Environmental factors, parental atopy and atopic eczema in primary-school children: A cross-sectional study in Taiwan

Environmental factors, parental atopy and atopic eczema in

primary-school children: a cross-sectional study in Taiwan

Departments of Occupational and Environmental Medicine, *Internal Medicine and §Dermatology, National Cheng Kung University, Tainan, Taiwan Institute of Environmental Health, China Medical University, Taichung, Taiwan

Department of Dermatology, Kaohsiung Medical University, Kaohsiung, Taiwan

–Department of Environmental and Occupational Medicine, College of Medicine, National Taiwan University (NTU) and NTU Hospital, 1 Sec 1, Jen-Ai Road, Taipei 100, Taiwan

Correspondence

Yueliang Leon Guo. E-mail: leonguo@ntu.edu.tw

Accepted for publication

1 July 2007

Key words

atopic eczema, children, environmental factors, parental atopy, population attributable risk

Conflicts of interest

None declared.

Summary

Background Parental atopy and environmental exposure are recognized risk factors for atopic eczema (AE) in childhood. However, the relative contributions of spe-cific risk factors and the overall contributions of hereditary and environmental exposure remain unexplored.

Objectives To identify risk factors, estimate the population attributable risk (PAR) of environmental exposure, and compare the AE data for boys vs. girls in primary-school children.

Methods During a February to June 2001 cross-sectional, Taiwan-based question-naire survey, we investigated 23 980 children from 22 primary schools, all located within 1 km of an air-monitoring station.

Results The 12-month prevalence of AE was reported as 6Æ1% in boys and 4Æ9% in girls. In both sexes, the risk of AE was strongly associated with parental atopy and perceived ambient air pollution. The presence of cockroaches [odds ratio (OR) 1Æ18, 95% confidence interval (CI) 1Æ00–1Æ40] and visible mould on walls at home (OR 1Æ46, 95% CI 1Æ22–1Æ70) were also significantly related to AE for girls; however, only visible mould on walls (and not the presence of cockroaches) at home was related to AE for boys (OR 1Æ40, 95% CI 1Æ18–1Æ66). While mutually adjusted models were applied, we found adjusted ORs and PARs were similar in boys and girls in hereditary and outdoor environmental factors. The PAR of indoor environmental factors was higher in girls (8Æ4%) than in boys (5Æ5%). There was no interaction between parental atopy and environmental factors. Conclusions Parental atopy contributed more to AE than indoor or outdoor environ-mental factors. Girls may be more susceptible to indoor environenviron-mental factors than boys.

Atopic eczema (AE) is now the most common inflammatory skin disease in children,1,2 and recently the prevalence of childhood AE has increased substantially in many countries.3–7 This increase has been too rapid to be accounted for by changes in gene frequencies. It is also unlikely to be accounted for totally by changes in either clinical diagnostic patterns or increased recognition of associated symptoms by the general population.8It does, however, suggest a role for environmen-tal factors in the aetiology of this evolving epidemic.9–11

Many factors are proven to be associated with AE, including personal factors (smoking habits, genetics, age, sex, nutritional status, number of siblings, lifestyle, allergy status, family history and occupation) and environmental stimuli (house dust-mite, animal danders, moulds, cockroach infestation,

occupa-tional exposure, environmental tobacco smoke (ETS), indoor ⁄ outdoor air pollution, heating systems, aeroallergens and climate).9–20 Both hereditary and environmental factors are believed to contribute to the relationship.21 However, epi-demiological evidence concerning different effects in boys and girls in relationships between environmental factors and AE was insufficient and thought to warrant further investigation.

To date, factors contributing to childhood AE have not been clearly documented in Taiwan. In this study, the relationship between AE and selected risk factors in a population-based sample of Taiwanese school children between the ages of 6 and 12 years was investigated. The population attributable risk (PAR) of each factor was estimated and compared for boys and girls. We also tested the hypothesis that the joint effects

of genetic predisposition and environmental factors on the risk of AE are greater than expected on the basis of their indepen-dent effects.

Materials and methods

The International Study of Asthma and Allergies in Childhood (ISAAC) is a multinational collaborative project developed to investigate variations in childhood atopic diseases at the popu-lation level.22 Between February and June 2001, we modified the ISAAC protocol and conducted a national, cross-sectional, school-based survey of primary-school children. Classroom incentives but not individual incentives were used to encour-age participation. The study protocol was approved by the Respiratory Health Screening Steering Committee of the Tai-wan Department of Health and the Institutional Review Board at our university hospital, and it complied with the principles outlined in the Helsinki Declaration.23 Parents of the school children consented to provide information by questionnaire.

