors for very preterm birth: a reference for clinicians . Midwifery . 2013 ; 29 ( 5 ): 434 – 439 . Google Scholar Crossref Search ADS PubMed 68 Morgan TK . Role of the placenta in preterm birth: a review . Am J Perinatol . 2016 ; 33 ( 3 ): 258 – 266 . Google Scholar Crossref Search ADS PubMed © The Author(s) 2018. Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png American Journal of Epidemiology Oxford University Press

Using a Mixture-of-Bivariate-Regressions Model to Explore Heterogeneity of Effects of the Use of Inhaled Corticosteroids on Gestational Age and Birth Weight Among Pregnant Women With Asthma

Loading next page...
 
/lp/ou_press/using-a-mixture-of-bivariate-regression-model-to-explore-heterogeneity-Yp5OUEru3I
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.
ISSN
0002-9262
eISSN
1476-6256
D.O.I.
10.1093/aje/kwy105
Publisher site
See Article on Publisher Site

Abstract

Abstract Asthma is a heterogeneous disease, and responses to asthma medications vary noticeably among patients. A substantively oriented objective of this study was to explore the potentially heterogeneous effects of exposure to maternal inhaled corticosteroids (ICS) on gestational age (GA) at delivery and birth weight (BW) using a cohort of 6,197 pregnancies among women with asthma (Quebec, Canada, 1998–2008). A methodologically oriented objective was to comprehensively describe the application of a Bayesian 2-component mixture-of-bivariate-regressions model to address this issue and estimate the effects of ICS on GA and BW jointly. Based on the proposed model, no association between ICS and GA/BW was found for a large proportion of asthmatic pregnancies. However, a positive association between ICS exposure and GA/BW was revealed in a small subset of pregnancies comprising mainly preterm and low–birth-weight infants. A novel application of this model was also subsequently performed using BW z score instead of BW as the outcome variable. In conclusion, the studied mixture-of-bivariate-regressions model was useful for detecting heterogeneity in the effect of ICS on GA and BW in our population of women with asthma. These analyses pave the way for analogous uses of this model for general assessment of exposure effect heterogeneity for these perinatal outcomes. asthma, birth weight, gestational age, heterogeneity, inhaled corticosteroids, mixture of regressions Asthma is a widespread and complex disease whose etiology involves both genetic and environmental factors (1–3). Worldwide, inhaled corticosteroids (ICS) are the first-line controller therapy for asthma at all ages (4–7). Common add-on therapies to ICS include leukotriene receptor antagonists and long-acting β2-agonists (LABA) (4, 5). Despite well-accepted guidelines, it is recognized that responses to asthma medications vary considerably among patients (8–10). While adherence to treatment, therapy combinations, and smoking explain some heterogeneity in response to asthma medication (11–16), it has been suggested to have a strong genetic foundation (8, 17, 18). Asthma is one of the most frequent chronic medical conditions to be reported during pregnancy (19). The prevalence of asthma among pregnant women has been observed to increase in the past several decades (20, 21), with recent reported numbers ranging from 7.8% to 13% (22–24). During pregnancy, current guidelines recommend continuing pharmacological therapy, as uncontrolled asthma puts the pregnant woman and the fetus at risk of adverse outcomes (5, 7, 25). Indeed, data from well-regarded studies have shown that optimally treated asthmatic women have fewer adverse perinatal and maternal outcomes than those without therapy (26–28). Moreover, difficult-to-control asthma requiring systemic corticosteroids during pregnancy has been associated with increased risk of maternal preeclampsia, perinatal mortality, hyperbilirubinemia, preterm birth (PTB; defined as gestational age (GA) at delivery <37 weeks), and low birth weight (LBW; defined as birth weight (BW) <2,500 g) (23, 29–34). The use of ICS at low-to-moderate doses by asthmatic pregnant women is generally regarded as safe with respect to adverse perinatal outcomes, including GA- and BW-related variables (22, 35). However, doubts remain concerning the effect of higher doses of ICS on GA and BW (22, 36, 37). While the dose-response relationship between ICS and GA and BW is one important pharmacoepidemiologic research direction (22), exploration of heterogeneity in the effects of ICS on BW and GA has not received due attention, being circumscribed to analyses stratified on asthma severity (38) or fetal sex (38, 39). Because recent data suggest that ICS treatment may lead to greater benefit for the control of asthma in certain subpopulations (40–44), it is reasonable to conjecture that ICS could exhibit heterogeneous effects on GA and BW due to patients’ or pregnancies’ characteristics. Finite mixture models are frequently used for the purpose of clustering (45–47). From this point of view, one desires to find homogeneous subpopulations among a hypothesized heterogeneous population (48). Specifically, this approach assumes the existence of K latent (unknown) subpopulations and aims to recover them by fitting a K-component mixture distribution, where the component-specific distributions are often taken to be from the same parametric family (e.g., such as normal distributions). Finite mixtures have been regularly proposed for modeling BW distributions (49–54). For this outcome, it has been suggested that the mixture’s components are the result of different types of births: The main component represents the subpopulation of normal births, while the other, minor components (typically 1 or 2 additional components) represent subpopulations of abnormal births (49, 55, 56). Recently, finite mixture models have also been applied in asthma pharmacogenetics to handle individual heterogeneity in response to the drug montelukast, a leukotriene receptor antagonist (57). In this paper, we use a finite mixture-of-regressions model to explore possible heterogeneity in the effects of ICS on GA and BW in a population of asthmatic pregnant women from the province of Quebec, Canada. In a mixture-of-regressions model, regression coefficients are allowed to vary from component to component, thus allowing for a given covariate, say ICS use, to have different effects on the outcome across the K subpopulations underlying the study population. Herein, rather than specifying a mixture of regressions for GA and BW separately, we use the Bayesian bivariate mixture-of-regressions model proposed by Schwartz et al. (56), which accounts for the strong correlation between these 2 perinatal outcomes. From a methodological perspective, our analyses serve as a worked example for illustrating the usefulness of this model for general heterogeneity of effect assessment. METHODS Bayesian finite mixture-of-bivariate-regressions model Prior to presenting our mixture-of-bivariate-regressions model, we refer the reader to Web Appendix 1 (available at https://academic.oup.com/aje) for a review of the univariate mixture-of-regressions model. Therein, we illustrate a key feature of mixtures of regressions, namely, how they model outcome heterogeneity and heterogeneity of effect due to unmeasured categorical predictors. Schwartz et al. (56) introduced a Bayesian finite mixture of bivariate regressions to jointly model GA and BW for effect estimation. They used a mixture specified by bivariate normal distributions, where each component-specific bivariate normal distribution is decomposed as a marginal form times a conditional form. Let G and B be the joint response variables corresponding to GA and BW, respectively. Let A be the exposure of interest, maternal ICS use during pregnancy, and let X be a set of p adjustment variables. Let (gi,bi) be the response data for the ith pregnancy (i=1,…,n), along with ai and xi′=(x1i,…,xpi), the corresponding observed exposure and covariates. Following Schwartz et al. (56), we specify the joint density function for (Gi,Bi) as f(gi,bi|xi,θ)=∑k=1Kπkfk(gi,bi)=∑k=1KπkN(gi|μg,k+aiτg,k+xi′βg,k,σg,k2)×N(bi|μb,k+aiτb,k+xi′βb,k+(gi−(μg,k+aiτg,k+xi′βg,k))β∗k,σb|g,k2), (1) where θ represents the set of unknown parameters of the model. In equation 1, τg,k and τb,k are the kth-component regression coefficients for the exposure (ICS use) associated with GA and BW, respectively. Moreover, the vectors of regression coefficients βg,k and βb,k are the kth-component regression coefficients associated with covariates X. Interestingly, the component-specific bivariate decomposition allows for the interpretation of BW conditional on GA within each component of the mixture: In component k, the coefficient β∗k in the conditional model for B given G encodes the mean change in B for a 1-unit increase in G after the contributions of A and X to explain G have been removed. In other words, it represents the residual association between GA and BW in each subpopulation. The residual variance associated with B(G) within component k is given by σb|g,k2(σg,k2). As such, model 1 (equation 1) allows for a component-specific residual variance in the marginal model for G and in the conditional model for B given G, thus permitting some heteroscedasticity in these residuals overall. To the notable exception of K, which is considered fixed, all parameters in model 1 are unknown and need to be estimated. Assuming independent units, the joint density function of B=(B1,…,Bn) and G=(G1,…,Gn) given x=(x1,…,xn) is f(g,b|x,θ)=Πi=1nf(gi,bi|xi,θ). Inference is based on the posterior distribution of θ, which is proportional to the likelihood times the prior distribution of θ. As in Schwartz et al. (56), this prior distribution is the product of standard prior distributions for the model parameters: a Dirichlet distribution for the mixing proportions (πk), multivariate normal distributions for the regression coefficients, and gamma distributions for the inverse of the residual variances. Data Data on medication prescriptions filled in community pharmacies, outpatient medical visits, emergency-department visits, medical procedures, and hospitalizations were retrieved from 2 administrative databases in Quebec: the Régie de l’assurance maladie du Québec and the Maintenance et exploitation des données pour l’étude de la clientèle hospitalière databases. These data were first used by Cossette et al. (36), who investigated the dose-response relationship between maternal use of ICS and LABA during pregnancy and the adverse perinatal outcomes LBW, PTB, and small size for GA (defined as the <10% BW quantile for a given GA in a population, by sex). Study design Our cohort was formed by applying the same inclusion and exclusion criteria as those used by Cossette et al. (36) to define their cohort. The inclusion criteria were: 1) singleton delivery (live birth or stillbirth) between 1998 and 2008; 2) maternal age ≤45 years at the beginning of pregnancy; 3) ≥1 diagnosis of asthma and ≥1 prescription for an asthma medication filled in the year before or during pregnancy; and 4) drug insurance coverage by the Régie de l’assurance maladie du Québec plan for at least 1 year before and throughout pregnancy. Exclusion criteria were: use of theophylline, cromoglycate, nedocromil, ketotifen, and LABA without ICS during pregnancy. Because our mixture-of-regressions model was defined for a set of independent observations, we kept only the most recent pregnancy if a woman had several pregnancies during follow-up (1,177 pregnancies were excluded). Our choice to select the last recorded pregnancy in the follow-up period rather than the first was made on the grounds that recent pregnancies are more likely to reflect current medical practice regarding the management of asthma during pregnancy. Our cohort comprised a total of 6,197 pregnancies, among which 5,021 (81%) were such that they corresponded to a woman’s unique pregnancy between 1998 and 2008, while the other 1,176 pregnancies (19%) corresponded to the last pregnancy for a woman who had more than 1 pregnancy during follow-up. Outcomes, exposure, and covariates The outcomes of interest were GA at delivery, measured in completed weeks, and BW, measured in grams. Maternal ICS exposure during pregnancy was measured with an algorithm based on prescription renewals that was used in prior studies (58, 59). ICS exposure was classified in 2 categories: use (>0 μg/day) and no use (0 μg/day). Our list of adjustment variables contained 26 identified risk factors for LBW, PTB, or small size for GA (36). These variables, some of which may be potentially confounding, can be divided into the following 4 categories: 1) mother’s and baby’s characteristics: maternal age, baby’s sex, receipt of social assistance, and rural residency; 2) maternal chronic conditions in the year before or during