Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2020 Feb 1.
Published in final edited form as: Pediatr Pulmonol. 2018 Dec 25;54(2):158–164. doi: 10.1002/ppul.24219

Longitudinal Analysis of Bronchodilator Response in Asthmatics and Effect Modification of Age-related trends by Genotype

Joanne E Sordillo 1, Michael McGeachie 2, Sharon M Lutz 1, Jessica Lasky-Su 2, Kelan Tantisira 2, Ching Hui Tsai 3, Amber Dahlin 2, Rachel Kelly 2, Ann Chen Wu 1,4
PMCID: PMC6818258  NIHMSID: NIHMS1052850  PMID: 30585438

Abstract

Background and objectives:

Genome Wide Association Studies (GWAS) have identified genetic polymorphisms associated with bronchodilator response (BDR), but it is unknown how these associations change across life stages. We examined the impact of genetic variants on BDR from childhood to adulthood in asthmatics to uncover potential effect modification by age.

Methods:

We searched the National Human Genome Research Institute (NHGRI) catalog of published GWAS to obtain a list of genetic associations with BDR, and tested them for effect modification by age in 604 subjects from the Childhood Asthma Management Program (CAMP), a clinical trial with longitudinal measures of BDR (age range 5 to 30 years). We performed longitudinal analyses using linear mixed models and visualized longitudinal changes in BDR using generalized additive models with repeated measures, adjusting for treatment group, sex, and main effects of age and additive genotype.

Results:

Increasing age was associated with decreased BDR (−0.24 % per year). Polymorphisms rs295137 near SPATS2L and rs2626393 near ASB3 demonstrated their strongest associations with BDR in early childhood through adolescence, with a large decrease in their magnitude of effect from adolescence onward. The effect estimate for % BDR associated with rs295137 genotype (Beta=1.3; 95% CI 0.6 to 2.1) was diminished by age (interaction term=−0.06, p=0.004). The effect estimate for rs2626393 (Beta=−0.92 (95% CI −1.7 to −0.2) was also modified by age (interaction term= 0.05, p = 0.0004).

Conclusions:

Polymorphisms associated with BDR in childhood may not be relevant for predicting adolescent and adult BDR, which could reflect age-related changes in asthma phenotypes.

Trial Registration:

This work uses data from the previous clinical trial of asthma, the Childhood Asthma Management Program (CAMP), registered at ClinicalTrials.gov, #

Keywords: Bronchodilator response, age, asthma

INTRODUCTION

Asthma affects over 350 million people world-wide, with high prevalence in both pediatric and adult populations.1 Responsiveness to asthma medications significantly impacts quality of life for most individuals with asthma. Genome wide association studies (GWAS) have sought to identify genetic variants that predict asthma treatment response, with precision medicine (individualized treatment tailored by genotype)2 as the ultimate goal. These studies may combine pediatric and adult populations, with little to no emphasis on the potential role of age in modifying the impact of genetic variants on asthma treatment response. Age may be an important modifier of pharmacogenetic associations, especially since age-related differences in asthma phenotypes may reflect different underlying pathogenic pathways,3 the genetic variation in which may affect also asthma treatment response. In other words, the genetic polymorphisms with utility for predicting asthma treatment responses may change along with age-related changes in asthma phenotype. The identification of age-related asthma phenotypes is not new; clinicians have long understood that childhood asthma tends to be dominated by Th2 mediated responses and eosinophilia, while adult asthma phenotypes are more heterogeneous, and can be either Th2 mediated/eosinophilic, neutrophilic or obesity-related.4

One of the most widely used asthma medications in both child and adult populations is albuterol, a short-term β2-agonist that serves as a bronchodilator. β2-agonists promote bronchodilation by stimulating β2 adrenergic receptors on airway smooth muscle cells to reduce bronchoconstriction via downstream increases in cyclic adenosine monophosphate (cAMP) and protein kinase A (PKA).5 Bronchodilator response (BDR) to albuterol as an asthma phenotype involves inflammation and interactions among multiple airway cell types including the epithelium,6 smooth muscle cells7 and cells of the autonomic nervous system.8 GWAS studies have identified genetic polymorphisms associated with BDR.9–11 Thus far, the impact of these genetic variants on longitudinal changes in BDR from childhood to adulthood in individuals with asthma has not been examined. In this study, our aim was to test GWAS polymorphisms associated with BDR for effect modification by age in the Childhood Asthma Management Program (CAMP) clinical trial which has genome-wide polymorphism data and longitudinal bronchodilator response outcomes from early childhood to young adulthood (ages 5 to 30).

