INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Recent epidemiologic evidence has suggested a protective effect of diets rich in flavonoids against cardiovascular diseases and possibly cancer (1). A protective effect of flavonoids against chronic lung disease has been hypothesized by Miedema and colleagues (5). They suggested that the stronger association with 25-yr incidence of asthma and COPD observed for solid fruits (apples, pears) than for other types of fruits, may be due to the high level of flavonoids in apples. In the recent study by Butland and coworkers (6), solid fruits were also more strongly associated with pulmonary function than were soft or citrus fruits.

Flavonoids are polyphenols naturally occurring in plant foods. The main sources in a western type diet are tea, fruits, and vegetables (7). Flavonoids have been observed to have both anti-inflammatory and antioxidative effects. By inhibiting lipoxygenase and cyclooxygenase, two enzymes involved in the arachidonic acid metabolism, flavonoids have been shown to reduce the production of two classes of proinflammatory factors (i.e., prostaglandins and leukotrienes) (4, 8- 10), which may be involved in the pathogenesis of COPD. Furthermore, a high intake of other dietary antioxidants such as vitamin C has been observed to be beneficially associated with COPD-related outcomes (11).

For three of the six flavonoid subclasses, i.e., catechins, flavonols, and flavones, data on their content in food are available (12). To determine whether they are protective against COPD, we studied intake of these flavonoids in relation to pulmonary function and COPD symptoms using data from 13,651 men and women examined between 1994 and 1997 in the MORGEN Study (the monitoring project on risk factors and health in The Netherlands).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects and Study Design

The MORGEN Study is a cross-sectional investigation into the prevalence of (risk factors for) chronic diseases in The Netherlands. Each year a new random sample of men and women 20 to 59 yr of age from three towns (Amsterdam, Doetinchem, and Maastricht) was examined through self-administered questionnaires and a physical examination. Between 1994 and 1997, 17,453 subjects were studied. For practical reasons, 1,635 of these subjects did not perform a lung function measurement. Excluded were those with missing dietary data (n = 394) or potential confounders (e.g., smoking habits; n = 198), a technically unacceptable or nonreproducible FEV₁ measurement (n = 1,493), and pregnant women (n = 82). The final study population consisted of 13,651 subjects.

Protocol

Habitual consumption of 178 food items during the previous year was estimated using a semiquantitative food frequency questionnaire developed for the MORGEN Study, which is part of the European Prospective Investigation into Cancer and Nutrition (16). Solid fruit intake (apples, pears) was calibrated using data from a single 24-h recall collected in a subsample (n = 2,689) (19). Vitamin C, -carotene, and energy intake were calculated using the 1996 Dutch food composition table (20) and flavonoid intake was calculated using specific food composition tables (12).

Questions on COPD symptoms were selected from the Dutch part of the European Community Respiratory Health Survey (21). Chronic cough and phlegm were defined as cough/bringing up phlegm during winter on most days for at least 3 mo a year, and breathlessness was defined as shortness of breath when walking on level ground with people of the same age.

FEV₁ was measured by trained paramedics using a heated pneumotachometer (E. Jaeger, Würzburg, Germany). Subjects had to achieve at least three technically acceptable maneuvers of which two had to be reproducible according to criteria of the European Respiratory Society (ERS) (23). The maximum value was used in the analysis.

Height (± 0.5 cm) and weight (± 0.1 kg) were measured and the body mass index (BMI) was calculated (weight/height²). Pack-years of cigarette smoking were calculated:

(years smoked · average number of cigarettes smoked per day)/20.

Statistical Analysis

Analyses were performed using SAS version 6.12 (SAS Institute, Cary, NC). FEV₁ was studied in relation to dietary factors using a multiple linear regression model with FEV₁/height² as the dependent and age, age² and sex as independent variables (24). The results are presented in milliliters for a standard height of 1.70m. Logistic regression analysis was used to study COPD symptoms in relation to dietary factors. The difference in FEV₁ (ml) and COPD symptoms (OR) was presented for the fifth versus the first quintile of dietary intake (Q5-Q1). Potential confounders considered were: age, sex, height (for pulmonary function only), BMI, energy intake, and smoking. Adjustment for educational level was not performed since this is likely to lead to overadjustment of the studied associations.

