UBC Faculty Research and Publications

Traffic-Related Air Pollution and Otitis Media Brauer, Michael; Gehring, Ulrike; Brunekreef, Bert; de Jongste, Johan; Gerritsen, Jorrit; Rovers, Maroeska; Wichmann, Heinz-Erich; Wijga, Alet; Heinrich, Joachim Sep 30, 2006

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-Brauer_M_et_al_Traffic_air_pollution_otitis_media.pdf [ 124.97kB ]
JSON: 52383-1.0074693.json
JSON-LD: 52383-1.0074693-ld.json
RDF/XML (Pretty): 52383-1.0074693-rdf.xml
RDF/JSON: 52383-1.0074693-rdf.json
Turtle: 52383-1.0074693-turtle.txt
N-Triples: 52383-1.0074693-rdf-ntriples.txt
Original Record: 52383-1.0074693-source.json
Full Text

Full Text

1414 VOLUME 114 | NUMBER 9 | September 2006 • Environmental Health PerspectivesResearch | Children’s HealthOtitis media is one of the most commonchildhood infections in young children. Threeof four children experience otitis media by 3years of age, with most infections occurringbefore age 2 (Bluestone and Klein 2001).Otitis media is one of the leading causes ofdoctor’s visits in childhood (Freid et al. 1998)and the main reason for children to consumeantibiotics or undergo surgery in developedcountries (Rovers et al. 2004). Otitis mediawith effusion (OME), in which fluid andmucus stay trapped in the ear after infection,may lead to conductive hearing loss that, ifpersistent, may lead to delays in the develop-ment of speech, language, and cognitive abili-ties (Klein 2000; Teele et al. 1990). Recurrentacute otitis media leads to decreased quality oflife measurements in children and is also stress-ful to their caregivers (Brouwer et al. 2005). Inaddition, the direct and indirect costs associ-ated with otitis media are high: In the UnitedStates, annual health care costs were estimatedat $3–5 billion (Bondy et al. 2000). Indirectcosts due to caregiver work loss are also sub-stantial and may in fact exceed direct costs. In1994, the total yearly cost for otitis media inCanada was estimated to be $611 million—60% of the total economic cost associated withall forms of diabetes (Coyte et al. 1999).Evidence also indicates a steady increase in theincidence of diagnoses (Bluestone and Klein2001; Lanphear et al. 1997). Consequently,identification of potentially preventable riskfactors for otitis media, such as air pollutionexposure, would have significant implicationsfor health care costs. Because air pollution isnot typically considered a risk factor for otitismedia, this illness is also not considered in airpollution health impact and cost–benefitassessments (Kunzli et al. 2000).Of particular relevance to a possible associa-tion between air pollution exposure and otitismedia is the strength of environmental tobaccosmoke (ETS) exposure as a risk factor (Jinotand Bayard 1996; National Cancer Institute1999; U.S. Department of Health and HumanServices 1986). A quantitative meta-analysispublished in 1998 concluded that consistentevidence of causal relationship between parentalsmoking in the home and acute otitis mediaexists (Strachan and Cook 1998). Despite thisfinding and the large number of studies assess-ing the impact of ambient air pollution expo-sures on upper respiratory infections, thepotential relationships between episodes of oti-tis media and ambient air pollution exposurehave not been examined in detail. Accordingly,we assessed the relationship between traffic-related air pollution and otitis media in twobirth cohorts, one in the Netherlands andanother in Munich, Germany.Previous analysis of this Dutch birthcohort indicated a significant associationbetween a combined measure of severe upperrespiratory tract and ear/nose/throat infec-tions for the 12-month period before thechild’s second birthday and exposure to traf-fic-related air pollutants (Brauer et al. 2002),but did not address otitis media specifically.This earlier analysis focused on the periodbetween 12 and 24 months of age, whereasotitis media incidence peaks between 6 and11 months of age (Rovers et al. 2004). Earlieranalysis of the German cohort indicated asso-ciations between air pollution exposure and(nocturnal dry) cough without respiratoryinfections (Gehring et al. 2002). Neither ofthese analyses included independent assess-ment of otitis media episodes in relation toair pollution exposure.Address correspondence to M. Brauer, School ofOccupational and Environmental Hygiene,University of British Columbia, 3rd Floor, 2206East Mall, Vancouver BC V6T1Z3 Canada.Telephone (604) 822-9585. Fax: (604) 822-9588.E-mail: brauer@interchange.ubc.caWe thank K. Meliefste, J. Cyrys, C. Harmath,M. Zeiler, K. Koschine, and M. Pitz for air pollutionsampling and measurement, P. van Vliet for prepara-tion of GIS data in the Netherlands, and G. Sedlmairfrom Stadt München, Referat für Umwelt undGesundheit, for providing GIS data for Munich. This study was supported by European UnionEnvironment contracts ENV4 CT97-0506 andQLRT 2000-00073 (exposure