Original Article

European Journal of Clinical Nutrition (2009) 63, S206–S225; doi:10.1038/ejcn.2009.82

Contribution of highly industrially processed foods to the nutrient intakes and patterns of middle-aged populations in the European Prospective Investigation into Cancer and Nutrition study

Guarantor: N Slimani.
Contributors: NS initiated and wrote this paper, taking into account comments from all co-authors, and was the overall coordinator of this project and of the EPIC Nutrient DataBase (ENDB) project. CB carried out the statistical analysis and preparation of tables and figures. GD was in charge of recoding dietary data according to the project-specific food reclassification, under the supervision of NS and DATS. DATS acted as an external expert on food chemistry and helped with the reclassification according to food processing methods. NS, GD, DATS, CB, VC, MMEvB, MCBR, AMcT, SG, JVK, IH, PA and MJ were members of the writing group and gave input on statistical analysis, drafting of the article and interpretation of results. MCBR, AMcT, SG, JVK, PA, PZ, AT, EW, GJ, SR, AKI, AB, LR, MT, MN, AM, FC, MCO, YTvdS, BB, CL, MB, AH, AT, AMJ and SB were local EPIC collaborators involved in collecting data, checking the project-specific food reclassification and documenting, compiling and evaluating the subset of their national nutrient databases used in the ENDB. ER is the overall coordinator of the EPIC study. All co-authors provided comments and suggestions on the article and approved the final version.

N Slimani1, G Deharveng1, D A T Southgate2,, C Biessy1, V Chajès1,3, M M E van Bakel1, M C Boutron-Ruault4, A McTaggart5, S Grioni6, J Verkaik-Kloosterman7, I Huybrechts1, P Amiano8, M Jenab9, J Vignat1, K Bouckaert1, C Casagrande1, P Ferrari1,28, P Zourna10, A Trichopoulou10, E Wirfält11, G Johansson12, S Rohrmann13, A-K Illner14, A Barricarte15, L Rodríguez16, M Touvier4,17, M Niravong4, A Mulligan5, F Crowe18, M C Ocké7, Y T van der Schouw19, B Bendinelli20, C Lauria21, M Brustad22, A Hjartåker23, A Tjønneland24, A M Jensen25, E Riboli26 and S Bingham5,27,

  1. 1Dietary Exposure Assessment Group, International Agency for Research on Cancer, Lyon, France
  2. 28 Penryn Close, Norwich, Norfolk, UK
  3. 3Institut Gustave Roussy, CNRS FRE 2939, Villejuif, France
  4. 4Inserm, ERI 20, Institut Gustave Roussy, Villejuif, France
  5. 5Department of Public Health and Primary Care, MRC Centre for Nutritional Epidemiology in Cancer Prevention and Survival, University of Cambridge, Cambridge, UK
  6. 6Department of Preventive & Predictive Medicine, Nutritional Epidemiology Unit, Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy
  7. 7National Institute for Public Health and the Environment (RIVM), Bilthoven, The Netherlands
  8. 8Public Health Department of Gipuzkoa, Basque Government, San Sebastian and CIBER Epidemiología y Salud Pública (CIBERESP), Spain
  9. 9Lifestyle and Cancer Group, International Agency for Research on Cancer, Lyon, France
  10. 10Department of Hygiene, Epidemiology and Medical Statistics, University of Athens Medical School, Athens, Greece
  11. 11Department of Clinical Sciences, Lund University, Malmö, Sweden
  12. 12Department of Nutritional Research, University of Umeå, Umeå, Sweden
  13. 13Division of Cancer Epidemiology, German Cancer Research Center (DKFZ), Heidelberg, Germany
  14. 14Department of Epidemiology, German Institute of Human Nutrition, Potsdam-Rehbrücke, Germany
  15. 15Institute of Public Health of Navarra, Pamplona and CIBER Epidemiología y Salud Pública (CIBERESP), Spain
  16. 16Public Health and Participation Directorate, Health and Health Care Services Council, Asturias, Spain
  17. 17AFSSA (French Food Safety Agency), DERNS/PASER, Maisons-Alfort, France
  18. 18Cancer Epidemiology Unit, University of Oxford, Oxford, UK
  19. 19Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, Utrecht, The Netherlands
  20. 20Molecular and Nutritional Epidemiology Unit, ISPO, Florence, Italy
  21. 21Cancer Registry, Azienda Ospedaliera ‘Civile-M.P.Arezzo’, Ragusa, Italy
  22. 22Institute of Community Medicine, University of Tromsø, Tromsø, Norway
  23. 23Cancer Registry of Norway, Oslo, Norway
  24. 24Institute of Cancer Epidemiology, Danish Cancer Society, Copenhagen, Denmark
  25. 25Institute of Preventive Medicine, Copenhagen, Denmark
  26. 26Department of Epidemiology, Public Health and Primary Care, Imperial College, London, UK
  27. 27Diet and Cancer Group, MRC Mitochondrial Biology Unit, Cambridge, UK

Correspondence: Dr N Slimani, Dietary Exposure Assessment Group, International Agency for Research on Cancer (IARC), WHO, Lyon, France. E-mail: Slimani@iarc.fr

The authors are deceased.

28Current address: Data Collection and Exposure Unit (DATEX), European Food Safety Authority, Parma, Italy.





To describe the contribution of highly processed foods to total diet, nutrient intakes and patterns among 27 redefined centres in the 10 countries participating in the European Prospective Investigation into Cancer and Nutrition (EPIC).



Single 24-hour dietary recalls were collected from 36034 individuals (aged 35–74 years) using a standardized computerized interview programme (EPIC-SOFT). Centre-specific mean food intakes (g/day) were computed according to their degree of food processing (that is, highly, moderately and non-processed foods) using a specifically designed classification system. The contribution (%) of highly processed foods to the centre mean intakes of diet and 26 nutrients (including energy) was estimated using a standardized nutrient database (ENDB). The effect of different possible confounders was also investigated.



Highly processed foods were an important source of the nutrients considered, contributing between 61% (Spain) and 78–79% (the Netherlands and Germany) of mean energy intakes. Only two nutrients, β-carotene (34–46%) and vitamin C (28–36%), had a contribution from highly processed foods below 50% in Nordic countries, in Germany, the Netherlands and the United Kingdom, whereas for the other nutrients, the contribution varied from 50 to 91% (excluding alcohol). In southern countries (Greece, Spain, Italy and France), the overall contribution of highly processed foods to nutrient intakes was lower and consisted largely of staple or basic foods (for example, bread, pasta/rice, milk, vegetable oils), whereas highly processed foods such as crisp bread, breakfast cereals, margarine and other commercial foods contributed more in Nordic and central European centres.