pregnancy: antiphospholipid syndrome, chronic hypertension, diabetes mellitus, and uterine defects; 3) pregnancy-related variables: gestational diabetes, eclampsia/preeclampsia, anemia, placental conditions, placental abruption, vaginal bleeding, maternal infections, fetal-maternal hemorrhage, pregnancy-induced hypertension, and use of β-blockers; and 4) asthma-related variables during pregnancy: leukotriene-receptor antagonists, short-acting β2-agonists, oral corticosteroids, intranasal corticosteroids, ≥1 emergency department visit for asthma, LABA, ≥1 hospitalization for asthma, and severity of asthma (in the year before conception). Statistical analyses Descriptive statistics were used to determine the characteristics of the pregnancies as a function of ICS use. Simple linear models regressing each outcome (GA, BW) as a function of each variable (exposure, adjustment variables) were fitted on our cohort of pregnancies. The associations of ICS with GA and BW were then estimated separately for each outcome using a standard adjusted linear regression model. Because GA is a potential intermediate variable in the pathway between ICS and BW, the BW-adjusted regression model did not include GA. Finally, the mixture model (model 1) with K=1 and K=2 was used to explore heterogeneity of the effects of ICS on GA/BW. The prior distribution on the parameters of model 1 is presented in Table 1. Table 1. Prior and Hyperparameter Values for the Parameters of Mixture-of-Bivariate-Regressions Models of Inhaled Corticosteroid Exposure During Pregnancy Among Asthmatic Women, Quebec, Canada, 1998–2008a Hyperparameter Mixture Model 1-Component Model (K = 1) 2-Component Model (K = 2) k = 1 k = 2 αk 1 1 1 μg,k0 37 37 32 τg,k0 0 0 0 βg,k0 0→ 0→ 0→ Σg,k 100,000I b 100,000I 100,000I μb,k0 3,500 3,100 1,900 τb,k0 0 0 0 βb,k0 0→ 0→ 0→ β∗k0 0 0 0 Σb,k 100,000I 100,000I 100,000I hg,k=hb,k 1 1 1 rg,k=rb,k 1 1 1 Hyperparameter Mixture Model 1-Component Model (K = 1) 2-Component Model (K = 2) k = 1 k = 2 αk 1 1 1 μg,k0 37 37 32 τg,k0 0 0 0 βg,k0 0→ 0→ 0→ Σg,k 100,000I b 100,000I 100,000I μb,k0 3,500 3,100 1,900 τb,k0 0 0 0 βb,k0 0→ 0→ 0→ β∗k0 0 0 0 Σb,k 100,000I 100,000I 100,000I hg,k=hb,k 1 1 1 rg,k=rb,k 1 1 1 a A Dirichlet distribution is specified for the mixing proportions: (π1,…,πK)~Dirichlet(α1,…,αK). Multivariate normal (N) distributions are specified for the gestational age and birth weight regression coefficients: (μg,k,τg,k,βg,k)~N((μg,k0,τg,k0,βg,k0),Σg,k), (μb,k,τb,k,βb,k,β∗k)~N((μb,k0,τb,k0,βb,k0,β∗k0),Σb,k). Gamma distributions are specified for the gestational age and birth weight residual precisions: (σg,k2)−1~Gamma(hg,k,rg,k);(σb|g,k2)−1~Gamma(hb,k,rb,k). bI = identity matrix. Table 1. Prior and Hyperparameter Values for the Parameters of Mixture-of-Bivariate-Regressions Models of Inhaled Corticosteroid Exposure During Pregnancy Among Asthmatic Women, Quebec, Canada, 1998–2008a Hyperparameter Mixture Model 1-Component Model (K = 1) 2-Component Model (K = 2) k = 1 k = 2 αk 1 1 1 μg,k0 37 37 32 τg,k0 0 0 0 βg,k0 0→ 0→ 0→ Σg,k 100,000I b 100,000I 100,000I μb,k0 3,500 3,100 1,900 τb,k0 0 0 0 βb,k0 0→ 0→ 0→ β∗k0 0 0 0 Σb,k 100,000I 100,000I 100,000I hg,k=hb,k 1 1 1 rg,k=rb,k 1 1 1 Hyperparameter Mixture Model 1-Component Model (K = 1) 2-Component Model (K = 2) k = 1 k = 2 αk 1 1 1 μg,k0 37 37 32 τg,k0 0 0 0 βg,k0 0→ 0→ 0→ Σg,k 100,000I b 100,000I 100,000I μb,k0 3,500 3,100 1,900 τb,k0 0 0 0 βb,k0 0→ 0→ 0→ β∗k0 0 0 0 Σb,k 100,000I 100,000I 100,000I hg,k=hb,k 1 1 1 rg,k=rb,k 1 1 1 a A Dirichlet distribution is specified for the mixing proportions: (π1,…,πK)~Dirichlet(α1,…,αK). Multivariate normal (N) distributions are specified for the gestational age and birth weight regression coefficients: (μg,k,τg,k,βg,k)~N((μg,k0,τg,k0,βg,k0),Σg,k), (μb,k,τb,k,βb,k,β∗k)~N((μb,k0,τb,k0,βb,k0,β∗k0),Σb,k). Gamma distributions are specified for the gestational age and birth weight residual precisions: (σg,k2)−1~Gamma(hg,k,rg,k);(σb|g,k2)−1~Gamma(hb,k,rb,k). bI = identity matrix. We implemented our mixture-of-bivariate-regressions model (equation 1) using a Gibbs sampling algorithm for augmented mixture (60) programmed with the R (R Foundation for Statistical Computing, Vienna, Austria) and C (61) languages. For each of the 2 mixture models considered, draws from the posterior distribution of the model parameters were obtained using the Gibbs sampler. A total of 100,000 iterations (burn-in = 100,000 iterations) were retained for each model and used to calculate posterior means and standard deviations and 95% credible intervals. Component allocation was monitored for the mixture model with K = 2. More precisely, while component membership was unknown for each pregnancy in our cohort, the proportion of times that each pregnancy is attributed to a given component over the Gibbs sampling iterations is an estimation of the posterior probability that the pregnancy belongs to that component. Hard clustering was done by assigning a pregnancy to the component with the highest posterior probability. An indicator variable Z was created and set to Zi=1 whenever pregnancy i was allocated more than 50% of times to the second component ( Zi=0 otherwise) over the Gibbs sampling iterations. We then used standard multiple logistic regression with outcome Z to investigate whether any variable was associated with this indicator. Sensitivity analyses regarding the choice of priors and adjustment covariates were performed subsequently; results are presented in Web Appendix 2. We also conducted unplanned analyses to obtain further insights on the problem and results: We performed 3 standard adjusted logistic regression analyses using LBW, PTB, and very PTB (GA <32 weeks) as outcome variables. Finally, because many perinatal studies rely on a GA-normalized BW variable as the primary outcome instead of or in addition to BW (62, 63), a novel application of the mixture of regressions with K = 2 was done to model ICS exposure effect heterogeneity using GA and BW z scores as response variables (with the priors for the BW z score intercepts centered at 0). The BW z scores were computed using a Canadian chart described by Kramer et al. (64). Approval from the Commission d’accès à l’information du Québec was obtained prior to requesting and linking the information from the Maintenance et exploitation des données pour l’étude de la clientèle hospitalière and Régie de l’assurance maladie du Québec databases. These analyses were approved by the ethics committee of the Hôpital du Sacré-Coeur de Montréal. RESULTS In our cohort, 3,461 babies out of 6,197 (55.8%) were exposed to ICS during their mother’s pregnancy. Among pregnancies exposed to ICS, 86.3% were exposed to a dose of less than 250 μg/day (fluticasone propionate equivalent) on average. Descriptive statistics are presented in Table 2 according to binary exposure status. A larger proportion of women treated with ICS received social assistance. Overall, asthma-related variables featured the greatest imbalance between pregnancies exposed to ICS and those unexposed. Notably, pregnancies exposed to ICS were associated with more severe asthma, were associated with more visits to emergency departments because of asthma, and were more exposed to leukotriene-receptor antagonists, short-acting β2-agonists, oral corticosteroids, and LABA (by design, no pregnancies unexposed to ICS were exposed to LABA in our cohort). Table 2. Distribution of Potentially Confounding Covariates According to Inhaled Corticosteroid Exposure During Pregnancy Among Asthmatic Women, Quebec, Canada, 1998–2008 Characteristic ICS Exposure Status Not Exposed Exposed No. % No. % No. of pregnancies 2,736 44.15 3,461 55.85 Mother’s and baby’s characteristics  Maternal age, years   <18 39 1.43 50 1.44   18–34 2,349 85.86 2,877 83.13   >34 348 12.72 534 15.43  Baby’s sex (female vs. male) 1,362 49.78 1,668 48.19  Receipt of social assistance 1,370 50.07 2,015 58.22  Rural residency 517 18.90 685 19.79 Maternal chronic conditions  Antiphospholipid syndrome 16 0.58 21 0.61  Chronic hypertension 75 2.74 118 3.41  Diabetes mellitus 87 3.18 136 3.93  Uterine defects 363 13.27 495 14.30 Pregnancy-related variables  Gestational diabetes 259 9.47 376 10.86  Eclampsia/preeclampsia 78 2.85 116 3.35  Anemia 385 14.07 509 14.71  Placental conditions 107 3.91 137 3.96  Placental abruption 95 3.47 125 3.61  Vaginal bleeding 381 13.93 448 12.94  Maternal infections 418 15.28 490 14.16  Fetal-maternal hemorrhage 5 0.18 10 0.29  Pregnancy-induced hypertension 167 6.10 241 6.96  Use of β-blockers 23 0.84 31 0.90 Asthma-related variables  Leukotriene-receptor antagonists 9 0.33 90 2.60  No. of SABA doses per week   0 1,549 56.62 434 12.54   0.01–3 762 27.85 1,346 38.89   >3 425 15.53 1,681 48.57  Oral corticosteroids 104 3.80 589 17.02  Intranasal corticosteroids 196 7.16 596 17.22  ≥1 ED visit for asthma 170 6.21 661 19.10  LABA 0 0 553 15.98  ≥1 hospitalization for asthma 6 0.22 63 1.82  Severity of asthma prior to pregnancy   Mild 2,479 90.61 2,499 72.20   Moderate 226 8.26 623 18.00   Severe 31 1.13 339 9.79 Characteristic ICS Exposure Status Not Exposed Exposed No. % No. % No. of pregnancies 2,736 44.15 3,461 55.85 Mother’s and baby’s characteristics  Maternal age, years   <18 39 1.43 50 1.44   18–34 2,349 85.86 2,877 83.13   >34 348 12.72 534 15.43  Baby’s sex (female vs. male) 1,362 49.78 1,668 48.19  Receipt of social assistance 1,370 50.07 2,015 58.22  Rural residency 517 18.90 685 19.79 Maternal chronic conditions  Antiphospholipid syndrome 16 0.58 21 0.61  Chronic hypertension 75 2.74 118 3.41  Diabetes mellitus 87 3.18 136 3.93  Uterine defects 363 13.27 495 14.30 Pregnancy-related variables  Gestational diabetes 259 9.47 376 10.86  Eclampsia/preeclampsia 78 2.85 116 3.35  Anemia 385 14.07 509 14.71  Placental conditions 107 3.91 137 3.96  Placental abruption 95 3.47 125 3.61  Vaginal bleeding 381 13.93 448 12.94  Maternal infections 418 15.28 490 14.16  Fetal-maternal hemorrhage 5 0.18 10 0.29  Pregnancy-induced hypertension 167 6.10 241 6.96  Use of β-blockers 23 0.84 31 0.90 Asthma-related variables  Leukotriene-receptor antagonists 9 0.33 90 2.60  No. of SABA doses per week   0 1,549 56.62 434 12.54   0.01–3 762 27.85 1,346 38.89   >3 425 15.53 1,681 48.57  Oral corticosteroids 104 3.80 589 17.02  Intranasal corticosteroids 196 7.16 596 17.22  ≥1 ED visit for asthma 170 6.21 661 19.10  LABA 0 0 553 15.98  ≥1 hospitalization for asthma 6 0.22 63 1.82  Severity of asthma prior to pregnancy   Mild 2,479 90.61 2,499 72.20   Moderate 226 8.26 623 18.00   Severe 31 1.13 339 9.79 Abbreviations: ED, emergency department; ICS, inhaled corticosteroids; LABA, long-acting β2-agonists; SABA, short-acting β2-agonists. Table 2. Distribution of Potentially Confounding Covariates According to Inhaled Corticosteroid Exposure During Pregnancy Among Asthmatic Women, Quebec, Canada, 1998–2008 Characteristic ICS Exposure Status Not Exposed Exposed No. % No. % No. of pregnancies 2,736 44.15 3,461 55.85 Mother’s and baby’s characteristics  Maternal age, years   <18 39 1.43 50 1.44   18–34 2,349 85.86 2,877 83.13   >34 348 12.72 534 15.43  Baby’s sex (female vs. male) 1,362 49.78 1,668 48.19  Receipt of social assistance 1,370 50.07 2,015 58.22  Rural residency 517 18.90 685 19.79 Maternal chronic conditions  Antiphospholipid syndrome 16 0.58 21 0.61  Chronic hypertension 75 2.74 118 3.41  Diabetes mellitus 87 3.18 136 3.93  Uterine defects 363 13.27 495 14.30 Pregnancy-related variables  Gestational diabetes 259 9.47 376 10.86  Eclampsia/preeclampsia 78 2.85 116 3.35  Anemia 385 14.07 509 14.71  Placental conditions 107 3.91 137 3.96  Placental abruption 95 3.47 125 3.61  Vaginal bleeding 381 13.93 448 12.94  Maternal infections 418 15.28 490 14.16  Fetal-maternal hemorrhage 5 0.18 10 0.29  Pregnancy-induced hypertension 167 6.10 241 6.96  Use of β-blockers 23 0.84 31 0.90 Asthma-related variables  Leukotriene-receptor antagonists 9 0.33 90 2.60  No. of SABA doses per week   0 1,549 56.62 434 12.54   0.01–3 762 27.85 1,346 38.89   >3 425 15.53 1,681 48.57  Oral corticosteroids 104 3.80 589 17.02  Intranasal corticosteroids 196 7.16 596 17.22  ≥1 ED visit for asthma 170 6.21 661 19.10  LABA 0 0 553 15.98  ≥1 hospitalization for asthma 6 0.22 63 1.82  Severity of asthma prior to pregnancy   Mild 2,479 90.61 2,499 72.20   Moderate 226 8.26 623 18.00   Severe 31 1.13 339 9.79 Characteristic ICS Exposure Status Not Exposed Exposed No. % No. % No. of pregnancies 2,736 44.15 3,461 55.85 Mother’s and baby’s characteristics  Maternal age, years   <18 39 1.43 50 1.44   18–34 2,349 85.86 2,877 83.13   >34 348 12.72 534 15.43  Baby’s sex (female vs. male) 1,362 49.78 1,668 48.19  Receipt of social assistance 1,370 50.07 2,015 58.22  Rural residency 517 18.90 685 19.79 Maternal chronic conditions  Antiphospholipid syndrome 16 0.58 21 0.61  Chronic hypertension 75 2.74 118 3.41  