modeling); by FederalMinistry for Education, Science, Research andTechnology grants 01EG 9732 and 01EG 9705/2(LISA-Munich cohort); and by the NetherlandsAsthma Fund (94.27), the Ministry of theEnvironment, ZorgOnderzoek Nederland, and theNational Institute of Public Health and theEnvironment (PIAMA cohort). M.B. was supportedin part by funding from the Michael SmithFoundation for Health Research to the Centre forHealth and Environment Research at the Universityof British Columbia. U.G. was supported by aresearch fellowship within the Postdoc-Programme ofthe German Academic Exchange Service (DAAD).The authors declare they have no competingfinancial interests.Received 13 February 2006; accepted 26 April 2006.Traffic-Related Air Pollution and Otitis MediaMichael Brauer,1 Ulrike Gehring,2 Bert Brunekreef,3 Johan de Jongste,4 Jorrit Gerritsen,5 Maroeska Rovers,6Heinz-Erich Wichmann,2 Alet Wijga,7 and Joachim Heinrich21University of British Columbia, School of Occupational and Environmental Hygiene, Vancouver, British Columbia, Canada;2GSF–National Research Center for Environment and Health, Institute of Epidemiology, Neuherberg, Germany; 3Institute for RiskAssessment Sciences, Utrecht University, Utrecht, the Netherlands; 4Department of Pediatrics, Division of Respiratory Medicine,Erasmus Medical Center–Sophia Children’s Hospital, Erasmus University, Rotterdam, the Netherlands; 5Department of PediatricRespiratory Medicine, University Medical Centre Groningen, University of Groningen, Groningen, the Netherlands; 6Julius Centre forHealth Sciences and Primary Care, University Medical Centre, Utrecht, the Netherlands; 7Centre for Prevention and Health ServicesResearch, National Institute for Public Health and the Environment (RIVM), Bilthoven, the NetherlandsBACKGROUND: Otitis media is one of the most common infections in young children. Althoughexposure to environmental tobacco smoke is a known risk factor associated with otitis media, littleinformation is available regarding the potential association with air pollution. OBJECTIVE: We set out to study the relationship between exposure to traffic-related air pollutionand otitis media in two birth cohorts. METHODS: Individual estimates of outdoor concentrations of traffic-related air pollutants—nitro-gen dioxide, fine particles [particulate matter with aerodynamic diameters ≤ 2.5 µm (PM2.5)], andelemental carbon—were calculated for home addresses of approximately 3,700 and 650 infantsfrom birth cohort studies in the Netherlands and Germany, respectively. Air pollution exposure wasanalyzed in relation to physician diagnosis of otitis media in the first 2 years of life. RESULTS: Odds ratios (adjusted for known major risk factors) for otitis media indicated positiveassociations with traffic-related air pollutants. An increase in 3 µg/m3 PM2.5, 0.5 µg/m3 elementalcarbon, and 10 µg/m3 NO2 was associated with odds ratios of 1.13 (95% confidence interval,1.00–1.27), 1.10 (1.00–1.22), and 1.14 (1.03–1.27) in the Netherlands and 1.24 (0.84–1.83), 1.10(0.86–1.41), and 1.14 (0.87–1.49) in Germany, respectively. CONCLUSIONS: These findings indicate an association between exposure to traffic-related air pollu-tants and the incidence of otitis media. Given the ubiquitous nature of air pollution exposure andthe importance of otitis media to children’s health, these findings have significant public healthimplications. KEY WORDS: air pollution, cohort studies, infant, otitis media, vehicle emissions. Environ HealthPerspect 114:1414–1418 (2006). doi:10.1289/ehp.9089 available via http://dx.doi.org/ [Online26 April 2006]Air pollution and otitis mediaEnvironmental Health Perspectives • VOLUME 114 | NUMBER 9 | September 2006 1415Materials and MethodsStudy populations. The Prevention andIncidence of Asthma and Mite Allergy(PIAMA) study is a prospective birth cohortstudy with an initial enrollment of 4,146 chil-dren (Brunekreef et al. 2002; Koopman et al.2001; Wijga et al. 2001). The cohort wasrecruited in 1997–1998 during the secondtrimester of pregnancy from a series ofcommunities varying from rural to largecities in the north, west, and center of theNetherlands. Mothers were classified as aller-gic or nonallergic on the basis of a validatedscreening questionnaire (Lakwijk et al. 1998).Nonallergic (based on a screening question-naire) pregnant women were invited to partic-ipate in a “natural history” study arm (initialenrollment of 3,291). Pregnant women iden-tified as allergic through a screening ques-tionnaire were allocated primarily to anintervention arm (initial enrollment of 855),with a random subset allocated to the naturalhistory arm. The intervention involved the useof mite-impermeable mattress and pillow cov-ers. The study protocol was approved by theinstitutional review boards of each participat-ing institute, and parents or guardians of allsubjects gave written informed consent. Allsubjects with completed questionnaires