Published ahead of print on November 19, 2004, doi:10.1164/rccm.200405-623OC
© 2005 American Thoracic Society doi: 10.1164/rccm.200405-623OC
Sex Differences in the Relation between Body Mass Index and Asthma and Atopy in a Birth CohortDunedin Multidisciplinary Health and Development Research Unit, Department of Respiratory Medicine, and Department of Preventive and Social Medicine, Dunedin School of Medicine, University of Otago, Dunedin, New Zealand; and Firestone Institute for Respiratory Health, Department of Medicine, McMaster University, Hamilton, Ontario, Canada Correspondence and requests for reprints should be addressed to R. J. Hancox, Dunedin Multidisciplinary Health and Development Research Unit, Department of Preventive and Social Medicine, Dunedin School of Medicine, University of Otago, PO Box 913, Dunedin, New Zealand. E-mail: bob.hancox{at}otago.ac.nz
Rationale: Several studies have identified an association between asthma and obesity in women. It remains unclear if this association is due to genuine asthma or to symptoms caused by overweight, at what age the association develops, and whether it is confined to females. Objective: To explore the relations between body mass index, asthma, and atopy in a birth cohort of approximately 1,000 individuals. Methods: Information on asthma and measurements of lung function, airway responsiveness, and atopy were obtained on multiple occasions between ages 9 and 26. Associations between these outcomes and body mass index were analyzed using generalized mixed linear regression models. Further analyses adjusted for potential covariates including breastfeeding, birth order, parental asthma, and personal and family smoking history. Main results: Body mass index was positively associated with asthma, wheeze, asthma treatment, atopy, immunoglobulin E, and inversely with the FEV1/FVC ratio in females. There was no significant association with airway responsiveness to methacholine or salbutamol. There was little evidence of an association between body mass index and asthma or atopy in males. Analyses adjusting for potential covariates showed similar findings. Asthma was not associated with a raised body mass index in childhood and childhood asthma did not lead to being overweight as an adult. Conclusions: A raised body mass index is associated with asthma and atopy in women but not men. Population attributable fraction calculations estimate that 28% (95% confidence interval 7–45) of asthma developing in women after age 9 is due to overweight.
Key Words: birth cohort • obesity • prevalence • respiratory epidemiology In recent decades, many parts of the world have observed marked increases in the prevalence of asthma (1–6) and obesity (7). Both conditions result from complex interactions between genetic traits and environmental influences. Although the environmental changes leading to the increase in obesity appear obvious, the causes of the increased prevalence of asthma remain obscure. Several studies have indicated that there may be a link between asthma and obesity (8). Although some commentators have questioned this association (9, 10), others have suggested biologically plausible mechanisms to explain it (8, 11). An association between asthma and obesity has been observed in both children and adults. In several studies, the association has been stronger, or only observed, in females (12–21). A causal relation between asthma and obesity is supported by data from cohort studies. Camargo and colleagues (22) and Beckett and colleagues (16) found that an increase in body mass index (BMI) is associated with an increased risk of developing asthma in women. Castro-Rodríguez and colleagues (17) found that girls who became overweight between ages 6 and 11 were more likely to have new asthma symptoms, peak flow variability, and bronchodilator responsiveness at age 11 or 13. This association was not observed in boys. Similarly, Gold and colleagues (23) reported that a higher baseline BMI and a greater increase in BMI were associated with developing asthma in 6- to 14-year-old girls followed for 5 years. The association in boys was more complex, with both the smallest and greatest changes in BMI being associated with asthma. However, our understanding of the issue remains incomplete (11). Many surveys have relied on responses to questionnaires to indicate asthma. This raises the possibility that respiratory symptoms caused by obesity, such as dyspnea, may have been mistaken for asthma. For example, Schachter and colleagues found significant associations between obesity and symptoms of wheeze and diagnosed asthma but not atopy, airflow obstruction, or airway responsiveness in adults (24). It is also uncertain if the association between overweight and asthma is confined to females. Gilliland and colleagues found that overweight boys had a higher risk of developing asthma over a 5-year follow-up than did overweight girls (25). Among the few studies to have used an "objective" marker of asthma, the European Community Respiratory Health Survey found an association between BMI and bronchial responsiveness to methacholine in males with only a weak and insignificant association in women (26), whereas Celedón and colleagues reported similar increases in symptomatic airway responsiveness to methacholine in both overweight and underweight Chinese men and women (27). We have explored the relation between BMI and measures of asthma, lung function, airway responsiveness, and atopy in a birth cohort of approximately 1,000 New Zealanders followed into adulthood. Because the association between asthma and body weight could be confounded by a number of other factors (e.g., we have previously found in this cohort that breastfeeding during infancy is associated with a lower risk of overweight and a higher prevalence of asthma [28, 29]), we further examined whether the association was confounded by a number of potential covariates.
