Abstract
Nonalcoholic fatty liver disease (NAFLD) comprises more than two thirds of patients with chronic liver disease in the United States. The effect of alcohol consumption on survival in patients with NAFLD is not clear.
We gathered data on National Health and Nutrition Examination Survey (NHANES) participants from 1988 to 2010, and linked them to National Death Index (NDI) for follow up of their survival. We diagnosed NAFLD based on a previously validated biochemical model (Hepatic Steatosis Index). We built multivariate Cox proportional hazards models to evaluate the effect of alcohol consumption on survival of patients with NAFLD.
After excluding participants with significant alcohol use, viral hepatitis, or increased transferrin Saturation 4568 participants with NAFLD were included in the analysis. In a Cox model adjusted for age, sex, and smoking history, drinking 0.5–1.5 drinks per day decreased the risk of overall mortality by 41% (HR=0.59, 95%CI 0.40–0.85, p value=0.005) compared to not drinking. Drinking ≥ 1.5 drinks per day showed a trend towards harm (HR=1.16, 95% CI 0.99–1.36, p value=0.119). After further adjustment for race, physical activity, educational level, diabetes, and fiber and polyunsaturated fatty acid intake, drinking 0.5–1.5 drinks per day continued to show a significant protective effect (HR=0.64, 95% CI 0.42–0.97, p value=0.035), and drinking ≥ 1.5 drinks per day showed a significant harmful effect on mortality (HR=1.45, 95% CI 1.01–2.10, p value=0.047).
Among patients with NAFLD modest alcohol consumption is associated with a significant decrease in all-cause mortality, while drinking ≥1.5 drinks per day is associated with an increase in mortality. These results help to inform the discussion of potential risks and benefits of alcohol use in patients with NAFLD.
Introduction
Non-Alcoholic Fatty Liver Disease (NAFLD) is defined as hepatic steatosis in the absence of secondary causes of steatosis, most importantly significant alcohol use. NAFLD has become the most common liver disease in the United States and comprises more than two thirds of patients with chronic liver disease (1).
Following the increased prevalence of metabolic risk factors such as obesity and diabetes mellitus in the past few decades, prevalence of NAFLD has gained epidemic proportions in the United States to the extent that between one fifth to just less than half of the population is estimated to have NAFLD (2, 3).
NAFLD comprises a continuum of disease severities, from steatosis to inflammation and steatohepatitis, and can potentially progress to fibrosis and cirrhosis (4, 5). NAFLD has become one of the main causes of cirrhosis and liver transplantation in United States(6, 7). However, cardiovascular disease (CVD) remains the most common cause of mortality and morbidity in patients with NAFLD (8–10). NAFLD and CVD share many common risk factors. Additionally, previous studies suggest that NAFLD might be an independent risk factor for CVD beyond the effect of their shared determinants (11). There is robust observational evidence that modest alcohol consumption, as compared to no or heavy alcohol intake, decreases the risk of adverse cardiovascular outcomes (12–16). There is also evidence for beneficial effects of modest alcohol consumption on risk of metabolic syndrome and insulin resistance which are important components of the NAFLD disease process (17–19). However, the effect of alcohol consumption on survival in patients with NAFLD is not well-described, as the available evidence provides conflicting information on whether modest alcohol consumption is beneficial or harmful in this patient population (20, 21).
We hypothesized that modest alcohol consumption will have a protective effect on overall survival in patients with NAFLD. This study was designed to evaluate the effect of alcohol consumption on survival in patients with NAFLD as diagnosed by a biochemical model, while adjusting for important potential confounders, in a cohort of United States general population.
Methods
Study population, alcohol use, and covariates
Participants in National Health and Nutrition Examination Survey III (NHANES III; 1988–1994) and 6 rounds of continuous NHANES (1990–2010) were included in this study. Follow up and mortality data were obtained by linking the NHANES data to National Death Index (NDI) through National Center for Health Statistics (NCHS) public-use linked datasets.
Participants’ individual level data were collected including demographic variables, body mass index (BMI), hepatitis B virus (HBV) and hepatitis C virus (HCV) serologies, transferrin saturation, liver transaminases, alcohol use, smoking history, blood glucose, diagnosis of diabetes mellitus, use of oral hypoglycemic agents or insulin, physical activity, dietary variables, duration of follow-up and cause of death in deceased subjects. Public-use linked mortality files from NCHS used in this study were updated with mortality follow-up data through December 31, 2011.
To identify patients with NAFLD, first NHANES participants with the following conditions were excluded: significant alcohol use (defined as >3 drinks/day for males, and >2 drinks/day for females)(22), positive serologies for Hepatitis B (positive hepatitis B surface antigen) or Hepatitis C (positive Hepatitis C antibody), or transferrin Saturation of more than or equal to 55% (Figure 1). Subsequently, NAFLD was defined using the Hepatic Steatosis Index (HSI) model in the remaining subjects (23). HSI is a biochemical model that predicts presence of NAFLD based on patient’s Aspartate and Alanine transaminase (AST and ALT) ratio as well as BMI, gender, and presence or absence of diabetes, defined as:
Figure 1.
Selection of the study population, from National Health and Nutrition Examination Survey participants from 1988 to 2010
† National Health and Nutrition Examination Survey, ‡ Significant alcohol use is defined as >3 drinks/day for male, and >2 drinks/day for females, ^ Aspartate Aminotransferase (ALT) and Alanine Aminotransferase (ALT), * Hepatic Steatosis Index.
This model has been previously validated with a sensitivity of 93% for ruling out NAFLD in patients with values less than 30, and specificity of 92.4% for detection of NAFLD in patients with values greater than 36, and an area under receiver-operating curve (ROC) of 0.81 (95% CI 0.80–0.82) (23, 24). We used the same cut off values used in the validation study and defined presence of NAFLD as HSI score >36, and absence of NAFLD as HSI score <30.