METHODS

CAMP Study Design and Participants

CAMP was a randomized, placebo-controlled trial of budesonide, nedocromil, or placebo for mild-to-moderate childhood asthma followed by three phases of observational follow-up; the trial and all follow-up phases included pre-bronchodilator and post-bronchodilator spirometry.12 A total of 1041 children, 5 to 12 years of age, were enrolled in the trial between December 1993 and September 1995.12 Inclusion was restricted to children with a history of chronic asthma who, during a 28-day run-in period, had asthma symptoms or a low morning peak flow on 8 or more days; inclusion also required a methacholine challenge with a concentration of 12.5 mg per milliliter or less that resulted in a reduction in FEV1 by at least 20% (since airway responsiveness is determined by the provocative concentration of methacholine required to reduce the FEV1 by at least 20% [methacholine PC20], with higher values indicating less airway responsiveness). Approval was obtained from the institutional review boards at each of the CAMP-participating institutions before initiation of the trial. Informed consent was obtained from the parent or guardian of the participant, and the child’s assent was obtained before study enrollment.

Participants were randomly assigned to receive budesonide, nedocromil or placebo, all by inhalation. Treatment was continued for a mean of 4.5 years (participants attended 3 clinical trial “visits” per year). The treatment component of the trial ended in 1999, and asthma care was transferred to each participants’ health care provider. Over 85% of the original 1041 participants participated in the 4 month transition phase (2 visits) following the trial, and 80% participated in all three phases (the treatment phase, the transition phase, and the continued follow-up phase (1–4 visits per year for 13 years)).13 We considered subjects of white race/ethnicity for our longitudinal analysis. Of the original 1041 participants, 711 were white, and 604 white subjects had genotyping data. The 604 subjects included both index children (N=570) and siblings (N=34). Follow-up procedures were identical for both index children and siblings. The age range for the longitudinal follow-up, including the clinical trial period, was 5 to 30 years of age. Longitudinal analyses included 13,997 observations for bronchodilator response in total with a maximum of 28 observations per subject. Follow-up time per subject ranged from 4 to 18 years, with a mean of 16 years ± standard deviation of 3 years. The number of BDR measurements per subject ranged from 6 to 28, with an average of 23 measurements.

SNP Selection and Genotyping

We used the NHGRI-EBI (National Health Genome Research Institute-European Bioinformatics Institute) catalog of published genome-wide association studies search tool (https://www.ebi.ac.uk/gwas/) to search for GWAS associations at a suggestive p value threshold of ≤ 1 × 10-5. We conducted our search with the following search terms “asthma (bronchodilator response)” and “bronchodilator response in asthma”. We identified 4 GWAS studies, 3 of which reported associations between genotypes and bronchodilator response. In these studies, four SNPs were associated with BDR9–11. These variants are shown in Table 1, and include SNPs associated with SPATS2L, ASB3, and COL22A1. Of these variants, 3 of the 4 were genotyped in CAMP. One of the SNPs, rs350729 (near the ASB3 gene), was not genotyped. Therefore, a proxy SNP (rs2626393) in LD with rs350729 (R2 = 0.99), was identified using the European populations (EUR) in LDlink, a web-based tool from NCBI (https://analysistools.nci.nih.gov/LDlink/). Genotyping was performed at the Channing Division of Network Medicine using Illumina Quad 610 microarray chips (Illumina, Inc., San Diego, CA). All of the SNPs considered met basic genotype quality control standards (minor allele frequency >0.05%, Hardy Weinberg equilibrium of at least 0.0001).

Table 1.