Whether or not a child was deemed to suffer from AE was determined by positive responses to the questions: ‘Has your child ever had an itchy rash which was coming and going for at least 6 months?’ and ‘Has your child had this itchy rash at any time in the past 12 months?’ If both answers were ‘yes’, the parent would be further asked: ‘Has this itchy rash at any time affected any of the following places: the folds of the elbows, behind the knees, in front of the ankles, under the buttocks, or around the neck, ears or eyes?’ In our study, chil-dren who were reported to suffer from a skin rash in the pre-vious year occurring at specific locations were defined as having AE. The ISAAC questions for symptoms of AE used in the present study have been validated in different parts of the world.24–26

In order to compare outdoor air pollution data with ques-tionnaire results, the study population was limited to children attending schools located within 1 km of Taiwan Environmen-tal Protection Agency (EPA) air-monitoring stations. Complete monitoring data for the air pollutants sulphur dioxide (SO2), nitrogen oxides (NOx), ozone (O3), carbon monoxide (CO) and particles with an aerodynamic diameter of 10 lm or less (PM10) were available from the Taiwan EPA. Twenty-two of the 2604 primary schools in Taiwan’s 22 counties were inves-tigated. Stratified sampling by grade was applied in each school. We believe our study population based on 22 different areas covering diverse parts of Taiwan to be representative of Taiwanese primary-school children.27,28

Genetic and environmental determinants

Literature was reviewed on the causes of childhood AE in order to identify the hereditary and environmental risk factors. Our focus was residential factors affected by climate and not directly related to human behaviour, including cockroaches, water damage or visible mould on walls at home. Parental

perception of ambient air pollution level was also considered as an outdoor factor. Parental atopy was a measure of genetic predisposition and defined by reports of the father or mother of the child ever having been diagnosed with AE, asthma or allergic rhinitis. To adjust for possible confounding, we also included host-related variables: the child’s age and sex, mater-nal smoking during pregnancy, the number of siblings at home and the educational level of the household head. Unfor-tunately, neither blood samplings nor skin tests could be per-formed in this large, nationwide study.

Statistical analysis

Previously reported analyses of ecological outcomes have dem-onstrated a larger intercity variation than would be predicted by interindividual variation.28,29 We used two-stage methods to correct for any excess between-site variability. In the first step, a logistic regression model was used to control for indi-vidual-level confounders. In the second step, the community-specific adjusted prevalences of perceived air pollution levels were regressed against the community-specific air pollutants; the regression used weights inversely proportional to the sum of the between-site and within-site variances.

Bivariate logistic models with community clustering were performed to determine associations with AE. All risk factors were categorized into three subgroups of factors—hereditary, indoor environmental and outdoor environmental factors— and we then developed multiple logistic regression models to assess the relative effectiveness of each on AE. Odds ratios (ORs) with 95% confidence intervals (CIs) were calculated after adjustment for potential confounders. PARs were also cal-culated to estimate the contribution of various risk factors for AE. The PAR represents the preventable AE cases if children were not exposed to specific agents or risk factors. PAR was calculated using the formula P(R) 1)/[P(R ) 1) + 1], where P is the prevalence of the exposure and R is the relative risk due to the exposure.30

We assessed potential effect modification by parental atopy by comparing effect estimates for children with and without atopic parents. Individual and joint effects of environmental factor and parental atopy on AE were estimated using indica-tor variables created for each category, omitting the hypothe-sized low–low risk category. Estimates for each of the three exposure categories with the reference group were derived from the same logistic regression model, after adjustment for confounders. Statistical significance was set at P < 0Æ05 based on a two-sided calculation.

Results

Our study surveyed 23 980 children from 22 primary schools. The total response rate was 88Æ8% (10 951 boys and 10 340 girls and their parents). All subjects were between 6 and 12 years old. Overall, 6Æ1% of boys and 4Æ9% of girls were reported to have had AE in the previous year. Younger sub-jects, higher parental education level, and maternal smoking

during pregnancy were found to be associated with the occur-rence of childhood AE (Table 1). In the year 2000 in Taiwan, there were approximately 1Æ04 million boys and 0Æ95 million girls between the ages of 6 and 12 years, and these propor-tions suggest that nationwide about 109 990 children of this age group are affected by AE.

After adjustment for parental education level, the effects of each outdoor air pollutant on parentally perceived ambient air pollution levels were assessed separately, and also expressed as ORs for a change by 1 SD (Table 2). In the regression model where P-values were calculated, statistically significant associa-tions were found for SO2, CO, PM10 and NOx in prevalences of any ambient air pollution and moderate to severe ambient air pollution in the 22 target communities. Relatively weak and nonsignificant associations were noted for O3. Levels of all outdoor air pollutants had relatively stronger predictive effects on the prevalence of perceived moderate to severe air pollution than on the prevalence of perceived mild to severe air pollution level (Table 2).

