Cross-sectional studies of children in the United States and Europe consistently have shown higher rates of bronchitis and bronchitic symptoms in areas with higher exposure to total suspended particulates (TSP). 1–4 Recently published reviews of air pollution effects 5–7 reported associations with chronic adverse health effects even at the relatively low levels of ambient particulates currently measured in most urban areas.
Since German reunification in 1990, the levels of ambient sulfur dioxide (SO2) and TSP in eastern Germany have declined tremendously. 8 In Erfurt, eastern Germany, the concentrations of accumulation-mode particles (100–500 nm) have decreased, whereas those of nucleation-mode particles (10–30 nm in aerodynamic diameter) have increased. 9,10 In the three areas of this study (all in eastern Germany), we found similar trends in size-specific particle number concentrations between 1993 and 1999. 11 These trends may be related to declining emissions from stationary sources, to increasing emissions from mobile sources, and to clean air regulations that selectively removed larger particles. 9–11
If ambient TSP and SO2 levels are associated with bronchitic symptoms in children, a decline in these air pollution concentrations should produce a corresponding decrease in symptom prevalence. We examined this relationship in surveys of children living in eastern Germany.
We conducted three cross-sectional surveys of children age 5–14 years in the metropolitan areas of Bitterfeld, Hettstedt, and Zerbst in the State of Sachsen-Anhalt. All of these areas were formerly part of East Germany (the German Democratic Republic). 12,13 These surveys were conducted in 1992–1993, 1995–1996, and 1998–1999. Within these surveys, we constructed three longitudinal cohorts of children who participated in two or three consecutive surveys. The three study areas initially had different levels of air pollution, but shared similar weather conditions.
Before German reunification in 1990, ambient air pollution in the Bitterfeld metropolitan area was caused primarily by emissions from chemical plants and power plants. Local power plants burned soft brown coal with a sulfur content of up to 4%. The predominant airborne pollutants were SO2, particulate matter, nitrogen oxides, and halogenated hydrocarbons.
The Hettstedt metropolitan area is situated about 100 km west of Bitterfeld in a valley surrounded by hills nearly 70 meters high. Since the 12th century, Hettstedt has been a center for the mining and smelting of nonferrous metals, primarily copper. Air pollution sources are primarily heavy metal-containing dust emissions from the county’s smelters and the domestic burning of high-sulfur brown coal.
The Zerbst metropolitan area is the agricultural and administrative center of the county and is located about 60 km north of Bitterfeld. Air pollution emissions are limited to combustion products from domestic heating with brown coal and are widely distributed over a broad plain. The children in the Zerbst area served as a control population with (initially) lower exposures to ambient air pollutants.
Before reunification, there were large differences among these three study areas with respect to industrial emissions of particulate matter, SO2, and oxides of nitrogen. All three areas experienced some air pollution from the domestic burning of high-sulfur brown coal. 10 After German reunification, the closure of most industrial plants and the replacement of brown coal by gas for domestic heating led to declines in SO2 and TSP emissions, especially during 1990 and 1991. 13 A more detailed description of these study areas has been published elsewhere. 13
Ambient Pollution Monitoring
Routine monitoring of ambient pollutants at all three sites 14–16 was supplemented by a special mobile particle measurement site. 11 Whereas the communities of Hett-stedt and Zerbst each had a single fixed measurement site, the Bitterfeld area had three sites within a 6-km radius, for which values were averaged. Annual means of SO2 were available from 1991 to 1998 (except in Zerbst; 1993–1998). Annual means of TSP were available from 1993 to 1998 (except in Bitterfeld; 1994–1998).
The data suggested a roughly linear association between log-transformed pollutant concentrations and time. Therefore, the missing values of annual means at the beginning of the 1990s were replaced using linear extrapolation for the logarithmic SO2 and TSP values. The average of each pollutant’s annual means for the 2 years preceding each survey (1991–1992, 1994–1995, and 1997–1998) were taken as exposure variables.
