Adverse effect of air pollution on respiratory health of primary school children in Taiwan.

This study is a part of the Study On Air Pollution and Health In Taiwan (SOAP&HIT), an ongoing research project involving cooperation of several universities in Taiwan. In this study, the objective was to evaluate the effects of ambient air pollution on respiratory symptoms and diseases of school children, in addition to considering indoor air pollution. Six communities were selected: one community located in a rural area (Taihsi), two in urban areas (Keelung and Sanchung), and the other three in petrochemical industrial areas (Toufen, Jenwu, and Linyuan). We sampled 5,072 primary school students in six communities from the main study population of SOAP&HIT. Respiratory health was assessed by evaluation of the children's respiratory symptoms and diseases using a parent-completed questionnaire. Data were analyzed using logistic regression analysis to compute odds ratios of adverse effect. The school children in the urban communities had significantly more respiratory symptoms (day or night cough, chronic cough, shortness of breath, and nasal symptoms) and diseases (sinusitis, wheezing or asthma, allergic rhinitis, and bronchitis) when compared with those living in the rural community. However, only nasal symptoms of children living in the petrochemical communities were more prevalent than in those living in the rural community. Although the association with ambient air pollution is suggestive, the cross-sectional study cannot confirm a causal relationship; thus further studies are needed.

the objctive was.. to evlt ef :c.tofa:mbient air pollution on resir ptoms and dises of hool cil to consider.ingindor ar Xuon . comm ies were selected: one communit located in a rurl area (Ta itwo in area e nd Sanchung), and the oer ee in pet eie i a Jin and Linuan).
We sampled 5,072 primary school students in sit communities from the main y population of SOAP&HIT. Respirtor eat was assessd by evautinof th childrens respiraoy symptoms and diseases using a pant-cpeted qusionaire. D .at wr alted usg ls-.
shortness of breath, and nasl sympt ) wa d diseases (sinuitis, wheezing or astma, alergic rhinitis, and ctis) w ose ing inhe rty Hoeve, only nsal sdof ch n in h e p et c it ere moe prvlent than in those living i .D,lthe rurl comuity.Altog th associationwit amhiet airplltFutiion.is sgestiv, the cross-etonal.stuycnnot cofirm a causlreltionship thius further studies are needed. Ky auord. air poluin cross-secinl std, home dmns,respiratory helh scho;ol cidre.EvirnHalh Pmic 106:33135 (1998), (Olin 1lMay 19981 hap :llebpnvsL niehs.nih.gev/daes/1998/106p331-S35ebealabnracat html In the last few years, several studies have reported significant associations between air pollution and daily mortality and various markers for acute respiratory morbidity, including hospital admissions, hospital emergency and outpatient clinic visits, exacerbation of respiratory symptoms, lung functions changes, and school absenteeism in Europe and America (1)(2)(3)(4)(5). However, there are still many issues to be clarified before we know the real causal relationship between air pollution and health effects. Specific air pollutants have not been identified as pauses of health effects, and effects of short-term peak air pollution exposures are not sufficiendy evident.
Also, there is no agreement on the economic impact of air pollution-related health effects. To develop the exposure-effect relationship between air pollution and health indices, we cooperated in an ongoing research project, The Study On Air Pollution and Health In Taiwan (SOAP&HIT). This is the first multicenter (five research institutes), multidisciplinary (health, exposure assessments, statistics, and economics), and multiyear (1994)(1995)(1996)(1997) cross-sectional air pollution study in Taiwan. The SOAP&HIT seeks to address three critical hypotheses: 1) What are the exposure-effect relationships between air pollutant mixtures and various adverse respiratory effects based on population and individualized air pollution exposure data? 2) What are the economic costs for air pollution-related health effects based on the dose-response relationship found in our research? and 3) Is there a subpopulation that is particularly susceptible to air pollution in Taiwan? To conduct SOAP&HIT efficiently, our disciplinary integrated study design was further classified into four components: exposure assessments, health effects, biological markers, and economic analysis. For efficiency of study design, we decided to use concentrations of three major air pollutants, SO2, 03, and particulate matter <10 pm (PM0o), in the year before this study as indicators to select study sites. A total of eight (23) communities were preliminarily selected based on these three pollutants using a factorial design.
