Original Article | Published:

Epidemiology

Diet and risk of chronic diseases: results from the first 8 years of follow-up in the EPIC-Potsdam study

European Journal of Clinical Nutrition volume 67, pages 412419 (2013) | Download Citation

Contributors: AvR performed the analyses and wrote the paper. HB and MMB were responsible for the study design, data collection and provision of advice or consultation.All authors were responsible for critical review and revision of the manuscript.

Abstract

Background/Objectives:

There is still a need for scientific evidence about which foods characterize a healthy diet in terms of primary prevention of major chronic diseases. Therefore, we aimed to give a comprehensive overview on health-related foods, based on 8 years of follow-up of the European Prospective Investigation into Cancer and Nutrition (EPIC)-Potsdam study.

Subjects/Methods:

We used data from 23 531 participants of the EPIC-Potsdam study to analyse the associations between 45 single food groups and risk of major chronic diseases, namely, cardiovascular diseases (CVD), type 2 diabetes and cancer using multivariable-adjusted Cox regression. Habitual dietary intake was assessed at baseline using food-frequency questionnaires. Incident chronic diseases were determined by self-administered follow-up questionnaires and medically verified, based on inquiry to treating physicians, cancer registries or through death certificates.

Results:

During follow-up, 363 incident CVD, 837 type 2 diabetes and 844 cancer cases were identified. Higher intakes of whole-grain bread, raw vegetables, coffee and cakes and cookies were found to be significantly associated with a lower risk of chronic diseases. Conversely, higher intakes of low-fat dairy, butter, red meat and sauce were associated with higher risks of chronic diseases.

Conclusion:

Overall, a healthy diet was characterized by a high consumption of whole-grain bread, raw vegetables and a low consumption of red meat and possibly butter, which is generally in line with previous findings. The paradoxical findings concerning the potential health benefit of coffee as well as cakes and cookies are interesting and should be investigated further.

Introduction

Many cohort studies have been established over the last three decades with the particular aim to prospectively investigate the role of diet in disease occurrence. In these cohorts, dietary habits are usually assessed on the food level by questionnaires, which reflect the practice of dietary intake. A major advantage of analyses on the food group level is that the results are better interpretable compared with nutrients or complex dietary patterns, and therefore easier to transfer into recommendations on primary prevention of non-communicable chronic diseases. With regard to major chronic diseases (including cardiovascular diseases (CVD), type 2 diabetes and cancer), the most often mentioned dietary candidates for primary prevention are a high intake of fruits, vegetables and whole-grain products and a low intake of red or processed meat.1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12 However, publications from cohort studies mostly reported selectively about the associations of specific food groups in relation to selected disease endpoints. There is a lack of studies with a complete picture about total food group intake in relation to major chronic disease risk. Hence, we do not know whether there are some further food groups in addition to those with already known health effects that could have a role for preventing chronic diseases. Furthermore, from the public health perspective, it is of particular interest which food items may contribute to the prevention of the overall burden of chronic diseases instead of one specific disease.

Consequently, we systematically investigated a complete list of 45 food groups in relation to a combined outcome defined by the most frequent major chronic diseases (that is, CVD, type 2 diabetes and cancer), using data from the first 8 years of the European Prospective Investigation into Cancer and Nutrition (EPIC)-Potsdam study. Thereby, we can identify candidate foods to be potentially considered within food-based dietary guidelines for chronic disease prevention. This could also help improve the current German food-based dietary guidelines, as they showed only weak to moderate associations with the risk of major chronic diseases.13, 14

Materials and Methods

Study population

The EPIC-Potsdam study is a prospective cohort study among 27 548 participants aged mainly 35–65 years. Between 1994 and 1998, participants were randomly selected from the general population of Potsdam and adjacent communities by using residents’ registration offices, and invited by mail to take part in a baseline examination. A total of 22.7% of the originally invited individuals agreed to participate in the study. Written informed consent was obtained from all participants at recruitment to the study. The study was approved by the Ethical Committee of the Federal State of Brandenburg.15 At baseline, computer-guided interviews on lifestyle and medical history were conducted, self-administered questionnaires on food consumption and lifestyle were filled out by the participants and physical examinations were conducted by trained staff.16 During follow-up, the participants received a mailed follow-up questionnaire every 2–3 years to identify incident cases of chronic diseases. The response rate for each follow-up round ranged between 93 and 96%.17