at2 years of age were included in the analyses.The participants of the LISA-Munich(Influence of Lifestyle factors on the develop-ment of the Immune system and Allergies inEast and West Germany) birth cohort studywere recruited during pregnancy. FromDecember 1997 to January 1999, newbornsfrom six obstetrical clinics in Munich whoseparents were born in Germany and hadGerman nationality were defined as the targetpopulation for the study. A detailed descrip-tion of selection, exclusion criteria, and char-acteristics of study population has beenpublished previously (Gehring et al. 2002).For this analyses, we selected all infants withbirth addresses in Munich (without surround-ing communities) for whom questionnairedata were available for the first year of life andwho did not move away from Munich withinthe first year of life. A total of 673 subjectsfrom the LISA cohort fulfilled these criteria.The study was approved by the ethics com-mission of the Landesaerztekammer Bavariaand was carried out in accordance with theinstitutional guidelines for the protection ofhuman subjects. Parents or guardians of allsubjects gave written informed consent.Exposure assessment. Air pollution concen-trations at the home address of each memberof the cohort were calculated by combining airpollution measurements with a geographicinformation system (GIS; Brauer et al. 2003;Gehring et al. 2002). Briefly, air pollutantswere measured at 40 individual sites in eachcountry, designed to capture the maximumvariability in pollution from traffic sources.Fine particles [particulate matter with aero-dynamic diameter ≤ 2.5 µm (PM2.5)] werecollected with Harvard impactors (AirDiagnostics and Engineering, Harrison, ME,USA), nitrogen dioxide was collected byPalmes tubes, and light-absorbing carbon wasmeasured as the reflectance of the PM2.5 fil-ters, a method shown to be highly correlatedwith thermal elemental carbon measurement(Cyrys et al. 2003). At each location, measure-ments were conducted for four 2-week periodsdispersed throughout 1 year and then adjustedfor temporal trends, based on the differencebetween concentrations measured during eachperiod to annual average measurements (Hoeket al. 2002), to calculate long-term averageconcentrations for the 40 locations.Geographic data were also collected regard-ing traffic, road, and population density in thevicinity of each monitoring location. We devel-oped regression models to relate the annualaverage concentrations at the 40 monitoringsites with the geographic variables. For exam-ple, in the Netherlands, the number of high-traffic roads within a 250-m radius of alocation, the presence of a major road within a50-m distance, the density of buildings(addresses) within a 300-m radius, and an indi-cator for the region of the country were used inthe model to predict light-absorbing carbonconcentrations. In Munich, the light-absorbingcarbon model used traffic intensity within a50-m distance, traffic intensity in a circulararea between 50 and 250 m, and the popula-tion density within a 300-m radius and in acircular area between 300 and 5,000 m to pre-dict concentrations. Models with similar vari-ables describing traffic intensity were developedfor PM2.5 and NO2. These models explained73, 81, and 85% of the variability in theannual average concentrations for PM2.5, light-absorbing carbon, and NO2, respectively, inthe Netherlands (Brauer et al. 2003) and 56,67, and 62% of the variability in the annualaverage concentrations for PM2.5, light-absorb-ing carbon, and NO2, respectively, inGermany (Gehring et al. 2002). We thenapplied these models to the same geographicvariables measured at the home addresses ofeach individual in the cohort to obtain uniquelong-term ambient air pollution concentrationsat the home address of each cohort member atthe time of birth. Given our interest in early-life exposures, we estimated air pollutant con-centrations only for the birth addresses ofcohort members. However, in both locationsonly a small (9% in the Netherlands, 11% inMunch) percentage of the study populationmoved within the first year of life (Brauer et al.2002; Gehring et al. 2002).Questionnaire data. We assessed informa-tion on otitis media using a parent-completedquestionnaire. Specifically, in the Netherlandsthe question “Did a doctor diagnose infectionof the middle ear in your child in the last 12months?” was asked when the child was 12months and 24 months of age. In Munich, thequestion “Did a doctor diagnose otitis media inyour child during the last 6 months?” was askedevery 6 months. A series of potential confound-ing variables (listed in Table 1) were selected ifexploratory data analysis suggested substantialvariability within the cohort or if variables weresuspected of being risk factors for otitis media.Confounder data were selected from the earliestquestionnaire that was available, to coincidewith the exposure data that were estimated