[Preventability estimates for colorectal and breast cancer in Germany. A methodological evaluation of the risk factors alcohol and overweight]. / Präventionspotenziale für Darm- und Brustkrebs in Deutschland. Eine methodische Evaluation von Schätzungen für die Risikofaktoren Alkohol und Übergewicht.

Population attributable fractions (PAFs) for the risk factors alcohol intake and overweight in the German population were calculated to estimate the preventability of colorectal and breast cancer attributable to these risk factors. Estimates were based on national alcohol consumption and overweight prevalence data in the German population. Comparative analyses were used to evaluate the variation of PAF estimates according to changes in the calculation parameters. PAFs quantify the preventive potential that could result from removing or reducing the risk factor exposure, respectively. Postmenopausal breast cancer was estimated to be preventable by 13-23% if the population were normal weight. Among German men, 10-25% of colon cancers were attributable to alcohol consumption and 7-13% of colorectal cancers to overweight. The comparative analyses demonstrated that preventability estimates vary considerably with the chosen data (risk estimates, reference categories) for PAF calculation and can differ by up to 50%. Thus, data selection should be evidence based, for example, based on meta-analyses, in order to increase the validity of preventability estimates.

Beer consumption and the 'beer belly': scientific basis or common belief?

BACKGROUND/OBJECTIVES: The term 'beer belly' expresses the common belief that beer consumption is a major determinant of waist circumference (WC). We studied the gender-specific associations between beer consumption and WC (partially in relation to body weight and hip circumference (HC) change). PARTICIPANTS/METHODS: Within the European Prospective Investigation into Cancer and Nutrition (EPIC)-Potsdam study (7876 men, 12 749 women), cross-sectional associations were investigated applying general linear models. Prospective analyses of baseline beer consumption and an 8.5-year WC change were assessed using multivariate general linear models and polytomous logistic regression. To test the site-specific effect of beer consumption on WC, an adjustment for concurrent changes in body weight and HC was carried out. In addition, the relationship between change in beer consumption and change in WC was studied. RESULTS: A positive association in men and no association in women were seen between beer consumption and WC at baseline. Men consuming 1000 ml/d beer were at 17% higher risk for WC gain compared with very light consumers. Significantly lower odds for WC gain (odds ratio=0.88; 95% confidence interval 0.81, 0.96) were found in beer-abstaining women than in very-light-drinking women. The adjustment for concurrent body weight and HC change diminished effect estimates notably, explaining most of the association between beer and change in WC. Decreasing beer consumption was related to higher relative odds for WC loss, although not statistically significant. CONCLUSIONS: Beer consumption leads to WC gain, which is closely related to concurrent overall weight gain. This study does not support the common belief of a site-specific effect of beer on the abdomen, the beer belly.

Associations of short-term weight changes and weight cycling with incidence of essential hypertension in the EPIC-Potsdam Study.

The aim of this study was to examine the relationships of short-term weight gain, weight loss, and weight cycling on the odds of developing hypertension. Normotensive middle-aged German men and women (n=12,362) of the European Prospective Investigation into Cancer and Nutrition-Potsdam Study were assigned to categories of 2-year short-term weight changes that were self-reported to have occurred prior to recruitment into the study (gain only, loss only, weight cycling, stable). After 2 years of follow-up after recruitment, 180 cases of incident essential hypertension were identified. In logistic regression models, odds ratios were estimated for the associations between short-term weight changes and risk of developing hypertension. Obesity status (BMI>or=30 or BMI<30 kg/m2) modified the associations between short-term weight change and incidence of diagnosed hypertension. Among obese individuals, short-term weight gain occurring during the 2 years prior to recruitment (OR=2.79, 95% CI 1.19-6.56), weight loss (OR=6.74, 95% CI 2.58-17.6) and weight cycling (OR=4.29, 95% CI 1.55-11.9) were strongly positively associated with incident hypertension, adjusted for age and gender, compared to obese individuals with short-term stable weight. No significant associations between short-term weight changes and risk of diagnosed hypertension were detected among non-obese individuals. Short-term weight changes appeared to present strong risk factors for developing hypertension among obese individuals. The effect seen for weight cycling supports the hypothesis that weight cycling increases the risk of hypertension. The finding for short-term weight loss may be explained by subsequent weight regain and needs further investigation.

[The problem of response in epidemiologic studies in Germany (Part II)]. / Zum Problem der Response in epidemiologischen Studien in Deutschland (Teil II).

