Nutritional epidemiology and cancer: A Tale of Two Cities

Abstract

Recently, Ioannidis questioned whether nutritional epidemiology could be reconciled with good scientific principles, and suggested that the field needs radical reform. One of the reforms he advocated was more randomized trials; though what diet would be tested and how it would be measured were unspecified, how adherence would be monitored was unclear or unimportant, and the length of time vaguely stated as “lifelong”. The other reform was reanalysis of shared data, which actually already exists in a large number of cohort consortia of individual pooled data. The 2018 report analysis of diet and cancer from the World Cancer Fund/American Institute of Cancer Research presents a sharply different picture of our knowledge of nutrition and cancer, which has evolved immensely in the past 3 decades. Based on current knowledge, factors related to energy balance, encompassed by body mass anthropometric measures might account for about 10–15% of the U.S. cancer burden. This 10–15% encompasses physical activity and obesogenic effects of diet. About 5% may be attributable to alcohol, and another 5% to specific dietary factors combined (e.g., red and processed meat, whole grains, fiber, calcium, fruits, and vegetables). Surrogates such as attained height and age at menarche are influenced by nutrition and are consistent risk factors for cancer, supporting the importance of early nutrition. Recent data suggest that early life dietary patterns, which may be modifiable, may be important. Though important questions remain, our understanding of nutrition and cancer over the life course is coherent and has been informed by nutritional epidemiology and other data.