European Journal of Clinical Nutrition (2002) 56, 638643 2002 Nature Publishing Group All rights reserved 09543007/02

2 $25.00 www.nature.com/ejcn

ORIGINAL COMMUNICATION Diet and 20-year chronic obstructive pulmonary disease mortality in middle-aged men from three European countries
sa nen3, F Fidanza4, A Menotti1,5, A Nissinen6,7, IC Walda1,2, C Tabak1,2*, HA Smit1, L Ra 1 8 EJM Feskens and D Kromhout
Department of Chronic Disease Epidemiology, National Institute of Public Health and the Environment, Bilthoven, The Netherlands; 2Institute for Risk Assessment Sciences, Division Environmental and Occupational Health, Universiteit Utrecht, Utrecht, The Netherlands; 3Division of Nutrition, University of Helsinki, Helsinki, Finland; 4Nutrition Section, Department of Internal Medicine and Endocrinological and Metabolical Sciences, University of Perugia, Perugia, Italy; 5Division of Epidemiology, School of Public Health, University of Minnesota, Minneapolis, Minnesota, USA; 6Department of Public Health and General Practice, University of Kuopio, Kuopio, Finland; 7Department of Neurology, Kuopio University Hospital, Kuopio, Finland; and 8 Division of Public Health Research, National Institute of Public Health and the Environment, Bilthoven, The Netherlands Objective: To investigate the relation of baseline antioxidant, fruit, vegetable and sh intake with 20 y chronic obstructive pulmonary disease (COPD) mortality in middle-aged men from three European countries. Design: Prospective study (1970 1990). Setting: Five population-based cohorts of middle-aged men from Finland, Italy and The Netherlands. Subjects: A total of 2917 men aged 50 69 y at baseline. Methods: Baseline information on diet was collected using the cross-check dietary history method. After 20 y of follow-up the underlying cause of death of those who died was established centrally. Survival analyses were performed using the Cox Proportional Hazards Model. Results: After adjustment for age, smoking and country, we observed an inverse trend (P-trend < 0.05) of 20 y COPD mortality across tertiles of fruit and vitamin E intake. No trend was observed for vegetables, sh, vitamin C and b-carotene. When modelled continuously, a 100 g increase in fruit intake was associated with a 24% lower COPD mortality risk (RR 0.76, 95% CI 0.60 0.92). For vitamin E intake (per 5 mg) the RR was 0.77 (95% CI 0.55 1.06), after adjustment for age, smoking and country. Additional adjustment for body mass index, total energy intake and alcohol consumption reduced the RR to 0.86 (95% CI 0.69 1.07, P 0.12) for fruit and 0.93 (95% CI 0.65 1.33) for vitamin E. Conclusions: Our results suggest a protective effect of fruit and possibly vitamin E intake against COPD. No effect was observed for intake of vitamin C, b-carotene, vegetables and sh. European Journal of Clinical Nutrition (2002) 56, 638 643. doi:10.1038=sj.ejcn.1601370 Keywords: diet; COPD mortality; longitudinal study; fruit; vegetables; antioxidants; sh
1

Introduction
*Correspondence: C Tabak, Department of Chronic Disease Epidemiology, National Institute of Public Health and the Environment, PO Box 1, 3720 BA Bilthoven, The Netherlands. Guarantor: C Tabak. Contributors: IW conducted the statistical analyses and wrote the initial version of the paper. CT wrote the nal versions of the paper. HAS and EJMF were responsible for the supervision on a day-to-day basis, while DK was at all stages involved at vital moments. AM, AN and DK were part of the team that organised the data collection. All authors were involved in the interpretation of the data and reviewed the nal version of the manuscript. Received 22 January 2001; revised 29 October 2001; accepted 30 October 2001

Several risk factors for the development of chronic obstructive pulmonary disease (COPD) have been identied, among which cigarette smoking is the most important one. High levels of free radicals in cigarette smoke cause direct (tissue oxidation) and indirect (release of oxidising agents and enzymes) damage to lung tissue (Chow, 1993). However, the fact that not all smokers develop signicant airow obstruction suggests that other factors are also involved. Diet is one of the potential factors. Several cross-sectional epidemiological studies have suggested that dietary antioxidants and foods rich in antioxidants (ie fruit and vegetables)