We excluded participants that reported prevalent type 2 diabetes, CVD or cancer at baseline. Furthermore, individuals with missing follow-up time or withheld information on diet and relevant covariates considered in the analysis (lifestyle, anthropometry and sociodemographic variables) and those who reported an implausibly high or low energy intake (<800 or >6000 kcal/day; equates to about <3350 and >25 120 kJ/day, respectively) were excluded. Thus, 23 531 participants (14 433 women and 9098 men) remained for the analyses.

Dietary assessment

At baseline, the habitual dietary intake of several food groups was assessed by self-administered semiquantitative food-frequency questionnaires (FFQ). The baseline FFQ inquired about the frequency of intake and portion sizes of 148 food items consumed during the preceding 12 months. In addition, questions concerning fat content of dairy products or types of fat used for food preparation and intake of supplements were included.

For the single food items, the frequency scale ranged from ‘never’ to ‘five times a day’. Portion sizes were estimated using photographs with varying portion sizes or standardized household measures (for example, teaspoons, pieces or cups). Serving sizes displayed or inquired were derived from a representative national nutrition survey (Nationale Verzehrsstudie). Finally, dietary intake of each food item in grams per day was calculated by multiplying the reported frequencies per day and portion size.

Details about reproducibility and validity of the FFQ data have been published previously.18, 19, 20, 21 In brief, repeated administration of the FFQ to a subset of participants after 6 months were used to assess reproducibility. Spearman test–retest correlation coefficients varied from 0.49 (bread) to 0.89 (alcoholic beverages) (median=0.70). Validity was assessed by comparing FFQ data with the mean of 12 repeated 24-hour dietary recalls administered at monthly intervals in the same subset of individuals. The Spearman correlation coefficient ranged from 0.14 (legumes) to 0.90 (alcoholic beverages) (median=0.45).20

Case ascertainment

Information on incident chronic diseases (that is, CVD (myocardial infarction and stroke), type 2 diabetes and cancer) were obtained during follow-up from self-reports of the respective condition, disease-relevant medication or reasons for a reported change in diet. Additional diagnoses were found by record linkages with the Common Cancer registry of the Federal States Berlin, Brandenburg, Mecklenburg-West Pomerania, Saxony-Anhalt, the Free State of Saxony and Thuringia and the database of the clinical center of Potsdam.

All potential incident diseases were verified with an inquiry to the treating physician, local cancer registries and information of death certificates.17 In case of multiple diseases, only the first verified chronic disease was considered for the present analyses. This approach may help in minimizing confounding because people who have been diagnosed with a major chronic disease are likely to change their diet and lifestyle.

Incident diseases were coded based on the International Classification of Diseases (ICD-10 codes: I21 for myocardial infarction, I60, I61, I63, I64 for stroke, E11 for type 2 diabetes and C00-97 for cancer (except C44: non-melanoma skin cancer)).

Statistical analyses

Consistent with previous dietary pattern analyses in the EPIC-Potsdam study, 45 food groups (see Table 1) were considered,22 which were created on the basis of prior knowledge on their different health effects or nutritional value. These 45 food groups have been analysed individually and systematically (that is, by using analogous models) in relation to the first incident event of a major chronic disease. Hazard ratios and 95% confidence intervals for the first incident chronic disease (CVD, type-2 diabetes and cancer) were estimated per one-serving-per-day increment in food group intake using multivariable-adjusted Cox proportional hazard regression.23 The dependent time variable was defined as the time period between the age of recruitment and the age of exit (accurate age of diagnoses, death or censoring). The models were stratified by age (in years) to be less sensitive to violations of the proportional hazards assumption. Furthermore, all models were adjusted for sex, smoking status (never, former and current), pack-years of smoking, alcohol consumption (g/day), waist-to-hip ratio, body mass index (BMI) (kg/m2), leisure-time physical activity (walking, cycling and sports in hour/week), education (vocational training or lower degree vs trade school, technical school or university), use of vitamin supplements (yes/no), non-consumption of the respective food group (yes/no) and total energy intake (kJ/day). If CVD or type 2 diabetes was the considered outcome, the models were additionally adjusted for prevalent hypertension (yes/no) and history of high blood lipid levels (yes/no).

