A study from the United States revealed that the peak prevalence of hospitalization for alcohol-related conditions was between 45 and 69 years of age[15]. In contrast, the prevalence of NAFLD in a large Japanese study peaked at 40-49 years of age in males and at age 60-69 years in females[12]; however, a large study in China found that the prevalence of NAFLD in both males and females peaked at 60-69 years of age[16]. In addition, NAFLD is commonly observed in children; a recent autopsy study from the United States revealed that approximately 10% of children had NAFLD[17].

ALD is predominantly observed in males. For example, a cross-sectional study revealed that the male to female ratio of patients with alcoholic liver cirrhosis was 9:1[18]. A nationwide study in the United States that assessed patients discharged from the hospital following a diagnosis of ALD found that 4.5 per 100000 persons had acute alcoholic hepatitis, with a male to female ratio of 1.83:1, and that 13.7 per 100000 persons had chronic alcoholic hepatitis with cirrhosis, with a male to female ratio of 2.64:1[19]. Studies have also revealed a male predominance in the prevalence of NAFLD. For example, a study from the United States found that 58.9% of patients with NAFLD were male[11], and a large study from Japan reported that the prevalence rates of NAFLD in males and females were 41.0% and 17.7%, respectively. The prevalence rates of NAFLD were higher at all ages in males than in females, but the rate gradually increased in females with age, from 3.3% in the second decade of life to 31.3% beyond the sixth decade[12]. Similar trends were observed in the Chinese population[20].

Studies have reported ethnic differences in the prevalence rates of ALD and NAFLD. For example, the prevalence rate of alcoholic liver cirrhosis was higher among South Asian males (32.8%) than among Afro-Caribbean males (1.1%) in the United Kingdom[21]. A recent large study of patients with NAFLD indicated that the prevalence of this disease was higher in Hispanics (24.2%) than in non-Hispanic Whites (17.8%) and non-Hispanic Blacks (13.5%)[22].

As mentioned above, ALD and NAFLD have similar pathological spectra, from simple steatosis to liver cirrhosis, which makes confident differentiation of the two diseases difficult. However, the following histological findings are helpful in the differential diagnosis[7]. The fatty degeneration of liver cells occurs to a greater degree in NAFLD than in ALD. In contrast, inflammatory cell infiltration is more pronounced in ALD than in NAFLD. Furthermore, venous or perivenular fibrosis, phlebosclerosis, and (less commonly) lymphocytic phlebitis are more common in ALD than in NAFLD.

Clinically, the differentiation between ALD and NAFLD is usually performed by taking a history of a patient’s alcohol intake combined with laboratory and imaging examinations; however, the reliability of these methods may not be high[23]. Therefore, discriminant indices consisting of clinical parameters have been developed[24-26]. For example, the ALD/NAFLD index (ANI) is calculated from a patient’s gender, BMI, aspartate aminotransferase (AST)/alanine aminotransferase (ALT) ratio, and mean corpuscular volume (MCV)[25]. This index has exhibited a high diagnostic accuracy, with an area under the receiver operating characteristic curve (AUROC) of 0.983 (cut-off value, 0; sensitivity, 93.5%; specificity, 92.0%) in the derivation set and AUROCs of 0.974, 0.989, and 0.767 in the three validation sets. A recent validation study confirmed the high accuracy of the index[27]. However, the ANI may be less reliable in patients with end-stage liver disease because these patients frequently have an elevated MCV and an increased AST/ALT ratio[25]. Because the ANI was formulated based on data sets mainly obtained from Caucasians, this index may not be applicable to other ethnic groups, suggesting the need for indices specifically for these groups.