Diet and COPD mortality IC Walda et al

639
may protect the airways against oxidant-mediated damage leading to COPD (Smit et al, 1999; Tabak et al, 1999, 2001a; Butland et al, 2000). Only three longitudinal studies on the relation between diet and COPD-related outcomes have been published. In the Zutphen Study an inverse relationship was observed between baseline consumption of total and solid ( apples, pears) fruit and 25 y incidence of chronic lung disease (Miedema et al, 1993). However, Miedema and coworkers did not observe a relationship with intake of antioxidants (vitamin C and b-carotene). Follow-up in the other two studies was relatively short (5 7 y). Butland et al (2000) observed no statistically signicant longitudinal relationship between vitamin C, vitamin E or apple intake with pulmonary function. However, baseline apple consumption did tend to be inversely related to change in FEV1. Carey et al (1998) observed no relation between average level of fresh fruit intake during follow-up and change in pulmonary function. In addition to the above-mentioned dietary factors with antioxidant capacities, a protective effect against development of COPD has been suggested for sh consumption. Fish oils are thought to have anti-inammatory effects, because of the inuence of the n-3 polyunsaturated fatty acids eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) on arachidonic acid metabolism (Sperling, 1991). Results of cross-sectional studies on this topic are inconclusive. Cross-sectional studies published in 1990 and 1994 suggested an inverse association between sh consumption and COPD related outcomes (Smit et al, 1999). However, all cross-sectional studies published in later years reported negative ndings (Cook et al, 1997; Fluge et al, 1998; Tabak et al, 1999, 2001a; Butland et al, 2000). In the only individuallevel longitudinal study, no association of sh consumption with 25 y incidence of chronic lung disease was observed (Miedema et al, 1993). To gain more insight into the longitudinal relationship between diet and COPD, we studied baseline antioxidant, fruit, vegetable and sh intake in relation to 20 y COPD mortality in 2917 middle-aged men. Data of the Finnish, Italian and Dutch cohorts of the Seven Countries Study were combined to provide a wide range in dietary intake. ytya and Mellila . Two cohorts of Italian men particiin Po pated in the study. One consisted of men living in the village of Crevalcore, which is located in the Po Valley. The participants of the other Italian cohort were inhabitants of Montegiorgio, a small country village in central Italy. The fth cohort was located in Zutphen, a small commercial town in the central-east of The Netherlands. In 1959 in Finland 1675 men were examined, about 98% of the total eligible population of men born from 1900 to 1919 in the two areas. After 10 y, 612 men in east Finland and 694 men in west Finland were re-examined, respectively 91 and 93% of the surviving men. Complete dietary information was obtained from 590 men in east Finland and 670 men in west Finland. In Zutphen a random sample of four out of each nine eligible men was drawn in 1960. Invited were 1088 men of which 872 men participated in both the medical examination and the dietary survey (80%). In 1970 625 men were re-examined, of whom 558 participated in individual dietary surveys. In both Italian cohorts 99% of the eligible men participated in 1960 (n 1712). In 1970, 570 participants in Montegiorgio and 753 men in Crevalcore were examined. From respectively 551 and 584 men complete dietary information was obtained. For 36 of the 2953 subjects with complete dietary data in the three countries, information on one or more potential confounders was missing. The nal study population therefore consisted of 2917 men.

Materials and methods

Study population From 1958 to 1964, 16 population samples of middle-aged men (40 59 y) from seven countries were enrolled and examined at baseline for the Seven Countries Study (Keys et al, 1967). During the second and third rounds (5 and 10 y of follow-up, respectively), dietary data were obtained in the Finnish, Italian and Dutch cohorts. Also 20 y follow-up mortality data were collected. In the present study two Finnish, two Italian and one Dutch cohort of the Seven Countries Study are involved. In Finland one cohort was located in the east, in Ilomantsi, a rural area close to the Russian border. The other Finnish cohort was situated in the south-western part of the country,