Highly industrially processed foods dominate diets and nutrient patterns in Nordic and central European countries. The greater variations observed within southern countries may reflect both a larger contribution of non/moderately processed staple foods along with a move from traditional to more industrialized dietary patterns.


24-h dietary recall, standardisation, processed foods, industrial foods, nutrient patterns, EPIC-SOFT



Two major historical periods have introduced profound changes in human diet and other lifestyle factors. The introduction of agriculture and animal husbandry in the neolithic period (~10000 years ago) and more recently the industrial revolution (~200 years ago) have led to an increased consumption of certain foods (for example, dairy products, cereals and cereal products, refined sugars, vegetable oils, salt) and of a myriad of processed foods that were virtually absent from pre-agricultural hunter-gatherer diets (Eaton et al., 1997; Cordain et al., 2005; Eaton, 2006). The substitution of unprocessed or modestly processed foods by more complex, refined (highly processed) foodstuffs may have affected several metabolic and nutritional characteristics of ancestral human diets that had remained unchanged over millions of years (for example, glycaemic load, fatty acid composition, macronutrient composition, micronutrient density, acid–base balance, sodium–potassium ratio and fibre content) (Cordain et al., 2005). The inability to adapt genetically to these recent changes is hypothesized to be one of the possible explanations for the increased incidence of obesity and chronic diseases (for example, type 2 diabetes, cardiovascular disease, cancer) from the mid twentieth century onwards (Eaton and Konner, 1985; Tooby and Cosmides, 1990; Kious, 2002; Cordain et al., 2005; Ulijaszek, 2007).

The development of intensive food production and industrialization was started in the eighteenth century in Europe and in the United States. The first objectives were to provide reliable food supplies, to improve microbiological quality and to devise means of preserving fresh and perishable foods. This was obtained through increasingly sophisticated preservation and processing techniques that changed food structure, nutritional content, texture and taste. These processing technologies varied according to food types, and involved packaging, moisture removal, heat treatments, chilling and freezing, acidity control, chemical additives and irradiation (Karmas and Harris, 1988). New and complex food products that combined natural and artificial ingredients, including additives, thus became widely available. Sugars, salt and fats, available at a relatively low cost, were also extensively used for preservation purposes, to make foods more palatable or as convenient ingredients to prevent rancidity and improve texture (for example, hydrogenated fats and margarine in cakes, biscuits and bakery products) (van Erp-Baart et al., 1998). In the first instance, these urban-industrialized food systems helped to improve life conditions and life expectancy and, with the increasing availability of elaborate ready-to-eat foods and dishes, responded to time scarcity in food preparation and cooking (Jabs and Devine, 2006). Since the mid twentieth century, however, a growing body of scientific evidence suggests that increased consumption of industrialized foods increases the risk of various chronic diseases (WHO/FAO, 2003; Schulze et al., 2004; Cordain et al., 2005; Mozaffarian et al., 2006; Pomerleau et al., 2006; Ulijaszek, 2007; WCRF/AICR, 2007).

Although several characteristics of urban-industrialized food systems have been investigated in relation to disease, there is a scarcity of data to evaluate these specific dietary patterns across populations. A better understanding of the contribution of (highly) processed foods to current diets across Europe will help to improve the design of future studies on the association between chronic diseases and dietary patterns rich in (highly) processed foods, and to formulate more targeted public health recommendations. The main objective of this study was to use a set of comparable, highly detailed dietary data to investigate the contribution of (highly) industrially processed foods to nutritional intakes and patterns in 27 middle-aged European population groups participating in the EPIC study.


Materials and methods

Study population

The population involved in this analysis comes from a calibration substudy nested in the European Prospective Investigation into Diet and Cancer (EPIC) study, a large cohort study undertaken in 23 centres in 10 countries: Denmark, France, Germany, Greece, Italy, Norway, Spain, Sweden, the Netherlands and the United Kingdom (Riboli et al., 2002; Bingham and Riboli, 2004). The main rationale of the calibration study was to correct for measurement errors in baseline country-specific dietary measurements and to attenuate bias in relative risk estimates through the use of a common standardized 24-h dietary recall (24-HDR) computerized interview programme (EPIC-SOFT) (Kaaks et al., 1995; Slimani et al., 1999; Ferrari et al., 2008). For the calibration study, a stratified random sample (36994 participants, ~8% of the total EPIC cohort) recorded a single 24-HDR with a trained interviewer between 1995 and 2000. Most of the EPIC participants were recruited from the general population, except in France (women state employees recruited from a local health insurance), in Turin and Ragusa in Italy and Spain (blood donors), Utrecht, the Netherlands and Florence, Italy (women participating in breast cancer screening), and a British cohort of vegans and ovo-lacto vegetarians (“health-conscious” cohort, recruited from around the United Kingdom). In Norway, only women from the general population were recruited. The initial 23 EPIC administrative centres were redefined into 27 geographical regions relevant to the analysis of dietary consumption patterns (Slimani et al., 2002a). In this paper, “central” (European) centres represent those located in Germany, the Netherlands and the United Kingdom, whereas southern centres are those in Greece, Italy, Spain and France, and northern centres are those in Sweden, Denmark and Norway. More details on the rationale and characteristics of the calibration study are provided elsewhere (Slimani et al., 2002a).

After a systematic exclusion of individuals under the age of 35 or over 74 years, because of low participation in these age categories (n=960), a total of 36034 individuals with 24-HDR data were finally included in this analysis. Approval for the study was obtained from the ethical review boards of the International Agency for Research on Cancer (Lyon, France) and the local EPIC collaborating centres. All participants provided written informed consent.

Dietary variables

Using the EPIC-SOFT computer programme, detailed, highly standardized dietary information was obtained through a single 24-HDR interview administered face-to-face in all centres (Slimani et al., 1999), except in Norway where it was obtained by telephone interviews (Brustad et al., 2003). Details on the rationale, structure and validity of EPIC-SOFT for between-population comparisons are reported elsewhere (Slimani et al., 2002b, 2003; Al-Delaimy et al., 2005; Ferrari et al., 2009; Saadatian-Elahi et al., 2009).