Following a differential diagnosis, the assessment of disease severity and stage is vital for optimal patient management. Although a liver biopsy is the gold standard, this procedure is invasive and costly and is accompanied by patient discomfort, which affects its feasibility. Non-invasive methods for assessing the stage of fibrosis have been recently developed, including fibrosis indices and transient elastography[28]. Many fibrosis indices have been devised based on clinical parameters, and the usefulness of these indices has been tested in patients with chronic liver diseases. The AST/platelet ratio index (APRI)[29], the FIB-4 index[30], and the FibroTest[31] have been validated for use in assessing the fibrosis stage in patients with both ALD and NAFLD. The AUROC of the FibroTest for the diagnosis of liver cirrhosis in patients with ALD was 0.95 (cut-off value, 0.7; sensitivity, 91%; specificity, 87%)[32], while a more recent study found that the AUROCs of the APRI, the FIB-4 index, and the FibroTest for liver cirrhosis in patients with ALD were 0.67, 0.80, and 0.94 (cut-off value, 0.7; sensitivity, 86.6%; specificity, 86.0%), respectively[33]. In patients with NAFLD, the AUROCs of the APRI and the FIB-4 index for the diagnosis of advanced fibrosis were 0.73 and 0.80 (cut-off value, 2.67; sensitivity, 33%; specificity, 98%), respectively[34], and in another study, the AUROCs for the APRI and the FIB-4 index were 0.82 (cut-off value, 1; sensitivity, 67%; specificity, 81%) and 0.87 (cut-off value, 3.25; sensitivity, 48%; specificity, 95%), respectively[35]. Another study in patients with NAFLD indicated that the AUROCs of the FibroTest for the diagnosis of advanced fibrosis in two independent cohorts were 0.92 (cut-off value, 0.7; sensitivity, 25%; specificity, 97%) and 0.81(cut-off value, 0.7; sensitivity, 25%; specificity, 99%), respectively[36].

Transient elastography has also been shown to be useful in assessing the disease stage. For example, the accuracy of this method in assessing the fibrosis stage was validated for ALD, with AUROCs of 0.94 (cut-off value, 11.60 kPa; sensitivity, 87%; specificity, 89%) for advanced fibrosis and 0.87 (cut-off value, 22.70 kPa; sensitivity, 84%; specificity, 83%) for liver cirrhosis[37]. The AUROCs of transient elastography, the APRI, and the FIB-4 index for the diagnosis of advanced fibrosis in patients with NAFLD were 0.93 (cut-off value, 9.6 kPa; sensitivity, 63.6%; specificity, 83.7%), 0.74 (cut-off value, 0.5; sensitivity, 65.1%; specificity, 72.3%), and 0.80 (cut-off value, 2.67; sensitivity, 20.6%; specificity, 95.5%), respectively[38]. Generally, transient elastography cannot differentiate simple hepatic steatosis from mild hepatic fibrosis but can assess the degree of hepatic fibrosis.

Recently, a novel measurement method for assessing the hepatic steatosis grade was developed, which has been designated as the controlled attenuation parameter (CAP)[39]. This method is based on transient elastography and permits a simultaneous evaluation of the fibrosis stage and the steatosis grade. Studies have demonstrated the usefulness of CAP in quantitatively assessing the steatosis grade with a high accuracy[40-42]. A large cohort study found that the AUROCs of the CAP for the diagnosis of steatosis > 10%, steatosis > 33%, and steatosis > 66% were 0.79, 0.84, and 0.84, respectively, in 440 patients who underwent a liver biopsy[42]. This study also found that elevated CAP values were significantly associated with excessive alcohol intake, indicating a possible clinical application of CAP in the diagnosis and management of ALD.

Further validation studies are required to establish the usefulness of these non-invasive methods in patients with ALD and NAFLD.

Scoring systems specific for ALD or NAFLD have been developed to assess disease severity and stage and/or to predict patient outcomes. Scoring systems for ALD[43,44] include Maddrey’s discriminant function (mDF), the glasgow alcoholic hepatitis score (GAHS), the Age-Bilirubin-international normalized ratio (INR)-Creatinine (ABIC) score, and the Lille score. Studies have demonstrated that these systems are useful in predicting the short-term survival of patients with alcoholic hepatitis[44,45]. The Brunt score[46] and the NAFLD activity score (NAS)[47] are histological scoring systems for NAFLD. Studies have suggested that the NAS has an excellent ability to differentiate simple hepatic steatosis from nonalcoholic steatohepatitis (NASH)[48]. Other scoring systems for the assessment of NAFLD stage, such as the NAFLD fibrosis score and the BARD score, have shown promising results[49].

The risks for the development and progression of ALD are increased as the intake of alcohol increases[18,57,58]. The type of beverage may also modify ALD progression. In a pooled cross-sectional time-series analysis, the consumption of spirits, rather than beer or wine, was associated with mortality from cirrhosis in primarily beer-drinking countries[59]. Drinking patterns are also factors associated with ALD. Studies have demonstrated that daily or near-daily heavy drinking, not episodic or binge drinking, is closely associated with ALD development[18,60]. Moreover, it was demonstrated that alcohol intake outside of mealtimes and the intake of multiple, different beverages increase the risk of developing ALD[18].