Data collection Individual dietary surveys were carried out around 1970 in all cohorts, except in Montegiorgio, where this survey took place in 1965. The dietary data gathered in 1965 in this Italian cohort were used as an approximation for dietary intakes in 1970. Individual dietary intake was estimated by the cross-check dietary history method (Burke, 1947). Although this method was adapted to the local situation in the specic country, the methodology was comparable for the three countries. The cross-check dietary history method provides information about the usual food consumption pattern during 6 12 months preceding the interview. The rst part of the method concerns questions about the foods used at breakfast, lunch, dinner and between meals to assess the usual food consumption pattern of a person during weekdays and weekends. For the second part of the survey a checklist with a number of foods was used. The frequencies and amounts consumed were recorded for the different food groups. The information about the food pattern was compared with the information from the checklist. In all cohorts the dietary interviews were carried out by experienced dieticians and nutritionists. The food intake was converted into energy and nutrient intake using the computerised versions of the local food tables for the three different countries (Den Hartog et al, 1965; Hautvast, 1975; Pekkarinen, 1981; AlbertiFidanza et al, 1988).
European Journal of Clinical Nutrition

Diet and COPD mortality IC Walda et al

640
In all cohorts information on age, height, weight and smoking status of the participants was collected in a standardised way (Keys et al, 1967). Body mass index (BMI) was calculated from measurements of body weight (kg) and height (m). Current smoking status was assessed using a standardised questionnaire. Possible answers were: current smoker, former smoker and never smoker. Current or former smokers were asked to report the amount of cigarettes they smoked or used to smoke daily. The number of years smoked was estimated by assuming that in all cohorts the men started smoking at the age of 16, as for the Dutch cohort. Pack years of smoking were calculated as the product of the number of years smoked and the number of packs of cigarettes smoked per day. It was assumed that a pack of cigarettes contained 20 cigarettes. The vital status of the participants was recorded during 20 y of follow-up. The underlying cause of death from those who died was established centrally by two investigators. They reviewed the information from clinical records, from family doctors, specialists and relatives and from other useful sources collected by local investigators. Usually the ofcial cause of death was not considered or only used as a preliminary indication. Primary mortality was coded according to the 8th revision of the International Classication of Diseases (ICD) of the World Health Organization (WHO, 1967). The ICD codes 491 493 refer to death from chronic bronchitis, lung emphysema and asthma, respectively. daily (Tabak et al, 2001b), and was therefore considered to be a potential confounder for the studied associations. For vitamin E, a fat-soluble antioxidant, adjustment for total fat intake did not alter the association with COPD mortality in a relevant way. Analyses stratied for country, which could have given insight into differences in potential protective effect of dietary factors between countries, were not possible due to the small number of COPD deaths within the countries. All analyses were performed using the SAS statistical package version 6.11 (Cary, USA). The term statistically signicant refers to P-values lower than 0.05 (two-sided tests).

Results
Table 1 presents general characteristics of the study population. Of the total research group only 21.3% were never smokers, 50.9% were current smokers and 27.8% were former smokers in 1970. Out of the total study population of 2917 men, 1712 men (58.7%) died during 20 y of followup, among which 73 deaths were from COPD. The overall COPD mortality rate was 171.1 per 100 000 person y with the highest rate in The Netherlands (224.2=100 000 person y) and the lowest rate in Italy (135.6 per 100 000 person y). Among never smokers the COPD mortality rate was 62 per 100 000 person y; in former smokers this rate reached 182 per 100 000 person y and 216 per 100 000 person y was the COPD mortality rate among current smokers. We observed a wide range in dietary intake (Table 2). Mean daily total energy intake ranged from 11.0 MJ in The Netherlands to 15.6 MJ in Finland. The level of alcohol consumption was much higher in Italy ( 8 drinks=day) compared to the other two countries ( < 1 drink=day). Mean daily vegetable intake varied strongly between the countries (68 181 g), but the range of mean fruit intake was smaller (153 175 g). After adjustment for country only, we observed an inverse trend of COPD mortality across tertiles (P-trend < 0.05) of baseline intake of fruit, vitamin C and vitamin E (Table 3).

