Non-alcoholic fatty liver disease (NAFLD) is the most prevalent chronic liver disorder affecting approximately a quarter of the adult population worldwide (1, 2). NAFLD comprises a continuum of disease severities from steatosis to non-alcoholic steatohepatitis (NASH). It can evolve into an advanced disease that progresses to cirrhosis, liver failure, and an increased risk of hepatocellular carcinoma (HCC) (3). Moreover, NAFLD patients have increased risk of cardiovascular events, other malignancies, and mortality (2, 3). Insulin resistance is a common feature in NAFLD patients, and it is considered the major contributor to the development and progression of the disease (4).

Alcohol consumption in modest quantity is believed to improve insulin resistance, lipid metabolism, and inflammatory status, thereby exerting cardiovascular and metabolic benefits (5). These effects have been shown to reduce the risk of diabetes, cardiovascular disease incidence, and mortality in a J-shape dose-response (6–8). Specific types of alcohols such as red wines and certain drinking patterns, for instance, modest consumption, not binge drinking displayed superior cardiometabolic benefits (9). The benefits of modest alcohol consumption also decrease the risk of developing NAFLD in the general population (10–12). However, recommendations on alcohol consumption among patients with pre-existing NAFLD, where metabolic syndrome and established cardiovascular disease are common comorbidities, remain a topic of vigorous debate, given that the evidence supporting the protective benefits of modest alcohol on liver-related outcomes is less consistent. To date, studies have reported varying results on histological severity, the natural course of liver disease, as well as liver-related outcomes, particularly the development of HCC (13–27). Consequently, there is no current consensus in clinical practice for counseling patients with NAFLD regarding modest alcohol consumption.

Therefore, this systematic review and meta-analysis were performed to comprehensively assess the effects of modest alcohol intake on histological severity, histological progression, and the risk of significant clinical outcomes, namely, the development of cirrhotic complications, HCC, and all-cause death among patients with NAFLD.

A total of potentially relevant 11,794 articles (9,023 from EMBASE and 2,771 from MEDLINE) were retrieved. After removing 2,405 duplicated articles, 9,389 articles remained for the first-round review. We then excluded 9,310 articles because they did not fulfill the inclusion criteria based on study design and types of articles resulting in 79 remaining articles for the second round full-text review. Fifteen studies fulfilled the inclusion criteria and were included in this study (13–27). However, Younossi et al. (27) and Hajifathalian et al. (25) used the identical database of National Health and Nutrition Examination Survey III (NHANES III); therefore, we selected Hajifathalian et al. due to the larger number of participants. Figure 1 provides an overview of the literature review and study selection process. Tables 1A–C summarizes the study design, characteristics of participants, and Newcastle-Ottawa Scale of the included studies.

The definition of both abstainers and modest alcohol drinkers varied considerably across included studies from lifetime abstainers to 0.5 standard drink per day and 20 g of alcohol per week to 21 standard drinks (210 g) of alcohol per week, respectively. The definition for maximum amount for modest drinking in men was higher than that of women in three studies (13, 17, 22). All cross-sectional studies were biopsy-based studies. We did not find any studies examining the effects of modest alcohol on the development of cirrhotic complications in the NAFLD population.

This comprehensive systematic review and meta-analysis suggest a possible association between modest alcohol consumption and decreased NASH and advanced fibrosis. However, moderate alcohol use may diminish the resolution of NASH and increase risk of HCC in NAFLD patients with advanced fibrosis. In contrast, the data from population-based samples show a protective effect of low-to-moderate alcohol consumption on mortality in patients with NAFLD.

The present meta-analysis observed that modest alcohol drinking was associated with a lower risk of having steatohepatitis and advanced fibrosis in biopsy-proven NAFLD patients. This phenomenon may be explained by the modulation of insulin sensitivity and anti-inflammatory effects of modest alcohol intake, resulting in the attenuation of intrahepatic lipid synthesis, accumulation, and subsequent hepatic steatosis (32–37). The improvement of insulin resistance could moderate lipotoxicity, organelle stress, and hepatocyte injury caused by toxic reactive oxygen species generated by lipid metabolism (32, 38). Interestingly, the effect on advanced fibrosis was only observed in higher-quality studies where ascertainment of alcohol consumption was conducted using validated tools, primarily AUDIT-C and Skinner Lifetime Drinking History. In contrast, the remaining studies involved some degree of self-reported questionnaires or interviews. This could potentially lead to recall bias in the latter group of studies and implies its results since under-reporting may be as high as 40–50% in an alcohol consumption survey and remains a barrier to accurate quantification (39).

The increased risk of HCC among patients with NAFLD who consumed a modest amount of alcohol in our analysis was not unexpected and further emphasized the potentially harmful effects of alcohol. It is known that alcohol is an independent risk factor for the development of HCC both directly via DNA damage from toxic metabolites, oxidative stress, and inflammation and indirectly via chronic liver disease and cirrhosis (20, 40, 41). Furthermore, alcohol in conjunction with diabetes and obesity, as are highly prevalent in the NAFLD study population, also exhibits a synergistic interaction and potentially augments the risk of HCC development (20, 23, 41). Our finding was consistent with a previous study by Kawamura et al. demonstrating that the elevated risk of hepatocarcinogenesis started trending with the daily consumption of 20–39 g of ethanol. However, Kawamura et al. used a light drinker of <20 g per day as the baseline comparator, which differed from our baseline group consisted of abstainers (42). It is also worth noting that all of the patients in Ascha et al., which was weighted at 77% in our analysis, were cirrhotic and were referred for liver transplantation listing due to hepatic decompensation. Therefore, there could be confounding factors for an increased risk of developing HCC via the omission of non-cirrhotic and compensated cirrhosis populations (23). In addition, Kimura et al. found in the multivariate analysis that HCC was associated with fibrosis but not with a mild drinking habit and that all HCC patients had advanced fibrosis (fibrosis stage 3–4) (20). As a result, the interpretation of the risk of HCC development should be made with caution due to the limitations of the NAFLD population. Furthermore, this analysis consisted of only two eligible studies in which HCC was identified exclusively in patients with advanced fibrosis/cirrhosis. Therefore, further research to clarify the actual effect of modest alcohol drinking on the development of HCC across the spectrum of NAFLD patients is needed.