The first part of this paper introduced various definitions of response and discussed their significance in the context of different study types. This second part addresses incentives as a method to increase response and evaluates the impact of non response or delayed response on the validity of the study results. Recruitment aims at minimising the proportion of refusal. To achieve this, incentives can be used and potential participants can be contacted in a sequence of increasing intensity. The effectiveness of different incentives was investigated within the pretest of the German survey on children and adolescents by the Robert Koch Institute. A low response is often interpreted in terms of non-response bias. This assumption, however, is as incorrect as would be opposite conclusion, that a high response guarantees valid results. Any study of the influence of nonresponse requires information on non-responders. The comparison between early and late responders as an indirect method to evaluate systematic differences between participants and non-participants by wave analysis is demonstrated within the Northern Germany Leukaemia and Lymphoma study (NLL). The German guidelines for Good Epidemiologic Practice recommend to solicit a minimum of information on the principal hypotheses of a study from non-participants. The example of a population-based health survey (Cooperative Health Research in the Region of Augsburg, KORA) illustrates how information on non-responders within a quantitative non-responder analysis can be achieved and used for the estimation of prevalences. Recommendations how to deal with the response in epidemiological studies in Germany are suggested.

[The problem of response in epidemiological studies in Germany (part I)]. / Zum Problem der Response in epidemiologischen Studien in Deutschland (Teil I).

To achieve high response rates in German epidemiological studies is growing more difficult. Low response in epidemiological studies may decrease the acceptance of the results. Response, however, is not identical with the quality of a study. In the first part of this paper various definitions of response (contact, cooperation, response, recruitment proportions) are introduced and discussed in the context of different study designs with reference to practical examples. A population-based survey such as the Study of Health in Pomerania (SHIP) investigates the distribution of risk factors and health-related endpoints. Surveys should yield representative results which can be generalised to apply to the entire population (external validity). This study design usually requires large participitation proportions. In a prospective cohort study such as the European Investigation into Cancer and Nutrition (EPIC) the emphasis is on internal validity. A stable study population willing to participate in regular follow-ups is a primary recruitment goal. If the response in a case-control study such as the Northern Germany Leukaemia and Lymphoma Study (NLL) is low, the priority is to achieve approximately equal response proportions for cases and controls. Simultaneous public relation and media activities can improve participitation in a study. Multidimensional strategies combining public communications, cooperation with local and regional officials and frequent press and media coverage are emphasised. The second part of this paper will discuss methods to quantify the effects of the response proportions on the validity of the study results.

History of the concept of 'levels of evidence' and their current status in relation to primary prevention through lifestyle interventions.

Primary prevention is a major option to reduce the burden of chronic disease in populations. Because lifestyle interventions have proved to be effective, lifestyle recommendations including nutritional advice are made abundantly. However, both their credibility and their effectiveness are often considered not to be high. Therefore, scientific evidence should form the basis of recommendations and, as in clinical medicine, a rational approach should be followed for the evaluation of evidence. In this paper, the development and current concepts of 'levels of evidence' as they are applied in clinical medicine are outlined and their impact on evidence-based recommendations is discussed. Next, the question is raised as to how far the existing schemes are applicable to the evaluation of issues pertaining to primary prevention through lifestyle changes. Current schemes were developed mainly for clinical research questions and therefore place major emphasis on randomised controlled trials as the main and most convincing evidence in the evaluation process. These types of study are rarely available for lifestyle-related factors and might even not be feasible to obtain. Arguments are advanced to support the notion that a modification of currently existing 'levels of evidence' as developed for clinical research questions might be necessary. Thereby, one might be able to accommodate the specific aspects of evidence-related issues of recommendations for primary prevention through lifestyle changes, like dietary changes.

Time trends in cigarette smoking in two German cohorts--results from EPIC Germany.

