Table 1.

Evidence synthesis of the association between diet and colorectal cancer in the EPIC study.

Wholegrains, Vegetables, and Fruit
No. of Cases (Non-Cases)	Mean Follow-Up (Years)	Results, Relative Risk (95% Confidence Interval (CI))	Reference
1329 (476,711)	5.1	The data showed no association between higher intake of nuts and seeds and risk of colorectal, colon, and rectal cancers in men and women combined, but a significant inverse association was observed in subgroup analyses for colon cancer in women at the highest (>6.2 g/d) vs. the lowest: HR 0.69 (0.50–0.95) category of intake, and for the linear effect of log-transformed intake: HR 0.89 (0.80–0.98), with no associations in men.	Jenab 2004 [26]
1721 (518,257)	6.2	The association between fiber and colorectal cancer was significant: HR 0.79 (0.63–0.99)	Bingham 2005 [24]
2819 (449,936)	8.8	Combined consumption of fruit and vegetables was inversely associated with colorectal cancer risk (highest vs. lowest EPIC-wide quintile of consumption): HR 0.86 (0.75–1.00), P trend = 0.04. No association between fruit or vegetable consumption was observed.	van Duijnhoven 2009 [21]
4517 (472,795)	11.0	Total dietary fibre: HR 0.87 (0.79–0.96) per 10 g/day increase in fibre.	Murphy 2012 [25]
3370 (518,078)	13.0	A lower risk of colon cancer was observed with higher self-reported consumption of fruit and vegetables combined (highest vs. lowest quartile): HR 0.87 (0.75–1.01) P trend = 0.02, but no consistent association was observed for separate consumption of fruits and vegetables. Variety in consumption of fruits and vegetables was not associated with a lower risk of colon or rectal cancer.	Leenders 2015 [22]
		Meat, Fish, Dairy Products, and Preservation/Processing of Foods
No. of Cases (Non-Cases)	Mean Follow-Up (YEARS)	Results, Relative Risk (95% Confidence Interval (CI))	Reference
1329 (476,711)	4.8	Intake of red and processed meat: HR 1.35 (0.96–1.88) highest vs. lowest intake; P trend = 0.03 Red and processed meat calibrated: HR 1.55 (1.19–2.02); P trend = 0.001 per 100-g increase. Intake of fish (>80 g/day vs. <10 g/day): HR 0.69 (0.54–0.88); P trend < 0.001. Fish calibrated: HR 0.46 (0.27–0.77) P trend = 0.003 per 100-g increase.	Norat 2005 [28]
1248 (518,752)	3.9	Greater dietary intake of calcium was associated with a lower colorectal cancer risk: IRR 0.69 (0.50–0.96).	Jenab 2010 [30]
861 (25,639)	NR	There was an interaction between red and processed meat intake and MGMT Ile143Val polymorphism on colorectal cancer risk (P interaction = 0.04): For individuals who carried the variant genotype with higher red and processed meat intake (above median) risk was increased: OR 1.43 (0.82–2.48), compared with those with the common genotype and lower red and processed meat intake (below median). Amongst the common genotype group with higher red and processed meat intake suggested an inverse association: OR 0.75 (0.55–1.01).	Loh 2010 [53]
2050 (23,490)	11.0	Dietary calcium intake was inversely but not statistically significantly associated with colorectal cancer mortality: HR for per 100 mg increase in intake 0.95 (0.88–1.02).	Li 2011 [54]
289 (44,952)	12.0	Yogurt intake: HR 0.65 (0.48–0.89) highest vs. lowest tertile.	Pala 2011 [32]
4513 (472,609)	11.0	Total milk consumption: HR 0.93 (0.89–0.98) per 200 g/day Whole-fat milk: HR 0.90 (0.82–0.99) per 200 g/day Skimmed milk: HR 0.90 (0.79–1.02) per 200 g/day Dietary calcium: HR 0.95 (0.91–0.99) per 200 mg/day; no association observed for non-dairy calcium sources (HR 1.00, 0.81–1.24) per 200 mg/day.	Murphy 2013 [31]
421 (46,297)	11.0	Consumption of yogurt (highest vs. lowest tertile): RR 0.65 (0.48–0.89) P trend = 0.002	Sieri 2015 [33]
3789 (516,189)	4.1	Pre-diagnostic red meat, processed meat or fibre intakes (defined as quartiles and continuous grams per day) were not associated with colorectal cancer mortality among colorectal cancer survivors; however, a marginal trend across quartiles of processed meat was detected (P = 0.053).	Ward 2016 [29]