Table 1: Definition of 45 food groups

However, the investigation of single food groups does not allow taking into account the potential effect of underlying dietary patterns. Therefore, a second model was tested where all 45 food groups were mutually adjusted in one model.

Moreover, interactions of the single food groups (in servings per day) with sex in relation to chronic diseases were evaluated using likelihood-ratio tests and a significance level of 5% was applied.

All analyses were performed with SAS software 9.3 (SAS Institute Inc., Cary, NC, USA).

Results

Descriptive results concerning the intake distribution of the 45 food groups in men and women of the EPIC-Potsdam study are presented in the online supplement (Supplementary Tables 1 and 2).

During an average follow-up period of 8 years, 363 incident cases of CVD, 837 cases of type 2 diabetes and 844 cancer cases were identified and verified as first events. A comprehensive overview of the associations between 45 food groups and the risk of these major chronic diseases is presented in Table 2.

Table 2: Increase of food group intake by one serving per day and risk of major chronic diseases (CVD, type 2 diabetes and cancer) in the EPIC-Potsdam study

Within the group of cereal products, intake of whole-grain bread was shown to be significantly inversely associated with risk of major chronic diseases, specifically type 2 diabetes. Furthermore, intake of grain flakes and muesli showed a tendency of an inverse association with risk of CVD but not overall chronic diseases.

Concerning fruits and vegetables, interestingly, the consumption of raw but not cooked vegetables was inversely associated with chronic diseases, in particular type 2 diabetes and CVD. Specifically, a one-serving increment (100 g) in intake of raw vegetables was associated with a 14% decreased risk of chronic diseases.

Within the group of dairy products, surprisingly, a borderline significant direct association of low-fat dairy intake with risk of major chronic diseases, specifically type 2 diabetes, was observed.

When investigating beverages, coffee consumption was found to be inversely associated with risk of major chronic diseases, specifically type 2 diabetes. Besides, intake of fruit juice was associated with risk of type 2 diabetes but not overall chronic diseases.

Inside the group of added fat and oil, a direct association of intake of butter with risk of major chronic diseases, in particular CVD, was observed. Furthermore, intake of sauces, a food group that is also often characterized by a high-fat content, was found to be directly associated with major chronic diseases, specifically cancer. A further subdivision of this heterogeneous group indicated that this finding is attributable to sauce eaten together with meat or fish (data not shown). Moreover, some other added fats and oils were found to be related to specific chronic diseases but not to the combined endpoint. For instance, intake of other vegetable fat was inversely related to CVD risk. Conversely, intake of other fats (that is, animal cooking fat) showed contradictory results (direct association with type 2 diabetes but inverse relation to CVD), which complicates a general conclusion about this food group. Moreover, the intake of nuts, a food rich in n-3 fatty acids, was also borderline significantly inversely related to the risk of type 2 diabetes.

Concerning animal products, a one-serving (100 g) increment in the consumption of red meat was associated with an increase in risk of major chronic diseases, including type 2 diabetes and cancer, by 30%.

Finally, when investigating subgroups of sweets and snacks, we observed an apparently paradox inverse association of cakes and cookies consumption with the risk of major chronic diseases, specifically type 2 diabetes and cancer. Furthermore, also intake of sweet bread spreads was shown to be inversely associated with the risk of type 2 diabetes but not major chronic diseases. For pizza, the results were inconsistent as pizza consumption showed an inverse relationship with the risk of cancer but a direct association with the risk of CVD. These opposing results might be because of the heterogeneous composition of pizza (pizza usually contains not only higher proportions of vegetables and tomato sauce but also a high amount of saturated fat).

Figure 1 represents a graphical summary of the relevant (borderline) significant associations of single food groups with risk of overall major chronic diseases that are also shown in Table 2.

Figure 1
Figure 1

Overview of the (borderline) significant dose–response relationships of single food groups with major chronic disease risk. P<0.05 considered as significant; 0.5<=P<0.10 considered as borderline significant (observed for low-fat dairy). Each food group was entered individually into the model, stratified by age and adjusted for sex, smoking status, pack-years of smoking, alcohol consumption, leisure-time physical activity, BMI, waist-to-hip ratio, prevalent hypertension at baseline, history of high blood lipid levels at baseline, education, vitamin supplementation, non-consumption of the respective food group and total energy intake.