BDR GWAS SNPs identified in NHGRI search

GWAS study SNPs Minor
Allele		 Genes Functional
Class		 GWAS
p value		 Minor allele
frequency in CAMP
Duan, QL et al., 2013
(PMID: 23508266) 		rs11252394 A PRKCQ, IL2RA, IL15RA, KLF6 Intergenic variant 2 × 10−7 0.08
rs6988229 T COL22A1 Intron variant 9 × 10−6 0.22
Himes BE et al, 2012
(PMID:22792082) 		rs295137 T SPATS2L Intergenic variant 3 × 10−6 0.41
Israel E et al, 2015
(PMID:25562107 ) 		rs2626393
(Proxy SNP for rs350729)		 C ASB3, SOCS, JUND, CEBPB Intron variant 2 × 10−10 0.34
Open in a new tab

Spirometry and Bronchodilator Response

Spirometry was performed at least 4 hours after short acting bronchodilator use and 24 hours after long-acting bronchodilator use and met American Thoracic Society (ATS) criteria for acceptability and reproducibility. At least 3 spirometric maneuvers were performed, with at least 2 reproducible maneuvers required for each test. The BDR test was performed at randomization and at subsequent visits during which a methacholine challenge was not administered (twice yearly during the clinical trial and at least annually during the post-trial follow-up). Post-bronchodilator spirometric values were obtained at least 15 minutes after the administration of 2 puffs of albuterol (90 mg per puff). BDR was expressed as the percentage change in FEV1 as follows: BDR=100 × [(Post-FEV1- pre-FEV1)/pre-FEV1].

Statistical Analysis

To assess the impact of age by genotype interactions on longitudinal BDR response, we used a mixed model approach (PROC MIXED) in SAS Statistical Software. Mixed models included main effects of age and additive genotype, a multiplicative interaction term (age x genotype), covariates (sex and treatment group), and random effects (subject specific intercept and slope). In additional models, we also adjusted for random effect of family (as a small number of children were siblings), but found that this made no difference in the results (see Supplement). Adjustment for the first six genotype principle components also did not change results in a meaningful way (see Supplement). For testing, we used the general linear mixed model with a compound symmetry covariance structure, since our data fit the model assumptions of this approach. (We also considered a first order auto-regressive covariance structure. We found that this model had a higher AIC than the compound symmetry model, therefore we used the compound symmetry covariance structure instead). For visualization, we used a generalized additive mixed model, due to the flexible nature of the linearity assumption.

For the generalized additive mixed model we used the gamm4 package in R, which allows for random effect of age, and a running smoother with no specified knots for the fixed effect of age. (Further details of the smoother can be found in Wood et al 2004).24 We created genotype-specific smooth plots of longitudinal BDR trends with age to visualize trends. Generalized additive mixed models included main effects of age and genotype, as well as covariate adjustment for sex and treatment group.

RESULTS

CAMP Demographics

In CAMP, the mean age of enrollment was 8.95 years (standard deviation (s.d.) ± 2.13). Characteristics of study subjects are shown in Table E1. The mean age at asthma diagnosis in CAMP subjects was 3.08 years (s.d. ± 2.44). Sixty percent of subjects were male, and 25% had a maternal history of asthma. Overall, the socioeconomic status of participants was relatively high; only 19% reported a household income level of less than $30,000 per year. We tested age by genotype interactions in white subjects (N=604) only.

BDR as a function of Age

We examined the main effect of age on BDR response prior to testing for age by genotype interactions. BDR response consistently decreased with age, as is shown in Figure 1. Visual investigation of the generalized additive mixed model (Figure 1) supports the linear assumption between age and BDR. On average, BDR decreased 0.24 % with each increasing year of age. (Table 2) Treatment group and sex did not confound the association between age and BDR over time (Table 2), and were not independent predictors of BDR. Since the CAMP longitudinal follow-up included the typical time period overlapping with the onset of puberty (which is later for males than females), we tested for interactions between age and sex, as a way of potentially accounting for sex-specific shifts in BDR that may occur as following puberty. Test of the interaction term for age x sex were non-significant (p=0.13).

Figure 1.

Figure 1.

Open in a new tab

Plot of predicted BDR (proportion) with age in CAMP subjects, adjusted for sex and treatment group (p<0.001 for smooth).

Table 2.

