Healthy and unhealthy plant-based diets in relation to the incidence of colorectal cancer overall and by molecular subtypes

Abstract

Background: Plant-based foods have been recommended for health. However, not all plant foods are healthy, and little is known about the association between plant-based diets and specific molecular subtypes of colorectal cancer (CRC). We examined the associations of healthy and unhealthy plant-based diets with the incidence of CRC and its molecular subtypes.

Methods: While 123 773 participants of the Nurses' Health Study and the Health Professionals Follow-up Study had been followed up (3 143 158 person-years), 3077 of them had developed CRC. Healthy and unhealthy plant-based diet indices (hPDI and uPDI, respectively) were calculated using repeated food frequency questionnaire data. We determined the tumoural status of microsatellite instability (MSI), CpG island methylator phenotype (CIMP), and BRAF and KRAS mutations.

Results: Higher hPDI was associated with lower CRC incidence (multivariable hazard ratio [HR] comparing extreme quartiles, 0.86, 95% confidence interval [CI]: 0.77, 0.96; P-trend = .04), whereas higher uPDI was associated with higher CRC incidence (multivariable HR comparing extreme quartiles, 1.16, 95% CI: 1.04, 1.29; P-trend = .005). The association of hPDI significantly differed by KRAS status (P-heterogeneity = .003) but not by other tumour markers. The hPDI was associated with lower incidence of KRAS-wildtype CRC (multivariable HR comparing extreme quartiles, 0.74, 95% CI: 0.57, 0.96; P-trend = .004) but not KRAS-mutant CRC (P-trend = .22).

Conclusions: While unhealthy plant-based diet enriched with refined grains and sugar is associated with higher CRC incidence, healthy plant-based diet rich in whole grains, fruits and vegetables is associated with lower incidence of CRC, especially KRAS-wildtype CRC.

Keywords: colorectal carcinoma; inverse probability weighting; molecular pathological epidemiology; sustainability.

FIGURE 1

Flow chart of study population. CIMP, CpG island methylator phenotype; FFQ, food frequency questionnaire; MSI, microsatellite instability

FIGURE 2

Association of individual plant food with colorectal cancer risk in the pooled cohort of Nurses’ Health Study and the Health Professionals Follow‐up Study. (A) Associations for healthy plant foods and (B) unhealthy plant foods. (C) Associations by equally substituting whole grains, fruits, or vegetables for two servings of refined grains. The associations in (A) and (B) were two servings/day for whole grains, fruits, vegetables and refined grains, and one serving/day for all other plant foods. All models were stratified by age (in month), calendar year and sex and adjusted for body mass index (continuous with a ceiling at 35 kg/m²), physical activity (continuous with a ceiling at 50 metabolic equivalent task score‐hours/week), smoking status (never, past, or current), regular use of aspirin or other non‐steroidal anti‐inflammatory drugs (≥2 tablets per week: yes or no), family history of colorectal cancer (yes or no), history of previous lower gastrointestinal endoscopy (yes or no), alcohol intake (continuous with a ceiling at 30 g/day), total energy intake (continuous) and intake of six animal food groups (continuous)