Catechin Intake and Associated Dietary and Lifestyle Factors in a Representative Sample of Dutch Men and Women

Catechin intake and associated dietary and lifestyle factors in a representative sample of Dutch men and women

ICW Arts1,2*, PCH Hollman2, EJM Feskens1, HB Bueno de Mesquita1 and D Kromhout3

1National Institute of Public Health and the Environment (RIVM), Department of Chronic Diseases Epidemiology, Bilthoven, The Netherlands; 2State Institute for Quality Control of Agricultural Products (RIKILT), Wageningen, The Netherlands; and 3National Institute of Public Health and the Environment (RIVM), Division of Public Health Research, Bilthoven, The Netherlands

Objective: To study the intake of catechins in the Dutch population and to assess the relation between catechin intake and other dietary factors. Catechins, dietary components that belong to the ¯avonoid family, potentially protect against chronic diseases such as cancer and cardiovascular diseases. Catechins are the major components of tea, but they are present in many other plant foods as well. Design: Data were used from a nationwide dietary survey carried out in 1998 among a representative sample of 6200 Dutch men and women aged 1 ± 97 y. Dietary data were collected using a 2 day dietary record method. Results: The average daily catechin intake was 50 mg (s.d. 56 mg=day). Catechin intake increased with age, and the intake was higher in women (60 mg=day) than in men (40 mg=day). Tea was the main catechin source in all age groups, whereas chocolate was second in children, and apples and pears were second in adults and elderly. Catechin intake was lower in smokers than in non-smokers, and increased with socio-economic status. A high intake was associated with a high intake of ®ber (r 0.20), vitamin C (r 0.17) and beta-carotene (r 0.10). Conclusions: Catechins are quantitatively important bioactive components of the daily diet, which should be taken into account when studying the relation between diet and chronic diseases. Catechin intake is only moderately associated with the intake of other nutrients, but much stronger with certain health behaviours such as smoking. Sponsorship: Commission of the European Communities Agriculture and Fisheries (FAIR) speci®c RTD Programme CT95 0653. Descriptors: catechin; tea; ¯avonoid; antioxidants; diet surveys; dietary intake European Journal of Clinical Nutrition (2001) 55, 76±81