Diabetes mellitus 87 3.18 136 3.93  Uterine defects 363 13.27 495 14.30 Pregnancy-related variables  Gestational diabetes 259 9.47 376 10.86  Eclampsia/preeclampsia 78 2.85 116 3.35  Anemia 385 14.07 509 14.71  Placental conditions 107 3.91 137 3.96  Placental abruption 95 3.47 125 3.61  Vaginal bleeding 381 13.93 448 12.94  Maternal infections 418 15.28 490 14.16  Fetal-maternal hemorrhage 5 0.18 10 0.29  Pregnancy-induced hypertension 167 6.10 241 6.96  Use of β-blockers 23 0.84 31 0.90 Asthma-related variables  Leukotriene-receptor antagonists 9 0.33 90 2.60  No. of SABA doses per week   0 1,549 56.62 434 12.54   0.01–3 762 27.85 1,346 38.89   >3 425 15.53 1,681 48.57  Oral corticosteroids 104 3.80 589 17.02  Intranasal corticosteroids 196 7.16 596 17.22  ≥1 ED visit for asthma 170 6.21 661 19.10  LABA 0 0 553 15.98  ≥1 hospitalization for asthma 6 0.22 63 1.82  Severity of asthma prior to pregnancy   Mild 2,479 90.61 2,499 72.20   Moderate 226 8.26 623 18.00   Severe 31 1.13 339 9.79 Abbreviations: ED, emergency department; ICS, inhaled corticosteroids; LABA, long-acting β2-agonists; SABA, short-acting β2-agonists. The crude results derived using standard univariate linear regression models are presented in Table 3. We observed that pregnancies exposed and unexposed to ICS were similar in terms of mean GA and BW. The standard univariate-adjusted analysis for GA (second and third columns in Table 4) revealed a positive association between ICS and GA (for GA associated with ICS, mean difference = 0.17 weeks, 95% confidence interval (CI): 0.05, 0.29). Although GA is arguably the strongest predictor of BW, the small positive association between ICS and GA did not translate to a significant association between ICS and BW (mean difference = 29.16 g, 95% CI: −4.81, 63.12). The regression coefficients for GA and BW obtained from the mixture model with K = 1 (last 2 columns in Table 4) were similar to those obtained from the standard univariate-adjusted regression models. Table 3. Associations Between Selected Characteristics of Asthmatic Pregnant Women and Their Pregnancy Outcomes, Quebec, Canada, 1998–2008a Characteristicb Pregnancy Outcome Gestational Age, weeks Birth Weight, g RC 95% CI RC 95% CI Gestational age, weeks N/A 185.97 180.52, 191.43 ICS exposure 0.08 −0.03, 0.18 1.54 −27.79, 30.87 Mother’s and baby’s characteristics  Maternal age, years   <18 0 Referent 0 Referent   18–34 −0.05 −0.47, 0.38 147.08 24.59, 269.57   >34 −0.38 −0.82, 0.06 111.91 −15.53, 239.36  Baby’s sex (female vs. male) −0.02 −0.12, 0.08 −147.41 −176.31, −118.51  Receipt of social assistance −0.29 −0.39, −0.19 −112.44 −141.56, −83.32  Rural residency −0.16 −0.29, −0.03 −40.03 −76.85, −3.21 Maternal chronic conditions  Antiphospholipid syndrome −1.09 −1.75, −0.43 −322.57 −511.45, −133.69  Chronic hypertension −0.54 −0.83, −0.25 −55.23 −139.06, 28.60  Diabetes mellitus −0.66 −0.93, −0.39 33.10 −45.09, 111.29  Uterine defects −0.59 −0.74, −0.44 −35.87 −78.03, 6.29 Pregnancy-related variables  Gestational diabetes −0.29 −0.46, −0.12 63.98 15.98, 111.98  Eclampsia/preeclampsia −1.33 −1.62, −1.04 −272.27 −355.63, −188.91  Anemia −0.12 −0.27, 0.02 58.25 16.83, 99.68  Placental conditions −0.80 −1.06, −0.54 −239.88 −314.53, −165.23  Placental abruption −1.85 −2.12, −1.58 −466.03 −543.88, −388.18  Vaginal bleeding −0.88 −1.02, −0.73 −207.58 −250.05, −165.10  Maternal infections −0.31 −0.46, −0.17 −33.40 −74.58, 7.78  Fetal-maternal hemorrhage −2.84 −3.87, −1.81 −492.31 −788.45, −196.18  Pregnancy-induced hypertension −0.30 −0.51, −0.10 −22.43 −81.15, 36.30  Use of β-blockers −0.48 −1.03, 0.07 −94.24 −250.93, 62.44 Asthma-related variables  Leukotriene-receptor antagonists −0.13 −0.53, 0.28 −76.21 −192.35, 39.94  No. of SABA doses per week   0 0 Referent 0 Referent   0.01–3 0.12 −0.01, 0.24 7.92 −27.93, 43.76   >3 −0.08 −0.20, 0.05 −41.16 −77.02, −5.31  Oral corticosteroids −0.07 −0.23, 0.09 −29.43 −75.63, 16.78  Intranasal corticosteroids −0.05 −0.20, 0.11 21.80 −21.82, 65.42  ≥1 ED visit for asthma 0.10 −0.05, 0.25 9.53 −33.21, 52.27  LABA −0.12 −0.30, 0.06 −20.40 −71.49, 30.68  ≥1 hospitalization for asthma −0.16 −0.65, 0.32 −101.28 −240.06, 37.49  Severity of asthma prior to pregnancy   Mild 0 Referent 0 Referent   Moderate 0.10 −0.05, 0.25 4.44 −38.10, 46.98   Severe −0.08 −0.30, 0.13 −98.31 −160.05, −36.58 Characteristicb Pregnancy Outcome Gestational Age, weeks Birth Weight, g RC 95% CI RC 95% CI Gestational age, weeks N/A 185.97 180.52, 191.43 ICS exposure 0.08 −0.03, 0.18 1.54 −27.79, 30.87 Mother’s and baby’s characteristics  Maternal age, years   <18 0 Referent 0 Referent   18–34 −0.05 −0.47, 0.38 147.08 24.59, 269.57   >34 −0.38 −0.82, 0.06 111.91 −15.53, 239.36  Baby’s sex (female vs. male) −0.02 −0.12, 0.08 −147.41 −176.31, −118.51  Receipt of social assistance −0.29 −0.39, −0.19 −112.44 −141.56, −83.32  Rural residency −0.16 −0.29, −0.03 −40.03 −76.85, −3.21 Maternal chronic conditions  Antiphospholipid syndrome −1.09 −1.75, −0.43 −322.57 −511.45, −133.69  Chronic hypertension −0.54 −0.83, −0.25 −55.23 −139.06, 28.60  Diabetes mellitus −0.66 −0.93, −0.39 33.10 −45.09, 111.29  Uterine defects −0.59 −0.74, −0.44 −35.87 −78.03, 6.29 Pregnancy-related variables  Gestational diabetes −0.29 −0.46, −0.12 63.98 15.98, 111.98  Eclampsia/preeclampsia −1.33 −1.62, −1.04 −272.27 −355.63, −188.91  Anemia −0.12 −0.27, 0.02 58.25 16.83, 99.68  Placental conditions −0.80 −1.06, −0.54 −239.88 −314.53, −165.23  Placental abruption −1.85 −2.12, −1.58 −466.03 −543.88, −388.18  Vaginal bleeding −0.88 −1.02, −0.73 −207.58 −250.05, −165.10  Maternal infections −0.31 −0.46, −0.17 −33.40 −74.58, 7.78  Fetal-maternal hemorrhage −2.84 −3.87, −1.81 −492.31 −788.45, −196.18  Pregnancy-induced hypertension −0.30 −0.51, −0.10 −22.43 −81.15, 36.30  Use of β-blockers −0.48 −1.03, 0.07 −94.24 −250.93, 62.44 Asthma-related variables  Leukotriene-receptor antagonists −0.13 −0.53, 0.28 −76.21 −192.35, 39.94  No. of SABA doses per week   0 0 Referent 0 Referent   0.01–3 0.12 −0.01, 0.24 7.92 −27.93, 43.76   >3 −0.08 −0.20, 0.05 −41.16 −77.02, −5.31  Oral corticosteroids −0.07 −0.23, 0.09 −29.43 −75.63, 16.78  Intranasal corticosteroids −0.05 −0.20, 0.11 21.80 −21.82, 65.42  ≥1 ED visit for asthma 0.10 −0.05, 0.25 9.53 −33.21, 52.27  LABA −0.12 −0.30, 0.06 −20.40 −71.49, 30.68  ≥1 hospitalization for asthma −0.16 −0.65, 0.32 −101.28 −240.06, 37.49  Severity of asthma prior to pregnancy   Mild 0 Referent 0 Referent   Moderate 0.10 −0.05, 0.25 4.44 −38.10, 46.98   Severe −0.08 −0.30, 0.13 −98.31 −160.05, −36.58 Abbreviations: CI, confidence interval; ED, emergency department; ICS, inhaled corticosteroids; LABA, long-acting β2-agonists; N/A, not applicable; RC, regression coefficient; SABA, short-acting β2-agonists. a Regression coefficients with 95% CIs were obtained from simple univariate linear models. b All variables are binary (yes = 1/no = 0) unless otherwise indicated. Table 3. Associations Between Selected Characteristics of Asthmatic Pregnant Women and Their Pregnancy Outcomes, Quebec, Canada, 1998–2008a Characteristicb Pregnancy Outcome Gestational Age, weeks Birth Weight, g RC 95% CI RC 95% CI Gestational age, weeks N/A 185.97 180.52, 191.43 ICS exposure 0.08 −0.03, 0.18 1.54 −27.79, 30.87 Mother’s and baby’s characteristics  Maternal age, years   <18 0 Referent 0 Referent   18–34 −0.05 −0.47, 0.38 147.08 24.59, 269.57   >34 −0.38 −0.82, 0.06 111.91 −15.53, 239.36  Baby’s sex (female vs. male) −0.02 −0.12, 0.08 −147.41 −176.31, −118.51  Receipt of social assistance −0.29 −0.39, −0.19 −112.44 −141.56, −83.32  Rural residency −0.16 −0.29, −0.03 −40.03 −76.85, −3.21 Maternal chronic conditions  Antiphospholipid syndrome −1.09 −1.75, −0.43 −322.57 −511.45, −133.69  Chronic hypertension −0.54 −0.83, −0.25 −55.23 −139.06, 28.60  Diabetes mellitus −0.66 −0.93, −0.39 33.10 −45.09, 111.29  Uterine defects −0.59 −0.74, −0.44 −35.87 −78.03, 6.29 Pregnancy-related variables  Gestational diabetes −0.29 −0.46, −0.12 63.98 15.98, 111.98  Eclampsia/preeclampsia −1.33 −1.62, −1.04 −272.27 −355.63, −188.91  Anemia −0.12 −0.27, 0.02 58.25 16.83, 99.68  Placental conditions −0.80 −1.06, −0.54 −239.88 −314.53, −165.23  Placental abruption −1.85 −2.12, −1.58 −466.03 −543.88, −388.18  Vaginal bleeding −0.88 −1.02, −0.73 −207.58 −250.05, −165.10  Maternal infections −0.31 −0.46, −0.17 −33.40 −74.58, 7.78  Fetal-maternal hemorrhage −2.84 −3.87, −1.81 −492.31 −788.45, −196.18  Pregnancy-induced hypertension −0.30 −0.51, −0.10 −22.43 −81.15, 36.30  Use of β-blockers −0.48 −1.03, 0.07 −94.24 −250.93, 62.44 Asthma-related variables  Leukotriene-receptor antagonists −0.13 −0.53, 0.28 −76.21 −192.35, 39.94  No. of SABA doses per week   0 0 Referent 0 Referent   0.01–3 0.12 −0.01, 0.24 7.92 −27.93, 43.76   >3 −0.08 −0.20, 0.05 −41.16 −77.02, −5.31  Oral corticosteroids −0.07 −0.23, 0.09 −29.43 −75.63, 16.78  Intranasal corticosteroids −0.05 −0.20, 0.11 21.80 −21.82, 65.42  ≥1 ED visit for asthma 0.10 −0.05, 0.25 9.53 −33.21, 52.27  LABA −0.12 −0.30, 0.06 −20.40 −71.49, 30.68  ≥1 hospitalization for asthma −0.16 −0.65, 0.32 −101.28 −240.06, 37.49  Severity of asthma prior to pregnancy   Mild 0 Referent 0 Referent   Moderate 0.10 −0.05, 0.25 4.44 −38.10, 46.98   Severe −0.08 −0.30, 0.13 −98.31 −160.05, −36.58 Characteristicb Pregnancy Outcome Gestational Age, weeks Birth Weight, g RC 95% CI RC 95% CI Gestational age, weeks N/A 185.97 180.52, 191.43 ICS exposure 0.08 −0.03, 0.18 1.54 −27.79, 30.87 Mother’s and baby’s characteristics  Maternal age, years   <18 0 Referent 0 Referent   18–34 −0.05 −0.47, 0.38 147.08 24.59, 269.57   >34 −0.38 −0.82, 0.06 111.91 −15.53, 239.36  Baby’s sex (female vs. male) −0.02 −0.12, 0.08 −147.41 −176.31, −118.51  Receipt of social assistance −0.29 −0.39, −0.19 −112.44 −141.56, −83.32  Rural residency −0.16 −0.29, −0.03 −40.03 −76.85, −3.21 Maternal chronic conditions  Antiphospholipid syndrome −1.09 −1.75, −0.43 −322.57 −511.45, −133.69  Chronic hypertension −0.54 −0.83, −0.25 −55.23 −139.06, 28.60  Diabetes mellitus −0.66 −0.93, −0.39 33.10 −45.09, 111.29  Uterine defects −0.59 −0.74, −0.44 −35.87 −78.03, 6.29 Pregnancy-related variables  Gestational diabetes −0.29 −0.46, −0.12 63.98 15.98, 111.98  Eclampsia/preeclampsia −1.33 −1.62, −1.04 −272.27 −355.63, −188.91  Anemia −0.12 −0.27, 0.02 58.25 16.83, 99.68  Placental conditions −0.80 −1.06, −0.54 −239.88 −314.53, −165.23  Placental abruption −1.85 −2.12, −1.58 −466.03 −543.88, −388.18  Vaginal bleeding −0.88 −1.02, −0.73 −207.58 −250.05, −165.10  Maternal infections −0.31 −0.46, −0.17 −33.40 −74.58, 7.78  Fetal-maternal hemorrhage −2.84 −3.87, −1.81 −492.31 −788.45, −196.18  Pregnancy-induced hypertension −0.30 −0.51, −0.10 −22.43 −81.15, 36.30  Use of β-blockers −0.48 −1.03, 0.07 −94.24 −250.93, 62.44 Asthma-related variables  Leukotriene-receptor antagonists −0.13 −0.53, 0.28 −76.21 −192.35, 39.94  No. of SABA doses per week   0 0 Referent 0 Referent   0.01–3 0.12 −0.01, 0.24 7.92 −27.93, 43.76   >3 −0.08 −0.20, 0.05 −41.16 −77.02, −5.31  Oral corticosteroids −0.07 −0.23, 0.09 −29.43 −75.63, 16.78  Intranasal corticosteroids −0.05 −0.20, 0.11 21.80 −21.82, 65.42  ≥1 ED visit for asthma 0.10 −0.05, 0.25 9.53 −33.21, 52.27  LABA −0.12 −0.30, 0.06 −20.40 −71.49, 30.68  ≥1 hospitalization for asthma −0.16 −0.65, 0.32 −101.28 −240.06, 37.49  Severity of asthma prior to pregnancy   Mild 0 Referent 0 Referent   Moderate 0.10 −0.05, 0.25 4.44 −38.10, 46.98   Severe −0.08 −0.30, 0.13 −98.31 −160.05, −36.58 Abbreviations: CI, confidence interval; ED, emergency department; ICS, inhaled corticosteroids; LABA, long-acting β2-agonists; N/A, not applicable; RC, regression coefficient; SABA, short-acting β2-agonists. a Regression coefficients with 95% CIs were obtained from simple univariate linear models. b All variables are binary (yes = 1/no = 0) unless otherwise indicated. Table 4. Associations of Inhaled Corticosteroid Exposure With Gestational Age and Birth Weight (Adjusted Models) Among Asthmatic Pregnant Women, Quebec, Canada, 1998–2008 Characteristica Univariate-Adjusted Model 1-Component Adjusted Mixture Model GA, weeks BW, g GA, weeks BW, g RC 95% CI RC 95% CI RC 95% CrI RC 95% CrI Intercept 38.98 38.55, 39.40 3,256.19 3,133.98, 3,378.40 39.02 38.60, 39.44 3,274.01 3,157.17, 3,390.80 ICS exposure 0.17 0.05, 0.29 29.16 −4.81, 63.12 0.17 0.05, 0.29 28.78 −5.14, 62.51 Mother’s and baby’s characteristics  Maternal age, years   <18 0 Referent 0 Referent 0 Referent 0 Referent   18–34 0.05 −0.36, 0.47 157.72 38.89, 276.55 0.01 −0.39, 0.42 140.05 26.61, 252.76   >34 −0.19 −0.62, 0.24 113.81 −10.33, 237.94 −0.23 −0.65, 0.19 95.80 −22.97, 213.66  Baby’s sex (female vs. male) 0.01 −0.09, 0.11 −142.11 −170.32, −113.91 0.01 −0.09, 0.11 −142.37 −170.51, −114.27  Receipt of social assistance −0.30 −0.40, −0.20 −118.13 −146.87, −89.39 −0.30 −0.40, −0.20 −118.02 −146.69, −89.44  Rural residency −0.21 −0.33, −0.08 −50.49 −86.44, −14.54 −0.21 −0.33, −0.08 −50.49 −86.41, −14.73 Maternal chronic conditions  Antiphospholipid syndrome −0.92 −1.55, −0.28 −341.40 −524.57, −158.22 −0.86 −1.48, −0.23 −313.12 −488.25, −138.52  Chronic hypertension −0.23 −0.52, 0.06 −32.30 −116.42, 51.82 −0.23 −0.52, 0.06 −32.02 −114.92, 50.86  Diabetes mellitus −0.38 −0.65, −0.12 105.14 28.58, 181.70 −0.39 −0.65, −0.12 102.76 26.36, 178.47  Uterine defects −0.47 −0.61, −0.32 −28.85 −70.22, 12.53 −0.47 −0.61, −0.32 −28.71 −69.94, 12.25 Pregnancy-related variables  Gestational diabetes −0.20 −0.36, −0.03 81.78 34.46, 129.11 −0.20 −0.36, −0.03 81.00 33.65, 128.08  Eclampsia/preeclampsia −1.22 −1.51, −0.92 −270.91 −356.08, −185.74 −1.21 −1.50, −0.91 −266.56 −350.73, −182.23  Anemia −0.04 −0.18, 0.10 74.53 34.18, 114.89 −0.04 −0.18, 0.10 73.70 33.66, 113.68  Placental conditions −0.33 −0.59, −0.07 −144.19 −218.69, −69.69 −0.33 −0.58, −0.07 −143.47 −216.75, −70.32  Placental abruption −1.34 −1.64, −1.03 −356.21 −443.47, −268.96 −1.32 −1.62, −1.02 −350.66 −436.76, −264.08  Vaginal bleeding −0.42 −0.59, −0.25 −92.93 −141.19, −44.68 −0.42 −0.59, −0.26 −94.39 −142.35, −46.30  Maternal infections −0.17 −0.31, −0.03 −9.77 −49.99, 30.45 −0.17 −0.31, −0.03 −9.91 −50.15, 30.26  Fetal-maternal hemorrhage −2.40 −3.40, −1.40 −335.49 −623.96, −47.02 −2.28 −3.23, −1.31 −277.66 −538.28, −16.82  Pregnancy-induced hypertension −0.01 −0.22, 0.20 17.90 −42.12, 77.92 −0.01 −0.22, 0.19 16.95 −42.28, 76.64  Use of β-blockers −0.12 −0.66, 0.42 −64.52 −219.89, 90.84 −0.11 −0.64, 0.42 −61.00 −210.51, 88.67 Asthma-related variables  Leukotriene-receptor antagonists 0.03 −0.38, 0.43 −30.25 −146.78, 86.27 0.03 −0.37, 0.43 −29.39 −144.01, 85.32  No. of SABA doses per week   0 0 Referent 0 Referent 0 Referent 0 Referent   0.01–3 0.06 −0.08, 0.19 −1.93 −39.56, 35.70 0.05 −0.08, 0.19 −2.15 −39.68, 35.51   >3 −0.18 −0.33, −0.04 −44.09 −86.45, −1.72 −0.18 −0.33, −0.04 −44.34 −86.88, −1.57  Oral corticosteroids −0.08 −0.26, 0.10 −29.51 −80.98, 21.96 −0.08 −0.26, 0.10 −28.86 −80.08, 22.19  Intranasal corticosteroids −0.01 −0.15, 0.14 23.75 −19.35, 66.85 −0.01 −0.15, 0.14 23.54 −19.33, 66.36  ≥1 ED visit for asthma 0.13 −0.03, 0.30 30.54 −16.32, 77.40 0.13 −0.03, 0.29 30.19 −16.68, 76.65  LABA −0.11 −0.30, 0.08 1.39 −53.35, 56.13 −0.11 −0.30, 0.08 1.60 −52.88, 56.02  ≥1 hospitalization for asthma −0.08 −0.56, 0.41 −81.84 −220.82, 57.14 −0.07 −0.54, 0.41 −78.29 −214.34, 57.85  Severity of asthma prior to pregnancy   Mild 0 Referent 0 Referent 0 Referent 0 Referent   Moderate 0.17 0.02, 0.33 11.06 −33.85, 55.97 0.17 0.02, 0.33 10.92 −34.01, 55.62   Severe 0.06 −0.17, 0.29 −71.50 −138.28, −4.72 0.06 −0.17, 0.29 −71.36 −137.81, −4.57 GA residual 184.83 179.31, 190.37 Characteristica Univariate-Adjusted Model 1-Component Adjusted Mixture Model GA, weeks BW, g GA, weeks BW, g RC 95% CI RC 95% CI RC 95% CrI RC 95% CrI Intercept 38.98 38.55, 39.40 3,256.19 3,133.98, 3,378.40 39.02 38.60, 39.44 3,274.01 3,157.17, 3,390.80 ICS exposure 0.17 0.05, 0.29 29.16 −4.81, 63.12 0.17 0.05, 0.29 28.78 −5.14, 62.51 Mother’s and baby’s characteristics  Maternal age, years   <18 0 Referent 0 Referent 0 Referent 0 Referent   18–34 0.05 −0.36, 0.47 157.72 38.89, 276.55 0.01 −0.39, 0.42 140.05 26.61, 252.76   >34 −0.19 −0.62, 0.24 113.81 −10.33, 237.94 −0.23 −0.65, 0.19 95.80 −22.97, 213.66  Baby’s sex (female vs. male) 0.01 −0.09, 0.11 −142.11 −170.32, −113.91 0.01 −0.09, 0.11 −142.37 −170.51, −114.27  Receipt of social assistance −0.30 −0.40, −0.20 −118.13 −146.87, −89.39 −0.30 −0.40, −0.20 −118.02 −146.69, −89.44  Rural residency −0.21 −0.33, −0.08 −50.49 −86.44, −14.54 −0.21 −0.33, −0.08 −50.49 −86.41, −14.73 Maternal chronic conditions  Antiphospholipid syndrome −0.92 −1.55, −0.28 −341.40 −524.57, −158.22 −0.86 −1.48, −0.23 −313.12 −488.25, −138.52  Chronic hypertension −0.23 −0.52, 0.06 −32.30 −116.42, 51.82 −0.23 −0.52, 0.06 −32.02 −114.92, 50.86  Diabetes mellitus −0.38 −0.65, −0.12 105.14 28.58, 181.70 −0.39 −0.65, −0.12 102.76 26.36, 178.47  Uterine defects −0.47 −0.61, −0.32 −28.85 −70.22, 12.53 −0.47 −0.61, −0.32 −28.71 −69.94, 12.25 Pregnancy-related variables  Gestational diabetes −0.20 −0.36, −0.03 81.78 34.46, 129.11 −0.20 −0.36, −0.03 81.00 33.65, 128.08  Eclampsia/preeclampsia −1.22 −1.51, −0.92 −270.91 −356.08, −185.74 −1.21 −1.50, −0.91 −266.56 −350.73, −182.23  Anemia −0.04 −0.18, 0.10 74.53 34.18, 114.89 −0.04 −0.18, 0.10 73.70 33.66, 113.68  Placental conditions −0.33 −0.59, −0.07 −144.19 −218.69, −69.69 −0.33 −0.58, −0.07 −143.47 −216.75, −70.32  Placental abruption −1.34 −1.64, −1.03 −356.21 −443.47, −268.96 −1.32 −1.62, −1.02 −350.66 −436.76, −264.08  Vaginal bleeding −0.42 −0.59, −0.25 −92.93 −141.19, −44.68 −0.42 −0.59, −0.26 −94.39 −142.35, −46.30  Maternal infections −0.17 −0.31, −0.03 −9.77 −49.99, 30.45 −0.17 −0.31, −0.03 −9.91 −50.15, 30.26  Fetal-maternal hemorrhage −2.40 −3.40, −1.40 −335.49 −623.96, −47.02 −2.28 −3.23, −1.31 −277.66 −538.28, −16.82  Pregnancy-induced hypertension −0.01 −0.22, 0.20 17.90 −42.12, 77.92 −0.01 −0.22, 0.19 16.95 −42.28, 76.64  Use of β-blockers −0.12 −0.66, 0.42 −64.52 −219.89, 90.84 −0.11 −0.64, 0.42 −61.00 −210.51, 88.67 Asthma-related variables  Leukotriene-receptor antagonists 0.03 −0.38, 0.43 −30.25 −146.78, 86.27 0.03 −0.37, 0.43 −29.39 −144.01, 85.32  No. of SABA doses per week   0 0 Referent 0 Referent 0 Referent 0 Referent   0.01–3 0.06 −0.08, 0.19 −1.93 −39.56, 35.70 0.05 −0.08, 0.19 −2.15 −39.68, 35.51   >3 −0.18 −0.33, −0.04 −44.09 −86.45, −1.72 −0.18 −0.33, −0.04 −44.34 −86.88, −1.57  Oral corticosteroids −0.08 −0.26, 0.10 −29.51 −80.98, 21.96 −0.08 −0.26, 0.10 −28.86 −80.08, 22.19  Intranasal corticosteroids −0.01 −0.15, 0.14 23.75 −19.35, 66.85 −0.01 −0.15, 0.14 23.54 −19.33, 66.36  ≥1 ED visit for asthma 0.13 −0.03, 0.30 30.54 −16.32, 77.40 0.13 −0.03, 0.29 30.19 −16.68, 76.65  LABA −0.11 −0.30, 0.08 1.39 −53.35, 56.13 −0.11 −0.30, 0.08 1.60 −52.88, 56.02  ≥1 hospitalization for asthma −0.08 −0.56, 0.41 −81.84 −220.82, 57.14 −0.07 −0.54, 0.41 −78.29 −214.34, 57.85  Severity of asthma prior to pregnancy   Mild 0 Referent 0 Referent 0 Referent 0 Referent   Moderate 0.17 0.02, 0.33 11.06 −33.85, 55.97 0.17 0.02, 0.33 10.92 −34.01, 55.62   Severe 0.06 −0.17, 0.29 −71.50 −138.28, −4.72 0.06 −0.17, 0.29 −71.36 −137.81, −4.57 GA residual 184.83 179.31, 190.37 Abbreviations: BW, birth weight; CI, confidence interval; CrI, credible interval; ED, emergency department; GA, gestational age; ICS, inhaled corticosteroids; LABA, long-acting β2-agonists; RC, regression coefficient; SABA, short-acting β2-agonists. a All adjustment variables are binary (yes = 1/no = 0) unless otherwise indicated. Table 4. Associations of Inhaled Corticosteroid Exposure With Gestational Age and Birth Weight (Adjusted Models) Among Asthmatic Pregnant Women, Quebec, Canada, 1998–2008 Characteristica Univariate-Adjusted Model 1-Component Adjusted Mixture Model GA, weeks BW, g GA, weeks BW, g RC 95% CI RC 95% CI RC 95% CrI RC 95% CrI Intercept 38.98 38.55, 39.40 3,256.19 3,133.98, 3,378.40 39.02 38.60, 39.44 3,274.01 3,157.17, 3,390.80 ICS exposure 0.17 0.05, 0.29 29.16 −4.81, 63.12 0.17 0.05, 0.29 28.78 −5.14, 62.51 Mother’s and baby’s characteristics  Maternal age, years   <18 0 Referent 0 Referent 0 Referent 0 Referent   18–34 0.05 −0.36, 0.47 157.72 38.89, 276.55 0.01 −0.39, 0.42 140.05 26.61, 252.76   >34 −0.19 −0.62, 0.24 113.81 −10.33, 237.94 −0.23 −0.65, 0.19 95.80 −22.97, 213.66  Baby’s sex (female vs. male) 0.01 −0.09, 0.11 −142.11 −170.32, −113.91 0.01 −0.09, 0.11 −142.37 −170.51, −114.27  Receipt of social assistance −0.30 −0.40, −0.20 −118.13 −146.87, −89.39 −0.30 −0.40, −0.20 −118.02 −146.69, −89.44  Rural residency −0.21 −0.33, −0.08 −50.49 −86.44, −14.54 −0.21 −0.33, −0.08 −50.49 −86.41, −14.73 Maternal chronic conditions  Antiphospholipid syndrome −0.92 −1.55, −0.28 −341.40 −524.57, −158.22 −0.86 −1.48, −0.23 −313.12 −488.25, −138.52  Chronic hypertension −0.23 −0.52, 0.06 −32.30 −116.42, 51.82 −0.23 −0.52, 0.06 −32.02 −114.92, 50.86  Diabetes mellitus −0.38 −0.65, −0.12 105.14 28.58, 181.70 −0.39 −0.65, −0.12 102.76 26.36, 178.47  Uterine defects −0.47 −0.61, −0.32 −28.85 −70.22, 12.53 −0.47 −0.61, −0.32 −28.71 −69.94, 12.25 Pregnancy-related variables  Gestational diabetes −0.20 −0.36, −0.03 81.78 34.46, 129.11 −0.20 −0.36, −0.03 81.00 33.65, 128.08  Eclampsia/preeclampsia −1.22 −1.51, −0.92 −270.91 −356.08, −185.74 −1.21 −1.50, −0.91 −266.56 −350.73, −182.23  Anemia −0.04 −0.18, 0.10 74.53 34.18, 114.89 −0.04 −0.18, 0.10 73.70 33.66, 113.68  Placental conditions −0.33 −0.59, −0.07 −144.19 −218.69, −69.69 −0.33 −0.58, −0.07 −143.47 −216.75, −70.32  Placental abruption −1.34 −1.64, −1.03 −356.21 −443.47, −268.96 −1.32 −1.62, −1.02 −350.66 −436.76, −264.08  Vaginal bleeding −0.42 −0.59, −0.25 −92.93 −141.19, −44.68 −0.42 −0.59, −0.26 −94.39 −142.35, −46.30  Maternal infections −0.17 −0.31, −0.03 −9.77 −49.99, 30.45 −0.17 −0.31, −0.03 −9.91 −50.15, 30.26  Fetal-maternal hemorrhage −2.40 −3.40, −1.40 −335.49 −623.96, −47.02 −2.28 −3.23, −1.31 −277.66 −538.28, −16.82  Pregnancy-induced hypertension −0.01 −0.22, 0.20 17.90 −42.12, 77.92 −0.01 −0.22, 0.19 16.95 −42.28, 76.64  Use of β-blockers −0.12 −0.66, 0.42 −64.52 −219.89, 90.84 −0.11 −0.64, 0.42 −61.00 −210.51, 88.67 Asthma-related variables  Leukotriene-receptor antagonists 0.03 −0.38, 0.43 −30.25 −146.78, 86.27 0.03 −0.37, 0.43 −29.39 −144.01, 85.32  No. of SABA doses per week   0 0 Referent 0 Referent 0 Referent 0 Referent   0.01–3 0.06 −0.08, 0.19 −1.93 −39.56, 35.70 0.05 −0.08, 0.19 −2.15 −39.68, 35.51   >3 −0.18 −0.33, −0.04 −44.09 −86.45, −1.72 −0.18 −0.33, −0.04 −44.34 −86.88, −1.57  Oral corticosteroids −0.08 −0.26, 0.10 −29.51 −80.98, 21.96 −0.08 −0.26, 0.10 −28.86 −80.08, 22.19  Intranasal corticosteroids −0.01 −0.15, 0.14 23.75 −19.35, 66.85 −0.01 −0.15, 0.14 23.54 −19.33, 66.36  ≥1 ED visit for asthma 0.13 −0.03, 0.30 30.54 −16.32, 77.40 0.13 −0.03, 0.29 30.19 −16.68, 76.65  LABA −0.11 −0.30, 0.08 1.39 −53.35, 56.13 −0.11 −0.30, 0.08 1.60 −52.88, 56.02  ≥1 hospitalization for asthma −0.08 −0.56, 0.41 −81.84 −220.82, 57.14 −0.07 −0.54, 0.41 −78.29 −214.34, 57.85  Severity of asthma prior to pregnancy   Mild 0 Referent 0 Referent 0 Referent 0 Referent   Moderate 0.17 0.02, 0.33 11.06 −33.85, 55.97 0.17 0.02, 0.33 10.92 −34.01, 55.62   Severe 0.06 −0.17, 0.29 −71.50 −138.28, −4.72 0.06 −0.17, 0.29 −71.36 −137.81, −4.57 GA residual 184.83 179.31, 190.37 Characteristica Univariate-Adjusted Model 1-Component Adjusted Mixture Model GA, weeks BW, g GA, weeks BW, g RC 95% CI RC 95% CI RC 95% CrI RC 95% CrI Intercept 38.98 38.55, 39.40 3,256.19 3,133.98, 3,378.40 39.02 38.60, 39.44 3,274.01 3,157.17, 3,390.80 ICS exposure 0.17 0.05, 0.29 29.16 −4.81, 63.12 0.17 0.05, 0.29 28.78 −5.14, 62.51 Mother’s and baby’s characteristics  Maternal age, years   <18 0 Referent 0 Referent 0 Referent 0 Referent   18–34 0.05 −0.36, 0.47 157.72 38.89, 276.55 0.01 −0.39, 0.42 140.05 26.61, 252.76   >34 −0.19 −0.62, 0.24 113.81 −10.33, 237.94 −0.23 −0.65, 0.19 95.80 −22.97, 213.66  Baby’s sex (female vs. male) 0.01 −0.09, 0.11 −142.11 −170.32, −113.91 0.01 −0.09, 0.11 −142.37 −170.51, −114.27  Receipt of social assistance −0.30 −0.40, −0.20 −118.13 −146.87, −89.39 −0.30 −0.40, −0.20 −118.02 −146.69, −89.44  Rural residency −0.21 −0.33, −0.08 −50.49 −86.44, −14.54 −0.21 −0.33, −0.08 −50.49 −86.41, −14.73 Maternal chronic conditions  Antiphospholipid syndrome −0.92 −1.55, −0.28 −341.40 −524.57, −158.22 −0.86 −1.48, −0.23 −313.12 −488.25, −138.52  Chronic hypertension −0.23 −0.52, 0.06 −32.30 −116.42, 51.82 −0.23 −0.52, 0.06 −32.02 −114.92, 50.86  Diabetes mellitus −0.38 −0.65, −0.12 105.14 28.58, 181.70 −0.39 −0.65, −0.12 102.76 26.36, 178.47  Uterine defects −0.47 −0.61, −0.32 −28.85 −70.22, 12.53 −0.47 −0.61, −0.32 −28.71 −69.94, 12.25 Pregnancy-related variables  Gestational diabetes −0.20 −0.36, −0.03 81.78 34.46, 129.11 −0.20 −0.36, −0.03 81.00 33.65, 128.08  Eclampsia/preeclampsia −1.22 −1.51, −0.92 −270.91 −356.08, −185.74 −1.21 −1.50, −0.91 −266.56 −350.73, −182.23  Anemia −0.04 −0.18, 0.10 74.53 34.18, 114.89 −0.04 −0.18, 0.10 73.70 33.66, 113.68  Placental conditions −0.33 −0.59, −0.07 −144.19 −218.69, −69.69 −0.33 −0.58, −0.07 −143.47 −216.75, −70.32  Placental abruption −1.34 −1.64, −1.03 −356.21 −443.47, −268.96 −1.32 −1.62, −1.02 −350.66 −436.76, −264.08  Vaginal bleeding −0.42 −0.59, −0.25 −92.93 −141.19, −44.68 −0.42 −0.59, −0.26 −94.39 −142.35, −46.30  Maternal infections −0.17 −0.31, −0.03 −9.77 −49.99, 30.45 −0.17 −0.31, −0.03 −9.91 −50.15, 30.26  Fetal-maternal hemorrhage −2.40 −3.40, −1.40 −335.49 −623.96, −47.02 −2.28 −3.23, −1.31 −277.66 −538.28, −16.82  Pregnancy-induced hypertension −0.01 −0.22, 0.20 17.90 −42.12, 77.92 −0.01 −0.22, 0.19 16.95 −42.28, 76.64  Use of β-blockers −0.12 −0.66, 0.42 −64.52 −219.89, 90.84 −0.11 −0.64, 0.42 −61.00 −210.51, 88.67 