Author:

Camilla Fiz

The European Prospective Investigation into Cancer and Nutrition (EPIC-Europe) is a long-term, large-scale collaborative project that studies different populations from countries across Europe to investigate the relationships between diet, nutrition, lifestyle, and environmental factors, and the incidence of cancer. It is one of the largest cohort studies in the world, with more than half a million participants recruited across 23 clinical centres in 10 countries in western Europe.

Much of the public health information that is now taken for granted, such as the protective role of fruit and vegetables and the Mediterranean diet, or the risks of body and abdominal fat, and consumption of red and processed meats, comes from epidemiological evidence revealed by the EPIC study.

That evidence helped to radically reshape the understanding of what causes cancer and to empower people to take action to lower their own cancer risk. It also led to the discovery of new approaches to therapy and prevention.

The question that launched EPIC

The idea behind the EPIC study originated in the 1980s, within the International Agency for Cancer Research (IARC), an interdisciplinary agency of the World Health Organization (WHO), with a remit to identify the causes of cancer with a view to promoting prevention. At that time, IARC’s research focused almost entirely on exogenous causes – exposure to industrial or environmental toxins. But there was a growing awareness of the extent to which the incidence of different cancer types varied from country to country, in a way that could not be explained. Emerging evidence was showing that people who migrated from one country to another would end up with the same cancer risk profile as their adopted country within one or two generations.

That finding prompted a group of epidemiologists at IARC to ask questions about how changes in lifestyles, such as diet, might affect a person’s cancer risk. Prominent among them was Elio Riboli, the current EPIC coordinator and professor at Imperial College London. In those days, recalled Riboli, neither the public nor the scientific community knew much about the causes of these phenomena. “At that time, everything was focused on chemicals,” – particularly at IARC.

This was not the first time that diet and lifestyle had been raised as potential risk factors for cancer, Riboli added. It had long been clear that the chemical factors studied at IARC could not alone explain the emergence of every type of cancer. The issue had cropped up periodically in the literature since the early 1900s. The discovery that cancer incidence drops in mice on restriction of their caloric intake was a seminal finding published in 1940 by Albert Tannenbaum, a cancer researcher at the Medical Research Institute, Michael Reese Hospital, in Chicago, Illinois. This was followed by a further paper in 1953, written together with his colleague Herbert Silvertone.

It wasn’t until the 1980s, however, that some retrospective studies began being published that compared cancer incidence around the world in different countries, “suggesting that lifestyle and environmental factors may explain very large variations,” said Riboli.

IARC decided to go further in the study of cancer risk in relation to diet and lifestyle. They started by setting up the Nutrition and Cancer Programme, and then in 1992 they established the EPIC study. They wanted to use a more reliable approach than the retrospective studies, even if it was longer and more complex to carry out. Before starting, they needed to develop new research approaches and practices, and create a well-coordinated international community of scientists in Europe.

How EPIC was set up

In Germany it was the Division of Cancer Epidemiology, at Deutsches Krebsforschungszentrum in Heidelberg that acted as the national collaborating centre, with Rudolf Kaaks appointed as principal investigator. “We designed EPIC as a prospective study because we wanted to avoid the biases of retrospective studies,” Kaaks recalled. The design of the retrospective studies was based on selecting people with cancer (patients) and without (controls), and asking them whether they had been exposed to various risk factors in their past, with the view to exploring associations between reported exposures and cancer incidence. “Every stage of this kind of analysis is at risk of biases,” Kaaks explained. “There can be an improper selection of control people as well as many errors in participants’ memories.”

Prospective studies, by contrast, start by recruiting healthy people, and asking them about diet and lifestyle factors, and then doing the follow-up to find out who among them gets cancer and who does not. This was quite a new way of doing research, according to Carlotta Sacerdote, the current principal investigator of the Italian EPIC study, and professor at the University of Eastern Piedmont. “In the 1990s, I was a young medical student and I had never heard about prospective studies,” she recalled. It wasn’t until she heard about the EPIC project, when she was doing her internship at Torino Medical School, that she came across the concept. “I immediately understood how innovative and visionary this programme was, and how much it would affect whole populations.”