Alcohol use was defined as the average number of drinks per day over a period of 12 months preceding the examination, calculated based on the average number of days, participants reported drinking and the average number of drinks consumed per day. A drink was defined as a 12-ounce beer, a 5-ounce glass of wine, or one and a half ounces of liquor. To assess the effects of alcohol on mortality, alcohol consumption was categorized into three categories: less than half a drink per day (reference group, called non-drinkers here), half a drink to less than one and a half drinks per day (called modest alcohol consumption here), and greater than or equal to one and a half drinks per day. Smoking was coded as either current smoking, never, or former smoking. Diabetes was defined as fasting plasma glucose of greater than or equal to 126 mg/dL, use of oral hypoglycemic agents or insulin, or reported diagnosis of diabetes by participant’s health care provider. Participants’ race was categorized as white, black, hispanic, and other. Education was categorized as less than 9th grade, 9 to 11th grade, high school graduate/GED or equivalent, and some college or above. Physical activity was assessed by asking the participants to compare themselves to other people of the same gender and age and categorize their physical activity as less active than peers (low), similar activity compared to peers (average), or more active than peers (high). For NHANES rounds 2007–2008, and 2009–2010 this information was not available and was replaced by 3 quantiles of frequency of vigorous physical activity per week corresponding to low, average, and high categories. Quality of participants’ diet was assessed by measuring fiber and Poly-Unsaturated Fatty Acids (PUFA) consumption, calculated as percentage of daily total calories. These two macronutrients were selected a priori based on the quality of available evidence of their association with cardiovascular adverse events (25–28). Fibrosis-4 (FIB-4) index for liver fibrosis was used to categorize patients with NAFLD into two groups: low fibrosis risk defined as FIB-4< 1.79 and high fibrosis risk defined as FIB-4> 1.79 based on prior validation studies in NAFLD patients (29). Upper limit of normal for ALT was defined as 33 IU/l for males and 25 IU/l for females according to 2017 American College of Gastroenterology Guidelines (30). Participants with missing data on covariates were excluded from the study (Figure 1).
Outcomes and statistical analysis
To follow patients longitudinally, NHANES Participants were linked with National Death Index (NDI), and data was extracted on length of follow up (from NHANES enrollment exam until death or last known alive). Main outcome in this study was defined as time from beginning of follow up (date of NHANES physical examination) to death. Time to non-cardiovascular death (death not due to disease of heart, ICD-10 codes I00-I09, I11, I13, I20-I51, or cerebrovascular diseases, ICD-10 codes I60-I69) was used in exploratory analysis.
Data are presented as Number (%), Mean (SD), or Median (IQR). T-test and Chi square tests were used to compare groups in univariate analysis. The main survival analysis for the effect of alcohol was done using unadjusted Kaplan-Meier curves for visual characterization and multivariate Cox proportional hazards regression models to estimate main effects, confidence intervals, and statistical significance. The adjusting variables were decided a priori according to their potential for confounding the association between alcohol consumption and mortality. It was decided that the analysis should be adjusted for age, gender, and smoking status as a minimum. To account for other potential confounders, a fully adjusted analysis was also performed which included race, physical activity, education, diabetes, and dietary variables as covariates in addition to age, gender, and smoking status. All tests are two-tailed with a significance level of alpha=0.05. All analyses were performed with Stata/SE 11.2 for Windows, StataCorp LP (Texas, USA).
Results
Population characteristics
After excluding participants with significant alcohol use, hepatitis B or C, elevated transferrin saturation, or missing data on covariates, 8,162 participants were eligible to be included in the study (Figure 1). Of these 4,568 (55.9%) had an HSI score of more than 36 and were classified as patients with NAFLD, while 1,001 (12.3%) had an HSI score of less than 30 and were classified as not having NAFLD. Participants were followed for a median of 70 months (IQR 38–112) during which 271 and 102 deaths occurred among patients with and without NAFLD, respectively. Table 1 summarizes the characteristics of the study population. Among patients with NAFLD, participants who drank alcohol were more likely to be male, current smokers, white, more physically active, have elevated ALT level, have lower BMI, and lower risk of diabetes compared to non-drinkers. Among patients with NAFLD, the unadjusted probability of death due to cardiovascular disease compared to other causes of death was minimally higher in drinkers compared to non-drinkers (p value=0.022).
Table 1.
