ABSTRACT
Objective
To examine whether exposure to fine particulate matter (PM2.5) during the first year after neonatal intensive care unit (NICU) discharge is associated with asthma by age 5 among infants with bronchopulmonary dysplasia (BPD).
Methods
We conducted a retrospective cohort study of 337 infants with BPD, born between 2010 and 2019, who survived to discharge with clinical follow‐up in the Children's Hospital of Philadelphia Care Network through age 5. Daily residential census block group PM2.5 exposures were estimated using a spatiotemporal machine‐learning model and averaged over the first year after NICU discharge. Modified Poisson regression models with robust standard errors quantified associations of PM2.5 with asthma by age 5, adjusting for neonatal clinical factors, insurance, neighborhood deprivation, and race/ethnicity.
Results
By age 5 years, 169 (50.1%) infants had an asthma diagnosis. Mean annual PM2.5 exposure was 8.8 µg/m3 (SD 1.1). Each 1 µg/m3 increment of PM2.5 was associated with higher asthma risk (unadjusted RR 1.14, 95% CI: 1.03–1.25; fully adjusted aRR 1.19, 95% CI: 1.03–1.37). Compared to the lowest exposure tertile (mean 7.6 µg/m3), adjusted rates of asthma tended to be higher as exposure increased: Tertile 2 (mean 8.7 µg/m3, aRR 1.31; 95% CI: 0.98–1.74), Tertile 3 (mean 10.0 µg/m3, aRR 1.68, 95% CI: 1.17–2.4).
Conclusions
Exposure to higher ambient PM2.5 in the year after NICU discharge was associated with asthma by age 5 among children with BPD. These findings highlight early‐life air quality as a modifiable determinant of long‐term respiratory outcomes in infants with BPD.
1. Introduction
Bronchopulmonary dysplasia (BPD) is a chronic lung disease primarily affecting preterm infants who require prolonged ventilatory support in the neonatal intensive care unit (NICU) [1, 2]. Among infants born extremely preterm (< 28 weeks of gestation), BPD represents the most prevalent chronic complication, affecting approximately half of survivors [3]. BPD is characterized by disrupted alveolar development and increased airway reactivity [4], leading to persistent pulmonary deficits that can extend into childhood [5]. As children with BPD grow, they are at significantly elevated risk of chronic respiratory morbidities, including recurrent wheezing, respiratory infections, and asthma [5, 6]. Notably, nearly half of infants with BPD require hospital readmission for respiratory illness within their first year after discharge, which highlights the significant respiratory vulnerability these infants face in the early post‐discharge period [1, 7, 8, 9]. Furthermore, as BPD impairs alveolarization and heightens airway reactivity, ambient air pollution represents a potential stressor and contributor to respiratory illness.
Exposure to ambient air pollutants, such as particulate matter with an aerodynamic diameter of ≤ 2.5 µm (PM2.5), has been extensively associated with adverse respiratory outcomes in children [9, 10, 11, 12, 13, 14]. Ambient PM2.5 can penetrate the pulmonary system, and thus, elevated levels have been linked to higher rates of asthma exacerbations and hospitalizations, inflammatory responses, impaired lung function, and disruption of normal respiratory development [15, 16]. While the general pediatric population is susceptible to ambient air pollution, infants with BPD may represent an especially vulnerable subgroup, given prolonged supplemental respiratory support requirement in the setting of pulmonary underdevelopment, inflammation, and damage. Consequently, even modest increments in PM2.5 exposure could exacerbate preexisting lung deficits and increase risk for subsequent respiratory complications [9, 14, 17, 18].
Understanding the role of environmental exposures during critical developmental windows is essential for mitigating long‐term respiratory risks in vulnerable populations. For infants and children with BPD, the year following NICU discharge represents a pivotal phase in lung maturation, characterized by rapid alveolar and vascular growth, remodeling, and recovery from neonatal injuries [19]. Although biologically plausible, evidence directly linking early‐life PM2.5 exposure to asthma risk among children with BPD remains limited, and this knowledge gap may hinder the development of targeted preventive strategies aimed at improving long‐term respiratory outcomes in this vulnerable population.
In this study, we evaluated whether higher mean ambient PM2.5 exposure in the first 12 months following hospital discharge is associated with asthma diagnosis at age 5 among children previously diagnosed with BPD. We hypothesized that elevated PM2.5 exposure during this critical postnatal window would be associated with a higher risk of asthma onset by age 5.
