The present study is a comparison of WTCHR children who were ages 0–8 years on the date of the disaster compared with a sociodemographically matched group. We have previously described our study population and recruitment strategy¹¹ (Appendix; available at www.jpeds.com). Briefly, eligibility for the WTCHR was based upon dates of birth on or between September 11, 1993 and September 10, 2001, residence or school attendance south of Canal Street (or presence south of Chambers Street) on September 11, 2001.

Comparisons were not eligible for the WTCHR, and had dates of birth on or between September 11, 1993 and September 10, 2001. In addition to online and web-based recruitment, participants for the comparison group were recruited from general primary care visits at clinics affiliated with the New York University School of Medicine and were excluded if they presented to these clinics for a clinical concern, especially previously diagnosed asthma.

Potential participants that indicated presence of serious lung or heart disease, heart or lung surgery, or an active upper respiratory infection, pregnancy, and inability to follow procedures at the time of their scheduled visit, were also excluded from participation. To maximize comparability between the 2 study populations, we developed a table of the desired frequencies of controls by age (0–2, 3–5, or 6–8 years old on September 11, 2001), sex, race (White, African-American, Asian, other), ethnicity (Hispanic, non-Hispanic), and income (<$25 000, ≥$25 000).

A flow diagram of participants into our study along with a detailed description of population data is presented (Figure 1). The WTCHR group (n = 180) was more likely to be male (P = .008) than comparison participants. Participants in the comparison group (n = 122) were more likely to be non-Hispanic White and Hispanic, and the WTCHR group was more likely to be Asian (Table I).Our analysis of BMI category showed that the comparison group was more overweight or obese compared with the WTCHR group (P = .045). We identified a significant difference in WTC exposures between groups (P < .0001; Table I), but even with exclusion criteria, WTC exposures were found in the comparison group: 0.5% for dust cloud, 7.7% for home dust, and 42.8% for traumatic exposure. WTCHR participants were different from their nonparticipant counterparts by income status (P < .001) and age (P = .004; Table II; available at www.jpeds.com). Self-reported WTC exposures were moderately correlated with each other and by study group, supporting separate treatment in multivariable models controlling for confounders (Table III; available at www.jpeds.com).

Figure 2 shows the incidence of new onset post-9/11 asthma by cohort and exposure groups. Univariate analysis showed that post-9/11 risk of incident asthma was higher in the WTCHR participants (OR 1.102, 95% CI 1.016, 1.195) than in the comparison group (P = .019). Comparing by exposure rather than by group, dust cloud (OR 1.229, 95% CI 1.103, 1.371; P < .001) and home dust (OR 1.121, 95% CI 1.027, 1.223; P = .011) exposures were associated with a greater incidence of post-9/11 asthma (Table IV; available at www.jpeds.com). There were no differences in post-9/11 risk of incident asthma by psychologically traumatic exposures (P = .661). Multivariable models confirmed these findings: WTCHR (OR 1.109, 95% CI 1.021, 1.206; P = .015), dust cloud (OR 1.223, 95% CI 1.095, 1.365; P < .001), home dust (OR 1.123, 95% CI 1.029, 1.226; P = .009), and traumatic exposure (OR 1.023, 95% CI 0.941, 1.112; P = .597) (Table IV).

In contrast to the significant difference in asthma, univariate analyses failed to yield any significant differences in spirometry, plethysmography, or oscillometry by any of the exposure variables (Figures 3–6; available at www.jpeds.com). Multivariable models confirmed absence of differences in lung function between exposure groups (Table V). In addition, subgroup analysis of individuals with new diagnosis of asthma post-September 11 showed no deficits in lung function using spirometry, plethysmography, or oscillometry (data not shown).

Models stratified by BMI category failed to reveal differences in spirometry, oscillometry, or plethysmography in relationship to reported exposure (data not shown).

In the present study, we identified increases in post-September 11 asthma persisting 13–15 years after the WTC disaster. We did not identify a significant difference in lung function. The frequent diagnosis of post-September 11 asthma (~25%) is a serious concern, consistent with other previous reports,⁴ and further supports the need for ongoing care in young adults exposed to the disaster in early life. Yet, the absence of differences in spirometry, oscillometry, and plethysmography may provide some reassurance to exposed children and their families regarding long-term consequences for their pulmonary health.

Strengths to this study include inclusion of a sociodemographically matched group, as well as careful and rigorously standardized measurement of pulmonary function, which was performed for the first time in a population exposed to the disaster in early life. Like other studies of disaster-exposed populations, residual confounding limits our interpretation. During the 13- to 15-year period after the disaster, participants in both arms of the study may have experienced adverse and beneficial environmental exposures that could have modified the effects of dust and stress related exposures, and thereby affected measurements of pulmonary function. Participation rates within the WTCHR were modest, limiting generalizability. However, this is a disaster-related recruitment sample and there is no other population well-poised to inform our research question. There were also differences in sociodemographic characteristics and adiposity (as evaluated with anthropometric measurements) between the WTCHR participants and matched comparisons that could contribute to explaining our results.

We also evaluated participants when they were asymptomatic. This approach to evaluation could have attenuated real differences that may otherwise have been detected. Although other studies have used bronchodilators to unmask deficits in lung function or even bronchoprovocation with mannitol or methacholine,³⁶ the NYU School of Medicine Institutional Review Board judged such methods to be significantly more than minimal risk, rather than a minor increase above minimal risk. We, therefore, decided to exclude bronchodilation or bronchoprovocation from the present study, recognizing that null findings might leave open the possibility that we may have missed a significant decrement in pulmonary function masked (or compensated) in an otherwise healthy adolescent.

The use of questionnaire data for evaluating exposure also creates difficulties with exposure imprecision, specifically with regard to dust cloud and home dust exposure. Typically, exposure imprecision biases associations toward the null for categorical outcomes,³⁷ though the direction of bias for continuous outcomes such as those measured in our study cannot be ascertained. Biomarkers of exposure may be better predictors of exposure this far out after the disaster. We have identified increases in perfluoroalkylchemicals (a chemical known to be in WTC dust) in the serum of children in the WTCHR and in relationship to self-reported exposure.¹¹ Insofar as a biomarker can be used as a better proxy for exposure to the disaster, especially in comparison to questionnaire-based recall 13–15 years later, associations of biomarkers of exposure with pulmonary function may be more revealing. That said, there will be interesting challenges, in part because perfluoroalkyl substances (PFAS) exposure can be both related to the WTC disaster but also due to dietary and other non-WTC-related exposures. Careful control for confounding will be needed because PFAS themselves may be immune disruptors and affect pulmonary development.^38,39 We anticipate examining PFAS biomarkers both separately and in addition to WTC dust and stress-related exposures.

We appreciate that we only evaluated asthma prevalence rather than current asthma prevalence. Our methodological approach was predicated on the concept that medical care under the World Trade Center Health Program is based upon diagnosis rather than current asthma. In addition, our reliance on self-reported indications of asthma in NHANES is a potential methodological error, a concern that other studies have previously raised.^40,41 In contrast, lung function measures can be more specific in their diagnostic value. However, we were limited by the inability to preform bronchoprovocation, which would have further resolved these issues.

The present study confirms differences in incident asthma, but did not identify presence of chronic lung function abnormalities among asymptomatic children exposed to the WTC disaster when studied together with a comparison group. Further study with bronchodilation and/or methacholine challenge may be needed to identify and further evaluate effects of WTC exposure. Biomarker studies may also be more informative in delineating exposure-outcome relationships.