Tobacco Documents Online

International Journal of Epidemiology ~ International Epidemiological Association 1997 Vol. 26, No. 3 Printed in Great Britain Childhood Asthma in Four Regions in Scandinavia: . Risk Factors and Avoidance Effects B FORSBERG,* J PEKKANEN,** J CLENCH-AAS,t M-B M.&RTENSSON,$ N STJERNBERG,~ A BARTONOVA,t K L TIMONEN** AND S SKERFVING== Forsberg B (Department of Environmental Health, Ume& University, S-901 87 UME/~, Sweden), Pekkanen J, Clench-Aas J, M&rtensson M-B, Stjemberg N, Bartonova A, Timonen K L and Skerfving S. Childhood asthma in four regions in Scandinavia: Risk factors and avoidance effects. International Journal of Epidemiology 1997; 26:610-619. Background. The high and increasing prevalence of childhood asthma is a major public health issue. Various dsk factors have been proposed in local studies with different designs. Methods. We have made a questionnaire study of the prevalence of childhood asthma, potential risk factors and their relations in four regions in Scandinavia (Ume& and Maim5 in Sweden, Kuopio in eastern Finland and Oslo, Norway). One urban and one less urbanized area were selected in each region, and a study group of 15 962 children aged 6-12 years was recruited. Resu/ts. The prevalence of symptoms suggestive of asthma varied considerably between different areas (dry cough 8-19%, asthma attacks 4-8%, physician-diagnosed asthma 4-9%), as did the potential risk factors. Urban residency was generally not a risk factor. However, dry c.ough was common in the most traffic polluted area. Exposure to some of the risk factors, such as smoking indoors and moisture stains or moulds at home during the first 2 years of life, resulted in an increased risk. However, current exposure was associated with odds ratios less than one. Conc/usions. Our findings were probably due to a combination of eady impact and later avoidance of these risk factors. The effects of some risk factors were found to differ significantly between regions. No overall pattern between air pollution and asthma was seen, but air pollution differed less than expected between the areas. Keywords: childhood asthma, home dampness, passive smoking, pets, air pollution, avoidance The high and increasing prevalence of childhood asthma is a major community problem. The increase probably reflects an altered use of the asthma label as well as a true increase in prevalence. Changes in liv- ing conditions and environmental exposure to potential causes, including diet and viral infections, probably account for the true increase. Epidemiological studies. mainly using questionnaires, have focused on various risk factors in the indoor or outdoor environments and comparisons of their results have been difficult. In the indoor environment, allergens from house-dust mitesl and pets such as cats,2 dampness and moutds,3-8 tobacco smoke8-1° and gas stoves6"s'11 have been found to be * Department of Environmental Health, Ume,~ University. S-901 87 UME,~, Sweden. ** National Public Health Institute, Kuopio. Finland. '~ Norwegian Institute for Air Research. Lillestr6m, Norway. * County Council of MalmOhus. Lund. Sweden. ~ National Institute for Working Life, Ume'~.'Sweden (deceasedl. " Department of Occupational and Environmental Medicine, Uni- versity Hospital. Lund. Sweden. 610 associated with an increased risk of childhood asthma and wheezing disorders. Air pollution from traffic and industry have been reported to be risk factors.~2-~4 However. in comparisons between eastern and western Europe, lower prevalences of atopy and asthma but a higher prevalence of cough have been reported in more polluted environments in the east than in less polluted areas in the west.15'16 The cumulative prevalence of physician-diagnosed asthma in children aged <10-12 years has earlier been reported to be 4-6% in Sweden,17-~9 and 3.5% in Oslo2° while in different parts of Finland it has been reported to be 1-3% among 3-8 year old children.21 The aim of this report is to describe the prevalence of childhood asthma indicators, potential risk factors and their relationships with asthma in four areas in Scandinavia: northern and southern Sweden, eastern Finland and southern Norway. All these areas were included in the European collaborative project (PEACE). "Pollution Effects on Asthmatic Children in Europe" focusing on short-term effects of air pollution on respir- atory symptoms. FIGUR MEfHOI Study A t'~ In each , area and (Figure 1 comparaL ing the Ume,h i 100 000 i in the cit and villa parts of t and three centre. metropoli inhabitan. of Swede two comr Since the cities, the ated abou as the co with 450 by hills ir sea level parts of tl tion, were