Introduction species, induction of enzyme systems involved in detox- i®cation, and protection against DNA damaging free radi- Catechins, also referred to as ¯avanols, are one of the six cals through their antioxidant activity (Rice-Evans et al, subclasses of ¯avonoids. Flavonoids are non-nutritive sec- 1996; Middleton & Kandaswami, 1994). Catechins may ondary plant metabolites, which are common components help prevent low-density lipoprotein (LDL) from oxidative of the human diet. Currently more than 4000 different damage through their free radical quenching and metal ¯avonoids have been identi®ed (Harborne, 1994). In vitro chelating abilities (Wiseman et al, 1997; Rice-Evans et al, and in vivo animal experiments have shown that catechins 1996), or play a role in the in¯ammatory and thrombotic have antimutagenic and anticarcinogenic properties, and processes involved in atherosclerosis (Middleton, 1998; that they may play a role in the prevention of cardiovas- Wollny et al, 1999). cular diseases. Proposed mechanisms for these potential Catechins are the major components of tea; they con- human health promoting properties include inhibition of the stitute about 30% of the dry weight of green tea, and 9% of metabolic activation of procarcinogens to DNA-reactive the dry weight of black tea (Harbowy & Balentine, 1997). Epidemiological studies on health effects of tea have suggested that a high tea intake may protect against certain *Correspondence: ICWArts, RIVM, CZE, PO Box 1, 3720BA Bilthoven, The Netherlands cancers and coronary heart disease, but results are incon- E-mail: [email protected] sistent (Bushman, 1998; Kohlmeier et al, 1997; Tijburg et al, Guarantor: ICW Arts. 1997; Blot et al, 1996). On the other hand, there is ®rm Contributors: PH, EF and DK formulated the hypothesis, IA and PH evidence from experimental studies in laboratory animals collected part of the data, and all ®ve authors were involved in the analysis that tea may prevent chronic diseases (Katiyar & Mukhtar, of the data and the writing of the report. Received 11 April 2000; revised 11 September 2000; accepted 1996; Dreosti et al, 1997). One possible explanation for this 14 September 2000 apparent inconsistency is that tea is not the only catechin Catechin intake and associated factors ICW Arts et al 77 containing food in the human diet. In particular in countries Catechin intake calculation were tea intake is low, other foods may be more important The catechin intake for each individual was calculated by sources of catechins than tea. multiplying the consumption of each food by its catechin Recently, we reported data on catechin contents of a content. Catechin contents were estimated either by direct comprehensive set of foods and beverages (Arts et al, chemical analysis of the food concerned (37%), calculated 2000a, b). These data enable us to estimate the intake of using standard recipes if one or more ingredients contained these compounds in a population-based sample of men and catechins (57%), or derived from similar foods (6%). Data women in The Netherlands, and to study dietary and life- on catechin contents of several varieties of apples and pears style factors that are associated with a high intake of were combined into single values according to auction catechins. supply data. Similarly, data on catechin contents of tea blends, and white and red wines were combined into single values according to supply data provided by, respectively, the Coffee and Tea Information Bureau, and the Dutch Methods Commodity Board of Wine. Standard recipes were used to calculate the catechin contents of foods that contained Food consumption survey one or more chemically analyzed ingredients, except for The Dutch National Food Consumption Survey 1998 chocolate-containing foods, where data were derived from was carried out among a sample of households that was the Conversion Model Primary Agricultural Products (Van representative of the Dutch population. A household was Dooren-Flipsen et al, 1996). Because seasonal variation de®ned as one or more persons living together in one was relatively low (Arts et al, 2000a), year average cate- house, eating together a home-prepared hot meal for at chin values were used. Since catechins are predominantly least 4 days a week. Based on this de®nition, institutiona- present in foods that are usually consumed raw, eg fruit, lized persons were excluded. Also, subjects who had not catechin loss due to home preparation is not an important mastered the Dutch language suf®ciently, and children issue. younger than 1 y of age were excluded. A total of 6250 persons (2885 men and 3365 women) aged 1 ± 97 y coming Statistical analysis from 2564 households participated in the study. The Statistical analyses were performed using the SAS statis- response rate was 68.5%. Trained dieticians collected diet- tical package (SAS, release 6.12, SAS Institute, North ary data between April 1997 and March 1998 using a 2 day Carolina, USA). Pregnant women (n 50) were excluded dietary record method. No data were collected on holidays; from the analyses, because they tended to have a deviant otherwise, record-days were distributed equally over the 7 dietary pattern. Wilcoxon rank sum tests were used to days of the week and over the year. Other measurements compare mean catechin intakes between groups. Partial included age, height, weight, socio-economic status and rank-order correlation coef®cients were calculated between life-style variables such as smoking. As indicators of socio- total catechin intake and intake of alcohol, saturated and economic status, (former) occupation and attained educa- polyunsaturated fatty acids, ®ber, vitamin C, vitamin E, and tional level of the head of the household were used beta-carotene, after adjusting for total energy intake. Mean (Hulshof & Van Staveren, 1991; Voedingscentrum, 1998). intakes strati®ed by smoking status, and socio-economic status were standardized for age and sex using analysis of covariance. Children (< 19 y of age) were excluded from Food analysis the smoking status analysis. Six major catechins (( )-catechin, (7)-epicatechin, ( )- gallocatechin (GC), (7)-epigallocatechin (EGC), (7)- epicatechin gallate (ECg), and (7)-epigallocatechin Results gallate (EGCg)) were determined in a comprehensive set of plant foods by reversed-phase high performance liquid The mean intake of catechins in the total population was chromatography with online ultraviolet and ¯uorescence 50 mg=day (s.d. 56 mg=day), and ranged from 0 (284 detection (Arts & Hollman, 1998; Arts et al, 2000a, b). subjects, 4.6%) to 958 mg=day on the two particular days Analyzed foods included 24 types of fruits, 27 types of on which subjects were surveyed. The distribution of vegetables and legumes, some staple foods, and a number catechin intake in the adult population (19 y and older), of processed foods such as chocolate, eight types of black was strongly skewed towards higher values (Figure 1). tea, 18 types of red and white wine, fruit juices, ice-tea, Only 123 subjects (2% of the total population) had an beer, chocolate milk and coffee. All perishable foods were average intake on the two survey days that was higher than purchased at three outlets: a nationwide supermarket 200 mg=day. chain, an open-air street market, and a grocery. To take Of the individual catechins, ECg contributed most to the into account seasonal and year-to-year variability, fruits total catechin intake (33%), followed by EGCg (24%) and and vegetables were purchased in August and December (7)-epicatechin (23%). GC (5%), EGC (8%), and ( )- 1997 and in April and August 1998, if available in that catechin (8%) were minor contributors. Women had a period. signi®cantly higher intake of all catechins than men: the