Statistical analyses The Cox Proportional Hazards Model (SAS Procedure PHREG; SAS Institute Inc., 1991) was used to assess the relationship between the consumption of the selected dietary variables and 20 y COPD mortality, since this method takes into account the unequal lengths of time that each participant was observed. First, dietary variables were ranked into tertiles and associations with COPD mortality were studied after adjustment for age and cigarette smoking (pack years) at baseline and for country. Adjustment for country was carried out by calculating a pooled relative risk (RR), using the strata option of the PHREG procedure that allows baseline hazards to vary between countries. Only dietary factors showing a statistically signicant trend in COPD mortality across the tertiles (P-trend < 0.05) were used in further analyses. In the subsequent analyses the potential confounding effects of total energy intake, BMI and alcohol consumption at baseline were evaluated. In these analyses the selected dietary factors were modelled as continuous variables to increase power. Total energy intake was considered to evaluate the effect of a dietary factor per se. BMI was considered because, in the Dutch cohort, BMI in 1960 was found to be inversely associated with 25 y incidence of chronic lung disease (Miedema et al, 1993). Alcohol consumption was associated with 20 y COPD mortality in the study population with the lowest risk in those consuming 1 30 g of alcohol
European Journal of Clinical Nutrition

Table 1 General characteristics of the Finnish, Italian and Dutch cohorts of the Seven Countries Study (1970)
Finland (n 1227) Italy (n 1132) The Netherlands (n 558)
59.6 (5.4) 25.1 (2.7) 20.2 (15.2) 52.0 40.3 7.7 315 19

Age in y (mean (s.d.)) 59.1 (5.5) 59.3 (4.9) 24.7 (3.8) 26.0 (3.9) BMI in kg=m2 (mean (s.d.))a Cigarette smoking Pack years (mean (s.d.)) 21.2 (17.2) 13.7 (13.9) Current smoker (%) 50.4 50.8 Former smoker (%) 29.7 19.5 Never smoker (%) 19.8 29.7 Twenty years of follow-up Number of deaths 769 628 b Number of COPD deaths 31 23
a b

Body mass index. Chronic obstructive pulmonary disease.

Diet and COPD mortality IC Walda et al

641
Table 2 Daily intake of selected nutrients and foods in the Finnish, Italian and Dutch cohorts of the Seven Countries Study (1970)
Finland (n 1227), mean (s.d.) Nutrients Energy intake (MJ) Ethanol (g) Vitamin C (mg) b-carotene (mg) Vitamin E (mg) Foods Fruit (g) Vegetables (g) Fish (g) Italy (n 1132), mean (s.d.) The Netherlands (n 558), mean (s.d.)

Table 4 The relation of intake of fruit and vitamin E (continuous) with 20 y COPD mortality, after adjustment for various additional potential confounders, in 2917 middle-aged men from Finland, Italy and The Netherlands
Fruit intake (per 100 g), RR (95% CI)
Adjusted for countrya only 1 adjusted for age and smokingb c 2 adjusted for BMI d 2 adjusted for energy intake 2 adjusted for alcohol e consumption 5 adjusted for BMI and energy intake
a

Vitamin E intake (per 5 mg), RR (95% CI)

0.71 0.77 0.78 0.88 0.79 (0.51 0.99) (0.55 1.06) (0.56 1.07) (0.62 1.25) (0.57 1.08)

15.6 6.6 91.4 2.0 7.0

(4.7) (13.1) (35.4) (1.9) (2.5)

12.3 82.1 43.0 0.9 8.5

(3.3) (57.7) (24.4) (0.5) (3.1)

11.0 9.4 95.7 1.2 15.5

(2.2) (13.5) (44.5) (0.3) (6.4)

0.74 0.76 0.80 0.80 0.79

(0.60 0.92) (0.62 0.95) (0.64 0.99) (0.64 0.99) (0.64 0.97)

175 (185) 78 (60) 40 (47)

153 (167) 68 (53) 20 (21)

170 (129) 181 (58) 17 (19)

0.86 (0.69 1.07)

0.93 (0.65 1.33)