Similarly, a high total energy intake is positively associated with the development of NAFLD[50], and specific dietary components affect the pathogenesis of this disease. A cross-sectional study found that a greater intake of soft drinks and meats was associated with an increased risk of NAFLD[61]. Fructose may contribute to disease progression[62] and disease development[63], whereas the ingestion of n-3 polyunsaturated fatty acids may decrease intrahepatic fat deposition[64].

Age: The ability to metabolize alcohol decreases with age because aging causes reductions in liver size and blood flow to the liver and decreases the activity of enzymes related to alcohol metabolism, such as alcohol dehydrogenase, acetaldehyde dehydrogenase, and cytochrome P-4502E1[65,66]. Thus, the livers of older subjects become more vulnerable to alcohol toxicity. Indeed, studies in patients with ALD revealed positive correlations between age and advanced fibrosis or liver cirrhosis[58,67]. Similarly, a systematic review of NASH patients revealed that aging was significantly associated with increased fibrosis[68]. The aging of the human population worldwide will likely have an impact on the progression of ALD and NAFLD.

Gender: Females have been found to be more susceptible to alcohol-induced liver damage than males. For example, a large prospective study found that the amount of alcohol intake at which the relative risk of ALD is greater than 1 was lower in females (7-13 beverages per week) than in males (14-27 beverages per week)[57], and multivariate analyses indicated that female gender was significantly associated with increased fibrosis in patients with ALD[58,67]. Studies of patients with simple alcoholic steatosis found that females were at a higher risk for cirrhosis than males[52,55]. In contrast, although a gender difference was observed in the prevalence of NAFLD, no significant gender difference was observed in the risk of increased fibrosis among patients with NASH[68].

Obesity, metabolic syndrome, and type 2 diabetes: Obesity predisposes the individual to the development of both ALD[58] and NAFLD[22]. Although a higher BMI was significantly associated with increased fibrosis in ALD patients[58,67], obesity was not significantly associated with increased fibrosis in patients with NAFLD[68]. Insulin resistance (IR) is a key factor in the development of metabolic syndrome[69] and is largely responsible for the development of type 2 diabetes. Recent studies have demonstrated the close relationship between ALD and IR[70] and have suggested that metabolic syndrome and type 2 diabetes are associated with the development of ALD[71]. In contrast, IR or metabolic syndrome and type 2 diabetes were found to be highly associated with NAFLD development[22,71,72] but not NAFLD progression[68].

Ethnicity: Ethnicity may influence the pathogenesis of ALD and NAFLD. For example, a large cross-sectional study revealed that, among current drinkers, the activities of the liver enzymes AST and γ-glutamyl transpeptidase (GGT) were likely to be 2-fold higher in black non-Hispanic and Mexican Americans than in white non-Hispanic Americans[73]. Another study suggested that African-Americans were more susceptible to alcohol-induced hepatotoxicity than whites[74]. Although the homeostasis model assessment of insulin resistance (HOMA-IR) was not a significant risk factor for NASH among Latinos (OR = 0.93; 95%CI: 0.85-1.02), the HOMA-IR was significant among non-Latino whites (OR = 1.06; 95%CI: 1.01-1.11), which suggests that ethnicity may modulate the effects of IR on the risk of NASH[75].

Genetic factors: Considerable effort has been expended to identify genetic factors that contribute to the pathogenesis of ALD and NAFLD[76,77]. Close links have been found between single nucleotide polymorphisms (SNPs) in DNA and the susceptibility to and severity of many diseases. The patatin-like phospholipase domain-containing 3 gene, PNPLA3, encodes adiponutrin[78], a protein thought to play an important role in lipid metabolism, such as the hydrolysis of triacylglycerols[79]. A recent genome-wide association study (GWAS) of NAFLD suggested that a PNPLA3 SNP, I148M (rs738409 C/G), is associated with increased hepatic fat content and hepatic inflammation[80]. Moreover, a recent meta-analysis[81] confirmed that this SNP strongly correlates with both hepatic fat content (patients with the GG genotype had 73% more hepatic fat than those with the CC genotype) and disease progression (the pooled ORs of the GG vs the CC genotype were 3.25 for having higher necroinflammatory scores and 3.26 for developing fibrosis). Notably, these findings were consistent across ethnic groups. Similar results were obtained in patients with ALD[82-84]. For example, a multivariate analysis revealed that the PNPLA3 SNP was more frequently found in ALD patients than in controls (OR = 1.54) and that this SNP was the strongest independent predictor of the progression of alcoholic liver cirrhosis (OR = 2.08)[84]. These findings regarding the PNPLA3 SNP may explain, at least in part, the ethnic differences in the susceptibility to and prevalence rates of ALD and NAFLD. For example, Hispanics have been reported to be more susceptible to ALD[73] and have the highest prevalence rate of NAFLD[11]; this is consistent with the frequency of the PNPLA3 SNP, which is highest among Hispanics[80].