European Journal of Clinical Nutrition Catechin intake and associated factors ICW Arts et al 78

Figure 1 Distribution of the total catechin intake of 4661 adults (19 y and older), the Dutch National Food Consumption Survey 1998.

total intake was 60 mg=day in women vs 40 mg=day in Table 1 Mean intakes of total catechins, ( )-catechin, ( 7 )- men (P-value 0.0001), (Table 1). This gender difference epicatechin, ( )-gallocatechin (GC), ( 7 )-epigallocatechin (EGC), ( 7 )-epicatechin gallate (ECg), and ( 7 )-epigallocatechin gallate persisted when we classi®ed subjects by age group and (EGCg) among 6200 men and women, the Dutch National Food standardized catechin intake on total energy intake (Figure Consumption Survey 1998 2). After the age of 16, women had a signi®cantly higher catechin intake per MJ energy intake than men. Further- Mean catechin intake (mg daily) Æ s.d. more, total catechin intake increased with age, even after Men Women subjects had attained adulthood (Figure 2). The most important source of catechins in children, n 2885 3315 adults and elderly people was tea (Table 2). Tea consump- Total catechins 39.5Æ 46.4* 59.8Æ 62.5 ( )-Catechin 3.5Æ 3.6* 4.4Æ 4.1 tion increased with age: the number of consumers, the (7)-Epicatechin 10.2Æ 9.0* 12.8Æ 10.1 mean intake, and the contribution to the total catechin GC 1.8Æ 2.6* 3.0Æ 3.6 intake all showed a steady increase, up to a contribution EGC 2.8Æ 4.1* 4.6Æ 5.6 of 87% in elderly subjects. Chocolate was an important ECg 12.1Æ 17.8* 20.1Æ 24.4 source of catechins in children only: among the 1 ± 18 y EGCg 9.0Æ 13.2* 14.9Æ 18.1 olds, 83% consumed chocolate on at least one of the two *P 0.0001 for difference between men and women. survey days (20% of total catechin intake). Chocolate and

Figure 2 Mean total catechin intake per MJoule total energy intake, by age group and sex among 6200 subjects, the Dutch National Food Consumption Survey 1998. Error bars indicate standard deviations.* P < 0.05 for difference between men and women.

European Journal of Clinical Nutrition Catechin intake and associated factors ICW Arts et al 79 Table 2 Contribution of various food groups to the mean catechin intake in the total population by age group

Children (1 ± 18 y) (n 1539) Adults (19 ± 64 y) (n 3954) Elderly ( > 65 y) (n 707)

Percentage Mean catechin Percentage Percentage Mean catechin Percentage Percentage Mean catechin Percentage users intake (mg=day) contribution users intake (mg=day) contribution users intake (mg=day) contribution