The Dunedin Multidisciplinary Health and Development Study is a cohort study of 1,037 children (52% male) born between April 1972 and March 1973 (30). Follow-up assessments have been conducted at ages 3, 5, 7, 9, 11, 13, 15, 18, 21, and 26 years when 980 (96%) of 1,019 living study members participated. The Otago Ethics Committee approved the study. Written informed consent was obtained at each assessment. At age 9, the accompanying adult was questioned about current and previous asthma, wheeze, and cough (31). This information was updated at subsequent assessments (32). Current asthma is defined as diagnosed asthma with symptoms within 12 months; asthma with airway hyperresponsiveness as both current asthma and airway hyperresponsiveness (see the following paragraph); current wheeze as more than two episodes of wheezing lasting more than 1 hour in the previous year; and asthma treatment as any bronchodilator, corticosteroid, or cromoglycate medication. Current smoking was defined as smoking daily for at least 1 month during the past year. At each age, height and weight in light clothing without shoes were measured to calculate BMI in kg/m2. Spirometry was performed at each assessment from age 9. Airway responsiveness was measured using methacholine at ages 9, 11, 13, 15, and 21 or salbutamol (albuterol) at ages 18 and 26 (33). A provocative concentration of methacholine to induce a 20% fall in FEV1 of 8 mg/ml or less, or an increase in FEV1 of 10% or greater, indicated airway hyperresponsiveness. Skin-prick testing at ages 13 and 21 included house dust mite (Dermatophagoides pteronyssinus), grass, cat, dog, horse, kapok, wool, Aspergillus fumigatus, Alternaria, Penicillium, and Cladosporium (34). A weal diameter 2 mm greater than saline control was considered positive, and atopy was defined as a positive response to one or more allergens. Total serum immunoglobulin E (IgE) was measured at ages 11 and 21 (35). The adult accompanying the study member at age 7 was asked whether the natural mother or father had asthma (36). Self-reported height and weight was obtained for 81% of mothers and 77% of fathers at the age 11 assessment. Family smoking history was ascertained at age 9. Duration of breastfeeding was reported at the age 3 assessment and verified from a prospective record (29). Birth order was categorized according to the number of older siblings (none, one, two, three, or more).
Statistical Analysis
The results of the generalized mixed model analyses are shown in Table 1. Asthma, wheeze, atopy, and log IgE were significantly associated with BMI in females, whereas there was a significant inverse association with the FEV1/FVC ratio. There was no significant association between BMI and airway responsiveness to either methacholine or salbutamol. In males, there was an association between BMI and wheeze that was of borderline statistical significance, but there were no other significant associations with BMI.
The results of the analyses with adjustment for the potential covariates are shown in Table 2. These show a similar pattern of findings with little change in the magnitude of the odds ratios, coefficients, or p values. In this analysis, the trend to an inverse relation between BMI and airway hyperresponsiveness to methacholine in females was still present. However, there was a significant positive association between BMI and asthma combined with airway responsiveness in females.
Model fitting for the generalized mixed logistic regression analyses was checked using different numbers of quadrature points. At lower numbers of quadrature points (8–16), some of the coefficients varied significantly depending on the number of points chosen. At higher number of points (greater than 20), the results were more stable. Although the size of the coefficients varied according to the number of quadrature points, the findings were similar and the interpretation of the results did not change according to the number of quadrature points used. Further exploration of the data using cross-sectional analyses at each age indicated that there was no association between BMI and any of the outcomes at age 9. Indeed, a diagnosis of asthma at age 9 was associated with a significantly lower BMI at age 26 (mean difference 1.47 kg/m2, 95% confidence interval [CI] 0.33–2.6), with similar differences in males and females. There was a significant association between adult BMI and developing asthma after age 9 in females (Figure 1). Analyses of data at ages 9, 11, 13, 15, 18, 21, and 26 suggest that the association between BMI and asthma in females emerged during the late teens. There were trends to an association between wheeze and BMI at all ages, although this was only significant from age 15 onward (Table 3). Population-attributable fraction calculations suggest that 18% (95% CI 2–32) of all asthma in 26-year-old women, and 28% (95% CI 7–45) of asthma developing since age 9, may be attributable to being overweight (BMI>25) at 26. Attributable fractions were smaller and not significant in males.