Characteristics of the study population
| NAFLD (N=4568) | Not NAFLD (N=1001) | |||||
|---|---|---|---|---|---|---|
| <0.5 drink per day† | >=0.5 drink per day† | P value | <0.5 drink per day† | >=0.5 drink per day† | P value | |
| Number | 3318 | 1250 | 644 | 357 | ||
| Female, N (%) | 1853 (55.9) | 294 (23.5) | <0.001 | 358 (56) | 117 (33) | <0.001 |
| Age, Mean (SD) | 48.9 (16.2) | 47.8 (15.9) | 0.024 | 46.1 (20.1) | 45.3 (19.3) | 0.546 |
| BMI, Mean (SD) | 33.0 (6.3) | 31.6 (4.9) | <0.001 | 20.8 (1.9) | 21.0 (1.8) | 0.070 |
| Diabetes, N (%) | 909 (27.4) | 289 (23.1) | 0.003 | 17 (2.6) | 4 (1.1) | 0.108 |
| Current smoker, N (%) | 476 (14.4) | 268 (21.4) | <0.001 | 178 (27.6) | 122 (34.2) | 0.031 |
| Fiber*, Mean (SD) | 16.2 (10.3) | 16.8 (9.9) | 0.126 | 17.0 (11.4) | 16.9 (10.1) | 0.889 |
| PUFA*, Mean (SD) | 7.3 (3.4) | 7.0 (3.4) | 0.068 | 7.1 (3.1) | 6.6 (2.8) | 0.011 |
| Race, N (%) | <0.001 | 0.117 | ||||
| White | 1517 (45.7) | 656 (52.5) | 383 (59.5) | 216 (60.5) | ||
| Black | 635 (19.1) | 217 (17.4) | 118 (18.3) | 75 (21.0) | ||
| Hispanic | 1061 (31.9) | 352 (28.2) | 104 (16.2) | 56 (15.7) | ||
| Other | 105 (3.2) | 25 (2.0) | 39 (6.1) | 10 (2.8) | ||
| Physical Activity˄, N (%) | <0.001 | 0.008 | ||||
| Low | 497 (25.4) | 130 (18.2) | 62 (15.9) | 19 (9.0) | ||
| Average | 851 (43.5) | 336 (46.9) | 155 (39.9) | 74 (35.1) | ||
| High | 608 (31.1) | 250 (34.9) | 171 (44.1) | 118 (55.9) | ||
| Education, N (%) | <0.001 | 0.112 | ||||
| <9th grade | 381 (11.5) | 128 (10.3) | 55 (8.5) | 22 (6.2) | ||
| 9-11th grade | 483 (14.6) | 214 (17.1) | 87 (13.5) | 59 (16.5) | ||
| High School | 795 (23.9) | 294 (23.5) | 155 (24.1) | 68 (19.1) | ||
| Some college | 1040 (31.4) | 323 (25.9) | 185 (28.7) | 102 (28.6) | ||
| College or higher | 616 (18.6) | 290 (23.2) | 162 (25.2) | 106 (29.7) | ||
| ALT level‡, N (%) | ||||||
| Less than ULN‡ | 2206 (66.5) | 746 (59.7) | <0.001 | 617 (95.8) | 341 (95.5) | 0.046 |
| More than ULN‡ | 1112 (33.5) | 504 (40.3) | 27 (4.2) | 16 (4.5) | ||
| Alcohol use, N (%) | ||||||
| <0.5 drink/day | 3318 (100) | NA | NA | 644 (100) | NA | NA |
| 0.5 -1.4 drinks/day | NA | 874 (69.9) | NA | 253 (70.9) | ||
| >=1.5 drinks/day | NA | 376 (30.1) | NA | 104 (29.1) | ||
| Cause of death, N (%) | ||||||
| Cardiovascular | 54 (26) | 18 (27) | 0.022 | 13 (18) | 11 (37) | 0.043 |
| Non-Cardiovascular | 151 (74) | 48 (73) | 59 (82) | 19 (63) | ||
| Prevalence of co-morbidities, N (%)** | ||||||
| Coronary artery disease | 225 (6.8) | 82 (6.6) | 0.791 | 33 (5.1) | 20 (5.6) | 0.751 |
| Congestive heart failure | 106 (3.2) | 21 (1.7) | 0.005 | 10 (1.6) | 5 (1.4) | 0.845 |
| Cerebrovascular disease | 103 (3.1) | 33 (2.6) | 0.410 | 13 (2.0) | 6 (1.7) | 0.707 |
| Chronic obstructive pulmonary disease | 262 (7.9) | 79 (6.3) | 0.071 | 41 (6.4) | 22 (6.2) | 0.899 |
| Malignancy | 270 (8.1) | 92 (7.4) | 0.386 | 66 (10.3) | 42 (11.8) | 0.459 |
Alcohol use was defined as the average number of drinks per day over a period of 12 months preceding the examination; participants with significant alcohol consumption (i.e. >3 drinks/day for males, and >2 drinks/day for females) are excluded in this study.
Consumption calculated as percentage of daily total calories for fiber and Poly-unsaturated Fatty Acids (PUFA).
Physical activity compared to average for similar age and gender.
Upper Limit of Normal (ULN) for Alanine Aminotransferase (ALT) is defined as 33 IU/l for males and 25 IU/l for females.
Co-morbidities are self reported by participants
Effect of alcohol consumption on mortality
Among patients with NAFLD (N=4,568) and compared to non-drinkers, modest alcohol consumption (here defined as drinking an average of half to one and a half drinks of alcohol per day) was associated with a significant 41% lower risk of death (HR=0.59, 95% CI 0.40–0.85, p value=0.005) after adjusting for age, gender, and smoking status (Figure 2). However, drinking an average of more than or equal to one and a half drinks per day showed a trend towards increased mortality among these patients (HR=1.16, 95% CI 0.99–1.36, p value=0.119). A similar but non-significant trend towards protection was observed among participants without NAFLD (N=1,001) with modest alcohol consumption (HR=0.80, 95% CI 0.49–1.31, p value=0.382). To control for the effect of potential confounders, the association between alcohol consumption and mortality was re-evaluated in a fully adjusted model (Table 2). Modest alcohol consumption continued to show a significant protective effect in patients with NAFLD (HR=0.64, 95% CI 0.42–0.97, p value=0.035), while drinking more than or equal to one and a half drinks per day showed a significant harmful effect in these patients (HR=1.45, 95% CI 1.01–2.10, p value=0.047). We failed to find a significant effect for alcohol consumption on mortality in participants without NAFLD in fully adjusted analysis (Table 2). Effect of alcohol was evaluated in men and women with NAFLD separately. Among men with NAFLD modest alcohol consumption was associated with a significant 36% lower risk of death (HR=0.64, 95% CI 0.42–0.97, p value=0.037) after adjusting for age, and smoking status (Figure 3), and drinking an average of more than or equal to one and a half drinks per day showed a trend towards increased mortality (HR=1.29, 95% CI 0.91–1.84, p value=0.158). Among women with NAFLD modest alcohol consumption was associated with a significant 60% lower risk of death (HR=0.40, 95% CI 0.16–0.98, p value=0.047) (Figure 3), and drinking an average of more than or equal to one and a half drinks per day showed a trend towards increased mortality (HR=1.47, 95% CI 0.63–3.41, p value=0.376). Effect of alcohol was also evaluated separately in older (>=65 years) and younger (< 65 years) participants with NAFLD. Among older patients with NAFLD modest alcohol consumption was associated with a significant 57% lower risk of death (HR=0.43, 95% CI 0.26–0.71, p value=0.001) after adjusting for gender, and smoking status, and drinking an average of more than or equal to one and a half drinks per day did not show a significant effect on mortality (HR=0.92, 95% CI 0.56–1.52, p value=0.754). Among younger patients with NAFLD modest alcohol consumption was not associated with a significant effect on mortality (HR=0.69, 95% CI 0.0.40–1.19, p value=0.184), and drinking an average of more than or equal to one and a half drinks per day was associated with a significant increase in risk of death (HR=1.62, 95% CI 1.03–2.54, p value=0.037).
