Decline of Ambient Air Pollution and Respiratory ... - ATS Journals

Download PDF
7 downloads 150 Views 80KB Size Report
Comment
East Germany during a phase of strong improvement in ambient air quality. .... addressed socioeconomic factors, medical history of the child, paren- tal atopy ...
Decline of Ambient Air Pollution and Respiratory Symptoms in Children JOACHIM HEINRICH, BERND HOELSCHER, and H. ERICH WICHMANN Forschungszentrum für Umwelt und Gesundheit, Institut fuer Epidemiologie, and Lehrstuhl für Epidemiologie, Institut für Medizinische Informationsverarbeitung, Biometrie und Epidemiologie, Ludwig-Maximilians Universitaet München, Neuherberg, Germany

Several regional cross-sectional studies have shown a consistently higher prevalence of respiratory disorders in children with exposure to total suspended particulates (TSP) than in children living in less polluted areas. The aim of the present study was to investigate the temporal changes in the prevalence of nonasthmatic respiratory symptoms and diseases in children living in three areas of East Germany during a phase of strong improvement in ambient air quality. Groups of 2,470 and 2,814 school children between 5 and 14 yr, respectively, participated in two regional cross-sectional studies in 1992–1993 and 1995–1996. In the three areas (Hettstedt, Bitterfeld, and Zerbst) examined in the study, the annual mean TSP decreased from 65, 48, and 44 ␮g/m3, respectively, in 1993 to 43, 39, and 36 ␮g/m3 in 1995. In the same time interval, the crude prevalence of bronchitis in the three respective areas decreased from 62%, 52%, and 50% to 47%, 40%, and 39%. During the 3-yr period between the two regional studies, prevalence decreased significantly for bronchitis (odds ratio [OR]: 0.55; confidence interval [CI]: 0.49 to 0.62), for otitis media (OR: 0.83; CI: 0.73 to 0.96), for frequent colds (OR: 0.74; CI: 0.64 to 0.86), and for febrile infections (OR: 0.76; CI: 0.66 to 0.88) after adjustment for several potential predictors. In conclusion, we found that the prevalence of nonasthmatic respiratory symptoms decreased from the first period to the second period in all three study areas.

Several regional cross-sectional studies done in the United States and Europe have shown consistently higher rates of bronchitis and bronchitic symptoms in children with high exposure to total suspended particulates (TSP) than in children living in less polluted areas (1–4). Recently published reviews of health effects of air pollution (5, 6) reported chronic adverse health effects even at relatively low levels of ambient particulates currently measured in urban areas. In the relatively short period since German reunification in 1990, the ambient sulfur dioxide (SO2) and TSP in East Germany have declined tremendously (7). If ambient TSP were associated with a higher prevalence of respiratory symptoms in children, one would expect a decrease in the prevalence of these symptoms corresponding to a decline in the suspended particle concentration. We examined this relationship by repeated examination of children living in East Germany in 1992–1993 and 1995–1996.

In 1992, the population of Bitterfeld county was approximately 117,400 (9). Ambient air pollution in Bitterfeld county was caused by emissions from chemical and power plants. Factories that formerly operated in Bitterfeld county produced several chemical products that were emitted into the local environment through leakage and insufficient air filtration. Furthermore, power plants burned local brown coal with a sulfur content up to 4%. The predominant air-borne pollutants were SO2, particulates, nitrogen oxides (NOx), and halogenated hydrocarbons. Hettstedt county is situated approximately 100 km west of the town of Bitterfeld. It includes the towns of Hettstedt and Mansfeld and two other villages, and had a population of approximately 52,000 in 1992. Before 1990, Hettstedt was a center for the mining and smelting of nonferrous metals, primarily copper, dating as far back as the 12th century. The study area in Hettstedt county was situated in a valley along a small river, that is surrounded by hills less than 70 m high. Air pollution in Hettstedt county is due to heavy metal-containing dust emissions from the county’s smelters in the past, and to the domestic burning of brown coal. Zerbst county, with approximately 36,800 inhabitants, includes the two towns of Zerbst and Loburg and four villages, and was selected as the control area for our study. The town of Zerbst is the agricultural and administrative center of the county and is located approximately 60 km north of Bitterfeld. Emissions in Zerbst county are minimal, and ambient air pollution is limited to combustion products from domestic heating with brown coal. Because Zerbst is a mountainless county, any pollutants emitted from these domestic sources are widely distributed. Large differences existed between the three counties with respect to industrial emissions of particles, SO2 and NOX, although all three counties experienced air pollution to some degree from domestic burning of brown coal (8). After the German reunification, the previously large amounts of SO2 and TSP emissions decreased greatly, especially during the years 1990 and 1991. The closure of most plants and the replacement of brown coal by gas used for domestic heating contributed to the decrease in SO2 and TSP. The three study areas share similar climatic conditions.