Recent studies have found other PNPLA3 SNPs associated with NAFLD[85,86]. A recent GWAS of Japanese NAFLD patients identified rs2896019 and rs381062, SNPs that the investigators suggested are associated with the NAFLD grade and/or stage[86]. Additionally, this GWAS also found SAMM50 and PARVB SNPs, both of which were considered to be involved in the second hit of NAFLD, leading to a shift from simple steatosis to NASH. In contrast, a GWAS of Caucasian NAFLD patients identified FDFT1 and COL13A1 SNPs[87]. Another GWAS of Caucasian NAFLD patients identified NCAN and PPP1R3B SNPs, as well as a PNPLA3 SNP (rs738409), and demonstrated that these SNPs correlate with hepatic steatosis. Furthermore, it was found that SNPs in the NCAN, GCKR, LYPLAL1, and PNPLA3 genes correlated with histological lobular inflammation/fibrosis[88]. In addition, accumulated data have suggested that some SNPs, other than the PNPLA3 SNP, may be associated with the pathogenesis of ALD. A meta-analysis revealed a close relationship between a TNFA SNP (rs361525) and ALD[89]. Moreover, several studies have found a higher frequency of the CD14 -159 C/T SNP in patients with alcoholic cirrhosis than in those without this disease[76]. Despite extensive genome-wide searching, the PNPLA3 rs738409 G genotype is one of the few confirmed genetic factors contributing to a patient’s susceptibility to both ALD and NAFLD and to the progression of both diseases. Because this SNP may be useful in screening and managing high-risk patients and may constitute a therapeutic target, further studies are warranted.

Although a recent study of patients with compensated alcoholic cirrhosis failed to identify any risk factors significantly associated with hepatocarcinogenesis[90], a study of patients with decompensated alcoholic cirrhosis found that older age was significantly associated with hepatocarcinogenesis[91]. Moreover, the PNPLA3 SNP was reported to be closely associated with the susceptibility to hepatocellular carcinoma (HCC) in patients with ALD[92,93].

Older age and alcohol intake were found to be independent risk factors for hepatocarcinogenesis in patients with NASH-related cirrhosis[94]. A multivariate analysis of a large number of Japanese patients with NAFLD found that older age, type 2 diabetes, an elevated serum AST concentration, and a low platelet count were significantly associated with hepatocarcinogenesis[95].

A recent study compared hepatocarcinogenesis in patients with alcoholic liver cirrhosis and NASH-related cirrhosis. A multivariate analysis revealed that older age, an elevated serum GGT concentration, and a higher Child-Pugh score were risk factors for the development of HCC in patients with NASH-related cirrhosis, whereas no factor was found to be significantly associated with hepatocarcinogenesis in patients with alcoholic cirrhosis[96].

Persistent alcohol intake has been reported to predict hepatic decompensation in patients with alcoholic cirrhosis[90]. To our knowledge, however, no studies to date have assessed the risk factors for hepatic decompensation in patients with compensated NASH-related cirrhosis.