After adjustment for host factors such as age, parental edu-cation level and maternal smoking during pregnancy, we found parental AE, parental asthma ⁄ allergic rhinitis, and parentally reported ambient air pollution were strongly related to AE in both sexes (Table 3). Girls who lived in homes with cockroaches present were 1Æ18 times (95% CI 1Æ00–1Æ40) more likely to develop AE. The presence of visible mould on walls at home was also significantly related to AE with OR 1Æ40 (95% CI 1Æ18–1Æ66) for boys and OR 1Æ46 (95% CI 1Æ22–1Æ70) for girls. However, water damage at home showed positive but not statistically significant effects for both sexes. When mutually adjusted models were applied, we found adjusted ORs and PARs were similar in boys and girls in hereditary and outdoor environmental factors (Table 3). Stron-ger association between indoor factors—defined as visible cockroaches, water damage or visible mould on walls at home—and AE was noted in girls (OR 1Æ18, 95% CI 0Æ96– 1Æ45) compared with boys (OR 1Æ11, 95% CI 0Æ89–1Æ38). The PAR of indoor environmental factors was higher in girls (8Æ4%) than in boys (5Æ5%). For all the hereditary and envi-ronmental factors we identified, the total PARs were 50Æ7% in the girls and 49Æ8% in the boys of our population.

Of the estimated 109 990 cases of atopic eczema in 6–12-year-old Taiwanese primary-school children, we estimated that more than 36 370 excess cases of AE were attributable to hereditary factors—defined as parental AE, asthma or allergic rhinitis. Exposure to ambient air pollution accounted for approximately 11 420 excess cases. Indoor factors accounted for 7400 excess cases (3489 boys and 3911 girls).

In order to elaborate the potential effect modification, we systematically conducted stratified analyses in categories of parental atopy. The stratum-specific relations were relatively consistent and no significant interaction was found between environmental factors and parental atopy on AE in both sexes (Table 4). This indicated that the environmental exposure and parental atopy both have independent effects on the preva-lence of AE in childhood.

Discussion

The present large questionnaire survey for AE among 6–12-year-old school children in Taiwan linked to local air-moni-toring data demonstrated that outdoor air pollutants were significantly associated with parentally perceived ambient air pollution. Although our study was cross-sectional, we analysed data using a case–control study method, which was very effi-cient compared with a cohort study yielding a similar amount of information. Both parental atopy and environmental expos-ure increased the risk of childhood AE but did not show sig-nificant interactive effects. In addition, it showed girls to be more susceptible to indoor factors than boys. We also found approximately 7400 excess cases of AE attributable to indoor factors, and 11 420 attributable to outdoor ambient air pollution.

Questionnaires have been widely used to assess the preva-lence of chronic illness such as atopic eczema. By using ISAAC questions, researchers have had good results in predicting AE diagnosed by dermatologists in the U.K.,24 Germany25 and Ethiopia.26 We used typical symptoms of AE during the past 12 months as the main outcome measurement in determining risk factors. The overall prevalences in our study were 6Æ1% in boys and 4Æ9% in girls, lower than that reported in other countries, such as Australia (16Æ3%),31 Singapore (20Æ8%)32 and Germany (10Æ5%).33The causality of these substantial dif-ferences is beyond the scope of this report, but some research-ers hypothesize that a Western life style is responsible:14,15 the higher the level of Westernization, the higher the preva-lence of AE.

In our data, younger subjects seem to have a higher preva-lence of AE, a finding consistent with a previous Australian study.31 We also found parental education level to be associ-ated with childhood AE (Table 1): the better educassoci-ated the par-ents, the more concerned they were for the health of their children, and hence the more likely these children were to be reported as having AE.16,17Moreover, maternal smoking dur-ing pregnancy exposes the fetus to more allergens,34 which would contribute to the occurrence of atopic diseases in later life. Because these factors were potential confounders in risk factor analyses, they were controlled as covariates in the fol-lowing regression models. Number of siblings at home was not associated with the prevalence of AE in our study (Table 1). Some studies showed that it was not the number of siblings that mattered, but was the birth order that really had effects on the occurrence of AE in children.9,17,35 Therefore, we did not consider number of siblings as a confounder in further analyses.

ETS and incense burning at home showed negative effects to the occurrence of AE (Table 1), a finding consistent with recent international studies.14,36 One possible explanation could be that ETS and incense use might be reduced by fami-lies with children with AE. Exposure to tobacco or incense might also provide protective effects for atopic diseases through selection mechanisms, as shown in a cross-sectional study.36 Unlike tobacco or incense exposures, the indoor

Table 1 Prevalence of atopic eczema among primary-school children and association with potential risk factors