Moreover, sensitivity analyses were performed without inclusion of BMI and waist-to-hip ratio into the models, as these variables are discussed to be potential intermediates in the relationship between diet and risk of chronic diseases. Interestingly, the majority of the investigated relationships between single food groups and major chronic diseases were not markedly changed, except the direct association with intake of low-fat dairy, eggs and fish, as well as an inverse relationship with sweet bread spreads, which were strengthened towards statistical significance after the exclusion of BMI and waist-to-hip ratio from the model. Furthermore, with regard to diabetes, a significant direct association with the intake of processed meat was observed in models without BMI and waist-to-hip ratio (data not shown).

In an additional sensitivity analysis, where we excluded all disease cases that occurred during the first 2 years of follow-up, the associations between the 45 food groups and chronic diseases were generally not markedly changed.

In a second model, where all 45 food groups were mutually adjusted, intake of raw vegetables, coffee and whole-grain bread was still inversely associated, whereas intake of red meat and sauces was directly associated with risk of major chronic diseases, indicating that these foods were most strongly related to chronic diseases (see Table 3). Conversely, the significant inverse association between consumption of cakes and cookies with risk of major chronic diseases, as well as the direct associations of intake of low-fat dairy and butter with chronic diseases that was observed from Table 2 were non-significant in the model where all food groups were mutually adjusted. This implies that the findings for butter, low-fat dairy and cakes and cookies might be partly attributable to underlying dietary patterns. Nevertheless, one should keep in mind that the inclusion of too many covariates into one model can also negatively affect the precision of the effect estimates.

Table 3: Increase of food group intake by one serving per day and risk of major chronic diseases in the EPIC-Potsdam study, with mutual adjustment for further food groups

Supplementary Table 3 represents a summary of the sex-specific results for which a significant interaction with sex had been identified.

Concerning specific disease endpoints, an inverse relationship between muesli and flakes with CVD risk was found to be restricted to women. In addition, the inverse association of cake and cookie consumption with cancer risk was observed only in women. Besides, odd effect modifications by sex with reference to the association of water with CVD risk (direct association in women, inverse association in men) and cooked vegetables with risk of cancer (direct association in men but not in women) were identified.

Discussion

In the first 8 years of follow-up of the EPIC-Potsdam study, the investigation of 45 food items revealed that higher intakes of whole-grain bread, raw vegetables, coffee and cakes and cookies were inversely related to major chronic disease risk. Conversely, higher intakes of red meat, butter, sauces and low-fat dairy showed direct associations with risk of major chronic diseases. All other food items showed no significant association with the risk of chronic diseases or the risk relation was confined to a specific disease without affecting overall chronic disease risk.

For the discussion of the results, we would like to focus on the combination of major chronic diseases instead of specific diseases. In general, prevention of overall chronic diseases by diet may be of higher interest for the healthy population than targeted recommendations for the prevention of specific diseases, as, for most subjects, occurrence of a specific chronic disease has a lower probability than the occurrence of any chronic disease. In addition, from the public health perspective, it is pursued that recommendations may contribute to primary prevention of several disease outcomes. Nevertheless, it is important to know which specific diseases are affected by a dietary factor to get further insight into potential disease etiologies.24 Thus, we presented overall and disease-specific results. However, one needs to be aware that this analysis includes many comparisons, which increases the probability of false-positive results. We tried to minimize this type of error by highlighting only those foods that are associated with the overall risk of major chronic diseases, showed no effect modification by gender and no obvious heterogeneity between the diseases such as divergent directions of risk for the specific diseases.

