[Hypothetical Alcohol Consumption Interventions and Hepatic Steatosis: A Longitudinal Study in a Large Cohort]
Ning Zhang,Yuan Zhang,Jun Wei,Yi Xiang,Yifan Hu,Xiong Xiao
DOI: https://doi.org/10.12182/20240560503
2024-05-20
Abstract:Objective: Non-alcoholic fatty liver disease (NAFLD) and alcohol-associated fatty liver disease (ALD) are the most common chronic liver diseases. Hepatic steatosis is an early histological subtype of both NAFLD and ALD. Excessive alcohol consumption is widely known to lead to hepatic steatosis and subsequent liver damage. However, reported findings concerning the association between moderate alcohol consumption and hepatic steatosis remain inconsistent. Notably, alcohol consumption as a modifiable lifestyle behavior is likely to change over time, but most previous studies covered alcohol intake only once at baseline. These inconsistent findings from existing studies do not inform decision-making concerning policies and clinical guidelines, which are of greater interest to health policymakers and clinician-scientists. Additionally, recommendations on the types of alcoholic beverages are not available. Usually, assessing the effects of two or more hypothetical alcohol consumption interventions on hepatic steatosis provides answers to questions concerning the population risk of hepatic steatosis if everyone changes from heavy drinking to abstinence, or if everyone keeps on drinking moderately, or if everyone of the drinking population switches from red wine to beer? Thus, we simulated a target trial to estimate the effects of several hypothetical interventions, including changes in the amount of alcohol consumption or the types of alcoholic beverages consumed, on hepatic steatosis using longitudinal data, to inform decisions about alcohol-related policymaking and clinical care. Methods: This longitudinal study included 12687 participants from the UK Biobank (UKB), all of whom participated in both baseline and repeat surveys. We excluded participants with missing data related to components of alcohol consumption and fatty liver index (FLI) in the baseline and the repeat surveys, as well as those who had reported liver diseases or cancer at the baseline survey. We used FLI as an outcome indicator and divided the participants into non-, moderate, and heavy drinkers. The surrogate marker FLI has been endorsed by many international organizations' guidelines, such as the European Association for the Study of the Liver. The calculation of FLI was based on laboratory and anthropometric data, including triglyceride, gamma-glutamyl transferase, body mass index, and waist circumference. Participants responded to questions about the types of alcoholic beverages, which were defined in 5 categories, including red wine, white wine/fortified wine/champagne, beer or cider, spirits, and mixed liqueurs, along with the average weekly or monthly amounts of alcohol consumed. Alcohol consumption was defined as pure alcohol consumed per week and was calculated according to the amount of alcoholic beverages consumed per week and the average ethanol content by volume in each alcoholic beverage. Participants were categorized as non-drinkers, moderate drinkers, and heavy drinkers according to the amount of their alcohol consumption. Moderate drinking was defined as consuming no more than 210 g of alcohol per week for men and 140 g of alcohol per week for women. We defined the following hypothetical interventions for the amount of alcohol consumed: sustaining a certain level of alcohol consumption from baseline to the repeat survey (e.g., none to none, moderate to moderate, heavy to heavy) and changing from one alcohol consumption level to another (e.g., none to moderate, moderate to heavy). The hypothetical interventions for the types of alcoholic beverages were defined in a similar way to those for the amount of alcohol consumed (e.g., red wine to red wine, red wine to beer/cider). We applied the parametric g-formula to estimate the effect of each hypothetical alcohol consumption intervention on the FLI. To implement the parametric g-formula, we first modeled the probability of time-varying confounders and FLI conditional on covariates. We then used these conditional probabilities to estimate the FLI value if the alcohol consumption level of each participant was under a specific hypothetical intervention. The confidence interval was obtained by 200 bootstrap samples. Results: For the alcohol consumption from baseline to the repeat surveys, 6.65% of the participants were sustained non-drinkers, 63.68% were sustained moderate drinkers, and 14.74% were sustained heavy drinkers, while 8.39% changed from heavy drinking to moderate drinking. Regarding the types of alcoholic beverages from baseline to the repeat surveys, 27.06% of the drinkers sustained their intake of red wine. Whatever the baseline alcohol consumption level, the hypothetical interventions for increasing alcohol consumption from the baseline alcohol consumption were associated with a higher FLI than that of the sustained baseline alcohol consumption level. When comparing sustained non-drinking with the hypothetical intervention of changing from non-drinking to moderate drinking, the mean ratio of FLI was 1.027 (95% confidence interval [CI]: 0.997-1.057). When comparing sustained non-drinking with the hypothetical intervention of changing from non-drinking to heavy drinking, the mean ratio of FLI was 1.075 (95% CI: 1.042-1.108). When comparing sustained heavy drinking with the hypothetical intervention of changing from heavy drinking to moderate drinking, the mean ratio of FLI was 0.953 (95% CI: 0.938-0.968). The hypothetical intervention of changing to red wine in the UKB was associated with lower FLI levels, compared with sustained consumption of other types of alcoholic beverages. For example, when comparing sustaining spirits with the hypothetical intervention of changing from spirits to red wine, the mean ratio of FLI was 0.981 (95% CI: 0.948-1.014). Conclusions: Regardless of the current level of alcohol consumption, interventions that increase alcohol consumption could raise the risk of hepatic steatosis in Western populations. The findings of this study could inform the formulation of future practice guidelines and health policies. If quitting drinking is challenging, red wine may be a better option than other types of alcoholic beverages in Western populations.