The epidemiology of childhood asthma in Red Deer and Medicine Hat, Alberta

Correspondence and reprints: Dr Patrick A Hessel, Department of Public Health Sciences, 13-103 Clinical Sciences Building, University of Alberta, Edmonton, Alberta T6G 2G3. Telephone 780-407-7135, fax 780-407-3608, e-mail pat.hessel@ualberta.ca PA Hessel, J Klaver, D Michaelchuk, S McGhan, MM Carson, D Melvin. The epidemiology of childhood asthma in Red Deer and Medicine Hat, Alberta. Can Respir J 2001;8(3):139-146.

A sthma is a common, chronic condition in Canada with an estimated 1.5 million or more persons affected (1). Of these, at least half a million are between birth and 19 years of age. A number of Canadian and international studies have documented increases in the prevalence of asthma (2)(3)(4), in hospital admissions due to asthma (5)(6)(7) and in asthmarelated mortality (5,(8)(9)(10). Several studies have indicated that although hospitalizations due to asthma are increasing in all age groups and for both sexes, children are experiencing the greatest increases (5)(6)(7).
Asthma is multifactorial in nature, and has been linked with environmental and host factors. Some studies have shown that persons with a genetic predisposition to asthma are more likely to develop the disease when exposed to environmental allergens (11). Host factors include age, sex, history of personal allergies, and diagnosed allergies or asthma in a biological parent (12). In industrialized countries, especially those with cold climates, people spend the majority of their time indoors (13). Furthermore, modern home construction produces indoor climates that favour the growth of certain allergens (house dust mite, and mould and/or mildew) (14). Indoor environmental factors potentially related to asthma include environmental tobacco smoke, home dampness or water damage, mould and/or mildew, and pets (12).
This study had three goals: to document the prevalence of asthma among school-aged children in Medicine Hat and Red Deer, Alberta; to understand the host and indoor environmental risk factors associated with the presence of childhood asthma in these cities; and to compare environmental risk factors and asthma triggers associated with asthma between the two communities.

PATIENTS AND METHODS
An initial screening questionnaire was distributed to students in 16 schools in Red Deer (central Alberta) and in 20 schools in Medicine Hat (southern Alberta). The numbers of high schools (grades 10 to 12), junior high schools (grades 7 to 9) and elementary schools (kindergarten to grade 6) were chosen so that the proportion of respondents in each stratum approximated the proportion within the total student population for each city. Schools in each category were selected randomly from lists of Catholic, public and private schools.
The questionnaires were distributed to the students, and completed by parents or guardians, except at one high school in each community where the students completed the questionnaires at school. Current asthma was defined as positive responses to each of the following three questions: Has your child ever had asthma? Was the asthma confirmed by a doctor? Does your child still have asthma? This definition was based on a previous Canadian study (12).
All respondents to the screening questionnaire were asked for permission to perform a follow-up interview. Follow-up interviews were conducted by telephone with the parents or guardians of a random sample of 1198 students who completed the screening questionnaire and had agreed to participate in the follow-up. Of these, 592 met the criteria for asthma from the screening questionnaire and had an 'attack' of asthma within the past year, or were taking asthma medication. These participants were classified as having asthma for the case-control study. There were 443 participants who had no history of asthma (classified as controls for the case-control study). There were 119 participants who had asthma according to the screening questionnaire but were not taking medication for asthma and did not have an attack in the past year. Forty-four participants had missing or inconsistent information. For the purposes of this study, only students with asthma according to the telephone interview criteria (cases) or with no history of asthma (controls) were included in the analysis (n=1035).
as a case or control that might have been part of the disease process (eg, history of pneumonia or bronchitis) were not entered into the multivariate model. Differences in the associations between asthma and risk factors between the communities were tested by using the final regression model for the two communities combined, adding an indicator variable for 'community' and an interaction term for each variable with the 'community' variable (ie, both main effects were included, as well as the interaction term). Each interaction was evaluated in a separate model. Ethics approval was obtained from the Health Ethics Review Board at the University of Alberta, Edmonton, Alberta.

Cross-sectional study:
Response rates for the cross-sectional study were 84% in Red Deer and 73% in Medicine Hat. The prevalence of current asthma among students included in the initial screening survey was 12.8% in Red Deer and 17.0% in Medicine Hat (Table 1). The prevalence was highest for females aged 15 to 19 years in Medicine Hat (20.0%) and lowest for females aged 5 to 9 years in Red Deer (8.3%). The prevalence for children younger than 15 years of age was higher in boys; however, girls had a higher prevalence in the 15-to 19-year-old age group. Case-control study: Approximately two-thirds (65%) of the respondents to the screening survey agreed to be contacted for the follow-up survey. Of those who were contacted for follow-up, 97% completed the interview. Asthma was more common among male than female respondents ( Table 2). Male sex was a significant risk factor in Red Deer (OR 1.7, 95% CI 1.2 to 2.4) and for both communities combined (OR 1.5, 95% CI 1.2 to 1.9). Positive histories of allergies and respiratory conditions were significantly more common among case subjects than among control subjects in