Smoking prevention is less advanced in Germany compared with other European and North American countries, and fewer data exist, especially on the consumption habits at the individual level over time. EPIC Germany, which is part of a European multicentre study on diet and cancer, collected data on individual smoking behaviour and allows for consideration of the changing consumption pattern for both centres and different age groups. Within EPIC 25 546 and 27 548 participants, respectively, were recruited in Heidelberg and Potsdam. Data on smoking habits were collected by means of a computer-guided interview during the baseline examination between 1994 and 1998. For each birth cohort smoking prevalence and mean number of cigarettes smoked per day at different ages were calculated. Additionally, the prevalence of non-filter cigarette smoking was computed. Smoking prevalence in the 1990s was still higher among men (Heidelberg 16.3-32.3%; Potsdam 18.2-29.3%) than among women (Heidelberg 12.8-32.0%; Potsdam 10.4-27.8%). However, the percentage of women smokers was still increasing. Filter cigarettes comprised a growing percentage of the cigarettes smoked, but especially among men differences between both German cohorts can still be seen: depending on age, 10.0-12.7% of men in the Heidelberg cohort smoked non-filter cigarettes, but only 1.1-2.3% in the Potsdam cohort. The quantity smoked was higher in the Heidelberg than in the Potsdam cohort with respect to the mean number of cigarettes smoked per day as well as the pack-years of smoking. In conclusion, smoking habits in the Potsdam and the Heidelberg cohorts did not strongly differ by smoking prevalence. However, they did differ according to the quantity and quality of smoking. These differences, as well as the changes over the last 40 years may contribute to a changing pattern of diseases in different groups of the German population.

Recent weight changes and weight cycling as predictors of subsequent two year weight change in a middle-aged cohort.

OBJECTIVE: To evaluate the influence of recent weight changes (weight gain, loss and cycling) on subsequent weight changes. DESIGN: Prospective cohort study with 2 y of follow-up. Data analysis with a polytomous logistic regression model. SUBJECTS: A total of 18 001 non-smoking subjects, 6689 men and 11 312 women, from the general population. MEASUREMENTS: Body height and weight measurements and interview data on lifestyle habits and medical history at baseline. For follow-up, self-administered questionnaires for assessment of body weight and incident diseases. RESULTS: Recent changes in body weight, that is weight gain, weight loss and weight cycling, were significant predictors of subsequent weight changes in both men and women after controlling for age, baseline BMI and several lifestyle and behavioural characteristics as potential confounding factors. Weight cycling before baseline was the strongest predictor of subsequent large weight gain (> or =2 kg) with an odds ratio (OR) of 4.84 (95% confidence interval (CI) 3.34-7.02) in men. In women, prior weight loss was the strongest predictor of subsequent large weight gain (OR 4.77; 95% CI 3.63-6.03), followed by weight cycling (OR 3.02; 95% CI 2.15-4.25). CONCLUSION: These data indicate the need for thorough weight history assessment to identify those who are most likely to gain weight. Effective weight control before the development of obesity or after intentional weight loss due to obesity should be a primary goal in the management of obesity.

European Prospective Investigation into Cancer and Nutrition (EPIC) calibration study: rationale, design and population characteristics.

The European Prospective Investigation into Cancer and Nutrition (EPIC), which covers a large cohort of half a million men and women from 23 European centres in 10 Western European countries, was designed to study the relationship between diet and the risk of chronic diseases, particularly cancer. Information on usual individual dietary intake was assessed using different validated dietary assessment methods across participating countries. In order to adjust for possible systematic over- or underestimation in dietary intake measurements and correct for attenuation bias in relative risk estimates, a calibration approach was developed. This approach involved an additional dietary assessment common across study populations to re-express individual dietary intakes according to the same reference scale. A single 24-hour diet recall was therefore collected, as the EPIC reference calibration method, from a stratified random sample of 36 900 subjects from the entire EPIC cohort, using a software program (EPIC-SOFT) specifically designed to standardise the dietary measurements across study populations. This paper describes the design and populations of the calibration sub-studies set up in the EPIC centres. In addition, to assess whether the calibration sub-samples were representative of the entire group of EPIC cohorts, a series of subjects' characteristics known possibly to influence dietary intakes was compared in both population groups. This was the first time that calibration sub-studies had been set up in a large multi-centre European study. These studies showed that, despite certain inherent methodological and logistic constraints, a study design such as this one works relatively well in practice. The average response in the calibration study was 78.3% and ranged from 46.5% to 92.5%. The calibration population differed slightly from the overall cohort but the differences were small for most characteristics and centres. The overall results suggest that, after adjustment for age, dietary intakes estimated from calibration samples can reasonably be interpreted as representative of the main cohorts in most of the EPIC centres.

Consumption of vegetables, fruit and other plant foods in the European Prospective Investigation into Cancer and Nutrition (EPIC) cohorts from 10 European countries.

OBJECTIVE: To describe and compare the consumption of the main groups and sub-groups of vegetables and fruits (V&F) in men and women from the centres participating in the European Prospective Investigation into Cancer and Nutrition (EPIC). DESIGN: Cross-sectional analysis. Dietary intake was assessed by means of a 24-hour dietary recall using computerised interview software and standardised procedures. Crude and adjusted means were computed for the main groups and sub-groups of V&F by centre, separately for men and women. Adjusted means by season, day of the week and age were estimated using weights and covariance analysis. SETTING: Twenty-seven centres in 10 European countries participating in the EPIC project. SUBJECTS: In total, 35 955 subjects (13 031 men and 22 924 women), aged 35-74 years, randomly selected from each EPIC cohort. RESULTS: The centres from southern countries had the highest consumption of V&F, while the lowest intake was seen in The Netherlands and Scandinavia for both genders. These differences were more evident for fruits, particularly citrus. However, slightly different patterns arose for some sub-groups of vegetables, such as root vegetables and cabbage. Adjustment for body mass index, physical activity, smoking habits and education did not substantially modify the mean intakes of vegetables and fruits. CONCLUSIONS: Total vegetable and fruit intake follows a south-north gradient in both genders, whereas for several sub-groups of vegetables a different geographic distribution exists. Differences in mean intake of V&F by centre were not explained by lifestyle factors associated with V&F intake.

Macronutrient, vitamin, and mineral intakes in the EPIC-Germany cohorts.

This article presents intakes of nutrients in the EPIC-Heidelberg and the EPIC-Potsdam (European Investigation into Cancer and Nutrition) studies. Estimates are based on standardized 24-hour dietary recalls. Recalls from 1,013 men and 1,078 women in Heidelberg and from 1,032 men and 898 women in Potsdam were included in the analysis. The estimated nutrient intake was based on the German Food Code and Nutrient Data Base version II.3. Analyses were carried out stratified by sex and weighted for the day of the week and age. Men in Potsdam reported significantly higher intakes of energy (mean Potsdam = 10,718 kJ, mean Heidelberg = 10,387 kJ) and higher intakes of vitamins and minerals as compared with men in Heidelberg. However, Heidelberg men consumed more alcohol, alpha-tocopherol, phosphorus, calcium, and magnesium. Potsdam women reported lower energy (mean Potsdam = 7,537 kJ, mean Heidelberg = 7,855 kJ), alcohol, and cholesterol intakes as compared with Heidelberg women. Vitamin and mineral intakes were lower too, except for retinol and ascorbic acid. The intakes of energy and most nutrients observed in the Potsdam and Heidelberg study populations were within the range reported from other German studies. The observed differences between both study populations indicate different dietary patterns, increasing the exposure variation in the EPIC study.

Comparison of self-reported alcohol intake with the urinary excretion of 5-hydroxytryptophol:5-hydroxyindole-3-acetic acid, a biomarker of recent alcohol intake.

Under-reporting of alcohol intake has been frequently reported. However, due to the lack of an objective reference method, e.g. a biomarker, information about the extent of under-reporting of alcohol intake obtained with dietary assessment instruments is not available. The objective of this study was to compare reported alcohol intake data derived from a 24 h recall with a biomarker of recent alcohol intake, the urinary excretion of 5-hydroxytryptophol (5-HTOL):5-hydroxyindole-3-acetic acid (5-HIAA). Embedded into the European Prospective Investigation into Cancer and Nutrition (EPIC)-Potsdam Study, Germany, a validation study that collected 24 h recall data and 24 h urine samples was conducted. Cohort study participants (n 107) volunteered to participate in this validation study. Among them were five subjects who reported no consumption of alcoholic beverages but had a 5-HTOL:5-HIAA ratio that indicated recent alcohol intake when the clinical cut-off point was taken as a judging criterion. After exclusion of these under-reporters, the Pearson's correlation coefficient between reported alcohol intake and the 5-HTOL:5-HIAA ratio was 0.92 (P<0.0001). Except for low alcohol intake of <0.1 g/kg body mass, a significant increase in 5-HTOL:5-HIAA excretion was observed with increasing amounts of alcohol intake. In conclusion, the 5-HTOL:5-HIAA excretion ratio appears to be a valuable quantitative biomarker of recent alcohol consumption. Denial of alcohol intake can be detected, but for the quantification of under-reporting of alcohol intake 24 h reference data are not yet available. With these data at hand, however, 5-HTOL:5-HIAA could become a biomarker for validation purposes in nutritional epidemiology.