Risk factor % of subjects N

% with AE OR 95% CI Sex Boys 51Æ4 10 951 6Æ1 1Æ00 Girls 48Æ6 10 340 4Æ9 0Æ79 0Æ70–0Æ89 Age (years) 6–7 21Æ7 4610 5Æ7 1Æ00 8 16Æ4 3487 6Æ3 1Æ10 0Æ91–1Æ32 9 16Æ4 3494 5Æ2 0Æ90 0Æ74–1Æ09 10 17Æ3 3689 6Æ2 1Æ08 0Æ90–1Æ30 11 16Æ4 3492 4Æ9 0Æ84 0Æ69–1Æ03 12 11Æ8 2519 4Æ6 0Æ80 0Æ64–0Æ99

Parental education level (years)

< 9 17Æ1 3637 4Æ0 1Æ00

9–11 46Æ2 9837 4Æ7 1Æ17 0Æ97–1Æ42

‡ 12 36Æ7 7817 7Æ3 1Æ88 1Æ57–2Æ28

Mother smoking during pregnancya

No 97Æ9 20 733 5Æ5 1Æ00 Yes 2Æ1 448 8Æ3 1Æ56 1Æ09–2Æ16 Number of siblingsa 0 11Æ7 2023 6Æ0 1Æ00 1 40Æ3 6967 6Æ1 1Æ03 0Æ84–1Æ27 ‡ 2 48Æ0 8304 4Æ7 0Æ78 0Æ63–0Æ96 AE in fathersa No 97Æ1 20 324 5Æ0 1Æ00 Yes 2Æ9 597 22Æ5 5Æ51 4Æ48–6Æ72 AE in mothera No 97Æ5 20 418 5Æ1 1Æ00 Yes 2Æ5 528 21Æ4 5Æ06 4Æ06–6Æ26 Asthma ⁄ AR in fathera No 82Æ2 17 199 4Æ9 1Æ00 Yes 17Æ8 3722 8Æ4 1Æ81 1Æ58–2Æ07 Asthma ⁄ AR in mothera No 84Æ8 17 752 4Æ8 1Æ00 Yes 15Æ2 3194 9Æ7 2Æ14 1Æ87–2Æ45 ETS at homea No 45Æ9 9674 5Æ9 1Æ00 Yes 54Æ1 11 421 5Æ2 0Æ88 0Æ79–0Æ99

Incense burning at homea

No 54Æ5 11 346 6Æ0 1Æ00

Yes 45Æ5 9458 5Æ1 0Æ85 0Æ75–0Æ95

Cockroaches seen monthly at homea

0 20Æ5 4309 5Æ1 1Æ00

1–2 44Æ2 9289 5Æ5 1Æ07 0Æ91–1Æ26

‡ 3 35Æ2 7397 5Æ9 1Æ17 0Æ99–1Æ38

Water damage at homea

No 93Æ3 19 769 5Æ5 1Æ00

Yes 6Æ7 1423 5Æ6 1Æ01 0Æ79–1Æ26

Number of walls with visible mould at homea

0 74Æ8 15 536 5Æ0 1Æ00

1 17Æ4 3605 6Æ6 1Æ34 1Æ15–1Æ55

‡ 2 7Æ8 1625 8Æ3 1Æ70 1Æ40–2Æ05

Perceived ambient air pollution levela

No 30Æ6 6386 4Æ8 1Æ00

Mild 63Æ0 13 145 5Æ8 1Æ22 1Æ06–1Æ40

Moderate to severe 6Æ5 1350 7Æ3 1Æ56 1Æ23–1Æ97

OR, odds ratio; CI, confidence interval; AE, atopic eczema; AR, allergic rhinitis; ETS, envi-ronmental tobacco smoke.aNumbers of subjects do not add up to total N because of mis-sing data. Some percentages do not total 100 because of rounding. ORs are crude odds ratios for each risk factor.

environmental factors we chose, including cockroaches, water damage and visible mould on walls at home, would not be easily changed by human behaviours and would not show negative effects on AE.

A family history of allergic diseases was associated with an increased risk of AE, suggesting that genetic factors play a cen-tral role in the development of childhood AE.9,17 Genetic markers could also increase susceptibility of children to the effects of environmental factors.9 A recent German cross-sec-tional study17 revealed that paternal and maternal histories of AE were equally strong determinants of AE in children. Besides AE in parents, maternal and paternal atopy were also found to be significantly associated with childhood AE in the U.K.9Our study demonstrated that paternal and maternal AE were stron-ger risk factors for childhood AE in both sexes than were asthma and allergic rhinitis in parents (Table 3). Children with

a parent carrying any atopic disease were also found to have a higher probability of developing AE in later life than those who lived at home with the selected environmental factors. In the mutually adjusted models, hereditary factors—defined as parental AE and asthma ⁄ allergic rhinitis—also possessed the highest attributable risks, which were consistent with 36 370 excess cases annually in Taiwan.