We would like to discuss, first, the findings on health effects of foods that are already confirmed in the literature; second, foods being controversially discussed in the literature for a long time; and, finally, foods that should be newly considered to potentially have a role in primary prevention. The findings regarding whole-grain bread and vegetables as well as red meat are well reflected in the literature. For instance, the inverse relationship between whole-grain consumption and risk of chronic diseases (including type 2 diabetes, CVD and colorectal cancer) have been reported in several systematic review articles.3, 4, 6, 8, 10, 11 In addition, the potential protective effects of vegetables on risk of chronic diseases are also well explored.10, 12, 25 However, our study as well as other studies26, 27 provided data to prefer raw over cooked vegetables. This finding might be because of the higher content of water-soluble and heat-sensitive nutrients (for example, vitamin C) in raw vegetables.27 In addition, the observation that the intake of red meat is linked to many types of chronic diseases, particularly type 2 diabetes and (colorectal) cancer, is supported by recent meta-analyses.1, 7, 10, 11

The potential health effects of low-fat dairy, butter and sauces were discussed rather controversially in the past. For dairy products, including milk, yoghurt and curd, the results from the EPIC-Potsdam study indicated that low-fat products might not be the healthier choice compared with high-fat dairies. We observed that the intake of low-fat dairy products showed a tendency of a direct association with the risk of chronic diseases, specifically type 2 diabetes. Maybe this finding might be explained by the fact that an increased intake of low-fat products, such as low-fat dairy, could indicate some type of dieting because of weight and health problems. Nevertheless, this view is not shared by others, as in a recent meta-analysis, it was reported that the intake of dairy, especially low-fat dairy, was inversely associated the with risk of type 2 diabetes.28 We also observed that intake of butter, another dairy product, was directly related to risk of chronic diseases, particularly CVD. The number of other studies that examined the associations of intake of butter and risk of chronic diseases, such as CVD, is surprisingly limited, despite the ongoing discussion in the past on this food item in relation to plasma cholesterol.29, 30, 31 In a systematic review, Ellwood et al.32 could identify only five cohort studies that investigated the intake of butter in relation to CVD risk. These studies, in general, did not show statistically significant associations. Conversely, in the Nutrition Evidence Library published by the USDA, it is reported that there is strong evidence from the literature that dietary intake of saturated fatty acids is directly associated with risk of CVD.33 Finally, intake of sauces, typically associated with a western lifestyle, was found to be directly associated with the risk of chronic diseases, specifically cancer, independent from the intake of other food groups. Sauces are a very heterogeneous food group, and when the food group of sauces was disaggregated into separate food items in our study, the overall health effect of sauce appears to be driven by sauce eaten together with meat or fish. As meat is eaten in much higher amounts than fish within the EPIC-Potsdam study, it can be assumed that the direct association of sauce intake with cancer risk is primarily due to fat and fat-soluble substances from meat, which migrate to the sauce. Nevertheless, more studies are necessary to investigate the potential health effects of sauce intake.

The significant associations of coffee and cakes and cookies with chronic disease risk can potentially generate new hypotheses for the primary prevention of chronic diseases by diet. Coffee contains many bioactive compounds and was shown to be inversely related to the risk of major chronic diseases, in particular type 2 diabetes. This finding is in line with a previous analysis within EPIC-Germany34 as well as a meta-analysis by Huxley et al.35, who reported that one additional cup of coffee per day was significantly associated with a 7% decreased type 2 diabetes risk. Interestingly, we could observe that the consumption of cakes and cookies was also inversely associated with the risk of chronic diseases, in particular type 2 diabetes and cancer, which is contrary to expectations. This finding might be because of selective under-reporting of the intake of socially undesirable foods36, 37, 38, 39 or could have a biological plausibility. Up to now, the scientific knowledge on the long-term health effects of such type of foods is generally scarce. We could identify only one study that reported an inverse association of cakes and cookies with the risk of type 2 diabetes in women with a BMI<29.40 Intake of cakes and cookies is often associated with light or in between meals and thus provides continuous supply of carbohydrates over the day. This could influence glycemic control and might be important for the utilization of the different macronutrients for the production of glucose being essential for brain functioning. High carbohydrate intake at breakfast was also shown to reduce appetite compared with other dietary regimens,41 which in turn may contribute to body weight control. Within the EPIC consortium, inverse relationships between intake of cakes and cookies with weight gain as well as type 2 diabetes were also observed but are still under statistical evaluation (Annika Steffen, Brian Buijsse; personal communication).

The EPIC-Potsdam study has several advantages including the high number of participants, the high response rate at the follow-up investigations (93–96%)17 and the medical verification of potential chronic disease cases. Nevertheless, we are confronted with the problem that, in general, habitual dietary intake cannot be precisely assessed by using food-frequency questionnaires.42 Measurement error, for example, due to misreporting, might either mask existing relationships or even lead to completely biased estimates of relative risk. In this way, particularly the non-significant findings are not easy to interpret as these could be due to the effect of measurement error and not a lack of biological effects. However, we are not aware that the food items showing relationships with chronic diseases are more precisely assessed or less affected by misreporting than the food items showing no relationships. A further potential limitation is that we did not correct for multiple testing. However, as we generally have prior knowledge or expectations, each food group–disease association had a specific underlying hypothesis that was tested. Furthermore, we have considered this issue to a certain extent in the second model where all 45 food groups are included simultaneously into one model, showing that more than half of the initially observed significant relationships with risk of major chronic diseases was stable.

To our knowledge, this is the first analysis that simultaneously investigates all relevant food groups in relation to major chronic diseases. Interestingly, the results of the first 8 years of the EPIC-Potsdam study indicated that only a few food groups may have a role in the primary prevention of major chronic diseases, such as increased intake of whole-grain bread and raw vegetables, as well as a reduced consumption of red meat and possibly butter, which is in line with previous knowledge. We could also add to the empirical evidence that drinking coffee and occasionally eating cakes and cookies are inversely associated with chronic diseases, an issue that should be investigated in further cohort or, ideally, interventional studies. For us, these foods are prime candidates to be potentially considered by recommendations targeted at primary prevention of major chronic diseases. This study might also intensify the discussion how to organize food-based dietary guidelines, as adherence to recent German food-based dietary recommendations was shown to be associated only weakly to moderately with risk of chronic diseases.13, 14 Food-based dietary guidelines not only serve the purpose of providing all nutrients in the recommended amounts but also reduce chronic disease risk.43 With this study in the background, we propose that appropriate and approved candidates for chronic disease prevention should be prominently considered in such recommendations.

References

  1. 1.

    , , . Meat consumption and the risk of type 2 diabetes: a systematic review and meta-analysis of cohort studies. Diabetologia 2009; 52: 2277–2287.

  2. 2.

    , , , , , et al. Critical review: vegetables and fruit in the prevention of chronic diseases. Eur J Nutr 2012; 51: 637–663.

  3. 3.

    , , , , . Whole grain, bran, and germ intake and risk of type 2 diabetes: a prospective cohort study and systematic review. PLoS Med 2007; 4: e261.

  4. 4.

    DGE. Kohlenhydratzufuhr und Prävention ausgewählter ernährungsmitbedingter Krankheiten—Evidenzbasierte Leitlinie. Bonn: Deutsche Gesellschaft für Ernährung (DGE) (Hrsg.) [German Nutrition Society] 2011, (assessed 2011 Nov); Available at: .

  5. 5.

    , , , , , . Healthy living is the best revenge: findings from the European Prospective Investigation into Cancer and Nutrition-Potsdam Study. Arch Intern Med 2009; 169: 1355–1362.

  6. 6.

    , . Grain foods and health: a primer for clinicians. Phys Sportsmed 2008; 36: 18–33.

  7. 7.

    , , , , , et al. Red meat consumption and risk of type 2 diabetes: 3 cohorts of US adults and an updated meta-analysis. Am J Clin Nutr 2011; 94: 1088–1096.

  8. 8.

    , , , . Whole grain foods for the prevention of type 2 diabetes mellitus. Cochrane Database Syst Rev 2008, CD006061.

  9. 9.

    , , , , , et al. An accurate risk score based on anthropometric, dietary, and lifestyle factors to predict the development of type 2 diabetes. Diabetes Care 2007; 30: 510–515.

  10. 10.

    WCRF/AICR. Food, Nutrition, Physical Activity, and the Prevention of Cancer: a Global Perspective. World Cancer Research Fund/American Institute of Cancer Research: Washington D., 2007.

  11. 11.

    WCRF/AICR. Continuous Update Project: Colorectal Cancer Report 2010 Summary: Food, Nutrition, Physical Activity, and the Prevention of Colorectal Cancer. World Cancer Research Fund/American Institute of Cancer Research, 2011, Available at: .