Size-fractionated particulate matter was measured for 6 months in 1993 and again in 1999 in conjunction with medical examinations in the three study areas. 11 Sulfur dioxide (SO2) was measured with an Ansyco model AF-21-M pulsed fluorescence analyzer (Environnement, Poissy, France). Total suspended particles (TSP) were measured with an FH-62-IN beta absorption monitor (FAG Kugelfischer, Schweinfurt, Germany). Number concentrations within specific size fractions of airborne particulate matter were measured by a mobile aerosol spectrometer. 11 The detailed methods for these measurements have been described elsewhere. 11,17,18
In each survey, we recruited first-, third-, and sixth-grade school children from the Zerbst, Hettstedt, and Bitterfeld metropolitan areas. In the Zerbst and Hett-stedt areas, we contacted all daycare centers and schools, whereas in the larger Bitterfeld metropolitan area, we randomly selected schools and daycare centers. The population in all three areas is ethnically homogenous. We excluded children if they had lived for less than 2 years in their current home and if their previous home was farther than 2 km away. Children who attended specialized institutions were not included. School entrants and third-graders who participated in the first and second surveys were eligible to be reexamined as third- and sixth-graders in the subsequent surveys. The study protocol was approved by the University of Rostock’s Ethics Committee. Informed consent was obtained from the parents of all participating children.
As part of the study, teachers distributed questionnaires to the children’s parents and collected completed questionnaires a week later. The questionnaire has been described previously. 13
We used a two-stage approach to examine the association between air pollution variables and respiratory symptoms. 1,2 The repeated-measures aspect of the design was taken into account by using the generalized estimating equations (GEE) approach. 19
In the first stage, adjusted respiratory symptom prevalences for the nine combinations of area and survey were computed with a GEE logistic regression model. As potential confounders, we included age, gender, parental education, parental atopy, and four indoor factors (home dampness or molds, gas cooking, environmental tobacco smoke (ETS) exposure at home, and contact with cats).
In the second stage, the logits of these adjusted prevalences were regressed against one air pollution variable in a mixed linear model that included a linear function of the air pollution variable, fixed area effects, random deviations, and the estimation errors from the first stage. We assumed a covariance matrix of the form ς2I + Σ, where I is the identity matrix and Σ is the estimated covariance matrix of the logits from the first stage. 2 For the pollution variable, the model results depend only on the temporal changes within each area. Identical results arise from modeling the logit differences within each area as a multiple of the corresponding changes in the pollutant, with an accordingly adapted covariance matrix.
We explored possible effect modification by computing adjusted logits for all combinations of area, survey, and the categories of the potential modifier, and we included different slopes of the pollutant in the second-stage model. We gave special consideration to children with or without any of four indoor exposures.
Ambient Air Pollution Trends
At the beginning of the 1990s, the Hettstedt area had the highest annual mean concentration of TSP, whereas the Bitterfeld area had the highest annual mean concentration of SO2 (Table 1). During the 1990s, the mean levels of SO2 and TSP decreased in all three study areas, with a 92% decrease in SO2 in Bitterfeld and an estimated 58% decrease in TSP in Hettstedt. By the end of the decade, differences in these air pollutants between the three areas had nearly disappeared. Over the surveys and sites, the Spearman correlation between the annual means of SO2 and TSP was r = 0.88.
Particle-number concentrations decreased by approximately 45% in the accumulation mode, but increased by 40% in the nucleation mode between two 6-month sets of measurements in 1993 and 1999 (Table 1). Zerbst had greater proportional changes in accumulation-mode particle numbers (−54%) and nucleation-mode particle numbers (+59%) than either Bitterfeld (−39% and +35%) or Hettstedt (−41% and +27%). The overall number concentration of fine particles changed very little between 1993 and 1999.
