Diet and cancer risk – tackling a meaty research topic
Judging by the newspaper headlines, you might think that there is a never-ending shopping list of foods that cause or protect against cancer (or, in some cases, manage to do both at the same time). Finding out how certain foods are linked to cancer risk is a complicated question, which Professor Tim Key has dedicated his career to answering.
When he first came to the University of Oxford’s Cancer Epidemiology Unit (CEU), now part of the Nuffield Department of Population Health (NDPH), as a PhD student in the 1980s, Key had no idea he would end up leading a research study of epic proportions, investigating the impact of diet, hormones and lifestyle on cancer risk.
Today he’s a professor of epidemiology, deputy director of the CEU, and the principal investigator of EPIC-Oxford – a study involving around 65,000 people in the UK that is aiming to unpick the complex relationships between lifestyle and health.
Science on an EPIC scale
The Oxford group is part of a much larger project known as EPIC (the European Prospective Investigation into Cancer and Nutrition) which includes more than half a million participants recruited from across ten countries in Europe who have been followed up for 20 years.
Launched in 1993, EPIC was the biggest study of diet and health that had ever been undertaken at the time, and it remains one of the largest prospective cohort studies in the world.
Participants were asked to fill out detailed questionnaires about their diet and lifestyle, which are being analysed alongside medical records, physical measurements and more than 9 million biological samples such as blood cells and plasma. They will be followed up in the future, ultimately until death, to see if there are any connections between their diet and health.
Studying diet and disease risks is hard because, unlike a specific behaviour like smoking, everyone has to eat. People’s diets can vary widely, and there may be multiple confounding factors such as genetics, geographical location and culture, and social inequalities.
Ranking the risks
Alcohol intake and obesity come through as significant diet-related risk factors for several types of cancer. After that, the study shows that bowel cancer risk is higher for people who eat more processed meat or less fibre, and there is some evidence that sticking to a traditional ‘Mediterranean diet’ is associated with a reduced risk of heart disease and longer survival for elderly people. After that, the evidence gets less clear.
‘Initially we were expecting to find big changes in cancer risk associated with specific foods, but actually that’s not usually the case,’ explains Key. ‘We don’t have definitive answers for things like fruit and vegetable intake, sugar, dairy products, vitamins and minerals.’
Professor Key points to inconclusive, negative or even potentially harmful results from clinical trials of various vitamins and minerals designed to prevent cancer. For example, studies investigating whether supplements of the vitamin A precursor beta-carotene could reduce the risk of lung cancer in smokers actually had the opposite effect, while multiple trials testing whether the mineral selenium can reduce prostate cancer risk have failed to show a significant benefit.
‘Studies like this tell us that there is probably a very wide range of “OK” levels for most of these substances,’ explains Professor Key. ‘Very low levels are bad and very high levels can be bad, but there’s a wide band in the middle where you’re fine.’
Faced with an ever-expanding multi-billion dollar supplement industry and celebrities promoting ‘wellness’ diets packed with all kinds of superfoods, large scale studies and analyses like those from Key and his NDPH colleagues are vital if we are to develop evidence-based recommendations on food and supplements that genuinely help to improve the health of the population.
The meat of the matter
Professor Key also has an interest in finding out whether following a vegetarian or vegan diet has benefits or detriments for health. EPIC-Oxford has a particular advantage here, as it recruited many participants who had been involved in a previous study known as the Oxford Vegetarian Study, which began in 1980.
Out of the 65,000 people in the Oxford cohort, half don’t eat meat. Some are vegetarian while others are pescatarian, eating fish and seafood but not other types of meat, and a smaller number are vegan. However, although Key has found that the risk of heart disease is reduced by around a third in vegetarians, this doesn’t seem to translate into longer lives for those who choose not to eat meat – a curious observation that needs further research.
Professor Key is also interested in the relationship between hormones and cancer risk, including sex hormones such as oestrogen, and insulin-like growth factors (IGFs). He’s been collaborating with researchers working on the Million Women Study, which has primarily focused on the relationship between hormones and breast cancer risk, to look at dietary information gathered from the participants. He is also expanding his scope of work to include other important health conditions, including cardiovascular disease, stroke, arthritis, bone fractures and diverticular (bowel) disease.
‘The most important things we know that definitely reduce cancer risk may sound simple – firstly, don’t smoke, then be a healthy weight and don’t drink too much alcohol,’ Professor Key points out. ‘Everyone wants the magic things from the newspaper headlines to be true – “eat turmeric and you won’t get cancer!” – but right now we don’t have the evidence.’
Key and his team are continuing to systematically examine the data from EPIC and other large cohort studies in search of relationships between disease, diet and other factors.
Professor Key says: ‘Although a lot of the previously published results are negative or inconclusive, especially around the role of specific foods in cancer risk, it doesn’t mean we’re not going to find some more solid results in the next few years, because having bigger studies and datasets with high quality measures, backed up with biomarkers, gives you a better chance to frame a good question and find the answers.’