Table 2: Association between specific pollutants and pediatric asthma.

Source/yearOutdoor variablesAge groupSample sizeClimate regionStudy designAssessment methodFindings and limits

Asthma diagnosis and symptom
Akinbami et al., 2010SO2, NO, O3, PM2.5, PM103–17 years34,073National, USCross-sectional/adjustedNational Health Interview Survey (NHIS) database; stratified multistage samplingaORs for current asthma for the highest quartile of estimated ozone exposure: 1.56 (95% CI: 1.15, 2.10) and for recent asthma attack 1.38 (95% CI: 0.99, 1.91). Limit: county-level 12-month averages of pollution are imprecise measures of children’s exposure to pollution.

Berhane et al., 2014NO2, PM10, PM2.5, O35–7 years1,211Western, USCohort/adjustedQuestionnaire, FeNO measurement, ambient air monitoring stationsIncreases in annual concentrations of 24-hr average NO2 and PM2.5 were associated with increase in FeNO. Limit: lack of information on time-activity patterns for the subjects could lead to misclassification of exposure

Cornell et al., 2012BC, PM2.5240Northeast, USCross-sectional/adjustedFeNO test, portable air sampling units, fixed BC monitor, PFT, Serum IgEBC higher in high-asthma neighborhoods (1.59 µg/m3 [95% CI 1.45–1.73]) than in low-asthma neighborhoods (1.16 µg/m3 [1.06–1.27]) with . Limit: sample was limited to middle-income households

Ebisu et al., 2011Urban land use, TRAP modeling, NO20-1 years680Northeast, USCross-sectional/adjustedInterview, asthma diary10% increase in urban land-use within 1,540 m buffer of infant’s residence associated with 1.09-fold increase in wheeze severity. This link becames insignificant with TRAP modeling added. Limit: NO2 as an indicator of overall TRAP misses other pollutants

Habre et al., 2014PM2.5, O3,6–14 years36Northeast, USCohort/adjustedSymptoms diary, skin test, air sampling, air monitoring, temperature, and humidity2 of the 3 highest frequency reactions were for ragweed (48%) and birch (39%). Exposure to O3 and particular matters was significantly associated with severe wheezing. Limit: reliance on central-site ambient measurements to assign outdoor exposure category

Jerschow, 2015Dichlorophenols (pesticide)≥6 years2,125 sample (% of children unclear)NationalCross-sectional/adjustedNHANES, dichlorophenols measured in urineHigher dichlorophenol levels were linked with asthma diagnosis, asthma prescriptions, missing work/school, exercise-induced wheezing in atopic wheezers. No association between dichlorophenol levels and asthma morbidity in nonatopic wheezers. Limit: reliance on self-reported data about wheezing problems.

Jung et al., 2012Polycyclic aromatic hydrocarbons (PAH)5-6 years354Northeast, USCohort/adjustedQuestionnaires, PAH air monitoring devices, blood samplesRepeated high exposure to pyrene was associated with report of asthma. Limit: PAH exposure was assessed only by 2 repeated measures 5 to 6 years apart, which could lead to misclassification

Lewis et al., 2013PM10, PM2.5, O35–12 years298Upper Midwest, USCohort/adjustedRespiratory symptom diary, ambient air monitoring, caregiver interviewOutdoor PM2.5, PM10, and O3 concentrations were associated with increased odds of respiratory symptoms, particularly in children using steroid medication. Similar associations were not realized with PM10-2.5. Limit: measuring symptoms using handwritten diaries by caregiver and the child could lead to errors.

Miller et al., 2010Polycyclic aromatic hydrocarbons (PAH)≥5 years222Northeast, USCohort/adjustedQuestionnaires, urine testing, immunoglobulintestingWidely varying levels of 10 PAH urinary metabolites were detected in all children. Levels of PAH metabolites were not associated with respiratory symptoms. Limit: the half-lives of PAH metabolites are short and thus variations in exposure across time may be large.

Nishimura et al., 2013O3, NO2, SO2, PM10, PM2.58–20 years4,320South, Northeast, West, Upper Midwest, Puerto Rico, USCohort/adjustedQuestionnaires, regional ambient air pollution data,Early life exposure to NO2 was associated with risk for asthma [OR = 1.17; 95% CI 1.04–1.31] in Latino and African American children across 5 US regions. Other pollutants’ impact varied across regions. Limit: measurement of PM2.5 was less complete than that of other pollutants, leading to a smaller sample.

Padula et al., 2015PAH9–18 years467West, USCross-sectional/adjustedPFTs, spirometry, skin testing, fixed air monitoring, wind and humiditySignificant association between PAH and lung function testing in nonasthmatic children: increase in PAH456 was associated with decrease in FEV1. Limit: change in pulmonary function over time wasn’t assessed

Patel et al., 2013O3, PM10, PM2.5, NO2, , BC14–19 years36Northeast, USCross-sectional/adjustedAethalometers to measure BC, EPA systems database, R-Tube, immunosorbent assaysBC and NO2 were positively associated with airway inflammation and oxidative stress. Limit: the use of central-site PM2.5 and O3 measurements could bias the effect estimate from them toward null.

