World Cancer Research Fund Global Network
Expert Report "Food, Nutrition, and the Prevention of Cancer: a Global Perspective"

M Wiseman
World Cancer Research Fund International

Introduction


The World Cancer Research Fund (WCRF) global network is an international alliance of organisations dedicated to the prevention and control of cancer through healthy diets and lifestyles. The network comprises WCRF International, an umbrella association based in the UK, and its member organisations national charities based in different countries. Current members operate in the USA, the UK, the Netherlands, Germany, and Hong Kong.

It is now generally accepted that among other major environmental factors, dietary patterns and nutritional status are important determinants of the risk of many cancers, including many of the cancers now most prevalent throughout the world. However, until relatively recently, this was not the case. While the causal link between smoking and lung cancer has been accepted since the 1950s, it was not until 1982 that the first authoritative statement on the relationship between food and nutrition and the causation and prevention of cancer was made, with publication of an expert report by the US National Academy of Sciences.¹

Serious consideration of the influence of diet on cancer risk had also been stimulated by the seminal paper by Doll and Peto, in which they estimated that approximately 35% of cancers in the USA may be related to dietary exposure.² Although confidence intervals around their estimates were wide between 10% and 70% this report nevertheless highlighted the potential for prevention across a wide range of cancer sites. Subsequent estimates by other workers have come to similar conclusions, and confidence in the estimates is increasing.^3,4 By the early 1990s, an impressive literature on diet, nutrition and cancer had accumulated. This diverse database comprised epidemiological studies, clinical studies, and laboratory data in animals and humans. In all areas, the quality of studies was varied and, in particular, data on diet itself was often unsatisfactory. Despite this diversity of quality and study type, some patterns of association became apparent such as the link between higher consumption of fruit and vegetables and lower risks of a number of cancers.⁵ Nevertheless, considerable heterogeneity in the data remained, and a number of individuals and groups considered it timely for a further review of the topic.

Expert Report


The World Cancer Research Fund/American Institute for Cancer Research (WCRF/AICR) first expert report Food, Nutrition and the Prevention of Cancer: a Global Perspective was published in 1997.⁶ The result of a 5-year process, which included a comprehensive review of the literature by a panel of independent experts from around the world, this report changed the landscape of how people perceive not only the links between diet, nutrition and cancer, but also the potential for prevention through dietary and other associated changes. The report provided in 1 volume a comprehensive analysis of the status of the evidence base up to 1996, incorporating assessments of epidemiological, clinical, and laboratory data.

Six years on, there has been a substantial accumulation of new evidence. Just as in the original evidence base for the 1997 WCRF report, not all studies come to identical conclusions, and their quality is variable. Clearly the new data will impact on the overall body of evidence, which must form the basis of any comprehensive interpretation.

Technical Developments

During the past few years, there has been an explosion of interest in the methods used to synthesise research evidence in the biological sciences. There have been major developments in both qualitative and quantitative aspects, including systematic reviews and meta-analyses. However, the principal focus for this activity has been in relation to healthcare interventions to inform clinical decisions. Relatively little attention has been given to the applicability of these principles to questions of causation and prevention of disease, including cancer.

One consequence of this approach has been that a hierarchy of the relative importance of different types of evidence, which now has considerable credibility in the context of healthcare interventions, has become regarded as universally applicable. However, for many questions, particularly those related to the factors involved in the development of chronic disease, this hierarchy has problems and may well be inapplicable.⁷

The main single issue with current concepts of hierarchy of types of research evidence is special reliance on randomised controlled trials (RCTs). These are often the best way of testing whether a particular intervention impacts on the progress of a disease, and therefore are conducted in relatively small numbers on groups of people at relatively high risk of developing clinical disease or some other identifiable pathology or 'end-point'. In contrast, studies concerned with the causation and, therefore the preventability, of disease are typically performed for decades in large numbers at low risk. The practical problems of conducting RCTs in such circumstances are insurmountable, so the vast majority of studies addressing such questions are observational epidemiology. In addition, the few direct intervention studies and the voluminous laboratory data often offer only indirect insights into the questions of relevance, or focus on a small component of the causal pathway. Negative results from such studies are difficult to interpret, but are often misinterpreted in too broad a context. On the other hand, positive results, although unusual, can, if replicated, provide compelling evidence. Consequently negative results from intervention trials should not be interpreted as cancelling positive results from other types of study.

The outcome is a misperception by health professionals and policy makers that the studies necessary to help develop rational public health strategies, designed to prevent and control disease, are not of adequate quality to underpin decisions. However, if prevention is to be part of the approach to managing cancer in the community, as it must, policy decisions need to made on the best evidence available. What is the best evidence in relation to the prevention and control of chronic diseases, especially those that typically develop for decades before becoming clinically evident?

