1 INTRODUCTION

Black and Hispanic children who live in urban areas are at particularly high risk of developing asthma.^{1, 2} Rhinovirus (RV) and respiratory syncytial virus (RSV) are important causes of wheezing illnesses during early life and are both associated with recurrent wheezing episodes and asthma.^3-6 Prior studies have found RV wheezing illnesses in early life to be more strongly associated with subsequent asthma than other viral illnesses, including RSV.^{3, 4, 7, 8} Children with the combination of allergic sensitization and RV wheezing are especially likely to develop atopic asthma.^{3, 9} In contrast, RSV wheezing illnesses may be more related to non-atopic asthma.⁷ Notably, not all wheezing illnesses are associated with viral infections, and non-viral respiratory illnesses may be more common in urban children.^{10, 11} Bacterial infections and non-allergic environmental irritants such as pollution may be contributing to these non-viral wheezing episodes.^12-14

Childhood asthma is a heterogeneous disorder with many different phenotypes. The URECA study previously described six phenotypes of respiratory health at 10 years of age based on longitudinal patterns of wheezing, allergic sensitization, and lung function.¹⁵ Of the four phenotypes associated with significant wheezing, the high wheeze-high atopy phenotype with low lung function [HW-HA-LF] had the greatest frequency of wheezing episodes from birth until 10 years of age. Furthermore, these phenotypes were associated with specific patterns of nasal cell gene expression. Early identification of this phenotype could enable interventions to prevent these adverse outcomes.

Given these findings, we hypothesized that the etiology of early-life wheezing illnesses is related to the probability of developing asthma and specific wheezing respiratory phenotypes. Specifically, we hypothesized that RV wheezing illnesses in early life will be associated with the high atopy phenotypes, whereas wheezing illnesses caused by other viruses (e.g. RSV) and non-viral wheezing episodes will be associated with the low atopy wheeze phenotypes. To test these hypotheses, we analyzed the relationships between the specific etiology of wheezing episodes within the first 3 years of life and subsequent asthma and wheezing respiratory phenotypes in URECA through 10 years of age.

2 METHODS

The design, methods, and study population of the URECA study have been previously reported.¹⁶ Briefly, URECA is a longitudinal birth cohort study in four U.S. urban areas: Baltimore, Boston, New York City, and St. Louis. Selection criteria included living in a census tract with more than 20% of families below the poverty level, mother or father with a history of allergic rhinitis, eczema, and/or asthma, and gestational age ≥34 weeks. Expectant families were recruited during the prenatal period, and written informed consent was obtained. Between February 2005 and March 2007, a total of 1850 families were screened, 776 met eligibility criteria and 560 newborns were enrolled at birth.

3 RESULTS

3.3 Wheezing illness etiology and risk of asthma

In a univariate analysis, we found that both viral and non-viral wheezing episodes were associated with an increased risk of asthma at 7 and 10 years (Figure 1). Similarly, wheezing with any of the specific virus types was also associated with an increased risk of developing asthma.

Using the hierarchical partitioning method, the overall sum of independent effects from all the viral etiologies and non-viral wheezing episodes explained 18.8% of the total variance for the 7-year asthma outcome and 14.5% of the total variance for the 10-year asthma outcome (Figure 2A,B). Considering all viruses together, any viral illness explained 11.0% of the asthma variance at 7 years and 8.7% at 10 years. Considering each etiology separately, non-viral wheezing episodes were more strongly associated with asthma at 7 and 10 years than any single viral etiology, explaining 8.0% and 5.8% of the variance, respectively. Among specific viral etiologies, RV had the greatest association with asthma at 7 years and multiple viruses at 10 years (6.3% and 3.9% of the variance, respectively). RSV had a relatively small independent effect on asthma outcome explaining <0.5% and 0.9% of the variance, respectively, at 7 and 10 years. We also analyzed this same data using multivariate logistic regression to estimate odds ratios (OR), 95% confidence intervals (CI), and p-values, in an unadjusted model and a model adjusted for gestational age and maternal smoking during pregnancy. This analysis supports our findings using the hierarchical partitioning method (Table S4).

3.4 Association between wheezing illness etiology and the wheezing respiratory phenotypes

We used multivariate analysis by hierarchical partitioning to determine how much of the variance of the wheezing respiratory phenotypes could be explained by knowing the etiology of early-life illnesses (Figure 3C). Viral wheezing episodes explained 10.1% of the total variance in phenotype outcome at 10 years of age. RV wheezing explained the most (4.9%) variance, followed by multiple viral wheezing episodes (3.1%), PIV (1.1%), and RSV (1%). Non-viral illnesses explained 3.8% of the variance in phenotypes.