No association was observed for vegetables, sh and b-carotene. After adjustment for country, age and smoking the trend remained statistically signicant for fruit and vitamin E intake. The adjusted RR of the highest vs the lowest tertile of intake was 0.49 (95% CI 0.26 0.93) for fruit and 0.54 (95% CI 0.27 1.05) for vitamin E (Table 3). In subsequent analysis fruit and vitamin E intake were modelled continuously (Table 4). After adjustment for country, age and smoking, a 100 g increase in fruit intake was associated with a 24% lower COPD mortality risk (RR 0.76, 95% CI 0.60 0.92). Additional adjustment for BMI, energy intake or alcohol consumption each reduced the estimated RR slightly (to 0.79 0.80). Additional adjustment for all three factors reduced the RR to 0.86 (95% CI 0.69 1.07, P 0.12). In these analyses vitamin E intake was not statistically signicantly associated with COPD mortality after adjustment for country, age and smoking (RR per 5 mg 0.77, 95% CI 0.55 1.06). Of the other three potential confounders energy intake had the strongest effect, reducing the RR to 0.88 (95% CI 0.62 1.25). After additional adjustment for the three factors the RR (per 5 mg) was 0.93 (95% CI 0.65 1.33; Table 4).

By calculating a pooled risk ratio, allowing baseline hazards to vary between countries. b Pack years of cigarette smoking. c Body mass index (kg=m2). d Total energy intake (MJ). e For alcohol consumption two dichotomous variables were entered into the model: 1 vs 1 30 g=day and > 30 vs 1 30 g=day.

Discussion
After adjustment for age, smoking and country, we observed an inverse trend (P-trend < 0.05) of 20 y COPD mortality across tertiles of baseline habitual intake of fruit and vitamin E. No trend was observed for vegetables, sh, vitamin C and b-carotene. When modelled continuously, a 100 g increase in fruit intake was associated with a 24% lower COPD mortality risk (RR 0.76, 95% CI 0.60 0.92). For vitamin E intake (per 5 mg) the RR was 0.77 (95% CI 0.55 1.06), after adjustment for age, smoking and country. Additional adjustment for BMI, total energy intake and alcohol consumption reduced the RR to 0.86 (95% CI 0.69 1.07, P 0.12) for fruit and 0.93 (95% CI 0.65 1.33) for vitamin E.

Table 3 Individual associations of baseline intake of different dietary factors (tertiles) with 20 y COPD mortality in 2917 middle-aged men from Finland, Italy and The Netherlands
Adjusted for a country Dietary factors
Fish (g) Fruit (g) Vegetables (g) Vitamin C (mg) b-carotene (mg) Vitamin E (mg)
a b

Adjusted for a b country, age and smoking

P-trend 0.60 0.02 0.31 0.03 0.15 0.01

RR2,1
0.56 0.91 0.71 0.54 0.50 0.51

RR3,1 (95% CI)

0.89 0.45 0.73 0.44 0.65 0.46 (0.52 1.54) (0.24 0.85) (0.36 1.96) (0.22 0.88) (0.36 1.16) (0.23 0.89)

RR2,1
0.61 1.00 0.77 0.64 0.54 0.57

RR3,1 (95% CI)

1.02 0.49 0.91 0.57 0.75 0.54 (0.59 1.78) (0.35 0.94) (0.44 1.85) (0.28 1.15) (0.42 1.34) (0.27 1.05)

P-trend 0.93 0.04 0.67 0.13 0.34 0.04

Adjusted for country by calculating a pooled risk ratio, allowing baseline hazards to vary between countries. Pack years of cigarette smoking. RR2,1 relative risk of middle vs lowest tertile. RR3,1 relative risk of highest vs lowest tertile. Tertile limits sh, < 11, > 28; fruit, < 77, > 192; vegetables, < 52, > 107; vitamin C, < 51, > 87; b-carotene, < 0.9, > 1.4; vitamin E, < 6.7, > 9.5.