Food definition and classification of industrially processed foods

To investigate the consumption of highly industrially processed foods in EPIC, as opposed to non- and moderately processed foods, common definitions and reclassification of the reported EPIC-SOFT food items were specifically developed with the support of an internationally recognized food chemistry expert, Professor DAT Southgate (Greenfield and Southgate, 2003). Each reported food was recoded according to its degree of processing and was classified into three main categories for which the food group-specific processes and examples are summarized in Table A1 of the Appendix.

Highly processed foods: Foods that have been industrially prepared, including those from bakeries and catering outlets, and which require no or minimal domestic preparation apart from heating and cooking (for example, bread, breakfast cereals, cheese, commercial sauces, canned foods including jams, commercial cakes, biscuits and sauces).

Moderately processed foods: This category includes two sets of foods. First, industrial and commercial foods involving relatively modest processing and consumed with no further cooking such as dried fruits, raw vacuum-packed or under controlled atmosphere foods (for example, salads), frozen basic foods, extra virgin olive oil, fruits and vegetables canned in water/brine or in own juice. Second, foods processed at home and prepared/cooked from raw or moderately processed foods (for example, vegetables, meat and fish cooked from raw fresh ingredients, or vacuum-packed, deep-frozen, canned in water/brine or in own juice).

Non-processed foods: Foods consumed raw without any further processing/preparation, except washing, cutting, peeling, squeezing (for example, fruits, non-processed nuts, vegetables, crustaceans, molluscs, fresh juices).

Foods with unknown process: Foods for which the processing involved is unknown, based on the information provided by the study subjects (for example, unknown preservation method for vegetables, milk, meat or information missing in homemade or commercially processed foods such as cakes and cream desserts). This category was relatively marginal, below 5% in most centres.

To increase comparability within and between centres, all the recoding and reclassification work was carried out at the food/ingredient level, after mixed recipes were broken down into ingredients. All the ingredients of ‘industrial/commercial’ recipes were coded as industrially processed foods, whereas those of homemade recipes were coded depending on whether the ingredients used were raw, moderately or industrially processed. Cakes, biscuits, sauces and soups, treated as foods in EPIC-SOFT, were broken down a posteriori using broad estimations of the relative proportions of non-industrially versus industrially processed foods, when they were made at home. Water was excluded from the comparisons as it was not systematically recalled in all centres.

Nutrient databases

Nutrient intakes and derived patterns were calculated by means of standardized nutrient databases developed through the EPIC Nutrient DataBase (ENDB) project. The rationale and procedures used to improve between-country comparability of the 26 nutrients included in this analysis are described elsewhere (Slimani et al., 2007).

Other lifestyle variables

The other lifestyle variables including education level, total physical activity and smoking history considered in this analysis were collected at baseline through standardized questionnaires and clinical examinations (Riboli et al., 2002), and have been described for the calibration sample elsewhere (Slimani et al., 2002a).

Data on age as well as body weight and height were self-reported by participants during the 24-HDR interview.

The mean time interval between these baseline questionnaire measures and the 24-HDR interview varied by country, from 1 day to 3 years later (Slimani et al., 2002a).

Statistical methods

All analyses, unless otherwise specified, were performed after stratification by gender and centre, and ordered according to a geographical south–north gradient. Tables 1a and b provide the mean consumption of foods and beverages (g/day and standard error (s.e.)) by centre and their relative contribution as a percentage of the total diet according to their degree of food processing. Centre-specific mean intakes and their percentages were adjusted for age, height, weight and energy intake and weighted by season and day of 24-HDR to control for differences in sampling procedures. This was computed in a multivariate regression model (analysis of covariance—ANCOVA) and weighted using generalized linear models (‘GENMOD’ procedure in SAS software). The effect of other possible confounders (smoking status, physical activity, BMI and education level) on mean intakes of highly processed foods (g/day) and centre rankings was also examined, by comparing models with or without the variable of interest (data not shown). The R-square of the model and the partial R-square of the additional co-variable were calculated. Significance was assessed by means of the partial F-test.

The contribution of different types of food subgroups to the centre mean energy intake from highly processed foods was investigated first (Figures 1a and b). Multi-dimensional graphic representations (Figures 2a–l) were then used to illustrate the percentage contribution of highly processed foods to the total centre mean intakes of 26 selected nutrients (including energy) after adjustment for gender, age, energy, height and weight and was weighted by season and weekday. Unless otherwise specified, the comparisons within or between centres are indicated in percentage points.

Figure 1.
Figure 1 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

(a) Contributions of highly processed staple/basic foods to centre mean total energy intakes (%) after adjustment for season, weekday, height, weight, age and gender. (b) Contributions of the other more complex highly processed foods to centre mean total energy intakes (%) after adjustment for season, weekday, height, weight, age and gender. (1) Flour, flakes, dough, pastries. (2) Fruit and vegetable juices, coffee, tea, herbal tea, chicory, non-alcoholic beer, alcohol for cooking. (3) Potato-, vegetable-, legume- and fruit-products (including olives and nuts), chocolate, candies, ice creams, sorbets, salty biscuits, sauces, vegetarian foods, dietetic products, creamers, snacks. (4) Yogurt, cheese, milk beverages, curd, cream dessert, dairy cream.

Full figure and legend (368K)

Figure 2.
Figure 2 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

(al) Percentage contribution of highly processed foods to the total centre mean intakes of 26 selected nutrients (including energy) after adjustment for gender, age, energy, height, weight and weighted by season and weekday.

Full figure and legend (338K)



Mean food and beverage consumption according to degree of food processing

The mean total dietary consumption ranged from 1980g/day and 1489g/day (Greece) to 3601g/day and 3176g/day (Heidelberg, Germany) for men and women, respectively (Tables 1a, Table 1b). In comparison with Nordic and central regions, southern countries (Greece, Spain and Italy), and to a lesser extent France, reported much lower contributions of beverages (water excluded) to total dietary consumption. In Spain and Greece, women systematically reported a 4–7% higher contribution of foods to total intakes as compared with men, with an equivalent lower contribution of beverages. In the other centres, the gender difference was less than 4% (Tables 1a and b).