Participation Rates and Demography
In total, parents completed 7,632 of 9,630 distributed questionnaires. The response rates ranged from 69% to 92% across the study areas and surveys. After exclusion of children living less than 2 years in a radius of 2 km from their current home, data for a total of 6,959 questionnaires (for 4,949 children) contributed to the descriptive analyses (Table 2). There were 3,264 children participating in only one survey, 1,685 children participating in at least two surveys, and 325 children participating in all three surveys. The survey-specific age distributions reflected the general decline in birth rates in eastern Germany after reunification. 20
Temporal Changes in Questionnaire Variables
For most respiratory outcomes, a continuous decline of prevalence was found between the three surveys and the temporal changes followed similar trends in all three study areas (Table 2, Figure 1).
Between 1992–1993 and 1998–1999, changes were also observed in several factors, especially housing characteristics (Table 3). Heating with coal or coke decreased by 24% as these stoves were replaced by gas central heating systems, whereas gas cooking stoves decreased by 23%. Dampness and mold in homes, contact to cats, and parental atopy all increased, whereas current exposure of children to ETS in their homes, attendance at daycare facilities before 1 year of age, and bedroom sharing dropped. Birth weight and percentage of children breastfed were unchanged.
Annual Means of TSP and SO2, and Respiratory Illnesses
The community-specific adjusted prevalence of bronchitis and frequent colds show strong associations with a 50-μg/m3 increment in the mean concentration of TSP during the prior 2 years (data not shown). The associations between the adjusted prevalence of respiratory illnesses and a 100-μg/m3 increment in the prior 2-year mean of SO2 are nearly the same as for TSP, but the scatter plot shows much more heterogeneity (Figure 1). Within each single survey, the prevalence of bronchitis was highest in the area with the highest biannual mean concentration of TSP, whereas the association with SO2 was not as clear.
For an increment of 50 μg/m3 in TSP, we found associations with lifetime bronchitis, sinusitis, and frequent colds during the past 12 months (Table 4). An increment of 100 μg/m3 in SO2 was associated with bronchitis, frequent colds, and febrile infections. Adjustment for several confounding factors changed the odds ratios only marginally.
We observed stronger associations of increments of TSP and SO2 with the prevalence of bronchitis and sinusitis among children who were not exposed to any of four indoor factors (living in damp houses with visible molds, ETS in the home, gas cooking emissions, and contact with cats;Table 4).
The changes in the adjusted prevalence of bronchitis and frequent colds (three or more during the past year) were similar in all three age groups and followed a similar pattern across the various birth cohorts (Table 5).
This study investigates the extent to which the abrupt decline of coal combustion-related air pollutants in selected areas of eastern Germany is associated with changes in prevalence of nonallergic respiratory disorders in children. This sudden decline in air pollution concentrations is unprecedented and provides a useful “natural experiment.”
Although we report the associations in terms of relative risks for an increase of a particular air pollutant, nonallergic respiratory health in children actually improved with the declines in TSP and SO2 concentrations (Figure 1). Higher effect estimates for these ambient pollutants were found among children not exposed to several indoor factors. Our results suggest that exposure to combustion-derived air pollution is causally related to nonallergic respiratory health in children.
Comparison with Other Studies
Direct comparisons with the results of other studies are limited because of differing measurements of ambient pollutants, sources of air pollutants, population exposures, temporal changes in air pollution levels, and health-outcome definitions. 1–4,21–28 Our longitudinal design provides an advantage in relating improvements in air quality with temporal changes of respiratory health. Despite differences among studies, our findings agree with results from other geographic settings (see review by Pope and Dockery 5).
For a 10-μg/m3 increment in TSP, our recalculated odds ratio of 1.25 for bronchitis is similar to the recalculated odds ratio of 1.20 (1.27 for a 10-μg/m3 increment in particulate matter less than 15 μm in aerodynamic diameter [PM15]) in the Harvard Six Cities Study, 1 and lower than the odds ratio of 1.40 for a 10-μg/m3 increment in PM10 in the SCARPOL children. 2 Among children entering school in Sachsen-Anhalt, a two-pollutant model with both SO2 and TSP showed an odds ratio for lifetime bronchitis of 1.63 for a 50-μg/m3 increment in TSP, 22 which is lower than the odds ratio we found.