Figure 2.
Unadjusted Kalpan-Meier survival curves for effect of alcohol consumption on all-cause mortality in patients with Non-Alcoholic Fatty Liver Disease (NAFLD).
Table 2.
Effect of Alcohol consumption on all-cause mortality in patients with or without Non-alcoholic Fatty Liver Disease (NAFLD), adjusted for age, gender, smoking, race, physical activity, diet, education, and diabetes status.
| NAFLD (N=2551) | Not NAFLD (N=546) | |||
|---|---|---|---|---|
| HR (95% CI) | P value | HR (95% CI) | P value | |
| 0.5 -1.4 drinks/day* | 0.64 (0.42-0.97) | 0.035 | 0.89 (0.51-1.55) | 0.678 |
| ≥1.5 drinks/day* | 1.45 (1.01-2.10) | 0.047 | 0.88 (0.46-1.67) | 0.687 |
| Age | 1.09 (1.08-1.10) | <0.001 | 1.09 (1.07-1.11) | <0.001 |
| Male sex | 1.55 (1.16-2.06) | 0.003 | 2.19 (1.37-3.53) | 0.001 |
| Current smoking | 2.33 (1.66-3.27) | <0.001 | 1.45 (0.85--492.49) | 0.174 |
| Race | 0.86 (0.73-1.01) | 0.068 | 1.14 (0.86-1.51) | 0.355 |
| Physical Activity˄ | 0.73 (0.62-0.88) | 0.001 | 0.71 (0.51-1.00) | 0.052 |
| Fiber† | 0.82 (0.72-0.93) | 0.002 | 0.96 (0.78-1.19) | 0.709 |
| PUFA† | 0.93 (0.83-1.04) | 0.211 | 0.75 (0.60-0.93) | 0.010 |
| Education | 0.96 (0.85-1.09) | 0.551 | 1.13 (0.91-1.39) | 0.264 |
| Diabetes | 1.62 (1.23-2.13) | 0.001 | 1.97 (0.30-3.22) | 0.973 |
Compared to non-drinkers (average of <0.5 drink/day).
HR for changing from less active to normally active or from normally active to more active.
HR for a quartile increase in percentage of daily total calories for fiber and Poly-unsaturated Fatty Acids (PUFA).
Figure 3.
Unadjusted Kalpan-Meier survival curves for effect of alcohol consumption on all-cause mortality in patients with Non-Alcoholic Fatty Liver Disease (NAFLD), separately by gender.
Additionally, we assessed the effects of alcohol consumption on mortality in patients with NAFLD according to risk of liver fibrosis based on FIB-4 index for liver fibrosis (Table 3). Among patients with NAFLD with low risk of fibrosis, modest alcohol consumption continued to show a protective effect on all-cause mortality (HR=0.46, 95% CI 0.26–0.81), and drinking more than or equal to one and a half drinks per day showed a trend towards harm (HR=1.42, 95% CI 0.92–2.18). The same analysis in NAFLD patient with high risk of fibrosis failed to show any harmful effect for modest alcohol consumption and showed a non-significant trend towards benefit (HR=0.64, 95 % CI 0.29–1.35). Drinking more than or equal to one and a half drinks per day in patients with NAFLD with high risk of fibrosis showed a trend towards harm (HR=1.08, 95% CI 0.51–2.28), similar to its effect in patient with low risk of fibrosis.
Table 3.
Effect of Alcohol consumption on all-cause mortality in patients with Non-alcoholic Fatty Liver Disease (NAFLD) according to risk of liver fibrosis.
| NAFLD with low fibrosis risk† (N=3612) |
NAFLD with high fibrosis risk† (N=328) |
|||
|---|---|---|---|---|
| HR (95% CI) | P value | HR (95% CI) | P value | |
| 0.5 -1.4 drinks/day* | 0.46 (0.26-0.81) | 0.007 | 0.64 (0.29-1.35) | 0.241 |
| ≥1.5 drinks/day* | 1.42 (0.92-2.18) | 0.112 | 1.08 (0.51-2.28) | 0.839 |
| Age | 1.09 (1.08-1.10) | <0.001 | 1.11 (1.07-1.14) | <0.001 |
| Male sex | 1.72 (1.23-2.42) | 0.002 | 1.23 (0.75-2.03) | 0.419 |
| Current smoking | 2.35 (1.61-3.45) | <0.001 | 1.68 (0.70-4.05) | 0.244 |
Low fibrosis risk defined as FIB-4< 1.79 and high fibrosis risk defined as FIB-4> 1.79.
Compared to non-drinkers (average of <0.5 drink/day)
We assessed the effect of alcohol consumption on non-cardiovascular death among patients with NAFLD in a fully adjusted analysis. Modest alcohol consumption continued to show a significant protective effect of Non-Cardiovascular mortality, similar to its effects on all-cause mortality with a 41% decrease in mortality (HR=0.59, 95% CI 0.36–0.96, p value=0.033). Similarly, drinking more than or equal to one and a half drinks per day on average showed a trend towards harm for non-cardiovascular death, similar to its effects on all-cause mortality (HR=1.45, 95% CI 0.96–2.21, p value=0.079).