Subjects

(Received in original form June 22, 1999 and in revised form December 8, 1999)

We invited all first-, third-, and sixth-grade school children who were residents of Zerbst and Hettstedt counties, and a subgroup of children from selected schools in Bitterfeld county, to participate in our study. All day-care centers and schools in the Zerbst and Hettstedt areas were contacted. Because of the larger area of Bitterfeld county, schools and day care centers in this county were randomly selected to represent all resident school-age children. Since the population in all three areas is very homogenous, all participants were white. Repeated cross-sectional studies were conducted in the years 1992–1993, and 1995–1996. In each survey, three age groups of school children (school entrants, third-graders, and sixth-graders) were examined. School entrants and third-graders in the first survey, in 1992–1993, were reexamined at the second survey in 1995–1996. The study protocol was approved by the University of Rostock Ethics Committee. Informed consent was obtained from the parents of all participating children.

Supported exclusively by grant 298 61 724 from the Federal Environmental Agency of Germany.

Questionnaire

Correspondence and requests for reprints should be addressed to Dr. Joachim Heinrich, GSF-Institute of Epidemiology, P.O. Box 1129, D-85758 Oberschleissheim, Germany. E-mail: [email protected] Am J Respir Crit Care Med Vol 161. pp 1930–1936, 2000 Internet address: www.atsjournals.org

As part of the study, teachers distributed questionnaires to the children’s parents, and collected completed questionnaires a week later. The questionnaire contained 78 items (8), and had been previously used and tested in several national and international studies, and further adapted to address east German living and housing conditions. It

METHODS Study Area and Sources of Air Pollution The study (8) was conducted in Bitterfeld county, in the State of Sachsen-Anhalt (formerly part of the German Democratic Republic). Bitterfeld county includes the towns of Bitterfeld and Wolfen and four adjoining villages located in a mountainless region of Sachsen-Anhalt.

1931

Heinrich, Hoelscher, and Wichmann, et al.: Air Pollution and Respiratory Health addressed socioeconomic factors, medical history of the child, parental atopy (whether the father or mother experienced asthma, hay fever, or eczema), respiratory and allergic diseases and symptoms, breast feeding history, attendance at day-care centers, bedroom sharing, crowding, housing characteristics, nutrition, indoor environmental exposures such as environmental tobacco smoke (ETS) and dampness, and contact to pets. The following health outcomes were analyzed in our study: parent-reported bronchitis and otitis media ever diagnosed by a physician; parent-reported frequent common colds (more than two within the previous 12 mo); and frequent febrile infections (more than common cold accompanied by fever within the previous 12 mo), and common cough in the morning during the fall and winter seasons (see APPENDIX).

Ambient Pollution Ambient pollutants were monitored by state authorities at all three sites for each investigated community from 1992 through 1996 (10–15). Both particulate matter of less than 10 ␮m in aerodynamic diameter (PM10) and of less than 2.5 ␮m (PM2.5), as well as black smoke (BS) were measured additionally for 6 mo by monitoring sites that were established for the study. Table 1 compares annual means of gaseous pollutants, particulate matter (PM), and black smoke in the three study areas. SO2 was measured with an Ansyco Model AF 21 M pulsed fluorescence analyzer (Environnement, Poissy, France). TSP was measured with an FH 62 IN ␤-ray absorption monitor (FAG Kugelfischer, Schweinfurt, Germany). The detailed methods for these analyses, and especially for measurements of BS and PM2.5 are described elsewhere (8, 16). The annual means of ambient air pollutants were calculated on the basis of hourly and daily measures (Table 1). Missing values in Table 1 are due to incomplete air monitoring. Comparisons of the two polluted regions with Zerbst show that the Hettstedt area had the highest annual mean concentration of TSP, whereas the Bitterfeld area had the highest annual mean concentration of SO2 (Table 2).

After the German reunification in 1990, mean SO2 and PM levels strongly decreased in Sachsen-Anhalt (10–15).

Statistical Methods Logistic regression was used to analyze spatial and temporal patterns of bronchitis, otitis media, and respiratory symptoms. A large set of possible predictors of respiratory health were considered in the analyses. Three different odds ratios (ORs) for changes with time (second versus first survey) are presented. The first is a crude OR, the second is adjusted for town of residence only, and the third is adjusted for all covariates. Because a reasonable fraction of all children participated in both surveys, their responses tended to be correlated, which had to be considered in the analyses. This could be achieved with marginal models, which assume the same form for the expectations and variances of the response as in logistic regression, but additionally incorporate a correlation structure for repeated observations of the same child. Estimates for temporal–spatial changes in the study measures and the corresponding standard errors, which account for the correlations, were computed with the generalized estimating equations (GEE) approach of Liang and Zeger (17). Values of p from the Wald statistic of the logistic model were reported, in order to compare several factors potentially influencing the development of airway diseases between the first and the second survey. Only observations without missing values for any of the covariates were used for the statistical modeling (n ⫽ 3,787). All computations were made with the GENMOD procedure of the SAS software package version 6.12 (SAS Institute, Cary, NC).