We next used multinomial logistic regression to analyze relationships between specific etiologies of preschool wheezing illnesses and the four wheezing respiratory phenotypes (Figure 3A,B). The results showed that either viral wheeze or non-viral wheeze was a risk factor for TW-LA, MW-HA, or HW-HA-LF. Neither viral nor non-viral preschool wheezing was associated with the MW-HA phenotype. This same pattern was noted for wheezing illnesses associated with the most common viral etiologies (multiple viruses or RV).

4 DISCUSSION

Preschool viral wheezing episodes have been strongly linked to recurrent wheezing and asthma^{1, 2}; however, not all early-life wheezing episodes are viral. We found that among urban children at high risk for asthma development, approximately 1/3 of wheezing episodes were not associated with a virus. Both non-viral and viral wheezing episodes were important risk factors for subsequent asthma, but non-viral wheezing episodes had a greater impact than any individual viral etiology. Similarly, both viral and non-viral wheeze were associated with three wheeze phenotypes. These included transient wheeze and atopic wheeze with or without low lung function, but not the MW-LA (non-atopic wheeze) phenotype. While RV wheezing illnesses followed by non-viral wheezing episodes had the greatest impact on differentiating the wheezing phenotypes, these relationships were of modest strength.

Non-viral and viral wheezing episodes had similar symptoms. The etiology of these non-viral episodes is unclear and may be related to environmental exposures unique to urban communities.²³ Since we used viral PCR testing with high sensitivity (94%) and specificity (99%), it is unlikely that non-viral wheezing illnesses represent false negative testing.²⁴ We have previously noted that the URECA birth cohort had a lower rate of viral detection during respiratory illnesses (68%) compared to the Childhood Origins of Asthma (COAST) cohort (89%),^{10, 11} which was a suburban US cohort of White children. Lewis et al. also identified high rates of virus-negative illnesses (66% of respiratory illnesses) in an urban population from Detroit, MI.²⁵ In that study, virus-negative illnesses were associated with higher exhaled nitric oxide levels, which can be related to allergic inflammation and exposure to particulate pollutants.^{26, 27} Air pollution due to traffic and other sources is a risk factor for recurrent wheezing and asthma in urban communities and disproportionately affects urban minority children.¹⁴ The airway microbiome also relates to wheezing illnesses in early life. Moraxella catarrhalis, Streptococcus pneumoniae, and Haemophilus influenzae are detected more often in wheezing illnesses of infants and school-aged children. Studies involving sequential sampling of nasal secretions indicate that viral infections can promote increased detection of bacterial pathogens in airway secretions.^{28, 29}

In URECA, both viral and non-viral wheezing episodes during the first 3 years of life were associated with a greater risk of asthma. Several previous studies have linked non-viral episodes to asthma exacerbations in established disease,^{10, 11, 25, 30} but there is relatively little information about their relationship with the development of asthma. Similar to several prior studies, we found RV to be more strongly associated with subsequent asthma compared to other viral etiologies.^{3, 4, 7, 8} Our findings suggest that non-viral wheezing episodes are an important risk factor for asthma in urban populations, potentially related to the unique environmental exposures within urban populations.

Similarly, both non-viral and viral wheezing episodes were related to the wheezing respiratory phenotypes, but the specific illness etiology did not differentiate very well between the wheezing phenotypes. These findings did not support our hypothesis that the high atopy phenotypes would be associated with RV wheezing and the low atopy phenotypes would be associated with non-viral wheezing episodes. RV wheezing illnesses had only a slightly higher impact on the wheezing respiratory phenotypes compared to other etiologies; this suggests that factors other than specific viral illnesses, such as atopy or patterns of airway inflammation, may play a larger role in the development of these phenotypes.³¹ For example, analysis of nasal epithelial cells in URECA demonstrated that the HW-HA-LF phenotype was associated with higher expression of gene networks for type 2 inflammation pathways and MUC5AC, and lower expression reduced expression of innate antiviral gene networks.²¹ These differences in gene expression could increase susceptibility to viral wheeze, irrespective of the virus. Similarly, decreased lung function during infancy is a risk factor for preschool wheezing.³² This relationship is true for both RV and RSV-wheezing illnesses, suggesting the association between lung function and wheezing may be independent of viral etiology.³³

Another factor to consider is that Black children growing up in disadvantaged urban environments are much more likely to develop asthma during infancy compared to White children, who typically develop asthma later in life.^{2, 34} This difference in asthma incidence suggests that there are differences in etiology and perhaps causative environmental exposures. For example, Black children are more likely to be exposed to higher levels of air pollution, and individual and neighborhood stressors, and have suboptimal nutrition compared to White children. Notably, maternal perceived stress and depression have been linked to increased wheezing, asthma, and the MW-LA phenotype.^35-38 We also found low income and low birthweight to be risk factors for wheezing. In our analysis, MW-LA was the only phenotype not associated with increased viral or non-viral preschool wheeze.