Conducting a prospective study on this scale was a real challenge. It meant monitoring a large population for a long period – 10 years at least. To throw light on the causes of dozens of cancer types meant collecting huge quantities of data on dietary, lifestyle, and socioeconomic factors from a sample of great heterogeneity. But in Europe, IARC understood it had the perfect natural laboratory for this kind of study.

On one hand, cancer incidence and mortality varied greatly between European countries – breast cancer incidence was the highest in the Netherlands, Belgium, Sweden, and Denmark, while it was markedly lower in Greece, Portugal, and Italy. On the other side, dietary and lifestyle habits also varied widely between countries, in terms of meat consumption, types of food and recipes, physical activity, smoking levels, and alcohol consumption. After a pilot prospective study conducted in Malmö in Sweden in the late 1980s, IARC researchers started recruiting the first research centres across Europe. This in itself was something of a novelty, as researchers from different countries were not used to working together.

The dawn of European collaboration

“The European Union has no coordinated research approach, and most of its national programmes are small and fragmented,” the journalist Sally Goodman wrote in Nature in 2002. “In those years, this was much true,” commented Riboli, “The EPIC study was one of the first attempts to do a large-scale European project in epidemiology, at least, federating some of the countries and centres.”

The project took shape mainly by word of mouth, recalled Kaaks. “Around the 1987, Elio Riboli with the help of Rodolfo Saracci – who then became the head of the IARC research unit of Analytical Epidemiology – started keeping in touch with some epidemiologists in Italy, Spain, and France. In parallel some contacts were kept with Sweden and the Netherlands and so on. Since I am originally from the Netherlands, we also got in touch with an institute in Utrecht, where I worked for a while.” Little by little, the EPIC project expanded to include 10 countries: Greece, Italy, Spain, France, Germany, the Netherlands, United Kingdom, Denmark, Sweden, and Norway. The Scandinavian cohort studies started independently, and joined the EPIC programme later.

The central base for the study was at the IARC offices in Lyon, France; Riboli was responsible for coordinating the collaborating centres. A large community of epidemiologists contributed to discussions on which techniques and tools to use, and how to carry out analyses using standardised metrics. Nick Day, the first principal investigator of the EPIC-Norfolk study, in east England, and the nutritionist Sheila Bingham, then based at the UK Medical Research Council in Cambridge, led on developing the feasibility of using a complex dietary assessment method in a prospective study.

“We did a lot of work on planning the study questionnaires and developing methods for recruiting the participants and collecting personal data, anthropometric measurements, and blood samples,” recalled Riboli. “It is worth noting that we developed a biorepository for blood sample collection, called IARC biobank. It was the largest one ever created at that time.”

The tools: questionnaires and blood samples

Between 1994 and 2001, more than 521,000 participants were recruited across 23 European centres. At the baseline, these were middle-aged people (aged 35–69 years) from the general population – voluntary workers, blood donors, staff at local school or companies. They were given dietary questionnaires, mainly focused on the frequency of food intake. Typical questions were: ‘What do you eat for breakfast? What do you eat for lunch?’ They used pictures to help people report, in a standardised fashion, the usual sizes of their food portions.

While the structure and the basic questions were the same, the type of food, the pictures, and the language varied from country to country, said Kaaks. “The diet is very different in Tromsø, in the north of Norway, from Greek islands in the Mediterranean. As each country has its own type of food, the questionnaires were also country specific.”

With the baseline data gathered, all these heterogeneous data from the questionnaire responses were compared across regions and countries. A second questionnaire, the ‘24-hour dietary recall’, was developed and given to a representative subset, comprising 8% of each cohort. This time the questions were: ‘What exactly did you eat in the last 24 hours? What was the size of your portions?’. “Unlike the first questionnaire, the ‘24-hour dietary recall’ made it easier to standardise and quantify, because it involves another kind of memory process, and releases a different kind of information,” Kaaks explained. The two questionnaires were then combined and harmonised. To facilitate comparisons between countries, each food item was broken down into its constituent ingredients, until a standardised list of more than 11,000 food items was created.

