Several studies have suggested that the increasing prevalence of symptoms of asthma, rhinitis and eczema, could be associated with dietary factors. In the present paper, a global analysis of prevalence rates of wheeze, allergic rhinoconjunctivitis and atopic eczema was performed in relation to diet, as defined by national food intake data.
Analyses were based on the International Study of Asthma and Allergies in Childhood (ISAAC) data for 6–7 and 13–14 yr old children. Symptoms of wheeze, allergic rhinoconjunctivitis and atopic eczema symptom prevalence were regressed against per capita food intake, and adjusted for gross national product to account for economic development. Dietary data were based on 1995 Food and Agriculture Organisation of the United Nations data for 53 of the 56 countries that took part in ISAAC phase I (1994/1995).
The 13–14 year age group showed a consistent pattern of decreases in symptoms of wheeze (current and severe), allergic rhinoconjunctivitis and atopic eczema, associated with increased per capita consumption of calories from cereal and rice, protein from cereals and nuts, starch, as well as vegetables and vegetable nutrients. The video questionnaire data for 13–14 yr olds and the ISAAC data for 6–7 yr olds showed similar patterns for these foods.
A consistent inverse relationship was seen between prevalence rates of the three conditions and the intake of starch, cereals, and vegetables. If these findings could be generalised, and if the average daily consumption of these foods increased, it is speculated that an important decrease in symptom prevalence may be achieved.
This work was supported by the Health Research Council of New Zealand, the Asthma and Respiratory Foundation of New Zealand, the National Child Health Research Foundation, the Hawke's Bay Medical Research Foundation, the Waikato Medical Research Foundation, GlaxoWellcome New Zealand and Astra New Zealand, as well as GlaxoWellcome International Medical Affairs for funding the regional coordination centres. The Wellington Asthma Research Group is supported by a Programme Grant from the Health Research Council of New Zealand.
The incidence of many noncommunicable diseases (NCDs) throughout the world, has been linked to diet 1, and there is a need for standardised multi-country studies of nutrition and disease 2. Many nutritional factors may influence disease, including changes to the production and availability of food, changing food preferences, increased supply of processed foods for consumption as well as supplies for food aid, use of fertilisers and pesticides, land availability, increased prices of staple foods, urbanisation, migration, economic factors and market fluctuations 3–6. Traditional nutritional patterns and patterns of health are changing as are other aspects of lifestyles of populations. Over the past few decades, “Westernization” has seen many countries moving away from the traditional diet of locally grown foods to a more western diet 4. In some countries, these patterns have been shown to be linked to the increase of NCDs 5–7 including obesity, cardiovascular diseases and some cancers 1, 6, 8. Others have reported a lower incidence of chronic disease to be associated with the consumption of a more traditional form of diet 9–12.
Asthma, a chronic respiratory disease, is classified as an NCD but there has been little study of its relationship to dietary factors. Recent hypotheses have focused on potential relationships between food content and the airway inflammatory response 13, 14. Decreased consumption of fruit and vegetables in the United Kingdom has been postulated to be associated with an increase in prevalence of asthma and rhinitis through their antioxidant properties, which may protect against inflammation 13, 15. Similarly, increased consumption of linoleic acid, found in polyunsaturated fatty acids (PUFAs), is thought to be linked to asthma, eczema and allergic rhinitis 12, 16. It is postulated that an increase in consumption of linoleic acid leads to an increase in the synthesis of prostaglandin-E2 (PGE2), resulting in allergic sensitisation 14. More recently, a European investigation indicated an association between trans-fatty acids and the prevalence of childhood asthma, rhinitis and eczema 17. Oily fish consumption has been associated with reduced risk of children having asthma and previous studies postulated that increased consumption of oily fish, rich in omega‐3 fatty acids, reduces the synthesis of proinflammatory cytokines 18. Migration and food aid (including processed food) have been associated with the rapid increase in asthma prevalence, suggesting a potent environmental factor 6, 19, 20. Thus, in seeking explanations for the large world-wide variations in the prevalence of asthma, rhinitis and eczema, and the increases within populations over the last few decades, dietary factors are a potentially important field for study. The findings of an ecological analysis of the relationship between per capita food intake and the prevalence of symptoms of asthma, allergic rhinoconjunctivitis and atopic eczema are reported here.