Main Effects of Age, Treatment Group and Sex on Longitudinal BDR Outcome in Linear Mixed Models

Characteristic BDR Outcome
(β = BDR % change
per level of variable)
Univariate Model
β (p value)		 Adjusted Model
β (p value)
Age (years) −0.24 (<0.0001) −0.24 (<0.0001)
Male Sex -- 0.12 (0.75)
Treatment Group (vs. Placebo)
Budesonide -- −0.42 (0.38)
Nedocromil -- −0.17 (0.72)
Open in a new tab

Age by Genotype Interactions for Longitudinal BDR

We tested age by genotype interactions for all four identified BDR GWAS variants in mixed models, and found that two of the four variants modified the effect of age on longitudinal BDR response (Table 3). Age interacted with each of these variants to mitigate their effects. With increasing age, the SPATS2L variant (T allele) is associated with a smaller increase in BDR, and the ASB3 variant (C allele) with a smaller decrease in BDR. P values for the interaction terms ranged from p=0.004 to 0.0004, and were both significant at the Bonferroni corrected threshold (p<0.0125). These polymorphisms were prevalent in the CAMP population. The minor allele frequency for rs295137 near SPATS2L was 41 % (T allele), and the minor allele frequency of rs2626393 near ASB3 was 34% (C allele). GWAS polymorphisms for BDR identified in the Duan et al (9) study (an intronic SNP in COL22A1 and a downstream variant of PRKCQ, IL2RA, IL15RA, KLF6) did not show evidence for effect modification of age related changes in BDR.

Table 3.

Longitudinal BDR and Age by Genotype Interactions in Linear Mixed Models

BDR Outcome*
(β = BDR % change
per level of variable)**
SNP
rs11252394
(A allele) 		SNP
rs6988229
(T allele) 		SNP
rs295137
(T allele) 		SNP
rs2626393
(C allele)
β 95% CI β 95% CI β 95% CI β 95% CI
SNP (Additive) 1.2 −0.2 to 2.6 0.35 −0.5 to 1.2 1.3 0.6 to 2.1 −0.92 −1.7 to −0.2
SNP*Age Interaction −0.02 −0.08 to 0.04 0.02 −0.02 to 0.05 −0.06 −0.09 to −0.03 0.05 0.02 to 0.08
Open in a new tab
*

Models included both additive genetic effect for SNP and the interaction term; all models adjusted for age, sex and treatment group

**

β values with p< 0.05 level are indicated in bold.

After identifying two SNPs as potential effect modifiers of age related BDR response in linear mixed models, we then used generalized additive mixed models to visualize smooths for longitudinal age-related BDR changes by genotype. For each comparison, there were three groups, individuals without the polymorphism, subjects with a heterozygous genotype, and those homozygous for the polymorphism. Longitudinal changes in BDR response by genotype are shown in SPATS2L and ASB3 polymorphisms are shown in figures 2 and 3. Smooths for longitudinal BDR trends with age (centered about the mean) are shown separately for each genotype, with shaded areas indicating the 95% confidence intervals around the smooths (Figure 2a–2c; Figure 3a–3c). For ease of interpretability and comparison across genotypes, we also plotted an overlay of genotype-specific smooths on a single plot, for changes in percent BDR (expressed as a proportion) with age (Figure 2d and Figure 3d). At the earliest ages, these two variants had opposite directions of effect; the rs295137 variant (T allele) near the SPATS2L gene was associated with increased BDR response, while the rs2626393 variant (C allele) near the ASB3 gene was associated with lower BDR. (These effects are visualized in figures 2 and 3. At the beginning of follow-up, there is a clear distinction by SPATS2L rs295137 genotype (figure 2d); those with two copies of the minor allele (TT genotype) had the highest BDR, with heterozygous individuals (CT genotype) showing lower BDR, and those homozygous for the C allele (CC genotype) with the lowest BDR. The ASB3 variant rs2626393 was also associated with different BDR at the youngest ages (figure 3d); those with two copies of the minor allele (CC genotype) had the lowest BDR, while heterozygous individuals (CT genotype) and those homozygous for the major allele (TT genotype) had higher BDR at the beginning of follow-up.) For both the SPATS2L and ASB3 variants, differential BDR response by genotype diminished with age, such that genotype-specific BDR response curves appeared to converge in adolescence at approximately age 15 years. To test whether these variants had significant main effects or age by genotype interactions after this time point, we performed sensitivity analyses by re-analyzing our mixed models using repeated measures of BDR from age 15 onward (see Supplemental Table E2). In CAMP, the effect of the SPATS2L variant (T allele) and its interaction with age were no longer statistically significant, suggesting that SPATS2L genotype mainly predicts BDR response in childhood/early adolescence, and that its effects on BDR no longer vary by age once subjects reach adolescence. We also examined the effects of ASB3 genotype in follow-up data after age 15, to determine if main effects or age related interactions would still be present. The main genetic effect of the ASB3 variant remained similar, as did the interaction term. These findings suggest that the ASB3 variant (C allele) may still have some association with BDR response beyond age 15, although its effects are mitigated by increasing age, and graphs of the smooths show convergence for all three genotypes in adolescence. Given the relatively high minor allele frequencies of the ASB3 and SPATS2L SNPs, a large percentage of our study population (41%) had one or more minor alleles for these variants (T for rs295137 near SPATS2L and C for rs2626393 near ASB3).

