Hepatic and Extrahepatic Malignancies in Cirrhosis Caused by Nonalcoholic Steatohepatitis and Alcoholic Liver Disease

Abstract

Background

To clarify the carcinogenic factors associated with steatohepatitis, we investigated the characteristic features of hepatic and extrahepatic malignancies in patients with cirrhotic nonalcoholic steatohepatitis (NASH-LC) and cirrhotic alcoholic liver disease (ALD-LC).

Methods

A total of 72 patients with NASH-LC and 85 with ALD-LC (both biopsy-proven steatohepatitis without hepatocellular carcinoma [HCC]) were assessed with regard to the development of hepatic and extrahepatic malignancies. Risk factors for HCC were analyzed.

Results

During follow-up, 10 NASH-LC patients and 6 ALD-LC patients developed HCC. The 5-year HCC development rate was similar for these 2 groups, being 10.5% in the NASH-LC group and 12.3% in the ALD-LC group. After adjusting for age and gender, the HCC development rates were also similar. Risk factors for HCC in the NASH-LC group were older age, higher γ-GTP level, and higher Child–Pugh score as determined by Cox hazards analysis. Regarding risk factors in the ALD-LC group, no risk factor was found by Cox hazards analysis, although diabetes mellitus led to a significantly higher HCC rate by log-rank test (p = 0.013). Regarding extrahepatic cancer, only 1 NASH-LC patient (1.4%) developed endometrial cancer. In contrast, 7 ALD-LC patients (8.2%) had other cancers (p = 0.052).

Conclusions

Comparison between NASH-LC and ALD-LC revealed similar HCC development curves. However, the risk factors for HCC and extrahepatic malignancies differed between the 2 diseases. In ALD-LC, the incidences of HCC and extrahepatic cancer are similar. When treating LC patients with NASH or ALD, the risk factors and extrahepatic malignancies associated with ALD-LC should be assessed.

Hepatocellular carcinoma (HCC) is the fifth-most-common cancer worldwide and the third-most-common cause of cancer mortality (Bosch et al., 2004). According to the most recent nationwide Japanese registration data, primary liver cancer ranks fourth in men and sixth in women as a cause of death from malignant neoplasm (Health and Welfare Statistics Association Japan, 2009). With regard to underlying liver disease, the latest nationwide report by the Liver Cancer Study Group of Japan, which registers such a report every 2 years, showed that hepatitis C virus (HCV) is the most common underlying liver disease in HCC (Ikai et al., 2007). HCV-related HCC accounts for 67% of all cases of HCC, followed by hepatitis B virus (HBV) at 16%. However, the incidence of HCV-related HCC has gradually decreased in recent years and that of HCC in cases of liver disease with a nonviral cause has gradually increased. According to our national survey of HCC, 15.8% of HCC cases were caused by nonviral liver diseases (Tokushige et al., 2011). Among the nonviral liver diseases, 45.6% were alcoholic liver diseases (ALDs), 32.3% were of unknown etiology, and 12.7% were nonalcoholic fatty liver diseases (NAFLDs). The increase in HCC related to NAFLD or ALD is becoming an especially urgent problem in Japan. It is important to accurately determine the development rate of HCC and the risk factors for HCC in both diseases.

Both ALD and NAFLD have similar liver histology, with changes such as steatosis or steatohepatitis, and NAFLD and ALD are distinguished only by alcohol consumption. However, the clinical features are clearly different between these 2 types of steatohepatitis (Scaglioni et al., 2011). Concerning extrahepatic malignancies, their development, such as prostate cancer, is becoming a problem in patients with NAFLD (Arase et al., 2012). It is well known that ALD occurs more frequently in western countries, but its prevalence is now increasing in Japan (Taniai et al., 2012). Alcohol is carcinogenic to humans, and alcohol intake is related to cancer at several sites, including the oral cavity, pharynx, esophagus, colorectum, and female breast (Boffetta and Hashibe, 2006). However, there have been few studies of extrahepatic malignancies in cirrhotic ALD (ALD-LC) patients, and it is unclear whether HCC or extrahepatic malignancies have a critical influence on the prognosis of ALD-LC.