The perception that RCTs offer a universally better approach bears examination. Well-conducted RCTs are the best way of avoiding various biases. However, experimental constraints may mitigate this advantage. The intervention applied may not be directly related to the question of relevance dietary fibre given as a synthetic medication the application of the intervention may be unrepresentative of usual practice high-dose
-carotene supplements given instead of -carotene given at lower doses representative of that in food and the population studied differing from the eventual target group a population selected for having colonic polyps.

In some healthcare situations, it has been possible to examine whether well conducted observational studies produce more biased results than randomised clinical trials. Concato et al compared the results of randomised intervention studies with observational evidence relating to a number of clinical situations in a systematic way.⁸ These authors concluded that, at least for the situations they studied, observational data from well conducted studies did not appear to produce biased results compared with randomised interventions. Possible reasons for this disparity from a long-standing perception include better control for confounding factors and definition of inclusion and exclusion criteria. While it may be difficult to extrapolate widely from the few situations Concato et al studied,⁸ in situations where randomised interventions can only bear indirectly on the aetiological question of relevance, it appears reasonable to use the existing data to draw conclusions.

The Next Steps


Given that observational data will provide most of the basis for clinical, public health, and other policy decisions, it is essential that these data are analysed in the most rigorous way. However carefully based on science, there is no such thing as an unchallengeable judgement, but it is possible to minimise bias in a number of ways.

Firstly it is highly desirable to separate the processes of collating and displaying the data from that of interpreting the evidence, and then drawing conclusions and making recommendations. Secondly, evidence should be collated in a systematic way. Thirdly, the process ought to be transparent, meaning that it should be open to external scrutiny and that any necessary judgements are overt. Finally, the evidence should be displayed in a uniformly detailed format.

The 1997 WCRF report gained a reputation for being remarkably objective in these and other respects, and its methods of assessing the evidence have been adapted by the World Health Organization (WHO).⁹ At the same time, WCRF International has always been committed to updating the evidence on which all its programmes are based. WCRF International has now begun the process of producing a new report, with publication projected for 2006.

As the first formal stage in this process, WCRF International convened a Methodology Task Force to address the question of how best to assimilate the evidence relating to the causation and prevention of cancer. We expect this to become a widely accepted convention when reviewing such data, and to impact on the perception of clinicians and policy makers with regard to the evidence for a role of diet in the development and prevention of cancer. As the next stage in the formal process, a new expert panel has been assembled, whose first full meeting will be in May 2003. The panel includes experts in food, nutrition, physical activity, and the translation of science into public policy from all parts of the world, as well as observers from the WHO, the Food and Agriculture Organization, the United Nations International Children's Emergency Fund, the International Union of Nutritional Sciences, the International Food Policy Research Institute, and the International Union Against Cancer. In addition, an advisory group whose members include selected members of the panel and the Methodology Task Force will guide the development of strategy and policy both for assessment of the science and for the new report itself. WCRF International believes that these and other agreed approaches should provide the most robust basis yet devised for developing dietary and policy recommendations on food nutrition, physical activity, and the prevention of cancer.


References

1. National Academy of Sciences. Diet, nutrition and cancer. Washington: National Academy Press; 1982.
   
   
2. Doll R, Peto R. The causes of cancer: quantitative estimates of avoidable risks of cancer in the United States today. J Natl Cancer Inst 1981;66:1191-1308.
   
   
3. Riboli E. Background and rationale of EPIC. Ann Oncol 1992;3:783-791.
   
   
4. Willett W. Diet, nutrition and avoidable cancer. Environ Health Perspect 1995;103 (Suppl 8):165-170.
   
   
5. Block G, Patterson B, Subar A. Fruit, vegetables, and cancer prevention: a review of the epidemiological evidence. Nutr Cancer 1992;18:1-29.
   
   
6. World Cancer Research Fund/American Institute for Cancer Research. Food, nutrition and the prevention of cancer: a global perspective. London, Washington: World Cancer Research Fund/American Institute for Cancer Research; 1997.
   
   
7. Harbour R, Miller J. A new system for grading recommendations in evidence based guidelines. Br Med J 2001;323:334-336.
   
   
8. Concato J, Shah N, Horwitz RI. Randomized, controlled trials, observational studies, and the hierarchy of research designs. N Engl J Med 2000;342:1887-1892.
   
   
9. World Health Organization. Draft report of the WHO/FAO expert consultation on diet, nutrition and the prevention of chronic diseases. Geneva: World Health Organization; 2002.
   
   

   For further information about World Cancer Research Fund in Hong Kong (WCRF HK), please contact: Ms Karen Sadler Unit 25B One Capital Place 18 Luard Road Wanchai, Hong Kong Tel: Fax: E-mail: