Table 3.

Characteristics of included studies (early-onset colorectal cancer).

Study	Setting and gender	Baseline participants	Study type	Dietary assessment	Study details	Results*
Yue et al. (24)	United States of America Female	Nurse's Health Study II 94,217 women included. Exclusion criteria: prior diagnosis of cancer (except non-melanoma skin cancer) or inflammatory bowel disease at baseline, those with implausible energy intakes.	Prospective Cohort Study	FFQ self-administered every 4 years since 1991. Prime Diet Quality Score (PDQS) is comprised of 21 food groups. Plant-based scores based on 18 food groups. Empirical dietary index for hyperinsulinemia = weighted sum of 18 food groups.	Primary endpoint = colorectal cancer. Average daily nutrient intake calculated by multiplying the frequency of intake by the nutrient content of each food and summing nutrient values across all foods. Averages computed from all questionnaires up to the start of each 2-year follow-up questionnaire PDQS score: 0–42 Plant-based scores: 18–90.	Plant-based diet indices and EOCRC risk stratified by age: HR = 1.24, CI = 0.74–2.08, p = 0.54. PDQS and EOCRC risk stratified by age: HR = 0.90, CI = 0.55–1.50.
Peters et al. (29)	Los Angeles County, California Male	147 case-control pairs. Cases: white males diagnosed with adenocarcinoma of colon or rectum at or before age 45. Controls: matched to cases based on race, sex, date of birth within 5 years, neighborhood of residence, marital status, height, birthplace, religion.	Case-Control Study	All participants given same interview with questionnaire. Subjects asked how often they had eaten several foods over most of their adult life (once a week or less, 2–4x/week, 5+x/week).	Pathology of each case was examined to determine subsite of primary lesion (transverse/ascending colon, sigmoid, right-sided, rectum).	Consumption of deep fried foods > 5x/week and risk for rectal cancer: OR = 4.3, CI = 1.5–12.1, p = 0.01; tumors on the right side of the colon: OR = 3.9, CI = 1.4–10.7, p = 0.008. Consumption of raw fruits and vegetables and colon cancer risk: p = 0.006. A greater consumption of red meat not associated with risk of any subsite.
Khan et al. (28)	Karachi, Pakistan Male and female	74 total colorectal cancer cases (39 female and 35 male). Cases: recruited from surgical oncology unit of Civil Hospital Karachi when colorectal cancer diagnosis confirmed by histopathology examination of biopsy. Controls: Two age and gender matched controls (one sibling, one neighbor; a friend in similar socio-economic situation and same age/gender used if there was an absence).	Case-Control Study	Structured questionnaire divided into 6 parts. Dietary section asked frequency of consumption of rice and rice powders, refined grains and sugars, high fat diet, vegetarian or non-vegetarian.	Vegetarian defined as refrained from eating any kind of meat from animals. Non-vegetarian defined as abstained from any food derived from vegetables. High-fat diet defined as continuous consumption of butter, cheese, whole eggs, nuts, yogurt, etc.	Refined foods and EOCRC risk: OR = 0.01, CI = 0.00–0.04 (unadjusted) A vegetarian diet and EOCRC risk: OR = 0.06, CI = 0.02–0.22 High fat diet and EOCRC risk compared to those who avoided this diet: OR = 1.98, CI-1.13–3.49, p = 0.017. (unadjusted)
Chang et al. (27)	Ontario, Canada Male and female	Cases: identified through Ontario Cancer Registry; between 20 and 49 at age of diagnosis and pathologically confirmed incident of invasive colorectal adenocarcinoma (between Jan 2018- May 2019). 175 cases total (58% female). Controls: Ontario residents aged 20–49, no history of CRC, contacted via telephone, 253 total.	Case-Control Study	Self-reported online questionnaire sent via email. Dietary intake assessed via frequency of consumption for specified serving size. FFQ asked about supplement use, consumption of various foods: fruits, vegetables, high-fiber/whole grain, red meat, processed meat, sugary desserts, fast food, canned foods, processed snacks, beverages, sugar substitutes. “Western-like” dietary pattern score calculated.	Dietary assessment was for “2 years ago” from present time of questionnaire. “Western-like” dietary pattern score between 0, 1, 2, 3 assigned for non-beneficial components (red meat, processed meat, sugary drinks, sugary desserts, fast food, processed snacks). Scores reversed for beneficial foods (fruits, vegetables, high fiber/whole grain). Final score calculated by summing up all values. Scores 0–27 (higher = more westernized).	Consumption of 7+ sugary drinks/week and EOCRC risk: OR = 2.99, CI = 1.57–5.68, p = 0.002. Consumption of desserts 3–6x/week and risk of EOCRC: OR = 2.28, CI = 1.28–4.04. Consumption of fast food 2+ times/week compared to <1x/week and risk of EOCRC: OR = 1.84, CI = 0.98–3.46, p = 0.07. 5+ servings of red meat/week and EOCRC risk: OR = 1.06, CI = 0.56–1.98. Higher Western diet score compared to lower and EOCRC risk: OR = 1.92, CI = 1.01–3.66, p = 0.047.
