Risk Factors for Hepatocellular Carcinoma in Patients With Alcoholic or Viral C Cirrhosis

CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2006;4:1062–1068

Risk Factors for Hepatocellular Carcinoma in Patients With Alcoholic or Viral C Cirrhosis GISÈLE N’KONTCHOU,* JACQUES PARIES,‡ MYINT TIN TIN HTAR,* NATHALIE GANNE–CARRIE,* LYDIE COSTENTIN,* VÉRONIQUE GRANDO–LEMAIRE,* JEAN–CLAUDE TRINCHET,* and MICHEL BEAUGRAND* *Department of Hepatogastroenterology and ‡Department of Public Health, Hôpital Jean Verdier, Assistance Publique-Hôpitaux de Paris, UPRES EA 3409, UFR SMBH, Université Paris 13, Bondy, France

Background & Aims: Influence of being overweight and diabetes mellitus on the occurrence of hepatocellular carcinoma (HCC) in patients with cirrhosis has not been evaluated prospectively. The aim of this study was to show the predictive value of these factors in a cohort of 771 patients with well-compensated alcohol- or hepatitis C (HCV)-related cirrhosis who were screened prospectively for HCC. Methods: The predictive value for HCC occurrence was assessed by using the log-rank test and the Cox proportional hazards model. At enrollment, the mean age was 61.4 ⴞ 10 years and 431 patients were men. Cirrhosis was caused by alcohol (n ⴝ 478), HCV (n ⴝ 220), or the association of both factors (n ⴝ 73). The mean body mass index (BMI) was 25.4 kg/m2 and 231 patients were diabetic. Results: During a mean follow-up period of 4.2 ⴞ 3 years, 220 patients developed HCC. In univariate analysis, a BMI of 25 kg/m2 or more, diabetes, male sex, age older than 60 years, and HCV infection were risk factors for HCC. In multivariate analysis, predictive factors were a BMI between 25–30 kg/m2 (hazard ratio [HR], 2.0; 95% confidence interval [CI], 1.4 –2.7), BMI of 30 kg/m2 or more (HR, 2.8; 95% CI, 2.0 – 4.0), diabetes (HR, 1.6; 95% CI, 1.2–2.1), age 60 –70 years (HR, 2.4; 95% CI, 1.3– 4.3), age older than 70 years (HR, 3.0; 95% CI, 1.7–5.5), male sex (HR, 2.0; 95% CI, 1.4 –2.7), HCV (HR, 1.6; 95% CI, 1.1–2.2), and mixed (HR, 2.6; 95% CI, 1.7– 4.0) etiology. We found a positive linear relationship between BMI level and HCC incidence during follow-up evaluation. Conclusions: Overweight and diabetes mellitus are associated with an increased risk of HCC occurrence in patients with HCVor alcohol-related cirrhosis.

verweight, an increasingly frequent condition in Western countries, is associated closely with type 2 diabetes and with insulin resistance syndrome.1–3 Evidence continues to mount linking this metabolic pattern to an increased risk of cancer in various organs such as the kidney, uterus, breast, gallbladder, pancreas, and colon.4 – 8 It already is well known that overweight and type 2 diabetes could induce liver lesions potentially leading to

O

cirrhosis and hepatocellular carcinoma (HCC).9,10 These conditions also could potentiate the risk of cirrhosis in alcoholic patients and in patients with hepatitis C virus (HCV) infection,11–13 which are the leading causes of cirrhosis in most Western countries.14,15 In addition, case-control studies have suggested that obesity and diabetes mellitus are risk factors for the occurrence of HCC in patients with or without cirrhosis.16 –18 A higher body mass index (BMI) has been reported in transplanted patients with HCC than in those without.19 Diabetes is also more frequent in patients with cirrhosis and HCC than in those without HCC.20 However, the role of these factors in patients with cirrhosis has not been shown prospectively. The purpose of this study was to assess the relative risk for the occurrence of HCC by being overweight or diabetic in a large cohort of patients with alcoholic and HCV cirrhosis who were followed-up prospectively.