2. Methods
2.1. Study Design and Population
We conducted a retrospective cohort study using electronic health records (EHR) of infants born at less than 32 weeks' gestation and diagnosed with BPD. Infants included in this analysis received neonatal care at the Children's Hospital of Philadelphia (CHOP) or the University of Pennsylvania Health System between January 1, 2010, and December 31, 2019. The diagnosis of BPD was determined based on criteria defined by the 2019 Neonatal Research Network (NRN) [20] which classifies BPD severity into three distinct grades according to the level of respiratory support received at 36 weeks postmenstrual age: Grade 1 involves nasal cannula at a flow of ≤ 2 L/min; Grade 2 involves nasal cannula at a flow of > 2 L/min or noninvasive positive airway pressure; and Grade 3 involves invasive mechanical ventilation. We included infants who survived until hospital discharge and who had residential information documented in the metropolitan Philadelphia region for the first 12 months following NICU discharge. We used residential data to estimate post‐discharge ambient air pollutant exposure at the census block group level. To ensure adequate clinical follow‐up and ascertain asthma diagnosis, we further restricted to infants with documented clinical follow‐up within the CHOP Care Network occurring through age 5; each included child had a least one outpatient visit after their fifth birthday. CHOP maintains over 30 pediatric primary care practices and over 15 specialty care and surgery centers throughout Pennsylvania and New Jersey [21]. Infants were included if they received any type of care including primary care, subspecialty care, acute care (e.g., urgent care or emergency department), or inpatient care at any CHOP‐affiliated site through age five. This allowed sufficient follow‐up time to reliably capture asthma diagnoses within the EHR. After applying these inclusion criteria, the final analytic cohort consisted of 337 infants. Characteristics comparing the included and excluded infants are displayed in Supporting Information S1: Table 1.
2.2. Outcome
The primary outcome was diagnosis of asthma by 5 years of age. Asthma diagnoses were identified through a two‐step process. First, asthma diagnoses were defined as at least one documented diagnosis using the International Classification of Diseases, Tenth Revision (ICD‐9/10), asthma codes (493.x/J45.x, available in Supporting Information S1: Table 2). Second, to enhance diagnostic accuracy, we then conducted manual chart reviews within the Epic electronic health system, using the search bar function to query for the term “asthma” across clinical documentation including provider notes, associated diagnoses for medication prescriptions, and clinical encounters. Chart review was performed by two independent reviewers (J.J.Z. and T.D.N.), with discrepancies resolved by consensus. Asthma diagnosis was assigned to infants who met either ICD‐9/10 or chart review criteria. All infants who were assessed to have asthma by ICD code also had confirmatory evidence on chart review.
2.3. Exposure Assessment
Residential addresses at the time of hospital discharge were extracted from the EHR and geocoded using ArcMap software (version 10.8; Esri, Redlands, CA), with ArcGIS StreetMap Premium North America 2021.1 address locator. We assigned each address to the corresponding census block group using 2010 US Census boundaries and estimated daily ambient PM2.5 concentrations with a spatiotemporal model that incorporates extreme gradient boosting and inverse‐distance weighting XGBoost‐IDW Synthesis (XIS) [23]. The XIS model uses satellite aerosol optic depth, meteorological, land use, and topographical parameters, yielding high‐resolution validated estimates across the contiguous United States from 2003 to 2023. Daily PM2.5 exposure predictions were averaged across the first 12 months post‐NICU discharge. For infants residing at multiple addresses, annual averages were calculated using address‐specific daily estimates accounting for moves. Residential moves and their respective durations were systematically captured and verified via the EHR and assumed to be the midpoint between address changes.
2.4. Covariates
We selected confounders a priori based on established associations in the literature and clinical judgment [7, 8, 24, 25, 26, 27, 28]. The relevant clinical neonatal covariates extracted from the EHR were gestational age, birth weight, infant sex, BPD grade, age at NICU discharge, respiratory support at the time of discharge, and inhaled medication use at the time of NICU discharge. In addition, we included birth period (2010–2014; 2015–2019) to reflect trends in PM2.5 exposure in the region and the updated National Ambient Air Quality Standard for PM2.5 (mean PM2.5 by birth year in supplement) [29, 30], infant insurance type (public vs. commercial), census tract‐level neighborhood material community deprivation using a validated measure that incorporates multiple neighborhood socioeconomic indicators [22], and race/ethnicity.
2.5. Statistical Analysis
Baseline cohort characteristics were summarized using standard descriptive statistics. Modified Poisson regression models with robust standard errors quantified the association of mean PM2.5 exposure during the first 12 months post‐NICU discharge with risk of asthma diagnosis through age 5 among infants with BPD. Regression models were sequentially built with initial adjustment for neonatal clinical covariates, gestational age, birth weight, infant sex, BPD grade, age at NICU discharge, respiratory support at the time of discharge, and inhaled medication use at the time of NICU discharge. Subsequent additional adjustment for insurance type and neighborhood community material deprivation, and final additional adjustment for race/ethnicity. This additive adjustment strategy was used to evaluate the extent to which the association of post‐discharge PM2.5 exposure with asthma risk was influenced by individual‐level clinical and sociodemographic factors, and neighborhood characteristics. We added race/ethnicity to the final model given that it is a social construct that serves as a proxy for multiple exposures, which may include exposure to PM2.5 [31, 32]. Model assumptions, including linearity of continuous predictors with the log link and potential multicollinearity among covariates, were evaluated prior to final modeling. To examine potential nonlinear exposure–outcome relationships, we categorized annual PM2.5 exposure into tertiles and re‐estimated the Poisson regression models, including indicator variables for Tertile 2 and Tertile 3, with Tertile 1 (lowest exposure) serving as the reference category.