RESULTS Participation Rates and Demographics

Parents of 2,470 of 2,773 eligible children between 5 and 14 yr of age completed a questionnaire (response rate: 89.1%) in the first

TABLE 1 ANNUAL MEAN OF SO2 PARTICULATES, AND BLACK SMOKE AT THE THREE AMBIENT MONITORING STATIONS IN ZERBST (CONTROL), BITTERFELD (POLLUTED), AND HETTSTEDT (POLLUTED): 1991 TO 1996 Ambient Concentration SO2, ␮g/m3 1991 1992 1993 1994 1995 1996 Dust fall, mg/m2/d 1991 1992 1993 1994 1995 1996 TSP, ␮g/m3 1991 1992 1993 1994 1995 1996 PM10, ␮g/m3 October 1993 to March 1994 PM2.5, ␮g/m3 January 1993 to June 1993 Black smoke, ␮g/m3 October 1993 to March 1994

Zerbst (Control)

(Reference No.)

Bitterfeld (Polluted)

(Reference No.)

Hettstedt (Polluted)

(Reference No.)

50* 53* 42 29 21 25

(11) (11) (12) (13) (14) (12)

130 83 66 44 41 25

(11) (11) (12) (13) (14) (12)

84 46 49 38 26 25

(11) (11) (12) (13) (14) (12)

110* 120* 60 80 70 60

(11) (11) (12) (13) (14) (15)

290–330 240–340 140–290 130–160 90 70

(11) (11) (12) (13) (14) (15)

240–380 130–180 140–210 110–140 110–120 80–90

(11) (11) (12) (13) (14) (15)

n.a. n.a. 44 34 36 37

(12) (13) (14) (15)

n.a. n.a. 48† 37 39 40

(12) (13) (14) (15)

n.a. n.a. 65 53 43 40

(12) (13) (14) (15)

34

(16)

n.a.

43

(16)

43‡

53‡

n.a.

41

45‡ 25

(16)

(16)

Definition of abbreviations: n.a. ⫽ not available; PM ⫽ particulate matter; TSP ⫽ total suspended particulates. * Data for Zerbst were not available; data presented were measured at a station in Stendal located in close proximity to Zerbst. † Data for Bitterfeld were not available; data presented were measured at a station in Greppin located in close proximity to Bitterfeld. ‡ Unpublished data.

1932

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE

VOL 161

2000

TABLE 2 RESPONSE, MOBILITY, AND CHARACTERISTICS OF THE STUDY POPULATION IN THE SURVEYS IN 1992–1993 AND 1995–1996 Survey 1 (1992–1993)

Respondents (questionnaire) Total Zerbst Bitterfeld Hettstedt Considered in analyses* Total Zerbst Bitterfeld Hettstedt Characteristics of participants Age group 5–7 yr Age group 8–10 yr Age group 11–14 yr Boys Higher parental education† Other characteristics (predictors for adjustment) Low birth weight No breast-feeding Parental atopy House built before 1960 House made of concrete Dwelling on ground floor Living space per person ⬎ 20 m2 Child shares bedroom Dampness or visible molds District or central heating Heating with coke/coal/briquettes Heating with gas Carpet in child’s room Current or prior ETS exposure at home Mother smoked during pregnancy Contact with cats Child ever attended day care center

Survey 2 (1995–1996)

Percent

Respondents/Total

Percent

Respondents/Total

89.1 88.3 87.0 92.1

2,470/2,773 857/971 799/918 814/884

74.7 68.6 80.4 72.8

2,814/3,765 748/1,090 1,255/1,561 811/1,114

84.2 84.2 80.3 88.2

2,335/2,773 818/971 737/918 780/884

67.4 61.0 71.7 66.6

2,536/3,765 675/1,090 1,119/1,561 742/1,114

31.1 32.1 36.8 50.7 42.2

726/2,335 750/2,335 859/2,335 1,184/2,335 958/2,272

25.3 34.1 40.5 53.6 42.2

642/2,536 866/2,536 1,028/2,536 1,360/2,536 1,020/2,417

6.5 20.3 26.6 49.2 40.1 41.5 43.7 48.8 18.9 50.6 36.8 7.0 93.3 57.8 5.4 32.6 89.1

146/2,240 471/2,321 620/2,335 1,110/2,257 901/2,245 948/2,284 944/2,158 1,121/2,297 438/2,312 1,178/2,328 840/2,284 160/2,284 2,173/2,329 1,336/2,312 123/2,295 750/2,300 2,062/2,314

6.1 18.5 30.7 49.3 38.8 39.5 52.9 41.0 22.9 61.5 16.3 17.6 81.8 53.3 6.1 32.3 88.7

145/2,390 465/2,517 779/2,536 1,217/2,470 956/2,466 1,000/2,529 1,235/2,335 1,036/2,529 577/2,524 1,499/2,438 409/2,510 442/2,510 2,042/2,497 1,339/2,511 152/2,496 817/2,529 2,233/2,518