  12. 12.

    WHO. Diet, nutrition and the prevention of chronic diseases. World Health Organ Tech Rep Ser 2003, 916: i-viii, 1-149, backcover.

  13. 13.

    , , . Does compliance with the recommendations of the DGE influence the risk of chronic diseases? [Article in German]. Ernahr-Umsch 2011; 58: 242–249.

  14. 14.

    , , , , . Adherence to recommendations of the German food pyramid and risk of chronic diseases: results from the EPIC-Potsdam study. Eur J Clin Nutr 2010; 64: 1251–1259.

  15. 15.

    , , . Recruitment procedures of EPIC-Germany. European Investigation into Cancer and Nutrition. Ann Nutr Metab 1999; 43: 205–215.

  16. 16.

    , , . EPIC-Germany—A source for studies into diet and risk of chronic diseases. European Investigation into Cancer and Nutrition. Ann Nutr Metab 1999; 43: 195–204.

  17. 17.

    , , . Follow-up procedures in EPIC-Germany—data quality aspects. European Prospective Investigation into Cancer and Nutrition. Ann Nutr Metab 1999; 43: 225–234.

  18. 18.

    , , , , . The relative validity of vitamin intakes derived from a food frequency questionnaire compared to 24-hour recalls and biological measurements: results from the EPIC pilot study in Germany. European Prospective Investigation into Cancer and Nutrition. Int J Epidemiol 1997; 26 (Suppl 1), S82–S90.

  19. 19.

    , , , . Reproducibility and relative validity of energy and macronutrient intake of a food frequency questionnaire developed for the German part of the EPIC project. European Prospective Investigation into Cancer and Nutrition. Int J Epidemiol 1997; 26 (Suppl 1), S71–S81.

  20. 20.

    , , , . Reproducibility and relative validity of food group intake in a food frequency questionnaire developed for the German part of the EPIC project. European Prospective Investigation into Cancer and Nutrition. Int J Epidemiol 1997; 26 (Suppl 1), S59–S70.

  21. 21.

    , , , , , et al. Validation of a self-administered food-frequency questionnaire administered in the European Prospective Investigation into Cancer and Nutrition (EPIC) Study: comparison of energy, protein, and macronutrient intakes estimated with the doubly labeled water, urinary nitrogen, and repeated 24-h dietary recall methods. Am J Clin Nutr 1999; 70: 439–447.

  22. 22.

    , , , . Dietary patterns and their association with food and nutrient intake in the European Prospective Investigation into Cancer and Nutrition (EPIC)-Potsdam study. Br J Nutr 2001; 85: 363–373.

  23. 23.

    . Regression models and life-tables. J R Stat Soc Ser B 1972; 34: 187–220.

  24. 24.

    , . Dietary pattern analysis for the evaluation of dietary guidelines. Asia Pac J Clin Nutr 2008; 17 (Suppl 1), 75–78.

  25. 25.

    DGE-Stellungnahme. Obst und Gemüse in der Prävention chronischer Krankheiten: Deutsche Gesellschaft für Ernährung (DGE) e.V.[German Nutrition Society] 2007, (assessed 2011 Nov); Available from .

  26. 26.

    , , , , . Raw and processed fruit and vegetable consumption and 10-year stroke incidence in a population-based cohort study in the Netherlands. Eur J Clin Nutr 2011; 65: 791–799.

  27. 27.

    , . Raw versus cooked vegetables and cancer risk. Cancer Epidemiol Biomarkers Prev 2004; 13: 1422–1435.

  28. 28.

    , , , , . Dairy consumption and risk of type 2 diabetes mellitus: a meta-analysis of cohort studies. Eur J Clin Nutr 2011; 65: 1027–1031.

  29. 29.

    , , , , , et al. Consumption of olive oil, butter, and vegetable oils and coronary heart disease risk factors. The Research Group ATS-RF2 of the Italian National Research Council. Jama 1990; 263: 688–692.

  30. 30.

    , . margarine and serum lipoproteins. Atherosclerosis 1997; 131: 7–16.

  31. 31.

    , , , . Effects of dietary fatty acids and carbohydrates on the ratio of serum total to HDL cholesterol and on serum lipids and apolipoproteins: a meta-analysis of 60 controlled trials. Am J Clin Nutr 2003; 77: 1146–1155.

