Results

The descriptive characteristics of the children studied in the cross sectional survey and in the follow up study by citrus/kiwi fruit consumption are presented in table 1. In the cross sectional study low citrus/kiwi fruit consumption (less than once a week) was associated with male sex (15.7% in males, 13.1% in females), rural area (16.5% in rural/urban areas, 13.0% in metropolitan areas), low paternal educational level (17.5% for elementary school, 10.1% for university degree), and maternal smoking (15.8% for current smokers, 13.5% for never smokers). The associations were similar in subjects participating in the follow up study.

In the cross sectional survey the occurrence of any wheeze and of shortness of breath with wheeze in the past 12 months was 7.6%, and 5.1%, respectively. Prevalence of any wheeze was 9.9%, 7.4%, 7.5%, and 6.6% for less than once a week, 1–2 times per week, 3–4 times per week, and 5–7 times per week, respectively (taking less than once a week as the reference, the crude odds ratios were 0.72, 0.73, and 0.64). For the same categories the prevalence of shortness of breath with wheeze was 6.8%, 4.8%, 5.1%, and 4.6%, respectively (crude odds ratios 0.69, 0.71, and 0.66). After controlling for potential confounders, regular citrus/kiwi fruit consumption was a highly significant protective factor for both outcomes (table 2). There were only minimal differences between adjusted and unadjusted odds ratios. Although the test for trend was significant for all comparisons, the protective effect of fresh fruit consumption was already evident among those in the 1–2 times per week category and no clear dose-response relationship was found. In fact, the likelihood test to assess departure from linearity was significant (p<0.01). All the other symptoms investigated, except for exercise related wheeze, were significantly associated with citrus/kiwi fruit consumption (table 2), with a stronger effect for severe wheeze. The results were practically unchanged when adjustment was made for table salt use (table salt use had a borderline association only with shortness of breath with wheeze, OR = 1.29, 95% CI 0.96 to 1.73).

In the follow up study the occurrence of any wheeze and of shortness of breath with wheeze in the 12 months preceding the second survey were 5.7% and 3.7%, respectively. Low fruit intake recorded at baseline was a strong and independent predictor of these two variables: the crude odds ratios were 0.45, 0.58, and 0.50 for any wheeze and 0.31, 0.41, and 0.36 for shortness of breath with wheeze. The adjusted odds ratios were practically similar and did not indicate a clear dose-response relationship among the three higher levels of intake, thus confirming the findings of the first survey (table 3). All the other symptoms investigated occurred less frequently among those eating citrus or kiwi fruit at least once a week, and statistical significance was achieved for exercise related wheeze, nocturnal cough, and severe wheeze. By contrast, no association was found between fruit intake and the reference question relative to the occurrence of toothache (reported in 10.3% of the subjects). Table salt use was not associated with wheezing symptoms in the follow up evaluation and the results were practically unchanged when adjustment was made for this factor (not shown).

In both the cross sectional and the follow up study the interaction terms related to sex, area of residence, father's education, parental smoking, parental asthma, and a lifetime history of asthma at baseline of the relationship between fruit intake and wheeze were far from significant (p>0.50), although relatively small p values were found for previous history of asthma (p = 0.13 and p = 0.19, respectively, in the cross sectional and follow up studies). When the cross sectional analysis was performed separately in children with or without a previous history of asthma, the protective effect of citrus fruit consumption on the crude prevalence of wheeze in the past 12 months was stronger in asthmatic than in non-asthmatic children: 52.1% of asthmatic children who ate fruit less than once per week had any wheeze and 41.9% of the others (OR = 0.62, 95% CI 0.48 to 0.82) compared with 4.7% and 3.9% for non-asthmatic children (OR = 0.86, 95% CI 0.69 to 1.06). Similarly, in the follow up study, wheezing occurred in 47% of asthmatic children with low fruit consumption and in 29.2% of those who ate fruit at least once a week (OR = 0.46, 95% CI 0.27 to 0.80). In non-asthmatic children wheeze occurred in 3.9% and 2.5%, respectively (OR = 0.67, 95% CI 0.40 to 1.14; table4).

There was a limited power to fully exploit the longitudinal component of the study—that is, to evaluate the occurrence of new symptoms in negative subjects at the baseline. Among 3612 children with no wheeze at the baseline, only for 91 subjects (2.5%) did the parents report wheeze during the 12 months of follow up. However, an association was found between eating fruit at least once a week and the incidence of wheeze (3.4% in 436 children eating fruit less than once per week and 2.4% in the others, OR = 0.69, 95% CI 0.39 to 1.21), which is consistent with the overall findings.

Discussion

Our data show a clear association between a low intake of oranges and other vitamin C containing fruits during winter and an increased risk of wheezing symptoms in children. The finding was confirmed in both the cross sectional and follow up studies. The protective effect does not appear to be dose related. A previous study had suggested a similar relationship,3 but it did not have the statistical power allowed by the large sample examined here: Cooket al 3 reported relative risks by fresh fruit consumption of a similar order to those reported here, although the number of subjects in their reference category was very small. Our data suggest, in particular, that consumption of citrus fruit prevents recurrence of symptoms in subjects with a previous history of asthma. These findings confirm previous evidence in adults that low dietary vitamin C is associated with increased symptoms of bronchitis and wheezing,6 and indications that the association between low consumption of fresh fruit and FEV₁decrement is stronger in already susceptible patients with asthma or bronchitis.1 ,3