Total (TSP) and accumulation-mode particles might contribute to the prevalence of nonallergic upper respiratory disorders. The increased number concentrations of nucleation-mode particles do not support the idea that long-term exposures to nucleation-mode particle number concentrations contribute to adverse effects on upper respiratory illnesses in childhood. Particle mass concentrations play a more important role in respiratory illnesses such as bronchitis, otitis, and infections than nucleation-mode particle number concentrations, which contribute only very little to total particle mass. 10
Whereas the prevalence of bronchitis and nonallergic respiratory symptoms consistently decreased in eastern German children during the 1990s, temporal changes of prevalence of asthma, hay fever, and allergic sensitization were inconsistent. 22–25
In line with some previous studies, 2,22,23 but not all, 26,27 this study does not support the hypothesis that long-term exposures to coal combustion-related particles are related to asthma and wheeze in children. 25 Whether exposures to high concentrations of nucleation-mode particles play a major role in the development of atopic diseases requires further study.
Time Frame for Improvement of Respiratory Health
Because most regional cross-sectional studies have included subjects who lived in the area for 2 or 3 years, they assessed an induction time for long-term effects of air pollutants of 2 to 3 years. 28 These studies cannot clearly determine the specific effects of air pollutant exposures at a particular age. Furthermore, the time frame relating reduced exposures to air pollutants with improved respiratory health has not yet been established. Although several studies showed long-term effects of exposure to environmental tobacco smoke during infancy and even in utero, 29 our results indicate an additional contribution to improvement of respiratory health in conjunction with improved air quality a few years before the examination (Table 5).
Our results are consistent with another cross-sectional study done in eastern Germany that reported a remarkable decline of prevalence of bronchitis in eastern German fourth-graders. 23 Although the children of both surveys (1991–1992 and 1995–1996) were highly exposed to SO2 and TSP during their first 4 or 5 years of life, their exposures differed during the years just before each examination. Presumably, the cumulative exposure a few years before the examination has a greater contribution to health than exposures in early infancy.
A major limitation of this study is the lack of data on other air pollutants, particularly fine particles that have shown strong associations with mortality. 5 However, we did conduct limited size-fractionated particle measurements in 1993 and in 1999. In the winter of 1993–1994, PM10 means were 40 μg/m3 in Hettstedt and 33 μg/m3 in Zerbst. 30 From January through June 1993, fine particle mass (PM2.5) was approximately 38–39 μg/m3 in the three study areas, but this value decreased to 19–21 μg/m3 in a similar 6-month period in 1999. 11 From these results, we can assume that 60% to 90% of TSP is from fine particle mass in our study areas. Measurements in the eastern German city of Erfurt (80 km south of our study areas) showed a similar contribution of fine particle mass to TSP during the 1990s. 9
Decreases in local lignite coal combustion and consequent decreased concentrations of SO2 and large particles may have led to decreased particle scavenging. This may be partially responsible for the observed increases in nucleation-mode particle concentration. 9–11 Sporadic measurements during the winters of 1993 and 1999 indicate a slight increase in nitric oxide and nitrogen dioxide concentrations in conjunction with a 22% increase in passenger cars per 1,000 inhabitants between 1991 and 1998. 11 Also, given the high correlation between SO2 and TSP in our study areas, we were not able to disentangle the effects of the two pollutants.
Another issue is the limited number of study areas. These communities differed with respect to other risk factors such as poverty, access to health care, nutrition, and smoking. To control regional confounding, we included area effects in our regression models. Thus, the computed odds ratios are based only on the temporal changes within each area.
A major strength of the present study was the dramatic improvement in air pollution levels within a short time period. However, because German reunification was accompanied by a sweeping change in East German living conditions toward a Western life-style, risk factors for respiratory symptoms other than air pollutants such as diet, medication, etc, might have changed in parallel with air pollution.
Cumulative exposures to air pollutants were assessed solely by central-site measurements in the city of residence. Fortunately, in this study, nearly half of the studied children have spent their entire lives in one home, fewer than 10% of the children in the initial cohort (1992–1993) left the area within the following 6 years, and each study area was only about 10 km in diameter. Thus, the children of each single study area share a rather homogenous exposure to air pollutants. Given the overall low mobility of our three study populations and their proximity to the central-site monitors, bias attributable to differential misclassification of exposure is unlikely.