Discussion
Although there is evidence suggesting potential beneficial effects for modest alcohol consumption in patients with NAFLD, the association between alcohol consumption and survival has not been adequately studied in these patients (31). In this prospective cohort of the United States population, we showed that modest alcohol consumption (here defined as drinking an average of half to one and a half drinks per day, equivalent to 7 to 21 g of alcohol per day) is associated with a robust and significant decrease in all-cause mortality, while drinking more than an average of one and a half drinks per day (≥21 g of alcohol per day) is associated with an increase in mortality in patients with NAFLD. We failed to find any harmful effect for modest alcohol consumption in Patients with NAFLD with high predicted risk of liver fibrosis. Interestingly, we showed that modest alcohol consumption has a protective effect on non-cardiovascular mortality in patients with NAFLD, similar to its effect on all-cause mortality.
The results of this study suggest that the effect of alcohol consumption on mortality in patients with NAFLD depends on the amount of drinking, at least partially. Previous studies on effects of alcohol in patients with NAFLD have generally evaluated incidence or severity of liver disease rather than association of alcohol consumption with mortality. However, our findings regarding protective effect of modest alcohol consumption on mortality in this population is consistent with the results of these studies which show a beneficial effect for modest alcohol consumption in development and severity of NAFLD. A study in a healthy Japanese population using data from annual health checkup data showed an inverse relationship between modest alcohol consumption (up to 40 g of alcohol per day) and risk of elevated transaminases (32). In a 2016 meta-analysis of mostly cross-sectional studies with a total of 76,608 participants, drinking alcohol up to 40 g per day was associated with a 23% reduction in prevalence of fatty liver disease (33). In a recent prospective study of a healthy Japanese population without liver disease at baseline who were followed for up to 3 years, drinking alcohol (up to more than 40 g per day for men and 20 g per day for women) was associated with decreased incidence of fatty liver diagnosed by ultrasound imaging (21). In a cross-sectional study of baseline data for adult patients in NASH Clinical Research Network (CRN) with biopsy proven NAFLD, after exclusion of heavy and binge drinkers, modest alcohol consumption (i.e. less than or equal to 20 g of alcohol per day) was associated with 44% lower risk of liver fibrosis and 34% less hepatocellular ballooning compared to non-drinkers (20). A second study from the same network also showed a beneficial effect on fibrosis stage in modest drinkers versus alcohol abstainers (34). Similarly, our results regarding increased mortality in patients with NAFLD with heavier alcohol consumption are consistent with the findings of previous studies which show increased steatosis and severity of liver disease in NAFLD associated with heavy drinking. In a cross-sectional study of Italian population where fatty liver was diagnosed using ultrasound imaging, drinking of more than 30 g of alcohol per day was associated with 2.8 fold increase in the risk of steatosis (35). In a study of seventy one patients with biopsy proven NAFLD who were followed for an average of about fourteen years, heavy episodic drinking was associated with increased risk of progression of fibrosis defined as progression of more than one fibrosis stage or development of end-stage liver disease over time (36).
In addition to the role of alcohol in development and progression of liver disease, the protective effect of modest drinking on mortality in this study can be mediated through the effect of alcohol on risk of cardiovascular events and all-cause mortality, and its potential beneficial effects on metabolic syndrome, and insulin resistance. Patients with NAFLD are at greater risk for cardiovascular disease compared to general population (37). In fact, patient with NAFLD are more likely to die from cardiovascular disease rather than liver disease (9, 38), and modest alcohol consumption has been shown to be associated with decreased risk of cardiovascular disease mortality in previous studies (13, 15, 16). Interestingly, our results suggest that the effect of alcohol cannot be exclusively explained through its effect on cardiovascular outcomes as we showed that modest alcohol consumption is associated with significant decrease in risk of non-cardiovascular mortality in patients with NAFLD. This finding is consistent with available evidence that suggests a decrease in all-cause mortality associated with modest alcohol consumption. A meta-analysis of 34 prospective studies on alcohol showed a J-shape relationship between alcohol and mortality in both men and women, and drinking up to 4 drinks per day for men and 2 drinks per day for women was associated with a decreased in total mortality (12). Similarly, a prospective study of a large United States population followed for nine year showed that general mortality rate was lowest among men and women who reported drinking about one drink per day, which is consistent with the findings of our study (14). Smoking is one of the most important predictors of all-cause mortality and adverse cardiovascular outcomes. Our analysis highlights the harmful effect of smoking in patients with NAFLD as smoking was associated with the largest increase in risk of death in this population for both overall mortality (HR=2.33) as well as non-cardiovascular mortality(HR=2.79).
NAFLD is associated with insulin resistance and increased triglyceride levels (39–41), decreased adiponectin levels (42–45), and metabolic syndrome (46, 47). Metabolic syndrome has been shown to be a risk factor for steatohepatitis and liver fibrosis in patients with NAFLD (47, 48). Alcohol consumption might have beneficial effects on these pathophysiologic processes which can contribute to the observed protective effect of modest alcohol use on mortality in patients with NAFLD in this study. Observational data suggests a significant inverse association between alcohol consumption and risk of metabolic syndrome (17, 19). Furthermore, interventional studies have shown that daily modest alcohol consumption is associated with increased adiponectin levels, and improved insulin sensitivity. A randomized crossover trial on 23 middle-aged male subjects who consumed an equivalent of 40 grams of ethanol or tap water daily during two successive periods of 17 days showed an average 11% increase in plasma adiponectin level which has been associated with improved insulin sensitivity index among insulin resistant participants (49). Similarly, a randomized controlled crossover trial on 63 postmenopausal women who were randomly assigned to consume 0, 15, or 30 grams per day of alcohol for 8 weeks showed a 19% percent decrease in insulin concentration and 7% increased insulin sensitivity and a 10% decrease in triglyceride level in participants who consumed 30 grams of alcohol daily, as well as a 8% decrease in triglycerides with 15 grams of alcohol consumption (18).