Rosato et al. (26)	Italy (greater, Milan, Pordenone, Forlì, Rome, Latina, Naples) and Switzerland (Swiss Canton of Vaud) Male and female	Cases: histologically confirmed colorectal cancer in major teaching and general hospitals. Controls: admitted to the same hospitals for wide spectrum of reasons. 329 cases and 1,361 controls all ≤ 45 YO were included. 208 colon cancer, 121 rectal cancer. Age range 19–45.	Case-Control Study	Trained interviewer administered questionnaire. Food info based on FFQ and aimed at assessing diet during the 2 years preceding diagnosis or hospital admission (for controls). Asked about 78 foods, food groups or recipes.	Data derived from 3 case-control studies conducted between 1985 and 2009.	EOCRC risk with highest tertile compared to lowest of beta-carotene: OR = 0.52, CI = 0.37–0.72, p < 0.001, vitamin C: OR = 0.68, CI = 0.49–0.94, p = 0.02, vitamin E: OR = 0.38, CI = 0.26–0.58, p < 0.001, and folate: OR = 0.59, CI = 0.40–0.86, p = 0.006. EOCRC risk and high consumption of vegetables: OR = 0.40, CI = 0.28–0.56, p < 0.001; fruits: OR = 0.75, CI = 0.54–1.02, p = 0.073 A greater consumption of red meat and risk: p = 0.63.
Archambault et al. (25)	US, Canada, Australia, Asia, Europe Male and female	Participants gathered from 3 large cohort studies: Colon Cancer Family Registry, Colorectal Transdisciplinary study, Genetics and Epidemiology of Colorectal Cancer Consortium. 3,767 cases < 50 YO, 4,049 controls < 50 YO 23,437 cases > 50 YO, 35,311 controls > 50 YO Cases: confirmed CRC with medical record, pathology report or death certificate Controls: identified based on study-specific eligibility and matching criteria.	Case-Control Study	FFQ used to measure dietary factors. Included: fruit intake, vegetable, red meat, processed meat (all servings/day) total calcium, (mg/d) total dietary fiber (g/day), total folate (mcg/day).	Analyses restricted to participants of genetically defined European descent. Exposures assessed before diagnosis by answering questions with mindset of 1–2 years prior to selection. All dietary variables modeled as sex and study-specific quartiles; reference level was category linked to lowest risk based on previously published studies.	Lower folate consumption and EOCRC risk: OR = 1.14, CI = 1.04–1.24, p = 0.003 for colon cancer; OR = 1.24, CI = 1.11–1.37, p < 0.001 for rectal cancer. Low fiber consumption and risk: OR = 1.14, CI = 1.02–1.27, p = 0.02 for colon cancer; OR = 1.30, CI = 1.14–1.48, p < 0.001 for rectal cancer. A multivariable model for CRC risk and low folate: OR = 1.08, CI = 0.98–1.18, p = 0.11 and fiber consumption OR = 1.11, CI = 1.00–1.23, p = 0.06. Low calcium intake and risk: OR = 1.15, CI = 1.05–1.26, p = 0.003 for colon; OR = 1.24, CI = 1.11–1.39, p < 0.001 for rectum. Multivariable model for low calcium intake: OR = 1.09, CI = 0.99–1.19, p = 0.08. Multivariable model high red meat consumption and risk: OR = 1.10, CI = 1.04–1.16, p < 0.001.