Definition of abbreviation: ETS ⫽ environmental tobacco smoke. * Criteria for exclusion: Child lived less than 2 yr in current home and previous home is more than 2 km distant. † Education of father or mother at least 12 yr.

survey; 2,814 of 3,765 eligible children participated in the second survey (response rate: 74.7%) (Table 2). We excluded children from analyses if they had lived for less than 2 yr in their current home and if their previous home was located more than 2 km from their current home (n ⫽ 135 in the first survey; n ⫽ 278 in the second survey; Table 2). Data for 2,335 and 2,536 children, respectively, in each of the two surveys contributed to the descriptive analyses. In total, we obtained information from 4,871 questionnaires, which came from 2,929 children who participated only once and from 971 children who participated in both surveys. Children participating in one of the two surveys were quite comparable with regard to sex and parental education to those participating in the other survey, but children who participated in the first survey were slightly younger (Table 2). Changes in the age distribution of the study populations in the first and second survey were affected by a general decrease in birth rates after the German reunification, which fell by as much as 60% throughout eastern Germany from 1990 to 1991 (9). Temporal–Spatial Changes

Figure 1 shows the crude community-specific prevalence of bronchitis and frequent colds plotted against annual mean concentrations of TSP and SO2. In the three study areas (Hettstedt, Bitterfeld, and Zerbst), the annual mean TSP decreased from 65, 48, and 44 ␮g/m3 in 1993 to 43, 39, and 36 ␮g/

m3, respectively, in 1995. The largest decrease occurred in Hettstedt, the most highly polluted city in the study. During the same time interval, the prevalence of bronchitis in the study population decreased from 62%, 52%, and 50% in Hettstedt, Bitterfeld, and Zerbst, respectively, to 47%, 40%, and 39%. Within each single survey, the prevalence of bronchitis was lowest in the least polluted city with the lowest annual mean concentrations of TSP, whereas the association with SO2 was not as clear. Additionally, the prevalence of bronchitis decreased between the first and second surveys, but the ranking of the involved communities remained unchanged. The results for other symptoms were less clear, and TSP showed a stronger effect than did SO2. Temporal changes in several factors that potentially influence the development of airway diseases and respiratory symptoms were studied over time (Table 2). Housing characteristics changed from the first to the second surveys. The proportion of available living space exceeding 20 m2 per person increased (43.7% versus 52.9%, p ⬍ 0.001). The relative number of single houses heated by ovens burning coal, briquettes, or coke decreased from 36.8% to 16.3% (p ⬍ 0.001) within the 3-yr period from the first to the second survey. Homes heated centrally with gas increased substantially (7.0% versus 17.6%, p ⬍ 0.001). Dampness and mold in homes, however, was reported more often in the second survey than in the first survey

Heinrich, Hoelscher, and Wichmann, et al.: Air Pollution and Respiratory Health

(22.9% versus 18.9%, p ⬍ 0.001). Current or prior exposure of children to ETS in their homes was somewhat less frequently reported by parents in 1995–1996 (53.3% versus 57.8%, p ⫽ 0.006) than in 1992–1993. Contact with cats, attendance at daycare facilities, and breast feeding remained unchanged. A remarkable decrease in the prevalence of bronchitis (54.3% versus 41.5%), otitis media (30.9% versus 26.3%), frequent colds (37.8% versus 32.0%), and febrile infections (33.1% versus 27.8%) was found between the 1992–1993 survey and the 1995–1996 survey. The temporal changes were not pronounced for cough in the morning(14.1% versus 12.9%) or for current wheezing (10.8% versus 9.2%). After adjustment for several potential predictors, the decrease in the prevalence of bronchitis (OR: 0.55; CI: 0.49 to 0.62), otitis media (OR: 0.83; CI: 0.73 to 0.96), frequent colds (OR: 0.74; CI: 0.64 to 0.86), and febrile infections (OR: 0.76; CI: 0.66 to 0.88) from the first to the second survey remained significant. Similar results were found for cough in the morning (OR: 0.86; CI: 0.70 to 1.05), but were not statistically significant (Table 3). Furthermore, no statistically significant temporal changes were found for current wheezing (OR: 0.82; CI: 0.66 to 1.01) after adjustment for several potential predictors. The temporal changes in the prevalence of bronchitis, and to a lesser degree in the prevalence of febrile infections, were more pronounced for the community of Hettstedt, the area that was most polluted by particulate matter. Excluding those children who had incomplete data on covariates, which decreases the number of observations to 3,787, did not change the crude ORs (data not shown). Possible interactions between areas and time were analyzed. We observed no statistically significant differences for

1933

temporal effects among the three study areas except in the case of febrile infections. The temporal changes in the prevalence of respiratory illnesses followed similar trends in all three study areas, but the prevalence of these illnesses decreased most strongly in the most polluted community, Hettstedt. Further sensitivity analyses were done for selected indoor factors, by stratification according to living in damp or moldy homes, exposure to ETS, cooking with gas, and contact with cats. Children from Hettstedt, the most polluted area, who were not living in damp houses with visible molds, who were not exposed to ETS, who were not exposed to emissions by gas cooking, and who did not report contact with cats showed a stronger temporal decrease in the prevalence of common frequent colds, frequent colds accompanied by fever, and frequent cough than did children who lived in the less polluted areas of Zerbst and Bitterfeld (Figure 2). Except for coughing, children exposed to indoor air pollution sources did not show a similar relation between temporal changes and study area to that shown by children exposed to outdoor air pollution. Summarizing these data, a more pronounced decrease in respiratory disorders was found in children living in the polluted areas who were not exposed in their homes to the indoor pollutants listed earlier.