~'tors. 12-~4 western ma but a in more polluted iagnosed tier been in Oslo-'0 reported evalence < factors areas in eastern aS were 'EACE), Europe" n respir- FIGURE 1 Stttdy areas in Sweden, Finland and Norway METHODS Stttdy Areas and Populations In each of the four areas included, one urban study area and one less urbanized control area were selected (Figure I), and were environmentally characterized by comparable air pollution measurements conducted dur- ing the panel studies within the PEACE study.-'~- UmeS, is the major city in northern Sweden with about 100 000 inhabitants, of whom almost three-quarters live in the city and the remainder in surrounding suburbs and villages. The urban study area included central parts of the town, and the control area was Holmsund and three other villages located 10-20 km from the city centre. MalmO is Sweden's third largest city and the metropolitan area of Malm6/Burl/Sv has about 250 000 inhabitants. It is situated in the most southern part of Sweden. A central area on the border between the two communities was selected as the urban study area, Since the surroundings are flat and affected by several cities, the rural communities Olofstr6m and ,Sdmhult situ- ated about 120 km north east of Maim0 were selected as the control area. Oslo, the capital city of Norway, with 450 000 inhabitants, lies on a t]ord surrounded by hills in southern Norway. The centre of the city is at sea level and was included as the urban area. Peripheral parts of the city, situated higher up and with less pollu- tion. were selected as the control area. Kuopio, situated CHILDHOOD ASTHMA AND RISK FACTORS 611 by a lake in eastern Finland, has about 80 000 inhabit- ants. The centre of the town was included as the urban area and the suburbs of Petonen and Jynkk~i, 5-10 km from the city centre, as the control area. In all the urban areas, motor vehicle traffic is the most important local source of air pollution. Exhaust fumes are lower in all the control areas than in the ur- ban areas, while ozone concentrations during the sum- mer should be somewhat lower in the cities because of the greater nitrogen oxide emissions there. The con- centrations of vehicle exhaust fumes are greatest during the winter, while ozone levels are higher in the summer. Questionnaire For this study the PEACE project's screening question- naire2z was extended to include questions on current risk factors as well as those present during the child's first 2 years of life, and on the presence of parental asthma. These exposure questions dealt with indoor factors in the child's home; including the presence of household pets with fur or feathers, fitted carpets, mois- ture stains or moulds, use of wood-stoves or fireplaces and smoking indoors. We also ascertained whether or not children had attended a day care centre during their first 2 years of life. The respiratory questions were adapted from a World Health Organization questionnaire~ and a questionnaire developed at the University of Groningen, The Nether- lands, based on the American Thoracic Society ques- tionnaire for children.~'4 In this study, we have mainly focused on three questions all relating to events in the 12 months prior to completion of the questionnaire which suggest recent asthma or asthma-like problems: attacks of shortness of breath with wheezing ('asthma attacks'), dry cough at night apart from colds, and asthma treatment by a physician. The cumulative (lifetime) pre- valences of asthma attacks and of physician-diagnosed asthma are also given in addition, as well as the preval- ence of current use of airway medication on a daily basis. In the Swedish centres, the questionnaires were mailed to the parents, and also returned by post, while they were handed out and returned through the schools in Kuopio and Oslo. [n Kuopio the questionnaire was distributed to all children in the first six grades in five schools in the centre of the town and in three schools in the control area. In Oslo ten schools in the urban area and six in the control area were included in the same way. In Sweden, children that were registered residents in the study areas were defined as the study population if they were born 1981-1987 (in northern Sweden) and in 1981-1986 (in southern Sweden), In the school- based samples some children were born before 1981 and have not been included in the analysis (Table I).