6291 (460,869)	14.9	Inversely associated with colorectal cancer incidence (highest vs. lowest quintile): Total fish: HR 0.88 (0.80–0.96) P trend = 0.005 Fatty fish: HR 0.90 (0.82–0.98) P trend = 0.009 Lean fish: HR 0.91 (0.83–1.00) P trend = 0.016 Total n-3 LC-PUFA: HR 0.86 (0.78–0.95) P trend = 0.010 Associated with increased risk of colorectal cancer (highest vs. lowest quintile): Dietary ratio of n-6:n-3 LC-PUFA: HR 1.31 (1.18–1.45) P trend < 0.001	Aglago 2020 [27]
1069 (469,869)	6.4	Subjects with higher concentrations of red blood cell stearic acid were at higher risk for colorectal cancer (per 1 mol%): OR 1.23 (1.07–1.42). Conversely, colorectal cancer incidence decreased with increasing proportions of red blood cell n-3 PUFA, particularly eicosapentaenoic acid (per 1 mol%): OR 0.75 (0.62–0.92).	Linseisen 2021 [55]
		Dietary Patterns
No. of Cases (Non-Cases)	Mean Follow-Up (Years)	Results, Relative Risk (95% Confidence Interval (CI))	Reference
172 (99,828)	6.3	The meat-eaters pattern (meat, poultry, and margarine) was positively associated with colorectal cancer risk (highest vs. lowest quintile): RR 1.58 (0.98–2.53) P trend = 0.02	Kesse 2006 [56]
290 (63,260)	NR	For colorectal cancer in vegetarians compared with meat eaters: IRR 1.39 (1.01–1.91). Comparing vegetarians with nonvegetarians, the risk of colorectal cancer was significantly higher among vegetarians: IRR 1.49 (1.09–2.03).	Key 2009 [57]
435 (44,840)	11.28	The Italian Mediterranean Index was inversely associated with colorectal cancer risk (highest vs. lowest category): HR 0.50 (0.35–0.71) P trend = 0.043. Highest Italian Mediterranean Index score was also significantly associated with reduced risks of any colon cancer: HR 0.54 (0.36–0.81), distal colon cancer: HR 0.44 (0.26–0.75), and rectal cancer: HR 0.41 (0.20–0.81), but not of proximal colon cancer.	Agnoli 2013 [34]
4355 (516,975)	11.6	A decreased risk of colorectal cancer was estimated when comparing the highest (scores 6–9) with the lowest (scores 0–3) adherence to the Centre-Specific Modified Mediterranean diet score and the Modified Mediterranean diet score: HR 0.92 (0.84–1.00) and HR 0.89 (0.80–0.99), respectively. A 2-unit increment in either Mediterranean scale was associated with a borderline statistically significant reduction in colorectal cancer risk (for the Modified Mediterranean diet score): HR 0.96 (0.92–1.00)	Bamia 2013 [35]
421 (46,297)	11.0	Adherence to Mediterranean diet (highest vs. lowest quartile): RR 0.50 (0.35–0.71) P trend = 0.043	Sieri 2015 [33]
5806 (421,701)	15.3	A higher Food Standards Agency nutrient profiling system dietary index (FSAm-NPS DI) score (lower nutritional quality diet) was associated with a higher risk of colorectal cancer (highest fifth vs. lowest quintile): HR 1.11 (1.01–1.22) P trend = 0.02	Deschasaux 2018 [36]
5991(470,169)	14.0	More proinflammatory diets were related to a higher colorectal cancer risk, particularly for colon cancer: Inflammatory Score of the Diet quartile (highest vs. lowest quintile): HR 1.15 (1.04–1.27) for colorectal cancer risk, HR 1.24 (1.09–1.41) for colon cancer, and HR 0.99 (0.83–1.17) for rectal cancer. Associations were more pronounced in men and not significant in women. The inflammatory profile score of the diet was associated with colorectal cancer risk, particularly colon cancer among men (highest vs. lowest quintile): HR 1.62 (1.31–2.01) for colon cancer overall, and HR 2.11 (1.50–2.97) for colon cancer in men.	Jakszyn 2020 [37]
		Alcoholic and Non-Alcoholic Drinks
No. of Cases (Non-Cases)	Mean Follow-Up (Years)	Results, Relative Risk (95% Confidence Interval (CI))	Reference