A long-term cohort study analyzed the predictors of mortality in patients with ALD and simple hepatic steatosis and reported that a low serum albumin concentration and severe steatosis were significantly associated with increased mortality[55]. Moreover, persistent alcohol intake was found to be the only factor predictive of mortality in patients with compensated alcoholic cirrhosis[90]. A long-term follow-up study, in which most of the patients suffered from decompensated alcoholic cirrhosis, found that older age, alcohol abuse, and elevated alkaline phosphatase levels were risk factors for mortality[97]. In another long-term follow-up study involving patients with decompensated alcoholic cirrhosis, older age, persistent alcohol intake, a low serum albumin concentration, and a higher baseline model for end-stage liver disease (MELD) score were predictive of mortality[98]. Similarly, older age and poorer liver function were significantly associated with increased mortality in patients with decompensated alcoholic cirrhosis[91]. More recently, an analysis of patients with advanced, non-decompensated ALD revealed that smoking, a low serum albumin concentration, persistent alcohol intake, and older age were associated with increased mortality[99].

In patients with NAFLD and simple hepatic steatosis, only a low serum albumin concentration was found to be significantly associated with patient mortality[55]. A long-term cohort study involving patients with NAFLD found that an AST/ALT ratio > 1 and older age were associated with overall mortality, whereas higher serum bilirubin concentrations and stage 4 fibrosis were associated with liver-related mortality[100]. Older age and accompanying type 2 diabetes were significantly associated with increased mortality in patients with biopsy-proven NAFLD, whereas NASH, older age and type 2 diabetes were risk factors for liver-related mortality[101]. A recent, large population-based study revealed that metabolic syndrome was independently associated with overall mortality, liver-specific mortality, and cardiovascular mortality and that older age, smoking, and black race were risk factors for overall mortality[102].

A large cohort study found that type 2 diabetes and/or IR were independent predictors of overall mortality in patients with both ALD and NAFLD. Furthermore, type 2 diabetes, IR, obesity, and metabolic syndrome were independent predictors of liver-related mortality in both patient subsets[103].

Excessive alcohol intake is highly associated with metabolic syndrome-related diseases, including hypertension, type 2 diabetes, and dyslipidemia[104-107], whereas NAFLD and metabolic syndrome are closely related[108]. Few studies to date have compared the associations of ALD and NAFLD with metabolic syndrome and/or related diseases. A recent study, however, found that an alcoholic fatty liver was more strongly associated with hypertension than a nonalcoholic fatty liver, whereas the latter was more strongly associated with dyslipidemia than the former[109]. Further studies are required to confirm these findings.

Although little is known regarding the relationships between ALD and cerebrovascular and cardiovascular diseases, excessive alcohol intake has been reported to enhance the risks of these diseases, whereas light-to-moderate alcohol intake lowers these risks[110-112]. In comparison, a systematic review confirmed that NAFLD is an independent risk factor for cardiovascular disease[113], but it remains unclear whether NAFLD is a risk factor for cerebrovascular disease.

Although alcohol per se is not a carcinogen, excessive alcohol intake has been associated with HCC and other cancers, including oral cavity, pharyngeal, laryngeal, esophageal, colorectal, and female breast cancer[96,114]. A recent systematic review suggested a close relationship between NAFLD and colorectal cancer[113].

The cumulative incidence rates of HCC were 6.8% at 10 years in ALD patients with compensated cirrhosis[90] and 7.1% at 5 years in patients with decompensated cirrhosis[91]. In contrast, a systematic review found that the cumulative incidence rate of HCC at up to 20 years in patients with NAFLD or NASH, of whom few or none had cirrhosis, was 0%-3%[115]. In patients with NASH and cirrhosis, the cumulative rates of HCC ranged from 2.4% over 7 years to 12.8% over 3 years[115]. A large Japanese study of patients with NAFLD found that the annual rate of new HCC cases was 0.043%[95]. More recently, the 5-year cumulative rates of HCC were found to be comparable in patients with alcoholic (12.5%) and NASH-related (10.5%) cirrhosis[96].

Collectively, the incidence of HCC seems to be somewhat higher in patients with ALD than in patients with NAFLD. Most (88.0%-94.7%) ALD-related HCCs arise from cirrhosis, whereas 41.7%-75.0% of NAFLD-related HCCs arise from a non-cirrhotic liver[116-118]. Moreover, there was no difference in age between the patients with ALD-related HCC and those with NAFLD-related HCC, although the percentage of females was lower in the former than in the latter[116,119].

To our knowledge, no studies to date have compared the outcomes of patients with ALD- and NAFLD-related HCC. A study of patients who had non-hepatitis B/non-hepatitis C HCC found that the survival rates were similar between the patients with ALD-related HCC and those with non-ALD-related HCC, whereas the tumor recurrence rate was higher in the ALD group[120]. A comparison of the outcomes of curative treatment for HCC in patients with NASH and those with ALD or hepatitis C-related liver disease revealed that the cause of liver disease did not influence the recurrence-free survival according to a multivariate analysis, whereas NASH was significantly associated with an improved overall survival[119].