612 TABLE 1 Study population INTERNATIONAL JOURNAL OF EPIDEMIOLOGY Variable Northern Sweden Southern Sweden Urban Control Urban Control Oslo Kuopio Urban Control Urban Control No. selected 4243 1496 2684 1989 2538 1983 1539 1456 Response rate (% I 94,5 96.1 88.0 92.9 81.9 89,0 87. I 84.0 No. of rospondents 4008 1437 2361 "1847 2084 1766 1341 1223 No. of respondents born 1981 - 1987 all all all all 2029 1740 1327 1213 TABLE 2 Prevalence of potential risk factors (%) Variable Northern Sweden Southern Sweden Urban Control Urban Control Oslo Kuopio Urban Control Urban Control :- V~'iable ' l~vcr asthma R~ent asthm~ . in the urban ~tween the u~an southe~ ~eas. espec~ the home we~ and more ~eas. Atten~ 2 years of life a~ most Background shown in Tab~ ~a, with the sulphur dioxid. ~culates wi~ Were higher in long-distance Were clearly 1o ence for patti, not seen in nc hOUses had oil- Total born 1981-1987 4008 1437 2361 1847 Parental asthma (%) 14 14 11 10 Currently (%) fitted carpets 15 16 19 24 household pets 40. 48 46 59 wood-stove (any use) 18 36 - 12 48 smoking indoors . 2-4 20 45 25 moisture stains or moulds 13 9 10 8 First 2 years of life (%) fitted carpets 25 28 45 50 household pets 31 40 38 51 wood-stove (any regular use) 9 15 6 24 smoking indoors 23 24 47 30 moisture stains or moulds I0 8 9 7 attendance at day care centre 57 50 52 41 The overall response rate was about 90% and resulted in a study group of 15 962 children aged 6-12 years. The questionnaires were answered during September (Oslo and Kuopio) and November-December (Sweden) 1993. Statistical Analysis Prevalence rates of respiratory problems were age- standardized by the direct method using four strata and weights according to the total study population's dis- tribution (children born 1981, 1982-1983, 1984-1985, t986-1987). Adjusted odds ratios (OR) (and 95% confidence intervals [CI]) were estimated using logistic regression (SPSS/PC), for the combined population and for each region separately since there were significant differences between regions in the relationship of some risk factors with prevalence: 2029 1740 1327 1213 10 9 9 7 42 46 3 2 32 42 42 36 44 64 25 47 62 49 13 6 8 5 6 3 52 56 10 7 24 23 22 18 30 36 16 15 49 38 15 9 8 4 4 2 53 49 33 33 RESULTS Exposure The patterns of risk factors in the eight study areas are shown in Table 2. Having a parent with asthma was most common in northern Sweden; 14% in both of these study areas, with 7-11% in the other areas. Fitted carpetS were especially common in Oslo but rare in Kuopio. In all areas such carpets were currently reported to be less common than during the child's first two years of life. Household pets were common, 36-59% of children had one, and, except in Kuopio. they were more common in the control than in the urban areas. Pets were less common during the child's first 2 years of life. Sire- ilarly, a wood-stove was more common in the control areas, with the highest occurrence in Oslo's control area. Smoking indoors at home occurred more frequently

Total g6 17 928 g4.0 89.6 23 16 067 '3 15 962 Kuopio n Control 1213 7 2 36 4.7 6 3 7 18 15 9 2 33 udy areas are ~ asthma was both of these Fitted carpets in Kuopio. In ted to be less years of life. 'children had lore common ~ts were less of life. Sim- n the control control area. • e frequently CHILDHOOD ASTHMA AND RISK FACTORS 613 TAI~LE 3 hulicators of exposure to air pollution (Idg/m"~} and temperature (°C) dttring measuremem period" Variable Northern Sweden Southern Sweden Oslo Kuopio Urban Control Urban Control Urban Control Urban Control Air pollution levelsb mean 24 h(SD) NO,. 25 (15) 1:5 (11) 20 (10) 9 (5) 49 (22) 20 (14) 28 (15) 14 (9) mean 24h{SD) SO2 3 (2) 4 (3) 6 (4) 4 (3) 12 (8) 4 (5} 6 (6) - mean24h(SD) BS 5 (5) 5 (4) 8 (7) 4 (3) 27 (18) 13 (10) 13 {I0) 8 {9) mean24htSD) PM.~ 13 (8) 12 (6) 22 (12) 16 (8) 19 (12) 12 161 18 (11) 13 (8) Temperature mean temperature (SD) -10 (7) -10 (7) +1 (3) -1 (3) --1. (4) -4 {41 -5 (7) -5 (71 Measurement period was 5 January-27 March 1994 for northern Sweden. 15 January-19 March 1994 for southern Sweden. I December 1993- I February 1994 for Oslo and 8 February-5 April 1994 for Kuopio. - Not measured. BS = black smoke, PM~o = particulate matter, aerodynamic diameter < 10 p.m. TABLE 4 Age-standardized prevalences of indicators for asthma Variable Northern Sweden Southern Sweden Oslo Kuopio Urban Control Urban Control Urban Control Urban Control Ever asthma attacks 17 16 15 Physician diagnosed asthma 8 8 8 Recent asthma attacks 7 8 7 Recent dry cough 8 8 12 Recent asthma treatment 4 5 5 Current medication use 3 3 4 15 14 14 12 12 6 9 9 4. 4. 6 6 6 4 5 10 19 12 12 12 3 5 5 3 3 2 2 2 3 4 in the urban areas, and there were large differences between the high prevalences in Oslo (both areas) and urban southern Sweden (45-62%) compared to the other areas, especially Kuopio. Moisture stains or moulds in the home were most often reported in Sweden (8-13%), and more frequently in the urban than in the control areas. Attendance at a day care centre during the first 2 years of life occurred most seldom in Kuopio (33%) and most common in urban northern Sweden (57%). Background air pollution levels in the study areas are shown in Table 3. Urban Oslo was the most polluted area, with the highest levels of nitrogen dioxide (NO_,), sulphur dioxide (SO.,) and black smoke (BS). Levels of particulates with an aerodynamic diameter < 10 I.tm (PMto) were higher in urban southern Sweden, probably due to long-distance transportation. Nitrogen dioxide levels were clearly lower in the control areas, while a differ- ence for particulate matter