European Journal of Clinical Nutrition

Diet and COPD mortality IC Walda et al

642
Due to the relatively small number of men dying from COPD (n 73) in our study derived condence intervals were rather wide. Still, the results suggest a protective effect of a high intake of fruit against COPD. This is in agreement with the results of most cross-sectional studies (Smit et al, 1999; Tabak et al, 1999, 2001a; Butland et al, 2000) and with the results of two of the three longitudinal studies (Butland et al, 2000; Miedema et al, 1993; Carey et al, 1998). Miedema et al (1993) observed an inverse relation between baseline consumption of total and especially solid fruit ( apples, pears) and 25 y chronic lung disease incidence in the Dutch cohort of the Seven Countries Study. Butland et al (2000) reported a nonsignicant trend towards an inverse relation between baseline apple consumption and change in FEV1. As has been observed earlier (Tabak et al, 2001a) the effect of fruit intake appeared to be largely independent of that of alcohol consumption. In none of the longitudinal studies that investigated vitamin C intake in relation to indicators of COPD (including our own) a clear association was observed (Miedema et al, 1993; Butland et al, 2000). This suggests that other fruit components than the antioxidant vitamin C may have a protective effect on the development of emphysema and chronic bronchitis, eg avonoids, which have strong antioxidant capacities (Bors & Saran 1987; Hertog et al, 1993). In a recent Dutch study, total intake of three of the six avonoid subclasses (ie catechins, avonols and avones) was observed to be positively associated with the FEV1 and inversely associated with chronic cough and breathlessness. More detailed analyses suggested that the association with COPD was likely to be due to an effect of catechins, and not of avonols and avones (Tabak et al, 2001c). Results of cross-sectional studies on the relationship between vitamin E intake and COPD-related outcomes are inconsistent; both results suggesting a protective and results suggesting no effect being reported (Smit et al, 1999; Tabak et al, 1999; Butland et al, 2000). Butland et al (2000) observed no longitudinal relation between vitamin E intake and pulmonary function. In our study the association was substantially weakened by adjustment for energy intake. Energy intake is not a known risk factor for COPD. It is strongly associated with physical activity, but physical activity, even training, does not affect lung volumes (McArdle et al, 1991). Energy intake may stand for another, unknown, factor associated with diet and COPD mortality, but in this case the factor may be an intermediate factor and not a confounder. Therefore, it is unclear whether the results are more valid after or before adjustment for energy intake. We did not observe a relationship between sh consumption and COPD mortality. In the early 1990s a protective effect of sh intake was suggested by several cross-sectional studies (Smit et al, 1999). These ndings were supported by a plausible biological mechanism, ie an anti-inammatory effect of the n-3 fatty acids abundantly present in sh oils. In accordance with our ndings, all more recent epidemiological studies observed no association of sh intake with
European Journal of Clinical Nutrition

COPD related outcomes (Cook et al, 1997; Fluge et al, 1998; Tabak et al, 1999, 2001a; Butland et al, 2000). It is not likely that this is due to a higher level of sh intake in the earlier studies. In our study, for instance, the mean daily sh intake was 27.4 g, which is comparable to the level of intake in the study by Shahar et al (1994; 1.9 servings=week or 85 142 g sh=week). Of the study population 89% had ever smoked cigarettes. Although we adjusted for smoking habits as well as possible, residual confounding by smoking cannot be excluded. It is known that smokers tend to eat less fruit, vegetables and whole grain products and have lower intakes of vitamin C and b-carotene (Margetts & Jackson, 1993). It is, however, not necessarily so that residual confounding by smoking leads to overestimation of the studied dietary effects. If heavy smokers tend to under-report the number of cigarettes they usually smoke, which is not unlikely, this would even cause dilution of observed dietary effects. Possibly even more important is the fact that in the majority of epidemiological studies on diet and COPD, that reported on whether observed effects varied by smoking status (smoker vs non-smokers, current vs former vs never smokers or interaction with smoking status), such variation was not observed (Strachan et al, 1991; Schwartz & Weiss, 1994; Dow et al, 1996; Chuwers et al, 1997; Hu et al, 1998; Butland et al, 2000). It is therefore unlikely that the observed associations in our study are substantially biased by residual confounding by smoking. A number of other methodological issues needs to be discussed. The validity and reproducibility of the crosscheck dietary history method has been observed to be comparable to that of other dietary assessment methods (Block, 1982; Bloemberg et al, 1989). Yet, random misclassication of intake of the selected dietary factors may have occurred, which is likely to lead to underestimation of the relative risks. Furthermore, it may be questioned whether food consumption around 1970 is a good indicator for average food consumption during 20 y of follow-up. Changes in the food consumption patterns in the countries participating in the Seven Countries Study have been observed between 1960 and 1985 (Kromhout et al, 1989). However, the relative position of the countries in the distribution of different foods was maintained (Kromhout et al, 1989). In survivors of our study population (from Finland, Italy and The Netherlands) fruit intake was observed to increase during 20 y of follow-up (Huijbregts et al, 1995). Since this increase was highest in the reference group, the lowest fruit tertile, this may have caused underestimation of the effect of fruit. Finally, although in our study we used the term COPD (dened as chronic bronchitis and emphysema), asthma does contribute to the mortality rates. However, chronic bronchitis or emphysema was often noted as the secondary cause of death in cases where asthma was reported as the underlying cause of death. Also considering the fact that the number of asthma cases was small (n 4), we decided against excluding them from the analyses.