Asthma (n [%]) Controls (n [%]) Asthma (n [%]) Controls (n [%]) Asthma (n [%]) Controls (n [%])
Males both communities (Table 3). Case subjects were also more likely than controls to have a maternal or paternal history of asthma and allergies. ORs for personal history variables varied from 9.8 (95% CI 7.1 to 13.6) for a history of a serious respiratory infection before the age of five years to 2.7 (95% CI 2.1 to 3.5) for a history of skin allergies or eczema (Table 4). Maternal asthma was the family history variable most strongly associated with childhood asthma. The OR for maternal asthma was higher in Medicine Hat than in Red Deer. Although personal smoking was less common in children with asthma than in control subjects, exposure to passive smoke did not differ overall or in Red Deer (Tables 5 and 6); however, children with asthma in Medicine Hat were marginally less likely than control subjects to be exposed to passive smoke. Children with asthma (especially in Red Deer) were more likely to have a gas stove in the home. Cats were less common in the homes of children with asthma.
The case-control difference was more pronounced in Medicine Hat. Children with asthma were more likely to have mould and/or mildew in the home in the past year, especially those in Red Deer. The presence of asthma was not associated with water damage or the age of the house. Children with asthma were more likely than control subjects to have no siblings. The difference was most prominent when comparing the proportions of cases and controls with four or more siblings (versus none) (OR=0.4; 95% CI 0.2 to 0.9). Other environmental factors investigated (but not presented) were parental smoking during childhood, home appliances used for cooking other than the stove, years in the home, type of home, type of home heating, and furry pets other than cats or dogs. None of these were associated with the presence of asthma.
Variables significantly related to asthma in the bivariate analysis were examined in a multivariate logistical regression analysis (

Asthma (n [%]) Control (n [%]) Asthma (n [%]) Control (n [%]) Asthma (n [%]) Control (n [%])
Personal  in Tables 3 and 4 were not included in this analysis. Variables such as personal smoking and presence of cats in the home were also excluded. Although these are known risk factors for asthma, some individuals may have reduced their exposures to these triggers as a result of having asthma. The analysis suggested that male sex, gas cooking, exposure to mould or mildew in the past year, maternal and paternal asthma and allergies, and the number of siblings (especially four siblings versus none) were significantly associated with asthma. Gas cooking and maternal allergies were only marginally significant in this analysis. In Red Deer, male sex, exposure to mould or mildew in the past year, maternal asthma, paternal allergies and/or eczema, and number of siblings were significantly associated with asthma. In Medicine Hat, family histories of asthma and allergies were the only factors associated with asthma in addition to age and sex. Maternal asthma remained the most significant factor in both cities.    When interaction terms were added to the final model for both communities (from Table 7) to compare risk factors between the communities, the interaction terms for mould or mildew exposure and maternal asthma were significant (P=0.04 and P=0.01, respectively).

Risk factor Asthma (n [%]) Controls (n [%]) Asthma (n [%]) Controls (n [%]) Asthma (n [%]) Controls (n [%])
The comparison of asthma triggers for children with asthma in the two communities showed that the most common trigger was colds or influenza (86%), followed by exposure to cigarette smoke (65%) and physical activity (65%) (Table 8). The triggers with the largest differences between the communities were airborne allergen exposures, specifically moulds, burning field stubble and air pollution; all were more commonly cited as triggers in Red Deer.