Our study is limited by the fact that diagnosis of NAFLD was made using HSI biochemical model, and not through tissue biopsy. We have excluded participants who have other common causes of chronic liver disease, and a validation study has shown that the HSI model can be very accurate in ruling in and out NAFLD with sensitivity and specificity of more than 90% for the cut off values used here (23, 24). However, it is inevitable to have misclassified some of the participants in our study regarding presence or absence of NAFLD. On the other hand, using a model based on available anthropomorphic and laboratory data has enabled us to study a large sample of the United States population with a long median follow up which increases the power and generalizability of our study. It is also plausible that the effect of alcohol on mortality in patients with NAFLD depends on the severity of liver disease and presence of advance fibrosis. We tried to clarify this by categorizing and analyzing patients based on their risk of liver fibrosis using a biochemical model. However, biochemical models are not a perfect predictor of the presence or absence of fibrosis on liver biopsy (29). As with most prospective cohort studies, our estimated can be affected by the fact that predictors of mortality, including alcohol intake, are measured only at baseline. Finally, investigating the relationship between alcohol and mortality is fraught with bias introduced through the effect of confounders, such as smoking, physical activity, social factors, diet, and presence or absence of major comorbidities. However, ethical considerations limit the use of alcohol as an experimental intervention in human studies, and the way forward remains to be careful analysis of high quality prospective observational data. In this study we paid particular attention to assessment of potential confounders, and adjusted for the effect of many known confounders which should improve the validity of our findings. Additionally, the self-reported rate of major comorbidities did not differ significantly between drinker and non-drinker participants in our study, which further decreases the risk of potential bias.
Results of this study suggest a protective effect of modest alcohol consumption on mortality in patients with NAFLD. In addition, our findings are consistent with previous studies showing a delicate J-shape relationship between alcohol consumption and mortality, where an increase in alcohol intake quickly becomes harmful (50). Although recommendations regarding health effects of alcohol remain controversial, our results can be useful when counseling the growing population of patients with NAFLD in the United States on benefits and harms of alcohol.
Acknowledgments
Funding in part by NIH UDK 505, NIAA1 U01 AA021893, NIH U11020821, 1P50AA024333
References
- 1.Younossi ZM, Stepanova M, Afendy M, Fang Y, Younossi Y, Mir H, Srishord M. Changes in the prevalence of the most common causes of chronic liver diseases in the United States from 1988 to 2008. Clin Gastroenterol Hepatol 2011;9:524–530 e521; quiz e560. [DOI] [PubMed] [Google Scholar]
- 2.Williams CD, Stengel J, Asike MI, Torres DM, Shaw J, Contreras M, Landt CL, et al. Prevalence of nonalcoholic fatty liver disease and nonalcoholic steatohepatitis among a largely middle-aged population utilizing ultrasound and liver biopsy: a prospective study. Gastroenterology 2011;140:124–131. [DOI] [PubMed] [Google Scholar]
- 3.Lazo M, Hernaez R, Eberhardt MS, Bonekamp S, Kamel I, Guallar E, Koteish A, et al. Prevalence of nonalcoholic fatty liver disease in the United States: the Third National Health and Nutrition Examination Survey, 1988–1994. Am J Epidemiol 2013;178:38–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Singh S, Allen AM, Wang Z, Prokop LJ, Murad MH, Loomba R. Fibrosis progression in nonalcoholic fatty liver vs nonalcoholic steatohepatitis: a systematic review and meta-analysis of paired-biopsy studies. Clin Gastroenterol Hepatol 2015;13:643–654 e641–649; quiz e639–640. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Farrell GC, Larter CZ. Nonalcoholic fatty liver disease: from steatosis to cirrhosis. Hepatology 2006;43:S99–S112. [DOI] [PubMed] [Google Scholar]
- 6.Wong RJ, Aguilar M, Cheung R, Perumpail RB, Harrison SA, Younossi ZM, Ahmed A. Nonalcoholic steatohepatitis is the second leading etiology of liver disease among adults awaiting liver transplantation in the United States. Gastroenterology 2015;148:547–555. [DOI] [PubMed] [Google Scholar]
- 7.Charlton MR, Burns JM, Pedersen RA, Watt KD, Heimbach JK, Dierkhising RA. Frequency and outcomes of liver transplantation for nonalcoholic steatohepatitis in the United States. Gastroenterology 2011;141:1249–1253. [DOI] [PubMed] [Google Scholar]
- 8.Ekstedt M, Franzen LE, Mathiesen UL, Thorelius L, Holmqvist M, Bodemar G, Kechagias S. Long-term follow-up of patients with NAFLD and elevated liver enzymes. Hepatology 2006;44:865–873. [DOI] [PubMed] [Google Scholar]
