Diet, Metabolomics, and Colorectal Cancer Risk

Abstract

Colorectal cancer (CRC) is the third most common cancer and the second leading cause of cancer death in 2020 globally. In the United States, although the estimated 2021 CRC incidence only ranks fourth among all cancers, CRC is still estimated to be the second leading cause of cancer death in 2021, placing a heavy burden on public health. According to the American Cancer Society, more than half (55%) of all CRCs in the United States are attributable to modifiable risk factors, such as obesity, physical inactivity, smoking, alcohol consumption, high red or processed meat intake, low dietary calcium intake, and low dietary fiber intake, half of which are diet-related. Among all cancers, CRC is also the most strongly associated with dietary behaviors. This dissertation adopted several research strategies, including epidemiology, metabolomics, and molecular pathology, to investigate the association between diet and CRC risk. In the first project (Chapter 1), we examined the association of folate intake with CRC risk, especially after folic acid fortification in the United States. We used Cox proportional hazards models to estimate the association between intake of four forms of folate and risk of CRC. We found a latency period, specifically, a period longer than 12 years, for the association of total folate, dietary folate, and synthetic folic acid with CRC risk. More importantly, in the post-fortification period, we did not observe an increased risk of CRC related to high total folate and synthetic folic acid intake, even among multivitamin users. In the second project (Chapter 2), we explored the plasma metabolomic profile related to intake of red meat, poultry, and fish and prospectively examined whether these metabolomic profiles were associated with CRC risk. We performed partial Spearman correlation analysis to assess the correlation between each metabolite and intake of red meat, poultry, and fish, and we developed metabolomic scores that were related to dietary consumption. The association of dietary intake and metabolomic scores with CRC risk was then examined by conditional logistic regression. We found systematic differences in plasma metabolite profiles, especially highly unsaturated lipids, according to red meat and fish consumption. Red meat consumption was inversely, and fish consumption was positively correlated to highly unsaturated lipids. The metabolomic score for fish intake was inversely associated with CRC risk. In the last project (Chapter 3), we incorporated molecular pathology and investigated the association between plant-based diets and the risk of CRC overall and by molecular subtypes. We conducted Cox proportional hazards regression to evaluate the association of healthful and unhealthful plant-based diets with CRC risk and whether these associations differed across various molecular subtypes. Our results showed that a plant-based diet rich in whole grains, fruits, and vegetables was associated with a lower CRC risk, especially the subtype non-MSI-high, CIMP-low/negative, BRAF-wildtype, and KRAS-wildtype. In contrast, a plant-based diet emphasizing less healthy plant foods such as refined grains was associated with higher CRC risk. This dissertation alleviated concerns about high folate intake and potentially higher CRC risk that has impeded folic acid fortification, emphasized the importance of differentiating healthy and less healthy plant foods when promoting plant-based diets for CRC prevention, and suggested the promising use of metabolomics in assessing diet-CRC association.