Figure 2.

Figure 2.

Open in a new tab

Longitudinal BDR response by SPATS2L (rs295137) genotype. (a-c) Centered smooths adjusted for age, sex, and treatment group, shaded area indicates 95% confidence interval for smooth (p<0.001 for all smooths) (a) Homozygous for Wild-type (CC) (N=206) (b) Heterozygous (CT) (N=297) (c) Homozygous for SNP (TT) (N=101) (d) Smooths for predicted BDR with age for all 3 genotypes overlaid.

Figure 3.

Figure 3.

Open in a new tab

Longitudinal BDR response by ASB3 (rs2626393) genotype. (a-c) Centered smooths for longitudinal BDR with age, adjusted for sex and treatment group, shaded area indicates 95% confidence interval for smooth (p<0.001 for all smooths) (a) Homozygous for Wild-type (TT)(N=259) (b) Heterozygous (CT) (N=274) (c) Homozygous for SNP (CC)(N=71) (d) Smooth for predicted BDR with age for 3 genotypes Overlaid.

DISCUSSION

In this work, we leveraged longitudinal follow-up of over 600 subjects with asthma, each of whom had multiple bronchodilator response measures spanning from early childhood to adulthood, to determine whether associations between genetic polymorphisms and BDR vary by age. Two previously identified BDR GWAS polymorphisms, one in SPATS2L11 and the other near ASB3,10 both demonstrated their strongest associations with BDR in early childhood through adolescence, with a large decrease in their magnitude of effects from adolescence into adulthood. These findings indicate that polymorphisms associated with BDR response in childhood may be different from polymorphisms that predict BDR response in adolescence and adulthood. Changes in the associations of these polymorphisms with BDR over time may reflect changes in underlying asthma phenotypes with age.

Age-related changes in asthma phenotypes have been well documented, with marked shifts occurring after puberty.14 The clinical trajectory of children with asthma typically shows improvement in asthma symptoms over time, a longitudinal trend most likely to occur in subjects with mild to moderate persistent asthma15. In fact, it is estimated that asthma symptoms persist from childhood to adulthood in approximately one third of cases.16 Age-related changes in gene expression patterns may underlie these longitudinal changes in asthma phenotypes. Genomic studies in multiple tissue types have identified dramatic shifts in both the transcriptome17 and the methylome18 that occur with increasing age; however the dearth of airway tissue samples in these studies makes it difficult to identify age related changes in the lung that may alter asthma phenotypes or treatment response. The relatively few studies to examine lung-specific endophenotypes19 have related sputum cellularity to asthma treatment response, showing that cell types (i.e. eosinophil vs. neutrophil dominant) are more variable in adults and can be used to guide treatment response,20 whereas cell types in children are not markedly different by severity status, and have yet to be proven a useful guide to treatment.21, 22