- 9.Rafiq N, Bai C, Fang Y, Srishord M, McCullough A, Gramlich T, Younossi ZM. Long-term follow-up of patients with nonalcoholic fatty liver. Clin Gastroenterol Hepatol 2009;7:234–238. [DOI] [PubMed] [Google Scholar]
- 10.Soderberg C, Stal P, Askling J, Glaumann H, Lindberg G, Marmur J, Hultcrantz R. Decreased survival of subjects with elevated liver function tests during a 28-year follow-up. Hepatology 2010;51:595–602. [DOI] [PubMed] [Google Scholar]
- 11.Targher G, Day CP, Bonora E. Risk of cardiovascular disease in patients with nonalcoholic fatty liver disease. N Engl J Med 2010;363:1341–1350. [DOI] [PubMed] [Google Scholar]
- 12.Di Castelnuovo A, Costanzo S, Bagnardi V, Donati MB, Iacoviello L, de Gaetano G. Alcohol dosing and total mortality in men and women: an updated meta-analysis of 34 prospective studies. Arch Intern Med 2006;166:2437–2445. [DOI] [PubMed] [Google Scholar]
- 13.Ronksley PE, Brien SE, Turner BJ, Mukamal KJ, Ghali WA. Association of alcohol consumption with selected cardiovascular disease outcomes: a systematic review and meta-analysis. BMJ 2011;342:d671. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Thun MJ, Peto R, Lopez AD, Monaco JH, Henley SJ, Heath CW Jr., Doll R , Alcohol consumption and mortality among middle-aged and elderly U.S. adults. N Engl J Med 1997;337:1705–1714. [DOI] [PubMed] [Google Scholar]
- 15.Fuchs CS, Stampfer MJ, Colditz GA, Giovannucci EL, Manson JE, Kawachi I, Hunter DJ, et al. Alcohol consumption and mortality among women. N Engl J Med 1995;332:1245–1250. [DOI] [PubMed] [Google Scholar]
- 16.Mukamal KJ, Chen CM, Rao SR, Breslow RA. Alcohol consumption and cardiovascular mortality among U.S. adults, 1987 to 2002. J Am Coll Cardiol 2010;55:1328–1335. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Freiberg MS, Cabral HJ, Heeren TC, Vasan RS, Curtis Ellison R. Alcohol consumption and the prevalence of the Metabolic Syndrome in the US.: a cross-sectional analysis of data from the Third National Health and Nutrition Examination Survey. Diabetes Care 2004;27:2954–2959. [DOI] [PubMed] [Google Scholar]
- 18.Davies MJ, Baer DJ, Judd JT, Brown ED, Campbell WS, Taylor PR. Effects of moderate alcohol intake on fasting insulin and glucose concentrations and insulin sensitivity in postmenopausal women: a randomized controlled trial. JAMA 2002;287:2559–2562. [DOI] [PubMed] [Google Scholar]
- 19.Yoon YS, Oh SW, Baik HW, Park HS, Kim WY. Alcohol consumption and the metabolic syndrome in Korean adults: the 1998 Korean National Health and Nutrition Examination Survey. Am J Clin Nutr 2004;80:217–224. [DOI] [PubMed] [Google Scholar]
- 20.Dunn W, Sanyal AJ, Brunt EM, Unalp-Arida A, Donohue M, McCullough AJ, Schwimmer JB. Modest alcohol consumption is associated with decreased prevalence of steatohepatitis in patients with non-alcoholic fatty liver disease (NAFLD). J Hepatol 2012;57:384–391. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Moriya A, Iwasaki Y, Ohguchi S, Kayashima E, Mitsumune T, Taniguchi H, Ando M, et al. Roles of alcohol consumption in fatty liver: a longitudinal study. J Hepatol 2015;62:921–927. [DOI] [PubMed] [Google Scholar]
- 22.Chalasani N, Younossi Z, Lavine JE, Diehl AM, Brunt EM, Cusi K, Charlton M, et al. The diagnosis and management of non-alcoholic fatty liver disease: practice guideline by the American Gastroenterological Association, American Association for the Study of Liver Diseases, and American College of Gastroenterology. Gastroenterology 2012;142:1592–1609. [DOI] [PubMed] [Google Scholar]
- 23.Lee JH, Kim D, Kim HJ, Lee CH, Yang JI, Kim W, Kim YJ, et al. Hepatic steatosis index: a simple screening tool reflecting nonalcoholic fatty liver disease. Dig Liver Dis 2010;42:503–508. [DOI] [PubMed] [Google Scholar]
- 24.Meffert PJ, Baumeister SE, Lerch MM, Mayerle J, Kratzer W, Volzke H. Development, external validation, and comparative assessment of a new diagnostic score for hepatic steatosis. Am J Gastroenterol 2014;109:1404–1414. [DOI] [PubMed] [Google Scholar]
- 25.Pereira MA, O’Reilly E, Augustsson K, Fraser GE, Goldbourt U, Heitmann BL, Hallmans G, et al. Dietary fiber and risk of coronary heart disease: a pooled analysis of cohort studies. Arch Intern Med 2004;164:370–376. [DOI] [PubMed] [Google Scholar]
- 26.Farvid MS, Ding M, Pan A, Sun Q, Chiuve SE, Steffen LM, Willett WC, et al. Dietary linoleic acid and risk of coronary heart disease: a systematic review and meta-analysis of prospective cohort studies. Circulation 2014;130:1568–1578. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.de Lorgeril M, Renaud S, Mamelle N, Salen P, Martin JL, Monjaud I, Guidollet J, et al. Mediterranean alpha-linolenic acid-rich diet in secondary prevention of coronary heart disease. Lancet 1994;343:1454–1459. [DOI] [PubMed] [Google Scholar]
- 28.Siscovick DS, Raghunathan TE, King I, Weinmann S, Wicklund KG, Albright J, Bovbjerg V, et al. Dietary intake and cell membrane levels of long-chain n-3 polyunsaturated fatty acids and the risk of primary cardiac arrest. JAMA 1995;274:1363–1367. [DOI] [PubMed] [Google Scholar]