foraddresses at birth. ETS exposure was assessedby questionnaire (Table 1) and has been vali-dated against home nicotine measurements forthe Dutch cohort (Brunekreef et al. 2000).Statistical analysis. We performed multiplelogistic regression analyses to analyze the rela-tionship between otitis media and estimated airpollution exposure for study subjects in each ofthe cohorts. Results are presented as crude andadjusted odds ratios (ORs) with 95% confi-dence intervals (CIs). We adjusted for potentialconfounding factors such as sex, parental atopy,maternal education, siblings, maternal smokingduring pregnancy, ETS exposure at home, useof gas for cooking, indoor moulds and damp-ness, number of siblings, breast-feeding, andpresence of pets in the home (Table 1). ORs arepresented for standardized (between the twostudy locations) differences (approximatinginterquartile range differences in the twolocations) in estimated exposures of 3 µg/m3for PM2.5, 0.5 × 10–5/m for particle lightabsorbance [corresponding to ~ 0.5 µg/m3 ele-mental carbon, based on colocated sampling(Cyrys et al. 2003)], and 10 µg/m3 for NO2.ResultsIn both cohorts, the prevalence of otitis mediaincreased in the second year and was nearlyidentical at 2 years of age (Table 1). By thatage, approximately 35% of each cohort had atleast one occurrence of otitis media. Thecohorts were similar with respect to maternalage and parental allergy/atopy. The rate ofbreast-feeding was much higher in the Germancohort, whereas ETS exposure, the use of gasfor cooking, and the presence of pets weremuch higher in the Dutch cohort. TheGerman cohort had a somewhat higher level ofeducation and a lower number of siblings.Exposures to traffic-related air pollutionare summarized in Table 2. Median and meanconcentrations of light-absorbing carbon andNO2 were similar between the Netherlandsand Germany whereas PM2.5 concentrationswere higher in the Netherlands, probably dueto higher regional background concentrations.Interquartile ranges were nearly twice as largein the Netherlands (3.2 µg/m3, 0.54 10–5/m,and 16.4 µg/m3 for PM2.5, light-absorbingcarbon, and NO2, respectively) than inMunich (1.5 µg/m3, 0.34 10–5/m, and 8.5µg/m3), reflecting the fact that the Dutchcohort included suburban and semirural areasthat were not heavily affected by traffic-relatedair pollutants, whereas the German cohort wasrestricted to the Munich metropolitan area inwhich traffic emissions contribute to theurban background as well as to variabilitywithin the area.Crude ORs indicated elevated risks for oti-tis media in association with all air pollutants,with associations reaching statistical signifi-cance in the larger Dutch cohort (Table 3).These ORs were similar for otitis media in thefirst year of life and for cumulative incidenceover the first 2 years. ORs increased slightlybut were largely unaffected by adjusting forcovariates in both cohorts. Given the differ-ences in prevalence of covariates between thetwo cohorts, this suggests sufficient control forpotential confounding variables or that thesecovariates did not have a major impact in thisanalysis. Further, we specifically investigatedthe impact of ETS exposure in both cohortsbut did not observe any association betweenETS exposure or smoking during pregnancyand otitis media, and adjustment for theseexposures exposure did not change the associa-tions between air pollution and otitis media.Because attendance at a child care facilityhas been independently associated with respira-tory tract infections in the Dutch cohort(Koopman et al. 2001) and is also a risk factorfor otitis media occurrence (Bluestone andKlein 2001), we conducted a sensitivity analysisto evaluate its potential impact on the associa-tion with air pollution. Because child careattendance in Munich was very low—3% inthe first year and < 15% at 2 years of age—andto retain similar analyses in both cohorts, wedid not include child care attendance in the pri-mary models. We therefore restricted the sensi-tivity analysis to the Dutch cohort that hadhigher levels of child care attendance. Of the981 children (26%) who reported attendingchild care at 1 year of age, the median numberof hours of child care attendance per week was18, and 735 of these children attended childcare > 10 hours per week. By 2 years of age,1,256 children (34%) had reported attendingchild care. Adjusted ORs for otitis media withmodels incorporating child care attendancewere somewhat reduced [e.g., for NO2 OR =1.13 (95% CI, 0.99–1.28) at 1 year and OR =1.10 (95% CI, 0.99–1.23) after 2 years of age]but still elevated. Additional stratified analysesindicate that the effect of air pollution on otitismedia occurrence was not restricted to thosechildren who attended child care.DiscussionThis analysis represents the first examinationof the relationship between air pollutionexposure and otitis media in a large cohortstudy. In two birth cohorts with a