The SPATS2L and ASB3 related polymorphisms associated with altered longitudinal bronchodilator responses are both biologically plausible modifiers of this lung function phenotype. In vitro functional experiments, published in conjunction with the BDR GWAS by Himes et al, show that miRNA knockdown on the SPATS2L gene leads to increased β2 adrenergic receptor expression, suggesting that this gene may be important for down regulation of β2 adrenergic receptors.11 ASB3 expression appears to be effected by β2 agonists, and is known to regulate differentiation in muscle cells.23 GWAS hits in both of these genes emerged in combined cohort studies of both children and adults; however, the primary discovery populations for these polymorphisms contained two clinical trials in childhood subjects (CAMP, CARE), as well as other trials that were a mix of child, adolescent and adult asthmatics (LOCCS, LODO, Sephracor, ACRN). We were unable to detect age modification of GWAS hits from the Duan et al study,9 which included polymorphisms in the COL22A1 gene. Of the genes identified as GWAS hits in our NHGRI search, COL22A1 is the least biologically plausible hit, with little known about this gene, except that it may encode for a cell adhesion ligand for skin fibroblasts and epithelial cells.

The strongest feature of our study is the longitudinal BDR phenotype data on asthmatics, where the effects of age on BDR trajectory are observed for each individual subject. We observed effect modification of associations by age, even after adjusting for sex and treatment group, and were able to visualize non-linear trends with age through the use of repeated measures generalized additive modeling techniques. Although there were many strengths to our study, a few caveats deserve mention. The CAMP study is unique in its extended, 16 year follow-up of asthmatic subjects and therefore we did not have access to a similar cohort for replication of our findings. We did not have Tanner stage assessment, which may have helped us tease apart age-related effects from those of pubertal onset. To optimize power, we focused only on GWAS hits for testing, but may have missed other important polymorphisms that interact with age to alter BDR. In addition, as with any analysis of BDR phenotypes, it is difficult to disentangle whether BDR shifts with age are due to alterations in treatment response to albuterol, or changes in asthma phenotype that reduce the baseline level of bronchoconstriction. Moreover, by screening only GWAS hits we may have missed some relevant polymorphisms, especially given that the GWAS studies we drew from combined child and adult populations together (potentially obscuring life-stage specific genetic associations for BDR that we could have tested in our longitudinal models). Nevertheless, focusing our longitudinal analysis on GWAS polymorphisms associated with BDR proved to be an effective method for identifying possible age-related modification of genetic effects.

CONCLUSIONS

In summary, our findings suggest that there are substantial biologic changes that occur in the airway as children age that may vary by genotype. These findings bear relevance for future precision medicine efforts, and suggest that genomic panels for identifying the most efficacious individualized treatment should rely on age-specific genomic makers.

Supplementary Material

supp Tables

Acknowledgments

Funded by National Institutes of Health (NIH) grants: R01HD085993, R00HL109162, U01HL065899

Abbreviations:

BDR

(Bronchodilator Response)

GWAS

(Genome Wide Association Study)

Footnotes

Conflict of Interest Statement: The authors have no conflicts of interest.