and sulphur dioxide was not seen in northern Sweden. probably because more houses had oil- or wood-fired heating in the control area. Prevalence and Risks The age-standardized prevalence rates for our asthma indicators varied considerably between the different study regions, but the pattern was not the same for the different indicators. The variation was less between urban areas and their corresponding control area (Table 4), The proportion of children who were reported to have ever had, or had had asthma attacks during the 12 months prior to completion of the questionnaire, was greatest in northern Sweden and lowest in Kuopio. Recent dry cough was most often reported in Oslo, and was much higher in urban Oslo than elsewhere. When studied in relation tO the measured air pollution levels, the indicators of recent asthma did not show any pattern (Figures 2-3). Combined Population After adjustment for age. sex, parental asthma and exposure indicators in the models for the combined population, the risk of dry cough was significantly

614 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY % Prevalence 2O 10 o o o o o o o o Treatment for asthma Dry cough Asthma attacks 0 10 12 14 16 18 20 22 ;,4 PMIO (pg/m3) FIGURE 2 Prevalence rates for indicators of recent asthma in relation to inhalable particulates (PMIO) levels 2O 15 10 % Prevalence o o 0 o 0 oo ~ o o 0 o o o 5 1'0 1'5 20 25 30 35 40 45 50 NO2 (I.lg/m3) FIGURE 3 Prevalence rates for indicators of recent asthma in relation to nitrogen dioxide (NO,_) levels o Treatment for asthma * Dry cough o Asthma attacks 55 TABI.E 5 At/,, potential ris. Variable Boy Parental asth= Current expt~" Urban resi~ Malmoh Kuopioh Oslo~ Fitted carF Household Wood-stov Smoking it Moisture s Daring the cl Fitted carl: Househokl Wood stov Smoking i Moisture ~ Attendanc~ Models alst Northern S~ greater in (OR = 1., treatment associated Parental z asthma tr,_ dry cough For cm risks asso, (OR = 0., statisticall dry cougl creased ri dry cough Conce= of life, tl- the home (OR = 1.15 home, wh dry cough and almo: attacks. A with a sli; risk of d~ a wood-si the OR fo

Wi CHILDHOOD ASTHMA AND RISK FAC'I~RS 6 15 . ,,~^BL~.5 Ad.juste~l o,!~ls ratios (with 95% confidence intervals) attd the association between indicators of recent asthma, study area and tors~: r- lennal risk ~! Vadable " " Boy . l~arenta/asthma ,'~3 ' Cart~nt exposure ~." i" Urban resident ,-,,~.~ Maim6~' ~i~ ' Ku°pi°~ Oslo" Fitted carpets " :.~ Household pets Woe-stove (any usel Asthma attacks Dry cough Asthma treatment Smoking indoors Moisture stains or moulds During the child's first 2 years Fitted carpets Household pets Wood stove (any regular use) Smoking indoors Moisture stains or moulds Attendance at a day care centre 1.4 (I.2-1.61 1.2 (I.1-1.4) 1.6 (1.4-1.9) 3.5 (3.0--4.1) ,,~2.3 (2.0-2.6) 4.6 (3.9-5.5) 1.0 (0.8-1.1) I.I (I.0-1.2) 1.0 (0.8-1.2) L I (0.9-1.31 1.7 ¢ 1.4-2.0) 1.2 (1.0-1.5) 0.7 (0.6-0.9) 2.1 (1.8-2.5) 1.2 (1.0-1.6) I. I (0.9-1.4) 2.9 (2.4-3.4) 1.6 (1.3-2.1) 0.7 (0.5-0,8) 0.8 (0.7.-0,9) 0.8 (0.6-1.0) 0.6 (0.6-0.8) 0.7 (0.6-0.8) 0.5 (0.5-0.7) 1.0 (0.8-1.2) 0.9 (0.8-I.OI 0.9 (0.7-1.I) 0.7 (0.6-0.8) 0.9 (0.8-1.0) 0.6 (0.5-0.8) 1.2 (0.9-1.5) 1.6 (1.3-1.9) 0.9 (0.7-1.2) 1,1 (I,0-1.3) 1.1 (0,9-1,2) 1,2 (1.0-1,4) 1.0 (0.8-1.2) 1.1 (1.0-1.2) 1.0 (0.9-1.2) 0.8 (0.7-1.0 ) 0.9 (0.8- I. 1 ) 0.9 (0.7- [, 1 I 1.2 (1.0-1.4) 1.3 ( 1.2-1.5) "~ 1.4 ( I. 1-1.7) 1.6 ( 1.3-2.1 ) ,~1.6 (1.3-2.0) ,1~ 2.0 ( 1.6-2.6) 1.1 (0.9-1.2) 1.1 I 1.0-1.3) 1.0 (0.8-1. I ) Models also include year of birth. Northern Sweden = reference group. greater in all other regions than in northern Sweden (OR = 1.7-2.9), and also in Oslo in regard to asthma treatment (Table 5). Being an urban resident was not associated with a significantly increased overall risk. Parental asthma and male sex gave the highest OR for asthma treatment (OR -- 4.6 and 1.6) and the lowest for dry cough (OR = 2.3 and 1.2). For current risk factors at home, there are decreased risks associated with pets (OR = 0.5-0.7), fitted carpets IOR = 0.7-0.8) and smoking indoors (OR = 0.6-0.9); statistically significant except for smoking at home and dry cough. Moisture stains or moulds produced in- creased risks that reached statistical significance for dry cough. Concerning exposure during the child's first 2 years of life, the presence of moisture stains or moulds in the home statistically significantly increased the risks (OR = 1.6-2.0). There was also an effect of smoking at home, which reached statistical significance for recent dry cough (OR = 1.3) and asthma treatment (OR = 1.4), and almost reached statistical significance for asthma attacks. Attendance at a day care centre was associated with a slight but statistically significant increase in the risk of dry cough (OR = 1.1). Pets. fitted carpets and a wood-stove did not have any significant effects, but the OR for fitted carpets, were all greater