commonexposure assessment approach, we have identi-fied associations between individual estimatesof traffic-related air pollution exposure and theincidence of otitis media. Given the wide-spread nature of air pollution exposure and thehigh prevalence of otitis media, these findingsindicate an important and previously unrecog-nized societal impact of air pollution. Further,given the high direct and indirect costs associ-ated with otitis media episodes, our resultssuggest a potentially important preventablerisk factor for this common childhood disease.In contrast to limited previous analyses thathave been conducted with small study popula-tions and have focused largely on cross-sectionalcomparisons between different geographicregions (Heinrich and Raghuyamshi 2004),we assessed exposure at the individual levelwith individual control for covariates. Perhapsthe strongest prior evidence relating air pollu-tion with otitis media comes from one of ourearlier studies in which prevalence rates forotitis media among 7,000 school-age childrenin three East German areas (two polluted andone control area) were compared in repeatedcross-sectional surveys (Heinrich et al. 2002).In addition, temporal changes of prevalencerates for otitis were studied in parallel withdramatic improvements in air quality [sulfurdioxide and total suspended particles (TSP)]after German reunification (Heinrich et al.2000). Although adjusted lifetime prevalenceBrauer et al.1416 VOLUME 114 | NUMBER 9 | September 2006 • Environmental Health PerspectivesTable 1. Prevalence of otitis media and selected potential confounders in the two cohorts.The Netherlands Munich, GermanyVariable (timing of questionnaire) n/N (%) n/N (%)Otitis mediaaAt 1 year of age 667/3,714 18.0 108/665 16.2At 2 years of age (cumulative) 1,262/3,650 34.6 226/650 34.8Confounding variablesbFemale sex 1,899/3,934 48.3 328/673 48.7Parental atopy 420/662 63.4Maternal allergy 1,281/4,114 31.1Paternal allergy 1,172/4,051 28.9Mother smoking during pregnancy 583/4,079 14.3 100/646 15.5Breast-feeding (first 3 months exclusively) 472/669 70.6Breast-feeding (any at 3 months of age) 1,827/3,883 47.0ETS at home (3 months or 6 months of age)c 1,121/3,903 28.7 97/669 14.5Maternal age at birth [median years (range)] 30 (17–42) 33 (19–44)Maternal education< 12 grades 1,743/3,709 47.0 219/670 32.7≥ 12 grades 1,966/3,709 53.0 451/670 67.3Siblings (at time of birth) 1,986/3,919 50.7 277/673 41.2Child care attendanceAt 1 year of age 981/3,734 26.3 22/667 3.3At 2 years of age (cumulative) 1,256/3,723 33.7 90/658 13.7Use of gas for cooking (at 3 months of age) 3,236/3,911 82.7 89/672 13.2Home dampness (at 3 months of age) 35/672 5.2Indoor molds (at 3 months of age) 1,215/3,704 32.8 209/673 31.1Pets (at 3 months of age) 2,011/3,905 51.5 118/672 17.6Cat 1,283/3,905 32.8 53/671 7.9Dog 629/3,905 16.1 23/671 3.4N refers to the total number of subjects who provided information on the specific variable; n refers to the number of sub-jects answering affirmatively with respect to each variable. aOtitis media responses refer to specific questions as described in “Materials and Methods.” bAtopy (parental history ofasthma and/or hay fever and/or eczema) was self-reported in the Munich cohort, and in the Netherlands atopy was self-reported allergy or reporting of physician-diagnosed allergy to house dust, house dust mite, pets, or hay fever/rhinitis) inthe (expecting) mothers. In the Netherlands, indoor molds refers to the (self-reported) presence of molds, water damage,and visible moisture in any of four specified rooms. In Munich, indoor molds refers to self-reported molds or mildew ormoisture spots anywhere in the dwelling. cQuestion regarding exposure to ETS at home (“Does anyone smoke in yourhouse?”) was asked at 3 months of age in the Netherlands and at 6 months of age in Munich. Table 2. Distribution of estimated annual average air pollution concentrations for the home (birth) addressin the cohorts.The Netherlands Munich, GermanyPM2.5 Light-absorbing carbon NO2 PM2.5 Light-absorbing carbon NO2(µg/m3) (PM2.5 absorbance, 10–5/m) (µg/m3) (µg/m3) (PM2.5 absorbance, 10–5/m) (µg/m3)Minimum 13.5 0.77 12.6 12.0 1.40 19.610th percentile 14.0 1.16 14.8 12.2 1.47 21.725th percentile 15.0 1.38 18.9 12.5 1.54 22.950th percentile 17.3 1.78 26.1 13.0 1.70 26.5Mean 16.9 1.72 25.6 13.4 1.76 27.775th percentile 18.2 1.92 29.2 14.0 1.88 31.490th percentile 19.1 2.19 35.3 14.8 2.10 34.8Maximum 25.2 3.68 58.4 21.9 4.39 64.4rates for otitis did not differ among childrenwho grew up in areas with different levels ofthese air pollutants (Heinrich et al. 2002),there were significant increases in prevalence ofother nonallergic respiratory illnesses (bronchi-tis, frequent colds, sinusitis, cough) anddecreased lung function in children from thepolluted areas (Frye et al. 2003; Heinrich2003; Heinrich et al. 2002). In parallel withthe decreases in SO2 and TSP in all three