However, we only selected six out of these communities due to financial constraints. These six sites included different pollution patterns representing rural, urban, and industrial areas in Taiwan.
The objective of the present study was to determine whether there was an adverse effect of air pollution on respiratory health in primary school children. This paper characterizes the respiratory symptoms and diseases of school children in urban and petrochemical communities and in a reference community, which is in a rural area and not in the immediate proximity of any petrochemical or other industrial emissions during the study period.

Materials and Methods
Study design. A cross-sectional survey was designed to determine the prevalence of respiratory symptoms and diseases of primary school children residing in two urban communities exposed to pollution caused by traffic (Keelung and Sanchung), and three petrochemical communities (Toufen, Jenwu, and Linyuan), which serve as the exposed group of stationary sources by petrochemical industries. These communities were compared to a rural community, Taihsi, which serves as the reference group in Taiwan (Table 1). According to population characteristics and air pollution patterns (6), the three petrochemical communities were classified further: both Toufen and Jenwu also had traffic pollution, the populations are ethnically Hakka in Toufen and Taiwanese in Jenwu, and Linynan represents typical petrochemical pollution. Subjects. Among 5,072 primary school students selected from the six communities, a total of 4,860 returned the questionnaire, for a response rate of 95.8%. Of this group of children, 81.6% had resided in their current houses for at least 3 years. Questionnaires with incomplete data (136) were excluded for the purpose of consistency throughout the analyses. A total of 4,697 valid questionnaires are presented in this study ( Protection Administration air quality moni-UV fluorescence for SO2; chemiluminestoring stations located in the campus of the cence for NOX, NO, and NO2; flame ionselected primary school in each community. ization for THC and NMHC; non-disperse The instrumentation used in the fixed-site infrared absorption for CO; and UV absorpmonitoring stations are Il-gauge for PM10; tion for 03. The air quality was measured by comparing the air pollution levels with appropriate National Ambient Air Quality Standards (NAAQS) for each air pollutant.
Two-year monitoring data from 1994 to 1995 are included in this study. In the preliminary analysis, the six public schools were similar with respect to indoor environment and timetable for teaching activity. Questionnaire. A respiratory health questionnaire was distributed through the schools to the children in December 1994. The questionnaire was completed by parent or guardian and returned to the school. One nurse for each school distributed and collected the questionnaires. The questionnaire used in this study was mainly adapted and modified from the World Health Organization (WHO) childhood respiratory questionnaire (7). The questionnaire was divided into the following five parts: demographic data, respiratory symptoms and diseases of the children, housing conditions, children's bedrooms, and possible sources of indoor air pollution such as household smoking, pets, fowls, coal stove used, tea gas-cooker used, incense burning the whole day, mosquito repellent burning, indoor plants, and home dampness. The dose gradient for home dampness was defined as the presence of any visible mold or its stuffy odor in the rainy season (usually about 3 months/year between April and June) only, when it is doudy or rains on several consecutive days only, or on most days.
In this paper, we defined morning cough as cough usually in the morning, day or night cough as cough usually in the rest of the day or at night, and chronic cough as the previous symptoms on most days for 3 consecutive months or more during the year. We defined shortness of breath as breathlessness when hurrying on the level or walking up a slight hill and wheezing as a wheezy or whistling sound in the chest at any time. Nasal symptoms were defined as stuffy and running nose or sneezing in the morning or before sleeping, apart from colds for 2 or 3 consecutive days or more. Self-reported respiratory diseases such as sinusitis, wheezing or asthma, allergic rhinitis, bronchitis, and pneumonia were included in a history of respiratory disease diagnosed by medical doctors. Statistics. We analyzed data concerning the prevalence of respiratory symptoms and diseases according to place of residence. To explore the relationship between respiratory symptoms and diseases and the exposure areas, multiple logistic regression models were used in which the potential confounding factors were controlled. The adjusted odds ratios (aORs) and their 95% confidence intervals (CIs) were computed. All statistical analyses were performed using the SPSS for Windows, Release 6.1. (8,9). Table 3, both mean concentrations and exceedance frequencies were significantly different for most air pollutants among the three community types. For two mixed-type communities (Toufen and Jenwu) and the typical petrochemical community (Linyuan), their average SO2 concentrations (16-46 ppb) were significantly higher than rural (3 ppb) and urban (9 ppb) communities. As expected, the average concentrations of CO in two urban communities (0.9-1.4 ppm) were significantly higher than the concentrations in both the rural (0.4 ppm) and the typical petrochemical (0.7 ppm) communities. For NO2, the average concentrations in rural communities (10 ppb) were significantly lower than those in typical petrochemical (20 ppb), urban The basic information of the study population is shown in Table 2. With regard to the children's sex, there was no  significant difference among those living in different communities. The proportion of children (44.5%) who were in high grades (grades 5 and 6) in the typical petrochemical industrial community (Linyuan) was slightly higher than proportions in the other communities. However, there were apparent differences in education of the children's parents. For example, education of fathers in Linyuan community was the lowest level on average of those in the six communities.