DISCUSSION
The cross-sectional study had the advantage of drawing from a large population of children with good response rates so that prevalence estimates were reliable. Similarly, the case-control follow-up had sufficient power to detect ORs of approximately 1.5 or greater.
The study was limited by the use of self-reports of asthma, symptoms, and personal and family history. Although selfreports are typical in these studies, measurements of lung function and/or bronchial responsiveness would have been helpful in verifying or at least supplementing the reports.
The use of two different criteria for asthma -one in the cross-sectional study and the other in the case-control follow-up -showed the problems associated with defining asthma in epidemiological studies. There were 711 respondents to the case-control follow-up who were classified in the cross-sectional study as having asthma. Of these, 119 children (16.7%) did not meet the criteria for asthma in the case-control follow-up (ie, they had not had an 'attack' within the past year and were not taking asthma medications). These 119 children were excluded from the casecontrol analysis. The characteristics of this group and their impact on prevalence estimates are being explored (18).
A study conducted by Health Canada examined asthma prevalence in children aged five to 19 years in nine health units across the country (1). The overall prevalence was 12.9%. The figure for Edmonton, Alberta (one of the participating health units), was also 12.9%. This prevalence estimate was comparable with that documented in the present study for Red Deer (12.8%) but lower than the estimate for  Medicine Hat (17.0%). The definition of asthma in the Health Canada study was similar to the definition used in the present case-control follow-up. Health Canada defined asthma in children as a physician's diagnosis of asthma and the presence of least one of the following three criteria in the past 12 months: an attack of asthma, use of asthma medications, or wheeze or whistling in the chest (1). In the present study, only one-quarter of the 119 children who met the criteria for asthma in the cross-sectional study -but not in the casecontrol study -had wheezed within the past year. Three Canadian studies have estimated asthma prevalence using administrative health data. The prevalence of asthma in children in Saskatchewan was estimated at 5.1% for those under five years of age and 4.4% in the five-to 14-year-old age group (3). The prevalence of asthma in Manitoba children ranged from 2.8% to 4.4% for various childhood age groups (4). On the basis of at least one contact with the health care system coded for asthma during a five-year period, asthma prevalence for children in Alberta from birth to nine years of age was approximately 12% and 8% for males and females, respectively (19). The figures for those aged 10 to 19 years were approximately 5.5% for both males and females. When based on at least three contacts with the health care system coded for asthma during the five-year period, the prevalence estimates for males and females from birth to nine years of age were 5% and 3%, respectively, and 2% for males and females aged 10 to 19 years. These estimates would be expected to be low (compared with population surveys) given that some patients may not contact the health care system regularly for asthma and some contacts with the health care system may be coded to another rubric (4).
The patterns of prevalence by age and sex were similar to those found in other studies -asthma was more common among young males and there was a female predominance after puberty (3,4,19,20). The reasons for the higher prevalence in Medicine Hat than in Red Deer are not readily apparent. Outdoor humidity is lower in Medicine Hat, suggesting that asthma prevalence should be lower (12,21). None of the indoor environmental risk factors was significantly associated with asthma in Medicine Hat. O'Hollaren et al (22) found that, among children younger than 17 years of age, sensitivity to Alternaria species was more common among those with asthma than those without asthma. They noted that Alternaria species reached high concentrations in the graingrowing areas of the mid-western United States. Medicine Hat is an area of intense grain farming, whereas the area around Red Deer is characterized by mixed farming and industrial activities. All of the schools, however, were within city limits in both communities. A comparison of immediate skin sensitivity to Alternaria species and/or an examination of specific immunoglobulin E would be required to determine whether Alternaria species or other airborne allergens may account for the relatively high prevalence of asthma in Medicine Hat. A comparison of the diagnostic habits of physicians in the two communities may also provide useful insights.
The host factors found to be associated with asthma in the present study were comparable with those shown in other studies. Associations with maternal and/or paternal asthma and allergies (23)(24)(25), a history of nasal allergies or eczema (21,25), and a history of a serious respiratory infection before five years of age (26) have previously been found. As in most other studies, maternal asthma appeared to be more predictive of asthma than paternal asthma (25). Although maternal asthma was significantly associated with asthma in both communities, the relationship in Medicine Hat was stronger, as indicated by the significant interaction term. Of the indoor environmental factors examined, only mould or mildew in the home and gas cooking were significantly associated with asthma. In both instances, the relation was stronger for Red Deer than for Medicine Hat. Both of these factors have been shown to be associated with asthma in previous studies (20,27). The significant interaction between 'community' and exposure to mould or mildew suggested a stronger relationship in Red Deer than in Medicine Hat. In fact, none of the environmental variables was significantly related to asthma in Medicine Hat, while the ORs for family history variables were generally higher in Medicine Hat than in Red Deer (significantly higher for maternal asthma). The comparison of asthma triggers for children with asthma in the two communities also showed that environmental allergens triggered asthma more commonly in Red Deer, while most nonallergen triggers (eg, colds, physical activity) did not differ between communities.
Children with siblings were less likely to have asthma than children without siblings. The relation was strongest for those with four or more siblings. It has been suggested that early exposure to respiratory infections inhibits the development of atopy and that the presence of siblings may increase opportunities for respiratory infections (28,29). Wickens et al (30) also found an association between smaller family size and risk of asthma.
The results of the present study differed from those of several others (20,31,32) in that environmental tobacco smoke was not associated with asthma. This may have been the result of increased societal awareness about asthma and its triggers, and actions on the part of parents and caregivers to limit exposures for children with asthma. Similarly, there was a negative association between asthma and the presence of cats in the house. Because cats are a known, common trigger of asthma, this result suggests that exposures had been intentionally limited. While this is encouraging, it should be noted that among children with asthma, 22.5% had a smoker in the home, 27.9% were regularly exposed to passive tobacco smoke and 20.0% had a cat in the house. This speaks to the continued need to inform children with asthma, caregivers and others responsible for the environments of children about the importance of limiting potentially harmful exposures.