Diet and COPD mortality IC Walda et al

In summary, our results suggest a protective effect of fruit and possibly vitamin E against dying from COPD. No effect was observed for intake of vitamin C, b-carotene, vegetables and sh. Acknowledgements The authors thank the many people that were involved in this longitudinal study. This includes the men who took part in the surveys and the organisers of the eldwork in the three countries, Professor A Menotti and Dr S Giampaoli in Italy, Professor A Nissinen in Finland and Professor D Kromhout in The Netherlands.

References
Alberti-Fidanza A, Seccareccia F, Torsello S & Fidanza F (1998): Diet of two rural population groups of middle aged men in Italy. Int. J. Vitam. Nutr. Res. 58, 442 451. Block G (1982): A review of validations of dietary assessment methods. Am. J. Epidemiol. 115, 492 505. Bloemberg BPH, Kromhout D, Obermann-de Boer GL & KampenDonker M van (1989): The reproducibility of dietary intake data assessed with the cross-check dietary history method. Am. J. Epidemiol. 130, 1047 1056. Bors W & Saran M (1987): Radical scavenging by avonoid antioxidants. Free Radic. Res. Commun. 2, 289 294. Burke BS (1947): The dietary history as a tool in research. J. Am. Diet. Assoc. 23, 1041 1046. Butland BK, Fehily AM & Elwood PC (2000): Diet, lung function, and lung function decline in a cohort of 2512 middle-aged men. Thorax 55, 102 108. Carey IM, Strachan DP & Cook DG (1998): Effect of changes in fresh fruit consumption on ventilatory function in healthy British adults. Am. J. Resp. Crit. Care. Med. 158, 728 733. Chow CK (1993): Cigarette smoking and oxidative damage in the lung. Ann. N.Y. Acad. Sci. 686, 289 298. Chuwers P, Barnhart S, Blanc P, Brodkin CA, Cullen M, Kelly T, Keogh J, Omenn G, Wiliams J & Balmes JR (1997): The protective effect of b-carotene and retinol on ventilatory function in an asbestosexposed cohort. Am. J. Respir. Crit. Care Med. 155, 1066 1071. Cook DG, Carey IM, Whincup PH, Papacosta O, Chirico S, Bruckdorfer KR & Walker M (1997): Effect of fresh fruit consumption on lung function and wheeze in children. Thorax 52, 628 633. Den Hartog C, Schaik Th FS, Dalderup LM, Drion EF & Mulder T (1965): The diet of volunteers participating in a long term epidemiological eld survey on coronary heart disease at Zutphen, The Netherlands. Voeding 26, 184 208. Dow L, Tracey M, Villar A, Coggon D, Margetts BM, Campbell MJ & Holgate ST (1996): Does dietary intake of vitamins C and E inuence lung function in older people? Am. J. Resp. Crit. Care Med. 154, 1401 1404. Fluge O, Omenaas E, Eide GE & Gulsvik A (1998): Fish consumption and respiratory symptoms among young adults in a Norwegian community. Eur. Respir. J. 12, 336 340. Hautvast JGAJ (1975): Commissie uniforme codering voedingsenquetes: ontwikkeling van een systeem om gegevens van voedingtes met behulp van een computer te verwerken. Voeding 36, senque 356 361 (in Dutch). Hertog MGL, Hollman PCH, Katan MB & Kromhout D (1993): Estimation of daily intake of potentially anticarcinogenic avonoids and their determinants in adults in The Netherlands. Nutr. Cancer 20, 21 29.