Asthma-related variables  Leukotriene-receptor antagonists 0.03 −0.38, 0.43 −30.25 −146.78, 86.27 0.03 −0.37, 0.43 −29.39 −144.01, 85.32  No. of SABA doses per week   0 0 Referent 0 Referent 0 Referent 0 Referent   0.01–3 0.06 −0.08, 0.19 −1.93 −39.56, 35.70 0.05 −0.08, 0.19 −2.15 −39.68, 35.51   >3 −0.18 −0.33, −0.04 −44.09 −86.45, −1.72 −0.18 −0.33, −0.04 −44.34 −86.88, −1.57  Oral corticosteroids −0.08 −0.26, 0.10 −29.51 −80.98, 21.96 −0.08 −0.26, 0.10 −28.86 −80.08, 22.19  Intranasal corticosteroids −0.01 −0.15, 0.14 23.75 −19.35, 66.85 −0.01 −0.15, 0.14 23.54 −19.33, 66.36  ≥1 ED visit for asthma 0.13 −0.03, 0.30 30.54 −16.32, 77.40 0.13 −0.03, 0.29 30.19 −16.68, 76.65  LABA −0.11 −0.30, 0.08 1.39 −53.35, 56.13 −0.11 −0.30, 0.08 1.60 −52.88, 56.02  ≥1 hospitalization for asthma −0.08 −0.56, 0.41 −81.84 −220.82, 57.14 −0.07 −0.54, 0.41 −78.29 −214.34, 57.85  Severity of asthma prior to pregnancy   Mild 0 Referent 0 Referent 0 Referent 0 Referent   Moderate 0.17 0.02, 0.33 11.06 −33.85, 55.97 0.17 0.02, 0.33 10.92 −34.01, 55.62   Severe 0.06 −0.17, 0.29 −71.50 −138.28, −4.72 0.06 −0.17, 0.29 −71.36 −137.81, −4.57 GA residual 184.83 179.31, 190.37 Abbreviations: BW, birth weight; CI, confidence interval; CrI, credible interval; ED, emergency department; GA, gestational age; ICS, inhaled corticosteroids; LABA, long-acting β2-agonists; RC, regression coefficient; SABA, short-acting β2-agonists. a All adjustment variables are binary (yes = 1/no = 0) unless otherwise indicated. The results for the mixture model with K = 2 are presented in Table 5. The mixing proportions π1 and π2, which represent the relative size of each component (subpopulation) in the population, were estimated to be 0.92 (95% credible interval (CrI): 0.90, 0.93) and 0.08 (95% CrI: 0.07, 0.10), respectively. We observed that the second component was associated with pregnancies with smaller GA and BW. The first component was centered at 39.51 weeks (95% CrI: 39.20, 39.82) and 3,358 g (95% CrI: 3,252, 3,464). The second component was centered at 33.24 weeks (95% CrI: 30.82, 35.44) and 2,182 g (95% CrI: 1,848, 2,511). In the first component, the mean difference associated with ICS for GA was −0.01 week (95% CrI: −0.10, 0.08), while the mean difference for BW was 1.27 g (95% CrI: −29.98, 32.15). Corresponding figures in the second component were 1.55 weeks (95% CrI: 0.67, 2.42) for GA and 200.3 g (95% CrI: 34.10, 366.0) for BW. In general, we observed large estimation uncertainty for many regression coefficients in the second component as a result of the relatively small representation of this component in our cohort of size 6,197 (equivalent of n times the estimated π2 value, i.e., 496). As reported in Web Appendix 2, our additional analyses showed robustness of the results to the choice of the prior distribution. Table 5. Associations of Inhaled Corticosteroid Exposure With Gestational Age and Birth Weight Among Asthmatic Pregnant Women in Each Component of an Adjusted 2-Component Adjusted Mixture-of-Bivariate-Regressions Model, Quebec, Canada, 1998–2008 Characteristica Component 1b Component 2c GA, weeks BW, g GA, weeks BW, g RC 95% CrI RC 95% CrI RC 95% CrI RC 95% CrI Intercept 39.51 39.20, 39.82 3,357.68 3,251.77, 3,463.80 33.24 30.82, 35.44 2,182.29 1,848.04, 2,511.27 ICS exposure −0.01 −0.10, 0.08 1.27 −29.98, 32.15 1.55 0.67, 2.42 200.33 34.10, 365.97 Mother’s and baby’s characteristics  Maternal age, years   <18 0 Referent 0 Referent 0 Referent 0 Referent   18–34 −0.27 −0.57, 0.02 110.16 7.03, 212.89 2.52 0.42, 4.85 396.48 84.92, 712.79   >34 −0.34 −0.65, −0.03 108.30 0.08, 215.71 1.02 −1.22, 3.43 33.24 −303.02, 369.88  Baby’s sex (female vs. male) 0.04 −0.03, 0.11 −143.34 −169.07,−117.62 −0.32 −1.08, 0.43 −142.10 −286.54, 1.54  Receipt of social assistance −0.20 −0.28, −0.13 −98.18 −124.44, −71.92 −0.70 −1.46, 0.07 −169.27 −314.51, −22.65  Rural residency −0.15 −0.24, −0.06 −38.53 −71.32, −5.54 −0.50 −1.45, 0.42 −90.59 −269.02, 81.75 Maternal chronic conditions  Antiphospholipid syndrome −0.53 −1.00, −0.07 −244.61 −410.75, −77.66 −0.54 −3.85, 2.51 −213.44 −694.67, 256.87  Chronic hypertension −0.33 −0.54, −0.13 −56.37 −132.13, 19.98 0.14 −1.94, 2.10 −8.04 −385.33, 353.93  Diabetes mellitus −0.51 −0.71, −0.32 63.96 −5.02, 133.07 −0.12 −2.24, 1.89 257.60 −89.35, 591.89  Uterine defects −0.52 −0.62, −0.42 −31.51 −69.19, 6.01 −0.73 −1.94, 0.44 −164.26 −389.89, 54.18 Pregnancy-related variables  Gestational diabetes −0.38 −0.50, −0.26 54.80 11.82, 98.01 1.22 −0.05, 2.44 207.57 −39.53, 443.16  Eclampsia/preeclampsia −0.52 −0.75, −0.28 −79.58 −162.51, 4.65 −1.87 −3.28, −0.47 −539.05 −804.27, −278.65  Anemia 0.10 0.00, 0.20 107.97 71.27, 144.83 −0.50 −1.53, 0.51 −80.34 −275.09, 109.39  Placental conditions −0.30 −0.50, −0.10 −151.93 −220.79, −82.81 −0.16 −1.77, 1.39 50.95 −240.65, 334.24  Placental abruption −0.62 −0.86, −0.38 −228.92 −311.94, −145.29 −2.10 −3.67, −0.51 −394.53 −675.78, −110.45  Vaginal bleeding −0.18 −0.30, −0.05 −46.73 −92.54, −0.82 −1.59 −2.70, −0.52 −261.21 −463.43, −62.13  Maternal infections −0.11 −0.21, 0.00 4.89 −32.26, 41.89 −0.49 −1.52, 0.52 −59.10 −252.09, 128.67  Fetal-maternal hemorrhage −0.24 −1.13, 0.71 14.26 −250.99, 285.98 −3.18 −6.34, −0.10 −275.13 −770.42, 207.84  Pregnancy-induced hypertension −0.16 −0.31, −0.01 2.59 −52.34, 57.77 0.22 −1.30, 1.67 −145.82 −435.69, 133.06  Use of β-blockers −0.09 −0.47, 0.29 −39.42 −180.71, 101.34 0.43 −2.52, 3.31 −210.51 −692.39, 281.47 Asthma-related variables  Leukotriene-receptor antagonists 0.03 −0.26, 0.32 −37.40 −141.50, 66.76 −0.73 −3.51, 1.88 −35.11 −490.04, 402.69  No. of SABA doses per week   0 0 Referent 0 Referent 0 Referent 0 Referent   0.01–3 0.04 −0.06, 0.13 −2.30 −36.50, 32.14 0.39 −0.57, 1.37 30.58 −152.98, 215.72   >3 −0.03 −0.14, 0.08 −21.05 −59.77, 17.69 −0.69 −1.74, 0.36 −40.12 −235.79, 154.51  Oral corticosteroids 0.00 −0.13, 0.13 −28.34 −75.45, 18.65 −0.11 −1.36, 1.12 149.64 −80.78, 375.93  Intranasal corticosteroids 0.00 −0.10, 0.11 25.19 −14.07, 64.54 −0.43 −1.59, 0.67 −61.50 −277.99, 147.23  ≥1 ED visit for asthma 0.06 −0.06, 0.18 25.28 −17.13, 68.02 0.24 −1.03, 1.50 −77.18 −312.31, 153.70  LABA −0.07 −0.21, 0.06 11.48 −38.86, 61.59 −0.15 −1.55, 1.22 −52.13 −322.39, 213.72  ≥1 hospitalization for asthma 0.07 −0.29, 0.43 −65.35 −190.57, 61.27 −0.35 −3.11, 2.18 3.94 −438.72, 426.41  Severity of asthma prior to pregnancy   Mild 0 Referent 0 Referent 0 Referent 0 Referent   Moderate 0.09 −0.02, 0.20 −6.30 −47.13, 34.96 0.68 −0.46, 1.81 133.34 −82.30, 348.92   Severe 0.04 −0.13, 0.21 −84.24 −144.89, −23.41 −1.18 −3.01, 0.58 −205.41 −523.13, 104.07 GA residual 156.44 146.77, 166.14 160.71 148.69, 172.67 Characteristica Component 1b Component 2c GA, weeks BW, g GA, weeks BW, g RC 95% CrI RC 95% CrI RC 95% CrI RC 95% CrI Intercept 39.51 39.20, 39.82 3,357.68 3,251.77, 3,463.80 33.24 30.82, 35.44 2,182.29 1,848.04, 2,511.27 ICS exposure −0.01 −0.10, 0.08 1.27 −29.98, 32.15 1.55 0.67, 2.42 200.33 34.10, 365.97 Mother’s and baby’s characteristics  Maternal age, years   <18 0 Referent 0 Referent 0 Referent 0 Referent   18–34 −0.27 −0.57, 0.02 110.16 7.03, 212.89 2.52 0.42, 4.85 396.48 84.92, 712.79   >34 −0.34 −0.65, −0.03 108.30 0.08, 215.71 1.02 −1.22, 3.43 33.24 −303.02, 369.88  Baby’s sex (female vs. male) 0.04 −0.03, 0.11 −143.34 −169.07,−117.62 −0.32 −1.08, 0.43 −142.10 −286.54, 1.54  Receipt of social assistance −0.20 −0.28, −0.13 −98.18 −124.44, −71.92 −0.70 −1.46, 0.07 −169.27 −314.51, −22.65  Rural residency −0.15 −0.24, −0.06 −38.53 −71.32, −5.54 −0.50 −1.45, 0.42 −90.59 −269.02, 81.75 Maternal chronic conditions  Antiphospholipid syndrome −0.53 −1.00, −0.07 −244.61 −410.75, −77.66 −0.54 −3.85, 2.51 −213.44 −694.67, 256.87  Chronic hypertension −0.33 −0.54, −0.13 −56.37 −132.13, 19.98 0.14 −1.94, 2.10 −8.04 −385.33, 353.93  Diabetes mellitus −0.51 −0.71, −0.32 63.96 −5.02, 133.07 −0.12 −2.24, 1.89 257.60 −89.35, 591.89  Uterine defects −0.52 −0.62, −0.42 −31.51 −69.19, 6.01 −0.73 −1.94, 0.44 −164.26 −389.89, 54.18 Pregnancy-related variables  Gestational diabetes −0.38 −0.50, −0.26 54.80 11.82, 98.01 1.22 −0.05, 2.44 207.57 −39.53, 443.16  Eclampsia/preeclampsia −0.52 −0.75, −0.28 −79.58 −162.51, 4.65 −1.87 −3.28, −0.47 −539.05 −804.27, −278.65  Anemia 0.10 0.00, 0.20 107.97 71.27, 144.83 −0.50 −1.53, 0.51 −80.34 −275.09, 109.39  Placental conditions −0.30 −0.50, −0.10 −151.93 −220.79, −82.81 −0.16 −1.77, 1.39 50.95 −240.65, 334.24  Placental abruption −0.62 −0.86, −0.38 −228.92 −311.94, −145.29 −2.10 −3.67, −0.51 −394.53 −675.78, −110.45  Vaginal bleeding −0.18 −0.30, −0.05 −46.73 −92.54, −0.82 −1.59 −2.70, −0.52 −261.21 −463.43, −62.13  Maternal infections −0.11 −0.21, 0.00 4.89 −32.26, 41.89 −0.49 −1.52, 0.52 −59.10 −252.09, 128.67  Fetal-maternal hemorrhage −0.24 −1.13, 0.71 14.26 −250.99, 285.98 −3.18 −6.34, −0.10 −275.13 −770.42, 207.84  Pregnancy-induced hypertension −0.16 −0.31, −0.01 2.59 −52.34, 57.77 0.22 −1.30, 1.67 −145.82 −435.69, 133.06  Use of β-blockers −0.09 −0.47, 0.29 −39.42 −180.71, 101.34 0.43 −2.52, 3.31 −210.51 −692.39, 281.47 Asthma-related variables  Leukotriene-receptor antagonists 0.03 −0.26, 0.32 −37.40 −141.50, 66.76 −0.73 −3.51, 1.88 −35.11 −490.04, 402.69  No. of SABA doses per week   0 0 Referent 0 Referent 0 Referent 0 Referent   0.01–3 0.04 −0.06, 0.13 −2.30 −36.50, 32.14 0.39 −0.57, 1.37 30.58 −152.98, 215.72   >3 −0.03 −0.14, 0.08 −21.05 −59.77, 17.69 −0.69 −1.74, 0.36 −40.12 −235.79, 154.51  Oral corticosteroids 0.00 −0.13, 0.13 −28.34 −75.45, 18.65 −0.11 −1.36, 1.12 149.64 −80.78, 375.93  Intranasal corticosteroids 0.00 −0.10, 0.11 25.19 −14.07, 64.54 −0.43 −1.59, 0.67 −61.50 −277.99, 147.23  ≥1 ED visit for asthma 0.06 −0.06, 0.18 25.28 −17.13, 68.02 0.24 −1.03, 1.50 −77.18 −312.31, 153.70  LABA −0.07 −0.21, 0.06 11.48 −38.86, 61.59 −0.15 −1.55, 1.22 −52.13 −322.39, 213.72  ≥1 hospitalization for asthma 0.07 −0.29, 0.43 −65.35 −190.57, 61.27 −0.35 −3.11, 2.18 3.94 −438.72, 426.41  Severity of asthma prior to pregnancy   Mild 0 Referent 0 Referent 0 Referent 0 Referent   Moderate 0.09 −0.02, 0.20 −6.30 −47.13, 34.96 0.68 −0.46, 1.81 133.34 −82.30, 348.92   Severe 0.04 −0.13, 0.21 −84.24 −144.89, −23.41 −1.18 −3.01, 0.58 −205.41 −523.13, 104.07 GA residual 156.44 146.77, 166.14 160.71 148.69, 172.67 Abbreviations: BW, birth weight; CrI, credible interval; ED, emergency department; GA, gestational age; ICS, inhaled corticosteroids; LABA, long-acting β2-agonists; RC, regression coefficient; SABA, short-acting β2-agonists. a All adjustment variables are binary (yes = 1/no = 0) unless otherwise indicated. b Mixing proportion π1: 0.92, 95% CrI: 0.90, 0.93. c Mixing proportion π2: 0.08, 95% CrI: 0.07, 0.10. Table 5. Associations of Inhaled Corticosteroid Exposure With Gestational Age and Birth Weight Among Asthmatic Pregnant Women in Each Component of an Adjusted 2-Component Adjusted Mixture-of-Bivariate-Regressions Model, Quebec, Canada, 1998–2008 Characteristica Component 1b Component 2c GA, weeks BW, g GA, weeks BW, g RC 95% CrI RC 95% CrI RC 95% CrI RC 95% CrI Intercept 39.51 39.20, 39.82 3,357.68 3,251.77, 3,463.80 33.24 30.82, 35.44 2,182.29 1,848.04, 2,511.27 ICS exposure −0.01 −0.10, 0.08 1.27 −29.98, 32.15 1.55 0.67, 2.42 200.33 34.10, 365.97 Mother’s and baby’s characteristics  Maternal age, years   <18 0 Referent 0 Referent 0 Referent 0 Referent   18–34 −0.27 −0.57, 0.02 110.16 7.03, 212.89 2.52 0.42, 4.85 396.48 84.92, 712.79   >34 −0.34 −0.65, −0.03 108.30 0.08, 215.71 1.02 −1.22, 3.43 33.24 −303.02, 369.88  Baby’s sex (female vs. male) 0.04 −0.03, 0.11 −143.34 −169.07,−117.62 −0.32 −1.08, 0.43 −142.10 −286.54, 1.54  Receipt of social assistance −0.20 −0.28, −0.13 −98.18 −124.44, −71.92 −0.70 −1.46, 0.07 −169.27 −314.51, −22.65  Rural residency −0.15 −0.24, −0.06 −38.53 −71.32, −5.54 −0.50 −1.45, 0.42 −90.59 −269.02, 81.75 Maternal chronic conditions  Antiphospholipid syndrome −0.53 −1.00, −0.07 −244.61 −410.75, −77.66 −0.54 −3.85, 2.51 −213.44 −694.67, 256.87  Chronic hypertension −0.33 −0.54, −0.13 −56.37 −132.13, 19.98 0.14 −1.94, 2.10 −8.04 −385.33, 353.93  Diabetes mellitus −0.51 −0.71, −0.32 63.96 −5.02, 133.07 −0.12 −2.24, 1.89 257.60 −89.35, 591.89  Uterine defects −0.52 −0.62, −0.42 −31.51 −69.19, 6.01 −0.73 −1.94, 0.44 −164.26 −389.89, 54.18 Pregnancy-related variables  Gestational diabetes −0.38 −0.50, −0.26 54.80 11.82, 98.01 1.22 −0.05, 2.44 207.57 −39.53, 443.16  Eclampsia/preeclampsia −0.52 −0.75, −0.28 −79.58 −162.51, 4.65 −1.87 −3.28, −0.47 −539.05 −804.27, −278.65  Anemia 0.10 0.00, 0.20 107.97 71.27, 144.83 −0.50 −1.53, 0.51 −80.34 −275.09, 109.39  Placental conditions −0.30 −0.50, −0.10 −151.93 −220.79, −82.81 −0.16 −1.77, 1.39 50.95 −240.65, 334.24  Placental abruption −0.62 −0.86, −0.38 −228.92 −311.94, −145.29 −2.10 −3.67, −0.51 −394.53 −675.78, −110.45  Vaginal bleeding −0.18 −0.30, −0.05 −46.73 −92.54, −0.82 −1.59 −2.70, −0.52 −261.21 −463.43, −62.13  Maternal