The role of diet in the aetiology of asthma

Asthma is a respiratory disease characterized by increased airway responsiveness, in ̄ammation and variable air ̄ow obstruction. The prevalence of asthma and associated atopy has increased substantially over the past 30 years in most developed countries [1±4] and appears to have risen in developing countries in relation to the degree of af ̄uence of the population [5±11]. These observations suggest that the prevalence of asthma and atopy is strongly dependent on environmental factors associated with an af ̄uent lifestyle, and have initiated a search for cultural in ̄uences that may be implicated. Although the explanation for the rise in asthma prevalence in developed countries is not known, it has coincided with a marked change in the diet of these nations, with decreased intake of fresh fruit, green vegetables, potatoes, meat, ®sh and milk, as seen in the United Kingdom [12], a trend that is also seen elsewhere in the developed world. Diet is an environmental factor that changes rapidly with the shift from a rural subsistence to an urban lifestyle, and from a controlled to a free market economy. The hypothesis that an alteration in the diet of nations may be causing the changes in the prevalence of asthma was initially proposed by Burney [13] and later developed by Seaton [14], and has since attracted a great deal of interest. Available evidence falls into two broad categories: epidemiological studies examining the role of differences in dietary intake of speci®c nutrients in explaining the distribution and incidence of asthma; and intervention studies utilizing dietary or parenteral supplementation to modify the disease. Identi®cation of a food constituent that in ̄uences the development or severity of asthma permits the possibility of dietary manipulation as a therapeutic intervention. This review will consider the published evidence for the effects of individual nutrients on asthma and the possibilities for treatment, concentrating particularly on sodium, magnesium, antioxidants including vitamins C and E, and fatty acids. The evidence reviewed is derived from various sources, including some large epidemiological datasets which have been used by several independent investigators to explore effects from different nutrients. For convenience, these datasets are summarized now. Four, large, crosssectional studies explored the role of nutrient intake in asthma by using dietary questionnaires to measure intake of food constituents, and different methods of measuring diseases and outcomes in different populations. The First National Health and Nutrition Examination Study (NHANES I) measured forced expiratory ̄ow rate in 1 s (FEV1) in 2526 randomly sampled adults aged 30±70 years from the USA [15], and the Second National Health and Nutrition Examination Survey (NHANES II) assessed the relationship between dietary intake and respiratory symptoms in 9074 adults aged 30 years or older [16]. The MORGEN Study measured FEV1 and forced vital capacity (FVC) in 6555 subjects aged 20±59 years from the Netherlands [17], and we assessed bronchial reactivity and spirometry in 2633 subjects in an adult population aged 18±70 years from Nottingham in the UK [18±20]. There are two large prospective studies of the effect of diet on lung disease. The Nurses' Health Study from The United States contains dietary questionnaire data on 77866 women aged 34±68 years followed since 1980 [21], and is the only prospective study to use diagnosed asthma as an endpoint. The Zuphen Study followed 793 middle aged men in the town of Zutphen in the Netherlands from 1960 to 1985, using an interview with a dietician to assess diet and having an endpoint of chronic non-speci®c lung disease, a diagnosis based on episodes of respiratory symptoms lasting more than 3 months, or a diagnosis of asthma, chronic bronchitis or emphysema [22]. All of these datasets have been used to address dietary and other hypotheses relating to asthma, and different investigators analysing individual nutrient effects have not always adjusted for effects of other nutrients shown to be related to the outcome in question in other studies. It is therefore sometimes unclear whether effects seen in these analyses are truly independent, although effects seen consistently in different datasets are perhaps more likely to be valid. The effects of different nutrients seen in these and other studies are as follows.