People were also asked for information about their lives, such as how much physical activity they do, whether they smoke, their occupational, reproductive, and clinical history, as well as some anthropometric data, such as weight, BMI, and the hip-to-waist ratio. “Moreover, they had the great idea of taking blood samples to measure biomarkers,” Sacerdote recalled. Blood samples were an objective and complementary tool for determining nutrient intake, but they could also be used for detecting hormones, genetic and epigenetic signatures, and more. “In a prospective study, if you identify a biomarker of cancer years before its onset, you can be more confident that it is involved in tumoral development, rather than being caused by the tumour,” she explained. In all, 3,800,000 blood samples are now stored in the IARC biobank.

Results in and beyond cancer

After completing the questionnaires and collecting data, the participants were followed up regularly monitored for almost 20 years. By 2016, almost 96,600 of the EPIC study population had been diagnosed with cancer – two-thirds of them men. By 2023, more than 3,000 scientific articles had been published analysing how these cancers correlated with the data on diet, lifestyle and blood samples. With such a treasure trove of valuable data, the studies soon widened out to explore risk factors for other diseases.

The spectrum of research gradually broadened to include studies of diabetes and cardiovascular, autoimmune, and neurological diseases. The NeuroEPIC4PD study, for instance, explored the relation between dietary factors and Parkinson’s disease, while the EPIC Europa InterACT study and the EPIC Heart study explored the correlation between socioeconomic factors and respectively diabetes and cardiovascular disease. “Validating all the patients’ diagnoses, in the EPIC Heart study, found significant correlations between diet, diabetes and cardiovascular diseases,” said Sacerdote. In 2011, this project demonstrated that people who eat more fruit and vegetables have a reduced risk of ischaemic heart disease.

The EPIC legacy

EPIC is a landmark study in epidemiological research. “Probably, it has been one of the most important studies in the world on this topic,” Riboli commented. “EPIC was also the beginning of an international movement aimed at developing large prospective studies.” Prominent examples that used the EPIC study as a reference, and took advantage of advancing molecular biology and informational technologies, include the UK Biobank study and the Multi-Ethnic Cohort in the United States.

The EPIC study itself still has more epidemiological evidence to reveal. A promising new line of research is now focusing on ultra-processed foods and their association with obesity and cancer. Thanks to the longitudinal nature of the study, it may also help throw light on important evidence regarding the role of ageing in cancer risk. In the Italian EPIC-Turin centre, for instance, Sacerdote and colleagues began work on recalling people who were surveyed and had their blood samples for the original EPIC cohort. “Our aim is to recall 3,000–3,500 participants from the EPIC-Turin study, at least. We want to give them the same questionnaire as we did 30 years ago and take new blood samples to monitor the effects of ageing in cancer risk.”

References

Ancel Keys. (1980). Seven countries: a multivariate analysis of death and coronary heart disease. Harvard University Press. Cambridge: Massachusetts.

Cordova R. et al., (2021). Consumption of ultra-processed foods associated with weight gain and obesity in adults: A multi-national cohort study. Clinical Nutrition, 40, 5079-5088.

Crowe F. L. et al., (2011). Fruit and vegetable intake and mortality from ischaemic heart disease: results from the European Prospective Investigation into Cancer and Nutrition (EPIC)-Heart study. European Heart Journal, 32, 1235-1243.

Gonzalez C. A., Riboli E. (2010). Diet and cancer prevention: Contributions from the European Prospective Investigation into Cancer and Nutrition (EPIC) study. European Journal of Cancer, 46, 2555-2562.

Goodman S. (2002). EU ponders joint action on cancer. Nature, 419, 101.

Morales-Berstein, F., Biessy, C., Viallon, V. et al. (2024). Ultra-processed foods, adiposity and risk of head and neck cancer and oesophageal adenocarcinoma in the European Prospective Investigation into Cancer and Nutrition study: a mediation analysis. European Journal of Nutrition, 63, 377–396.

Riboli E., Kaaks R. (1997). The EPIC Project: rationale and study design. European Prospective Investigation into Cancer and Nutrition, 26, S6-S14.

Saracci R., Wild P. Christopher. (2015). International Agency for Research on Cancer: The first 50 years, 1965-2015 (pp. 160-179). International Agency for Research on Cancer. Lyon: France.

Tannenbaum A. (1940). The initiation and growth of tumors. The American Journal of Cancer, 38, 335-350.

Tannenbaum A., Silverstone H. (1953). Nutrition in Relation to Cancer. Advances in Cancer Research, 1, 451-501.