In sensitivity analyses, we tested potential bias arising from differential referral patterns and geographic variability in access to clinical sites; we refitted the primary analysis stratifying on residential addresses within and outside of Philadelphia County. All statistical tests were two‐sided, and significance was set at p < 0.05. This study was approved by the Institutional Review Board of the Children's Hospital of Philadelphia (IRB #20‐018358). Given the retrospective nature of the study, the IRB did not require informed consent. This study is reported in accordance with the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) statement, using the checklist for cohort studies to guide transparent reporting. All analyses were conducted using Stata statistical software (Version 18, StataCorp, College Station, TX). Code is available at https://github.com/tnelin/PM2.5_Asthma.git.
3. Results
3.1. Population Characteristics at Baseline
Population characteristics at baseline are shown in Table 1. We identified 337 infants who had a geocoded address and clinical follow‐up in the CHOP Care Network through age 5 (Figure 1). The cohort predominantly included infants who were male (59%), born after 2014 (61%), and with public insurance (62%). Over half were identified as non‐Hispanic Black in the EHR (52%). The median [IQR] gestational age at birth was 27 [25–28] weeks, with the majority classified as extremely preterm (< 28 weeks'). The median [IQR] birth weight was 771 g and approximately three‐quarters of the infants had extremely low birth weights (< 1000 g). At the time of NICU discharge, infants had a mean (SD) age of 156 (89) days. While most infants (72%) were discharged without inhaled respiratory medications, 17% received albuterol alone, and 11% were prescribed albuterol/inhaled corticosteroid in combination. Approximately one‐fourth required some form of respiratory support at discharge, including supplemental oxygen (12%) or tracheostomy (14%). The mean (SD) material community neighborhood deprivation was 0.41 (0.16), and the mean (SD) PM2.5 was 8.8 (1.1) μg/m3 (Table 2). Annual mean PM2.5 exposure significantly varied by tertile of PM2.5 with a mean (SD) PM2.5 exposure of 7.6 (0.4) in the lowest tertile and 10.0 (0.6) μg/m3 in the highest tertile of exposure, p < 0.001 (Table 2). Annual mean PM2.5 exposure decreased over the study period (Supporting Information S1: Figure 1). There were 159 infants (47%) with a discharge address in Philadelphia County (Supporting Information S1: Table 3).
Table 1.
Demographic baseline characteristics of included infants.
| Characteristic | Overall | No asthma diagnosis by age 5 | Asthma diagnosis by age 5 |
|---|---|---|---|
| Demographic features | n = 337 | n = 168 | n = 169 |
| Sex (male)—n (%) | 199 (59) | 99 (59) | 100 (59) |
| Birth period > 2014—n (%) | 207 (61) | 104 (62) | 103 (61) |
| Insurance—n (%) | |||
| Commercial | 129 (38) | 72 (43) | 57 (34) |
| Public | 208 (62) | 96 (57) | 112 (66) |
| Race/ethnicity—n (%) | |||
| Non‐Hispanic Black | 174 (52) | 75 (44) | 99 (59) |
| Non‐Hispanic White | 88 (26) | 53 (32) | 35 (21) |
| Other | 75 (22) | 40 (24) | 35 (21) |
| Clinical features | |||
| Gestational age—weeks, median [IQR] | 27 [25–28] | 27 [25–29] | 26 [25–27] |
| Extremely preterm—n (%) | 227 (67) | 98 (58) | 129 (76) |
| Birth weight—g, median [IQR] | 771 [630–1015] | 820 [643–1073] | 737 [609–930] |
| ELBW+—n (%) | 249 (74) | 114 (68) | 135 (80) |
| BPD Grade—n (%)a | |||
| Grade 1 | 120 (36) | 64 (38) | 56 (33) |
| Grade 2 | 133 (39) | 62 (37) | 71 (42) |
| Grade 3 | 84 (25) | 42 (25) | 42 (25) |
| Readmission in year after NICU discharge—n (%) | 164 (49) | 57 (34) | 107 (63) |
| Discharge age—days, mean (SD) | 156 (89) | 154 (93) | 158 (86) |
| Medications at NICU discharge—n (%) | |||
| Albuterol | 58 (17) | 31 (18) | 27 (16) |
| Albuterol/inhaled corticosteroid | 37 (11) | 11 (6) | 26 (15) |
| Discharge respiratory support—n (%) | |||
| No support | 251 (74) | 127 (76) | 124 (73) |
| Supplemental oxygen | 40 (12) | 18 (10) | 22 (13) |
| Tracheostomy | 46 (14) | 23 (14) | 23 (13) |
Note: Extremely preterm infants correspond to infants born before 28 weeks of gestation.
BPD grade criteria are defined by the 2019 Neonatal Research Network. Grade 1 = nasal cannula at a flow of ≤ 2 L/min; Grade 2 = nasal cannula at a flow of > 2 L/min or noninvasive positive airway pressure; and Grade 3 = to infants requiring invasive mechanical ventilation.
Abbreviations: BPD, bronchopulmonary dysplasia; ELBW, extremely low birth weight < 1000 g; NICU, neonatal intensive care unit.
Figure 1.