There was a strong geographic gradient for the contribution of foodstuffs to total mean consumption of foods and beverages according to the degree of processing. In both men and women, the mean contribution of highly processed foods ranged from 35–43% in Murcia (Spain) and Ragusa (Italy) to ~60% in the Netherlands, Sweden, Norway, Denmark and the UK general population. Moderately processed foods represented 20–25% of total food intake in Italy, the United Kingdom, the Netherlands, Germany, Granada (women, Spain), Sweden, Norway and Denmark. Slightly higher percentages (27–38%) were observed in Greece, France and in most Spanish centres. For men, the contribution of non-processed foods ranged from 11–14% of total intake in the UK general population, Sweden, Denmark and the Netherlands (Bilthoven), to 25–32% in several southern centres in Spain, Italy and Greece. In women, there was a similar trend, although in almost all centres they had a higher relative contribution of non-processed foods than men, up to 5–7% higher in Varese and Ragusa (Italy), Granada (Spain), the UK general population, Germany, Sweden and Denmark.

Beverage consumption (excluding water) consisted almost exclusively of highly processed commercial beverages (for example, alcoholic beverages, tea and coffee), with values above 85% in most centres and higher values in men than in women, by up to 11–13% in Granada (Spain) and Greece. The contribution of non-processed beverages (for example, fresh fruit juices) was marginal in most centres (<3% in both genders), except in Spain (men 5%, women 9.5%, in the southern centres of Granada and Murcia) and Greece (men 10%, women 21%).

Main food sources of energy intake from highly processed foods

Energy intake was selected as the most relevant nutritional indicator for assessing the contribution of highly processed foods to the overall diet, taking into account both food and beverage energy sources. We consistently observed that three main food groups contributed the greatest amount to energy intake from highly processed foods. Cereals and cereal products were by far the most important energy providers in all centres, with values ranging from ~30% (Potsdam, Germany) to ~58% (Ragusa, Italy), followed by dairy products (10–23%) or alternatively fats (5–27%), depending on centre. Greater differences were observed across centres in the types of the next food groups contributing most to energy intake (<1–12%). Two complementary figures (Figures 1a–b) show comparisons of the contribution to the centre mean energy intake of highly processed foods from relatively staple/basic foods (for example, bread, pasta, rice, milk, vegetable oils) (Figure 1a) versus other more highly processed foods (for example, cakes, biscuits, breakfast cereals, crisp bread, confectionery, processed meat and fish, milk beverages, yoghurt, cheese, cream desserts, margarines and other hardened fats and alcoholic beverages) (Figure 1b). It seemed that ~31–45% of the mean energy intake in the southern centres (Greece, Spain, Italy) was provided by highly processed staple/basic foods. In contrast, in the Nordic and central European countries, 50–56% of energy was provided by other more highly processed foods. France (44–46%) and the UK health-conscious group (48%) reported patterns similar to the Nordic and central European countries.

Contribution of highly processed foods to centre mean nutrient intake and patterns

The contribution (%) of highly industrially and commercially processed foods to centre mean nutrient intakes and patterns is summarized by country in Figures 2a–j. After adjustment for age, gender, weekday, season, energy, height and weight, a strong geographical gradient was observed, with two distinct sets of nutrient patterns, namely, the Nordic and central European countries versus the southern European countries.

Nordic and central European centres

In the UK (general population), Germany, the Netherlands, Sweden, Denmark and Norway (Figure 2e–j), highly processed foods provided less than 50% of total intake for only two nutrients: β-carotene (34–46%) and vitamin C (28–36%). For all other nutrients and energy, highly processed foods provided between 50 and 99% of intake. A large proportion (76–79%) of the intake of energy, fat and carbohydrate components, magnesium, phosphorus, potassium, calcium, retinol and alcohol was provided by highly processed foods, and a moderate proportion for vitamin B6 (50–55%), dietary fibre (53–66%), potassium (57–62%), vitamin B1 (58–68%), cholesterol (58–69%), iron (60–70%) and protein (62–68%). Greater variability was observed among the Nordic and central European centres for vitamin D (62–80%) and vitamin E (62–78%).

There were virtually no differences in the overall nutrient patterns (less than or equal to3% for most nutrients) between centres within Denmark, Sweden or Norway (Figures 2h–j). More notable between–centre differences (3–9%) were observed for certain nutrients in the Netherlands (for example, simple sugars, fibre, vitamin E, vitamin B12 and β-carotene) and in Germany (for example, β-carotene, vitamin D, vitamin B12 and cholesterol). For the United Kingdom, the health-conscious group and the general population sample had a similar contribution of highly processed foods for a number of nutrients such as total fat, fatty acids and phosphorus (centre difference less than or equal to1%), and for energy, starch, fibre, iron, magnesium, alcohol, retinol, β-carotene and most B-vitamins (centre difference 2–5%). Differences were greater between these two British population groups for vitamin B2 (6%), potassium (7%), vitamin C (8%), calcium and carbohydrates (10%) and sugar (16%), with higher contributions from highly processed foods in the general population sample, whereas for protein (6%), vitamin D (11%) and cholesterol (20%), the proportion of intake was higher for the health-conscious group.

When considered together, the overall contribution of highly processed foods to mean nutrient intakes showed similar overlapping shapes among all Nordic and central European countries (Figure 2k).

Southern European centres

In contrast to Nordic and central European countries, southern centres in Spain, Italy, Greece and, to a lesser extent, France showed different nutrient patterns and greater variability within and between countries (Figures 2a–d and l). Overall, we observed a much lower contribution to the mean nutrient intake from highly processed foods, particularly in Spain and Italy, and for vitamin C (3–24%), β-carotene (14–27%), vitamin B6 (26–43%), iron (29–47%), fibre (28–53%), cholesterol (26–60%) and potassium (26–60%).

For energy, highly processed foods contributed 61–65% in all Spanish centres and Ragusa (Italy) and 72–74% in the French regions and Greece. For macronutrients, alcohol came essentially from highly processed foods (94–98%), whereas greater variability within and between countries was reported for the other macronutrients. For protein, the range varied from 38% (San-Sebastian, Spain) to 61% (Naples, Italy). For total fat (53–86%) and fat subtypes (48–89%), the contribution of highly processed foods varied widely across southern countries, with values in Greece (81–89%) similar to those reported in Nordic and central European centres. However, highly processed foods contributed much less to cholesterol intakes, with values ranging from 26% (San Sebastian, Spain) to 60% (Varese, Italy). In contrast to iron (35–56%) and potassium (29–47%), larger contributions from highly processed foods were observed for calcium (58–79%), magnesium (43–67%) and phosphorus (43–64%), although they were still significantly lower than those reported in Northern and Central Europe. Retinol (54–86%), vitamin D (39–67%) and vitamin B1, B2 and B12 (36–68%) showed greater contributions from highly processed foods and greater variability across southern countries, as compared with vitamin C (3–24%), β-carotene (14–27%) and vitamin B6 (26–43%). In contrast to Nordic and central centres (Figure 2k), nutrient patterns in southern countries do not overlap, but show distinct local dietary habits and greater heterogeneity in the contribution of highly processed foods within and between countries (Figure 2l).