References

  • 1.GBD 2015 Chronic Respiratory Disease Collaborators. Global, regional, and national deaths, prevalence, disability-adjusted life years, and years lived with disability for chronic obstructive pulmonary disease and asthma, 1990–2015: a systematic analysis for the Global Burden of Disease Study 2015. Lancet Respir Med. 2017; [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Cardon LR, Harris T. Precision medicine, genomics and drug discovery: Table 1. Human Molecular Genetics. 2016;25(R2):R172. [DOI] [PubMed] [Google Scholar]
  • 3.Tan DJ, Walters EH, Perret JL et al. Age-of-asthma onset as a determinant of different asthma phenotypes in adults: a systematic review and meta-analysis of the literature. Expert Review of Respiratory Medicine. 2014;9(1):109–123. [DOI] [PubMed] [Google Scholar]
  • 4.Martinez F, Vercelli D. Asthma. Lancet. 2013;382(9901):1360–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Nelson HS. β-Adrenergic Bronchodilators. The New England Journal of Medicine. 1995;333(8):499. [DOI] [PubMed] [Google Scholar]
  • 6.Salathe M. Effects of beta-agonists on airway epithelial cells. J Allergy Clin Immunol.110:275. [DOI] [PubMed] [Google Scholar]
  • 7.Shore SA, Moore PE. Regulation of β-adrenergic responses in airway smooth muscle. Respiratory Physiology & Neurobiology. 2003;137(2):179–195. [DOI] [PubMed] [Google Scholar]
  • 8.Jartti T. Asthma, asthma medication and autonomic nervous system dysfunction. Clinical Physiology. 2001;21(2):260–269. [DOI] [PubMed] [Google Scholar]
  • 9.Duan QL, Lasky-Su J, Himes BE et al. A genome-wide association study of bronchodilator response in asthmatics. The pharmacogenomics journal. 2014;14(1):41. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Israel E, Lasky-Su J, Markezich A et al. Genome-Wide Association Study of Short-Acting β2-Agonists: A Novel Genome-Wide Significant Locus on Chromosome 2 near ASB3. American Journal of Respiratory and Critical Care Medicine. 2015;191(5):530. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Himes BE, Jiang X, Hu R et al. Genome-Wide Association Analysis in Asthma Subjects Identifies SPATS2L as a Novel Bronchodilator Response Gene. PLOS Genetics. 2012;8(7):e1002824. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.The Childhood Asthma Management Program Research Group. Long-Term Effects of Budesonide or Nedocromil in Children with Asthma. The New England Journal of Medicine. 2000;343(15):1054–1063. [DOI] [PubMed] [Google Scholar]
  • 13.McGeachie MJ, Yates KP, Zhou X et al. Patterns of Growth and Decline in Lung Function in Persistent Childhood Asthma. The New England journal of medicine. 20169;374(19):1842. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Fuseini H, Newcomb D. Mechanisms Driving Gender Differences in Asthma. Curr Allergy Asthma Rep. 2017;17(3):1–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Fuchs O, Bahmer T, Rabe KF, von Mutius E. Asthma transition from childhood into adulthood. The Lancet Respiratory Medicine. 2017;5(3):224–234. [DOI] [PubMed] [Google Scholar]
  • 16.Stern DA, Morgan WJ, Halonen M, Wright AL, Martiez FD.. Wheezing and bronchial hyper-responsiveness in early childhood as predictors of newly diagnosed asthma in early adulthood. . Lancet. 2008; 372(9643):1058–64.. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Peters MJ, Joehanes R, Pilling LC et al. The transcriptional landscape of age in human peripheral blood. Nature Communications. 2015;6:8570. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Horvath S. DNA methylation age of human tissues and cell types. Genome biology. 2013;14(10):R115. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Anderson GP. Endotyping asthma: new insights into key pathogenic mechanisms in a complex, heterogeneous disease. The Lancet. 2008;372(9643):1107–1119. [DOI] [PubMed] [Google Scholar]
  • 20.Nair P, Pizzichini MMM, Kjarsgaard M et al. Mepolizumab for Prednisone-Dependent Asthma with Sputum Eosinophilia. The New England Journal of Medicine. 2009;360(10):985–993. [DOI] [PubMed] [Google Scholar]
  • 21.Fleming L, Tsartsali L, Wilson N, Regamey N, Bush A. Sputum inflammatory phenotypes are not stable in children with asthma. Thorax. 2012;67(8):675–681. [DOI] [PubMed] [Google Scholar]
  • 22.Petsky HL, Li A, Chang AB. Tailored interventions based on sputum eosinophils versus clinical symptoms for asthma in children and adults. Cochrane Database Syst Rev. 2017; 8: CD005603. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.McDaneld Tara G., Hancock Deana L., Moody Diane E.. Altered mRNA abundance of ASB15 and four other genes in skeletal muscle following administration of β-adrenergic receptor agonists. Physiological Genomics. 2004;16(2):275–283. [DOI] [PubMed] [Google Scholar]
  • 24.Wood SN Stable and efficient multiple smoothing parameter estimation for generalized additive models. Journal of the American Statistical Association. 2004; 99:673–686 [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

supp Tables
NIHMS1052850-supplement-supp_Tables.docx (16.4KB, docx)

RESOURCES