The association between asthma and BMI in the parents of the study members was similar to that observed in the study members themselves. That is, asthma was significantly associated with BMI in the mothers of the study members (n = 813, odds ratio [OR] for each unit increase in BMI = 1.06, 95% CI 1.01–1.12), but not in the fathers (n = 770, OR = 0.99, 95% CI 0.93–1.05). (Additional RESULTS are provided in the online supplement.)
This study found that a high BMI is associated with asthma, wheeze, and asthma treatment in females. There were also associations between BMI and atopy on skin-prick testing and serum IgE, and a negative association with the FEV1/FVC ratio. In males, there was an association between a high BMI and wheeze that was of borderline statistical significance, but otherwise the associations were weaker and not significant. This association appears to be due to the development of asthma in females who become overweight during adolescence and early adulthood. Cross-sectional analyses indicate that there was no evidence of an association in girls at age 9. By age 26 the association between BMI and asthma was statistically significant. Our findings confirm and extend those of a number of studies that have reported an association between BMI and asthma in women (13–19, 21). We found no association between a raised BMI and asthma in children; the association appeared to emerge in late adolescence (Table 3). It has been reported that an early onset of puberty and a high BMI are both risk factors for developing asthma during adolescence in females (17, 38). We did not find an association between asthma and age of menarche in this study, nor did it modify the relationship between BMI and asthma (data not shown). There was also no evidence that the association between asthma and adult BMI arose because girls with childhood asthma put on extra weight during puberty. In fact, those who had already developed asthma by age 9 tended to have lower BMIs as adults. The lack of association between BMI and asthma in prepubertal children is consistent with large surveys of 4- to 11-year-old Canadian and American children (17, 39). By contrast, the 1993–1994 survey of the UK National Study of Health and Growth found that BMI was associated with asthma in 4- to 11-year-old children of both sexes (15). An analysis of earlier samples of children in the same study suggested that the association between BMI and asthma is of recent onset (40). The authors speculated that this may reflect a recent change in nutrition or lifestyle increasing the prevalence of both obesity and asthma and thereby confounding the association. If similar changes occurred in New Zealand, this might explain why the association between BMI and asthma in the Dunedin Study was not apparent until the 1987–1988 assessment when the study members were age 15. However, such a sudden change in environmental or nutritional factors seems unlikely, and it is unclear how this explanation would account for the differences in the association between asthma and BMI in males and females. Furthermore, we are able to analyze the relationship between estimated BMI and reported asthma in the parents of the study members using data obtained between 1979 and 1984. This shows a significant association between asthma and BMI in the mothers of the study members, but not in their fathers. This indicates that the association between BMI and asthma has been stable across two generations of New Zealand women. The question remains as to whether this association between BMI and asthma is real. It is possible that there is an apparent association because symptoms of overweight, such as exercise-induced dyspnea, are mistaken for asthma. The definition of asthma used in this study—a self-report of diagnosed asthma with symptoms in the previous year—does not exclude this possibility. In support of this explanation, BMI was not associated with airway responsiveness to either methacholine or salbutamol. In fact, there was a trend to an inverse association between BMI and methacholine responsiveness in females. However, a higher BMI was associated with a lower FEV1/FVC ratio in females, indicating that overweight was associated with airflow obstruction and with respiratory symptoms. This inverse association between BMI and the FEV1/FVC ratio in females was independent of an asthma diagnosis (coefficient –0.07, 95% CI –0.13––0.01). A recent study by Aaron and colleagues (41) found that weight loss in obese asthmatic women was associated with improvements in lung function, including both the FEV1 and FVC, but had no significant effect on airway responsiveness to methacholine. Because airway responsiveness is regarded as a hallmark of asthma (42), the lack of association between BMI and methacholine responsiveness suggests that obese women develop airflow obstruction and respiratory symptoms through nonasthmatic mechanisms. Schachter and colleagues also found that adult obesity was associated with asthma and wheeze, but not airway responsiveness (24). By contrast, the European Community Respiratory Health Survey found obesity to be associated with methacholine responsiveness, but paradoxically this association was stronger in men than women (26). The