Gender differences

When men and women were considered together, the P-value of the interaction between gender and the contribution of highly processed foods to centre mean intakes was statistically significant for energy and all nutrients except thiamine. In most cases, these gender differences did not exceed 10%. For fibre and β-carotene, for example, women had statistically lower contributions from highly processed foods than did men in most centres, with differences of 2–9% for fibre and 1–13% for β-carotene.

Influence of covariates

The influence of possible confounders (smoking status, physical activity and education level) on centre rankings and mean intakes of highly processed foods (g/day) was also examined by comparing the baseline model (adjusted for age, energy, height, weight, weekdays, seasons) with and without variables of interest. For BMI, the effect was estimated using a model without height and weight to remove the dependence between these co-variables. In men, level of education, physical activity, BMI and smoking, each considered separately, had no effect on the centre mean intakes of highly processed foods (<1%) and ranking, and could not explain the total variability in the mean intake of highly processed foods. In women, smoking and physical activity had a higher impact on centre ranking, although the effect on centre mean intakes of highly processed foods was relatively modest for smoking (up to 2–3% in the Spanish centres) and for physical activity (less than or equal to1%).



The main objective of our analysis was to investigate how highly industrially processed foods actually contribute to overall diet and nutrient intakes in middle-aged European populations, using a unique dataset of detailed, standardized 24-HDR measurements. This study showed that highly industrially processed foods dominate the diets and particularly nutrient patterns in Nordic and central European countries, whereas non-processed and highly processed staple foods contribute more to the dietary and nutrient intakes in southern countries.

For this study, we devoted much effort to developing ad hoc common definitions and food classifications of processed foods to enable dietary comparisons across the 27 centres participating in the EPIC study. In this detailed descriptive analysis, the definitions and terminology used for highly industrially processed foods, as opposed to raw or moderately processed foods, were conservative and independent of any a priori knowledge of diet–disease associations. Only the food processes involved and criteria used to discriminate between the different degrees of food processing were considered. For example, staples such as cereal-based foods (for example, bread, pasta, white rice), milk and vegetable oils that undergo relatively complex food processing are included among the highly processed foods, although their consumption tends to be inversely associated with several chronic diseases (Hu and Willett, 2002; Mann, 2007; van Dam and Seidell, 2007).

This study shows that in middle-aged populations from Nordic and central European regions, highly processed foods provide less than 50% of total intake of only two of the nutrients considered in the analysis (vitamin C and β-carotene), whereas the figures range from 50 to 91% for the others (excluding alcohol). Surprisingly, despite large differences in the qualitative and quantitative dietary patterns reported elsewhere in the same populations (Slimani et al., 2002b), highly industrially processed foods contribute in relatively similar proportions within and between the Nordic and central European countries for a large series of nutrients, including energy. Greater variability was reported in southern centres, with women in France showing contributions in between southern and Nordic/central regions. Furthermore, when considering the qualitative types of highly processed foods, we observed that ~31–45% of total energy intake in Greece, Spain and Italy comes from basic or staple foods (for example, bread, pasta/rice, milk, vegetable oils). Similar results have been reported elsewhere (Karamanos et al., 2002; Tur et al., 2004; Garcia-Closas et al., 2006). This suggests that staple foods still make important contributions to the diet in these countries, although dietary patterns are observed to be changing, with a move away from traditional to commercial foods, particularly among younger generations (Cruz, 2000; Parizkova, 2000). In contrast, (variants of) the so-called ‘Western’ dietary patterns, characterized by a diet rich in (saturated) fats, red meat, sugary desserts and refined grains and low in fresh fruits and vegetables, poultry and/or fish, were clearly established in Nordic and central European countries, with relatively modest contributions to total energy intake from highly processed staple foods (21–29%) and non-processed foods. In Umeå (Sweden), for example, more highly processed foods such as crispbread, breakfast cereals, cakes, biscuits, margarine, dairy products excluding milk, processed meats, alcoholic drinks and soft drinks comprise up to 56% of total energy intakes.

Currently, insufficient specifically designed epidemiological and intervention studies have been carried out to draw firm conclusions on the effects of industrially processed foods on disease risk. Increasingly, however, direct and indirect evidence points to adverse effects of industrially processed foods on the pandemic of obesity (Swinburn et al., 2004; Astrup et al., 2008) and of various chronic diseases such as type II diabetes (van Dam et al., 2002), cardiovascular disease (Hu et al., 2000; Fung et al., 2004) and cancer (Kesse et al., 2006; Ambrosini et al., 2008; Campbell et al., 2008; Chajes et al., 2008). Of particular concern are specific features of highly processed foods such as the fact that they are high in energy, fats, sugar and salt and poor in dietary fibre (Astrup et al., 2008), that other compounds may be added or generated during food processing (for example, colourants, additives, acrylamide, trans-fatty acids) (Dybing et al., 2005; Astrup et al., 2008; McCarthy et al., 2008) and that they are usually consumed in large portion sizes (Matthiesen et al., 2003). Trans-fatty acid formation, for example, results from industrial partial hydrogenation of vegetable oils (hardened vegetable oils) (Sommerfeld, 1983). These partially hydrogenated oils are consumed in margarine, fast foods and highly processed foods (cakes, rolls, confectionery, biscuits, chocolate, potato crips and chips) and have been associated with different chronic diseases (Mozaffarian et al., 2006; Chajes et al., 2008). Many processed foods, including foods from major companies, still contain high levels of industrially produced trans-fatty acids, despite efforts to reduce them (Aro et al., 1998; McCarthy et al., 2008). A cross-sectional study on plasma fatty acid levels among a sub-sample of our study population (N=3003) shows a high correlation (r=0.72, P<0.01) between margarines and elaidic acid (trans 18:1 n-9), a specific biomarker of hardened fats and their related industrial foodstuffs (Saadatian-Elahi et al., 2009). Higher levels of plasma elaidic acid were observed in northern and central than in southern regions, which is compatible with the geographical differences in the consumption of highly processed foods reported in our analysis.