areas,prevalence rates for otitis decreased from 31%in 1992–1993 to 26% in 1995–1996 and27% in 1998–1999 (Heinrich et al. 2002).The adjusted OR for a 50-µg/m3 change inTSP was 1.45 (95% CI, 0.89–2.37) and 1.42(95% CI, 0.94–2.15) for a 100-µg/m3 changein SO2 concentration (Heinrich et al. 2002).However, these increased risk estimates ofambient air pollutants for otitis media weredriven mainly by the temporal improvement ofair quality and therefore may also parallel otherunmeasured lifestyle changes, whereas theregional gradient of air pollution concentra-tions were not consistent with area-specific dif-ferences in prevalence rates. For several othernonallergic respiratory outcomes, temporalchanges of prevalence rate (decreasing) wereconsistent with spatial differences (highestprevalence rate in the most polluted area).In a somewhat similar design, ear infec-tions were examined in two cross-sectionalsurveys of approximately 400 children 11–13years of age from three districts of São Paulo,Brazil, conducted in 1986 and in 1998(Ribeiro and Cardoso 2003). The three dis-tricts experienced different levels of SO2 in1973–1983, and these differences were associ-ated with crude prevalence rates (unadjustedfor potential confounders) for current and fre-quent ear infection. Social indicators such asparental education and literacy also differedbetween the three districts. Further, temporalchanges in ambient particulate matter concen-trations were also associated with changes infrequent ear infection prevalence. Althoughthis study has several limitations (small samplesizes, descriptive presentations of methods andmain findings), the results are consistent withan association between prevalence of otitismedia and ambient air pollution.Other studies that have evaluated airpollution and otitis media typically lack expo-sure measurements or were conducted onsmall sample sizes (Caceres Udina et al. 2004;da Costa et al. 2004; Dostal et al. 2001;Holtby et al. 1997) but also suggest associa-tions. For example, a study of 1,156 childrenwith OME in the United Kingdom did notmeasure exposures but used distances of thehome of these children from known industrialemission points as an exposure proxy (Holtbyet al. 1997). A significantly greater portion ofstudy entrants with OME lived within a1,000-m buffer of an industrial point source,but no trend of decreasing prevalence rate ofOME with increasing distance was observed.Although our study has several majoradvantages over previous investigations (indi-vidual level exposure assessment and largestudy population), there are also inherent lim-itations to our approach. First, as is commonwith air pollution epidemiologic analyses, weestimated exposures instead of directly meas-uring them using personal monitoring.Additionally, for those children who movedor attended child care, estimating exposures atthese locations may reduce exposure misclassi-fication. Second, we asssessed otitis media byquestionnaire-based self-reporting of physi-cian diagnosis and not by any objective mea-sure. Third, we did not assess severity oraddress issues such as recurrent otitis mediaand interactions between air pollution anddifferent treatment regimes. Such analyses,although important to understanding thepublic health significance of our findings, areprobably best addressed in a longitudinalstudy design in which detailed informationregarding otitis media occurrence and treat-ment is evaluated in relation to short-termchanges in air pollution exposure. Finally,although we adjusted analyses for a largenumber of potential risk factors, the possibil-ity for residual confounding remains, espe-cially given that exposure estimates werebased on spatial contrasts in air pollution thatmay also lead to spatial contrasts for otherunmeasured otitis media risk factors. Theselimitations are, however, largely unavoidablefor cohort studies of large populations.These findings indicate an associationbetween exposure to traffic-related air pollu-tants and the incidence of otitis media. Thisassociation is supported by a wealth of evi-dence linking exposure to high levels of airpollution indoors in developing countries withacute lower respiratory infections (Smith et al.2000), more limited evidence of associationsbetween levels outdoor air pollution in devel-oped countries and upper respiratory tractinfections (Chauhan and Johnston 2003; Linet al. 2005; Romieu et al. 2002), and the factthat some upper respiratory tract infectionsmay progress to otitis media (Rovers et al.2004). The strong evidence linking otitismedia with ETS exposure and the similaritiesbetween ETS and ambient air pollution addfurther support to our findings. The specificair pollutants that affect respiratory infectionshave not been clearly identified, althoughsome evidence suggests that NO2 and coarseparticles may be especially active in this regard(Chauhan and Johnston 2003; Lin et al.2005). Additionally, the