DISCUSSION Although several cross-sectional and cohort studies consistently showed a higher prevalence of respiratory illness in children with high exposure to ambient particulate matter, few data have been published on improvements in respiratory

Figure 1. Community-specific prevalence of bronchitis (lifetime prevalence) and frequent colds (more than three in the previous 12 mo) in 5- to 14yr-old children in 1992–1993 (solid circles, n ⫽ 2,335) and 1995–1996 (open triangles, n ⫽ 2,536) versus annual mean of TSP and SO2 in 1993 and 1995. (Children who lived less than 2 yr in current home and had a previous home more than 2 km distant were excluded from analyses.)

1934

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE

VOL 161

2000

TABLE 3 TEMPORAL AND SPATIAL CHANGES OF PREVALENCE OF BRONCHITIS, OTITIS MEDIA, FREQUENT COLDS* (⬎ 2, PAST YEAR) AND FEBRILE INFECTIONS† (⬎ 1, PAST YEAR) IN CHILDREN AGED 5 TO 14 yr BETWEEN SURVEYS IN 1992–1993 AND 1995–1996 Bronchitis (n ⫽ 3,778) OR

95% CI

Otitis media (n ⫽ 3,785) OR

Frequent colds (n ⫽ 3,686)

Febrile infections (n ⫽ 3,736)

95% CI

OR

95% CI

OR

95% CI

Temporal changes Survey 2 versus Survey 1 Crude Adjusted for area Adjusted for area and potential predictors (listed in this table) Potential predictors Girls Boys Age group 5–7 yr Age group 8–10 yr Age group 11–14 yr Parental education ⭐ 10 yr Parental education ⬎ 10 yr No low birth weight Low birth weight (⬍ 2,500 g) Breastfeeding No breastfeeding No parental atopy Parental atopy House built after 1960 House built before 1960 House not made of concrete House made of concrete Dwelling not on ground floor Dwelling on ground floor Living space per person ⭐ 20 m2 Living space per person ⬎ 20 m2 Child does not share bedroom Child shares bedroom Neither dampness nor visible molds Dampness or visible molds No district or central heating District or central heating No heating with coke/coal/briquettes Heating with coke/coal/briquettes No heating with gas Heating with gas No carpets in child’s home Carpet in child’s room No ETS exposure Current or prior ETS exposure Mother did not smoke during pregnancy Mother smoked during pregnancy No contact with cats Contact with cats Child never attended day care center Child ever attended day care center

0.59 0.60

(0.53, 0.66) (0.54, 0.66)

0.80 0.80

(0.71, 0.91) (0.71, 0.90)

0.73 0.74

(0.64, 0.83) (0.65, 0.84)

0.74 0.73

(0.64, 0.84) (0.63, 0.83)

0.55

(0.49, 0.62)

0.83

(0.73, 0.96)

0.74

(0.64, 0.86)

0.76

(0.66, 0.88)

1.00 1.15 1.00 1.02 1.15 1.00 1.47 1.00 1.06 1.00 0.74 1.00 1.18 1.00 1.15 1.00 1.44 1.00 0.95 1.00 0.87 1.00 1.00 1.00 1.40 1.00 0.99 1.00 0.83 1.00 0.87 1.00 0.86 1.00 1.02 1.00 1.06 1.00 0.91 1.00 1.47

(1.00, 1.33) (0.88, 1.18) (1.00, 1.35) (1.27, 1.70) (0.78, 1.43) (0.62, 0.89) (1.02, 1.37) (0.94, 1.41) (1.16, 1.79) (0.82, 1.10) (0.74, 1.02) (0.87, 1.16) (1.19, 1.65) (0.69, 1.42) (0.57, 1.19) (0.59, 1.29) (0.71, 1.04) (0.89, 1.18) (0.80, 1.43) (0.79, 1.05) (1.16, 1.85)

1.00 0.93 1.00 1.05 0.88 1.00 1.33 1.00 0.86 1.00 0.90 1.00 1.25 1.00 1.17 1.00 1.08 1.00 1.00 1.00 1.05 1.00 0.93 1.00 1.20 1.00 1.15 1.00 1.24 1.00 1.00 1.00 1.10 1.00 1.19 1.00 0.97 1.00 0.88 1.00 1.17

