The present authors investigated whether wheezing is less common in children who consume more apples and other fruits.
A population-based survey of 2,640 primary school children aged 5–10 yrs was carried out in Greenwich (South London, UK). Information about asthma symptoms and fruit consumption was obtained by means of a questionnaire.
After controlling for potential confounding variables, eating bananas at least once a day (compared with less than once a month) was negatively associated with current wheeze (odds ratio 0.66; 95% confidence interval 0.44–1.00) and ever wheeze (0.69 (0.50–0.95)), but not with ever asthma (0.80 (0.56–1.14)). Drinking apple juice from concentrate at least once a day (compared with less than once a month) was also negatively associated with current wheeze (0.53 (0.34–0.83)), weakly associated with ever wheeze (0.74 (-0.54–1.02)), but not associated with ever asthma. Consumption of apples, other fruits and orange juice was not significantly associated with asthma symptoms.
No association was found between eating fresh apples and asthma symptoms in the study population, but some evidence was found to suggest that a higher consumption of apple juice from concentrate and bananas may protect against wheezing in children.
The increase in incidence, prevalence and associated medical and economic costs of asthma among children is a worldwide concern 1, 2. Repeat surveys in the UK have shown that the prevalence of current wheezing and diagnosed asthma in children is still high 3, 4. It has been suggested that the rise in asthma may partly reflect changes in the population susceptibility resulting from alterations in diet, especially a fall in antioxidant intake, rather than increasing environmental toxicity 5. A number of observational studies in adults have found an association between low fruit intake and asthma or lower lung function 6–11.
The present authors’ group has reported evidence for a protective effect of apple intake on adult asthma in Greenwich (London, UK) 8 and other communities in South East London 12. Several other studies in children have observed similar beneficial relationships between a higher fruit intake and improvement in lung function, but not asthma symptoms 13–17. The relationship between consumption of apples and other fruits and prevalence of wheeze was examined in a cross-sectional survey of young children attending primary schools in Greenwich.
All 64 schools included in the 2004/2005 Greenwich Primary schools list were invited to participate in the present cross-sectional study. These comprised 46 community and 18 voluntary aided schools spread over the London Boroughs of Greenwich and Bromley. Of these, 36 (56.3%) centres agreed to participate. Invitation letters and information sheets explaining the study in more detail were sent to parents or carers of 5–10-yr-old children (school years 2–5) in April 2005.
Data on asthma symptoms were collected by structured questionnaires, which were completed by parents and returned via the children's schools between April and June 2005. The questionnaire comprised sections on asthma symptoms and risk factors for childhood asthma. The asthma screening questions were based on the International Study of Asthma and Allergies in Childhood (ISAAC) 18.
Definition of outcomes
Primary outcome measures were defined as follows. 1) Current wheeze: answered yes to the question “Has your child had wheezing or whistling in the chest in the last 12 months?” 2) Ever wheeze: answered yes to the question “Has your child ever had wheezing or whistling in the chest at any time in the past?” 3) Ever-asthma: answered yes to the question “Has your child ever had asthma?”
Secondary outcomes included exercise-related wheeze, sleep disturbance due to wheeze and nocturnal cough.
Definition of nutritional exposures: assessment of fruit intake
Consumption of apples and other fruits in the previous 12 months was assessed using a fruit frequency questionnaire. Based on previous studies 19, 20, which had shown that food frequency measures are sensitive, a set of relevant questions were extracted from a food frequency questionnaire used previously by the present authors’ group 8. Questions were asked to determine the intake of fresh apples, apple juice, bananas and other types of fruit. Secondary exposures of interest included soft, stoned, citrus, tinned and tropical fruit grouped together as miscellaneous fruit. Parents were asked to report how often, on average, their children had consumed specific fruit or juices during the previous 12 months. Questions regarding consumption of apple juice distinguished between three types: freshly prepared apple juice, juice from concentrate and any other apple juice. For each fruit or juice type, eight possible responses were available: more than once a day; once per day; 5–6 times per week; 2–4 times per week; once a week; 1–3 times per month; less than once per month; and never. Estimates of daily consumption of these fruits were then calculated as described by Carey et al. 21. To avoid baseline groups <10% and other small categories, the groups were aggregated as follows: 1) consumption of fruit or juice less than once per month; 2) once per month to once per week; 3) 2–6 times per week; and 4) at least once per day.
Logistic regression analysis was used to evaluate the association between asthma symptoms and fruit consumption (in four levels, using “less than once per month or never” as reference) after controlling for potential confounding variables.