Selection bias attributable to the lower participation rates for the second and third surveys appears unlikely because parental educational level (a marker for socioeconomic status) showed a similar distribution in all three surveys. Furthermore, as young adults who participated in a respiratory health survey tended to overreport respiratory symptoms, 31 the lower participation rates observed in our later surveys should have led to higher reported prevalence in these surveys.
In conclusion, the prevalence of nonasthmatic respiratory disorders may be reduced within a short time period by improving air quality. Our data on long-term exposures to air pollution indicate the reversibility of adverse health effects in children and add further evidence of a causal association with combustion-related air pollutants.
We thank Helgard Bach, Horst Adam, Bernhard Wilde, Hannelore Wolff, Doris Bodesheim, Irina Hörhold, Ingo Keller, Gerd Burmester, Johannes Rudzinski, Brigitte Hollstein, Helga Machander, Regina Müller, Doris Albrecht, and Christa Boettcher for gathering regional data and for local assistance; Hubert Schneller for data handling; Thomas Tuch for air pollution measurements in 1993; all of the teachers in Hettstedt, Zerbst, and Bitterfeld, as well as the local school authorities and health care centers for their support; and all parents and children for their participation.
1. Dockery DW, Speizer FE, Stram DO, Ware JH, Spengler JD, Ferris BG. Effects of inhalable particles on respiratory health
. Am Rev Respir Dis 1989; 139: 587–594.
2. Braun-Fahrländer C, Vuille JC, Sennhauser FH, et al
. Respiratory health
and long-term exposure to air pollutants in Swiss schoolchildren. SCARPOL Team. Swiss Study on Childhood Allergy and Respiratory Symptoms with Respect to Air Pollution, Climate and Pollen. Am J Respir Crit Care Med 1997; 155: 1042–1049.
3. Jedrychowski W, Flak E. Effects of air quality on chronic respiratory symptoms adjusted for allergy among preadolescent children
. Eur Respir J 1998; 11: 1312–1318.
4. Peters J, Avol E, Navidi W, et al
. 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 1999; 159: 760–767.
5. Pope CA, Dockery DW. Epidemiology of particle effects. In: Holgate ST, Samet JM, Koren HS, Maynard RL, eds. Air Pollution and Health. New York: Academic Press, 1999.
6. Health effects of outdoor pollution. Part 1. Committee of the Environmental and Occupational Health Assembly of the American Thoracic Society. Am J Respir Crit Care Med 1996; 153: 3–50.
7. Health effects of outdoor pollution, part 2. Committee of the Environmental and Occupational Health Assembly of the American Thoracic Society. Am J Respir Crit Care Med 1996; 153: 477–498.
8. Umweltbundesamt. Daten zur Umwelt 1995/96. Berlin: Erich Schmidt Verlag, 1997.
9. Ebelt S, Brauer M, Cyrys J, et al
. Air quality in post-unification Erfurt, East Germany
: apportioning changes in pollutant concentrations to changes in emissions. Environ Health Perspect 2001; 109: 325–333.
10. Wichmann HE, Spix C, Tuch T, et al
. Daily Mortality and Fine and Ultrafine Particles in Erfurt, Germany. Part I. Role of Particle Number and Particle Mass
. HEI Pub. No. 98. Boston: Health Effects Institute, 2000.
11. Pitz M, Kreyling WG, Hoelscher B, Cyrys J, Wichmann HE, Heinrich J. Change of the ambient particle size distribution in East Germany
between and 1999. Atmospheric Environ
12. Heinrich J, Hoelscher B, Wjst M, Ritz B, Cyrys J, Wichmann HE. Respiratory diseases and allergies in two polluted areas in East Germany
. Environ Health Perspect 1999; 107: 53–62.