reason for these differences between studies is unclear. Although the lack of association between BMI and airway responsiveness suggests that that the association between obesity and asthma may be due to diagnostic or reporting bias, other findings indicate that this is unlikely to be the case. First, a more stringent definition of asthma requiring both current diagnosed asthma and significant airway responsiveness to either methacholine or salbutamol was positively associated with BMI in women. Second, there were significant associations between BMI and both atopy and serum IgE levels in females. Previous reports on obesity and atopy are conflicting. Neither Schachter and colleagues (24) nor Jarvis and colleagues (43) found an association between BMI and atopy in adults. By contrast, other surveys found a high BMI to be a risk factor for atopy in 7- to 12-year-old girls (20), both boys and girls ages 4–17 (44), and young adults of both sexes (45). Although we found that the associations between BMI and atopy/IgE were of borderline statistical significance after adjusting for potential confounding factors, these are well recognized risk factors for asthma and suggest that there may be an association between adiposity and the immunologic responses typical of allergic asthma. Consistent with several other studies, we did not find a relation between BMI and asthma in males. Why the association between BMI and asthma is more often observed in females is unknown, but may be partially explained by the fact that for a given BMI, women have a greater percentage of body fat than men (46). Thus the relation between BMI and asthma/atopy may be weaker in males and this study may have had insufficient power to demonstrate it. It is possible that a relation between BMI and asthma will emerge as these men age and adiposity increases. An alternative explanation is that the association between BMI and asthma is mediated by female sex hormones (8, 11, 47). Regardless of the underlying cause, these observations may help to explain the widely recognized but poorly understood sex difference in asthma epidemiology. Asthma is more prevalent among preadolescent boys than girls, but females are more likely to develop asthma after puberty. Our findings suggest that part of the explanation for this may be the increase in body fat in females after puberty. Population attributable fractions indicate that 28% (95% CI 7–45) of incident asthma between age 9 and 26 in women may be attributable to becoming overweight by age 26 (BMI > 25 kg/m2). This study has a number of strengths for assessing the changing risks for asthma through the life course. These include frequent assessments during childhood to early adulthood, a low attrition rate, and direct measurement of objective variables including height, weight, lung function, atopy, and airway responsiveness. Even so, our main outcome measure—diagnosed asthma with symptoms in the past year—is self-reported and open to error. Undoubtedly there will be a few study members who misreport this, but we think it is unlikely that reporting errors would bias our findings. Furthermore, reporting error could not account for the findings of lower FEV1/FVC ratios and increased atopy and serum IgE. In summary, we have found a significant association between BMI and the development of asthma and atopy in females. A positive association with atopy and an inverse association with the FEV1/FVC ratio suggest that this association may be due to "real" asthma, rather than a mislabeling of symptoms caused by overweight and obesity. We found little evidence for this association in males. There was no evidence of reverse causation because having asthma as a child was not a risk factor for becoming overweight as an adult.
The authors are grateful to the study members and their parents for their continued support. They also thank Air New Zealand and Dr. Phil A. Silva, the study founder.
The Dunedin Multidisciplinary Health and Development Research Unit is funded by the Health Research Council of New Zealand. The respiratory section of the study was funded by the Health Research Council, the Otago Medical Research Foundation, the New Zealand Lottery Grants Board, and the Asthma Foundation of New Zealand. Collection of data used in this report were also funded by the National Heart Foundation (NZ) and the U.S. National Institute of Mental Health grant MH45070. This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org Conflict of Interest Statement: R.J.H. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; B.J.M. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; R.P. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; D.R.T. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; J.M.G. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; C.R.M. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; J.O.C. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; E.M.F. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; G.P.H. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; M.R.S. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Received in original form May 14, 2004; accepted in final form November 17, 2004
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
METHODS