Tea 47 16.6Æ 25.6 65.2 67 46.3Æ 58.8 83.3 84 65.6Æ 62.3 87.3 Chocolate 83 5.1Æ 5.8 20.1 65 3.1Æ 4.8 5.7 61 2.0Æ 3.4 2.6 Apples and pears 50 3.1Æ 4.5 12.0 44 3.3Æ 5.3 5.9 58 4.0Æ 5.0 5.3 Other fruits 15 0.4Æ 1.5 1.5 20 0.8Æ 3.0 1.4 29 1.6Æ 4.6 2.1 Wine 2 0.0Æ 0.3 0.1 20 1.6Æ 5.0 2.8 20 1.1Æ 3.7 1.5 Legumes 3 0.1Æ 0.6 0.2 4 0.2Æ 1.4 0.3 5 0.5Æ 3.6 0.7 Other foods 37 0.2Æ 0.6 0.9 45 0.4Æ 0.9 0.7 56 0.4Æ 1.0 0.6 All foods 97 25.5Æ 27.2 100 95 55.6Æ 60.2 100 96 75.1Æ 63.4 100

Table 3 Partial rank-order correlation coef®cients between total catechin After adjusting for total energy intake, we observed intake and other dietary factors strati®ed by age group and adjusted for weak positive associations between intake of catechins total energy intake and intake of ®ber, vitamin C and beta-carotene among Children (1 ± 18 y) Adults (19 ± 64 y) Elderly ( > 65 y) children, adults and the elderly (Table 3). The strongest (n 1539) (n 3954) (n 707) associations were observed for ®ber (r 0.21 for children). Catechin intake was not associated with intake of vitamin Alcohol 7 0.01 7 0.08* 7 0.15* Saturated fatty 0.05 7 0.03 0.02 E, saturated fatty acids and polyunsaturated fatty acids, and acids was inversely associated with intake of alcohol in the Polyunsaturated 7 0.02 7 0.02 0.01 elderly. The mean catechin intake standardized for age fatty acids and sex was signi®cantly lower among smokers compared Fiber 0.21* 0.20* 0.13* with non-smokers, and increased with socio-economic Vitamin C 0.07* 0.17* 0.11* Vitamin E 0.01 0.05* 0.07 status (Table 4). Beta-carotene 0.06* 0.10* 0.11*

*P < 0.05 Discussion

In a representative sample of 6200 Dutch men and women aged 1 ± 97 y, the intake of six major catechins was on average 50 mg=day. KuÈhnau (1976) ®rst estimated the Table 4 Mean total catechin intake according to smoking status (adults only) and socio-economic status, standardized for age and sex dietary intake of catechins using data on the average American diet which were based on food disappearance n Mean catechin intake (mg daily) values (Food Consumption Statistics 1955 ± 1971, OECD, Paris, 1973). The estimated catechin intake in the US Smoking status Yes 1431 43.7* was 220 mg=day expressed as quercitrin (molecular No 3226 65.2 weight 464), which equals 140 mg catechins per day. This is roughly three times our estimate for the Dutch Socio-economic status A (high) 794 59.8{ population. Although such an intake is possible (in our B1 1630 54.0 study 7% of the population had an intake of at least B2 1298 53.0 140 mg=day), KuÈhnau most likely overestimated the aver- C 2230 44.5 age catechin intake in the US, because he used food D (low) 242 33.5 disappearance data, which are known to overestimate the *P < 0.05 for difference between smokers and non-smokers. true food intake. In addition he only had available catechin {P-value for a linear trend: 0.0001. data obtained with methods now considered obsolete. The most advanced technique used at that time was thin-layer chromatography with spectrophotometric measurement, apples=pears were equally important sources of catechins in which is likely to overestimate the catechin content of a adults (6%), while apples and pears were more important in food. Most likely the American daily catechin intake is elderly people. Wine, other fruits such as cherries, straw- lower than the intake in The Netherlands because the berries and peaches, and legumes such as kidney beans consumption of tea, a major catechin source, is relatively contributed only marginally to the total catechin intake of low in the US (Zheng et al, 1996; Rimm et al, 1996). More this population, and were generally consumed by only a recently a paper was published on the catechin intake in the small part of the population. Within consumers, however, Danish population (Dragsted et al, 1997). An estimated they may be important sources. average daily intake of 20 ± 50 mg was reported using data