than one, and for asthma attacks and asthma treatment almost reached statistical significance. Excluding questions about current exposure from the models for the combined population did not sub- stantially alter the results or the conclusions about the relative importance of early risk factors. When region-risk factor interaction terms were included in the models for the combined population, the effects of some risk factors were found to differ sig- nificantly among regions. Such significant differences existed t'or current exposure to household pets in all three models, for parental asthma and fitted carpets dur- ing the first 2 years of life in the asthma attack model and for urban residence in the dry cough model. These findings also motivate region-specific models. Region-specific Models The logistic regression models for each region showed only small differences in risk between the urban and control area in all four areas (Table 6). Urban residence was associated with a significantly higher risk of asthma treatment in southern Sweden and of dry cough in Oslo. in northern Sweden and Kuopio urban children had an OR of 0.9 for all three indicators of recent asthma. Both parental asthma (either parent with asthma) and sex were important determinants. Parental asthma was

T^t~I.E 6 Adjusted odds ratios Orith 95% cm!lidence intert,als) for the associations between indicators of recent asthma and potential risk factors" Variable Northern Sweden Southern Sweden Oslo Kuopio asthma dry cough asthma asthma dry cough asthma asthma dry cough asthma asthma dry cough asthma attacks treatment attacks treatment attacks treatment attacks trealment Boy Parental aslhma Currenl cxpos;ure Urban resident Fiued carpels I Itmxehold pets Wood-slove tany use1 1.3 (1.0-1.6) 1.4 ( I. I-I.8) ~" 3.212.5-4.1) 2.411.9-3.1) 0.9 (0,7-1.1) 0.9 (0.7-1.1) 1.0 (0.7-1.4) 0.7 (0.5-1.0') 0.7 (0,5-0.9) 0,7 (0.(~0.9) 1.0 (0.7-1.3) 0,7 (0.5-1,0) Smoking ind~rs 0.5(0.4-0.8) 0.8 (0,~--I,I) Moisture stains or moulds I.I (0.8-1.6) 1.4 (I,0-1.9) During the child's first 2 years Fitted carpets 0,7 (0.6-1.01 0.9 (0.7-1.21 Hou~hold pets 1,010.7-1.2) I.I (0.8-1.3) Wt~l-slovc (any regular use) 0.7 (0.6-1.0) I.I (0.8-1.~) Smoking ind~s)rs 1.2 (0~9-1.71 1.6 11.2-2.21 Moisture stains ¢~ rmmlds 1.7 (I,2-2.41 1,7 (I.3-2.41 Auemlancc al a day care centre 1,0 (0.8-1,2) 1.2 (1~0-1.5) " Mt~lels al~ include year of birth. 1.3 (I,0-1.71 1.5 (I,1-1.91 1.2 (0.9-1.4) 4.1 (3.0-5.4) 3.4(2.5-4.61 2.011.5-2.7) 0.9 (0,6-1.21 I,I (0.8-1.51 I.I (0.9-1.41 1.0 10.7-1.51 0.5 (0.4-0.8) 0,7 (0.6-1.01 0.5 (0.4-0.7) 0.8 (0.6-1.0) 0.8 (0.7-1.0) 0.6 10.4-1.01 1.0 (0.7-1.5) 0.9 (0.7-1.21 0.5 (0.3--0t8) 0.7 (0.5-1.0) 0.8 (0.5-1,0) I.I (0.7-1.7) 1.5(I.0-2.2) 1.7(I.3-2.4) 0.9 (0.6-1.3) 1.4 (I.1-1.9) 1.3 (I.0-1.61 1.010.7-1.4) 1.110.8-1.41 0.910.7-1.21 0.9 10,5-1.61 0,7 (0,5-1,1) 0.9 (O.(P-I.21 1.6 (I.I-2,31 I.I (0.8-1,51 1.2(I.0-1.61 2.011.3-3.0) 1.3(0.9-2.1) 1.611.2-2.21 1.0 (0,7-1.31 I,I 10.8-1,41 1.0 t018-1.3) 1.7 11.2-2.41 1.7 (I.2-2.3) 1.2 (I.0-1.5) 1.9 (I.4-2.81 1.2 (0.8-1.71 1.3 11.0-1.61 1.8 11.2-2.71 5.213.7-7.3) 5.3(3.7-7.5) 2.2(I.6-2.91 6.414.4-9.11 2.411.4-3.91 2.511.8-3.61 3.2(2.0-5.31 1.611.1-2.31 0.9(0.7-1.3) 1.5(I.2-1.8) 1.0(0.7-1.41 0.910.6-1.4) 0.910.7-1.21 0.910.6-1A) 0.5(0,3-0.8) 0.6(0.4-0.8) 0.910.7-1.1) 0.8(0,5-1,21 1,2(0.3-4.21 0,6111,2-1.61 0.8(0.2-3.71 0.6 (0~4-0.81 03 10,5-1.0) 0,9 (0.7-1.1) 0.9 (().(~-1,3} 0.2 (0.1-0.4) (L5 (0.4-0,6) 0.2 (0.1-0.41 0.9 (0.6-1.4) 0.9 (1).6-1.3) 0.9 (0.7-1.1) I,I (0.7-1.5) I.I 10.7-1.7) 0,910.7-1.3) 1.010,6-1.6) 0.5(0.3-0.7) 0.710,5-1.11 0,9(0.7-1.21 0.7(0.5-1.11 0.810.3-2.11 1.1(0.7-1.91 IJ1(0,4-2.41 1.0 (0,6-1.81 0.910.5-1.81 1,9(I.3-2.7) 0.6(0.3-t.4) 0.8(0.3-2.01 1.510.9-2.61 0.310.1-1.01 1.4 (I.0-2.0) 1.3 (0.9-1.8) I.I (0.9-1.4) 1.3 (0.9-1.9) 1.6 (0.8-3.1) 0.9 (0.5-1.5) 1,4 ((I.7-2.9) 1.0 (0~7-1.4) 0~8 10.6-1.2) I;I (0.9-1.4) (L9 (0.6-1.31 1.3 (0.8-2.31 1.4 (I.0-1.91 1.6 (0.9-2.6) 1.0 (0,6-1.7) 0~9 (0.7-1.31 0.9 (0,7-1.1) 0.8 10.6-l.21 0.6 (0.2-l.21 1.5 (I.0-2.31 1.3 (0.9-1.91 1,2 (0,9-1.5) I.I 10.7-1,71 0.7 (0.3-1.61 1.6 (0.9-2.6) 1,9 (I.1-3,4) 1.3 (0.9-2.0) 1.8 (0.8-3.4) 2.5 (I.0-6.1) 0.9 (0.6-1.31 0.7 10.4-1.31 0.9 10.6-1.51 0.7 10.3-1.61 2.2 t 1.2--4.01 5.0 12.2-11.21 I.I (0.8-1.51 1.310,9-1,8) I,I 10.9-1.3) 0.9(0.7-1.31 0.910.6-1.4) 1.311.0-1.61 0.910.6-1.41 "7 r- r~