It is well-established that light to moderate alcohol consumption is associated with lower mortality for all-cause, cardiovascular, and cerebrovascular deaths via moderation of metabolic profiles (43, 44). However, studies demonstrating these protective effects were primarily conducted in the general population in national surveys (44). In line with these data, our analyses focused on patients diagnosed with NAFLD and found that modest alcohol consumption was associated with a reduction in all-cause mortality. This outcome could be driven by the decrease as mentioned earlier in the prevalence of advanced fibrosis, which was a significant predictor for long-term overall mortality among biopsy-proven NAFLD patients (45). Consequently, as a knock-on effect of advanced fibrosis reduction, mortality from cirrhosis as the leading cause of death (46) might be attenuated as a result. In addition, cardiovascular death is the leading cause of deaths among NAFLD patients, given the shared atherosclerotic risk factors such as age, diabetes, hypertension, dyslipidemia, insulin resistance, and metabolic syndrome (46, 47). Hence, NAFLD is unsurprisingly considered a risk for cardiovascular disease (48), and it is possible that the cardiometabolic benefits of modest alcohol consumption extended from the general population onto this particular group of NAFLD patients accompanied by atherosclerotic risks. This is particularly evident in Aberg et al. that cardiovascular outcomes, albeit not death-exclusive, were lower among very light drinkers (24). Similarly, we found that all-cause mortality benefits only persisted in light drinkers when patients were grouped according to light or moderate drinking habits. This finding implies that alcohol may not confer its protective effects when consumed beyond a very low threshold. Different types of alcoholic beverages would also need to be accounted for, and further studies to prove this causal relationship in the NAFLD population.

Our meta-analysis has some limitations, which are inherent to the design of the included studies. First, the cross-sectional design of the studies evaluating the effects of alcohol use on the severity of liver disease limits our ability to establish causality of the observed associations. Several studies displayed that moderate alcohol drinkers tended to have higher socioeconomic status, increased physical activity, and less obesity than abstainers (49, 50). These factors have been demonstrated to influence drinking patterns and may affect the severity of liver disease, thereby confounding the association between alcohol use and NAFLD. Second, some of these studies reported incomplete adjustments for potential confounders, and thus reliability of the findings is diminished. Third, although longitudinal cohort studies provided the high quality of the prognostic relevance of modest alcohol use on clinical outcomes in NAFLD, these studies failed to obtain lifetime drinking histories to evaluate past heavy alcohol use. Thus, the population abstaining from alcohol drinking may be enriched for former heavy drinkers, leading to selection bias and more severe liver disease. Fourth, another potential limitation of population-based studies is that NAFLD diagnosis was made using serum biomarkers of steatosis such as fatty liver index and hepatic steatosis index. Accordingly, it is inevitable to have misclassified some of the participants in these studies concerning the presence or absence of NAFLD. Finally, although a random-effects model was applied in this meta-analysis, some findings need to be interpreted cautiously, given the high heterogeneity observed. From the results of the sensitivity analyses, it is assumed that high heterogeneity reflects differences in the tools used for alcohol assessment and characteristics of study populations.

Despite these considerations, this meta-analytic study also has important strengths. First, we believe that the topic of our meta-analysis is clinically relevant, given the conflicting literature on the effects of modest alcohol use in NAFLD and emerging data regarding possible mechanisms of modest alcohol protection for NAFLD. Second, we included studies that performed a liver biopsy to diagnose NASH and assess the liver fibrosis stage, and thus, the histological severity association was ascertained by the gold standard. Third, for the clinical outcomes, the included cohorts had long-term follow-up duration for the pre-specified outcomes to occur adequately. Follow-up time for HCC development, especially in patients with advanced fibrosis and cirrhosis, was as long as six years, while the mean follow-up for mortality was up to 11.1 years. Finally, we used standardized risk estimates from all eligible studies to combine estimates across studies.

Conflicting results from high heterogeneity of studies and evidence on whether modest alcohol consumption is detrimental or beneficial make clinicians uncertain for counseling abstinence or allowing modest alcohol drinking for potential health benefits. Thus, medical advice should be made cautiously in the context of individual clinical implications. Undoubtedly, patients with NASH and advanced fibrosis should be considered as high-risk groups for progressing to end-stage liver disease; hence, alcohol drinking should be avoided. On the contrary, NAFLD patients with low fibrosis risk may be allowed for modest and safe drinking. Thus, there is an urgent need to clarify possible variable impacts of modest alcohol use across the different stages of NAFLD.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

WW, SN, and PC were involved in the study design, data search and collection, and writing of the manuscript. PC performed statistical analysis. All authors approved the final version of the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.