Alcohol, Cancer and Cardiometabolic Health

Abstract

Alcoholic beverages rank among the most popular drinks worldwide. While the adverse effects of heavy alcohol consumption are well-established, health effects of moderate consumption—defined as one drink or less per day for women and two drinks or less for men—have long been the subject of controversy. For instance, moderate wine consumption is considered a healthy component in Mediterranean diet, based on potential benefits on cardiovascular system, whereas some guidelines recommend abstaining from alcohol altogether for cancer prevention. The biological plausibility of alcohol benefiting health lies in ethanol's supposed ability to improve blood lipid profiles, reduce insulin resistance, and modulate systemic inflammation. Polyphenols, found in alcoholic beverages, are also known for their anti-inflammatory and antioxidative properties. These mechanisms play crucial roles in both cardiometabolic conditions and cancer development. On the other hand, the first metabolite of ethanol, acetaldehyde, is carcinogenic. Therefore, the dose-response relationship between alcohol and different health outcomes, influenced collectively by these mechanisms, requires meticulous investigation. Examining associations between alcohol consumption and health outcomes in observational setting face several challenges: confounding by other lifestyle factors and socioeconomic status, reverse causation as people quitting drinking due to illness, and modification by drinking frequency, beverage choice, drinking with meals or not. Long-term changes in alcohol consumption further complicate the analysis. The Nurse’s Health Study (NHS) and Health Professionals Follow-up Study (HPFS) serve as great resources for addressing these challenges as they have comprehensive covariates assessment, information collected on multifaceted drinking pattern, and have repeated assessment of alcohol intake. In Chapter 1, our analysis confirmed that moderate alcohol intake was associated with favorable profiles of insulinemic/glycemic and inflammatory markers and blood lipids. Using generalized linear regression, we observed lower levels of hemoglobin A1c, tumor necrosis factor receptor 2 and intracellular adhesion molecule 1, along with higher adiponectin, HDL-cholesterol and lower triglyceride among drinkers compared to nondrinkers. These associations were more pronounced if the alcohol was consumed from wine (for inflammation) and done regularly (spreading over the week better than binge drinking). In Chapter 2, we conducted Cox proportional hazard regression and found an inverse relationship between alcohol consumption and risk of type 2 diabetes regardless of whether nondrinkers or those consuming 0.1-4.9 g/day served as the reference group. This inverse association persisted after mitigating the impact of reverse causation and confounding through the exclusion of former drinkers, a series of sensitivity analyses, repeated assessments, and adjusting for time-varying confounders. Notably, we found that the inverse association was primarily driven by drinking frequency rather than the quantity consumed per drinking day, indicating that regular light drinking may confer the greatest benefits. In Chapter 3, we focused on colorectal cancer (CRC), the etiology of which is particularly intertwined with insulin resistance and systemic inflammation. Our Cox regression analysis indicated a positive association between total alcohol intake and CRC. Individuals with both high drinking frequency and daily intake exhibited the highest CRC risk, suggesting total alcohol intake is the critical factor. We found the positive association was primarily driven by the consumption of liquor and beer rather than wine. Additionally, we utilized an analytic strategy to separate effects of alcohol intake over time. We estimated the time lag between alcohol consumption and CRC occurrence was around 8 to 12 years. Former drinkers did not experience a significant reduction in CRC risk even after 10 years of quitting or reducing consumption.