~/i ~C~iI~s~O(~RAI~MA AND RISK FACTORS innificant in all models and showed th /'." ~,ma treatment (3.2-6.4) and the lowest for dry cough ~'~'~i .,~:v-.-~l 9, but this did not always reach statistical significance. "' Most statistically significant OR for potential risk ~ factors were found in the Swedish study populations. current exposures at home mainly showed OR less than o,~ .~ one that were often statistically significant. For example. ~~' all OR were less than one for pets, and, except for ,:,~,,,~ Kuopio, also for smoking indoors at home. For smoking • :.~ tMs effect was significant for asthma attacks and treat- ment for asthma in the Swedish study populations (OR = 0.5-0.7), but weaker for dry cough. This pattern • " also seems, at least in southern Sweden and Oslo, to include fitted carpets, significantly so in southern Swe- den (OR = 0.5-0.7) and significantly for asthma attacks in Oslo (OR = 0.6). Moisture stains or moulds in the home were the only current exposure that was asso- ciated with increased risks, of dry cough consistently (OR = 1.4-1.9) and was statistically significant in all areas except in Kuopio. Significant OR for exposures during the first 2 years of life, on the other hand, were greater than one, with only one exception: there was a negative association between the presence of fitted carpets with treatment for asthma in UmeL The most consistent findings were the results for moisture stains or moulds with OR greater than one in all models (OR = 1.3-5.0) and all values reached significance in northern Sweden and Kuopio, as did those for dry cough in southern Sweden and for asthma attacks in Oslo. All OR were greater than one for smoking indoors at home. except in Kuopio. All OR were greater than one for symptoms associated with fitted carpets, in southern Sweden (significantly so) and in Oslo, while they were less than one in north- ern Sweden. The presence of household pets during the child's first 2 years of life gave OR close to one, except in Kuopio where the association with dry cough (OR = t,4) was significant. A wood-stove at home or attendance at a day care centre gave OR not deviating significantly from one. DISCUSS/ON The response rates in the different study areas ranged from 82 to 96%, with the highest rates in Sweden where the questionnaires were mailed to the parents. There was no difference in the estimated prevalence of symptoms before and after the two reminders used in Ume~t. which suggests that response bias may not be a serious problem. There was no general geographical pattern in age- adjusted prevalences of the asthma indicators between the regions or between urban and control areas. After 617 adjustment for determinants such as parental asthma and exposure such as smoking in the home, there were only a few significant differences. In southern Sweden recent asthma treatment was more common in the urban area than in the control area, but since the difference was smaller for symptoms this may be an artefact of the health-care system, with more private physicians in the city than in the control area. Recent dry cough was common in urban Oslo, which is the most traffic polluted area. However, in the combined analysis there was no overall effect of urban residency on the preval- ences. Nor was there any overall pattern between air pollution levels and any of the prevalence rates, but the levels of air pollution concentrations in the urban and control areas were more similar than we had expected, especially the particle concentrations. It has generally been difficult to link asthma prevalence to air pollution.2~ There might however exist a selection of families with asthmatics from city centres to less traffic polluted areas. The differences in prevalences between regions that were still seen after adjustment could to some extent be an artefact of translation, parental reporting or diagnostic intensity. However, clinical examinations were beyond the scope of the study although the questionnaire was validated in Finland with a clinical examination of a sub-sample of children and a high degree of agreement was found in the prevalence of physician-diagnosed asthma or asthma symptoms in the questionnaire and the clinical diagnoses.~-6 For most current potential risk factors OR were below one, indicating either a protective effect, or more likely, an avoidance of these factors in families with symptomatic children. The only exceptio.n was for mois- ture stains or moulds in the home. which is a problem that is more difficult for the family to tackle. The avoid- ance pattern is confirmed by a study of the changes in the environmental factors e.g. smoking indoors. Of the children with smoking at home during their first 2 years of life, 74% are still exposed if not recently treated for asthma, while 60% are still exposed if recently treated for asthma. Among children who where not exposed dur- ing their first two years of life, 13% are now exposed if not recently treated and 9% if recently treated for asthma. However. avoidance effects may also reflect selective reporting and are then probably most strongly determined by general advice given about avoidance measures. On the other hand, exposures reported during the child's first 2 years of life were associated with increased risks, even when current exposure was omitted from the models. Dampness at home has also previously been reported to be a risk factor.3-8 The mechanism for this may be related to many unspecified indicators including house- dust mites, moulds, poor ventilation that concentrates