  32. 32.

    , , , . The consumption of milk and dairy foods and the incidence of vascular disease and diabetes: an overview of the evidence. Lipids 2010; 45: 925–939.

  33. 33.

    USDA. What is the effect of saturated fat (SFA) intake on increased risk of cardiovascular disease or type 2 diabetes, including effects on intermediate markers such as serum lipid and lipoprotein levels?, US Department of Agriculture: Nutrition Evidence Library; 2010 (cited December, 30th 2011); Available at .

  34. 34.

    , , , , , . Coffee consumption and risk of chronic disease in the European Prospective Investigation into Cancer and Nutrition (EPIC)-Germany study. Am J Clin Nutr 2012; 95: 901–908.

  35. 35.

    , , , , , et al. Coffee, decaffeinated coffee, and tea consumption in relation to incident type 2 diabetes mellitus: a systematic review with meta-analysis. Arch Intern Med 2009; 169: 2053–2063.

  36. 36.

    , , , , , et al. Low energy reporting related to lifestyle, clinical, and psychosocial factors in a randomly selected population sample of Greek adults: the ATTICA Study. J Am Coll Nutr 2007; 26: 327–333.

  37. 37.

    , . Underreporting of energy intake among Japanese women aged 18-20 years and its association with reported nutrient and food group intakes. Public Health Nutr 2004; 7: 911–917.

  38. 38.

    , , , . Obesity as a major determinant of underreporting in a self-administered food frequency questionnaire: results from the EPIC-Potsdam Study. Z Ernährungswiss 1997; 36: 229–236.

  39. 39.

    , , , . Is macronutrient composition of dietary intake data affected by underreporting? Results from the EPIC-Potsdam Study. European Prospective Investigation into Cancer and Nutrition. Eur J Clin Nutr 1998; 52: 119–126.

  40. 40.

    , , , , , . Diet and risk of clinical diabetes in women. Am J Clin Nutr 1992; 55: 1018–1023.

  41. 41.

    , , , . Meal timing and composition influence ghrelin levels, appetite scores and weight loss maintenance in overweight and obese adults. Steroids 2012; 77: 323–331.

  42. 42.

    , , . Is it time to abandon the food frequency questionnaire? Cancer Epidemiol Biomarkers Prev 2005; 14: 2826–2828.

  43. 43.

    EFSA. Panel on Dietetic Products, Nutrition, and Allergies: Scientific Opinion on establishing Food-Based Dietary Guidelines. EFSA Journal 2010; 8: 1460 [42 pp].

Download references

Acknowledgements

We thank all of the EPIC-Potsdam study participants. Furthermore, we thank Ellen Kohlsdorf and Wolfgang Bernigau for data management, and Wolfgang Fleischhauer for case ascertainment. We also offer our special thanks to Wolfgang Bernigau for his support in peforming the statistical analyses with SAS. The recruitment phase of the EPIC-Potsdam Study was supported by the Federal Ministry of Science, Germany (grant 01 EA 9401) and the ‘Europe against Cancer’ program of the European Community (grant SOC 95 201408 05F02). The EPIC-Potsdam Study is now supported by the German Cancer Aid (grant 70-2488-Ha I) and the European Community (grant SOC 98 200769 05F02).

Author information

Affiliations

  1. Department of Epidemiology, German Institute of Human Nutrition Potsdam-Rehbruecke, Nuthetal, Germany

    • A von Ruesten
    • , S Feller
    • , M M Bergmann
    •  & H Boeing
  2. Department of Public Health and Community Medicine, University of Gothenburg, Gothenburg, Sweden

    • A von Ruesten

Authors

  1. Search for A von Ruesten in:

  2. Search for S Feller in:

  3. Search for M M Bergmann in:

  4. Search for H Boeing in:

Competing interests

The authors declare no conflict of interest.

Corresponding author

Correspondence to A von Ruesten.

Supplementary information

About this article

Publication history

Received

Revised

Accepted

Published

DOI

https://doi.org/10.1038/ejcn.2013.7

Supplementary Information accompanies the paper on European Journal of Clinical Nutrition website (http://www.nature.com/ejcn)