(0.85, 1.19)

(0.90, 1.50)

1.00 0.77 1.00 0.66 0.52 1.00 0.91 1.00 1.07 1.00 0.79 1.00 1.29 1.00 1.11 1.00 1.06 1.00 0.93 1.00 1.01 1.00 0.96 1.00 1.38 1.00 1.08 1.00 1.01 1.00 1.10 1.00 0.96 1.00 1.06 1.00 0.96 1.00 0.99 1.00 0.93

(0.80, 1.08) (0.90, 1.23) (0.74, 1.05) (1.14, 1.56) (0.61, 1.20) (0.73, 1.10) (1.07, 1.46) (0.94, 1.45) (0.85, 1.37) (0.85, 1.17) (0.88, 1.24) (0.79, 1.08) (1.02, 1.43) (0.75, 1.75) (0.80, 1.91) (0.63, 1.57) (0.87, 1.38) (1.02, 1.38) (0.71, 1.34) (0.75, 1.04)

(0.92, 1.30)

(0.74, 1.19)

1.00 0.91 1.00 0.67 0.46 1.00 0.93 1.00 1.30 1.00 0.75 1.00 1.03 1.00 1.01 1.00 1.17 1.00 1.12 1.00 1.09 1.00 1.11 1.00 1.27 1.00 0.89 1.00 0.89 1.00 0.86 1.00 0.96 1.00 1.11 1.00 0.76 1.00 1.08 1.00 0.74

(0.67, 0.89) (0.57, 0.78) (0.44, 0.61) (0.78, 1.05) (0.78, 1.47) (0.65, 0.97) (1.11, 1.50) (0.80, 1.38) (0.84, 1.34) (0.79, 1.08)

(0.83, 1.12) (1.16, 1.65) (0.73, 1.59) (0.68, 1.50) (0.73, 1.67) (0.77, 1.20) (0.91, 1.23) (0.69, 1.33) (0.84, 1.15)

(0.79, 1.05) (0.56, 0.80) (0.38, 0.54) (0.79, 1.08) (0.95, 1.77) (0.61, 0.93) (0.88, 1.21) (0.81, 1.27) (0.92, 1.49) (0.95, 1.31)

(0.95, 1.30) (1.06, 1.52) (0.60, 1.30) (0.60, 1.32) (0.56, 1.30) (0.76, 1.20) (0.95, 1.29) (0.53, 1.09) (0.92, 1.27) (0.58, 0.93)

Geographic area Bitterfeld versus Zerbst, adjusted for potential predictors listed in this table Hettstedt versus Zerbst, adjusted for potential predictors listed in this table

0.99

(0.83, 1.17)

0.92

(0.77, 1.11)

0.92

(0.77, 1.11)

1.30

(1.08, 1.56)

1.43

(1.19, 1.71)

0.84

(0.70, 1.02)

1.05

(0.88, 1.27)

1.38

(1.14, 1.67)

Definition of abbreviations: CI ⫽ confidence interval; ETS ⫽ environmental tobacco smoke; OR ⫽ odds ratio.

health paralleling decreases in ambient pollution levels. The tremendous decline of air pollution levels in east Germany within the relatively short period of from 1 to 3 yr after German reunification, allowed us to study changes in health status paralleling such an improvement in air quality. In summary, our study found a higher prevalence of nonasthmatic respiratory illness for children growing up in areas of increased levels of TSP and SO2 than in those growing up in the

less polluted control area. Furthermore, the prevalence of parent-reported, physician-diagnosed bronchitis and otitis media, parent-reported frequent colds, and febrile infections decreased in parallel with the reduction in annual mean TSP and SO2 between 1992–1993 and 1995–1996. Although we lacked power to show effects of additional exposures to several indoor sources, we observed a slight trend toward an improvement in respiratory health in Hettstedt, the most polluted study area (Figure 2).

Heinrich, Hoelscher, and Wichmann, et al.: Air Pollution and Respiratory Health

The decrease in nonasthmatic respiratory illnesses seems to have been stronger in children not exposed to dampness or emissions from gas use for cooking, ETS in the home, or contact with cats. This let us to conclude that children with less exposure to indoor pollutants are more likely to respond to a change in ambient pollutant levels, such as ambient levels of particulate matter. Adjustment for other potential risk factors did not have much impact on the ORs for the effect on respiratory diseases and symptoms over time of living in more polluted areas. The location of their residence was used as a proxy measure for long-term exposure of children to ambient pollutants. The overall mobility of the study population was extremely low. Nearly 50% of the children had spent their entire lives in one home. We believe that the low mobility of this eastern German population helped to reduce a potential diluting effect from migration. A limitation of our study was the small number of study areas involved. Communities could differ with respect to risk factors for health outcomes other than ambient pollutants, such as poverty status, access to health care, nutritional habits, or smoking. However, the three study areas are located in relative proximity to each other, and their inhabitants are homogenous with respect to ethnic composition, general social and lifestyle factors, and access to medical care. Temporal