Differences in the carcinogenic factors among steatohepatitis, nonalcoholic steatohepatitis (NASH), and ALD are also unknown. Some clinical entities of fatty liver disorder, including alcohol, insulin resistance, and others, have been proposed (Neuschwander-Tetri and Caldwell, 2003). When managing patients with NASH or ALD, it is important to know the carcinogenic features of both diseases.

To clarify the factors related to carcinogenesis in patients with steatohepatitis, we investigated carcinogenesis in the liver and other organs, as well as the risk factors for HCC associated with NASH and ALD.

Materials and Methods

Subjects

Between 1990 and 2010, 513 Japanese patients were diagnosed with biopsy-proven NAFLD at Tokyo Women's Medical University. Among them, 72 patients had LC (NASH-LC). Data on 85 patients with ALD-LC were collected retrospectively. None of the LC patients had HCC at the time of LC diagnosis. Patients with HCC at the beginning of this study were excluded. Informed consent was obtained in writing from each patient, and the study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki. Clinical data on the NASH-LC patients were collected prospectively.

Definition

The diagnosis of NASH was based on the following criteria: detection of steatohepatitis by liver biopsy; intake of <100 g of ethanol per week (as confirmed by the attending physician and family members in close contact with the patient); appropriate exclusion of other liver diseases (Brunt, 2004; Neuschwander-Tetri and Caldwell, 2003). All patients were negative for hepatitis B surface antigen and HCV antibody and/or HCV RNA as determined by polymerase chain reaction analysis.

Liver biopsy specimens were stained with hematoxylin and eosin, as well as with silver impregnation for reticulin fibers, Mallory stain for collagen fibers, Victoria blue stain for copper-binding protein and elastin fibers, and Perls’ Prussian blue stain for iron. All liver biopsy specimens were examined for fibrosis and for histological features of NASH, such as inflammation, ballooning degeneration, and Mallory bodies. Steatohepatitis was diagnosed by the presence of steatosis >10% lobular inflammation, and ballooning degeneration (Brunt et al., 1999).

ALD was diagnosed according to the modified criteria proposed by Takada and Tsutsumi (1995), which were habitual alcohol consumption of over 70 g daily and negativity for markers of HBV and HCV. In both groups, LC was diagnosed on the basis of clinicopathological findings.

HCC was diagnosed histologically or through consistent findings based on at least 2 of the following imaging modalities: ultrasonography (US), computed tomography (CT), magnetic resonance imaging (MRI), and selective hepatic arteriography.

Obesity was defined as body mass index >25 kg/m² according to the Japanese criteria. The diagnosis of type II diabetes mellitus (DM) was based on the American Diabetes Association criteria or current treatment for type II diabetes (Genuth et al., 2003). Dyslipidemia was diagnosed if the patient was being treated with lipid-lowering agents or had elevated levels of total cholesterol (>220 mg/dl) and/or triglycerides (>150 mg/dl) on at least 3 occasions. Hypertension (HT) was diagnosed if the patient was receiving antihypertensive therapy or had a blood pressure >140/90 mmHg on at least 3 occasions.

Management

A complete history was obtained, and physical examination was carried out for all patients. None of the patients were on hormone replacement therapy for menopause, and none had undergone liver transplantation. The following laboratory parameters were measured at the time of the diagnosis of LC and every 3 months during follow-up: aspartate aminotransferase (AST), alanine aminotransferase (ALT), total bilirubin, alkaline phosphatase, gamma-glutamyltranspeptidase, albumin, white blood cell count, red blood cell count, hemoglobin, platelet count, prothrombin time, alpha-fetoprotein (AFP), and des-gamma-carboxy prothrombin (DCP).

LC patients were monitored regularly (at least every 3 months) by clinical examination, liver function tests, measurement of AFP and DCP, and by US.

Imaging with CT, MRI, and/or selective hepatic angiography was performed if HCC was suspected on the basis of elevated tumor markers and/or US findings. Extrahepatic cancers were diagnosed by appropriate examinations such as endoscopy and CT.

We did not perform any special therapy for NASH. Advice regarding diet was given to NASH patients who were obese, while medications were given to treat concurrent DM, HT, and dyslipidemia. We did not perform any special therapy for ALD, but patients were advised to stop drinking.

