Abstract
BackgroundMultiple diet patterns play a crucial role in the development of colorectal cancer and its precursor, colorectal adenoma, but mediating effect of plasma metabolite profiles is unclear.MethodsA total of 95,275 participants from UK Biobank with plasma metabolomics and dietary information were analyzed. Metabolite profile scores for 14 dietary patterns were estimated through elastic net regression. Cox regression analysis assessed the associations of dietary patterns and their metabolite profile scores with colorectal tumor risk. Mediating effects of identified metabolite profile scores were estimated in the associations.ResultsFourteen metabolite profile scores, including a range of 28 to 205 signatures, were weak to moderate correlation with dietary patterns (all p < 0.001). Multivariable Cox regression analyses revealed that five dietary patterns were significantly correlated with a decreased risk of colorectal tumor after FDR correction and adjustment for covariates. HRs (95% CIs) per 1 SD for these diet patterns were as follows: WCRF (0.93, 0.90-0.96), CRC score (0.94, 0.92-0.97), AHEI-2010 (0.95, 0.92-0.97), DASH (0.94, 0.91-0.97), and hPDI (0.95, 0.93-0.98). Similarly, metabolite profile scores for these five dietary patterns were inversely associated with colorectal tumor risk, with HRs (95% CIs) per 1 SD as follows: WCRF (0.59, 0.49-0.70), CRC score (0.67, 0.58-0.77), AHEI-2010 (0.73, 0.65-0.80), DASH (0.75, 0.66-0.84), and hPDI (0.56, 0.47-0.67). The mediation proportions of five metabolite profile scores between dietary patterns and colorectal tumor risk ranged from 6.37 to 27.23% (all p < 0.001).ConclusionsFive dietary patterns and their metabolite profile scores, were inversely correlated with colorectal tumor risk. These findings highlight the potential of metabolite profiles as mediators in the association between dietary patterns and the risk of colorectal tumor, further contributing to the prevention of colorectal cancer or adenoma and providing new insights for future research.