1833 (476,899)	6.2	Lifetime alcohol intake was significantly positively associated to colorectal cancer risk (for 15 g/day increase): HR 1.08 (1.04–1.12). Baseline alcohol was significantly positively associated to colorectal cancer risk (for 15 g/day increase): HR 1.09 (1.05–1.13). The colorectal cancer risk for beer (HR 1.38, 1.08–1.77) was higher than wine (HR 1.21, 1.02–1.44).	Ferrari 2007 [42]
407 (23,837)	11.0	Total alcohol consumption: HR 0.70 (0.44–1.13) for alcohol consumption of ≥21 units/week compared with non-drinkers; P trend = 0.14 (not associated with colorectal cancer). Daily consumption of ≥1 unit of wine: HR 0.61 (0.40–0.94) P trend = 0.04	Park 2009 [43]
1367 (NR)	3.6	Among individuals drinking <30 g alcohol/day (highest vs. lowest quintile of folate status): For males: RR 0.79 (0.52–1.23) P trend = 0.19) For females: RR 0.96 (0.67–1.37) P trend = 0.73) Among those drinking >30 g alcohol/day (highest vs. lowest quintile of folate status): For males: RR 0.91 (0.47–1.75) P trend = 0.87 For females: RR 2.59 (0.53–1.75) P trend = 0.47	Eussen, Vollset, Igland 2010 [41]
3759 (343,478)	12.0	Alcohol consumption: HR 0.87 (0.81–0.94)	Aleksandrova 2014 [40]
NR (521,330)	16.4	Total soft drink consumption was positively associated with colorectal cancer deaths (≥1 glass per day vs. <1 glass per month): HR 1.25 (1.07–1.47) P = 0.0047, with statistically non-significant associations found for sugar-sweetened and artificially sweetened soft drinks.	Mullee 2019 [44]
6291 (515,039)	14.9	Greater alcohol consumption was associated with an increased risk of colorectal cancer (per 15-g/day increment): HR 1.05 (1.03–1.07).	Murphy 2019 [39]
154 (45,339)	14.0	An increase in rectal cancer risk among subjects drinking more than 3 drinks/day of alcohol compared with drinkers of less than 1 drink/day of alcohol: HR 1.74 (1.08–2.80)	Bendinelli 2020 [38]
		Other Dietary Exposures
No. of Cases (Non-Cases)	Mean Follow-Up (Years)	Results, Relative Risk (95% Confidence Interval (CI))	Reference
1078 (518,922)	3.7	Serum C-peptide concentration was positively associated with an increased colorectal cancer risk (highest vs. lowest quintile): OR 1.37 (1.00–1.88) P trend = 0.10 The cancer risk was stronger for colon: OR 1.67 (1.14–2.46) P trend < 0.01; than for rectal cancer: OR 1.42 (0.90–2.25) P trend = 0.35	Jenab 2007 [58]
1365 (NR)	3.6	The relative risks comparing highest to lowest quintile were: Vitamin B2: RR 0.71 (0.56–0.91) P trend = 0.02 Vitamin B6: RR 0.68 (0.53–0.87) P trend < 0.001 Vitamin B12: RR 1.02 (0.80–1.29) P trend = 0.19	Eussen, Vollset, Hustad 2010 [47]
1367 (NR)	3.6	Folate status (highest vs. lowest quintile): RR 0.94 (0.74–1.20) P trend = 0.44 Individuals living in Northern European countries showed an inverse association between plasma folate and rectal cancer risk (highest vs. lowest folate concentrations): RR 0.56 (0.29–1.09) P trend = 0.04	Eussen, Vollset, Igland 2010 [41]
1248 (518,752)	3.9	The cancer risks associated with 10% higher level of circulating 25-(OH)D were: colorectal HR 0.97 (0.95 to 0.99); colon HR 0.95 (0.93 to 0.98); rectum HR 1.00 (0.97 to 1.03). Lower levels of concentration of 25-(OH)D were associated with higher colorectal cancer risk: <25.0 nmol/L: IRR 1.32 (0.87–2.01); and 25.0–49.9 nmol/L: IRR 1.28 (1.05–1.56) Higher concentrations of 25-(OH)D were associated with lower risk: 75.0–99.9 nmol/L: IRR 0.88 (0.68–1.13); and ≥100.0 nmol/L: 0.77 (0.56–1.06)	Jenab 2010 [30]
861 (25,639)	NR	Individuals who carried the variant genotype with higher vitamin E intake had a lower OR of 0.46 (0.26–0.82) whereas those with lower vitamin E intake had an OR of 1.46 (0.98–2.18) compared with those with the common genotype and lower vitamin E intake (P interaction = 0.009). Similarly, the variant genotype group with higher intake of carotene had an inverse association for colorectal cancer in contrast to the common genotype group, with lower carotene intake: OR 0.39 (0.21–0.71) P interaction = 0.005	Loh 2010 [53]