The ecological exposure assessment had many advantages in our study. The density of primary schools in Taiwan is very high, and almost all the surveyed children attended schools within 1 km of their homes. Monitoring stations located near the schools were therefore also likely to be near the children’s homes, and thus provided good indicators for both school and home exposure. A Taiwanese study has suggested that parental ranking of the air pollution level was a good pre-dictor for childhood asthma,27,37 which also demonstrated an

Table 2 Odds ratios (ORs) and 95% confidence intervals (CIs) for the relationship between each outdoor air pollutant and prevalence rates of perceived ambient air pollution level in 22 communities

Mild to severe air pollution Moderate to severe air pollution

OR 95% CI P-value OR 95% CI P-value SO2 1Æ42 1Æ16–1Æ73 0Æ002 1Æ62 1Æ24–2Æ13 0Æ001 CO 1Æ37 1Æ14–1Æ65 0Æ002 1Æ56 1Æ21–2Æ02 0Æ002 O3 1Æ05 0Æ81–1Æ35 0Æ72 1Æ11 0Æ79–1Æ55 0Æ54 PM10 1Æ34 1Æ10–1Æ63 0Æ006 1Æ50 1Æ14–1Æ99 0Æ006 NOx 1Æ42 1Æ21–1Æ66 < 0Æ001 1Æ56 1Æ23–1Æ99 < 0Æ001

Prevalence rates are adjusted for parental education level and results are obtained from sin-gle pollutant model. ORs are calculated by using those reporting no ambient air pollution as reference group, and expressed for a change in each pollutant by 1 SD (SO2, 2Æ25 ppb; CO, 158 ppb; O3, 3Æ44 ppb; PM10, 17Æ9 lg m)3and NOx, 8Æ60 ppb).

Table 3 Odds ratios (ORs) with 95% confidence intervals (CIs), mutually adjusted ORs and population attributable risks for parental atopy and environmental factors associated with atopic eczema in primary-school children

Boys Girls Prevalence (%) OR 95% CI PAR (%) aOR 95% CI PAR (%) Prevalence (%) OR 95% CI PAR (%) aOR 95% CI PAR (%) Hereditary factors AE in father 3Æ0 6Æ20 4Æ75–8Æ04 10Æ8 2Æ7 4Æ38 3Æ12–6Æ03 7Æ4 AE in mother 2Æ7 5Æ35 4Æ00–7Æ07 8Æ5 2Æ4 4Æ09 2Æ85–5Æ74 6Æ0 Asthma ⁄ AR in father 18Æ0 1Æ73 1Æ44–2Æ07 11Æ3 17Æ6 1Æ69 1Æ37–2Æ07 10Æ5 Asthma ⁄ AR in mother 15Æ5 1Æ98 1Æ64–2Æ37 12Æ7 15Æ0 2Æ22 1Æ80–2Æ73 14Æ9

Any parental atopy 32Æ5 2Æ65 2Æ26–3Æ11 33Æ9 2Æ60 2Æ21–3Æ05 33Æ2 31Æ6 2Æ63 2Æ19–3Æ16 33Æ1 2Æ61 2Æ17–3Æ14 32Æ9 Indoor environmental factors

Cockroaches 79Æ4 1Æ08 0Æ89–1Æ33 6Æ0 79Æ5 1Æ18 1Æ00–1Æ40 12Æ5

Water damage 6Æ6 1Æ10 0Æ79–1Æ48 0Æ7 6Æ9 1Æ01 0Æ74–1Æ39 0Æ1

Visible mould 25Æ5 1Æ40 1Æ18–1Æ66 9Æ1 24Æ9 1Æ46 1Æ22–1Æ70 10Æ1

Any indoor factor 83Æ4 1Æ11 0Æ89–1Æ38 8Æ4 1Æ07 0Æ86–1Æ34 5Æ5 83Æ6 1Æ18 0Æ96–1Æ45 13Æ1 1Æ11 0Æ90–1Æ39 8Æ4 Outdoor environmental factor

Perceived ambient air pollution

69Æ4 1Æ26 1Æ06–1Æ52 15Æ2 1Æ18 0Æ98–1Æ42 11Æ1 69Æ4 1Æ25 1Æ02–1Æ54 14Æ8 1Æ15 0Æ93–1Æ41 9Æ4

All ORs are adjusted for age, parental education level and maternal smoking during pregnancy. The risk factors are not mutually exclusive and PARs are not additive in this table. PAR, population attributable risk; aOR, mutually adjusted odds ratio; AE, atopic eczema; AR, allergic rhinitis.

apparent dose–response relationship. In this study, we also proved that the parentally perceived ambient air pollution level was associated with outdoor air pollutants and the associ-ation was stronger if a more severe air pollution level was reported (Table 2). Limited studies have been conducted on the effect of air pollution on AE and the results were still inconclusive,14,19but in our study, outdoor air pollution was found to play an important role in the occurrence of AE in childhood (Table 3). The greater the ambient air pollution perceived by parents, the higher the risks of AE in children.