A study of the natural history of alcoholic cirrhosis found that the cumulative 10 year hepatic decompensation rate was 37.4%[90]. During the follow-up, only approximately one-fourth of these patients abstained from alcohol. In contrast, 45% of patients with compensated NASH-related cirrhosis developed hepatic decompensation during a follow-up of 10 years[121]. Ascites, a manifestation of hepatic decompensation, is frequently observed in patients with alcoholic cirrhosis[90,98,122], but the incidence of ascites is relatively low in patients with NASH-related cirrhosis[121].

The cumulative 5-year survival rates of patients with ALD were reported to be 83.3% in patients with simple hepatic steatosis and 74.9% in those with alcoholic steatohepatitis[52]. Furthermore, the cumulative 5-year survival rates in patients with compensated and decompensated alcoholic cirrhosis were reported to be 83.9%[90] and 71.3%[91], respectively. Abstinence from alcohol has the potential to increase the survival of patients with alcoholic cirrhosis[123]. In a study of the survival of patients with decompensated alcoholic and hepatitis C-related cirrhosis, a multivariate analysis revealed that the cause of liver disease did not affect survival[91].

A comparison of patients with NASH-related cirrhosis (mean Child-Pugh score 6.1) and hepatitis C-related cirrhosis (mean Child-Pugh score 6.1) revealed that the cumulative 5-year survival rates were 75.2% and 73.8%, respectively[124]. A similar study examining the natural history of advanced fibrosis or compensated cirrhosis due to either NASH or hepatitis C found that the 10-year cumulative survival rates were 81.6% and 82.0%, respectively[100].

A recent study of patients with NASH-related or alcoholic cirrhosis who underwent liver transplantation found no significant differences in the post-transplant survival and cardiovascular mortality rates. However, cardiovascular causes of post-transplant death were more frequent in the NASH group, while malignancies were more frequent in the alcoholic group[125]. In addition, studies have reported that the overall survival after liver transplantation was excellent in patients with HCC and ALD or NAFLD[126,127].

The causes of death have been found to differ in patients with ALD and NAFLD. A long-term follow-up of patients with simple hepatic steatosis revealed that many of the patients with ALD and NAFLD died of extrahepatic, rather than hepatic, causes. For example, the main causes of death in ALD patients were arteriosclerosis (20%), liver cirrhosis (17%), unknown causes (16%), and extrahepatic cancers (14%), whereas the main causes of death in NAFLD patients were arteriosclerosis (38%), unknown causes (19%), extrahepatic cancers (17%), and infections (8%)[55]. Similarly, studies of patients with alcoholic cirrhosis indicated that a significant percentage of individuals died of extrahepatic causes, with the leading causes being liver failure (25.0%-36.0%), bacterial infection (11.5%-25.0%), extrahepatic cancers (8.0%-25.0%), and HCC (12.5%-13.0%)[90,98]. A population-based study found that ALD contributed to liver-related but not cardiovascular mortality[128]. Moreover, recent studies of patients with histologically proven NAFLD, of whom 41.6%-59.2% had NASH, found that the leading causes of death were cardiovascular disease (27.8%-28.2%), liver-related causes (15.4%-26.1%), and extrahepatic malignancy (15.7%-17.9%)[101,129]. Similar patterns were observed in other studies of patients with NAFLD, with the leading causes of death being extrahepatic malignancy (28.0%-33.3%), liver-related complications (12.8%-19.0%), and cardiovascular disease (19.0%-25.0%)[130-133]. In one study of patients with NASH-related cirrhosis, the major causes of death were infection (41.4%), cardiovascular disease (27.6%), and liver-related complications (24.1%)[121], whereas the causes were HCC (47.4%) and liver failure (31.6%) in a second study[124].

In summary, the mortality rates of patients with ALD and NAFLD were found to be similar, although the causes of death differed somewhat. When compared with patients with hepatitis C, larger percentages of patients with ALD and NAFLD died of extrahepatic causes; many ALD patients died of infection and extrahepatic cancers, and many NAFLD patients died of cardiovascular disease.