Statistical Analysis

Analysis was performed using the SPSS statistical package (SPSS Inc., Chicago, IL). Mann–Whitney test or chi-square test was used to compare baseline variables between the NASH-LC and ALD-LC groups. The baseline date was the date of the initial diagnosis of LC. Patients in both groups were followed until they died and were censored at the time of their last clinic visit. The primary endpoint was the development of HCC. Time to failure analysis (Kaplan–Meier) was performed, and the log-rank test was used to compare the NASH and ALD groups as well as to make comparisons between LC patients with and without risk factors. To identify risk factors for the development of HCC, Cox proportional hazards analysis was used. The prevalence of extrahepatic cancer was analyzed by chi-square test. In all analyses, p < 0.05 was considered statistically significant.

Results

Table 1 lists the clinical data for the NASH-LC and ALD-LC groups. The gender difference was significant (percentage of women, NASH-LC: 68%; ALD-LC: 12%).The median age of the NASH-LC group was significantly higher than that of the ALD-LC group (NASH-LC: 63.5 ± 12.9 years; ALD-LC: 55.0 ± 10.7 years). The smoking rate of the ALD-LC group was significantly higher than that of the NASH-LC group, while the NASH-LC group had significantly higher prevalence rates of obesity, DM, dyslipidemia, and HT than the ALD-LC group. Regarding biochemical data, serum albumin, prothrombin time, and ALT levels were significantly higher in NASH-LC, while serum total bilirubin and γ-GTP levels were significantly higher in ALD-LC. Child–Pugh grade and platelet count were not significantly different between the 2 groups. The histological features are shown in Table S1. The grade of steatosis was higher in NASH-LC, but the grade of ballooning hepatocytes and the frequencies of pericellular and perivenular fibrosis were higher in ALD-LC.

Table 1. Clinical Profiles of the 2 LC Groups

	NASH-LC	ALD-LC	p-Value
Child, Child–Pugh classification; ALD-LC, cirrhotic alcoholic liver disease; ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; NASH-LC, cirrhotic nonalcoholic steatohepatitis. Data are expressed as median ± standard deviation (SD) or as the number of patients. Percentages are shown in parentheses.
No.	72	85
Gender (F/M)	49/23 (68%/32%)	10/75 (12%/88%)	0.002
Age	63.5 ± 12.9	55.0 ± 10.7	0.000
Follow-up (months)	50 ± 43.1	36 ± 40.5	NS
Smoking	16 (22.2%)	57 (67.1%)	0.000
Obesity	53 (74%)	15 (17.6%)	0.000
BMI (kg/m^2)	27.7 ± 5.3	22.2 ± 3.2	0.000
Diabetes mellitus	42 (58%)	30 (35%)	0.004
Dyslipidemia	30 (42%)	14 (16%)	0.0005
Hypertension	39 (54%)	20 (23.5%)	0.00001
Albumin (g/dl)	3.8 ± 0.69	3.3 ± 0.65	0.002
Total bilirubin (mg/dl)	0.8 ± 0.81	1.1 ± 2.36	0.005
AST (IU/l)	50.5 ± 42.3	55 ± 208.3	NS
ALT (IU/l)	46.0 ± 43.2	31 ± 91.6	0.006
γ-GTP (IU/l)	80.5 ± 68.7	159 ± 313.4	0.000
Platelet count (×10^4/μl)	11.8 ± 6.1	11.8 ± 7.1	NS
Prothrombin time (%)	81.0 ± 15.8	68.0 ± 20.9	0.005
Child (A/B or C)	56/16 (78%/22%)	57/28 (67%/33%)	NS

During follow-up, 10 NASH-LC patients and 6 ALD-LC patients developed HCC, but there were no cases of cholangio-cellular carcinoma. The clinical data of the 10 HCCs from NASH-LC and 6 HCCs from ALD-LC are shown in Table S2. There was no difference between the 2 groups. As all LC patients were checked every 3 months, most HCCs were found at a relatively early stage. Representative images of NASH-HCC and ALD-HCC are shown in Fig. S1.

