Genome-Wide Interaction Study of Dietary Intake and Colorectal Cancer Risk in the UK Biobank.

IMPORTANCE: Candidate gene analysis approaches have shown that colorectal cancer (CRC) risk attributable to diet may differ according to genotype. A genome-wide approach further allows for the exploration of underlying pathways for associations between diet and CRC risk across the genome.

OBJECTIVES: To identify genetic variants that modify diet-CRC associations and to further explore the underlying pathways in the cause of CRC.

DESIGN, SETTING, AND PARTICIPANTS: This nested case-control study used data on White British participants from the prospective cohort UK Biobank. Participants were recruited between March 13, 2006, and October 1, 2010, and data were censored June 25, 2021.

EXPOSURES: The average frequency intake of 11 dietary factors in the year preceding baseline was obtained via a touchscreen questionnaire. After quality control for more than 93 million variants of imputed genetic data, 4 122 345 variants remained.

MAIN OUTCOMES AND MEASURES: Colorectal cancer cases were identified according to the International Statistical Classification of Diseases and Related Health Problems, Tenth Revision. Genome-wide interaction analysis was performed to test interactions between dietary factors and variants using a conditional logistic regression model. Summary statistics of interactions at the variant level were used to calculate empirical P values for interactions at gene and gene-set levels in gene-based and gene-set enrichment analyses.

RESULTS: A total of 4686 participants with CRC (mean [SD] age, 60.7 [6.6] years; 2707 men [57.8%]) received a new diagnosis during a median of 12.4 years (IQR, 11.6-13.1 years) of follow-up. Once a case was detected, 3 matched controls were identified, for a total of 14 058 controls (mean [SD] age, 60.4 [6.6] years; 8121 men [57.8%]). A total of 324 variants were identified that interacted with diet consumption at the suggestive threshold (P < 1 × 10-5). In gene-based analysis, aggregation of multiple EPDR1 gene variants was found to interact with fish intake regarding CRC risk. Furthermore, gene-set enrichment analysis found that several sets of protein-coding genes, which were overrepresented with particular functions and pathways, interacted with the consumption of milk (ART), cheese (OR), tea (KRT), and alcohol (PRM and TNP).

CONCLUSIONS AND RELEVANCE: In this nested case-control study, the risk of CRC associated with fish intake was modified by multiple single-nucleotide polymorphisms of the EPDR1 gene. The findings further suggested possible functions and pathways that might link the consumption of milk, cheese, tea, and alcohol with CRC development.