mechanism by whichair pollution may lead to otitis media is notknown. Air pollution exposure may result in amore severe or persistent infection—for exam-ple, by decreasing mucociliary clearance(Chauhan and Johnston 2003; Thomas andZelikoff 1999)—making progression to otitismedia more likely. For example, an interac-tion between respiratory syncytial virus infec-tion and NO2 exposure before infection hasbeen demonstrated to lead to increased sever-ity of asthma exacerbations (Chauhan et al.2003). Alternatively, air pollution may activelypromote progression to otitis media.Addressing these or other possibilities willrequire further research. Although replicationof our results in similar cohort studies isneeded, the ubiquitous nature of air pollutionexposure and the importance of otitis media tochildren’s health suggest that these findingshave significant public health implications.Air pollution and otitis mediaEnvironmental Health Perspectives • VOLUME 114 | NUMBER 9 | September 2006 1417Table 3. Association between long-term exposure to air pollution and otitis and respiratory infections in the two cohorts: crude and adjusted ORs and 95% CIs.The Netherlands Munich, GermanyUnadjusted Adjusteda Unadjusted AdjustedaOtitis media OR (95% CI) N OR (95% CI) N OR (95% CI) N OR (95% CI) NAt 1 year of agePM2.5 1.13 (1.00–1.29)* 3,705 1.13 (0.98–1.32) 2,984 1.09 (0.68–1.75) 665 1.19 (0.73–1.92) 620Light-absorbing carbon 1.11 (1.00–1.23)* 3,705 1.11 (0.98–1.26) 2,984 1.07 (0.80–1.44) 665 1.12 (0.83–1.51) 620NO2 1.14 (1.02–1.27)* 3,705 1.17 (1.03–1.34)* 2,984 1.03 (0.74–1.43) 665 1.09 (0.78–1.54) 620At 2 years of age (cumulative)PM2.5 1.10 (0.99–1.22) 3,642 1.13 (1.00–1.27)* 2,970 1.18 (0.81–1.75) 650 1.24 (0.84–1.83) 605Light-absorbing carbon 1.08 (0.99–1.18) 3,642 1.10 (1.00–1.22)* 2,970 1.08 (0.85–1.37) 650 1.10 (0.86–1.41) 605NO2 1.10 (1.01–1.21)* 3,642 1.14 (1.03–1.27)* 2,970 1.10 (0.85–1.42) 650 1.14 (0.87–1.49) 605ORs are calculated as described in “Materials and Methods.”aAdjusted for mother smoking during pregnancy, ETS exposure, mother’s/father’s education, sex, gas for cooking/heating, siblings, breast-feeding, molds, pets, parental allergy,mother’s age; in the Netherlands only, adjusted for ethnicity, study arm (intervention/natural history), and use of allergen-impermeable mattress cover. *Statistically significant elevatedORs (p < 0.05). Brauer et al.1418 VOLUME 114 | NUMBER 9 | September 2006 • Environmental Health PerspectivesREFERENCESBluestone CD, Klein JO. 2001. Otitis Media in Infants andChildren. 3rd ed. Philadelphia:W.B. Saunders.Bondy J, Berman S, Glazner J, Lezotte D. 2000. Direct expendi-tures related to otitis media diagnoses: extrapolations froma pediatric medicaid cohort. Pediatrics 105(6):72–78.Brauer M, Hoek G, van Vliet P, Meliefste K, Fischer P, Gehring U,et al. 2003. Estimating long-term average particulate air pollu-tion concentrations: application of traffic indicators and geo-graphic information systems. Epidemiology 14(2):228–239.Brauer M, Hoek G, Van Vliet P, Meliefste K, Fischer PH,Wijga A, et al. 2002. Air pollution from traffic and thedevelopment of respiratory infections and asthmatic andallergic symptoms in children. Am J Respir Crit Care Med166(8):1092–1098.Brouwer CN, Rovers MM, Maille AR, Veenhoven RH, GrobbeeDE, Sanders EA, et al. 2005. The impact of recurrent acuteotitis media on the quality of life of children and their care-givers. Clin Otolaryngol 30(3):258–265.Brunekreef B, Leaderer BP, van Strien R, Oldenwening M, SmitHA, Koopman L, et al. 2000. Using nicotine measurementsand parental reports to assess indoor air: the PIAMA birthcohort study. Prevention and Incidence of Asthma andMite Allergy. Epidemiology 11(3):350–352.Brunekreef B, Smit J, de Jongste J, Neijens H, Gerritsen J,Postma D, et al. 2002. The Prevention and Incidence ofAsthma and Mite Allergy (PIAMA) birth cohort study: designand first results. Pediatr Allergy Immunol 13(suppl 15):55–60.Caceres Udina MJ, Alvarez Martinez JA, Argente del Castillo J,Chumilla Valderas MA, Fernandez Alvarez E, Garrido RomeraA, et al. 2004. Incidence, air pollution and risk factors ofacute otitis media in the first year of life: a prospective study.An Pediatr (Barc) 60(2):133–138.Chauhan AJ, Inskip HM, Linaker CH, Smith S, Schreiber J,Johnston SL, et al. 2003. Exposure to nitrogen dioxide(NO2) and the severity of virus-induced asthma in children.Lancet 361(9373):1939–1944.Chauhan AJ, Johnston SL. 2003. Air pollution and infection inrespiratory illness. Br Med Bull 68:95–112.Coyte PC, Asche CV, Elden LM. 1999. The economic cost of otitismedia in Canada. Int J Pediatr Otorhinolaryngol 49(1):27–36.Cyrys J, Heinrich J, Hoek G, Meliefste K, Lewne M, Gehring U,et al. 2003. Comparison between different traffic-relatedparticle indicators: elemental carbon (EC), PM2.5 mass, andabsorbance. J Expo Anal Environ Epidemiol 13(2):134–143.da Costa JL, Navarro A, Neves