Figure 1 A shows the development rate of HCC in the NASH-LC and ALD-LC groups. The curves of the 2 groups were similar, with the 5-year HCC development rate being 10.5% in the NASH-LC group and 12.3% in the ALD-LC group. After adjusting for age and gender (ALD-LC, n = 32; mean age, 55.8 ± 8.2 years; male, 69%; NASH-LC, n = 32; median age, 55.6 ± 13.7 years; male, 69%), the HCC rates were also quite similar (Fig. 1 B).

Figure 1. Development of HCC in NASH-LC and ALD-LC groups. (A) HCC rate in all patients. HCC incidence was similar. The 5-year HCC development rate was 10.5% in the NASH-LC group and 12.3% in the ALD-LC group. (B) HCC rate in adjusted patients. After adjusting for age and gender, HCC rates were still similar. ALD, alcoholic liver disease; HCC, hepatocellular carcinoma; NASH, nonalcoholic steatohepatitis.

By Cox hazards model, 3 significant risk factors for HCC development, namely older age (p = 0.024), higher serum γ-GTP level (p = 0.023), and higher Child–Pugh score (p = 0.006), were identified in NASH-LC (Table 2A). On the other hand, no significant risk factor for HCC was found in the ALD-LC group.

Table 2. Risk Factors for HCC in NASH-LC and ALD-LC Groups According to Cox Hazards Model

	p-Value	HR	95% CI
ALD-LC, cirrhotic alcoholic liver disease; ALT, alanine aminotransferase; ALP, alkaline phosphatase; AST, aspartate aminotransferase; BMI, body mass index; DM, diabetes mellitus; HCC, hepatocellular carcinoma; NASH-LC, cirrhotic nonalcoholic steatohepatitis; HR, hazard ratio; HT, hypertension; Varies, esophageal varies. There was no significant risk factor in ALD-LC.
(A) NASH-LC
Factors: age, gender, BMI, ascites, Varies, encephalopathy, DM, HT, dyslipidemia, total bilirubin, albumin, AST, ALT, ALP, γ-GTP, platelets, prothrombin time, Child–Pugh score
Age	0.024	1.12	1.014 to 1.226
γ-GTP	0.023	1.01	1.002 to 1.022
Child–Pugh score	0.006	3.09	1.374 to 6.934
(B) ALD-LC
Factors: age, gender, alcohol intake, BMI, DM, HT, ascites, Varies, encephalopathy, dyslipidemia, total bilirubin, albumin, AST, ALT, ALP, γ-GTP, platelets, prothrombin time, Child–Pugh score
Age	0.434	2.26	0.29 to 17.6
DM	0.705	4.239	0.00 to >10,000

These risk factors for HCC in the NASH-LC group were confirmed by Kaplan–Meier analysis using the log-rank test (Fig. 2 A–C). NASH-LC patients aged ≥58 years had a significantly higher HCC rate compared with NASH-LC patients aged ≤57 years (p = 0.017). In addition, NASH-LC patients with a serum γ-GTP level ≥115 U/l also had a higher HCC rate compared with those with γ-GTP ≤114 U/l (p = 0.077). Furthermore, NASH-LC patients with Child–Pugh B or C showed a significantly higher HCC rate compared with patients with Child–Pugh A (p = 0.037).

Figure 2. Comparison of HCC rates by risk factors. (A) Association between age and HCC rate in NASH-LC. NASH-LC patients aged ≥58 years had a significantly higher HCC rate compared with NASH-LC patients aged ≤57 years (p = 0.017). (B) Association between γ-GTP and HCC rate in NASH-LC. NASH-LC patients with serum γ-GTP ≥115 U/l had a higher HCC rate compared with NASH-LC patients with γ-GTP ≤114 U/l (p = 0.077). (C) Association between Child–Pugh grade and HCC rate in NASH-LC. NASH-LC patients with Child–Pugh B or C had a significantly higher HCC rate compared with NASH-LC patients with Child–Pugh A (p = 0.037). (D) Association between diabetes mellitus (DM) and HCC rate in ALD-LC. ALD-LC patients with DM had a significantly higher HCC rate compared to those without DM as determined by Kaplan–Meier analysis (p = 0.013). ALD-LC, cirrhotic alcoholic liver disease; HCC, hepatocellular carcinoma; NASH-LC, cirrhotic nonalcoholic steatohepatitis.