Dietary patterns and their association with food and nutrient intake in the European Prospective Investigation into Cancer and Nutrition (EPIC)-Potsdam study.

Dietary pattern analysis has recently received growing attention, as it might be more appropriate in studies of diet-disease associations than the single food or nutrient approach that has dominated past epidemiological research. Factor analysis is a technique which is commonly used to identify dietary patterns within study populations. However, the ability of factor solutions to account for variance of food and nutrient intake has so far remained unclear. The present study therefore explored the statistical properties of dietary patterns with regard to the explained variance. Food intake of 8975 men and 13 379 women, assessed by a food-frequency questionnaire, was aggregated into forty-seven separate food groups. Dietary patterns were identified by principal component analysis and subsequent varimax rotation. Seven factors were retained for both men and women, which accounted for about 31 % of the total variance. The explained variance was relatively high (>40 %) for cooked vegetables, sauce, meat, dessert, cake, bread other than wholemeal, raw vegetables, processed meat, high-fat cheese, butter and margarine. Factor scores were used to investigate associations between the factors and nutrient intake. The patterns accounted for relatively large proportions of variance of energy and macronutrient intake, but for less variance of alcohol and micronutrient intake, especially of retinol, beta-carotene, vitamin E, Ca and ascorbic acid. In addition, factors were related to age, BMI, physical activity, education, smoking and vitamin and mineral supplement use.

Association between glycated hemoglobin and diet and other lifestyle factors in a nondiabetic population: cross-sectional evaluation of data from the Potsdam cohort of the European Prospective Investigation into Cancer and Nutrition Study.

BACKGROUND: Glycation reactions of proteins and other compounds, depending on blood glucose concentrations, have a detrimental effect on health. OBJECTIVE: The association of diet and other lifestyle factors with glycated hemoglobin (Hb A(1c)) values was examined in a nondiabetic population. DESIGN: This was a cross-sectional study of 1773 middle-aged men and women. Mean Hb A(1c) values were calculated for categories of diet and lifestyle factors, and odds ratios (ORs) for the highest versus lowest tertiles of Hb A(1c) were determined and compared. RESULTS: The OR of being in the highest Hb A(1c) tertile compared with the lowest increased with greater age [age 40-44 y compared with >60 y: men (OR: 2.86; 95% CI: 1.60, 5.20) and women: (6.11; 3.15, 12.30)] and greater obesity [body mass index (in kg/m(2)) >25 and waist-hip ratio >1.0 in men and >0.8 in women): men (2.80; 1.48, 5.45) and women (1.73; 1.15, 2.61)]. High energy and energy-adjusted saturated fat intakes were associated with increased risk of being in the highest tertile of Hb A(1c) [highest compared with lowest quintile: (1.53; 1.04, 2.26; P for trend = 0.013) and (1. 98; 1.33, 2.95; P for trend = 0.003), respectively]. No significant associations were observed for intakes of carbohydrates, protein, dietary fiber, or beta-carotene; however, some of the associations were nearly significant. Alcohol, vitamin C, and vitamin E intakes were inversely related to risk [highest compared with lowest quintile: (0.56; 0.38, 0.83; P for trend = 0.001), (0.50; 0.33, 0. 74; P for trend = 0.003), and (0.65; 0.43, 0.96; P for trend = 0. 036), respectively]. CONCLUSION: Hb A(1c) values might be modifiable by diet and other lifestyle factors.

Differences of blood pressure estimates between consecutive measurements on one occasion: implications for inter-study comparability of epidemiologic studies.

Currently, substantial variation in epidemiologic studies exists regarding the number of blood pressure (BP) readings obtained and the way in which they are combined. This might result in systematically different BP estimates. We therefore analysed data from 25,891 subjects (10,124 men and 15,767 women) of the EPIC-Potsdam Study (European Prospective Investigation into Cancer and Nutrition) to estimate the magnitude of differences between consecutive BP readings and their combinations. Three measurements with 2 min intervals were performed in the sitting position on the right arm with the supported arm elevated at heart level by trained interviewers using oscillometric devices. Mean BP declined from first to second reading and further to third reading by systolic 5.0/0.9 mmHg in men and 4.9/0.8 mmHg in women and by diastolic 1.5/0.3 mmHg in men and 1.9/0.5 mmHg in women, as well as pulse pressure (PP) (3.5/0.6 in men, 3.0/0.3 in women) and hypertension prevalence (9.1/1.7%-points). The magnitude of BP decline depended on BP level, age, body mass index (BMI), and BP medication. Combinations including the first reading lead to generally higher estimates than subsequent readings or their combination. Published data on mean BP, PP and hypertension prevalence depend on the number and subsequent handling of BP readings which might introduce bias to the comparison of different studies unless the same defined readings were used. The combination of the second and third reading seems to be favourable over any single reading or other combinations.