More research is required to understand the effects of industrially processed foods on health. Furthermore, rapid dietary changes and the massive introduction of industrial foods to diet raise questions on how to measure and monitor them properly. These include the following:

(1) The shortcomings of dietary assessment methods used so far in nutritional studies. There is cumulating methodological evidence that traditional food frequency questionnaires used in nutritional epidemiology have several limitations for measuring current dietary exposure and its association with diseases compared with open-ended methods such as repeated food records and recalls (Bingham et al., 2003; Kristal et al., 2005). New approaches that use a combination of dietary assessment techniques, including specific biomarkers and calibration, are increasingly recommended for estimating dietary exposures (Kaaks et al., 1997; Subar et al., 2006, Ferrari et al., 2008).

(2) The difficulty in obtaining reliable information from the food industry on the composition of foodstuffs, including commercial recipes, new products such as food supplements and functional foods, is currently a major limitation for estimating and monitoring the consumption of processed foods and their association with diseases.

(3) Subsequently, and despite important efforts to consolidate existing data and generate new data on commercial foods (www.eurofir.net), current food composition tables are inadequate for accurately measuring the levels of exposure to industrially processed foodstuffs or to bioactive components added or generated during food processing (for example, hardened and other fats, trans-fatty acids, acrylamide, colourants and additives). However, although imprecise, nutrient databases remain essential for identifying the main sources of food components of interest and providing relevant public health recommendations. The increasing use of specific biomarkers of industrial food exposures (for example, to trans-fatty acids, acrylamide, colourants, etc. for example) should be privileged and further investigated in addition to, or as a substitute for, dietary exposure measurements.

(4) The lack of specifically designed epidemiological studies to investigate the role of food processing—broadly defined as including all the processes involved in transforming basic ingredients into manufactured foods and beverages (production, processing and preservation methods)—in the development of chronic diseases. The limited evidence from epidemiological studies was the main reason provided recently by a World Cancer Research Fund panel for not drawing firm conclusions on the relation between food processing and cancer (WCRF/AICR, 2007). Only certain specific domestic or industrial food processing methods (for example, processed meat preserved by smoking, curing, salting or addition of chemicals, Cantonese-style salted preserved fish, frying/grilling, barbecuing and refining), known to be associated with certain cancers, were reported. More specific targeted studies on processed foods should be promoted. This research should also include gene–gene and gene–environment interaction studies, as different hypotheses suggest that a possible explanation for the epidemic of chronic diseases is the inability of humans to adapt genetically to the major recent changes in diet (Eaton and Konner, 1985; Baschetti, 1998; Cordain et al., 2005; Ulijaszek, 2007).

In conclusion, this study shows that highly industrially processed foods dominate diets and nutrient patterns in Nordic and central European countries. The wider range in the consumption of highly processed foods observed in southern countries and the greater contribution of non- or moderately processed staple foods may reflect changes in dietary patterns moving from traditional to more industrialized diets. However, our study, conducted from 1995 to 2000 in middle-aged populations, most likely underestimates the current situation in strictly representative populations, particularly among younger and poorer, more vulnerable populations. In view of the high consumption of industrialized foods in western European populations, the quality of these foods, particularly in terms of possibly harmful components, should be better monitored and evaluated through concerted actions between policy makers, public health actors, scientists and the food industry.

Supplementary information

Contribution of staple versus more complex highly processed foods for all nutrients, except energy, which is provided in this paper, is available on the EPIC website (http://epic.iarc.fr).


Conflict of interest

M Jenab has received grant support from the World Cancer Research Fund. S Bingham has received grant support from MRC Centre. The remaining authors have declared no financial interests.