Regarding risk factors for HCC in the ALD-LC group, patients with DM showed a significantly higher HCC development rate compared with patients without DM as determined by Kaplan–Meier analysis using the log-rank test (p = 0.013). These Kaplan–Meier analyses in adjusted patients (n = 32) showed the same tendency, but without significant difference.

Regarding extrahepatic malignancies, only 1 NASH-LC patient developed endometrial cancer. In the ALD-LC group, 2 patients had esophageal cancer, 2 had gastric cancer, 1 had pancreatic cancer, 1 had pharyngeal cancer, and 1 had both esophageal and pharyngeal cancer. Thus, a total of 7 ALD-LC patients had extrahepatic cancers (Table 3), and the ALD-LC group showed a higher malignancy rate compared with the NASH-LC group (NASH-LC, 1.4%; ALD-LC, 8.2%, p = 0.052). The incidences of HCC and extrahepatic malignancies appeared to be similar in ALD-LC. After being adjusted for age and gender, there was no extrahepatic tumor in NASH-LC. In contrast, the adjusted ALD-LC group had 2 extrahepatic tumors, 1 gastric cancer, and 1 esophageal cancer. Regarding extrahepatic tumors between the 2 adjusted groups, there was no significant difference, but the same tendency was observed.

Table 3. Extrahepatic Cancers

ALD-LC, cirrhotic alcoholic liver disease; NASH-LC, cirrhotic nonalcoholic steatohepatitis. ^a One patient had both esophageal and pharyngeal cancer. ^b p = 0.052 by chi-square test.
(A) NASH-LC (n = 72)
Endometrial cancer	1
Total	1 (1.4%)
(B) ALD-LC (n = 85)
Gastric cancer	2
Esophageal cancer	3a
Pharyngeal cancer	2a
Pancreatic cancer	1
Total	7 (8.2%)b

Discussion

ALD-LC patients without biopsy were excluded to avoid 2-group bias, because NASH is diagnosed by liver biopsy. Baseline data of the NASH-LC and ALD-LC groups differed with respect to gender and age at diagnosis. Concerning gender difference, the low number of women in the ALD-LC group can be attributed to the very low alcohol consumption of older Japanese women. These differences were also observed for patients with HCC related to NAFLD and ALD in our national survey (Tokushige et al., 2011). Age at ALD-LC diagnosis was younger than that at NASH-LC diagnosis, although the reason for this difference was not clear. It is possible that the severity of liver damage and the progression rate differ between the 2 diseases. Another factor may be that obesity and metabolic syndrome were uncommon in Japan before 2000. In the future, the age at NASH-LC diagnosis might become younger by 10 or 20 years because of the dramatic increase in obesity and metabolic syndrome in Japan. The difference in smoking rates might be associated with gender difference, as smoking rate is not common among elderly Japanese women. The meanings of age and gender differences in LC in Japan will be clarified in the future.

This is the first report about the assessment of hepatic and extrahepatic malignancies in NASH-LC and ALD-LC in Japan. The present study showed that the development rates of HCC of these 2 conditions were quite similar. In contrast, the risk factors for HCC and the rates of extrahepatic malignancies were strikingly different. Although our NASH-LC patients were older and included many females, analysis in which adjustments were made to compensate for these differences yielded the same HCC development rate. To compare the real HCC rates, larger patient populations of the same gender and similar age will be needed. Ascha and colleagues (2010) and Sanyal and colleagues (2006) reported that the 5-year HCC rate in NASH-LC patients was 10 to 13.6% in the United States. Therefore, the HCC rate of NASH-LC patients is almost the same in the United States and Japan. We had previously compared NASH-LC to LC induced by HCV infection (Yatsuji et al., 2009). It was found that the HCC development rate was lower in NASH-LC than in LC associated with HCV, while HCC and Child–Pugh B or C were important prognostic factors for NASH-LC (Yatsuji et al., 2009). The fact that steatohepatitis caused by insulin resistance and that caused by alcohol showed similar HCC development rates is very important in assessing the carcinogenesis of ALD and NASH. Hashimoto and colleagues (2012) reported that inflammatory cytokines, oxidative stress, and insulin resistance might induce hepatic carcinogenesis in NASH. Inflammatory cytokines and oxidative stress were shared in ALD-LC. Therefore, it was considered that a common mechanism might be responsible for inducing almost the same hepatic carcinogenesis between NASH-LC and ALD-LC.