Measures of quality control in the German component of the EPIC study. European Prospective Investigation into Cancer and Nutrition.

Quality control is an indispensable part of quality assurance in any study, intending to ensure high standards during data acquisition. The aim of this paper is to describe the measures of quality control undertaken in the German EPIC study centers and to present selected results of these procedures (EPIC = European Prospective Investigation into Cancer and Nutrition). For all data assessment tools applied in the German EPIC study, procedures were developed to monitor both the personnel as well as the technical instruments. These procedures combined quantitative and qualitative measurements of quality control. Interviewer performance was evaluated through direct observation and rated according to an evaluation score. Blood pressure and anthropometric measurements were both controlled through direct observation of measurement procedures as well as through periodical technical control of measurement devices. Blood sampling procedures were directly monitored and subsequent handling of the probes tightly recorded, including information on time sequence of work-up and room temperature. With these diverse control measurements and the obtained rating of assessment procedures a broad pool of information has been made available to support a critical evaluation of the data obtained in the EPIC study centers in Heidelberg and Potsdam.

Dietary habits in the German EPIC cohorts: food group intake estimated with the food frequency questionnaire. European Investigation into Cancer and Nutrition.

In the baseline assessment of the two EPIC-Germany cohorts Heidelberg and Potsdam, dietary information was obtained with an identical food frequency questionnaire (FFQ). The optically readable FFQ was designed to assess the usual food and nutrient intake of individuals during the past 12 months. The present analysis was based on dietary data from 25,212 participants in Heidelberg (11,776 men, 13,436 women) and 26,270 participants in Potsdam (10,249 men, 16,021 women). This paper presents the first results of a descriptive dietary analysis on a food group level based on 16 food groups and selected subgroups. Each of these food groups and subgroups was divided into quintiles, and the age-adjusted mean intake for each quintile was calculated. The comparison of dietary habits between the two cohorts, as well as the comparison between men and women within each cohort showed clear differences both in the quintiles of most food groups as well as in the range between the lowest and highest quintile. Except for the food groups non-alcoholic and alcoholic beverages, sugar and confectionery, sauces, and soups, men and women participating in Potsdam reported higher intakes of all the other food groups. The amount of food intake was generally lower in women than in men, with the exception of vegetables, fruit, dairy products, and non-alcoholic beverages. Further differences between the study centers were observed regarding the use of cooking fat for meat and vegetable preparation. In conclusion, the dietary variation, e.g. the exposure variation, was increased by recruiting two geographically distinct cohorts, instead of only one, in Germany.

Quantitative food intake in the EPIC-Germany cohorts. European Investigation into Cancer and Nutrition.

The EPIC-Heidelberg and the EPIC-Potsdam studies with about 53,000 study participants represent the German contribution to the EPIC (European Investigation into Cancer and Nutrition) cohort study. Within the EPIC study, standardized 24-hour dietary recalls were applied as a quantitative calibration method in order to estimate the amount of scaling bias introduced by the varying center-specific dietary assessment methods. This article presents intake of food items and food groups in the two German cohorts estimated by 24-hour quantitative dietary recalls. Recalls from 1,013 men and 1,078 women in Heidelberg and 1,032 men and 898 women in Potsdam were included in the analysis. The intake of recorded food items or recipe ingredients as well as fat used for cooking was summarized into 16 main food groups and a variety of different subgroups stratified by sex and weighted for the day of the week and age. In more than 90% of the recalls, consumption of dairy products, cereals and cereal products, bread, fat, and non-alcoholic beverages, particularly coffee/tea, was reported. Inter-cohort evaluations revealed that bread, potatoes, fruit and fat were consumed in higher amounts in the Potsdam cohort while the opposite was found for pasta/rice, non-alcoholic, and alcoholic beverages. It was concluded that the exposure variation was increased by having two instead of one EPIC study centers in Germany.