Indoor dampness markers, such as visible mould and water damage, increase the risks of AE in childhood.17,20A German study demonstrated that water-related damage at home was associated with the amount of house dust-mite antigen in the dust vacuumed from the children’s mattresses.38 In Finland, children exposed to mould in a school building were found to have higher risks of IgE elevation.39 Our results also showed significant associations between AE and visible mould at home after adjustment for covariates in both sexes. Although boys had a higher prevalence of AE than girls, when we estimated the effect of any indoor environmental factor on childhood AE, a stronger relationship was noted in girls than in boys (Table 3). The mechanism of such female-led suscep-tibility is not well understood. One possible explanation con-sidered was that girls might have more extensive exposures because they are relatively inactive and spend more time at home and, therefore, are more influenced by the indoor envi-ronment. Sex-differences in the pathogenesis of AE might be partly due to difference in lifestyle, or differences in skin mor-phology and physiology.

Our data did not show any significant interactive effect between parental atopy and environmental factors on the prev-alence of AE (Table 4). In a recent questionnaire survey of Finnish adolescents, Kilpelainen et al. also found that no inter-active effects existed between indoor dampness and parental atopic diseases.20 Parental atopy would not modify the effects of environmental factors on the risk of AE in childhood.

It is difficult to target preventative efforts on childhood AE. Environmental factors showed a relatively small but substantial effect on childhood AE in our study. Only 10Æ4% boys and 8Æ5% girls with parental atopy were reported as having AE. If such efforts were to target only families with a history of atopy, then only 9Æ5% children could potentially benefit. However, it seems easier to eliminate such exposures on a national scale than to attempt to counter hereditary factors. Additional research is necessary to prove that the elimination of indoor ⁄ outdoor environmental exposures will result in lower rates of childhood AE.

Our study has some limitations. This is a cross-sectional study with relatively weak inference in causal relationship. The questionnaire-based assessment based on symptom report-ing is not as precise as doctors’ diagnoses. However, in the presence of a true association, misclassification of AE that was random with respect to other study variables would weaken the observed association rather than lead to false-positive results. In addition, this report is a prevalence study, rather

Tab le 4 Preva lence (%) of atopic eczema (AE) and assoc iation with environmental factors in primary-school children, stratifie d b y parental atop y Environmental factor Pa rental atop y Boys Gir ls AE (%) with factor A E (%) without fact or OR 95% CI P-value fo r interaction A E (%) with fact or AE (%) wi thout factor OR 95% CI P-val ue for interaction Cockro aches Yes 10 Æ51 0 Æ21 Æ05 0 Æ79–1 Æ40 0 Æ88 8 Æ68 Æ31 Æ08 0 Æ79–1 Æ50 0 Æ48 No 4 Æ14 Æ01 Æ02 0 Æ77–1 Æ37 3 Æ42 Æ71 Æ28 0 Æ91–1 Æ84 Wate r damage Yes 10 Æ61 0 Æ31 Æ06 0 Æ66–1 Æ62 0 Æ83 7 Æ48 Æ60 Æ91 0 Æ52–1 Æ47 0 Æ59 No 4 Æ54 Æ01 Æ14 0 Æ71–1 Æ74 3 Æ43 Æ21 Æ10 0 Æ63–1 Æ79 Visib le mould Yes 12 Æ89 Æ51 Æ39 1 Æ10–1 Æ75 0 Æ37 9 Æ28 Æ41 Æ13 0 Æ86–1 Æ47 0 Æ07 No 4 Æ73 Æ91 Æ19 0 Æ90–1 Æ54 4 Æ62 Æ91 Æ62 1 Æ21–2 Æ16 Any indoor factor Yes 10 Æ51 0 Æ01 Æ08 0 Æ80–1 Æ50 0 Æ84 8 Æ58 Æ71 Æ03 0 Æ73–1 Æ48 0 Æ43 No 4 Æ14 Æ01 Æ03 0 Æ76–1 Æ43 3 Æ42 Æ71 Æ27 0 Æ89–1 Æ89 Percei ved amb ient air pollu tion Yes 10 Æ51 0 Æ01 Æ06 0 Æ83–1 Æ38 0 Æ25 8 Æ78 Æ31 Æ06 0 Æ80–1 Æ42 0 Æ49 No 4 Æ53 Æ41 Æ32 1 Æ02–1 Æ72 3 Æ42 Æ81 Æ22 0 Æ91–1 Æ65 Any environmental factor Yes 10 Æ65 Æ32 Æ13 1 Æ10–4 Æ78 0 Æ70 8 Æ67 Æ31 Æ20 0 Æ67–2 Æ39 0 Æ59 No 4 Æ22 Æ41 Æ75 1 Æ02–3 Æ33 3 Æ32 Æ21 Æ52 0 Æ86–2 Æ98 OR, odds rati o; CI, con fidence interval. All ORs are adjusted for age, par ental education le vel and maternal smoking during pregn ancy.

than an incidence study. A small number of families relocated, and some of the factors we studied might have affected the prevalence of AE through effects on disease duration rather than disease incidence. However, our findings are interesting and real, regardless of whether or not the observed associa-tions were caused by effects arising from incidence or dura-tion. In fact, if factors were found to be associated with the prevalence of AE, then they are of major interest in them-selves, irrespective of whether the aetiological mechanism involved the increase in disease incidence or in duration.