European Journal of Clinical Nutrition Catechin intake and associated factors ICW Arts et al 80 from the Danish Household Consumption Survey and more often non-smokers. There was a clear increase in literature data on catechin contents of foods. The calculated catechin intake with increasing socio-economic status. range was wide because of lack of detail in the survey data When studying the health effects of catechins using epide- and wide ranges of reported catechin contents of foods in miological data, these dietary and lifestyle factors should the literature. be taken into account. In our study a 2 day dietary record method was used to measure food intake. This is a reliable method to estimate Acknowledgements ÐThe authors would like to thank Betty van de Putte population intakes. At the individual level, however, the for excellent technical assistance, and Jacob van Klaveren and Margo van two particular days included in the study probably do not Dooren-Flipsen at RIKILT for use of data from the Conversion Model represent the habitual dietary pattern. We determined Primary Agricultural Products. catechin contents of more than 100 plant foods and beverages commonly consumed in The Netherlands using References state-of-the-art high performance liquid chromatography methods, taking into account seasonal and year-to-year Arts ICW & Hollman PCH (1998): Optimization of a quantitative method variation (Arts et al, 2000a, b). This comprehensive data- for the determination of catechins in fruits and legumes J. Agric. Food base represents catechin contents of foods as consumed. It Chem. 46, 5156 ± 5162. is unlikely that we have missed important catechin sources Arts ICW, Van de Putte B & Hollman PCH (2000a): Catechin contents of in the Dutch diet because of the extensive sampling scheme foods commonly consumed in The Netherlands. 1. Fruits, vegetables, staple foods, and processed foods. J. Agric. Food Chem. 48, 1746 ± (Arts et al, 2000a,b). 1751. The major sources of catechins were tea, chocolate and Arts ICW, Van de Putte B & Hollman PCH (2000b): Catechin contents of apples=pears. Absolute consumption levels of tea were foods commonly consumed in The Netherlands. 2. Tea, wine, fruit much higher than those of chocolate, which resulted in a juices, and chocolate milk. J. Agric. Food Chem. 48, 1752 ± 1757. Blot WJ, Chow WH & McLaughlin JK (1996): Tea and cancer: a review of lower contribution of chocolate to the total catechin intake, the epidemiological evidence. Eur. J. Cancer Prev. 5, 425 ± 438. despite the fact that catechin concentrations are higher in Bushman JL (1998): Green tea and cancer in humans: a review of the chocolate than in tea (Arts et al, 2000a, b). The major literature. Nutr. Cancer 31, 151 ± 159. catechins in our study were ECg, EGCg and (7)-epicate- Dragsted LO, Strube M & Leth T (1997): Dietary levels of plant phenols chin. This partly justi®es the relatively large research effort and other non-nutritive components: could they prevent cancer? Eur. J. Cancer Prev. 6, 522 ± 528. directed at EGCg in in vitro and animal experimental Dreosti IE, Wargovich MJ & Yang CS (1997): Inhibition of carcinogenesis studies. However, it also warrants more attention for ECg by tea: The evidence from experimental studies. Crit. Rev. Food Sci. and (7)-epicatechin, and comparative studies would be Nutr. 37, 761 ± 770. extremely useful. Harborne JB, ed. (1994): The Flavonoids: Advances in Research Since Catechin intake was higher among adult women than 1986. London: Chapman & Hall. Harbowy ME & Balentine DA (1997): Tea chemistry. Crit. Rev. Plant Sci. men. There were more females that consumed tea (75% vs 16, 415 ± 480. 58% of adult males), other fruits than apples=pears (23% vs Hulshof KFAM & Van Staveren WA (1991): The Dutch National Food 16%), and wine (24% vs 15%). Additionally, adult female Consumption Survey: design, methods and ®rst results. Food Policy 16, tea consumers drank on average more tea than adult male 257 ± 260. Katiyar SK & Mukhtar H (1996): Tea in chemoprevention of cancer: tea consumers. In elderly, gender differences were similar, epidemiologic and experimental studies. [review.] Int. J. Oncol. 8, but less pronounced. Catechin intake increased with age 221 ± 238. among both sexes. Again, the observed differences were Kohlmeier L, Weterings KGC, Steck S & Kok FJ (1997): Tea and cancer largely due to variations in tea intake. On the basis of these prevention: an evaluation of the epidemiologic literature. Nutr. Cancer cross-sectional data, it is not possible to conclude that, with 27, 1 ± 13. KuÈhnau J (1976): The ¯avonoids: a class of semi-essential food increasing age, people increase their catechin intake: we components: their role in human nutrition. World Rev. Nutr. Diet. 24, cannot distinguish between age and cohort effects. The 117 ± 191. increasing catechin intake with age may be due to an Middleton E (1998): Effect of plant ¯avonoids on immune and in¯amma- increase in tea consumption with aging, but it may also tory cell function. Adv. Exp. Med. Biol. 439, 175 ± 182. be due to decreasing popularity of tea among youths Middleton E & Kandaswami C (1994): The impact of plant ¯avonoids on mammalian biology: implications for immunity, in¯ammation and nowadays. If the latter would be true, the adults of today cancer. In The Flavonoids: Advances in Research Since 1986, ed. JB would have a lower catechin intake by the time they Harborne, pp 619 ± 652. London: Chapman & Hall. become elderly than the current elderly. Rice-Evans CA, Miller NJ & Paganga G (1996): Structure-antioxidant In epidemiological studies, it is important to consider activity relationships of ¯avonoids and phenolic acids. Free Radical lifestyles and health behaviors associated with the exposure Biol. Med. 20, 933 ± 956. Rimm EB, Katan MB, Ascherio A, Stampfer MJ & Willett WC (1996): of interest to take into account potential confounders or Relation between intake of ¯avonoids and risk for coronary heart effect modi®ers. Several studies have reported that cate- disease in male health professionals. Ann. Intern. Med. 125, 384 ± 389. chin-rich beverages such as wine and tea are associated Tijburg LBM, Mattern T, Folts JD, Weisgerber UM & Katan MB (1997): with a healthy dietary pattern (Tjonneland et al, 1999; Tea ¯avonoids and cardiovascular diseases: a review. Crit. Rev. Food Sci. Nutr. 37, 771 ± 785. Tsubono et al, 1997). In our study, subjects with a high Tjonneland A, Gronbaek M, Stripp C & Overvad K (1999): Wine intake catechin intake tended to have a high intake of vitamin C, and diet in a random sample of 48763 Danish men and women. Am. J. beta-carotene and ®ber, consumed less alcohol, and were Clin. Nutr. 69, 49 ± 54.