Flowchart of cohort inclusion and exclusion criteria in the participant selection process.
Table 2.
Socioenvironmental features of the cohort.
| Characteristic | Overall | No asthma diagnosis by age 5 | Asthma diagnosis by age 5 |
|---|---|---|---|
| n = 337 | n = 168 | n = 169 | |
| PM2.5, mean (SD), µg/m3 | 8.8 (1.1) | 8.6 (1.1) | 8.9 (1.1) |
| Neighborhood community material deprivationa, mean (SD) | 0.41 (0.16) | 0.38 (0.16) | 0.43 (0.16) |
| PM2.5 exposure and asthma diagnosis by PM2.5 tertile | PM 2.5 Tertile 1 | PM 2.5 Tertile 2 | PM 2.5 Tertile 3 |
| n = 113 | n = 112 | n = 112 | |
| PM2.5, mean (SD), µg/m3 | 7.6 (0.4) | 8.7 (0.3) | 10.0 (0.6) |
| Neighborhood community material deprivationa, mean (SD) | 0.38 (0.16) | 0.41 (0.15) | 0.43 (0.17) |
| Asthma diagnosis by age 5, n (%) | 45 (40) | 58 (52) | 66 (59) |
Abbreviation: PM2.5, fine particulate matter with aerodynamic diameter ≤ 2.5 μm.
Six variables from the American Community Survey are used to create the Community Material Deprivation Index: fraction of the population with income in the past 12 months below the poverty level, median household income in the past 12 months, fraction of the population 25 and older with attainment of at least a high school degree, fraction of the population with no health insurance, fraction of households receiving public assistance income or food stamps or SNAP in the past 12 months, and fraction of houses that are vacant [22]. The nationwide mean for the 2018 deprivation index was 0.35. The cohort mean for the 2018 deprivation index in metropolitan Philadelphia was 0.41.
Table 3.
Modified Poisson regression models quantifying the association of mean annual PM2.5 exposure during the first 12 months post‐NICU discharge with asthma diagnosis by age 5 among infants with BPD (n = 337).
| PM2.5 exposure | Model 0 | Model 1ᵃ | Model 2ᵇ | Model 3ᶜ |
|---|---|---|---|---|
| RR (95% CI) | aRR (95% CI) | aRR (95% CI) | aRR (95% CI) | |
| Continuous, per 1 µg/m3 PM2.5 | 1.14 (1.03, 1.25) | 1.24 (1.08, 1.42) | 1.19 (1.03, 1.37) | 1.19 (1.03, 1.37) |
| Tertile 1 | Reference | Reference | Reference | Reference |
| Tertile 2 | 1.30 (0.97, 1.74) | 1.35 (1.01, 1.81) | 1.31 (0.98, 1.74) | 1.31 (0.98, 1.74) |
| Tertile 3 | 1.48 (1.12, 1.95) | 1.81 (1.26, 2.58) | 1.68 (1.17, 2.42) | 1.68 (1.17, 2.41) |
Note: Estimates obtained using Poisson regression with robust standard errors.
Abbreviations: aRR, adjusted risk ratio; PM2.5, fine particulate matter with aerodynamic diameter ≤ 2.5 μm; RR, risk ratio.
Adjusted for gestational age, birth weight, birth period, sex, BPD grade, discharge age, discharge support, and discharge medication.
Model 1 + insurance type and census‐tract neighborhood community material deprivation.
Model 2 + race/ethnicity.
3.2. Asthma by Age 5
By age 5, approximately 50% of the cohort had received a physician diagnosis of asthma. PM2.5 exposure in the first 12 months post‐NICU discharge was significantly associated with asthma diagnosis. Each µg/m3 increment of PM2.5 was associated with 14% higher risk of asthma (RR 1.14; 95% CI: 1.03, 1.25) (Table 3). The association persisted after adjustment for clinical covariates (aRR 1.24; 95% CI: 1.08, 1.42). The risk of asthma remained elevated after adjustment for insurance status, neighborhood community material deprivation, and race/ethnicity (aRR 1.19; 95% CI: 1.03, 1.37) (Table 3).
When PM2.5 exposure was modeled in tertiles, we observed a potential exposure dose–response pattern. Compared with infants in the lowest tertile (mean 7.6 µg/m3), those in the second tertile (mean 8.7 µg/m3) had a modestly elevated radjusted risk of asthma that was significant with adjustment for clinical covariates (aRR 1.35; 95% CI: 1.01, 1.81) but confidence intervals included the null with further adjustment for insurance, neighborhood deprivation, and race/ethnicity (aRR 1.31; 95% CI: 0.98, 1.74). In contrast, compared to the lowest tertile, effect estimates for infants in the highest tertile (mean 10.0 µg/m3) were significant in unadjusted and all adjusted models, with 68% higher risk in the fully adjusted model (aRR 1.68; 95% CI: 1.17, 2.41) (Table 3).