Because we were unable to measure personal environ-mental exposures or sensitization to various allergens, such as dust mites, fungi or cockroaches, we might have under-estimated the effects of these indoor factors on childhood AE. Outdoor environmental factors were measured only by perceived air pollution level, which must be an imprecise way of measuring the influence of environmental exposure. Another potential source of bias was in the interpretation of parental history for atopy as an indicator of a genetic pre-disposition to childhood AE. Although the importance of parental history as a predictor of disease has been demon-strated,9,17 not every child in the family inherits the allergic tendency. Ecological confounders such as urbanization and socialization actually could exist in data analysis and there might be incomplete adjustment and residual confounding. However, more complete personal risk factors are very diffi-cult to obtain in such a large-scale survey. Investigators decided not to try to obtain more personal information as it would have reduced the participation rate and introduced greater bias into the study.

In conclusion, we identified a number of hereditary and environmental factors associated with AE in 6–12-year-old pri-mary-school children in Taiwan. Parental atopy contributed more to childhood AE than environmental factors. Exposure to environmental factors increased the risk of AE in children regardless of the coexisting hereditary factors. Girls may be more susceptible to indoor environmental factors than boys. The present findings suggest that public health policies for eliminating certain environmental factors are needed, which could contribute not only to children’s health but also to medical costs in Taiwan.

Acknowledgments

The authors thank the field workers, teachers and other school staff who supported data collection, and all the parents and children who participated in this study. This study was par-tially supported by grant DOH90-TD-1138 from Department of Health in Taiwan.

References

1 International Study of Asthma and Allergies in Childhood Steering Committee. Worldwide variation in prevalence of symptoms of asthma, allergic rhinoconjunctivitis, and atopic eczema: ISAAC. Lan-cet 1998; 351:1225–32.

2 Williams HC. On the definition and epidemiology of atopic derma-titis. Dermatol Clin 1995; 13:649–57.

3 Grize L, Gassner M, Wuthrich B et al. Trends in prevalence of asthma, allergic rhinitis and atopic dermatitis in 5–7-year-old Swiss children from 1992 to 2001. Allergy 2006; 61:556–62.

4 Galassi C, De Sario M, Biggeri A et al. Changes in prevalence of asthma and allergies among children and adolescents in Italy: 1994–2002. Pediatrics 2006; 117:34–42.

5 Wang XS, Tan TN, Shek LP et al. The prevalence of asthma and allergies in Singapore; data from two ISAAC surveys seven years apart. Arch Dis Child 2004; 89:423–6.

6 Selnes A, Bolle R, Holt J et al. Cumulative incidence of asthma and allergy in north-Norwegian schoolchildren in 1985 and 1995. Pedi-atr Allergy Immunol 2002; 13:58–63.

7 Lee YL, Li CW, Sung FC et al. Increasing prevalence of atopic eczema in Taiwanese adolescents from 1995 to 2001. Clin Exp Allergy 2007; 37:543–51.

8 Hopkin J. Genetics of atopy. Pediatr Allergy Immunol 1995; 6:139–44. 9 Harris JM, Cullinan P, Williams HC et al. Environmental association

with eczema in early life. Br J Dermatol 2001; 144:795–802. 10 McNally NJ, Williams HC, Phillips DR. Atopic eczema and the

home environment. Br J Dermatol 2001; 145:730–6.

11 McNally NJ, Williams HC, Phillips DR et al. Is there a geograph-ical variation in eczema prevalence in the U.K.? Evidence from the 1958 British birth cohort study. Br J Dermatol 2000; 142: 712–20.

12 Annesi-Maesano I, Oryszczyn MP, Raherison C et al. Increased prev-alence of asthma and allied diseases among active adolescent tobacco smokers after controlling for passive smoking exposure. A cause for concern? Clin Exp Allergy 2004; 34:1017–23.

13 Mitchell EA, Stewart AW, ISAAC Phase One Study Group. Interna-tional Study of Asthma and Allergy in Childhood. The ecological relationship of tobacco smoking to the prevalence of symptoms of asthma and other atopic diseases in children: the International Study of Asthma and Allergies in Childhood (ISAAC). Eur J Epidemiol 2001; 17:667–73.

14 Dotterud LK, Odland JO, Falk ES. Atopic disease among school-children in Nikel, Russia, an Arctic area with heavy air pollution. Acta Derm Venereol (Stockh) 2001; 81:198–201.