JB, Martin M. 2004. Householdwood and charcoal smoke increases risk of otitis media inchildhood in Maputo. Int J Epidemiol 33(3):573–578.Dostal M, Hertz-Picciotto I, James R, Keller J, Dejmek J,Selevan S, et al. 2001. Childhood morbidity and air pollutionin the Teplice program. Cas Lek Cesk 25;140(21):658–661.Freid VM, Makuc DM, Rooks RN. 1998. Ambulatory health carevisits by children: principal diagnosis and place of visit.Vital Health Stat 13 137:1–23.Frye C, Hoelscher B, Cyrys J, Wjst M, Wichmann HE, Heinrich J.2003. Association of lung function with declining ambientair pollution. Environ Health Perspect 111:383–387.Gehring U, Cyrys J, Sedlmeir G, Brunekreef B, Bellander T,Fischer P, et al. 2002. Traffic-related air pollution and res-piratory health during the first 2 yrs of life. Eur Respir J19(4):690–698.Heinrich J. 2003. Nonallergic respiratory morbidity improvedalong with a decline of traditional air pollution levels: areview. Eur Respir J 21(suppl 40):64s–69s. Heinrich J, Hoelscher B, Frye C, Meyer I, Pitz M, Cyrys J, et al.2002. Improved air quality in reunified Germany anddecreases in respiratory symptoms. Epidemiology13(4):394–401.Heinrich J, Hoelscher B, Wichmann HE. 2000. Decline of ambi-ent air pollution and respiratory symptoms in children. AmJ Respir Crit Care Med 161(6):1930–1936.Heinrich J, Raghuyamshi VS. 2004. Air pollution and otitismedia: a review of evidence from epidemiologic studies.Curr Allergy Asthma Rep 4(4):302–309.Hoek G, Meliefste K, Cyrys J, Lewné M, Brauer M, Fischer P, et al.2002. Spatial variability of fine particle concentrations in threeEuropean countries. Atmos Environ 36(25):4077–4088.Holtby I, Elliott K, Kumar U. 1997. Is there a relationship betweenproximity to industry and the occurrence of otitis mediawith effusion in school entrant children? Public Health111(2):89–91.Jinot J, Bayard S. 1996. Respiratory health effects of exposureto environmental tobacco smoke Exposure. Rev EnvironHealth 11(3):89–100. Klein JO. 2000. The burden of otitis media. Vaccine 19:S2–S8.Koopman LP, Smit HA, Heijnen ML, Wijga A, van Strien RT,Kerkhof M, et al. 2001. Respiratory infections in infants:interaction of parental allergy, child care, and siblings—the PIAMA study. Pediatrics 108(4):943–948.Kunzli N, Kaiser R, Medina S, Studnicka M, Chanel O, Filliger P,et al. 2000. Public-health impact of outdoor and traffic-related air pollution: a European assessment. Lancet356(9232):795–801.Lakwijk N, Van Strien RT, Doekes G, Brunekreef B, Gerritsen J.1998. Validation of a screening questionnaire for atopywith serum IgE tests in a population of pregnant Dutchwomen. Clin Exp Allergy 28(4):454–458.Lanphear BP, Byrd RS, Auinger P, Hall CB. 1997. Increasingprevalence of recurrent otitis media among children in theUnited States. Pediatrics 99(3):1–7.Lin M, Stieb DM, Chen Y. 2005. Coarse particulate matter andhospitalization for respiratory infections in children youngerthan 15 years in Toronto: a case-crossover analysis.Pediatrics 16(2):235–240.National Cancer Institute. 1999. Health Effects of Exposure toEnvironmental Tobacco Smoke: The Report of theCalifornia Environmental Protection Agency. NIH 9-4645.Bethesda, MD:National Cancer Institute.Ribeiro H, Cardoso MR. 2003. Air pollution and children’s healthin São Paulo (1986–1998). Soc Sci Med 57(11):2013–2022.Romieu I, Samet JM, Smith KR, Bruce N. 2002. Outdoor air pol-lution and acute respiratory infections among children indeveloping countries. J Occup Environ Med 44:640–649.Rovers MM, Schilder AG, Zielhuis GA, Rosenfeld RM. 2004.Otitis media. Lancet 363(9407):465–473.Smith KR, Samet JM, Romieu I, Bruce N. 2000. Indoor air pollu-tion in developing countries and acute lower respiratoryinfections in children. Thorax 55(6):518–532.Strachan DP, Cook DG. 1998. Health effects of passive smok-ing. 4. Parental smoking, middle ear disease and adeno-tonsillectomy in children. Thorax 53(1):50–56.Teele DW, Klein JO, Chase C, Menyuk P, Rosner BA, and theGreater Boston Otitis Media Study Group. 1990. Otitismedia in infancy and intellectual ability, school achieve-ment, speech, and language at age 7 years. J Infect Dis162:685–694.Thomas PT, Zelikoff JT. 1999. Air pollutants: modulators of pul-monary host resistance against infection. In: Air Pollutionand Health (Holgate ST, Samet JM, Koren HS, MaynardRL, eds). New York:Academic Press, 357–379.U.S. Department of Health and Human Services. 1986. TheHealth Consequences of Involuntary Smoking: A Report ofthe Surgeon General. DHHS Publication No. [CDC] 87-8398.Washington, DC:U.S. Department of Health and HumanServices.Wijga A, Smit HA, Brunekreef B, Gerritsen J, Kerkhof M,Koopman LP, et al. 2001. Are children at high familial riskof developing allergy born into a low risk environment?The PIAMA Birth Cohort Study. Prevention and Incidenceof Asthma and Mite Allergy. Clin Exp Allergy 31(4):576–581.


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items