Sources of data
Symptom prevalence data
The International Study of Asthma and Allergies in Childhood (ISAAC) programme 21, has reported major world-wide variation in the prevalence of symptoms of asthma, allergic rhinoconjunctivitis and atopic eczema 22–25. The methods have been described in detail elsewhere 21 and they are mentioned only briefly here. ISAAC Phase One studied a total of 721,601 children of two age groups (6–7 and 13–14 yr olds) from 156 collaborating centres in 56 countries. Each centre was responsible for obtaining its own funding and obtained the necessary approval from an ethics committee before the start of the study. Three one-page core questionnaires on asthma, rhinitis and eczema symptoms for both age groups were used and an asthma symptoms video questionnaire used for the older age group. The basic question used for assessing asthma symptom prevalence was “Have you had wheezing or whistling in the chest in the last 12 months?” (current wheeze). The prevalence of asthma severity was assessed by using the question “In the last 12 months, how often, on average has your sleep been disturbed due to wheezing?” and identifying respondents who reported such sleep disturbance “one or more nights per week”. The allergic rhinoconjunctivitis prevalence analyses were based on positive answers to two questions; “In the past 12 months have you had a problem with sneezing or a runny or blocked nose when you DID NOT have a cold or the flu?” If yes: “In the past 12 months has this nose problem been accompanied by itchy watery eyes?” The atopic eczema prevalence analyses used positive answers to three questions relating to an itchy rash: “Have you ever had an itchy rash which was coming and going for at least six months?” If yes: “Have you had this itchy rash at any time in the last 12 months?” If yes: “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?” The older age group self completed the questions and also answered a video questionnaire, for which the first scene, wheezing at rest, was used for analysis. Questions for the younger age group were reworded accordingly and answered by the parent/guardian of the child.
Food Balance Sheets
We used the 1995 food balance sheet and food supply data, from the Food and Agriculture Organisation (FAOSTAT) website 26 to obtain comparative figures of food and nutrient intake among the large number of countries who took part in ISAAC Phase One. FAOSTAT data was obtained for 53 out of the 56 ISAAC countries (all except Oman, Singapore and Taiwan). FAOSTAT data describe amounts of food produced per country, and amounts exported and imported, but wastage is not accounted for. Food consumption per capita is calculated as the sum of food production and food imports, minus food exports, divided by the population, and the resulting food balance sheets subjected to rigorous internal and external consistency checks 26. This information was analysed using an international Food and Nutrient database (www.healthnz.co.nz) which estimated per capita consumption of macro nutrients and micro nutrients in appropriate units or as a percentage of total energy consumption (%TEC) corrected for edibility obtained from food composition tables 27, 28.
The data for the three conditions and dietary variables were entered onto a computer and analysed using PC SAS (SAS Institute, Cary, NC, USA). Analyses of dietary variables were performed both by units of kg and by %TEC. The results given are for %TEC unless otherwise stated, to enable a focus on dietary patterns rather than total food intake (which may be more prone to inaccuracies due to wastage or miss reporting). Mixed model regression analyses were carried out to incorporate the different structure of each of the data sets, with FAOSTAT data representing whole countries and ISAAC data using centres within countries. All analyses were repeated with gross national product (GNP) in the model to account for the level of economic development of countries, a factor related to symptom prevalence 29. The analysis did not examine combinations and interactions of the diet variables.
There were 53 countries with both FAOSTAT and ISAAC data with a wide range of per capita food supply. TEC (expressed as kilocalorie (kCal) for consistency with the data obtained from FAOSTAT) ranged from 1602–4013 kCal·day−1; protein 48–119 g·day−1 (5–14%TEC); carbohydrate from 311–587 g·day−1 (42–74%TEC) and fat 24–181 g·day−1 (13–43%TEC). Table 1⇓ shows these and the range of other dietary variables.
The results of the regression analysis for the 13–14 yr age group are shown in table 2⇓. For kCal and protein, a negative pattern of association with wheeze (current and severe), allergic rhinoconjunctivitis and atopic eczema was seen, significantly negative between kCal and atopic eczema. Carbohydrate showed a negative association, except with atopic eczema, which became positive. Fat and saturated fat showed a generally consistent positive association (GNP adjusted data).
Starch (g), total fibre (g), calories from cereal and rice (%TEC), protein from cereals and nuts (g), total seafood (g), and all fresh/frozen fish (g), showed a consistent negative association with current and severe wheeze, allergic rhinoconjunctivitis and atopic eczema. These negative associations were significant between starch and current wheeze; total fibre and current wheeze; calories from cereal and rice and current/severe wheeze and allergic rhinoconjunctivitis; protein from cereals and nuts and current wheeze, allergic rhinoconjunctivitis and atopic eczema; total seafood and severe wheeze; all fish and severe wheeze. The strongest findings were for calories from cereals and rice. Figure 1⇓ shows the %TEC of calories from cereals and rice by country, ranging from 21% (Germany) to 72% (Ethiopia). Figure 2⇓ plots, for the 13–14 year age group, the percentage of current wheeze against calories from cereals and rice as %TEC, expressed as averaged country data. As an example, an increase in calories from cereal and rice of 10%TEC was associated with a decrease of 3.2% (3.1% if not adjusted for GNP) for current wheeze, 0.4% (0.2%) for severe wheeze, 2.4% (1.9%) for allergic rhinoconjunctivitis and 0.6% (0.9%) for atopic eczema.
For vegetables (g and %TEC), and vegetable derived nutrients of vitamin A (μg), E (mg), saturated fat (%TEC), calcium (mg) and protein (g), consistent negative associations were seen with all three conditions. These negative associations were significant for vegetables (g) and current wheeze and atopic eczema; vegetables (%TEC) and current wheeze, allergic rhinoconjunctivitis and atopic eczema; vegetable derived nutrients, vitamin A and current wheeze, atopic eczema; vitamin E and atopic eczema; saturated fat and current wheeze, allergic rhinoconjunctivitis and atopic eczema; calcium and current wheeze, atopic eczema; protein and current wheeze, atopic eczema. For example, an increase of vegetable consumption by l00 g (country range 18–470 g·day−1) was associated with a decrease of current wheeze by 2% (1% unadjusted for GNP) and atopic eczema by 2% (1%).
There was no association between total monounsaturated fatty acid (MUFA) and polyunsaturated fatty acid (PUFA) intake %TEC, as well as MUFA and PUFA from vegetable oils and animals, for current wheeze, severe wheeze, allergic rhinoconjunctivitis and atopic eczema. However, significant negative associations were seen between MUFA from vegetables and current wheeze, atopic eczema and PUFA from vegetables and current wheeze, allergic rhinoconjunctivitis and atopic eczema (GNP adjusted).
For oils, olive oil (g) showed a negative association with current wheeze, severe wheeze and allergic rhinoconjunctivitis, becoming significantly negative with atopic eczema. Soya oil (g) was positively associated with all three conditions and significantly positive with severe wheeze.
The video questionnaire showed generally the same pattern of associations as for current wheeze. Although most were insignificant, significant negative associations were seen with starch and protein from cereals and nuts. For example, a 50 g increase in starch was associated with a 1.4% (1.6% unadjusted for GNP) decrease in symptoms from the video questionnaire.
The 6–7 yr age group (data not shown), showed the same pattern as for the 13–14 yr age group data. Significant negative associations were seen between starch (g) and current wheeze and atopic eczema; calories from cereal and rice (%TEC) and current wheeze and severe wheeze; as well as significant negative associations between fat (g), vitamin E from vegetables (mg), PUFA (%TEC); olive oil (g); oleic fat (%TEC); total fibre (g) and atopic eczema. For example a 10%TEC increase of calories from cereals and rice was associated with a 2.8% decrease in current wheeze and a 50 g increase of starch with a 2.7% decrease. A significant positive association was seen between soya oil and current wheeze and severe wheeze (GNP adjusted).
