Introduction

It has been estimated for Germany that dietary modifications might lead to a decrease in cancer mortality by 4–11% and an even higher impact of diet has been estimated for cancer incidence [3]. However, the results of many investigations suggest that dietary patterns rather than individual nutrients might account for such effects, taking into account the possibility of diverse interactions between food components [19, 24]. Thus, whether the use of dietary supplements such as vitamins, minerals, fibers, and protein, but also food extracts like yeast or garlic products could have an effect on the risk of chronic diseases is still a matter of debate [2, 6, 17, 25, 31, 35]. In contrast, for subgroups of the population, a high intake of specific supplements might even be harmful [1, 30, 32, 40]. However, the use of dietary supplements has increased for years and 383 million Euros were spent on vitamin and mineral supplements in Germany in 2004 [15].

It is the aim of this investigation to describe the consumption of different dietary supplements in the EPIC-Heidelberg cohort, and to further characterize the consumers by lifestyle and dietary behaviours.

Subjects and methods

Subjects

During June 1994 and October 1998, 13,615 women aged 35–65 and 11,929 men aged 40–65 were recruited in the EPIC-Heidelberg (Germany) cohort [11, 36]. The target population was the general population from Heidelberg and surrounding communities. The participation rate was 38.3%. A detailed comparison between the cohort and the survey population is provided elsewhere [10]. The study was approved by the local ethical committee and all participants gave their informed consent prior to the study participation. All study participants were included in the present evaluation.

Data collection

All participants completed a validated semiquantative food frequency questionnaire (FFQ) with 148 food items, which covered dietary habits during the past year [9, 12–14]. Recorded food items were divided into 22 main food groups, and nutrient intake was calculated by means of the German food composition data base BLS version II.1 (BgVV, Berlin, Germany). Questions on regular dietary supplement use, which was defined as ‘for at least 4 weeks during the past year’, were included in the FFQ. The participants were asked for their use of vitamins, minerals, protein products, yeast products, garlic, and fibre supplements.

By means of a lifestyle questionnaire and a computer-assisted face-to-face interview . smoking habits, physical activity, education, employment, health status as well as family and reproductive life was assessed. Weight, height, and waist circumference were measured following standardised procedures [21]. Participation in cancer screening programs was assessed during the first follow-up between 1998 and 2000.

Statistical analysis

Associations with supplement use were examined for education (≤9 years of schooling and no vocational training; 10 years of schooling and/or vocational or technical training; 13 years of schooling and/or graduated from university [38]), physical activity (average number of hours per week spent on leisure time activities in the last year), occupational activity (mainly sitting; mainly standing or walking; physically exhausting; heavy physical work), smoking status (non-smokers, ex-smokers, current smokers), marital status (living with their spouse/partner versus living without), body mass index (BMI) (<25 kg/m2, ≥25–<30 kg/m2, ≥30 kg/m2), and participation in cancer screening measures (Hemoccult® test, palpation of the rectum, digital rectal examination (DRE), PAP smear test, mammography, palpation of the breast).

All analyses were stratified by sex. Associations between socio-demographic, lifestyle, and dietary factors and supplement use were examined using unconditional logistic regression models. Supplement use was modelled as a dichotomous variable (user/non-user of vitamin and/or mineral supplements). Dietary variables were analysed by quintiles of intake separately for each gender. Quintiles were defined based on the intake of non-supplement users. Models on the association of supplementation and food group or nutrient intake were adjusted for age, education, physical activity and smoking. To test for trend, each food variable was included as an interval-scaled variable in the regression model. For all analyses, SAS version 6.12 (SAS Institute Inc. Cary NC, USA) was used.

Results

Regular use of any of the listed dietary supplements was reported by 47.1% of female and 41.3% of male participants (Table 1). The majority of supplement users took vitamins, followed by minerals. Overall, 40.1% of women and 33.5% of men used vitamin and/or mineral nutrient supplements. Supplementation prevalence was significantly higher in women than in men, especially for the use of minerals. Likewise, there was a significant tendency towards a more frequent use of nutrient supplements (all kinds, vitamins and/or minerals, minerals) with older age (Table 1).

Table 1 Regular use of nutrient supplements in the EPIC-Heidelberg cohort at recruitment, by gender and age groups
Full size table

Regular use of only one type of supplement was reported by 33.2% of women and 29.5% of men, two different supplements were taken by 10.3% of women and 8.6% of men, 3.6% of women and 3.2% of men consumed three or more supplements. The most common combination in women was vitamins with minerals, which was consumed by 5.2% (men 3.3%), while men preferred vitamins with fibres 3.7% (women 3.9%).

Because vitamin and mineral supplements were the most frequently used types of dietary supplements in this cohort, our analyses on the associations of lifestyle and diet with supplement use refer to these two groups only.

Demographic and lifestyle factors

Women, older, more physically active as well as better educated subjects were more likely to use vitamin and mineral supplements (Table 2). In women, quitting of smoking is associated with a supplement use. The likelihood of supplement use is increased in women living without spouse or partner. Obese women took vitamins and/or mineral less often than women with normal BMI (Table 2). In addition, women, but not men, who had had a Hemoccult® test, were more likely to use supplements compared with subjects who had no test (Table 2).

Table 2 Association of vitamin and mineral supplementation with socio-demographic and lifestyle factors in the EPIC-Heidelberg cohort at recruitment
Full size table

Dietary factors

Neither total energy nor total fat intake was related to the use of vitamin or mineral supplements (Table 3). However, supplementation is associated with a lower intake of saturated fat and a higher intake of n-3 fatty acids, the latter resulting in a lower n6/n3-fatty acid ratio In addition, a high intake of the antioxidants vitamin C, vitamin E, and β-carotene was positively, although not always statistically significantly, associated with the use of vitamin or mineral supplements in men and women (Table 3).