As indicated in
The reported prevalences of respiratory symptoms and diseases are summarized in Table 4. Children living in the urban area (Keelung and Sanchung) had consistently higher rates of respiratory symptoms and diseases (with the exception of pneumonia) than did those living in the rural community (Taihsi). However, nasal symptoms were more prevalent in children living in the petrochemical communities (Toufen, Jenwu, and Linyuan) than in the rural community. For morning cough and day or night cough, there was a steady rise with the Hakka (Toufen), mixed Janwu), and typical (Linyuan) petrochemical industrial communities. Table 5 shows the conditions of the indoor environment of the study subjects. There were obvious differences among children living in different communities with Environmental Health Perspectives * Volume 106, Number 6, June 1998 regard to the level of crowding, the amount of household smoking in the dwelling, the presence of feathered or hairy pets, fowls kept in the home, the use of a coal stove, the use of a tea gas-cooker in the home, the burning of incense the whole day, the burning of a mosquito repellent, plants inside the home, and self-reported home dampness. Children in Taihsi, Jenwu, and Linyuan lived in more crowded households than did those in the other communities.
The amount of household smoking in Taihsi and Linyuan was higher than in the other communities. Home dampness was reported more frequently in Keelung.
The results of the logistic regression analysis are shown in Table 6. The school children in the urban communities (Keelung and Linyuan) had higher risks of having respiratory symptoms (day or night cough, chronic cough, shortness of breath, and nasal symptoms) or diseases (sinusitis, wheezing or asthma, allergic rhinitis, and bronchitis) when compared with those living in the rural community (Taihsi). For example, compared to children in the rural community, children in the urban communities had a higher risk of having day or night cough (aOR = 1.67; CI, 1.21-2.29), chronic cough (aOR = 1.88; CI, 1.07-3.31), shortness of breath (aOR = 1.40; CI, 1.04-1.91), and nasal symptoms (aOR = 1.96; 95% CI, 1.54-2.48). On the other hand, children living in the petrochemical communities Jenwu and Linyuan) had higher odds of having nasal symptoms than did those living in the rural community. Compared to children in the rural community, children in the petrochemical industrial communities had a higher risk ofhaving nasal symptoms (mixed type: aOR = 1.59; CI, 1.21-2.09 and typical: aOR = 1.59; CI, 1.19-2.14).
Among indicators of indoor environmnent, home dampness was the most significant factor in relation to children's respiratory health. The crowding index, household smoking, pets, fowls, use of a coal stove, use of a tea gas-cooker, incense burning the whole day, mosquito repellent burning, and indoor plants are also significantly associated with some of respiratory symptoms and diseases. There were dose-response relationships between home dampness and most of the respiratory symptoms and diseases after controlling for the potential confounders mentioned above (see Table 7). For example, compared to children whose parent did not report any sign of dampness in the home, children with reported home dampness had a higher risk of having morning cough, which increased with an increase in dampness (dampness in rainy season: aOR = 1.62; CI, 1.32-1.98; dampness with continuous

Discussion
The school children in the urban communities had significantly more respiratory symptoms (day or night cough, chronic cough, shortness of breath, and nasal symptoms) and diseases (sinusitis, wheezing or asthma, allergic rhinitis, and bronchitis) when compared with those living in the rural community. On the other hand, nasal symptoms were more prevalent in children living in the petrochemical communities than in those living in the rural community. When evaluating the effects of individual sources of pollution, an epidemiologic approach best eliminates the effects of other variables (10). Unfortunately, such an approach is impossible in this cross-sectional survey. In this study, we evaluated the health effects in six communities with different pollution patterns, based on community monitoring (6). We used a spatial approach in which multivariate statistical analyses were performed to control for possible confounding factors.