- 29.Shah AG, Lydecker A, Murray K, Tetri BN, Contos MJ, Sanyal AJ. Comparison of noninvasive markers of fibrosis in patients with nonalcoholic fatty liver disease. Clin Gastroenterol Hepatol 2009;7:1104–1112. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Kwo PY, Cohen SM, Lim JK. ACG Clinical Guideline: Evaluation of Abnormal Liver Chemistries. Am J Gastroenterol 2017;112:18–35. [DOI] [PubMed] [Google Scholar]
- 31.Ajmera VH, Terrault NA, Harrison SA. Is moderate alcohol use in nonalcoholic fatty liver disease good or bad? A critical review. Hepatology 2017;65:2090–2099. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Suzuki A, Angulo P, St Sauver J, Muto A, Okada T, Lindor K. Light to moderate alcohol consumption is associated with lower frequency of hypertransaminasemia. Am J Gastroenterol 2007;102:1912–1919. [DOI] [PubMed] [Google Scholar]
- 33.Cao G, Yi T, Liu Q, Wang M, Tang S. Alcohol consumption and risk of fatty liver disease: a meta-analysis. PeerJ 2016;4:e2633. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Zein CO, Unalp A, Colvin R, Liu YC, McCullough AJ. Smoking and severity of hepatic fibrosis in nonalcoholic fatty liver disease. J Hepatol 2011;54:753–759. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Bellentani S, Saccoccio G, Masutti F, Croce LS, Brandi G, Sasso F, Cristanini G, et al. Prevalence of and risk factors for hepatic steatosis in Northern Italy. Ann Intern Med 2000;132:112–117. [DOI] [PubMed] [Google Scholar]
- 36.Ekstedt M, Franzen LE, Holmqvist M, Bendtsen P, Mathiesen UL, Bodemar G, Kechagias S. Alcohol consumption is associated with progression of hepatic fibrosis in non-alcoholic fatty liver disease. Scand J Gastroenterol 2009;44:366–374. [DOI] [PubMed] [Google Scholar]
- 37.Adams LA, Lymp JF, St Sauver J, Sanderson SO, Lindor KD, Feldstein A, Angulo P. The natural history of nonalcoholic fatty liver disease: a population-based cohort study. Gastroenterology 2005;129:113–121. [DOI] [PubMed] [Google Scholar]
- 38.Jepsen P, Vilstrup H, Mellemkjaer L, Thulstrup AM, Olsen JH, Baron JA, Sorensen HT. Prognosis of patients with a diagnosis of fatty liver--a registry-based cohort study. Hepatogastroenterology 2003;50:2101–2104. [PubMed] [Google Scholar]
- 39.Marchesini G, Brizi M, Morselli-Labate AM, Bianchi G, Bugianesi E, McCullough AJ, Forlani G, et al. Association of nonalcoholic fatty liver disease with insulin resistance. Am J Med 1999;107:450–455. [DOI] [PubMed] [Google Scholar]
- 40.Utzschneider KM, Kahn SE. Review: The role of insulin resistance in nonalcoholic fatty liver disease. J Clin Endocrinol Metab 2006;91:4753–4761. [DOI] [PubMed] [Google Scholar]
- 41.Kelley DE, McKolanis TM, Hegazi RA, Kuller LH, Kalhan SC. Fatty liver in type 2 diabetes mellitus: relation to regional adiposity, fatty acids, and insulin resistance. Am J Physiol Endocrinol Metab 2003;285:E906–916. [DOI] [PubMed] [Google Scholar]
- 42.Xu A, Wang Y, Keshaw H, Xu LY, Lam KS, Cooper GJ. The fat-derived hormone adiponectin alleviates alcoholic and nonalcoholic fatty liver diseases in mice. J Clin Invest 2003;112:91–100. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Bugianesi E, Pagotto U, Manini R, Vanni E, Gastaldelli A, de Iasio R, Gentilcore E, et al. Plasma adiponectin in nonalcoholic fatty liver is related to hepatic insulin resistance and hepatic fat content, not to liver disease severity. J Clin Endocrinol Metab 2005;90:3498–3504. [DOI] [PubMed] [Google Scholar]
- 44.Pagano C, Soardo G, Esposito W, Fallo F, Basan L, Donnini D, Federspil G, et al. Plasma adiponectin is decreased in nonalcoholic fatty liver disease. Eur J Endocrinol 2005;152:113–118. [DOI] [PubMed] [Google Scholar]
- 45.Targher G, Bertolini L, Rodella S, Zoppini G, Scala L, Zenari L, Falezza G. Associations between plasma adiponectin concentrations and liver histology in patients with nonalcoholic fatty liver disease. Clin Endocrinol (Oxf) 2006;64:679–683. [DOI] [PubMed] [Google Scholar]
- 46.Hamaguchi M, Kojima T, Takeda N, Nakagawa T, Taniguchi H, Fujii K, Omatsu T, et al. The metabolic syndrome as a predictor of nonalcoholic fatty liver disease. Ann Intern Med 2005;143:722–728. [DOI] [PubMed] [Google Scholar]
- 47.Marchesini G, Marzocchi R, Agostini F, Bugianesi E. Nonalcoholic fatty liver disease and the metabolic syndrome. Curr Opin Lipidol 2005;16:421–427. [DOI] [PubMed] [Google Scholar]
- 48.Marchesini G, Bugianesi E, Forlani G, Cerrelli F, Lenzi M, Manini R, Natale S, et al. Nonalcoholic fatty liver, steatohepatitis, and the metabolic syndrome. Hepatology 2003;37:917–923. [DOI] [PubMed] [Google Scholar]
- 49.Sierksma A, Patel H, Ouchi N, Kihara S, Funahashi T, Heine RJ, Grobbee DE, et al. Effect of moderate alcohol consumption on adiponectin, tumor necrosis factor-alpha, and insulin sensitivity. Diabetes Care 2004;27:184–189. [DOI] [PubMed] [Google Scholar]
- 50.Xi B, Veeranki SP, Zhao M, Ma C, Yan Y, Mi J. Relationship of Alcohol Consumption to All-Cause, Cardiovascular, and Cancer-Related Mortality in U.S. Adults. J Am Coll Cardiol 2017;70:913–922. [DOI] [PubMed] [Google Scholar]