The Paper of the Month for April is 'Mechanistic evidence underpinning dietary policy: bringing the jigsaw pieces together?'. The blog is written by author Christine M Williams. The paper is published in Proceedings of the Nutrition Society and is free to access for one month.

Most of the evidence linking diet with complex diseases such as heart disease and cancer (non-communicable diseases (NCD)) is based on findings from epidemiological cohort studies which follow large populations of people to determine whether groups of people who have been diagnosed with or died from the disease, have diets that differ from those who are free from the disease. A limiting factor in these studies is their observational nature which limits the certainty about causal relationships between the exposure (diet) and the outcome (NCDs). Randomised controlled trials (RCTs), where a dietary factor is fed under carefully controlled circumstances, have a greater ability to prove the dietary factor is causing the disease and provide confidence for policymaking. However, due to the costs and complexity of dietary trials, such data is very rarely available. Nevertheless, advances in approaches to selecting suitable cohort studies (systematic reviews) and combining and visually representing analysis of data from cohort studies (Meta-analysis and Forest plots), have allowed much deeper analyses of their findings. These approaches allow assessment of causative characteristics in the combined cohort data, such as consistency of effect, large effect size, specificity (free from confounding) and dose-response. These types of analyses are now widely used in policymaking.

The PNS paper published in 2022 (1) ‘Mechanistic evidence underpinning dietary policy: bringing the jigsaw pieces together’, discussed the benefit to policymaking of greater use of mechanistic findings, since plausible biological mechanisms are a key causal characteristic that cannot be derived from cohort or RCT studies. Although large volumes of mechanistic studies (cell, animal, human) are published in the literature, unlike the case for epidemiological studies, systematic approaches to quality-selecting and combining and visualising outcomes from different types of mechanistic studies, have not yet been fully developed.  This limits the ability to discriminate mechanistic studies that can help address current diet and health policy questions from those that cannot. The PNS paper describes some of the research from a Bristol group (2) currently developing novel approaches to combining data from mechanistic studies concerning diet and cancer. In the case illustrated in the PNS paper the dietary exposure under scrutiny was milk or dairy, the chosen intermediate was IGF1 and the outcome was prostate cancer.  The most important outcome from this study was the development of a novel approach for determining the effect size of milk or dairy and IGF1, and for visualising the strength and direction of the relationship between milk and dairy foods and circulating concentrations of IGF1 (an identified intermediate for prostate cancer). Their findings lend support to the possibility of using similar approaches to mechanistic studies as that which is currently used for cohort studies and NCDs.

Bibliography

(1). Williams CM (2022).  doi: 10.1017/S0029665122002750

(2) Lewis et al (2017). doi.org/10.1158/1055-9965.EPI-17-0232