Results were similar when the cohort was restricted to within and outside of Philadelphia County; point estimates for continuous exposure were similar for continuous exposures and for the highest tertile compared to the lowest tertile (Supporting Information S1: Table 4). To assess potential confounding, we compared demographic and clinical characteristics across exposure tertiles, compared to infants in the lowest tertile of exposure, infants in the highest tertile of exposure were more likely to be diagnosed with asthma, be male, born 2014 or before, receive public insurance, be diagnosed with grade 1 BPD, and not require supplemental respiratory support at time of NICU discharge (Supporting Information S1: Table 5).
4. Discussion
In this retrospective cohort of 337 children with BPD, we found that average ambient PM2.5 exposure during the first 12 months after NICU discharge was associated with asthma diagnosis by 5 years of age. The observed associations remained robust after adjustment for clinical neonatal factors, socio‐economic factors, and race/ethnicity. Importantly, our results build upon evidence that links ambient PM2.5 to wheeze, respiratory deficits, and asthma in infants born preterm [22, 30], and demonstrate that, among children with compromised lung function, air pollution may influence asthma susceptibility later on in childhood. Specifically, our findings add to the small but growing literature suggesting that post‑discharge PM2.5 exposure may influence long‐term respiratory outcomes among infants with BPD.
The impacts of air pollution exposure on preterm infant lung health have been previously studied. For instance, a study by Decrue et al. [33] found that children born preterm had lower FEV1 than healthy controls and that prenatal exposure to air pollutants, such as PM2.5, PM10, and NO2, can further impair postnatal lung function. They note that preterm infants may have a reduced capacity to handle oxidative stress, making them especially susceptible to respiratory inflammation and long‐term lung impairment from pollution. However, while these findings corroborate our observed associations, the study relied on prenatal air pollution exposure to estimate postnatal effects.
Few studies isolate postnatal ambient air pollutant exposures as the primary exposure of interest, and those that do typically encompass short‐term exposures or short‐term outcomes. Our findings are consistent with a study by Kelchtermans et al. [17] in which short‑term post‐discharge pollution exposures among infants with BPD were associated with emergency department visits with systemic steroid administration. These findings were similar to our prior studies demonstrating associations of the CDC's environmental justice index and PM2.5 in the year after NICU discharge with medically attended acute respiratory illness (i.e., an ED visit or inpatient readmission) in the first year after NICU discharge among infants with BPD [8, 34] [9, 14]. However, these studies focused on acute episodes rather than how children fared in terms of long‐term respiratory outcomes. Our research builds upon these findings by examining early‐life ambient PM2.5 exposure with a longer‐term outcome, asthma diagnosis by age 5.
Our findings are biologically plausible. Lungs of preterm infants can exhibit limited alveolarization, dysregulated angiogenesis, and heightened oxidative stress [35, 36]. Ultrafine and fine particles penetrate distal airways, generate reactive oxygen species, and amplify Th2‑skewed inflammation, pathways that are already primed in infants with BPD [37, 38]. Models that categorized PM2.5 in tertiles revealed an exposure dose–response gradient, with the highest tertile of PM2.5 exposure conferring a significantly increased asthma risk among children compared to the lowest tertile of PM2.5 exposure. This dose–response relationship is consistent with biological synergies reflected in a 2025 study of asthma in Fresno, California, Cisneros et al., in which PM2.5 exposure above 9 μg/m3, the current standard set by the Environmental Protection Agency (EPA), over 14 days was associated with increased risk of an acute care visit for asthma [29, 39]. Average annual PM2.5 exposure in the cohort was less than the current EPA threshold of 9 μg/m3 [29]. Therefore, it is important to note that the adverse health effects of fine PM exposure may remain even at relatively low concentrations. Further, individual factors such as birth weight, gestational age, and BPD diagnosis have been shown to be associated with asthma risk in children [24, 25, 28]. Specifically, in a study of 131,783 children hospital with asthma in Denmark, Sweden, and Finland, Liu et al. reported that risk of childhood asthma hospitalization increased with each 1000 g decrease in birth weight (RR 1.17; 95% CI: 1.15, 1.18) and with each decrease in gestational age (RR 1.05; 95% CI: 1.04, 1.06) [28]. We detected an association of a potentially modifiable environmental exposure associated with increased risk of asthma, even when adjusting for individual clinical characteristics. Importantly, the effect sizes highlight that even small reductions in average exposure (e.g., 1 µg/m3) may provide meaningful long‐term respiratory benefits for children with BPD. Findings that have important implications for health policy and law as it pertains to air quality and the long‐term health outcomes of medically vulnerable infants and children.
4.1. Strengths and Limitations
Our study had several strengths. We utilized a cohort of infants with Grade 1–3 BPD with continued follow‐up in the CHOP Care Network through the age of 5. Ambient PM2.5 was obtained from a high‑resolution spatiotemporal machine learning XIS model and aggregated at the census block group level. We employed robust adjustment for clinical, demographic, and socioeconomic factors, in combination with sensitivity analyses, to strengthen the internal validity of our findings. In future studies, effort should be allocated to studies with personal air‑pollution monitors and where household triggers are captured, and whether other morbidities mediate the association presented here.
This study has several limitations that warrant consideration. PM2.5 exposure estimates used in these analyses did not capture indoor or personal environments, including exposures to secondhand smoke; therefore, non‑differential misclassification of the exposure is plausible. Furthermore, our observed associations are likely influenced by survivor and referral biases as only infants surviving to discharge and returning for follow‑up within the CHOP Care Network were studied. The impact of these biases can result in potentially under‑representing the sickest or most socio‑marginalized children. ICD‐9/10 codes may lead to outcome misclassification; however, we utilized a two‐step approach including manual chart reviews within the EHR to minimize this limitation. However, a provider diagnosed asthma by age 5 remains a proxy for respiratory symptoms and not an objective measure, such as performance on a pulmonary function test. Therefore, whether these findings represent true biologic asthma versus a childhood manifestation of BPD remains unclear. Infants included in the present study resided in one geographic region with its own PM2.5 exposure trends, and from one healthcare system with its unique practice variations; therefore, these findings may not be generalizable to other parts of the country and warrant further study. The associations presented here were estimated across a heterogeneous cohort spanning multiple counties in one geographic region. To address potential biases arising from differential referral patterns and clinical sites, we performed sensitivity analyses based on discharge address in Philadelphia County and outside of Philadelphia County. Although these models yielded similar effect estimates in both the magnitude and directionality of the associations, the smaller sample size likely accounts for the imprecise estimates.
5. Conclusion
We found associations of ambient PM2.5 exposure during the first 12 months after NICU discharge with asthma by age 5 among infants with BPD. Our analyses were robust to adjustment for clinical and socioeconomic confounders and exhibited a potential dose‐response relationship. These findings suggest that creating guidance to decrease fine PM exposure during a pivotal phase of lung growth may offer an actionable pathway to mitigate long‑term respiratory morbidity among high‐risk neonates with BPD. Given that environmental exposures may be modifiable with personal exposure mitigation and environmental policy, these results have important implications for developing targeted interventions to improve long‐term health outcomes in children with BPD.
Author Contributions
Lizbeth F. Gomez, Timothy D. Nelin, and Heather H. Burris conceptualized and designed the study, coordinated, supervised, and collected data, carried out analysis, interpreted the data, and analyzed the results. Lizbeth F. Gomez wrote the initial manuscript draft and reviewed and revised the manuscript throughout the submission process. Timothy D. Nelin and Heather H. Burris reviewed and revised the initial draft and were involved with revisions throughout the submission process. Jonathan J. Szeto, Joshua K. Radack, Sara B. DeMauro, Chén C. Kenyon, Nicolas P. Novick‐Goldstein, Kristan A. Scott, Daria C. Murosko, Kathleen A. Gibbs, Ella Whitman, Jonathan C. Levin, Allan C. Just, and Scott A. Lorch helped conceptualize the study, reviewed, and revised the manuscript throughout the submission process.
Funding
L.F.G. supported by NHLBI R01HL162354; T.D.N. supported by NIH T32 HL160493.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
AsthmaPMBPD_Supplements.docx.
Gomez L. F., Szeto J. J., Radack J. K., et al., “Associations of Early Life Ambient PM2.5 Exposure With Asthma Risk in a Cohort of Preterm Infants With Bronchopulmonary Dysplasia,” Pediatric Pulmonology 60 (2025): 1‐9. 10.1002/ppul.71432.
Data Availability Statement
The data sets generated during/or analyzed during the current study and the code used to analyze and manage the data are available from the corresponding author on reasonable request.
References
- 1. Watterberg K. L., Walsh M. C., Li L., et al., “Hydrocortisone to Improve Survival Without Bronchopulmonary Dysplasia,” New England Journal of Medicine 386, no. 12 (2022): 1121–1131, 10.1056/NEJMoa2114897. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Thébaud B., Goss K. N., Laughon M., et al., “Bronchopulmonary Dysplasia,” Nature Reviews Disease Primers 5, no. 1 (2019): 78, 10.1038/s41572-019-0127-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Bhandari A. and Panitch H. B., “Pulmonary Outcomes in Bronchopulmonary Dysplasia,” Seminars in Perinatology 30, no. 4 (2006): 219–226, 10.1053/j.semperi.2006.05.009. [DOI] [PubMed] [Google Scholar]