Because there are several types of health care in Taiwan, including Western, Chinese, and herbal medicine and self-medication bought from drug stores, we could not use medical records from general practice only. Thus, we selected respiratory symptoms of primary school children as indices of health effects resulting from air pollution. Parents of the subjects completed questionnaires concerning the status of the children's respiratory health. The limitations of this type of research tool have been reviewed (11), and reporting errors may have occurred in this study. All questions concerning respiratory health were selected from a standardized questionnaire (7). This cross-sectional study offers weak evidence for causality because one cannot be confident that exposure preceded the disease. Most of the children in this study had lived in their current houses for at least 3 years, although the etiologically meaningful time between exposure and occurrence of respiratory symptoms and disease is still a matter of debate. The overall high response rate achieved in the present study increases confidence in the data and makes an important selection bias unlikely. Also, we collected the data in the same month in order to avoid differential inclusion of study subjects into the survey according to season, which can bias the results.
The study communities included rural, urban, and petrochemical industrial areas and therefore may have shown apparent differences in socioeconomic status. In this study, the father's education at a primary school level gave lower odds for chronic cough, allergic rhinitis, and bronchitis (more severe respiratory conditions) than those at higher educational levels. We conduded that these respiratory conditions in questionnaires were underreported by parents or guardians at low levels of education. The father's education as a surrogate was induded in the final models to control for differences of healthcare seeking behavior and disease perception.
Hakka people (Toufen) are known for their hard work and high tolerance to difficult or hazardous environments. They may be less likely to express their complaints of bodily discomfort or they may have a higher threshold of discomfort to consider it a symptom. Our results in Tables 4 and 6 show this trend before and after controlling for indoor air pollution, compared to the Volume 106, Number 6, June 1998 X Environmental Health Perspectives Table 7. Adjusted odds ratios (95% confidence intervals) for respiratory symptoms and diseases by home dampness Respiratory condition In raining season When continuously raining On most days Symptoms Morning cough other two schools located in petrochemical industrial area.
The study communities were located in different regions of Taiwan and therefore exhibited differences in indoor air pollution. However, it is unlikely that there were major disparities in indoor air pollution because data on a number of known or potential determinants of indoor air pollution were collected and adjusted for, including household smoking in the dwelling and exposure to other major potential sources.
The observed high prevalence of home dampness and mold indicated that dampness in the home was very common in the subtropical region studied, and home dampness was a strong predictor of respiratory symptoms (12). The prevalence of most of the reported respiratory conditions was found to be higher (day or night cough, chronic cough, shortness of breath, and bronchitis significantly so) among children whose fathers or mothers were smokers, compared with children of nonsmoking parents. The effect of parental smoking is in accordance with the literature (13). This study was cross-sectional in design and, although it shows that respiratory symptoms and diseases were significantly more common among children from the urban communities, it may not be attributed to a specific pollutant. Particulate SO2, NO2, and acid aerosols have been associated with increased prevalences of these respiratory symptoms in children (14-1. The concentrations of NO2 and CO in these two urban communities were apparently higher than those in the rural community. Thus, exposure to NO2 may contribute to increased risks of respiratory diseases, but the result is probably also related to traffic pollution mixtures (18,19).
On the other hand, only irritant symptoms, such as nasal symptoms other than colds, were significantly more prevalent among children from the petrochemical communities where SO2 was evidently higher than in the rural community. These pollutants may explain the higher prevalences of nasal symptoms in the petrochemical communities; however, increased levels of one pollutant (such as SO2) are often associated with high concentrations of other pollutants, such as particulate and acid aerosols. Therefore, the contribution of each compound to impaired respiratory health remains inconclusive.
Exposure to a mixture of air pollutants from mobile sources may be associated with adverse respiratory health of the children living in the urban areas, but air pollutants predominantly from the petrochemical industrial areas may only induce nasal irritation symptoms. Although the association with known air pollutants is suggestive, a cross-sectional study cannot confirm a causal relationship and further studies are needed to determine the exposure-effect relationship between individualized air pollution exposure and various adverse respiratory effects.