In this study, older age, higher γ-GTP level, and higher Child–Pugh grade were detected as risk factors in NASH-LC patients. Old age is well known as an important risk factor for many kinds of malignancies. Several authors have reported that the serum γ-GTP level is influenced by oxidative stress (Abdul-Rasheed et al., 2010), and oxidative stress with the release of reactive oxygen species is likely to contribute to the development of both NASH and HCC (Hashimoto and Tokushige, 2012). Oxidative stress may favor tumorigenesis by promoting inflammation and cell proliferation, or it may directly induce cancer-promoting gene mutations (Hu et al., 2002). Fibrosis is the single most important risk factor for HCC. Ratziu and colleagues (2002) reported that patients with suspected burnt-out NASH had a high rate of HCC (27%). It was thought that our NASH-LC patients with Child–Pugh B or C were close to the burnt-out state. Thus, it is reasonable that NASH-LC patients with Child–Pugh B or C had a significantly higher HCC rate.

DM was a risk factor for HCC in ALD-LC patients according to the log-rank test. We could not detect risk factors using Cox hazards analysis, perhaps because it included too many factors for analyzing 85 samples. In addition, in ALD-LC, 88% were males. This obvious male gender prevalence made it difficult to find other significant factors. Several authors have reported that DM is a risk factor for HCC and other malignancies (El-Serag et al., 2004; Jiang et al., 2011). In the case of NASH, patients without HCC often have DM, which may explain why it was not detected as a risk factor for HCC.

Regarding extrahepatic cancer, ALD-LC patients developed esophageal, pharyngeal, gastric, and pancreatic cancers. In particular, Asians often show slower conversion of acetaldehyde to acetic acid, making it even more important to check for extrahepatic cancers in Asian ALD-LC patients. As all 7 ALD-LC patients with extrahepatic cancer were smokers, a synergistic effect of alcohol and smoking might be associated with extrahepatic carcinogenesis. Six ALD-LC patients developed HCC during follow-up, meaning that the incidences of HCC and extrahepatic cancer were similar in ALD-LC. In contrast, only 1 NASH-LC patient had extrahepatic (endometrial) cancer, suggesting that the risk of cancer outside the liver is relatively low in NASH-LC.

It is well known that alcohol intake is a significant risk factor for esophageal and pharyngeal cancer (Boffetta and Hashibe, 2006). Synergy among alcohol intake, genetic background, and tobacco smoking has been reported for squamous cell carcinoma of the esophagus (Yokoyama et al., 2008). As for gastric cancer, a positive association with alcohol intake was found in 2 of 12 cohort studies and 8 of 40 case–control studies (Franceschi and La Vecchia, 1994). However, the prevalence of Helicobacter pylori infection was not clarified in these studies. Most investigations of an association between alcohol and pancreatic cancer have come to a negative conclusion (Bagnardi et al., 2001). However, in our study, the 1 case in which pancreatic cancer developed was complicated with alcoholic chronic pancreatitis.

A limitation of this paper was the drop-out rate of ALD-LC patients during follow-up. It is very difficult to follow-up ALD-LC patients perfectly owing to their unpredictable behavior. In conclusion, a comparison of NASH-LC and ALD-LC in terms of HCC incidence revealed similar carcinogenesis curves. However, the risk factors and rate of extrahepatic cancer differed between these 2 diseases. The incidences of HCC and malignancies were similar, especially in ALD-LC. When managing LC patients with NASH or ALD, these risk factors and the extrahepatic malignancies associated with ALD-LC should be considered.

Acknowledgment

This work was supported in part by a grant-in-aid from the Ministry of Health, Labour and Welfare of Japan.