- 4. Cui X. and Fu J., “Reinitiating Lung Development: A Novel Approach in the Management of Bronchopulmonary Dysplasia,” Respiratory Research 25, no. 1 (2024): 384, 10.1186/s12931-024-02996-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Collaco J. M., Eldredge L. C., and McGrath‐Morrow S. A., “Long‐Term Pulmonary Outcomes in BPD Throughout the Life‐Course,” Journal of Perinatology, ahead of print, April 29, 2024, 10.1038/s41372-024-01957-9. [DOI] [PubMed] [Google Scholar]
- 6. Um‐Bergström P., Hallberg J., Pourbazargan M., et al., “Pulmonary Outcomes in Adults With a History of Bronchopulmonary Dysplasia Differ From Patients With Asthma,” Respiratory Research 20, no. 1 (2019): 102, 10.1186/s12931-019-1075-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Hannan K. E., Hwang S. S., and Bourque S. L., “Readmissions Among NICU Graduates: Who, When and Why?,” Seminars in Perinatology 44, no. 4 (2020): 151245, 10.1016/j.semperi.2020.151245. [DOI] [PubMed] [Google Scholar]
- 8. Nelin T. D., Yang N., Radack J., et al., “Associations of Neighborhood Social Vulnerability With Emergency Department Visits and Readmissions Among Infants With Bronchopulmonary Dysplasia,” Journal of Perinatology 43 (2023): 1308–1313, 10.1038/s41372-023-01735-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Nelin T. D., Radack J. K., Yang N., et al., “Associations of PM2.5 Exposure With Emergency Department Visits and Readmissions Among Preterm Infants With Bronchopulmonary Dysplasia,” Pediatric Pulmonology 59 (2024): 2947–2955, 10.1002/ppul.27164. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Ortega‐García J. A., Sánchez‐Solís M., and Ferrís‐Tortajada J., “Air Pollution and Children's Health,” Anales de Pediatría (Engl Ed) 89, no. 2 (2018): 77–79, 10.1016/j.anpedi.2018.04.017. [DOI] [PubMed] [Google Scholar]
- 11. Ierodiakonou D., Zanobetti A., Coull B. A., et al., “Ambient Air Pollution, Lung Function, and Airway Responsiveness in Asthmatic Children,” Journal of Allergy and Clinical Immunology 137, no. 2 (2016): 390–399, 10.1016/j.jaci.2015.05.028. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Orellano P., Quaranta N., Reynoso J., Balbi B., and Vasquez J., “Effect of Outdoor Air Pollution on Asthma Exacerbations in Children and Adults: Systematic Review and Multilevel Meta‐Analysis,” PLoS One 12, no. 3 (2017): e0174050, 10.1371/journal.pone.0174050. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Bi J., D'Souza R. R., Moss S., et al., “Acute Effects of Ambient Air Pollution on Asthma Emergency Department Visits in Ten U.S. States,” Environmental Health Perspectives 131, no. 4 (2023): 47003, 10.1289/EHP11661. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Nelin T., Radack J. K., DeMauro S. B., et al., “Modelling the Association of Reduction in PM 2.5 With Medically Attended Acute Respiratory Illness in Infants With Mild Bronchopulmonary Dysplasia,” Archives of Disease in Childhood‐Fetal and Neonatal Edition 110 (2025): 521–522, 10.1136/archdischild-2025-328533. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Altman M. C., Kattan M., O'Connor G. T., et al., “Associations Between Outdoor Air Pollutants and Non‐Viral Asthma Exacerbations and Airway Inflammatory Responses in Children and Adolescents Living in Urban Areas in the USA: A Retrospective Secondary Analysis,” Lancet Planetary Health 7, no. 1 (2023): e33–e44, 10.1016/S2542-5196(22)00302-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Bouazza N., Foissac F., Urien S., et al., “Fine Particulate Pollution and Asthma Exacerbations,” Archives of Disease in Childhood 103, no. 9 (2018): 828–831, 10.1136/archdischild-2017-312826. [DOI] [PubMed] [Google Scholar]
- 17. Kelchtermans J., Aoyama B. C., Rice J. L., Martin A., Collaco J. M., and McGrath‐Morrow S. A., “Ambient Air Pollution and Outpatient Morbidities in Bronchopulmonary Dysplasia,” Annals of the American Thoracic Society 21 (2023): 88–93, 10.1513/annalsats.202302-096oc. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Ruran H. B., Adamkiewicz G., Cunningham A., et al., “Air Quality, Environment and Respiratory Outcomes in Bronchopulmonary Dysplasia, the AERO‐BPD Cohort Study: Design and Adaptation During the SARS‐CoV‐2 Pandemic,” BMJ Open Respiratory Research 8, no. 1 (2021): e000915, 10.1136/bmjresp-2021-000915. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Callahan K. P., Farrell K., Gibbs K., et al., “Childhood Outcomes Following Discharge From a Referral Bronchopulmonary Dysplasia Program,” Journal of Perinatology 44, no. 12 (2024): 1832–1838, 10.1038/s41372-024-02035-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Jensen E. A., Dysart K., Gantz M. G., et al., “The Diagnosis of Bronchopulmonary Dysplasia in Very Preterm Infants an Evidence‐Based Approach,” American Journal of Respiratory and Critical Care Medicine 200 (2019): 751–759, 10.1164/rccm.201812-2348OC. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Children's Hospital of Philadelphia, CHOP's Pediatric Network (2025), https://www.chop.edu/about-us/our-pediatric-network.