618 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY indoor emissions, including environmental tobacco smoke, as well as reporting bias. In this study mites are not a likely explanation, particularly in the northern regions. When samples of asthmatic children included in the PEACE study were skin-prick tested, the preval- ence of positive reactions for house-dust mites was only 1% in Ume5 and 9% in Kuopio. Current exposure to moisture stains or moulds in the home shows the high- est OR for dry cough, but is not a risk factor for asthma treatment. This could fit with Strachan's theory of biased parental symptom reporting,"s which should be less likely for asthma treatment than symptoms such as cough. Exposure during the first 2 years of life. however, had the strongest impact on recent treatment for asthma. Increased awareness or overreporting of exposure among parents with children with recent asthma problems, should affect the results for current exposure more than for early exposure. This would suggest that such report- ing bias probably is an unlikely explanation. Early exposure to passive smoking at home was a risk factor with the strongest effect in northern Sweden. Odds ratios less than one in Kuopio might be explained by reporting bias or avoidance measures in high risk families. The evidence concerning environmental tobacco smoke as a risk factor for asthma has. recently been re- viewed.~-7 Common knowledge about the adverse effects of passive smoking in Scandanavia makes reporting bias more likely than for other exposures. This might have given an underestimation of effects from environ- mental tobacco smoke. Some significant differences in the effects of the individual risk factors studied seem to exist between the regions. Early. exposure to fitted carpets was a signific- ant risk in southern, but not in northern, Sweden. This might be explained by the higher prevalence of allergy to house-dust mites in southern Sweden and may be a reason for a stronger avoidance effect seen for current exposure to fitted carpets in southern Sweden, as well as in the Oslo region. Similarly, early exposure to pets resulted in the highest, and the only significant OR in Kuopio. where the avoidance effect for current ex- posure also was much stronger than elsewhere. A study that only gathered information about cur- rent exposure would obviously yield misleading results because of the general avoidance effect. Avoidance measures have been reported for pets in children with respiratory allergy-~8 and for change of floor covering and mattresses among asthmatic children.29 It has also been shown that current smoking at home is less com- mon in family members of children with asthma in The Netherlands~-9 and in Sweden.~9 The analyses based on the combined study popula- tion assumed homogeneous relations in the different regions, which as indicated by the significant region. risk factor differences was not present. This suggests difficulties in comparing risks in different regions with simultaneous adjustment for individual risk factors in one model without prior information. Interaction analyses are essential if possible differences in the relationships between regions are to be detected. ACKNOWLEDGEMENTS The data were collected within the frame-work of the 'Pollution Effects on Asthmatic Children in Europe' (PEACE) study. The PEACE study is a study on effects of short-term variations in urban air pollution on the respiratory health of children with chronic respiratory symptoms. The study was funded by the EU Environ- ment Programme Contract EV5V-CT92-0220 (seven centres) and two additional EU PECO contracts to allow participation of five centres in Central and Eastern Europe. The Finnish, Norwegian and two Swedish centres were funded by grants from the respective gov- ernments. The study was co-ordinated by the Depart- ment of Epidemiology and Public Health. University of Wageningen, The Netherlands. This report has been supported by The Nordic Council of Ministers, The Swedish and Norwegian Asthma and Allergy Founda- tion and The Medical Faculty, Lunds University. REFERENCES ~ Sporik R. Chapman M D, Platts-Mills T A. House-dust mite exposure as a cause of asthma. Clin ~'rp Allergy 1992; 22: 897-906. -' Morgan W J, Martinez F D. Risk factors for developing wheezing and asthma in childhood. Pediatr Clin Nm'th Am 1992: 39:1185-203. 3 Brunekreef B, Dockery D W. Speizer F E, Ware J H, Spengler J D, Ferris B G. Home dampness and respiratory morbidity in children. Am Rev Respir Dis 1989; 140: 1363-67. ") Platt S D. Martin C J, Hunt S M, Lewis C W. Damp housing. mould growth, and symptomatic health state. Br Med J 1989; 298: 1673-78. ,s Strachan D P, Ftannigan B, McCabe I~ M, McGarry F. Quantification of airborne moulds in the homes of children with and without wheeze. Thorax 1990; 45: 382-87. 