Tea is the main dietary source of the studied flavonoids. To reduce the correlation between catechin intake and that of flavonols and flavones (from r > 0.90 to r = 0.55), thereby allowing us to study the independent effects, the intake of these substances from other sources than tea was calculated. Vitamin C and -carotene intake were observed to be associated with FEV₁ in the MORGEN Study (24). These factors were added to the adjusted models, to assess potential confounding. Finally, no clear effect modification by smoking status was observed, and the stratified results are therefore not presented.

If the 95% CI did not include the expected value under the null hypothesis (for FEV₁:  = zero and for symptoms OR = 1) the observed differences were considered to be statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The characteristics of the study population are described in Table 1. Mean FEV₁ was 3.6 L, and 16% of the subjects reported one or more COPD symptoms. The average number of pack-years smoked was 10.3 (Table 1).

Average catechin, flavonol, and flavone intake was 58 mg/d (SD = 46). The main sources of catechins were tea (72%), apples (12%), and chocolate (4%), and the main sources of flavonols and flavones were tea (47%), apples (14%), and onions (14%). Intake of catechins and intake of flavonols and flavones were highly correlated (r > 0.90). To be able to study the independent effects, we calculated the intake of catechins (mean = 12 mg/d, SD = 6) and of flavonols and flavones (mean = 8 mg/d, SD = 4) from sources other than tea (r = 0.55).

Catechin, flavonol, and flavone intake was positively associated with age ( = 0.30 mg/yr) and energy intake ( = 1.6 mg/ MJ) and inversely associated with BMI [ = 0.57 mg/(m/kg²)] and pack-years of smoking ( = 0.55 mg/py) (p < 0.001). On average women had a 14.6 mg/d higher intake of the studied flavonoids than did men. Smoking was inversely associated with the FEV₁ ( = 7.6 ml/py) and positively associated with COPD symptoms (OR = 1.03 per py) (p < 0.001).

In Table 2, per quintile of catechin and/or flavonol and flavone intake, the mean FEV₁ and the prevalence of COPD symptoms are given. In further analyses subjects in the first and fifth quintile of intake were compared. The difference between the median intake in these quintiles was 102 mg/d for total catechin, flavonol, and flavone intake and, respectively, 14.8 and 8.3 mg/d for catechin and for flavonol and flavone intake not derived from tea.

Before adjustment for smoking, total intake of catechins, flavonols, and flavones was positively associated with the FEV₁ (_Q5-Q1 = 106 ml) and inversely associated with all COPD symptoms (OR_Q5-Q1 = 0.54-0.57). Adjustment for pack-years of smoking, however, strongly reduced the observed associations, to _Q5-Q1 = 44 ml for the FEV₁ and OR_Q5-Q1 = 0.74-0.90 for COPD symptoms. The association with the prevalence of chronic phlegm was no longer statistically significant (Table 3).

In Table 4, the results of the analyses on the independent effects of catechins and of flavonols and flavones not derived from tea are given. After adjustment for all potential confounders and for flavonol and flavone intake, catechin intake showed a beneficial association with the FEV₁ (_Q5-Q1 = 130 ml) and all COPD symptoms (OR_Q5-Q1 = 0.60-0.72). Flavonol and flavone intake was independently associated with the prevalence of chronic cough only (OR_Q5-Q1 = 0.77).

Additional adjustment for -carotene intake did not alter the associations regarding the FEV₁, presented in Tables 3 and 4, in a relevant way. Adjustment for vitamin C intake reduced the effect of total catechin, flavonol, and flavone intake from 44 to 27 ml (95% CI = 1-53) and the independent effect of catechin intake from 130 to 122 ml (95% CI = 92-152).