infections −0.11 −0.21, 0.00 4.89 −32.26, 41.89 −0.49 −1.52, 0.52 −59.10 −252.09, 128.67  Fetal-maternal hemorrhage −0.24 −1.13, 0.71 14.26 −250.99, 285.98 −3.18 −6.34, −0.10 −275.13 −770.42, 207.84  Pregnancy-induced hypertension −0.16 −0.31, −0.01 2.59 −52.34, 57.77 0.22 −1.30, 1.67 −145.82 −435.69, 133.06  Use of β-blockers −0.09 −0.47, 0.29 −39.42 −180.71, 101.34 0.43 −2.52, 3.31 −210.51 −692.39, 281.47 Asthma-related variables  Leukotriene-receptor antagonists 0.03 −0.26, 0.32 −37.40 −141.50, 66.76 −0.73 −3.51, 1.88 −35.11 −490.04, 402.69  No. of SABA doses per week   0 0 Referent 0 Referent 0 Referent 0 Referent   0.01–3 0.04 −0.06, 0.13 −2.30 −36.50, 32.14 0.39 −0.57, 1.37 30.58 −152.98, 215.72   >3 −0.03 −0.14, 0.08 −21.05 −59.77, 17.69 −0.69 −1.74, 0.36 −40.12 −235.79, 154.51  Oral corticosteroids 0.00 −0.13, 0.13 −28.34 −75.45, 18.65 −0.11 −1.36, 1.12 149.64 −80.78, 375.93  Intranasal corticosteroids 0.00 −0.10, 0.11 25.19 −14.07, 64.54 −0.43 −1.59, 0.67 −61.50 −277.99, 147.23  ≥1 ED visit for asthma 0.06 −0.06, 0.18 25.28 −17.13, 68.02 0.24 −1.03, 1.50 −77.18 −312.31, 153.70  LABA −0.07 −0.21, 0.06 11.48 −38.86, 61.59 −0.15 −1.55, 1.22 −52.13 −322.39, 213.72  ≥1 hospitalization for asthma 0.07 −0.29, 0.43 −65.35 −190.57, 61.27 −0.35 −3.11, 2.18 3.94 −438.72, 426.41  Severity of asthma prior to pregnancy   Mild 0 Referent 0 Referent 0 Referent 0 Referent   Moderate 0.09 −0.02, 0.20 −6.30 −47.13, 34.96 0.68 −0.46, 1.81 133.34 −82.30, 348.92   Severe 0.04 −0.13, 0.21 −84.24 −144.89, −23.41 −1.18 −3.01, 0.58 −205.41 −523.13, 104.07 GA residual 156.44 146.77, 166.14 160.71 148.69, 172.67 Characteristica Component 1b Component 2c GA, weeks BW, g GA, weeks BW, g RC 95% CrI RC 95% CrI RC 95% CrI RC 95% CrI Intercept 39.51 39.20, 39.82 3,357.68 3,251.77, 3,463.80 33.24 30.82, 35.44 2,182.29 1,848.04, 2,511.27 ICS exposure −0.01 −0.10, 0.08 1.27 −29.98, 32.15 1.55 0.67, 2.42 200.33 34.10, 365.97 Mother’s and baby’s characteristics  Maternal age, years   <18 0 Referent 0 Referent 0 Referent 0 Referent   18–34 −0.27 −0.57, 0.02 110.16 7.03, 212.89 2.52 0.42, 4.85 396.48 84.92, 712.79   >34 −0.34 −0.65, −0.03 108.30 0.08, 215.71 1.02 −1.22, 3.43 33.24 −303.02, 369.88  Baby’s sex (female vs. male) 0.04 −0.03, 0.11 −143.34 −169.07,−117.62 −0.32 −1.08, 0.43 −142.10 −286.54, 1.54  Receipt of social assistance −0.20 −0.28, −0.13 −98.18 −124.44, −71.92 −0.70 −1.46, 0.07 −169.27 −314.51, −22.65  Rural residency −0.15 −0.24, −0.06 −38.53 −71.32, −5.54 −0.50 −1.45, 0.42 −90.59 −269.02, 81.75 Maternal chronic conditions  Antiphospholipid syndrome −0.53 −1.00, −0.07 −244.61 −410.75, −77.66 −0.54 −3.85, 2.51 −213.44 −694.67, 256.87  Chronic hypertension −0.33 −0.54, −0.13 −56.37 −132.13, 19.98 0.14 −1.94, 2.10 −8.04 −385.33, 353.93  Diabetes mellitus −0.51 −0.71, −0.32 63.96 −5.02, 133.07 −0.12 −2.24, 1.89 257.60 −89.35, 591.89  Uterine defects −0.52 −0.62, −0.42 −31.51 −69.19, 6.01 −0.73 −1.94, 0.44 −164.26 −389.89, 54.18 Pregnancy-related variables  Gestational diabetes −0.38 −0.50, −0.26 54.80 11.82, 98.01 1.22 −0.05, 2.44 207.57 −39.53, 443.16  Eclampsia/preeclampsia −0.52 −0.75, −0.28 −79.58 −162.51, 4.65 −1.87 −3.28, −0.47 −539.05 −804.27, −278.65  Anemia 0.10 0.00, 0.20 107.97 71.27, 144.83 −0.50 −1.53, 0.51 −80.34 −275.09, 109.39  Placental conditions −0.30 −0.50, −0.10 −151.93 −220.79, −82.81 −0.16 −1.77, 1.39 50.95 −240.65, 334.24  Placental abruption −0.62 −0.86, −0.38 −228.92 −311.94, −145.29 −2.10 −3.67, −0.51 −394.53 −675.78, −110.45  Vaginal bleeding −0.18 −0.30, −0.05 −46.73 −92.54, −0.82 −1.59 −2.70, −0.52 −261.21 −463.43, −62.13  Maternal infections −0.11 −0.21, 0.00 4.89 −32.26, 41.89 −0.49 −1.52, 0.52 −59.10 −252.09, 128.67  Fetal-maternal hemorrhage −0.24 −1.13, 0.71 14.26 −250.99, 285.98 −3.18 −6.34, −0.10 −275.13 −770.42, 207.84  Pregnancy-induced hypertension −0.16 −0.31, −0.01 2.59 −52.34, 57.77 0.22 −1.30, 1.67 −145.82 −435.69, 133.06  Use of β-blockers −0.09 −0.47, 0.29 −39.42 −180.71, 101.34 0.43 −2.52, 3.31 −210.51 −692.39, 281.47 Asthma-related variables  Leukotriene-receptor antagonists 0.03 −0.26, 0.32 −37.40 −141.50, 66.76 −0.73 −3.51, 1.88 −35.11 −490.04, 402.69  No. of SABA doses per week   0 0 Referent 0 Referent 0 Referent 0 Referent   0.01–3 0.04 −0.06, 0.13 −2.30 −36.50, 32.14 0.39 −0.57, 1.37 30.58 −152.98, 215.72   >3 −0.03 −0.14, 0.08 −21.05 −59.77, 17.69 −0.69 −1.74, 0.36 −40.12 −235.79, 154.51  Oral corticosteroids 0.00 −0.13, 0.13 −28.34 −75.45, 18.65 −0.11 −1.36, 1.12 149.64 −80.78, 375.93  Intranasal corticosteroids 0.00 −0.10, 0.11 25.19 −14.07, 64.54 −0.43 −1.59, 0.67 −61.50 −277.99, 147.23  ≥1 ED visit for asthma 0.06 −0.06, 0.18 25.28 −17.13, 68.02 0.24 −1.03, 1.50 −77.18 −312.31, 153.70  LABA −0.07 −0.21, 0.06 11.48 −38.86, 61.59 −0.15 −1.55, 1.22 −52.13 −322.39, 213.72  ≥1 hospitalization for asthma 0.07 −0.29, 0.43 −65.35 −190.57, 61.27 −0.35 −3.11, 2.18 3.94 −438.72, 426.41  Severity of asthma prior to pregnancy   Mild 0 Referent 0 Referent 0 Referent 0 Referent   Moderate 0.09 −0.02, 0.20 −6.30 −47.13, 34.96 0.68 −0.46, 1.81 133.34 −82.30, 348.92   Severe 0.04 −0.13, 0.21 −84.24 −144.89, −23.41 −1.18 −3.01, 0.58 −205.41 −523.13, 104.07 GA residual 156.44 146.77, 166.14 160.71 148.69, 172.67 Abbreviations: BW, birth weight; CrI, credible interval; ED, emergency department; GA, gestational age; ICS, inhaled corticosteroids; LABA, long-acting β2-agonists; RC, regression coefficient; SABA, short-acting β2-agonists. a All adjustment variables are binary (yes = 1/no = 0) unless otherwise indicated. b Mixing proportion π1: 0.92, 95% CrI: 0.90, 0.93. c Mixing proportion π2: 0.08, 95% CrI: 0.07, 0.10. Figure 1 presents the clustering of the 6,197 pregnancies on the GA-BW plane. All but 1 of the pregnancies allocated to the second component were preterm, while the vast majority (82%) corresponded to LBW infants. In parallel, most PTB pregnancies allocated to the first component did not correspond to LBW infants. At the 5% level, a high probability of membership in the second component was found to be negatively associated with exposure to ICS and gestational diabetes and positively associated with receipt of social assistance, fetal-maternal hemorrhage, placental abruption, vaginal bleeding, anemia, eclampsia/preeclampsia, and use of more than 3 doses of short-acting β2-agonists per week (see Table 6). Figure 1. View largeDownload slide Allocation of pregnancies in the first (gray dots) and second (black dots) components of a 2-component mixture-of-bivariate-regressions model of inhaled corticosteroid exposure among asthmatic pregnant women, Quebec, Canada, 1998–2008. Figure 1. View largeDownload slide Allocation of pregnancies in the first (gray dots) and second (black dots) components of a 2-component mixture-of-bivariate-regressions model of inhaled corticosteroid exposure among asthmatic pregnant women, Quebec, Canada, 1998–2008. Table 6. Associations Between Selected Characteristics of Asthmatic Pregnant Women and Membership in the Second Component of a Mixture-of-Bivariate-Regressions Model of Inhaled Corticosteroid Exposure (Multiple Logistic Regression Analysis), Quebec, Canada, 1998–2008 Characteristica OR 95% CI P Value ICS exposure 0.62 0.47, 0.82 <0.001 Mother’s and baby’s characteristics  Maternal age, years   <18 1.00 Referent   18–34 0.49 0.23, 1.04 0.06   >34 0.75 0.34, 1.66 0.48  Baby’s sex (female vs. male) 0.94 0.75, 1.18 0.59  Receipt of social assistance 1.43 1.13, 1.82 <0.01  Rural residency 1.18 0.89, 1.58 0.25 Maternal chronic conditions  Antiphospholipid syndrome 1.57 0.44, 5.57 0.49  Chronic hypertension 0.75 0.36, 1.55 0.44  Diabetes mellitus 0.69 0.36, 1.31 0.25  Uterine defects 0.81 0.57, 1.14 0.23 Pregnancy-related variables  Gestational diabetes 0.60 0.38, 0.94 0.02  Eclampsia/preeclampsia 3.70 2.29, 5.98 <0.0001  Anemia 1.35 1.01, 1.82 0.04  Placental conditions 1.24 0.77, 2.01 0.37  Placental abruption 2.03 1.27, 3.26 <0.01  Vaginal bleeding 1.91 1.38, 2.60 <0.001  Maternal infections 1.16 0.86, 1.57 0.34  Fetal-maternal hemorrhage 5.18 1.52, 17.6 <0.01  Pregnancy-induced hypertension 0.74 0.45, 1.23 0.24  Use of β-blockers 1.60 0.54, 4.71 0.40 Asthma-related variables  Leukotriene-receptor antagonists 0.82 0.31, 2.14 0.68  No. of SABA doses per week   0 1.00 Referent   0.01–3 1.06 0.78, 1.45 0.71   >3 1.56 1.11, 2.18 <0.01  Oral corticosteroids 1.35 0.91, 2.00 0.14  Intranasal corticosteroids 1.00 0.71, 1.42 0.99  ≥1 ED visit for asthma 0.77 0.52, 1.15 0.20  LABA 1.27 0.83, 1.96 0.28  ≥1 hospitalization for asthma 1.49 0.60, 3.72 0.39  Severity of asthma prior to pregnancy   Mild 1.00 Referent   Moderate 0.95 0.67, 1.37 0.80   Severe 0.83 0.49, 1.41 0.50 Characteristica OR 95% CI P Value ICS exposure 0.62 0.47, 0.82 <0.001 Mother’s and baby’s characteristics  Maternal age, years   <18 1.00 Referent   18–34 0.49 0.23, 1.04 0.06   >34 0.75 0.34, 1.66 0.48  Baby’s sex (female vs. male) 0.94 0.75, 1.18 0.59  Receipt of social assistance 1.43 1.13, 1.82 <0.01  Rural residency 1.18 0.89, 1.58 0.25 Maternal chronic conditions  Antiphospholipid syndrome 1.57 0.44, 5.57 0.49  Chronic hypertension 0.75 0.36, 1.55 0.44  Diabetes mellitus 0.69 0.36, 1.31 0.25  Uterine defects 0.81 0.57, 1.14 0.23 Pregnancy-related variables  Gestational diabetes 0.60 0.38, 0.94 0.02  Eclampsia/preeclampsia 3.70 2.29, 5.98 <0.0001  Anemia 1.35 1.01, 1.82 0.04  Placental conditions 1.24 0.77, 2.01 0.37  Placental abruption 2.03 1.27, 3.26 <0.01  Vaginal bleeding 1.91 1.38, 2.60 <0.001  Maternal infections 1.16 0.86, 1.57 0.34  Fetal-maternal hemorrhage 5.18 1.52, 17.6 <0.01  Pregnancy-induced hypertension 0.74 0.45, 1.23 0.24  Use of β-blockers 1.60 0.54, 4.71 0.40 Asthma-related variables  Leukotriene-receptor antagonists 0.82 0.31, 2.14 0.68  No. of SABA doses per week   0 1.00 Referent   0.01–3 1.06 0.78, 1.45 0.71   >3 1.56 1.11, 2.18 <0.01  Oral corticosteroids 1.35 0.91, 2.00 0.14  Intranasal corticosteroids 1.00 0.71, 1.42 0.99  ≥1 ED visit for asthma 0.77 0.52, 1.15 0.20  LABA 1.27 0.83, 1.96 0.28  ≥1 hospitalization for asthma 1.49 0.60, 3.72 0.39  Severity of asthma prior to pregnancy   Mild 1.00 Referent   Moderate 0.95 0.67, 1.37 0.80   Severe 0.83 0.49, 1.41 0.50 Abbreviations: CI, confidence interval; ED, emergency department; ICS, inhaled corticosteroids; LABA, long-acting β2-agonists; OR, odds ratio; SABA, short-acting β2-agonists. a All variables are binary (yes = 1/no = 0) unless otherwise indicated. View Large Table 6. Associations Between Selected Characteristics of Asthmatic Pregnant Women and Membership in the Second Component of a Mixture-of-Bivariate-Regressions Model of Inhaled Corticosteroid Exposure (Multiple Logistic Regression Analysis), Quebec, Canada, 1998–2008 Characteristica OR 95% CI P Value ICS exposure 0.62 0.47, 0.82 <0.001 Mother’s and baby’s characteristics  Maternal age, years   <18 1.00 Referent   18–34 0.49 0.23, 1.04 0.06   >34 0.75 0.34, 1.66 0.48  Baby’s sex (female vs. male) 0.94 0.75, 1.18 0.59  Receipt of social assistance 1.43 1.13, 1.82 <0.01  Rural residency 1.18 0.89, 1.58 0.25 Maternal chronic conditions  Antiphospholipid syndrome 1.57 0.44, 5.57 0.49  Chronic hypertension 0.75 0.36, 1.55 0.44  Diabetes mellitus 0.69 0.36, 1.31 0.25  Uterine defects 0.81 0.57, 1.14 0.23 Pregnancy-related variables  Gestational diabetes 0.60 0.38, 0.94 0.02  Eclampsia/preeclampsia 3.70 2.29, 5.98 <0.0001  Anemia 1.35 1.01, 1.82 0.04  Placental conditions 1.24 0.77, 2.01 0.37  Placental abruption 2.03 1.27, 3.26 <0.01  Vaginal bleeding 1.91 1.38, 2.60 <0.001  Maternal infections 1.16 0.86, 1.57 0.34  Fetal-maternal hemorrhage 5.18 1.52, 17.6 <0.01  Pregnancy-induced hypertension 0.74 0.45, 1.23 0.24  Use of β-blockers 1.60 0.54, 4.71 0.40 Asthma-related variables  Leukotriene-receptor antagonists 0.82 0.31, 2.14 0.68  No. of SABA doses per week   0 1.00 Referent   0.01–3 1.06 0.78, 1.45 0.71   >3 1.56 1.11, 2.18 <0.01  Oral corticosteroids 1.35 0.91, 2.00 0.14  Intranasal corticosteroids 1.00 0.71, 1.42 0.99  ≥1 ED visit for asthma 0.77 0.52, 1.15 0.20  LABA 1.27 0.83, 1.96 0.28  ≥1 hospitalization for asthma 1.49 0.60, 3.72 0.39  Severity of asthma prior to pregnancy   Mild 1.00 Referent   Moderate 0.95 0.67, 1.37 0.80   Severe 0.83 0.49, 1.41 0.50 Characteristica OR 95% CI P Value ICS exposure 0.62 0.47, 0.82 <0.001 Mother’s and baby’s characteristics  Maternal age, years   <18 1.00 Referent   18–34 0.49 0.23, 1.04 0.06   >34 0.75 0.34, 1.66 0.48  Baby’s sex (female vs. male) 0.94 0.75, 1.18 0.59  Receipt of social assistance 1.43 1.13, 1.82 <0.01  Rural residency 1.18 0.89, 1.58 0.25 Maternal chronic conditions  Antiphospholipid syndrome 1.57 0.44, 5.57 0.49  Chronic hypertension 0.75 0.36, 1.55 0.44  Diabetes mellitus 0.69 0.36, 1.31 0.25  Uterine defects 0.81 0.57, 1.14 0.23 Pregnancy-related variables  Gestational diabetes 0.60 0.38, 0.94 0.02  Eclampsia/preeclampsia 3.70 2.29, 5.98 <0.0001  Anemia 1.35 1.01, 