13. Heinrich J, Hoelscher B, Wichmann HE. Decline of ambient air pollution
and respiratory symptoms in children
. Am J Respir Crit Care Med 2000; 161: 1930–1936.
14. Landesamt für Umweltschutz Sachsen-Anhalt. Immissions-schutzbericht
, 1995. Report No. 27. Halle: Landesamt für Umwelt-schutz, 1998.
15. Landesamt für Umweltschutz Sachsen-Anhalt. Immissions-schutzbericht
, 1992. Report No. 32. Halle: Landesamt für Umwelt-schutz, 1999.
16. Landesamt für Umweltschutz Sachsen-Anhalt. Immissions-schutzbericht
, 1999. Report No. 37. Halle: Landesamt für Umwelt-schutz, 2000.
17. Tuch T, Brand P, Wichmann HE, Heyder J. Variation of particle number and mass concentration in various size ranges of ambient aerosols in Eastern Germany
. Atmospheric Environ 1997; 31: 4193–4197.
18. Brand P, Ruoβ; K, Gebhart J. Performance of a mobile aerosol spectrometer for an in situ
characterization of environmental aerosol in Frankfurt city. Atmospheric Environ 1992; 26A: 2451–2457.
19. Liang KY, Zeger SL. Longitudinal data analysis using generalized linear models. Biometrika 1986; 73: 13–22.
20. Statistisches Bundesamt. Statistisches Jahrbuch 1994 für die Bundesrepublik Deutschland. Wiesbaden: Metzler-Poeschel, 1994.
21. Peters JM, Avol E, Gauderman WJ, et al
. A study of twelve Southern California communities with differing levels and types of air pollution. II. Effects on pulmonary function. Am J Respir Crit Care Med 1999; 159: 768–775.
22. Kramer U, Behrendt H, Dolgner R, et al
. Airway diseases and allergies in East and West German children
during the first 5 years after reunification: time trends and the impact of sulphur dioxide and total suspended particles. Int J Epidemiol 1999; 28: 865–873.
23. von Mutius E, Weiland SK, Fritzsch C, Duhme H, Keil U. Increasing prevalence of hay fever and atopy among children
in Leipzig, East Germany
. Lancet 1998; 351: 862–866.
24. Heinrich J, Wjst M. Increase of allergy in East Germany
. Lancet 1998; 351: 1813.
25. Heinrich J, Hoelscher B, Jacob B, Wjst M, Wichmann HE. Trends in allergies among children
in a region of former East Germany
between 1992–1993 and 1995–1996. Eur J Med Res 1999; 4: 107–113.
26. Abbey DE, Ostro BE, Petersen F, Burchette RJ. Chronic respiratory symptoms associated with estimated long-term ambient concentrations of fine particulates less than 2.5 microns in aerodynamic diameter (PM2.5
) and other air pollutants. J Expo Anal Environ Epidemiol 1995; 5: 137–159.
27. Abbey DE, Hwang BL, Burchette RJ, Vancuren T, Mills PK. Estimated long-term ambient concentrations of PM10
and development of respiratory symptoms in a nonsmoking population. Arch Environ Health 1995; 50: 139–152.
28. Zemp E, Elsasser S, Schindler C, et al
. Long-term ambient air pollution
and respiratory symptoms in adults (SAPALDIA study). Am J Respir Crit Care Med 1999; 159: 1257–1266.
29. Gergen PJ, Fowler JA, Maurer KR, Davis WW, Overpeck MD. The burden of environmental tobacco smoke exposure on the respiratory health
2 months through 5 years of age in the United States: Third National Health and Nutrition Examination Survey, 1988 to 1994. Pediatrics 1998; 101: E8.
30. Beyer U, Franke K, Cyrys J, et al
. Air pollution and respiratory health
: the PEACE panel study in Hettstedt and Zerbst, Eastern Germany
. Eur Respir Rev
31. de Marco R, Verlato G, Zanolin E, Bugiani M, Drane JW. Nonresponse bias in EC Respiratory Health
Survey in Italy. Eur Respir J 1994; 7: 2139–2145.