- 22. Brokamp C., Beck A. F., Goyal N. K., Ryan P., Greenberg J. M., and Hall E. S., “Material Community Deprivation and Hospital Utilization During the First Year of Life: An Urban Population–Based Cohort Study,” Annals of Epidemiology 30 (2019): 37–43, 10.1016/j.annepidem.2018.11.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Just A. C., Arfer K. B., Rush J., Lyapustin A., and Kloog I., “XIS‐PM2.5: A Daily Spatiotemporal Machine‐Learning Model for PM2.5 in the Contiguous United States,” Environmental Research 271 (2025): 120948, 10.1016/j.envres.2025.120948. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Sun T., Yu H. Y., Yang M., Song Y. F., and Fu J. H., “Risk of Asthma in Preterm Infants With Bronchopulmonary Dysplasia: A Systematic Review and Meta‐Analysis,” World Journal of Pediatrics 19, no. 6 (2023): 549–556, 10.1007/s12519-023-00701-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Shin J. E., Jang H., Han J. H., et al., “Association Between Bronchopulmonary Dysplasia and Early Respiratory Morbidity in Children With Respiratory Distress Syndrome: A Case–Control Study Using Nationwide Data,” Scientific Reports 12, no. 1 (2022): 7578, 10.1038/s41598-022-11657-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Grant T., Croce E., and Matsui E. C., “Asthma and the Social Determinants of Health,” Annals of Allergy, Asthma & Immunology: Official Publication of the American College of Allergy, Asthma, & Immunology 128, no. 1 (2022): 5–11, 10.1016/j.anai.2021.10.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Mowitz M. E., Mangili A., Han L., et al., “Prevalence of Chronic Respiratory Morbidity, Length of Stay, Inpatient Readmissions, and Costs Among Extremely Preterm Infants With Bronchopulmonary Dysplasia,” Expert Review of Pharmacoeconomics & Outcomes Research 21, no. 5 (2021): 1117–1125, 10.1080/14737167.2021.1848554. [DOI] [PubMed] [Google Scholar]
- 28. Liu X., Olsen J., Agerbo E., et al., “Birth Weight, Gestational Age, Fetal Growth and Childhood Asthma Hospitalization,” Allergy, Asthma, and Clinical Immunology: Official Journal of the Canadian Society of Allergy and Clinical Immunology 10, no. 1 (2014): 13, 10.1186/1710-1492-10-13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Particulate Matter (PM2.5) Trends, (United States Environmental Protection Agency, 2025), https://www.epa.gov/air-trends/particulate-matter-pm25-trends. [Google Scholar]
- 30. Li C., van Donkelaar A., Hammer M. S., et al., “Reversal of Trends in Global Fine Particulate Matter Air Pollution,” Nature Communications 14, no. 1 (2023): 5349, 10.1038/s41467-023-41086-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Flanagin A., Frey T., and Christiansen S. L., “Updated Guidance on the Reporting of Race and Ethnicity in Medical and Science Journals,” Journal of the American Medical Association 326, no. 7 (2021): 621–627, 10.1001/jama.2021.13304. [DOI] [PubMed] [Google Scholar]
- 32. Tessum C. W., Paolella D. A., Chambliss S. E., Apte J. S., Hill J. D., and Marshall J. D., “PM2.5 Polluters Disproportionately and Systemically Affect People of Color in the United States,” Science Advances 7, no. 18 (2021): 4491, 10.1126/sciadv.abf4491. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Decrue F., Gorlanova O., Salem Y., et al., “Increased Impact of Air Pollution on Lung Function in Preterm Versus Term Infants the BILD Study,” American Journal of Respiratory and Critical Care Medicine 205, no. 1 (2022): 99–107, 10.1164/rccm.202102-0272OC. [DOI] [PubMed] [Google Scholar]
- 34. Szeto J. J., Radack J. K., DeMauro S. B., et al., “Environmental Determinants of Post‐Discharge Acute Respiratory Illness Among Preterm Infants With Bronchopulmonary Dysplasia,” International Journal of Environmental Research and Public Health 21, no. 5 (2024): 648, 10.3390/ijerph21050648. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Wright R. J., “Preterm Birth Enhances Ambient Pollution Toxicity Oxidative Stress and Placental Function,” American Journal of Respiratory and Critical Care Medicine 205, no. 1 (2022): 10–12, 10.1164/rccm.202110-2338ED. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Ferrante G., Carota G., Li Volti G., and Giuffrè M., “Biomarkers of Oxidative Stress for Neonatal Lung Disease,” Frontiers in Pediatrics 9 (2021): 618867, 10.3389/fped.2021.618867. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Restori K. H., Srinivasa B. T., Ward B. J., and Fixman E. D., “Neonatal Immunity, Respiratory Virus Infections, and the Development of Asthma,” Frontiers in Immunology 9, no. June (2018): 1249, 10.3389/fimmu.2018.01249. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Velten M., Heyob K. M., Rogers L. K., and Welty S. E., “Deficits in Lung Alveolarization and Function After Systemic Maternal Inflammation and Neonatal Hyperoxia Exposure,” Journal of Applied Physiology 108, no. 5 (2010): 1347–1356, 10.1152/japplphysiol.01392.2009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Cisneros R., Gharibi H., Amiri M., et al., “The Dose‐Response of Asthma to PM2.5 Exposure in the Underserved Community of Fresno, California,” Atmospheric Environment 360 (2025): 121414, 10.1016/j.atmosenv.2025.121414. [DOI] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
AsthmaPMBPD_Supplements.docx.
Data Availability Statement
The data sets generated during/or analyzed during the current study and the code used to analyze and manage the data are available from the corresponding author on reasonable request.