1.82 0.04  Placental conditions 1.24 0.77, 2.01 0.37  Placental abruption 2.03 1.27, 3.26 <0.01  Vaginal bleeding 1.91 1.38, 2.60 <0.001  Maternal infections 1.16 0.86, 1.57 0.34  Fetal-maternal hemorrhage 5.18 1.52, 17.6 <0.01  Pregnancy-induced hypertension 0.74 0.45, 1.23 0.24  Use of β-blockers 1.60 0.54, 4.71 0.40 Asthma-related variables  Leukotriene-receptor antagonists 0.82 0.31, 2.14 0.68  No. of SABA doses per week   0 1.00 Referent   0.01–3 1.06 0.78, 1.45 0.71   >3 1.56 1.11, 2.18 <0.01  Oral corticosteroids 1.35 0.91, 2.00 0.14  Intranasal corticosteroids 1.00 0.71, 1.42 0.99  ≥1 ED visit for asthma 0.77 0.52, 1.15 0.20  LABA 1.27 0.83, 1.96 0.28  ≥1 hospitalization for asthma 1.49 0.60, 3.72 0.39  Severity of asthma prior to pregnancy   Mild 1.00 Referent   Moderate 0.95 0.67, 1.37 0.80   Severe 0.83 0.49, 1.41 0.50 Abbreviations: CI, confidence interval; ED, emergency department; ICS, inhaled corticosteroids; LABA, long-acting β2-agonists; OR, odds ratio; SABA, short-acting β2-agonists. a All variables are binary (yes = 1/no = 0) unless otherwise indicated. View Large Lastly, when reanalyzing our cohort using LBW as opposed to BW in a standard adjusted logistic regression model, we obtained an odds ratio associated with ICS of 0.79 (95% CI: 0.63, 1.00). For the outcomes PTB and very PTB, the adjusted ICS odds ratios were 0.84 (95% CI: 0.68, 1.04) and 0.28 (95% CI: 0.15, 0.51), respectively. The results for the 2-component mixture secondary analysis using GA and BW z scores are presented in Web Appendix 3 (Web Table 1). The relative sizes of the 2 subpopulations were similar to those found in the corresponding analysis using GA and BW. The association between ICS and GA also remained similar in both components; however, ICS was not found to be associated with BW z score in either component. In the smallest component, only diabetes mellitus and eclampsia/preeclampsia were found to be positively and negatively associated with BW z score, respectively. Unlike in the main analyses, the within-component residual association between GA and BW z score (β∗k) did not significantly differ from zero. DISCUSSION Adjusting for a large set of covariates, including many variables directly related to asthma severity and control, the credible intervals for exposure to ICS associated with GA and BW in the first, major component of our K = 2 mixture model were found to be centered approximately at zero and relatively narrow. These results, which concern a large proportion of pregnancies, are consistent with those of many other studies that have failed to associate ICS exposure during pregnancy with adverse perinatal outcomes, including GA- and BW-related variables (35). Interestingly, our 2-component mixture model revealed that there is a small group of asthmatic women, representing about 8% of the population, in whom ICS is positively associated with GA and BW, with a magnitude which could be qualified as clinically significant. In this subpopulation, pregnancies exposed to ICS were, on average, approximately 11 days longer than those that were not (95% CrI: 4.70, 16.0). This increase in GA reverberated on BW, as was seen from the estimated average increase of about 200 g in BW when pregnancies were exposed to ICS (95% CrI: 34.10, 366.0). However, our bivariate model makes it possible to note that this mean difference of about 200 g is somewhat smaller than what would be expected from an increase of 1.5 weeks in GA, since, in this component, the residual mean change for the effect of 1 additional week in GA on BW was estimated to be 160.7 g (95% CrI: 148.7, 172.7). This observation is consistent with the BW z score analysis, where the point estimate for the association between ICS and BW z score in the second component was found to be small and negative (estimate = −0.07, 95% CrI: −0.35, 0.21). Differences in conclusions between the standard linear regressions (or K = 1 mixture model) for GA and BW and the standard logistic regressions which instead parameterized PTB and LBW may suggest that the effect of exposure to ICS takes place at low quantiles of the joint GA/BW distribution. This is more strongly supported by the results from the main 2-component mixture analysis, for which a positive association between ICS and GA/BW was seen in the most problematic asthmatic pregnancies. Indeed, nearly all hard-clustered pregnancies in the smaller component were preterm, which would explain why inconclusive results were obtained in our standard adjusted logistic regression analysis using PTB as the outcome (odds ratio = 0.84, 95% CI: 0.68, 1.04). The significant results obtained for very PTB (odds ratio = 0.28, 95% CI: 0.15, 0.51) could also be anticipated from the mixture model’s results, since 32 weeks is a central point for GA in the second component (recall Figure 1). To our knowledge, only 1 previous study has examined the association between maternal exposure to ICS and very PTB using ICS-unexposed asthmatic women as the reference group: Using very PTB as the outcome variable, Schatz et al. (65) found a nonsignificant unadjusted odds ratio for ICS of 0.94 (P > 0.05) based on a sample of 2,123 asthmatic women. Our study was a reanalysis of Cossette et al.’s cohort of pregnancies among women with asthma (36); as such, the same limitations regarding the data pertain to the present work. Generalization of the results to a general obstetrical population is questionable, since the Régie de l’assurance maladie du Québec prescription insurance plan is a compulsory insurance plan for Quebec residents who are not eligible for a private one. Overall, people enrolled in this plan have a less favorable socioeconomic status, which may influence ICS adherence. Indeed, recall that exposure to ICS was based on dispensed prescriptions, which might not reflect actual use. Moreover, our list of measured covariates omitted important potential confounders or risk factors, such as maternal cigarette smoking during pregnancy, prepregnancy body mass index, ethnicity, and parity. We also lacked genetic information, which could be related to global responsiveness to ICS and the occurrence of adverse perinatal outcomes such as PTB and LBW. However, additional analyses which used the number of pregnancies in the follow-up period to define a proxy for parity yielded essentially the same results (not reported). Moreover, Cossette et al. found, in sensitivity analyses, that their estimates of the effect of ICS on the perinatal outcomes LBW, PTB, and small size for GA were relatively robust to unmeasured smoking status (36). A similar conclusion was reached for maternal obesity status (see Web Appendix 4, including Web Table 2, for results and discussion). Because one can define the mixture-of-regression model to represent homogenous subpopulations indexed by common values for unmeasured binary or categorical covariates, it is, however, possible that these missing covariates were partially accounted for by our model. Finally, a limitation specific to this study is the possibility that the ever/never exposure classification may have attenuated true risks associated with the effect of ICS use on GA/BW, or even created a protective effect. This bias would arise from increased opportunity for exposure among women remaining pregnant longer and has been shown to explain an apparent protective effect of maternal influenza immunization in pregnancy on risk of PTB (66). Many of our study’s strengths directly follow from those of Cossette et al. (36)—notably the large number of pregnancies, medication-exposure assessment from data prospectively collected independently of the outcomes and recorded in pharmacy records, the inclusion of multiple potential confounders, and the use of previously validated algorithms to measure ICS and short-acting β2-agonist exposure and the control and severity of asthma. While the size of our cohort was deemed sufficiently large to apply a mixture-of-regressions model, a larger number of pregnancies would have allowed a more thorough assessment of the heterogeneity of the associations between exposure to ICS and GA and BW. Indeed, because the regression coefficients’ CrIs were generally wide for the smallest subpopulation, we chose not to study the effect of ICS on GA and BW as a function of dosage. Similarly, we did not consider adding a third mixture component, which, for a larger cohort, may have revealed more specific insights on the role of ICS in the most problematic asthmatic pregnancies. Indeed, because it is well-known that prematurity is a multifactorial syndrome (67, 68), it is likely that pregnancies in the smallest component still exhibit significant individual heterogeneity with regard to the ICS effect. As it stands, we nonetheless believe that our mixture-of-regressions model enabled us to carry out a unique and highly valuable exploration of our cohort data, which will likely serve to refine or generate new hypotheses as to how treatment of asthma with ICS during pregnancy affects fetal growth. In conclusion, we have introduced a finite mixture-of-bivariate-regressions model to explore the heterogeneity of ICS exposure effects on BW (or BW z score) and GA in a large cohort of asthmatic pregnant women, but this model can certainly be used for risk assessment for other exposures. When using this model with BW z score instead of BW, it is expected that the coefficients β∗k(k=1,2,…,K) will be zero if the GA-dependent normalization scale for BW is appropriate. The proposed mixture is thus overly flexible for exploring effect heterogeneity with a BW z score outcome, since it models the theoretical null dependence between GA and BW z score. While this could be viewed as undesirable, one may find it useful to be able to check for practical deviation of β∗k from zero, indicating a residual correlation between GA and BW z score and thus suggesting an external reference population inadequate for the study population. With a BW z score outcome, a straightforward modification of the model to strictly enforce the β∗k coefficient values to zero could alternatively be done, thereby reducing by K the number of parameters to estimate. ACKNOWLEDGMENTS Author affiliations: Département de mathématiques, Faculté des sciences, Université du Québec à Montréal, Montréal, Québec, Canada (Mariia Samoilenko, Geneviève Lefebvre); Faculté de pharmacie, Université de Montréal, Montréal, Québec, Canada (Lucie Blais, Geneviève Lefebvre); Centre de recherche, Hôpital du Sacré-Coeur de Montréal, Montréal, Québec, Canada (Lucie Blais); Centre de recherche Clinique Étienne-Le Bel, Centre hospitalier universitaire de Sherbrooke, Sherbrooke, Québec, Canada (Lucie Blais); and Département d’obstétrique-gynécologie, Centre hospitalier universitaire de Sainte-Justine, Université de Montréal, Montréal, Québec, Canada (Isabelle Boucoiran). This work was supported by grants from the Fonds de recherche Québec–Santé (FRQ-S) and the Natural Sciences and Engineering Research Council of Canada. G.L. is an FRQ-S Research Scholar. G.L. thanks Calcul Québec (Montréal, Québec, Canada) and Dr. Daniel Stubbs (Calcul Québec) for programming support. Conflict of interest: none declared. Abbreviations BW birth weight CI confidence interval CrI credible interval GA gestational age ICS inhaled corticosteroids LABA long-acting β2-agonists LBW low birth weight PTB preterm birth REFERENCES 1 Ober C , Yao TC . The genetics of asthma and allergic disease: a 21st century perspective . Immunol Rev . 2011 ; 242 ( 1 ): 10 – 30 . Google Scholar Crossref Search ADS PubMed 2 Noval Rivas M , Crother TR , Arditi M . The microbiome in asthma . Curr Opin Pediatr . 2016 ; 28 ( 6 ): 764 – 771 . Google Scholar Crossref Search ADS PubMed 3 Borish L , Culp JA . Asthma: a syndrome composed of heterogeneous diseases . Ann Allergy Asthma Immunol . 2008 ; 101 ( 1 ): 1 – 8 ; quiz 8–11, 50. Google Scholar Crossref Search ADS PubMed 4 Lougheed MD , Lemiere C , Ducharme FM , et al. . Canadian Thoracic Society 2012 guideline update: diagnosis and management of asthma in preschoolers, children and adults . Can Respir J . 2012 ; 19 ( 2 ): 127 – 164 . Google Scholar Crossref Search ADS PubMed 5 Global Initiative for Asthma . Global Strategy for Asthma Management and Prevention. Updated 2016. Bethesda, MD : National Heart, Lung, and Blood Institute ; 2016 . https://ginasthma.org/wp-content/uploads/2016/04/wms-GINA-2016-main-report-final.pdf. Accessed September 7, 2016. 6 National Asthma Education and Prevention Program . Expert Panel Report 3 (EPR-3): guidelines for the diagnosis and management of asthma—summary report 2007 . J Allergy Clin Immunol . 2007 ; 120 ( 5 suppl ): S94 – S138 . PubMed 7 Healthcare Improvement Scotland; Scottish Intercollegiate Guidelines Network (SIGN); British Thoracic Society . British Guideline on the Management of Asthma: a National Clinical Guideline. (SIGN 153). Edinburgh, United Kingdom : Healthcare Improvement Scotland ; 2016 . https://www.brit-thoracic.org.uk/document-library/clinical-information/asthma/btssign-asthma-guideline-2016/. Accessed October 11, 2016. 