European Journal of Clinical Nutrition Catechin intake and associated factors ICW Arts et al 81 Tsubono Y, Takahashi T, Iwase Y, Iitoi Y, Akabane M & Tsugane S Wiseman SA, Balentine DA & Frei B (1997): Antioxidants in tea. Crit. (1997): Dietary differences with green tea intake among middle-aged Rev. Food Sci. Nutr. 37, 705 ± 718. Japanese men and women. Prev. Med. 26, 704 ± 710. Wollny T, Aiello L, Di Tommaso D, Bellavia V, Rotilio D, Donati MB, De Van Dooren-Flipsen MMH, Boeijen I, Van Klaveren JD & Van Donkers- Gaetano G & lacoviello L (1999): Modulation of haemostatic function goed G (1996): Conversie van consumeerbare voedingsmiddelen naar and prevention of experimental thrombosis by red wine in rats: a role primaire agrarische producten. Berekening inname residuen en contam- for increased nitric oxide production. Br. J. Pharmac. 127, 747 ± 755. inanten. Voeding 57,6±9. Zheng W, Doyle TJ, Kushi LH, Sellers TA, Hong CP & Folsom AR Voedingscentrum (1998): Zo eet Nederland; Resultaten van de Voedselcon- (1996): Tea consumption and cancer incidence in a prospective cohort sumptiepeiling 1997 ± 1998. [in Dutch.] Den Haag: Voedingscentrum. study of postmenopausal women. Am. J. Epidemiol. 144, 175 ± 182.

European Journal of Clinical Nutrition