Table 3 Association of dietary fatty acid and antioxidant intake with vitamin and mineral supplementation in the EPIC-Heidelberg cohort at recruitmenta
Full size table

A high intake of milk, milk products, fish, and cereals was positively associated with the use of vitamin or mineral supplements (Table 4). Men and women in the highest quintiles of meat, meat products, or animal fat intake were less likely to use supplements than subjects in the lowest quintiles of the respective food groups. In women, we found an inverse relation of cake, bread, and vegetable consumption with the use of supplements (Table 4). No statistically significant associations were seen between intake of fruits and fruit juices and the use of supplements.

Table 4 Association of selected food group consumption with vitamin and mineral supplementation in the EPIC-Heidelberg cohort at recruitmenta
Full size table

Discussion

In Germany, supplement use was shown to contribute up to 50% of the total intake of selected vitamins or minerals [4, 43]. It was the aim of this investigation to describe nutrient supplementation and to identify supplement-use related behaviour that might influence the risk of cancer and chronic diseases.

Our analysis conducted in the EPIC-Heidelberg cohort (1994–1998) is comparable to a previous investigation in the EPIC-Potsdam cohort (1995–1996), which relies on the same study design and, thus, shows good methodological comparability [22]. Regular dietary supplement use in EPIC-Heidelberg was reported by 47% of female and 41% of male participants and use of vitamin and/or mineral nutrient supplements by 40% of women and 33% of men. In EPIC-Potsdam, these numbers were distinctly lower, although relying on the same definition of regular supplement use. There, only 33% of women and 26% of men used supplements, vitamins and/or minerals were consumed by 25% of women and 18% of men. Participants in Heidelberg used more often fibres (16%) than those in Potsdam (10%), who, instead, more often used garlic supplements (22%) than subjects in the Heidelberg cohort (11%). We can only speculate about the reasons for these differences, but they may include factors such as differences in income, education, availability of different types of supplements, or dietary habits between Potsdam and Heidelberg.

The definition of regular supplement use and the observed time period has a distinct impact on the results. Thus, it is not surprising that in the present study, covering a period of 1 year and defining regular supplement use as ‘at least 4 weeks during the past year’, the prevalence of supplement use in the EPIC-Heidelberg cohort was higher than in other German studies. In the MONICA-Augsburg project, which examined supplement use within the last 7 days, the prevalence of vitamin and/or mineral supplement use was 25% in women and 18% in men [37]. In the German Nutrition Survey, the prevalence of vitamin and/or mineral supplement use, defined as regular use at least once a week during the last 12 months, was 22% in women and 18% in men [29]. Evaluation of a comparable variable in EPIC-Heidelberg (‘regular use of supplement during the last week’) revealed similar results for short-term nutrient supplementation (data not shown).

As reported in other studies [8, 16, 22, 23, 26–29, 34–29, 37–29, 41, 42], vitamin and mineral supplement use in the EPIC-Heidelberg cohort was significantly higher in women than in men, significantly more frequent in older subjects and in subjects with a higher education. Participants of the present study are better educated than a representative selection of German citizens of the same age [39] and better educated subjects are usually more health conscious. Regular engagement in leisure time physical activity is more common in supplement users than in non-users, an association that is confirmed by several European studies [22, 29, 33, 37]. The impact of leisure time physical activity on health is important and has to be kept in mind when comparing relative risks for different diseases among dietary supplement users and non-users [7, 44]. As in other studies [8, 18, 20, 26–29, 34–29, 41], supplement use was less common among current or heavy smokers and in obese subjects. Women who had had a Hemoccult® test were more likely to be users of vitamin and mineral supplements in our study. Similar associations have previously been reported in the French EPIC cohort [41] and two US studies [18, 34]. All these associations confirm the assumption that dietary supplements users tend to have a healthier lifestyle and a better awareness of health risks than non-users.

Higher consumption of milk, milk products, and fish was generally associated with a higher likelihood of using vitamin and mineral supplements, whereas high meat and meat product consumption was inversely associated. Other studies reported similar associations between food choices and use of supplements [5, 26, 28, 41]. However, in contrast to other reports [16, 20, 22, 26, 28, 34], the positive association between fruit or fruit juice consumption and the use of supplements in our cohort did not reach statistical significance. In female participants an even inverse association between vegetable consumption and supplement use was observed. Interestingly, supplement use did not differ by vegetable (men and women) or fruit (men only) consumption in the German Nutrition Survey either [5]. The associations between nutrient intake and use of vitamin and mineral supplements in our study reflect the differences in food choices. The inverse associations with saturated fat intake and n6/n3-fatty acid ratio can be seen as a consequence of the lower consumption of meat, meat products and animal fats and a higher consumption of fish and vegetable oils. In addition, a higher vitamin C supply in supplement users may result from the increased consumption of fruit and fruit juices. Similar results have been described in a French study [41] and US studies [20, 26, 28]. The higher likelihood of having a higher β-carotene intake in supplement users can only be explained by a preference for β-carotene-rich vegetables. A high intake of β-carotene, vitamin C (in men), and vitamin E in participants using supplements was also observed in the EPIC-Potsdam cohort. However, they did not find differences in fatty acid intake between supplement users and non-users [22].

In conclusion, data of our study indicate that the use of vitamin and mineral supplements in EPIC-Heidelberg is related to a more health conscious behaviour and can be regarded as one marker of a health conscious lifestyle. Studies on the protective effect of nutrients (including supplementation) on cancer and chronic disease risk should always be aware of this source of confounding.