8 Lima JJ . Treatment heterogeneity in asthma: genetics of response to leukotriene modifiers . Mol Diagn Ther . 2007 ; 11 ( 2 ): 97 – 104 . Google Scholar Crossref Search ADS PubMed 9 Dahlin A , Tantisira KG . Integrative systems biology approaches in asthma pharmacogenomics . Pharmacogenomics . 2012 ; 13 ( 12 ): 1387 – 1404 . Google Scholar Crossref Search ADS PubMed 10 Farzan N , Vijverberg SJ , Arets HG , et al. . Pharmacogenomics of inhaled corticosteroids and leukotriene modifiers: a systematic review . Clin Exp Allergy . 2017 ; 47 ( 2 ): 271 – 293 . Google Scholar Crossref Search ADS PubMed 11 Simpson KR , Creehan PA . AWHONN’s Perinatal Nursing . 4th ed . Philadelphia, PA : Wolters Kluwer/Lippincott Williams & Wilkins Health ; 2013 . 12 Barnes PJ . Glucocorticosteroids: current and future directions . Br J Pharmacol . 2011 ; 163 ( 1 ): 29 – 43 . Google Scholar Crossref Search ADS PubMed 13 Barnes PJ . Corticosteroid resistance in patients with asthma and chronic obstructive pulmonary disease . J Allergy Clin Immunol . 2013 ; 131 ( 3 ): 636 – 645 . Google Scholar Crossref Search ADS PubMed 14 Matsumura Y . Heterogeneity of glucocorticoid resistance in patients with bronchial asthma . Int J Biomed Sci . 2010 ; 6 ( 3 ): 158 – 166 . Google Scholar PubMed 15 Lazarus SC , Chinchilli VM , Rollings NJ , et al. . Smoking affects response to inhaled corticosteroids or leukotriene receptor antagonists in asthma . Am J Respir Crit Care Med . 2007 ; 175 ( 8 ): 783 – 790 . Google Scholar Crossref Search ADS PubMed 16 Bourdin A , Halimi L , Vachier I , et al. . Adherence in severe asthma . Clin Exp Allergy . 2012 ; 42 ( 11 ): 1566 – 1574 . Google Scholar Crossref Search ADS PubMed 17 Silverman ES , Liggett SB , Gelfand EW , et al. . The pharmacogenetics of asthma: a candidate gene approach . Pharmacogenomics J . 2001 ; 1 : 27 – 37 . Google Scholar Crossref Search ADS PubMed 18 Davis JS , Weiss ST , Tantisira KG . Asthma pharmacogenomics: 2015 update . Curr Allergy Asthma Rep . 2015 ; 15 ( 7 ): 42 . Google Scholar Crossref Search ADS PubMed 19 Murphy VE . Managing asthma in pregnancy . Breathe (Sheff) . 2015 ; 11 ( 4 ): 258 – 267 . Google Scholar Crossref Search ADS PubMed 20 Katz O , Sheiner E . Asthma and pregnancy: a review of two decades . Expert Rev Respir Med . 2008 ; 2 ( 1 ): 97 – 107 . Google Scholar Crossref Search ADS PubMed 21 Murphy VE , Clifton VL , Gibson PG . Asthma exacerbations during pregnancy: incidence and association with adverse pregnancy outcomes . Thorax . 2006 ; 61 ( 2 ): 169 – 176 . Google Scholar Crossref Search ADS PubMed 22 Gregersen TL , Ulrik CS . Safety of bronchodilators and corticosteroids for asthma during pregnancy: what we know and what we need to do better . J Asthma Allergy . 2013 ; 6 : 117 – 125 . Google Scholar PubMed 23 Namazy JA , Murphy VE , Powell H , et al. . Effects of asthma severity, exacerbations and oral corticosteroids on perinatal outcomes . Eur Respir J . 2013 ; 41 ( 5 ): 1082 – 1090 . Google Scholar Crossref Search ADS PubMed 24 Grzeskowiak LE , Clifton VL . Asthma management during pregnancy: how long before we can all breathe a little easier? J Asthma . 2015 ; 52 ( 10 ): 1020 – 1022 . Google Scholar Crossref Search ADS PubMed 25 National Heart, Lung, and Blood Institute; National Asthma Education and Prevention Program Asthma and Pregnancy Working Group . NAEPP Expert Panel Report. Managing asthma during pregnancy: recommendations for pharmacologic treatment—2004 update . J Allergy Clin Immunol . 2005 ; 115 ( 1 ): 34 – 46 . Crossref Search ADS PubMed 26 Schatz M , Zeiger RS , Hoffman CP , et al. . Perinatal outcomes in the pregnancies of asthmatic women: a prospective controlled analysis . Am J Respir Crit Care Med . 1995 ; 151 ( 4 ): 1170 – 1174 . Google Scholar PubMed 27 Schatz M , Zeiger RS , Harden KM , et al. . The safety of inhaled beta-agonist bronchodilators during pregnancy . J Allergy Clin Immunol . 1988 ; 82 ( 4 ): 686 – 695 . Google Scholar Crossref Search ADS PubMed 28 Perlow JH , Montgomery D , Morgan MA , et al. . Severity of asthma and perinatal outcome . Am J Obstet Gynecol . 1992 ; 167 ( 4 ): 963 – 967 . Google Scholar Crossref Search ADS PubMed 29 Schatz M . The safety of asthma and allergy medications during pregnancy . Can J Allergy Clin Immunol . 1998 ; 3 ( 5 ): 242 – 254 . 30 Schatz M , Zeiger RS , Harden K , et al. . The safety of asthma and allergy medications during pregnancy . J Allergy Clin Immunol . 1997 ; 100 ( 3 ): 301 – 306 . Google Scholar Crossref Search ADS PubMed 31 Dombrowski MP , Schatz M , Wise R , et al. . Randomized trial of inhaled beclomethasone dipropionate versus theophylline for moderate asthma during pregnancy . Am J Obstet Gynecol . 2004 ; 190 ( 3 ): 737 – 744 . Google Scholar Crossref Search ADS PubMed 32 Murphy VE , Namazy JA , Powell H , et al. . A meta-analysis of adverse perinatal outcomes in women with asthma . BJOG . 2011 ; 118 ( 11 ): 1314 – 1323 . Google Scholar Crossref Search ADS PubMed 33 Rejnö G , Lundholm C , Gong T , et al. . Asthma during pregnancy in a population-based study—pregnancy complications and adverse perinatal outcomes . PLoS One . 2014 ; 9 ( 8 ): e104755 . Google Scholar Crossref Search ADS PubMed 34 Namazy JA , Chambers C , Schatz M . Safety of therapeutic options for treating asthma in pregnancy . Expert Opin Drug Saf . 2014 ; 13 ( 12 ): 1613 – 1621 . Google Scholar Crossref Search ADS PubMed 35 Namazy JA , Schatz M . Pharmacotherapy options to treat asthma during pregnancy . Expert Opin Pharmacother . 2015 ; 16 ( 12 ): 1783 – 1791 . Google Scholar Crossref Search ADS PubMed 36 Cossette B , Forget A , Beauchesne MF , et al. . Impact of maternal use of asthma-controller therapy on perinatal outcomes . Thorax . 2013 ; 68 ( 8 ): 724 – 730 . Google Scholar Crossref Search ADS PubMed 37 Samoilenko M , Blais L , Cossette B , et al. . Assessing the dose-response relationship between maternal use of inhaled corticosteroids therapy and birth weight: a generalized propensity score approach . Obs Stud . 2016 ; 2016 ( Article 8 ): 90 – 118 . 38 Murphy VE , Gibson PG , Giles WB , et al. . Maternal asthma is associated with reduced female fetal growth . Am J Respir Crit Care Med . 2003 ; 168 ( 11 ): 1317 – 1323 . Google Scholar Crossref Search ADS PubMed 39 Hodyl NA , Stark MJ , Osei-Kumah A , et al. . Fetal glucocorticoid-regulated pathways are not affected by inhaled corticosteroid use for asthma during pregnancy . Am J Respir Crit Care Med . 2011 ; 183 ( 6 ): 716 – 722 . Google Scholar Crossref Search ADS PubMed 40 Huffaker MF , Phipatanakul W . Pediatric asthma: guidelines-based care, omalizumab, and other potential biologic agents . Immunol Allergy Clin North Am . 2015 ; 35 ( 1 ): 129 – 144 . Google Scholar Crossref Search ADS PubMed 41 Langmack EL , Martin RJ . Heterogeneity of response to asthma controller therapy: clinical implications . Curr Opin Pulm Med . 2010 ; 16 ( 1 ): 13 – 18 . Google Scholar Crossref Search ADS PubMed 42 Drazen JM . A step toward personalized asthma treatment . N Engl J Med . 2011 ; 365 ( 13 ): 1245 – 1246 . Google Scholar Crossref Search ADS PubMed 43 Blake K , Raissy H . Pharmacogenomic testing in the asthma clinic: will inhaled corticosteroids lead the way? Pediatr Allergy Immunol Pulmonol . 2012 ; 25 ( 1 ): 44 – 47 . Google Scholar Crossref Search ADS 44 Tantisira KG , Lasky-Su J , Harada M , et al. . Genomewide association between GLCCI1 and response to glucocorticoid therapy in asthma . N Engl J Med . 2011 ; 365 ( 13 ): 1173 – 1183 . Google Scholar Crossref Search ADS PubMed 45 McLachlan GJ , Basford KE . Mixture Models: Inference and Applications to Clustering . New York, NY : Marcel Dekker, Inc. ; 1988 . 46 Melnykov V , Maitra R . Finite mixture models and model-based clustering . Stat Surv . 2010 ; 4 : 80 – 116 . Google Scholar Crossref Search ADS 47 Andrews JL , McNicholas PD . Model-based clustering, classification, and discriminant analysis via mixtures of multivariate t-distributions . Stat Comput . 2012 ; 22 ( 5 ): 1021 – 1029 . Google Scholar Crossref Search ADS 48 Frühwirth-Schnatter S . Finite Mixture and Markov Switching Models . New York, NY : Springer Science+Business Media LLC ; 2006 . 49 Charnigo R , Chesnut LW , Lobianco T , et al. . Thinking outside the curve, part I: modeling birthweight distribution . BMC Pregnancy Childbirth . 2010 ; 10 : 37 . Google Scholar Crossref Search ADS PubMed 50 Gage TB , Therriault G . Variability of birth-weight distributions by sex and ethnicity: analysis using mixture models . Hum Biol . 1998 ; 70 ( 3 ): 517 – 534 . Google Scholar PubMed 51 Deng W , Charnigo R , Dai H , et al. . Characterizing components in a mixture model for birthweight distribution . J Biom Biostat . 2011 ; 2 ( 4 ): 118 . Google Scholar Crossref Search ADS 52 Tentoni S , Astolfi P , De Pasquale A , et al. . Birthweight by gestational age in preterm babies according to a Gaussian mixture model . BJOG . 2004 ; 111 ( 1 ): 31 – 37 . Google Scholar Crossref Search ADS PubMed 53 Oja H , Koiranen M , Rantakallio P . Fitting mixture models to birth weight data: a case study . Biometrics . 1991 ; 47 ( 3 ): 883 – 897 . Google Scholar Crossref Search ADS PubMed 54 Platt RW , Abrahamowicz M , Kramer MS , et al. . Detecting and eliminating erroneous gestational ages: a normal mixture model . Stat Med . 2001 ; 20 ( 23 ): 3491 – 3503 . Google Scholar Crossref Search ADS PubMed 55 Whitmore GA , Su Y . Modeling low birth weights using threshold regression: results for US birth data . Lifetime Data Anal . 2007 ; 13 ( 2 ): 161 – 190 . Google Scholar Crossref Search ADS PubMed 56 Schwartz SL , Gelfand AE , Miranda ML . Joint Bayesian analysis of birthweight and censored gestational age using finite mixture models . Stat Med . 2010 ; 29 ( 16 ): 1710 – 1723 . Google Scholar Crossref Search ADS PubMed 57 York TP , Vargas-Irwin C , Anderson WH , et al. . Asthma pharmacogenetic study using finite mixture models to handle drug-response heterogeneity . Pharmacogenomics . 2009 ; 10 ( 5 ): 753 – 767 . Google Scholar Crossref Search ADS PubMed 58 Blais L , Beauchesne MF , Rey E , et al. . Use of inhaled corticosteroids during the first trimester of pregnancy and the risk of congenital malformations among women with asthma . Thorax . 2007 ; 62 ( 4 ): 320 – 328 . Google Scholar Crossref Search ADS PubMed 59 Martel MJ , Rey E , Beauchesne MF , et al. . Use of inhaled corticosteroids during pregnancy and risk of pregnancy induced hypertension: nested case-control study . BMJ . 2005 ; 330 ( 7485 ): 230 . Google Scholar Crossref Search ADS PubMed 60 Diebolt J , Robert CP . Estimation of finite mixture distributions through Bayesian sampling . J R Stat Soc Series B Stat Methodol . 1994 ; 56 ( 2 ): 363 – 375 . 61 Loosemore S , McGrath R , Oram A , et al. . The GNU C Reference Manual . Boston, MA : Free Software Foundation, Inc. ; 2001 . https://www.gnu.org/software/gnu-c-manual/). Accessed July 2, 2018. 62 Arnold CC , Kramer MS , Hobbs CA , et al. . Very low birth weight: a problematic cohort for epidemiologic studies of very small or immature neonates . Am J Epidemiol . 1991 ; 134 ( 6 ): 604 – 613 . Google Scholar Crossref Search ADS PubMed 63 Hutcheon JA , Platt RW . The missing data problem in birth weight percentiles and thresholds for “small-for-gestational-age.” Am J Epidemiol . 2008 ; 167 ( 7 ): 786 – 792 . Google Scholar Crossref Search ADS PubMed 64 Kramer MS , Platt RW , Wen SW , et al. . A new and improved population-based Canadian reference for birth weight for gestational age . Pediatrics . 2001 ; 108 ( 2 ): E35 . Google Scholar Crossref Search ADS PubMed 65 Schatz M , Dombrowski MP , Wise R , et al. . The relationship of asthma medication use to perinatal outcomes . J Allergy Clin Immunol . 2004 ; 113 ( 6 ): 1040 – 1045 . Google Scholar Crossref Search ADS PubMed 66 Vazquez-Benitez G , Kharbanda EO , Naleway AL , et al. . Risk of preterm or small-for-gestational-age birth after influenza vaccination during pregnancy: caveats when conducting retrospective observational studies . Am J Epidemiol . 2016 ; 184 ( 3 ): 176 – 186 . Google Scholar Crossref Search ADS PubMed 67 Watson LF , Rayner JA , Forster D . Identifying risk factors for very preterm birth: a reference for clinicians . Midwifery . 2013 ; 29 ( 5 ): 434 – 439 . Google Scholar Crossref Search ADS PubMed 68 Morgan TK . Role of the placenta in preterm birth: a review . Am J Perinatol . 2016 ; 33 ( 3 ): 258 – 266 . Google Scholar Crossref Search ADS PubMed © The Author(s) 2018. Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Journal

American Journal of EpidemiologyOxford University Press

Published: Sep 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off