Perez et al., 2012NO2, O3<18 years2.54M +Western, USCross-sectional/adjustedACS, local surveys, EPA air quality system, ambient air monitors, proximity to traffic8% of asthma cases were partially caused by resident proximity to major road. Link between proximity to major road and asthma exacerbations is positive. Limit: traffic density and vehicular emissions are not reflected in this metric of traffic proximity

Pongracic et al., 2010Fungal allergen exposure5–11 years936 children (moderate-severe asthma)National, USCohort/Adjusted for covariatesInterviews, portable air sampling, site inspections, dust samplesExcess symptom days per 2 weeks associated with increase in outdoor fungi level; increases in total fungal exposure was associated with increases in symptom days and asthma-related unscheduled visits. Limit: the study did not have children not sensitized to fungal allergens

Ratnapradipa et al., 2013Soot, exhaust, wood or oil smoke<5-6 (pre-school)691Northeast, USCross-sectional/adjustedStructured interviewsExposure to soot, exhaust, wood, or oil smoke was associated with higher risk of asthma than those never exposed. Limit: the cross-sectional nature of the study and the recall bias were associated with interview-based data

Sarnat et al., 2012BC, PM, , PM2.5,6–12 years58South, US, MexicoCross-sectional/adjustedeNO testing, air monitoring, air monitors, passive badge samplers, BMI measurementThere exists significant link between eNO and measures of PM and BC. PM pollutant levels predict acute respiratory responses better than NO2 measurements. Limit: clinical significance of the estimated increases in eNO with pollutant levels as observed here is unclear.

Spira-Cohen et al., 2011PM2.5, SO2, Elemental carbon (EC)10–12 years40Northeast, USCohort/adjustedQuestionnaires, air monitoring, time-activity daily diary, aethalometer, spirometryElevated risk of wheeze, shortness of breath, and total symptoms were associated with same-day increased personal EC, but not with personal PM2.5 mass. No associations with school-site PM2.5 or, SO2. Limit: a small sample size of only 40 study participant

Vette et al., 2013PM2.5, BC, NO2, NOx, CO, , VOCs14–16 years139Midwest, USCohort/adjustedFeNO testing, nasal lavage, F2-isoprostances, air monitoring, diaries, air monitoringThis paper is a protocol, yet preliminary data provide evidence of roadway impacts on the measured concentrations and indicate that variations in exposures between study participants are evident. Limit: full detailed results are yet to come, not in this paper

Zora et al., 2013, PM2.5, PM10, markers for TRAP (BC, NO2)6–11 years36South, USCross-sectional/adjustedQuestionnaire, ambient air monitoring, meteorology data, pulmonary function testingPositive (but not statistically significant) association between asthma and each single pollutant. Limit: use the questionnaire-based data as outcome variable could bring in recall bias, social desirability bias, etc.

Asthma cost
Brandt et al., 2012NO2, O30–17 years1,290Western, USCross-sectional/adjustedMEPS, CHIS, NHTS, HCUP, published averages of NO2 and O3Nearly 50% is due to regional air pollution-attributable exacerbations among children with asthma. Limit: costs are usually difficult to measure

Asthma-related symptoms and care utilization
Strickland et al., 2010, O3, NO2, SO2, CO, as markers for TRAP5–17 years91,386Southeast, USCross-sectional/adjustedAdministrative data (ICD-9) from ED visits, ambient air quality monitors, pollen countsAsthma ED visits associated with O3 during warm season and cold season (Nov–Apr), several TRAP measures in warm season, PM2.5 and SO2 in warm season, in cold season; associations with ED visits present at relatively low ambient concentrations of studied variables. Limit: difficult to draw causal inference from cross-sectional design

Tse et al., 2015Wildfire exposure2,195, 3,965West, USCross-sectional/adjustedShort-acting β-agonist (SABA) use in obese childrenSABA use increased (+16%, ) in obese children (BMI > 30) compared to nonobese (BMI < 30) in 2003; increased but nonsignificant difference (+10.5%, N.S.) in SABA use in 2007. Limit: asthmatic patients may have taken preventive action to minimize the exposure

Lemke et al., 2014NO2, SO2, VOC, PM10, PAH5–89 years2,900Upper Midwest, US & CanadaCross-sectional/adjustedGeospatial data, air sampling station data, ICD-9 codes with ED visits and hospitalizationsIntraurban air quality variations related to adverse respiratory events; NO2, PM10, and VOC positively correlated with ED visits. Limit: relatively coarse temporal resolution in study design compromises generalizability

Evans et al., 2014PM2.5, CO, SO2, O33–10 years74Northeast, USCross-sectional/adjustedPhysician visits, ER visitsIncreases in UFP and CO concentration were associated with pediatric asthma visits. Increases in O3 were associated with less asthma visits. No associations for mode particles, BC, fine particles, or SO2. Limit: the monitoring station is located on a diesel bus route, which could lead to higher measured pollutant concentrations than the actual exposure among some of the study subjects.

Delfino et al., 2014CO, N, PM2.5, O3, as markers for TRAP0–18 years11,390 visits/7,492 patientsWestern, USCross-sectional/adjustedEmergency Department visits, inpatient admissions; ambient air station dataED visits and admissions for asthma were positively associated with ambient air pollution (i.e., O3, PM2.5) during the warm season, and CO, NO2, PM2.5 in the cool season. Limit: insurance status is the only individual-level sociodemographic information

BC: black carbon; ED/ER: emergency department/emergency room; eNO: exhaled nitric oxide; SABA: Short-Acting Beta-Agonists; UFP: ultrafine particles; VOC: volatile organic compound. Note. (a) estimated exposure levels using LUR modeling.