  1. Al-Delaimy WK, Slimani N, Ferrari P, Key T, Spencer E, Johansson I et al. (2005). Plasma carotenoids as biomarkers of intake of fruits and vegetables: ecological-level correlations in the European Prospective Investigation into Cancer and Nutrition (EPIC). Eur J Clin Nutr 59, 1397–1408. | Article | PubMed | ISI | ChemPort |
  2. Ambrosini GL, Fritschi L, de Klerk NH, Mackerras D, Leavy J (2008). Dietary patterns identified using factor analysis and prostate cancer risk: a case control study in Western Australia. Ann Epidemiol 18, 364–370. | Article | PubMed
  3. Aro A, Van Amelsvoort J, Becker W, van Erp-Baart M-A, Kafatos A, Leth T et al. (1998). Transfatty acids in dietary fats and oils from 14 European countries: The TRANSFAIR Study. J Food Comp Anal 11, 137–149. | Article | ChemPort |
  4. Astrup A, Dyerberg J, Selleck M, Stender S (2008). Nutrition transition and its relationship to the development of obesity and related chronic diseases. Obes Rev 9(Suppl 1), S48–S52. | Article
  5. Baschetti R (1998). Diabetes epidemic in newly westernized populations: is it due to thrifty genes or to genetically unknown foods? J R Soc Med 91, 622–625. | PubMed | ChemPort |
  6. Bingham SA, Luben R, Welch A, Wareham N, Khaw KT, Day N (2003). Are imprecise methods obscuring a relation between fat and breast cancer? Lancet 362, 212–214. | Article | PubMed | ISI
  7. Bingham S, Riboli E (2004). Diet and cancer--the European Prospective Investigation into Cancer and Nutrition. Nat Rev Cancer 4, 206–215. | Article | PubMed | ISI | ChemPort |
  8. Brustad M, Skeie G, Braaten T, Slimani N, Lund E (2003). Comparison of telephone vs face-to-face interviews in the assessment of dietary intake by the 24 h recall EPIC SOFT program--the Norwegian calibration study. Eur J Clin Nutr 57, 107–113. | Article | PubMed | ChemPort |
  9. Campbell PT, Sloan M, Kreiger N (2008). Dietary patterns and risk of incident gastric adenocarcinoma. Am J Epidemiol 167, 295–304. | Article | PubMed
  10. Chajes V, Thiebaut AC, Rotival M, Gauthier E, Maillard V, Boutron-Ruault MC et al. (2008). Association between serum trans-monounsaturated fatty acids and breast cancer risk in the E3N-EPIC Study. Am J Epidemiol 167, 1312–1320. | Article | PubMed
  11. Cordain L, Eaton SB, Sebastian A, Mann N, Lindeberg S, Watkins BA et al. (2005). Origins and evolution of the Western diet: health implications for the 21st century. Am J Clin Nutr 81, 341–354. | PubMed | ChemPort |
  12. Cruz JA (2000). Dietary habits and nutritional status in adolescents over Europe--Southern Europe. Eur J Clin Nutr 54(Suppl 1), S29–S35. | Article | PubMed | ISI | ChemPort |
  13. Dybing E, Farmer PB, Andersen M, Fennell TR, Lalljie SPD, Müller DJG et al. (2005). Human exposure and internal dose assessments of acrylamide in food. Food Chem Toxicol 43, 365–410. | Article | PubMed | ChemPort |
  14. Eaton SB, Konner M (1985). Paleolithic nutrition. A consideration of its nature and current implications. N Engl J Med 312, 283–289. | PubMed | ISI | ChemPort |
  15. Eaton SB, Eaton III SB, Konner MJ (1997). Paleolithic nutrition revisited: a twelve-year retrospective on its nature and implications. Eur J Clin Nutr 51, 207–216. | Article | PubMed | ISI | ChemPort |
  16. Eaton SB (2006). The ancestral human diet: what was it and should it be a paradigm for contemporary nutrition? Proc Nutr Soc 65, 1–6. | Article | PubMed | ChemPort |
  17. Ferrari P, Day NE, Boshuizen HC, Roddam A, Hoffmann K, Thiebaut A et al. (2008). The evaluation of the diet/disease relation in the EPIC study: considerations for the calibration and the disease models. Int J Epidemiol 37, 368–378. | Article | PubMed
  18. Ferrari P, Roddam A, Fahey MT, Jenab M, Bamia C, Ocké M et al. (2009). A bivariate measurement error model for nitrogen and potassium intakes to evaluate the performance of regression calibration in the European Prospective Investigation into Cancer and Nutrition study. Eur J Clin Nutr 63(Suppl 4), S179–S187. | Article
  19. Fung TT, Stampfer MJ, Manson JE, Rexrode KM, Willett WC, Hu FB (2004). Prospective study of major dietary patterns and stroke risk in women. Stroke 35, 2014–2019. | Article | PubMed | ISI
  20. Garcia-Closas R, Berenguer A, Gonzalez CA (2006). Changes in food supply in Mediterranean countries from 1961 to 2001. Public Health Nutr 9, 53–60. | Article | PubMed | ISI
  21. Greenfield G, Southgate DAT (2003). Food Composition Data: Production, Management and Use. FAO: Rome.
  22. Hu FB, Rimm EB, Stampfer MJ, Ascherio A, Spiegelman D, Willett WC (2000). Prospective study of major dietary patterns and risk of coronary heart disease in men. Am J Clin Nutr 72, 912–921. | PubMed | ISI | ChemPort |
  23. Hu FB, Willett WC (2002). Optimal diets for prevention of coronary heart disease. JAMA 288, 2569–2578. | Article | PubMed | ISI | ChemPort |
  24. Jabs J, Devine CM (2006). Time scarcity and food choices: an overview. Appetite 47, 196–204. | Article | PubMed
  25. Kaaks R, Riboli E, van Staveren W (1995). Calibration of dietary intake measurements in prospective cohort studies. Am J Epidemiol 142, 548–556. | PubMed | ISI | ChemPort |
  26. Kaaks R, Riboli E, Sinha R (1997). Biochemical markers of dietary intake. In: P Toniolo, P Boffetta, D Shuker, N Rothman, B Hulka, N Pearce (eds) Application of Biomarkers in Cancer Epidemiology. IARC Sci Publ No 142. International Agency for Research on Cancer: Lyon. pp 103–126.
  27. Karamanos B, Thanopoulou A, Angelico F, Assaad-Khalil S, Barbato A, Del Ben M et al. (2002). Nutritional habits in the Mediterranean Basin. The macronutrient composition of diet and its relation with the traditional Mediterranean diet. Multi-centre study of the Mediterranean Group for the Study of Diabetes (MGSD). Eur J Clin Nutr 56, 983–991. | Article | PubMed | ISI | ChemPort |
  28. Karmas E, Harris RS (1988). Nutritional Evaluation of Food Processing 3rd edn. Van Nostrand Reinhold: New York.
  29. Kesse E, Clavel-Chapelon F, Boutron-Ruault MC (2006). Dietary patterns and risk of colorectal tumors: a cohort of French women of the National Education System (E3N). Am J Epidemiol 164, 1085–1093. | Article | PubMed | ISI | ChemPort |
  30. Kious BM (2002). Hunter-gatherer nutrition and its implications for modern societies. eScholarship Repositories, UCLA Biol Chem (http://repositories.cdlib.org/uclabiolchem/nutritionnoteworthy/
    ). Nutrition Noteworthy 5, Article 3, pp 1–5.
  31. Kristal AR, Peters U, Potter JD (2005). Is it time to abandon the food frequency questionnaire? Cancer Epidemiol Biomarkers Prev 14, 2826–2828. | Article | PubMed
  32. Mann J (2007). Dietary carbohydrate: relationship to cardiovascular disease and disorders of carbohydrate metabolism. Eur J Clin Nutr 61(Suppl 1), S100–S111. | Article | PubMed | ChemPort |
  33. Matthiesen J, Fagt S, Biltoft-Jensen A, Beck AM, Ovesen L (2003). Size makes a difference. Public Health Nutr 6, 65–72. | Article | PubMed
  34. McCarthy J, Barr D, Sinclair A (2008). Determination of trans fatty acid levels by FTIR in processed foods in Australia. Asia Pac J Clin Nutr 17, 391–396. | PubMed | ChemPort |
  35. Mozaffarian D, Katan MB, Ascherio A, Stampfer MJ, Willett WC (2006). Trans fatty acids and cardiovascular disease. N Engl J Med 354, 1601–1613. | Article | PubMed | ISI | ChemPort |
  36. Parizkova J (2000). Dietary habits and nutritional status in adolescents in Central and Eastern Europe. Eur J Clin Nutr 54(Suppl 1), S36–S40. | Article | PubMed
  37. Pomerleau J, Lock K, McKee M (2006). The burden of cardiovascular disease and cancer attributable to low fruit and vegetable intake in the European Union: differences between old and new member states. Public Health Nutr 9, 575–583. | Article | PubMed
  38. Riboli E, Hunt KJ, Slimani N, Ferrari P, Norat T, Fahey M et al. (2002). European Prospective Investigation into Cancer and Nutrition (EPIC): study populations and data collection. Public Health Nutr 5, 1113–1124. | Article | PubMed | ISI | ChemPort |
  39. Saadatian-Elahi M, Slimani N, Chajès V, Jenab M, Goudable J, Biessy C et al (2009). Plasma phospholipid fatty acid profiles and their association with food intakes: results from a cross-sectional study within the European Prospective Investigation into Cancer and Nutrition (EPIC). Am J Clin Nutr 89, 331–346. | Article | PubMed | ChemPort |
  40. Schulze MB, Manson JE, Ludwig DS, Colditz GA, Stampfer MJ, Willett WC et al. (2004). Sugar-sweetened beverages, weight gain, and incidence of type 2 diabetes in young and middle-aged women. JAMA 292, 927–934. | Article | PubMed | ISI | ChemPort |
  41. Slimani N, Deharveng G, Charrondiere RU, van Kappel AL, Ocke MC, Welch A et al. (1999). Structure of the standardized computerized 24-h diet recall interview used as reference method in the 22 centers participating in the EPIC project. European Prospective Investigation into Cancer and Nutrition. Comput Methods Programs Biomed 58, 251–266. | Article | PubMed | ISI | ChemPort |
  42. Slimani N, Kaaks R, Ferrari P, Casagrande C, Clavel-Chapelon F, Lotze G et al. (2002a). European Prospective Investigation into Cancer and Nutrition (EPIC) calibration study: rationale, design and population characteristics. Public Health Nutr 5, 1125–1145. | Article | PubMed | ISI | ChemPort |
  43. Slimani N, Fahey M, Welch AA, Wirfalt E, Stripp C, Bergstrom E et al. (2002b). Diversity of dietary patterns observed in the European Prospective Investigation into Cancer and Nutrition (EPIC) project. Public Health Nutr 5, 1311–1328. | Article | PubMed | ISI | ChemPort |
  44. Slimani N, Bingham S, Runswick S, Ferrari P, Day NE, Welch AA et al. (2003). Group level validation of protein intakes estimated by 24-h diet recall and dietary questionnaires against 24-h urinary nitrogen in the European Prospective Investigation into Cancer and Nutrition (EPIC) calibration study. Cancer Epidemiol Biomarkers Prev 12, 784–795. | PubMed | ISI |
  45. Slimani N, Deharveng G, Unwin I, Southgate DA, Vignat J, Skeie G et al. (2007). The EPIC nutrient database project (ENDB): a first attempt to standardize nutrient databases across the 10 European countries participating in the EPIC study. Eur J Clin Nutr 61, 1037–1056. | Article | PubMed | ChemPort |
  46. Sommerfeld M (1983). Trans unsaturated fatty acids in natural products and processed foods. Prog Lipid Res 22, 221–233. | Article | PubMed | ChemPort |
  47. Subar AF, Dodd KW, Guenther PM, Kipnis V, Midthune D, McDowell M et al. (2006). The Food Propensity Questionnaire: concept, development, and validation for use as a covariate in a model to estimate usual food intake. J Am Diet Assoc 106, 1556–1563. | Article | PubMed
  48. Swinburn BA, Caterson I, Seidell JC, James WP (2004). Diet, nutrition and the prevention of excess weight gain and obesity. Public Health Nutr 7, 123–146. | Article | PubMed | ISI | ChemPort |
  49. Tooby J, Cosmides L (1990). The past explains the present—emotional adaptations and the structure of ancestral environments. Ethol Sociobiol 11, 375–424. | Article
  50. Tur JA, Romaguera D, Pons A (2004). Food consumption patterns in a Mediterranean region: does the Mediterranean diet still exist? Ann Nutr Metab 48, 193–201. | Article | PubMed | ISI | ChemPort |
  51. Ulijaszek SJ (2007). Obesity: a disorder of convenience. Obes Rev 8(Suppl 1), S183–S187. | Article
  52. van Dam RM, Rimm EB, Willett WC, Stampfer MJ, Hu FB (2002). Dietary patterns and risk for type 2 diabetes mellitus in U. S. men. Ann Intern Med 136, 201–209.
  53. van Dam RM, Seidell JC (2007). Carbohydrate intake and obesity. Eur J Clin Nutr 61(Suppl 1), S75–S99. | Article | PubMed | ChemPort |
  54. van Erp-Baart MA, Couet C, Cuadrado C, Kafatos A, Stanley J, van Poppel G (1998). Trans fatty acids in bakery products from 14 European countries: The TRANSFAIR Study. J Food Compos Anal 11, 161–169. | Article | ChemPort |
  55. WCRF/AICR (2007). Food, Nutrition, Physical Activity, and the Prevention of Cancer: A Global Perspective. American Institute for Cancer Research/World Cancer Research Fund: Washington DC.
  56. WHO/FAO (2003). Diet, Nutrition and the Prevention of Chronic Diseases. Report of a Joint WHO/FAO Expert Consultation WHO Technical Report 916. World Health Organization: Geneva.