1935

changes in nutritional habits before and after the German reunification were reported for adults, but changes did not continue to happen after the early 1990s (18). The annual mean level of TSP in our study ranged from 34 to 65 ␮g/m3, a range slightly narrower than the range of community differences reported in the Harvard Six cities study (1) (PM2.5: 20 to 59 ␮g/m3), and slightly wider than the range in the Swiss Study of Childhood Allergy and Respiratory Symptoms with Respect to Air Pollution (SCARPOL) (2) (PM10: 10 to 33 ␮g/m3). Furthermore, the levels of SO2 were substantially higher than in the Harvard and Swiss studies. Our data do not permit an evaluation of the time course of improvement in respiratory health in our study population. Involvement of only three communities, and the high correlation between all pollutants, makes it difficult to evaluate effects of a single pollutant, but TSP seems to have shown the strongest changes paralleling the change in prevalence of respiratory illness with time in both the cross-sectional and imbedded cohort parts of the study. This interpretation is in agreement with results of previous studies of the relationship between air pollution and nonasthmatic respiratory symptoms and diseases. A recently published review of studies done in central and eastern European countries reported positive associations between “traditional” air pollutants, such as SO2 and TSP, and the prev-

Figure 2. Adjusted ORs for temporal changes between 1992 and 1995 in the prevalence of respiratory disorders in children aged 5 to 14 yr (GEE models stratified for a total of 3,770 children, with [n ⫽ 3,144] and without [n ⫽ 626] one of the following indoor exposures: damp or moldy home [n ⫽ 808], ETS [n ⫽ 2,055], cooking with gas [n ⫽ 1,573], contact with cats [n ⫽ 1,203]). (A) Temporal–spatial changes in children with at least one of the listed indoor exposures (n ⫽ 3,144). (B) Temporal–spatial changes in children without any of the listed indoor exposures (n ⫽ 626).

1936

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE

alence of respiratory symptoms and diseases and reduced lung function in children (19). According to a review by Pope and coworkers (20), studies done in the United States with the most sophisticated analytic approaches showed clear adverse health effects of long-term exposure to particulate matter, and to a lesser degree to NO2 and O3. The 24-cities study done in the United States examined the health effects of a wide range of air pollutants, consisting of particulate matter (PM10 and PM2.5), sulfate, particle strong acidity, ozone, SO2, ammonia, and nitrous acid (21). A comparison of the respiratory health profiles of 13,000 children (8 to 12 yr of age) residing in these 24 communities showed that children living in communities with the highest particle strong acidity more often reported at least one episode of bronchitis in the preceding year (OR: 1.66). In addition, fine particulate sulfate was associated with higher reporting of bronchitis (OR: 1.65). No other respiratory symptoms were found to be associated with the air pollutants measured in the 24 studied cities. Recently published results of the Swiss SCARPOL study (2) provide further evidence that rates of respiratory illness and symptoms, but not of allergies, are associated with moderately increased levels of air pollution. Additionally, a recently published Polish cross-sectional study of 1,129 children reported an association between outdoor air pollutants and chronic phlegm (OR: 4.2) (3). Our findings are consistent with those of the Harvard Six Cities study (1) and the Swiss SCARPOL study (2), which also showed an association between air pollution and bronchitis, but not with wheezing. A major strength of the present study was that an identical questionnaire was used in both surveys, and that children and their parents were approached in the same manner. Yet the participation rate in the second survey was slightly lower. It did not appear that this lower participation rate caused selection bias: parental educational level, a marker for socioeconomic status, showed a similar distribution in both surveys. If the lower participation rate in the second survey introduced any bias, one would expect reports of a higher prevalence of respiratory disorders. It has been previously reported that young adults who are most willing to participate in a respiratory health survey tend to overreport respiratory symptoms (22). In conclusion, the prevalence of nonasthmatic respiratory symptoms and diseases was higher in children living in more polluted communities, especially with respect to TSP and SO2, suggesting that disease occurrence may be reduced within a short period by improvement of air quality. Acknowledgment : The authors wish to thank H. Schneller and K. Honig-Blum for data handling; G. Burmester, J. Rudzinski, B. Hollstein, H. Machander, D. Albrecht, and C. Boettcher for gathering regional data and local assistance; Dr. B. Ritz and A. Ibald-Mulli for critical reading of the manuscript; all teachers in Hettstedt, Zerbst, and Bitterfeld, and the local school authorities and health care centers for their support; and all parents and children for their participation.

References 1. Dockery, D. W., F. E. Speizer, D. O. Stram, J. H. Ware, J. D. Spengler, and B. G. Ferris, Jr. 1989. Effects of inhalable particles on respiratory health of children. Am. Rev. Respir. Dis. 139:587–594. 2. Braun-Fahrländer, J., C. Vuille, F. H. Sennhauser, U. Neu, T. Künzle, L. Grize, M. Gassner, C. Minder, C. Schindler, H. S. Varonier, B. Wüthrich, and the SCARPOL Team. 1997. Respiratory health and long-term exposure to air pollutants in Swiss schoolchildren. Am. J. Respir. Crit. Care Med. 155:1042–1049. 3. Jedrychowski, W., and E. Flak. 1998. Effects of air quality on chronic

4.

