Intake of Dietary Fruit, Vegetables, and Fiber and Risk of Colorectal Cancer According to Molecular Subtypes: A Pooled Analysis of 9 Studies

Abstract

Protective associations of fruits, vegetables, and fiber intake with colorectal cancer risk have been shown in many, but not all epidemiologic studies. One possible reason for study heterogeneity is that dietary factors may have distinct effects by colorectal cancer molecular subtypes. Here, we investigate the association of fruit, vegetables, and fiber intake with four well-established colorectal cancer molecular subtypes separately and in combination. Nine observational studies including 9,592 cases with molecular subtypes for microsatellite instability (MSI), CpG island methylator phenotype (CIMP), and somatic mutations in BRAF and KRAS genes, and 7,869 controls were analyzed. Both case-only logistic regression analyses and polytomous logistic regression analyses (with one control set and multiple case groups) were used. Higher fruit intake was associated with a trend toward decreased risk of BRAF-mutated tumors [OR 4th vs. 1st quartile = 0.82 (95% confidence interval, 0.65-1.04)] but not BRAF-wildtype tumors [1.09 (0.97-1.22); P difference as shown in case-only analysis = 0.02]. This difference was observed in case-control studies and not in cohort studies. Compared with controls, higher fiber intake showed negative association with colorectal cancer risk for cases with microsatellite stable/MSI-low, CIMP-negative, BRAF-wildtype, and KRAS-wildtype tumors (P _trend range from 0.03 to 3.4e-03), which is consistent with the traditional adenoma-colorectal cancer pathway. These negative associations were stronger compared with MSI-high, CIMP-positive, BRAF-mutated, or KRAS-mutated tumors, but the differences were not statistically significant. These inverse associations for fruit and fiber intake may explain, in part, inconsistent findings between fruit or fiber intake and colorectal cancer risk that have previously been reported. SIGNIFICANCE: These analyses by colorectal cancer molecular subtypes potentially explain the inconsistent findings between dietary fruit or fiber intake and overall colorectal cancer risk that have previously been reported.

Figure 1.. Forest plot of the association between fiber intake and colorectal cancer risk by combined molecular subtypes using polytomous logistic regression analysis.

Odds ratios (ORs) and 95% confidence intervals (95% CIs) for multivariate adjusted models are presented for each increasing quartile of fiber intake and each combined molecular subtypes of colorectal cancer (CRC) risk. A total of 3,697 cases and 6,485 controls were included. Gray boxes are centered at multivariate adjusted ORs, and lines depict their 95% CIs. Subtype classifications with less than 50 cases were excluded from analyses. The number of cases per molecular subtype are listed under Cases. Ptrend was calculated by assigning ordinal values for quartile categories of fiber intake and modeling that variable continuously. P-difference is the degree of difference in P-value of multivariate adjusted OR between Type 4 and each of the other Types.