Environmental Health Respiratory Symptoms in Children Living near Busy Roads and Their Relationship to Vehicular Traffic: Results of an Italian Multicenter Study (sidria 2)

Background: Epidemiological studies have provided evidence that exposure to vehicular traffic increases the prevalence of respiratory symptoms and may exacerbate pre-existing asthma in children. Self-reported exposure to road traffic has been questioned as a reliable measurement of exposure to air pollutants. The aim of this study was to investigate whether there were specific effects of cars and trucks traffic on current asthma symptoms (i.e. wheezing) and cough or phlegm, and to examine the validity of self-reported traffic exposure. Published: 18 June 2009 Environmental Health 2009, 8:27 doi:10.1186/1476-069X-8-27 Received: 9 April 2008 Accepted: 18 June 2009 This article is available from: http://www.ehjournal.net/content/8/1/27 © 2009 Migliore et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Environmental Health 2009, 8:27 http://www.ehjournal.net/content/8/1/27 Page 2 of 16 (page number not for citation purposes) Methods: The survey was conducted in 2002 in 12 centers in Northern, Center and Southern Italy, different in size, climate, latitude and level of urbanization. Standardized questionnaires filled in by parents were used to collect information on health outcomes and exposure to traffic among 33,632 6–7 and 13–14 years old children and adolescents. Three questions on traffic exposure were asked: the traffic in the zone of residence, the frequency of truck and of car traffic in the street of residence. The presence of a possible response bias for the self-reported traffic was evaluated using external validation (comparison with measurements of traffic flow in the city of Turin) and internal validations (matching by census block, in the cities of Turin, Milan and Rome). Results: Overall traffic density was weakly associated with asthma symptoms but there was a stronger association with cough or phlegm (high traffic density OR = 1.24; 95% CI: 1.04, 1.49). Car and truck traffic were independently associated with cough or phlegm. The results of the external validation did not support the existence of a reporting bias for the observed associations, for all the self-reported traffic indicators examined. The internal validations showed that the observed association between traffic density in the zone of residence and respiratory symptoms did not appear to be explained by an over reporting of traffic by parents of symptomatic subjects. Conclusion: Children living in zones with intense traffic are at higher risk for respiratory effects. Since population characteristics are specific, the results of validation of studies on self-reported traffic exposure can not be generalized. Background Vehicular traffic is a major source of outdoor air pollution. Several studies have reported associations between exposure to traffic pollutants in the zone of residence and increased frequency of respiratory tract illnesses [1-10]. The specific role of diesel exhaust from heavy traffic has been suggested in some of these studies [4,5,7], and airway inflammation due to exposure to diesel exhaust seems the likely biological mechanism [11,12]. With respect to the type of respiratory disorder, consistent associations have been found between exposure to traffic fumes and bronchitis symptoms, while the role of these exposures in the etiology of asthma is still unclear [13]. The SIDRIA project – Studi Italiani sui Disturbi Respiratori nell'Infanzia e l'Ambiente [14-16] – is a large multicenter, population-based study conducted in the framework of the International Study of Asthma and Allergies in Childhood (ISAAC) [17]. The findings from SIDRIA-1, conducted in 1994–1995, showed a positive association between indicators of air pollution from heavy vehicular traffic in the street of residence and a wide range of respiratory disorders in children living in highly urbanized areas [18]. However, it was not possible to assess the independent effects of car and truck traffic in that study as the information was not available. A second phase of SIDRIA was conducted in 2002 to evaluate time trends in the prevalence of respiratory disorders in childhood, according to the ISAAC Phase III protocol [19], to confirm the role of several potential risk factors identified in SIDRIA-1 and to explore some new hypotheses. Self-reported exposure to road traffic has been recently questioned as a reliable measurement of exposure to air pollutants [20-22]. Heinrich et al [20] compared parental report of traffic intensity at the home address with a combination of air pollution measures and GIS based modelled exposure in two different European countries (Germany and the Netherlands). They found that the degree of agreement between the two methods was relatively low, but the reasons for these discrepancies were not analyzed. Kuehni et al [21] estimated the association between selfreported exposure to road traffic and respiratory symptoms in preschool children in Leicestershire, UK, and investigated whether the effect could have been caused by reporting bias. The association between traffic exposure and respiratory outcomes was assessed using unconditional logistic regression and conditional regression models (matching by postcode). Matched analysis comparing symptomatic and asymptomatic children living at the same postcode (thus theoretically exposed to similar road traffic) suggested that parents of children with respiratory symptoms reported more road traffic than parents of asymptomatic children. In this paper, based on the data from SIDRIA-2, we present an analysis of the relation between indicators of road traffic pollution and several chronic respiratory symptoms while evaluating the potential effect of information bias. Methods Population and study design The SIDRIA-2 study design has been described elsewhere [15,16,23] and it will be only summarized here. The survey was conducted in 2002, between January and May, in Environmental Health 2009, 8:27 http://www.ehjournal.net/content/8/1/27 Page 3 of 16 (page number not for citation purposes) 12 centers (Bari, Colleferro, Emilia-Romagna, Empoli, Florence, Mantova, Milan, Palermo, Rome, Siena, Turin, Trento) of Northern, Center and Southern Italy, different in size, climate, latitude and level of urbanization. Eight of these centers had already participated in SIDRIA-1 [18]. The protocol of the study was approved by Ethics Committee of the Catholic University in Rome. The sample included 22,442 children (6–7 years old) and 16,336 adolescents (13–14 years old) attending respectively the first two grades of the primary school and the last year of the middle school. The primary sampling units were schools, both public and private, weighted for the number of attending subjects. Each center contributed with at least 1,000 subjects for each age group. Data collection To collect information on the medical history of the children, we used standardized, self administered questionnaires that included also the relevant ISAAC-Phase III questions on asthma, rhinitis and eczema symptoms, and questions on various known or suspected risk factors. For children (6–7 years old), all questionnaires were completed by parents. According to the standard ISAAC protocol [17], the questionnaires for adolescents were filled by adolescents themselves. However, in the SIDRIA study, another questionnaire, including questions on both symptoms and risk factors, was also completed by the adolescents' parents. For reasons of consistency and comparability between age groups, the current analyses are based on the parental questionnaires for both age-groups. Children with "asthma symptoms" were defined as those reporting in the past 12 months at least one of the following: one or more wheezing episodes, wheeze with exercise, morning chest tightness; or if they reported night dry cough in the last 12 months and had a reporting of lifetime asthma; or those reporting treatment for medically diagnosed asthma or had a hospital admission for asthma in the last 12 months; or if they reported a life time asthma and a positive answer to the question "Is your child still suffering from asthma?". Children were defined as having "severe asthma" if in the past 12 months at least one of the following were reported: 4 or more wheezing attacks, waking at night with wheezing one or more times a week, an attack severe enough to limit speech to only one or two words at a time between breaths, or a hospital admission for asthma. Children were defined as having "cough or phlegm" if they reported cough or phlegm for at least 4 days a week (in the absence of a cold) for one or more months a year. Questions on traffic included a parental subjective evaluation of traffic density in the zone of residence ("absent", "low", "moderate" or "high") and of the daily frequency of passing cars and trucks in the street of residence ("never or seldom", "sometimes", "frequently", "continuously"). The exact wording is reported in the Additional file 1. Data analyses Odds ratio (OR) and 95% confidence intervals (95% CI) were estimated with multiple logistic regression analyses. The basic analyses were for "asthma symptoms" and "cough or phlegm", but we also conducted multinomial logistic regression analyses comparing: i) those with asthma symptoms without cough or phlegm; ii) those with cough or phlegm without asthma symptoms; and iii) those with the combination of the two conditions, with the remaining subjects, negative for both symptoms. For "cough or phlegm", the analyses focused on the possible independent effect of cars and trucks transit in the street of residence. In some of the analyses, in order to obtain sufficient numbers in each category, car transit was recoded into three categories ("absent/sometimes", "frequently" and "continuously") and truck transit was also recoded into three categories ("absent", "sometimes" and "frequently/continuously"). Potential confounding factors included in the multiple logistic regression models were: sex, age, parental asthma or allergy (rhinitis or eczema), parental education (higher educational level between parents as a proxy of socioeconomic status – SES), passive smoke at home (at least one smoker-mother, father or others – in the household), indoor mould/dampness, season, person filling the questionnaire, floor of the apartment, change of residence and study area. Robust estimate of standard errors were used to control for clustering within schools [24]. In order to explore the role of potential effect