5.

6.

7. 8.

9. 10. 11. 12. 13. 14. 15. 16.

17. 18.

19.

20.

21.

22.

VOL 161

2000

respiratory symptoms adjusted for allergy among preadolescent children. Eur. Respir. J. 11:1312–1318. Peters, J., E. Avol, W. Navidi, S. London, W. Gaudermann, F. Lurmann, W. Linn, H. Margolis, E. Rappaport, H. Gong, Jr., and D. Thomas. 1999. A study of twelve southern California communities with differing levels and types of air pollution: I. Prevalence of respiratory morbidity. Am. J. Respir. Crit. Care Med. 159:760–767. Bascom, R., P. Bromberg, D. Costa, R. Devlin, D. Dockery, M. Frampton, W. Lambert, J. Samet, F. Speizer, and M. Utell. 1996. Health effects of outdoor pollution. Part 1. Am. J. Respir. Crit. Care Med. 153:3–50. Bascom, R., P. Bromberg, D. Costa, R. Devlin, D. Dockery, M. Frampton, W. Lambert, J. Samet, F. Speizer, and M. Utell. 1996. Health effects of outdoor pollution. Part 2. Am. J. Respir. Crit. Care Med. 153:477–498. Umweltbundesamt 1997 Daten zur Umwelt 1995/96. Erich Schmidt Verlag. Heinrich, J., B. Hoelscher, M. Wjst, B. Ritz, J. Berlin, and H. E. Wichmann. 1999. Respiratory diseases and allergies in two polluted areas in East Germany. Environ. Health Perspect. 107:53–62. Statistisches Bundesamt, Wiesbaden. 1994. Landesamt für Umweltschutz Sachsen-Anhalt. 1992. Immissionsschutzbericht. Report No. 8. Landesamt für Umweltschutz, Halle; 1993. Landesamt für Umweltschutz Sachsen-Anhalt. 1993. Immissionsschutzbericht. Report No. 12. Landesamt für Umweltschutz, Halle; 1994. Landesamt für Umweltschutz Sachsen-Anhalt. 1994. Immissionsschutzbericht. Report No. 17. Landesamt für Umweltschutz, Halle; 1995. Landesamt für Umweltschutz Sachsen-Anhalt. 1995. Immissionsschutzbericht. Report No. 19. Landesamt für Umweltschutz, Halle; 1996. Landesamt für Umweltschutz Sachsen-Anhalt. 1992. Immissionsschutzbericht. Report No. 22. Landesamt für Umweltschutz, Halle; 1997. Landesamt für Umweltschutz Sachsen-Anhalt. 1997. Immissionsschutzbericht. Report No. 27. Landesamt für Umweltschutz, Halle; 1998. Tuch, T., P. Brand, H. E. Wichmann, and J. Heyder. 1997. Variation of particle number and mass concentration in various size ranges of ambient aerosols in Eastern Germany. Atmospher. Envir. 31:4193–4197. Liang K. Y., and S. L. Zeger. 1986. Longitudinal data analysis using generalized linear models. Biometrika 73:13–22. Winkler, G., S. Brasche, and J. Heinrich. 1997. Trends in food intake in adults from the city of Erfurt before and after the German reunification. Ann. Nutr. Metab. 41:283–290. Jedrychowski, W. 1995. Review of recent studies from central and Eastern Europe associating respiratory health effects with high levels of exposure to “traditional” air pollutants. Environ. Health Perspect. 2:15– 21. Pope, C. A., III, D. V. Bates, and M. E. Raizenne. 1995. Health effects of particulate air pollution: time for reassessment? Environ. Health Perspect. 103:472–480. Dockery, D. W., J. Cunningham, A. J. Damokosh, L. M. Neas, J. D. Spengler, P. Koutrakis, J. H. Ware, M. Raizenne, and F. E. Speizer. 1996. Health effects of acid aerosols on North American children: respiratory symptoms. Environ. Health Perspect. 104:500–505. de Marco, R., G. Verlato, E. Zanolin, M. Bugiani, and J. W. Drane. 1994. Nonresponse bias in EC Respiratory Health Survey in Italy. Eur. Respir. J. 7:2139–2145.

APPENDIX Questions asked in respiratory health questionnaire. Has a doctor ever diagnosed your child with one of the following diseases: bronchitis, otitis media. How often did your child suffer from a common cold during the past 12 months? About __ times, not at all. How often did your child suffer from a common cold accompanied with fever? About __ times, not at all. Has your child had wheezing or whistling in the chest in the last 12 months? Never, once, several times, do not know. Did your child cough early in the morning on a regular basis during fall and winter season? Yes, no, do not know.