Role of Diet in Colorectal Cancer Incidence

Key Points Question How credible is the evidence behind the association of dietary factors with colorectal cancer (CRC) risk in published meta-analyses of prospective observational studies? Findings This umbrella review of 45 meta-analyses describing 109 associations found convincing evidence for an association between lower CRC risk and higher intakes of dietary fiber, dietary calcium, and yogurt and lower intakes of alcohol and red meat. Meaning This study suggests that dietary factors may have a role in the development and prevention of CRC, but more research is needed on specific foods for which the evidence remains suggestive.

analysis, we noted whether relevant confounders were accounted for in adjusted summary estimates and reported. We communicated with authors to obtain data for evidence synthesis if it was not clearly reported in the published meta-analysis.

Definition of important dietary patterns:
Dietary pattern Study Exposure definition Dietary calcium Meng 2019 (218) The comparison of elemental intake of dietary calcium from each study included in the meta-analysis classified as highest categories (Q3, Q4, and Q5 [up to 2057 mg/day]) and the lowest categories (Q1 and Q2 [<228 mg/day]).

Dietary glycemic load
Reynolds 2019 (182) Based on WHO Nutrition Guidance Expert Advisory Group

Dietary glycemic load
Reynolds 2019 (182) Based on WHO Nutrition Guidance Expert Advisory Group

Light alcohol drinking
Fedirko 2011 (201) Consumption of ≤1 drink/day (≤12.5 g/day of ethanol) Mediterranean diet Schwingshackl 2017 (178) High consumption of plant-based foods, especially whole grain products, vegetables, fruits, nuts, and legumes with regular intake of fish and seafood. Eggs, red and processed meat as well as high-fat dairy products are consumed in low amounts Moderate alcohol drinking Fedirko 2011 (201) Consumption of 2-3 drinks/day (12.6-49.9 g/day of ethanol) Western diet Feng 2017 High consumption of red and/or processed meat, refined grains, sweets, high-fat dairy products, butter, potatoes and high-fat gravy, and low intake of fruits and vegetables Non-vegetarian diet Godos 2016 (181) Eating meat more than once per week

Pesco-vegetarian diet
Godos 2016 (181) Consumption of fish more than once per month in those following vegetarian diet Semi-vegetarian diet Godos 2016 (181) Low consumption of meat (more than once per month but less than once per week)

Supplemental calcium
Heine-Bröring 2015 (216) Use of calcium in supplement form. Mean level of intake:145 mg/day to 1,130 mg/day Unhealthy diet Grosso 2017 (180) High intakes of red and processed meat, sugary drinks and salty snacks, starchy foods, and refined carbohydrates Vegetarian diet Godos 2016 (181) Eating meat less than once per month eTable 1. Class III Low § Not performed due to limited number of primary studies *Sensitivity analysis is not possible because no information on quality assessment of primary studies # Sensitivity analysis is not possible because meta-analysis only included good-quality studies NP = not pertinent, because estimated number is larger than observed, and there is no evidence of excess significance based on assumption made for plausible effect size; O/E = observed/expected number of studies with significant results; OR = odds ratio; RR = risk ratio. eTable 6: Evidence Criteria: Difference Between WCRF and Present Review WCRF (223)  Evidence from more than one study  Evidence from at least two independent cohort studies  No substantial heterogeneity  Good quality studies to exclude with confidence the possibility that the observed association results from random or systematic error, including confounding, measurement error, and selection bias  Dose-response relationship  Strong and plausible experimental evidence Umbrella review  (195) 5 Explanation: Although WCRF is the latest report, the meta-analyses they used for the intake of red meat and total dietary fibre are different from ours. According to the published methodology by WCRF, the search for articles was updated to April 2015. For our current review paper, a systematic literature search up to September 2019 was performed. Hence, the meta-analyses included for the intake of red meat, total dietary fibre, yogurt, and dietary calcium are from recent papers published between 2018 to 2019, except for heavy alcohol intake which was published in 2011. The meta-analyses included in our review are chosen based on specified selection criteria: meta-analysis with the largest number of primary studies and the one with the largest number of colorectal cancer cases. However, the selection criteria in the WCRF report are unclear. We followed exactly the protocol as suggested by recent umbrella reviews for selection of meta-analysis for evidence grading. We excluded the meta-analyses used by WCRF and they are shown in the exclusion references in this supplementary material. For all of the associations, we used summary estimates for high versus low intake instead of dose-response meta-