This world wide analysis of food intake and symptoms of asthma, allergic rhinoconjunctivitis and atopic eczema has shown, for the 13–14 yr age group, a consistent negative association between all three atopic conditions and kCals, protein, vegetables and specific vegetable nutrients (vitamin A, Vitamin E, protein, MUFA, PUFA and saturated fats), starch, calories from cereal and rice, total fibre, protein from cereals and nuts, total seafood, all fish (fresh and frozen) and olive oil. There were no consistent associations for other foods. The video questionnaire results showed similar, but weaker, associations with the diet variables. For the 6–7 yr age group the negative and positive associations generally remained as for the older age group, but were also weaker. This may indicate that the diet of 13–14 year old children is more accurately reflected in the average per capita per day intake of food and may account for the stronger associations seen for the older age group.
The analysis used reliable data from standardised sources (ISAAC and FAOSTAT). FAOSTAT data must be interpreted with care, however, as its accuracy depends on the reliability of the underlying statistics of population, supply and utilisation of foods and of their nutritive value. Food balance data provide direct information about the availability of foods, but only indirect information about dietary intake. Furthermore, FAOSTAT is at times, updated retrospectively when more reliable information becomes available. An example would be 1995 census data for a country becoming available in 1999, allowing legitimate changes to the data (already submitted) to be made, or perhaps food groups being separated as more accurate country data becomes available. The associations seen therefore may change slightly once the FAOSTAT data is updated.
The limitations of ecological analyses are well known 30. In particular, an association at the country level may be due to complex biases and may not apply at the individual level, since dietary variations between countries are likely to reflect broader cultural and environmental differences and may not be directly causal. Nevertheless, an ecological approach has certain advantages in examining possible relationships between diet and disease. In so far as diet varies more widely between countries than within them, associations may be detectable when comparisons are made between different countries that would not be evident within any one country 31. Ecological studies also avoid the regression dilution bias that arises in studies of individuals whose diet varies over time.
There was a large between country variability within some food groups, as shown in table 1⇑, illustrating the minimum and maximum range of country food supply per capita per day. There was over a 1500 fold difference between the two countries at the extreme for olive oil consumption, a 25 fold difference for the two countries at the extremes for vegetable consumption and nearly a 3 fold difference for the two countries at the extremes for calories from cereal and rice. The between country variability of calories from cereal and rice, as illustrated in figure 1⇑, shows the country with the lowest consumption (Germany, 21%TEC) to the country with the highest consumption (Ethiopia 72%TEC). Although the country range is relatively small, the analysis showed a significant negative association with current and severe wheeze, allergic rhinoconjunctivitis and atopic eczema.
It is striking that when adjusted for GNP, the negative association of total calories, protein, vegetables, vegetable fats and vegetable micronutrients, starch, total fibre, calories from cereal and rice, protein from cereals and nuts, total seafood, all fish and olive oil was consistent for symptoms of all three conditions, asthma, allergic rhinoconjunctivitis and atopic eczema. There was no consistent pattern for any other food category, so it seems unlikely that these findings simply reflect systematic bias or confounding by other unknown dietary factors. Do these foods have a protective effect in the expression of symptoms of these conditions? A possibility for the association seen between vegetables and vegetable micronutrients is that this could be due to their antioxidant effects. In particular, Vitamins A and E are known antioxidants that are essential in the lung defence mechanisms against reactive oxidant propagation, inhibitor of oxidant propagation and scavenging of hydroxyl radicals and singlet oxygen 13. These antioxidant defences may be important in reducing airway inflammation and lung permeability in response to cigarette smoke, air pollutants and ozone 13, and may also prove to be important in allergic rhinoconjunctivitis and atopic eczema. Although vitamin C may have similar properties 32, no consistent association was found in this analysis. However, it would be premature to focus on specific micronutrients at this stage. By analogy, there is clear evidence that vegetable consumption is a negatively associated cancer risk, but interventions based on specific micronutrient supplementation have been ineffective 33. Thus, other potential mechanisms for a possible protective effect of vegetables also deserve attention. One such possibility is indicated by differences in the intestinal microflora between countries with a high and low prevalence of allergies 34 and between allergic and nonallergic children 35.