221 (886)	9	Phyto-oestrogen intake not associated with colorectal cancer among men. Among women: Enterolactone intake: OR 0.33 (0.14–0.74); P trend = 0.008 Total enterolignans intake: OR 0.32 (0.13–0.79); P trend = 0.013 Secoisolariciresinol intake: OR 1.60 (0.96–2.69); P trend = 0.074	Ward 2010 [59]
1202 (518,798)	6.1	Participants with 25(OH)D levels in the highest quintile had an HR of 0.69 (0.50–0.93) for colorectal cancer mortality.	Fedirko 2012 [46]
1372 (384,375)	12.0	Incidence rate ratio of distal colon cancer of plasma total alkylresorcinols (highest vs. lowest quartile): IRR 0.48 (0.28–0.83).	Kyrø 2014 [49]
1399 (520,049)	4.5	An association was observed between higher prediagnostic plasma retinol concentration and a lower risk of colon cancer (highest vs. lowest quartile): IRR 0.63 (0.46–0.87) P trend = 0.01. Dietary b-carotene showed an inverse association with colon cancer (highest vs. lowest quartile): OR 0.69 (0.52–0.94) P trend = 0.02. Dietary vitamin C was inversely associated with risk of distal colon cancer (highest vs. lowest quartile): OR 0.60 (0.39–0.93) P trend = 0.02. Dietary vitamin E showed an inverse association with risk of distal colon cancer (highest vs. lowest quartile): OR 0.65 (0.42–0.99) P trend = 0.04.	Leenders 2014 [48]
1367 (NR)	3.7	Plasma methionine: OR 0.79 (0.63–0.99); P trend = 0.05 Plasma choline: OR 0.77 (0.60–0.99); P trend = 0.07 Plasma betaine: OR 0.85 (0.66–1.09); P trend = 0.06	Nitter 2014 [50]
966 (520,482)	3.9	Higher selenium concentrations were associated with a non-significant lower colorectal cancer risk (per 25 lg/L increase): IRR 0.92 (0.82–1.03)	Hughes 2015 [60]
421 (46,297)	11.0	Glycemic index (highest vs. lowest category): RR 1.35 (1.03–1.78) P trend = 0.031	Sieri 2015 [33]
434 (44,758)	11.28	No significant association between dietary total antioxidant capacity and colorectal cancer: HR 0.88 (0.65–1.19) highest category vs. lowest; P trend = 0.353 Dietary total antioxidant capacity in colon cancer (highest vs. lowest tertile): HR 0.63 (0.44–0.89) P trend = 0.008 Dietary total antioxidant capacity in rectal cancer (highest vs. lowest tertile): HR 2.48 (1.32–4.66) P trend = 0.007 Intakes of vitamin C, vitamin E, and ß-carotene not significantly associated with colorectal cancer risk.	Vece 2015 [61]
4517 (472,795)	11.3	Nutrient pattern characterised by high intakes of vitamins and minerals: HR 0.94 (0.92–0.98) per 1 SD. Pattern characterised by total protein, riboflavin, phosphorus and calcium: HR 0.96 (0.93–0.99) per 1 SD. The remaining two patterns were not significantly associated with colorectal cancer risk.	Moskal 2016 [62]
966 (NR)	NR	Circulating concentration of copper (highest vs. lowest quintile): OR 1.50 (1.06–2.13) P trend = 0.02 Circulating concentration of zinc (highest vs. lowest quintile): OR 0.65 (0.43–0.97) P trend = 0.07 Ratio of copper/zinc (highest vs. lowest quintile): OR 1.70 (1.20–2.40) P trend = 0.0005	Stepien 2017 [52]
4517 (472,795)	11	No association between total flavonoid intake and the risk of overall colorectal cancer or any subtype. Total dietary flavonoid intake (highest vs. lowest quintile): HR 1.05 (0.93–1.18) P trend = 0.58 No association with any intake of individual flavonoid subclasses.	Zamora-Ros 2017 [63]
5991 (470,169)	13.9	Total dietary polyphenol intake (as a continuous variable) in women: HR 1.06 (0.99–1.14) Total dietary polyphenol intake (as a continuous variable) in men: HR 0.97 (0.90–1.11)	Zamora-Ros 2018 [64]
1043 (518,957)	8.3	Results for colorectal cancer mortality associated with deficient relative to sufficient 25(OH)D concentrations were: HR 2.24 (1.44–3.49) among cases with the vitamin D-binding protein isoform. HR 0.94 (0.68–1.22) among cases without vitamin D-binding protein.	Gibbs 2020 [45]
1608 (NR)	7.7	Fatty acids: OR 0.51 (0.29–0.90) per unit increase Endogenous metabolites: OR 0.62 (0.50–0.78) per unit change.	Rothwell 2020 [51]

HR: hazard ratio; IRR: incidence rate ratio; NR: not reported; OR: odds ratio; RR: risk ratio.