Consumption of tea (per 200 ml), the main source of the studied flavonoids, was not associated with the FEV₁, ( = 0.3 ml, p = 0.92) or the prevalence of COPD symptoms (OR = 0.98, p = 0.33), after adjustment for all potential confounders. Intake of solid fruits (mean = 55 g/d, SD = 46) was positively associated with the FEV₁ (29 ml/SD, p < 0.001) and inversely associated with the prevalence of COPD symptoms (one or more: OR = 0.79 per SD, p < 0.001). Similar, although slightly weaker, associations were observed with the intake of, respectively, citrus and other fruits.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Our study was the first to investigate the association of catechin, flavonol, and flavone intake with COPD. Intake of the studied flavonoids was positively associated with the FEV₁ (_Q5-Q1 = 44 ml) and inversely associated with chronic cough (OR_Q5-Q1 = 0.80) and breathlessness (OR_Q5-Q1 = 0.74). Catechin intake (not derived from tea) showed strong beneficial associations with both the FEV₁ (_Q5-Q1 = 130 ml) and all COPD symptoms (OR_Q5-Q1 = 0.60-0.72), independent of the effects of flavonols and flavones. The latter were independently associated only with cough. Of the main dietary sources of the studied flavonoids, solid fruit (apples, pears) intake was, but tea consumption was not, beneficially associated with COPD.

Both for catechins and for flavonols and flavones anti- inflammatory and antioxidant activity has been observed, and a beneficial effect on COPD seems theoretically equally likely (25). We only observed an effect of catechins and not of flavonols and flavones. Furthermore, since tea is the main dietary source of the studied flavonoids, a beneficial effect of tea consumption on COPD would be expected, even based on the effect of catechins alone. The fact that we did not observe such an association could indicate that the observed effect of the catechins is not causal. The association may be confounded by a substance, which intake from foods other than tea is related to that of the catechins.

A few parallels with associations between other dietary factors and COPD can, however, be observed. In the literature there is reasonably consistent evidence for a beneficial effect of vitamin C, a hydrophilic antioxidant, on COPD (11). For vitamin E (lipophilic) the evidence is scarce and inconsistent (11), despite an equally plausible mechanism of action (26). In line with this, we observed a beneficial effect of the catechins that are hydrophilic, but not with the lipophylic flavonols and flavones (25). Furthermore, other fruits and chocolate are the main dietary sources of catechins besides tea and apples. For flavonols and flavones, the main remaining dietary source is vegetable intake. It may not be coincidental that in contrast to fruits, the evidence for a beneficial association of vegetables with COPD is scarce (11).

The two catechins that contribute 60% to the catechin content of tea, and are rare in other foods, are epigallocatechin gallate (ECGg) and epicatechine gallate (ECg). In vitro, the antioxidant capacities of these catechins were observed to be equal to or stronger than that of the other catechins present in tea (8, 27). However, if ECGg and ECg are less biologically active in vivo than are other catechins, this would be an alternative explanation for the fact that no association with tea was observed.

Residual confounding by smoking habits does not seem to explain our findings, since in never-smokers similar associations were observed between intake of the studied flavonoids and COPD (results not shown). Finally, 15 mg of catechins is present in two apples or 30 g of plain chocolate. If valid, the associated increase in FEV₁ (130 ml) may compensate for the effect of smoking 1 pack of cigarettes per day for ± 17 yr (= 130 ml / 7.6 ml/py). In healthy adults the FEV₁ decreases with 20 to 30 ml/yr (28, 29). The observed effect of catechins on FEV₁ is, therefore, also roughly equivalent to the effect of aging 4 to 6 yr.

In conclusion, this first study on intake of catechins, flavonols, and flavones in relation to COPD, suggests a beneficial effect of a high intake of catechins. However, the fact that we observed no association with tea consumption raises the question whether the association is causal. Surprising was also the lack of association between COPD and flavonol and flavone intake. Further research is required to confirm our findings and to get more insight into the biologic effects of different flavonoids.