The biological plausibility of a protective role of vitamin C in asthma has been extensively reviewed.1 ,10 Reactive oxygen species, released from eosinophils, alveolar macrophages, and neutrophils, seem to play a key role in asthma. They may directly contract airways smooth muscle, stimulate histamine release from mast cells and mucus secretion,10 and interact with α₁-protease inhibitor.22 ,23 Hatch has suggested that vitamin C is the major antioxidant present in the airway surface liquid of the lung, and it may protect against endogenous oxidants as well as against exogenous agents such as cigarette smoke and environmental air pollutants.10 Asthma is associated with an oxidant-antioxidant imbalance and patients with asthma have low plasma concentrations of vitamin C and increased oxidative stress.24 ,25 So far, however, experimental studies conducted in asthmatic subjects have yielded conflicting results regarding the effect of vitamin C supplementation on lung function and bronchial responsiveness.26 ,27

Most of the exacerbations of asthma in children have been associated with upper respiratory viral infections, especially rhinovirus.28 Although the ability of vitamin C supplementation to reduce the incidence of common colds is controversial, there are suggestions that vitamin C supplementation may affect the susceptibility to common cold in subjects with a low intake of vitamin C.29 Moreover, vitamin C reduces the duration of episodes and the severity of symptoms of the common cold.30 In a longitudinal study of adults, biological markers of infection (C reactive protein and α₁-antichymotrypsin concentrations) have been found to be inversely related to serum ascorbate concentration, an indicator of low dietary intake.31 It may be postulated that, in asthmatic subjects, vitamin C modulates the biological response to viral infections. However, given the observed link between a low intake of foods containing vitamin C and a low ventilatory function,3 and that children with lower pulmonary function levels are more likely to wheeze than children with higher levels, we cannot exclude the hypothesis that the association that we have found reflects the occurrence of symptoms in children with lower lung function on average.

We found that the protective effect of citrus fruit consumption did not follow a dose-response relationship. Instead, a positive effect was already evident among children eating fruit at least once a week. It might therefore be possible that wheezing symptoms are more likely to appear only below a threshold of vitamin C intake. There are few studies on the association between vitamin C and asthma symptoms, and there are suggestions that the relationship may be non-linear. Schwartz and Weiss first reported in 1990 an inverse association between bronchitis and wheezing and both dietary and plasma vitamin C levels in a representative sample of the adult population in the USA.6 They found that the relation was more or less linear on a logistic scale—that is, exponential on a linear scale. Troisiet al 7 found no association between dietary vitamin C intake and the incidence of asthma over a 10 year period in the longitudinal Nurses' Health Study. However, they noted that the adult female population under study had a greater than average mean vitamin C consumption, few having a low intake. In a case-control study of adults in Scotland Soutar et al 8 reported a fivefold increased risk of bronchial hyperreactivity for the lowest tertile of vitamin C intake compared with the highest tertile, but no association was found for the intermediate level of vitamin C intake. Finally, Cook and colleagues3 found an extremely high prevalence of wheeze in children who never ate fresh fruit in a cross sectional sample of 2650 children in the UK, but they failed to find a dose-response relationship between wheezing and fresh fruit consumption. Fresh fruit consumption had a beneficial effect on lung function, especially among wheezers.

Information bias is a possible problem in this study since we used parental questionnaires to assess both the exposure and the outcomes. However, questionnaire data on dietary items recorded by the parents are generally considered to be reliable, especially regarding vitamin C intake.32 The possibility that parents may have under-reported fruit consumption because their children were symptomatic is unlikely because (1) the news regarding the possible protective effect of vitamin C on lung health had not been reported in the media at the time of the study; (2) the questionnaire covered a number of other environmental exposures in addition to the dietary items; (3) the cross sectional association was confirmed longitudinally in the follow up study; and (4) no association was found for the reference symptom (toothache) used to evaluate reporting bias in this study.

These analyses were designed specifically to focus on vitamin C, and we tend to interpret our findings as being due to this nutrient. Caution is needed, however, when interpreting the results. An infrequent consumption of fresh fruit may be an indicator of a poor diet, which may lack several other factors (including other vitamins such as A and E) that may also have a protective effect on respiratory health. Vitamin E has been associated with a protective effect against the onset of asthma7 and current wheeze in adults.9 Intakes of vitamins C and E are usually correlated,4 and both are likely to play a synergistic role in the antioxidant defence mechanism.9 Low magnesium intake from a diet based on processed foods has been associated with increased wheezing symptoms, bronchial hyperreactivity, and decreased lung function in adults.33 It has been suggested that omega-3-fatty acids, contained in fish oil, may protect children from asthma.34 Olive oil, an important component of the Italian diet which is rich in vitamin E and other antioxidants, could also have a protective effect.34 ,35 We were able to control only for table salt use, which turned out to be a marginal risk factor. We cannot therefore completely address the issue of confounding from other dietary factors in the present investigation. Although this is a clear limitation, it should be considered that, if confounding is present in this study, it should be from a dietary factor that exerts an even more powerful effect than fruits rich in vitamin C. Finally, a possibility of residual confounding from socioeconomic status should be mentioned since we controlled only for father's education. However, this possibility is unlikely since the protective effect of citrus/kiwi fruit consumption on wheezing symptoms in this study was found even in the lowest category of father's education, even if maternal education was controlled for (data not shown). Moreover, despite the fact that we controlled for several potential confounders, crude and adjusted odds ratios were very similar.

The novel aspects of this study compared with other related published papers should be underlined. The large sample of young children enabled us to find an association which has been previously suggested but never detected. After controlling for several potential confounders, our results suggest that the relation between vitamin C intake and respiratory symptoms may be non-linear and that a history of asthma exerts an effect modification. In the prospective study the explanatory variables were recorded at the start of the period, reducing the possibility of reporting bias and allowing for temporality.

In conclusion, low intake of fruit rich in vitamin C is associated with an increased frequency of wheezing symptoms in children, especially among already susceptible individuals. These findings may have important scientific and public health implications.