Of interest is the strong negative association between calories from cereal and rice and protein from cereal and nuts and the three atopic conditions. Throughout human development dietary patterns have evolved and now vary greatly in different parts of the world. The diet of economically developed countries and urban areas of the developing world are now relatively low in cereals, tubers and other starchy foods and relatively high in sugar, fat, protein of animal origin, salt, meat, dairy produce and alcohol. Plant based carbohydrate-rich diets have been associated with lower rates of other NLDs 9, 10, 11, 33, 36 and it is of interest that this association has been found for asthma, allergic rhinoconjunctivitis and atopic eczema. Although it is speculated that antioxidants contained in cereal, rice and vegetables may be a reason for this protective effect, other mechanisms need to be explored.
Recent exploration of PUFA and the potential for promoting allergic sensitisation and asthma 14, 16, 37, found a positive relationship. The present analysis saw a significant association between vegetable derived PUFA and decreased symptom prevalence of the three conditions, as well as vegetable derived MUFA and decreased symptom prevalence of current wheeze and atopic eczema. Similarly, a relationship between the consumption of oily fish and asthma symptoms in children 18, and fish intake and lung function 38 have been observed, although clinical studies (of fish oil supplements) have not shown the expected benefits on airway functioning [39, 40]. The present analysis showed a consistent negative association between total seafood and all fish (fresh and frozen) and the three conditions, although insignificant except for severe wheeze.
In summary, the consistent findings of a negative association between starch, calories from cereals and rice, protein from cereals and nuts and vegetables and symptoms of asthma, allergic rhinoconjunctivitis and atopic eczema, especially in the older age group, warrants further exploration. Antioxidant effects may be important, although other explanations, such as effects on the intestinal microflora, should also be explored. There has been little exploration of the association between starchy foods and these conditions and although it is generally acknowledged that a diet containing large amounts of fruit and vegetables is associated with improved health, the association with cereal should be explored further. From this analysis, it is speculated that, if the daily per capita amount of calories from cereal and rice consumed were to be increased by 10% of total energy consumption, it may be possible to achieve a 3.2% decrease in the prevalence of current wheeze, a 0.4% decrease in severe wheeze, a 2.4% decrease in allergic rhinoconjunctivitis and a 0.6% decrease in atopic eczema. The relationship seen between the intake of cereals, starch and vegetables and symptom prevalence of asthma, allergic rhinoconjunctivitis and atopic eczema is strong enough to be explored further in strategies for prevention of these conditions. However it is likely that other nondietary factors are also of importance.
The writing committee consisted of P.E. Ellwood, M.I. Asher, B. Björkstén, M.L. Burr, N. Pearce and C.F. Robertson. The ISAAC Steering Committee consisted of, N. Aït-Khaled, G. Anabwani, H.R. Anderson, M.I. Asher (Chair), R. Beasley (Coordinator Phase One and Phase Three), B. Björkstén, M.L. Burr, J. Crane, U. Keil, C.K.W. Lai, J. Mallol, F.C. Martinez, E.A. Mitchell, S. Montefort, N. Pearce, C.F. Robertson, J.R. Shah, A.W. Stewart, D.P. Strachan, E. von Mutius, S.K. Weiland (Coordinator Phase Two), and H.C. Williams. The ISAAC International Data Centre consists of M.I. Asher (Director), T.O. Clayton, P. Ellwood, E.A. Mitchell and A.W. Stewart.
The collaborators in the participating centres, all parents, children, teachers and other staff as well as the many field workers, who participated in the surveys and data collection have contributed enormously to the success of ISAAC. The authors would particularly like to thank them. The authors also wish to thank the many funding agencies that have supported ISAAC, especially the funding bodies who supported the ISAAC International Data Centre. Our grateful thanks to R. Akroyd, Paediatric Dietitian, Starship Children's Hospital, Auckland, New Zealand, for her dietary assistance, to Dr Murray Laugesen, Waiheke Island, New Zealand, who supplied the database for the analysis and to T. Clayton from the ISAAC International Data Centre for his statistical assistance.
- Received September 28, 1999.
- Accepted November 13, 2000.
- © ERS Journals Ltd