15 Diepgen TL. Atopic dermtatitis: the role of environmental and social factors, the European experience. J Am Acad Dermatol 2001; 45:S44–8.

16 Heinrich J, Popescu MA, Wjst M et al. Atopy in children and paren-tal social class. Am J Public Health 1998; 88:1319–24.

17 Zutavern A, Hirschw T, Leupoldz W et al. Atopic dermatitis, extrin-sic atopic dermatitis and the hygiene hypothesis: results from a cross-sectional study. Clin Exp Allergy 2005; 35:1301–8.

18 Schafer T, Heinrich J, Wjst M et al. Indoor risk factors for atopic eczema in school children from East Germany. Environ Res 1999; 81:151–8.

19 Dotterud LK, Odland JO, Falk ES. Atopic diseases among adults in the two geographically related Arctic areas Nikel, Russia and Sor-Varanger, Norway: possible effects of indoor and outdoor air pollution. J Eur Acad Dermatol Venereol 2000; 14:107–11.

20 Kilpelainen M, Terho EO, Helenius H et al. Home dampness, cur-rent allergic diseases, and respiratory infections among young adults. Thorax 2001; 56:462–7.

21 Kluken H, Wienker T, Bieber T. Atopic eczema ⁄ dermatitis syn-drome—a genetically complex disease. New advances in discover-ing the genetic contribution. Allergy 2003; 58:5–12.

22 Asher MI, Keil U, Anderson HR et al. International Study of Asthma and Allergies in Childhood (ISAAC): rationale and methods. Eur Respir J 1995; 8:483–91.

23 41st World Medical Assembly. Declaration of Helsinki: recommen-dations guiding physicians in biomedical research involving human subjects. Bull Pan Am Health Organ 1990; 24:606–9.

24 Williams HC, Burney PGJ, Pembroke AC et al. Validation of the U.K. diagnostic criteria for atopic dermatitis in a population set-ting. U.K. Diagnostic Criteria for Atopic Dermatitis Working Party. Br J Dermatol 1996; 135:12–17.

25 Kramer U, Schafer T, Behrendt H et al. The influence of cultural and educational factors on the validity of symptom and diagnosis questions for atopic eczema. Br J Dermatol 1998; 139:1040–6. 26 Haileamlak A, Lewis SA, Britton J et al. Validation of the

Interna-tional Study of Asthma and Allergies in Children (ISAAC) and U.K. criteria for atopic eczema in Ethiopian children. Br J Dermatol 2005; 152:735–41.

27 Lee YL, Lin YC, Hsiue TR et al. Indoor ⁄ outdoor environmental exposures, parental atopy, and physician-diagnosed asthma in Taiwanese schoolchildren. Pediatrics 2003; 112:e389–95.

28 Lee YL, Shaw CK, Su HJ et al. Climate, traffic-related air pollutants, and allergic rhinitis prevalence in middle-school children in Taiwan. Eur Respir J 2003; 21:964–70.

29 Dockery DW, Cunningham J, Damokosh AI et al. Health effects of acid aerosols on North American children: respiratory symptoms. Environ Health Persp 1996; 104:500–5.

30 Coughlin S, Benichou J, Weed D. Attributable risk estimation in case–control studies. Epidemiol Rev 1994; 16:51–64.

31 Marks B, Kilkenny M, Plunkett A et al. The prevalence of common skin conditions in Australian school students: 2. Atopic dermatitis. Br J Dermatol 1999; 140:468–73.

32 Tay YK, Kong KH, Khoo L et al. The prevalence and descriptive epi-demiology of atopic dermatitis in Singapore school children. Br J Dermatol 2002; 146:101–6.

33 Werner S, Buser K, Kapp A et al. The incidence of atopic dermatitis in school entrants is associated with individual life-style factors but not with local environmental factors in Hannover, Germany. Br J Dermatol 2002; 147:95–104.

34 van Bever HP. Early events in atopy. Eur J Pediatr 2002; 161:542–6. 35 Strachan DP. Allergy and family size: a riddle worth solving.

Clin Exp Allergy 1997; 27:235–6.

36 Hjern A, Hedberg A, Haglund B et al. Does tobacco smoke prevent atopic disorders? A study of two generations of Swedish residents. Clin Exp Allergy 2001; 31:908–14.

37 Lin RS, Sung FC, Huang SL et al. Role of urbanization and air pollu-tion in adolescent asthma: a mass screening in Taiwan. J Formos Med Assoc 2001; 100:649–55.

38 Nicolai T, Illi S, von Mutius E. Effect of dampness at home in childhood on bronchial hyperreactivity in adolescence. Thorax 1998; 53:1035–40.

39 Savilahti R, Uitti J, Roto P et al. Increased prevalence of atopy among children exposed to mold in a school building. Allergy 2001; 56:175–9.