The work described in this paper was carried out with the financial support of the European Commission: Public Health and Consumer Protection Directorate 1993–2004; Research Directorate-General 2005; Ligue contre le Cancer (France); Société 3M (France); Mutuelle Générale de l'Education Nationale; Institut National de la Santé et de la Recherche Médicale (INSERM); Institut Gustave Roussy; German Cancer Aid; German Cancer Research Center; German Federal Ministry of Education and Research; Danish Cancer Society; Health Research Fund (FIS) of the Spanish Ministry of Health; Spanish Regional Governments of Andalucía, Asturias, Basque Country, Murcia and Navarra and the Catalan Institute of Oncology; and ISCIII RETIC (RD06/0020), Spain; Cancer Research UK; Medical Research Council, UK; the Stroke Association, UK; British Heart Foundation; Department of Health, UK; Food Standards Agency, UK; the Wellcome Trust, UK; Greek Ministry of Health; Hellenic Health Foundation; Italian Association for Research on Cancer; Italian National Research Council, Regione Sicilia (Sicilian government); Associazione Iblea per la Ricerca Epidemiologica—ONLUS (Hyblean association for epidemiological research, NPO); Dutch Ministry of Health, Welfare and Sport; Dutch Prevention Funds; LK Research Funds; Dutch ZON (Zorg Onderzoek Nederland); World Cancer Research Fund (WCRF); Swedish Cancer Society; Swedish Research Council; Regional Government of Skane and the County Council of Vasterbotten, Sweden; Norwegian Cancer Society; the Norwegian Research Council and the Norwegian Foundation for Health and Rehabilitation. We thank Sarah Somerville, Nicole Suty and Karima Abdedayem for assistance with editing and Kimberley Bouckaert and Heinz Freisling for technical assistance.

Extra navigation