6 Spengler J, Nea,,¢ L. Nakai Set al. Respiratory symptoms and housing characteristics. Indoor Air 1994; 4: 72-82. 7 Dales R E, Zwanenburg H. Burnett R, Franklin C A. Respiratory health effects of home dampness and molds among Canadian children. Am J Epidemiol 1991: 134: 196-203. s Dekker C. Dales R. Bartlett S, Brunekreef B. Zwanenburg H. Childhood asthma and the indoor environment. Chest 1991; 100: 922-28. '~ Weitzman M, Gortmaker S. Walker D K. Sobol A. Maternal smoking and childhood asthma. Pediatr 1990: 85:505-1 t. u~ Lindfors A, Hedlin, G. Rietz H. Nordvall L, Wickman M. Indoor risk factors in young children with asthma. J Allergy Clin Immunol 1994; 93: 290. ))NeasLM, D Ferris B respirat, Am J El )-" Schmitzbergt pollutio Puhnon ~ Goren A I. I disease: pollute, ~'~ Weiland S b wheezi~ ity on s ~'~ yon Mutius Thiem~: Med [ '~ ~6 Brfib~.ck L. Bj~rks~ Allerg3 ~7 Holmgren ~ asthm;~ 18 Croner S, K in chil years t )9 Burr M L. Childh lnt J 1- "-o Skarpaas I respir: 1985:

ic~nt re'gi6,n- "his suggests regions with sk factors in 'tio~alyses rel~lhships -work of the ~ in Europe' dy on effects ution on the c respiratory EU Environ- 0220 (seven 'acts to allow and Eastern wo Swedish pective gov- the Depart- Jniversity of ,rt has been nisters, The ~'gy Founda- ersity. ouse-dust mite ergy 1992; 22: rloping J H. Spengler tory morbidity 363-.67. ~)amp housing, tte. Br Med J McGarry F. ~es of children 382-87. ,ymptoms and 72-82. A. Respiratory molds among ~4:196-203. wanenburg H. ~nment. Chest A. Maternal 85: 505-I 1. ~lan M. Indoor Allergy Clin CHILDHOOD ASTHMA AND RISK FACTORS 619 Neas L M, Dockcry D W, Ware J H, Spengler J D. Speizer F E, Ferris B G .It. Association of indoor nitrogen dioxide with respiratory symptoms and pulmonary function in children. Am J Epidemiol 1991 ; 134: 204-19. Schmitzberger R. Rhomberg K, Buchele H et al. Effects of air pollution on the respiratory tract of children. Pediatric Pubnonology 1993; 15: 68-74. Goren A I, Hellman S. Prevalence of respiratory symptoms and diseases in schoolchildren living in a polluted and in a low polluted area in Israel. Enviromnenttd Res 1988; 45: 28-37. Weiland S K. Mundt K A, Ritckman A. Keil U. Self-reported wheezing and allergic rhinitis in children and traffic dens- ity on street of residence. Ann Epidemiol 1994: 4: 243--47. yon Mutius E, Martinez F D, Fritzsch C, Nicolai T, Roell G. Thiemann H-H. Prevalence of asthma and atopy in two arcus of West and East Germany. Am J Respir Crit Care Med 1994; 149: 358-64. Br.'tb~ick L, Breborowicz A, Dreborg S, Knutsson A. Pielik H, Bj~rkst~n B. Atopie sensitization and respiratory symp- toms among Polish and Swedish school children. Cl#t E.rp Allergy 1994; 24: 826-35. Holmgren D, Aberg N, Lindberg U, Engstr6m t. Childhood asthma in a rural county. Allergy 1989; 44: 256-59. Croner S. Kjellman N-I M. Natural history of bronchial asthma in childhood. A prospective study from birth up to 12-14 years of age. Allergy 1992; 47: 150-57. Burr M L. Limb E S. Andrae S. Barry D M J, Nagel F. Childhood asthma in four countries: A comparative survey. htt J Epidemiol 1994; 23: 341-.-47. Skarpaas I J K, Gulsvik A. Prevalence of bronchial asthma and respiratory symptoms in school children in Oslo. Allergy 1985; 40: 295-99. -'~ P~Systt L, Korppi" M, Pietikfiinen M, Juntunen-Backman K. Asthma, allergic rhinitis and atopic eczema in Finnish children and adolescents. Allergy 1991 ; 46:161-65. .,z Brunekreef B (ed.}. Effects of short-term changes in urban air pollution on the respiratory health of children with chronic respiratory symptoms: Study procedures for collaborative study funded by the Commission of the European Com- munities in the framework of the 'Environment" RDT Programme. Wageningen Agricultural University, 1993. ,-3 Florey C du V, Leeder S R. Methods for Cohort Studies of Chronic Airflow Limitation. WHO regional publications. European series No. 12, Copenhagen 1982. "-'~ Brunekreef B, Groot B, Rijcken B et al. Reproducibility of childhood respiratory symptoms questions. Eur Respir J 1992; 5: 930-35. -'~ Wardlaw A J. The role of air pollution asthma. Clin Exp Allergy 1993; 23: 81-96. .'6 Remes S T, Korppi M, Remes K, Pekkanen J. Prevalence of asthma at school age: a clinical population-based study in Eastern Finland. Acta Pediatr Scand 1996; 85: 59-63. -'7 Halken S, H~st A, Nilsson L, Taudorf E. Passive smoking as a risk factor for development of obstructive respiratory disease and allergic sensitization. Allergy 1995; 50: 97-105. .,s Brunekreef B, Groot B, Hock G. Pets, allergy and respiratory symptoms in children, b~t J Epidemiol 1992; 21: 338-42. z9 Verhoeff. A P, van Strien R T, van Wijnen J H, Brunekreef B. House dust mite allergen (Der p I) and respiratory symp- toms in children; a case-control study. Clin ~rp Allergy 1994; 24: 1061-69. (Revised version received October 1996)