Strengths of our study include the birth cohort design with high subject retention that has allowed for prospective frequent evaluations of wheezing episodes, asthma, allergy, and lung function. Our study population consists of predominantly Black and Hispanic children from urban neighborhoods who are at high risk for asthma based on environmental exposures and family history. While studies in these populations are critically needed, our data may not be generalizable to other populations. One other limitation is that we obtained viral diagnostics on a subset of wheezing illnesses,¹⁰ which limits the power of the analyses comparing etiologies and outcomes.

5 CONCLUSIONS

Non-viral wheezing episodes were frequently identified within our high-risk urban-residing population and were associated with a high risk of asthma. Urban environmental exposures (e.g., air pollutants and stress) could contribute to non-viral wheezing episodes and high asthma incidence rates in exposed children. Notably, specific etiologies of wheezing episodes provided relatively little information about subsequent specific wheezing respiratory phenotypes. Further research is needed to better understand the cause of non-viral wheezing episodes, which could enable new preventive strategies.

AUTHOR CONTRIBUTIONS

Tara N. Havens, James E. Gern, Daniel J. Jackson, Leonard B. Bacharier, Cindy Visness, and Agustin Calatroni had full access to all the data in the study and took responsibility for the integrity of the data and the accuracy of the data analysis. Tara N. Havens: writing- original draft (lead); review and editing (equal); conceptualization (equal); supervision (equal); validation (equal); data curation(equal); formal analysis (equal). Petra LeBeau: review and editing (equal); formal analysis (equal) and data curation (equal). Agustin Calatroni: review and editing (equal); formal analysis (equal) and data curation (equal). James E. Gern: review and editing (equal); conceptualization (equal); supervision (equal); validation (equal); funding (equal); data curation(equal); formal analysis (equal). George T. O'Connor: review and editing (equal); supervision (equal); validation (equal); funding (equal). Robert A. Wood: review and editing (equal); supervision (equal); validation (equal); funding (equal). Carin Lamm: review and editing (equal); supervision (equal); validation (equal); funding (equal). Rebecca Z. Krouse: review and editing (equal); formal analysis (equal) and data curation (equal). Cynthia M. Visness: review and editing (equal); formal analysis (equal) and data curation (equal). Peter J. Gergen: review and editing (equal); supervision (equal); validation (equal); funding (equal). Daniel J. Jackson: review and editing (equal); conceptualization (equal); supervision (equal); validation (equal); funding (equal); data curation(equal); formal analysis (equal). Leonard B. Bacharier: review and editing (equal); conceptualization (equal); supervision (equal); validation (equal); funding (equal); data curation(equal); formal analysis (equal).

FUNDING INFORMATION

This project has been funded with Federal funds from the National Institute of Allergy and Infectious Diseases, National Institutes of Health, under Grant/Contract numbers 1UM1AI160040, NO1-AI-25496, NO1-AI-25482, HHSN272200900052C, HHSN272201000052I, 1UM1AI114271-01, and UM2AI117870. Additional support was provided by the National Center for Research Resources, National Institutes of Health, under grants RR00052, M01RR00533, 1UL1RR025771, M01RR00071, 1UL1RR024156, and 5UL1RR024992-02, and the National Center for the Advancement of Translational Research, National Institutes of Health, under grants UL1TR001079 and UL1TR000040. The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

CONFLICT OF INTEREST STATEMENT

All authors, with the exception of P. Gergen and T. Havens report grants from NIH/NIAID during the conduct of study. T. Havens, C. Visness, P. LeBeau, A. Calatroni, and R. Krouse have nothing to disclose outside the submitted work. D. Jackson reports personal fees from Novartis, Pfizer, Regeneron, AstraZeneca, Sanofi, and Vifor Pharma, grants and personal fees from GlaxoSmithKline, and grants from NIH/NHLBI, outside the submitted work. G. O'Connor reports grants from NIH/NIAID/DAIT, personal fees from AstraZenec, and grants from Janssen Pharmaceuticals, outside the submitted work. J. Gern reports personal fees from AstraZeneca, and Gossamer Bio and personal fees and stock options from Meissa Vaccines Inc. In addition, J. Gern has a patent Methods of Propagating Rhinovirus C in Previously Unsusceptible Cell Lines issued, and a patent Adapted Rhinovirus C issued. R. Wood reports grants from DBV, Aimmune, Astellas, HAL-Allergy, and Regeneron and royalties from Up to Date, outside the submitted work. L. Bacharier reports personal fees from GlaxoSmithKline, Genentech/Novartis, DBV Technologies, Teva, Boehringer Ingelheim, AstraZeneca, WebMD/Medscape, Sanofi/Regeneron, Vectura and Circassia and personal fees and non-financial support from Merck, outside the submitted work.

DISCLAIMER

The content of this article is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.