modifiers, statistical significance of appropriate interaction terms were evaluated through likelihood-ratio tests. Eventually we performed subgroup analyses for different factors (age, gender, latitude, parental education, smoking, parental asthma or allergies, level of urbanization, indoor mould/ dampness, floor of the apartment, change of residence), in each case considering the associations of truck traffic exposure (frequent or continuous vs never) and the investigated respiratory symptoms. All analyses were conducted using STATA 9 (Stata Corporation, College Station, Texas). Data validation External validation In the city of Turin data on traffic flows were obtained from the local company of public transportation (5T S.C.R.L), that measures hourly traffic mostly in street segments with medium-high volume of traffic. The address of each subject included in the survey was geocoded and linked to data on traffic using a Geographical Information Environmental Health 2009, 8:27 http://www.ehjournal.net/content/8/1/27 Page 4 of 16 (page number not for citation purposes) System (GIS). We excluded 204 subjects (out of 3,453) attending a school in the municipality but resident outside the city of Turin, for which traffic information and geocoding was not available. For each segment of street we calculated average daily vehicle count in work hours (h.07–19) weighted by the number of observations available; this information was linked to the subject using a computer based GIS, excluding subjects living in internal civic numbers. We positively matched 887 subjects out of 3,249 (328 street segments out of 1067). All GIS analyses were performed with the ArcGis software version 9 (ESRI, Redlands, California, USA). We analyzed the frequency distribution for some characteristics of the subjects, including reported traffic, separately for children linked with traffic count and for non-linked subjects. We performed the Kruskal-Wallis test to assess the differences between median traffic counts and calculated Jackknife confidence intervals for the median of the differences [25]. The same analysis was performed for subjects with and without a reporting of current respiratory symptoms, separately for asthmatic symptoms and cough or phlegm (ANOVA) [26]. Internal validation In the 3 metropolitan areas of Turin, Milan and Rome (N = 10,285), we matched the children by census block. The assumption is that within this small area (for example, in Turin a census block covers a mean of 250 subjects – [27]) the true exposure to road traffic in an urban context would be similar; therefore, if the association between traffic exposure and symptoms is true, it should disappear when the comparisons are stratified by census block (matched analysis). For this analysis, only data from census blocks where at least one symptomatic and one asymptomatic subject were living could be used; therefore, a varying number of census block was included in the analysis for each of the respiratory symptoms examined. To further investigate the possibility of a reporting bias, in the subgroup of subjects that could be matched by census block we performed a further analysis having attributed to symptomatic subjects the mean traffic exposure reported by parents of asymptomatic subjects for the same census block; in this way, we theoretically excluded the possibility of a bias due to over-reporting of traffic by parents of symptomatic subjects. In these analyses, reported life time asthma was also included as an endpoint. Results The parental questionnaire was completed for 20,016 6– 7 years old (response rate 89.2%) and for 13,616 13–14 years old (response rate 83.3%). Table 1 shows the combined prevalence of asthma symptoms and cough or phlegm. Overall, 13.5% of children (95% CI: 13.2%, 13.9%) were reported to have asthma symptoms, and 6.8% (95% CI: 6.6%, 7.1%) to have cough or phlegm. For 2.9% (95% CI: 2.7%, 3.0%) of the subjects, cough or phlegm and asthma symptoms were concomitant. The frequency of each traffic indicator is presented in the Additional file 2. The prevalences of respiratory symptoms in the various subgroups are shown in Table 2. Respiratory symptoms were more frequent if a parental history of asthma or allergies was reported, among subjects exposed to passive tobacco smoke and in the presence of moulds/dampness in the child's bedroom. The prevalence of respiratory symptoms increased inversely with parental education, and directly with urbanization level. Table 3 shows the associations of traffic indicators with asthma symptoms and with cough or phlegm. Reported high traffic density, continuous car transit and continuous truck transit in the street of residence were weakly associated with asthma symptoms. There were stronger associations of reported high traffic density, continuous car transit and continuous truck transit with cough or phlegm. The associations were generally stronger for truck transit than for car transit. For comparison with the literature, we also present the results of the associations for different asthma symptoms (see Additional file 3). In the same table, we also report the associations separately for light and severe asthma symptoms, and for symptoms of Table 1: Prevalence of respiratory outcomes investigated in a sample of Italian schoolchildren.