Hu G, Zhang X, Chen J, Peto R, Campbell TC & Cassano PA (1998): Dietary vitamin C intake and lung function in rural China. Am. J. Epidemiol. 148, 594 599. sa nen L, Alberti-Fidanza A, MutaHuijbregts PPCW, Feskens EJM, Ra nen M, Fidanza F & Kromhout D (1995): Dietary intake in ve aging cohorts of men in Finland, Italy and The Netherlands. Eur. J. Clin. Nutr. 49, 852 860. Keys A, Aravanis C & Blackburn HW (1967): Epidemiological studies related to coronary heart disease: characteristics of men aged 40 59 in seven countries. Acta Med. Scand. 460, S191 208. Kromhout D, Keys A, Aravanis C, Buzina R, Fidanza F, Giampaoli S, Jansen A, Menotti A, Nedeljkovic S, Pekkarinen M, Simic BS & Toshima H (1989): Food consumption patterns in the 1960s in seven countries. Am. J. Clin. Nutr. 49, 889 894. Margetts BM & Jackson AA (1993): Interaction between peoples diet and their smoking habits: the dietary and nutritional survey of British adults. Br. Med. J. 307, 1381 1384. McArdle WD, Katch I & Katch VL (1991): Exercise Physiology: Energy, Nutrition, and Human Performance, 3rd edn. Philadelphia, PA: Lea and Febiger. Miedema I, Feskens EJM, Heederik D & Kromhout D (1993): Dietary determinants of long-term incidence of chronic nonspecic lung disease: the Zutphen study. Am. J. Epidemiol. 138, 37 45. Pekkarinen M (1981): Dietary surveys in connection with coronary heart disease studies in Finland. In New Trends in Nutrition, Lipid Research, and Cardiovascular Diseases, ed. NG Bazan, R Paoletti & JM Iacono, pp 243 61. New York: Alan R Liss. SAS Institute Inc. (1991): SAS technical report P-217; The PHREG procedure, version 6. Cary, NC: SAS Institute Inc. Schwartz J & Weiss ST (1994): Relationship between dietary vitamin C intake and pulmonary function in the First National Health and Nutrition Examination Survey (NHANES 1). Am. J. Clin. Nutr. 59, 110 114. Shahar E, Folsom AR, Melnick SL, Tockman MS, Comstock GW, Gennaro V, Higgins MW, Sorlie PD, Ko W-J & Szklo M (1994): Dietary n-3 polyunsaturated fatty acids and smoking-related chronic obstructive pulmonary disease. New Engl. J. Med. 331, 228 233. Smit HA, Grievink L & Tabak C (1999): Dietary inuences on chronic obstructive lung disease and asthma: a review of the epidemiological evidence. Proc. Nutr. Soc. 58, 309 319. Sperling RI (1991): Effects of dietary sh oil on leucocyte leukotriene and PAF generation and on neutrophil chemotaxis. World Rev. Nutr. Diet. 66, 391 400. Strachan DP, Cox BD, Erzinclioglu SW, Walters DE & Wichelow MJ (1991): Ventilatory function and winter fresh fruit consumption in a random sample of British adults. Thorax 46, 624 629. sa nen L, Fidanza F, Menotti A, Nissinen A, Tabak C, Smit HA, Ra Feskens EJM, Heederik D & Kromhout D (1999): Dietary factors and pulmonary function: a cross sectional study in middle aged men from three European countries. Thorax 54, 1021 1026. MC & Kromhout D (2001a): Diet Tabak C, Smit HA, Heederik D, Ocke and chronic obstructive pulmonary disease: independent benecial effects of fruits, whole grains, and alcohol (the MORGEN study). Clin. Exp. All. 31, 747 755. sa nen L, Fidanza F, Menotti A, Nissinen A, Tabak C, Smit HA, Ra Feskens EJM, Heederik D & Kromhout D (2001b): Alcohol consumption in relation to 20-year COPD mortality and pulmonary function in middle-aged men from three European countries. Epidemiology 12, 239 245. Tabak C, Arts ICW, Smit HA, Heederik D & Kromhout D (2001c): COPD and intake of catechins, avonols and avones: the MORGEN-study. Am. J. Respir. Crit. Care Med. 164, 61 64. WHO (1967): International Classication of Diseases: Manual of the International Statistical Classication of Diseases, Injuries and Causes of Death, Vol 1. Geneva: World Health Organization.

643

European Journal of Clinical Nutrition