Adjustment for confounders
For each fruit exposure (apples per day, pears per day, bananas per day, miscellaneous fruit per day, apple juice from concentrate per day, other apple juice per day, orange juice from concentrate per day, other orange juice/squash per day), a propensity score was defined 22, 23 based on the following confounders. 1) Sex. 2) Age group: 5–6, 7, 8, 9, 10–11 yrs or unknown. 3) Paracetamol exposure group: less than once per week, more than once per week or unknown. 4) Ibuprofen exposure group: less than once per week, more than once per week or unknown. 5) Vitamin, iron or other supplement use: no, yes or unknown. 6) Ever lived on a farm: yes, no or unknown. 7) Mould or mildew in bedroom: no, yes or unknown. 8) Mould or mildew in living areas: no, yes or unknown. 9) Mould/mildew in hallways: no, yes or unknown. 10) Mould or mildew in kitchen: no, yes or unknown. 11) Finance source for home repairs: family, landlord, housing association, council or unknown. 12) Current exposure to passive smoking: no, yes or unknown. 13) Ever exposed to passive smoking: no, yes or unknown. 14) Ethnic group: White British, Black British, Black African, Black Caribbean, Asian-British, Asian, other European, other or unknown. 15) Birth weight group (kg): <2.5, ≥2.5–3, ≥3–3.5, ≥3.5–4, ≥4–4.5, ≥4.5 or unknown. 16) Breastfeeding: no, yes or unknown. 17) Number of parents living with child: 1, 2 or unknown. 18) Number of other children at home: 0, 1, 2, 3, ≥4 or unknown. 19) Mother's educational level: primary school, secondary school, A levels, university, post-graduate or unknown. 20) Father's educational level: primary school, secondary school, A levels, university, post-graduate or unknown.
The propensity score for each exposure was defined using a regression model of the exposure with respect to all the confounders, which was a generalised linear model with a γ-variance function and an inverse link for miscellaneous fruit per day, and an ordinal logistic regression model for all other exposures. The values of each propensity score were grouped into 20 equal-propensity groups. For each outcome and exposure, the present authors fitted the parameters of four logistic regression models, these were: an unadjusted per-unit model, containing a baseline odds for zero exposure and an odds ratio (OR) per unit exposure; a propensity-adjusted per-unit model, containing a baseline odds for zero exposure in each of the 20 propensity groups and a common per-unit OR; an unadjusted grouped model, containing a baseline odds for zero exposure and an OR for each nonzero exposure group; and a propensity-adjusted grouped model, containing a baseline odds for zero exposure in each of the 20 propensity groups and an OR for each nonzero exposure group.
Ethical approval was obtained from Bexley and Greenwich Local Research Ethics Committee. Headteachers of all participating schools also approved the survey.
The parents of 5,470 primary school children were sent questionnaires to complete. A total of 2,640 (48.3%) responded and returned completed questionnaires. Five of these children, whose ages and sex could not be determined, were removed from the analysis.
Current wheeze was reported in 314 (11.9%) children, ever wheeze in 24.5% and a history of ever having asthma in 18%. Exercise-induced wheeze was reported in 7.8% of children and 9% of all children were reported as having both ever-asthma and current wheeze. Overall, 4% of the study population reported wheeze at least once a month and 2.6% had sleep disturbance on one or more nights per week in the previous 12 months. Of those with current wheeze, 207 (66%) wheezed only three or less times in the preceding year.
Onset of asthma symptoms
A total of 51.5% of those with a history of asthma symptoms had their first asthma symptoms in infancy, 40% at 1–5 yrs of age and 8.5% after 5 yrs of age.
Table 1⇓ shows the distribution of background characteristics of children whose parents responded. The mean (range) age of pupils in the study was 8 (5–11) yrs. A total of 40% of children lived in a household where at least one adult had either smoked in the past or was currently a smoker. In univariate analyses, living with a smoker, being male and British, having a family history of asthma and lack of breastfeeding were significantly associated with a history of ever-asthma.
Fruit eating habits
Table 2⇓ shows the consumption of apples, apple juice and bananas by age, sex, ethnicity and parents' educational status. There were significant differences in the proportion of children taking at least one fruit or juice serving a day between ethnic groups, sex and according to parents' educational status, but not across age groups. White British children and children whose parents had less than A-level qualifications consumed less fruit.
Frequency of apple consumption
Table 3⇓ and figure 1⇓ show the association between asthma symptoms and frequency of apple consumption. In univariate analyses, eating apples twice or more in a week, compared with eating them less than once a month, was negatively associated with current wheeze, ever wheeze and ever-asthma. With a propensity-adjusted model, these associations became less significant; whilst there was some evidence for a protective effect of eating apples 2–6 times per week compared with less than once a month, there was no evidence for a significant trend.
In a multivariate logistic regression model, apple juice from concentrate was significantly negatively associated with current wheezing and less strongly associated with ever wheezing, with evidence of a dose–response effect (table 4⇓). There was no significant relationship between other apple (table 4⇓) or orange juice consumption (data not shown) and asthma symptoms after controlling for confounders.
Further analysis of other fruits (pears, bananas and miscellaneous fruits) showed that intake of bananas, but not other fruits, was negatively associated with ever wheeze and current wheeze (fig. 1⇑). There was a weak association with ever-asthma. However, there was no evidence of a dose–response effect. Analysis of intake of fruits as a continuous variable (per portion per day) showed no significant associations with asthma symptoms. The continuous adjusted OR (95% confidence interval) for miscellaneous fruits (0.94 (0.80–1.0)), pears (0.89 (0.54–1.45)) and bananas (0.85 (0.58–1.24)) showed no significant protective effects against current wheeze. Similarly, there were no significant continuous effects for any of the fruits on ever wheeze or ever-asthma (data not shown).
A relationship has not been found between apple intake and asthma in the present cross-sectional study in children. However, there is some evidence to suggest that children who have a higher consumption of apple juice from concentrate and bananas exhibit a lower prevalence of current wheeze than children with a lower intake.
In contrast to studies in adults 8, 9, the present authors have not been able to replicate a significant negative association between apple intake and asthma in children. However, the present negative findings are consistent with previous studies 13–15, which showed no clear relationship between wheezing in children and fresh fruit intake despite a beneficial effect on lung function. It is not clear why the link observed in adult asthma was not replicated in children. The variation in the choice of outcome measures between studies and in the validity and reliability of food frequency questionnaires in children might have been responsible for the differences observed. Although the use of similar questionnaires to collect data on dietary items was considered to be a sensitive method in previous studies 19, 20, none was validated in children. It is possible that the bluntness of the fruit frequency questionnaire caused misclassification. This is especially likely in school age pupils in whom parents might not be able to account fully for fruit intake during school hours. It is also possible that the mild nature of disease in the population studied could have obscured or affected any relationship between fruit eating and asthma symptoms. Furthermore, the low response rate in the present study could have biased the results obtained and since no information is available on the difference between responders and nonresponders, the direction and extent of such bias is uncertain.
The present study observed that consumption of apple juice from concentrate was negatively associated with prevalence of current wheeze. Whilst this association may have occurred by chance, the strong dose–response relationship with current wheeze and the weaker association with ever wheeze favour a causal interpretation. Much of the protective effect of apples has previously been attributed to phytochemicals, which include flavonoids, isoflavonoids and phenolic acids 24, 25. In fact, apples are the largest source of free phenolic acids in people’s diet in the USA and Europe 26–29. Thus, flavonoids in apples and apple juice could plausibly reduce asthma inflammation and consequently lead to improvement in disease severity 8. However, if this were the case, it is unclear why high consumption of fresh apples was not also negatively associated with asthma symptoms in the present population. The current authors had expected to see a stronger effect of fresh fruit intake than juice consumption, as studies 26, 29 have suggested that the processing of apples for juice results in a very significant decrease in phenolic acids. However, a recent study 30 has suggested that the increase in plasma antioxidant capacity after apple consumption may be attributable to an increase in urate, rather than apple-derived antioxidant flavonoids, which tend to be poorly absorbed.
This survey incidentally found that banana intake might be beneficial for asthma symptoms. Bananas have long been recognised for their health benefit. For example, a recent study suggested an inverse relationship between cancer and banana intake 31, 32, but the present study is the first to show a link between wheeze and intake of bananas in young children, although a clear dose–response relationship was not observed. One explanation for this link might be their antioxidant content. Bananas have a higher content of water soluble phenolic acids than other fruits, including apples 31, and could plausibly reduce asthma inflammation. Furthermore, bananas have been shown to increase the absorption of other nutrients and are rich in pro-vitamin A carotenoids, which have been shown to protect against some chronic diseases 33, 34. Additionally, they are one of the best sources of potassium 35, which has been reported to be negatively related to lung volumes and flows in children 36.
Although the present authors have controlled as rigorously as possible for potential non-nutritional confounders using propensity scores, the possibility of residual confounding cannot be excluded. Furthermore, some potential confounders, such as body mass index and other foods and nutrients, were not measured in the present study.
In conclusion, the present authors were unable to show a link between eating fresh apples and asthma symptoms in a population of young children. Further studies are needed to confirm the protective effects of apple juice from concentrate and bananas observed in the present study.
The authors would like to thank A. Wong and M. Tumilty for carrying out the fieldwork, and S. Chinn, R. Rona and D. Jarvis for their inputs at different stages of the present study. The authors are also grateful to the heads of the schools, parents and children that participated in the survey.
- Received July 26, 2006.
- Accepted January 24, 2007.
- © ERS Journals Ltd