Patients and Methods Patients We included in this study all patients with compensated alcoholic or viral C cirrhosis who were screened prospectively for HCC from November 1994 to March 2004. Additional selection criteria were as follows: age 40 years or older, available recorded weight and height, absence of other causes of liver disease such as hepatitis B virus infection (negative serum hepatitis B surface antigen and anti-HBc antibody), or hemochromatosis, autoimmune hepatitis, primary biliary cirrhosis, primary sclerosing cholangitis. Patients were not included in case of human immunodeficiency virus infection, the presence of or suspected HCC, the presence of hepatic decompensation, particularly ascites, or other extrahepatic life-threatAbbreviations used in this paper: BMI, body mass index; CI, confidence interval; HCC, hepatocellular carcinoma; HCV, hepatitis C virus; HR, hazard ratio. © 2006 by the American Gastroenterological Association Institute 1542-3565/06/$32.00 doi:10.1016/j.cgh.2006.05.013

August 2006

RISK FACTORS FOR HCC WITH ALCOHOLIC OR VIRAL C CIRRHOSIS

ening disease at the time of inclusion. Patients belonging to Child–Pugh class C also were excluded. Cirrhosis was proven histologically in all cases. Patients were classified according to the cause of cirrhosis into the following categories: alcohol (daily ethanol intake ⱖ80 g in men and ⱖ60 g in women during a period longer than 10 years in the absence of another cause of cirrhosis), HCV (presence of serum anti-HCV antibodies and HCV RNA), and mixed when these 2 criteria were associated. Age, sex, cause of cirrhosis, diabetes status, and BMI were recorded at enrollment and collected retrospectively. Patients treated for diabetes or those with fasting blood glucose levels of 126 mg/dL or more were considered diabetic. Diabetes status was recorded as a binary parameter (yes/no). BMI was calculated as weight divided by squared height (kg/m2). No patients had ascites at inclusion.

Follow-up Evaluation HCC was screened by abdominal ultrasonography and determination of serum level of ␣-fetoprotein level performed every 3– 6 months. The diagnosis of HCC was assessed through histology or noninvasive criteria according to Barcelona conference recommendations.21 The duration of follow-up evaluation was defined as the time between the date of enrollment in the screening program and the date of HCC occurrence or the last ultrasonography scan in case of liver transplantation, death, or drop-out before the reference date of June 2004.

Statistical Analysis Data were expressed as mean ⫾ SD or percentages. The end point was the occurrence of HCC. The differences of variables between subgroups according to BMI or cause of liver disease were determined by analysis of variance. The univariate analysis was performed using the

1063

log-rank test and occurrence of HCC was estimated by the Kaplan–Meier method. All variables with a P value of less than .05 were selected for the Cox proportional hazard model for multivariate analysis. All possible interactions of variables in the final Cox model were tested. Finally, a composite index was established by combining 2 variables of significant interaction (BMI and diabetes). All probabilities were 2-sided with a 95% confidence interval. Analysis was performed with the SPSS statistical package (SPSS Inc., Chicago, IL) and Stat X (StatXact Cyten Software Corporation, Cambridge, MA).

Results Patient Characteristics A total of 771 of 795 patients selected were included (478 with alcoholic cirrhosis, 220 with HCV cirrhosis, and 73 with mixed causes). Twenty-four patients were excluded secondarily because of other neoplasms (9 patients), lack of data concerning height (10 patients), or weight (5 patient) at the time of inclusion. Patients’ characteristics are shown in Table 1. For the global population, the mean age was 61.4 ⫾ 10 years, the mean BMI was 25.4 ⫾ 4.0 kg/m2, and 231 (25%) patients were diabetic. Among them, 48 patients required insulin therapy but none was considered as having type 1 diabetes on the basis of their medical files. Follow-up Evaluation and Incidence of Hepatocellular Carcinoma The mean overall follow-up period was 4.2 ⫾ 3.0 years. At the reference date, 64 patients were lost to follow-up evaluation and were censored at the date of the

Table 1. Baseline Characteristics of Patients According to Cirrhosis as the Cause Cause Characteristic

Alcohol (n ⫽ 478)

HCV (n ⫽ 220)

Mixeda (n ⫽ 73)

P

Time since enrollment, yb Age, yb Age ⬍50 y 50–60 y 60–70 y ⬎70 y Male sex BMI, kg/m2b BMI, kg/m2 BMI ⬍25 BMI 25–30 BMI ⱖ30 Diabetes

4.1 ⫾ 4.1 59.7 ⫾ 9.7

4.7 ⫾ 3.7 66.4 ⫾ 9.7

4.0 ⫾ 3.2 57.8 ⫾ 10.7

NS ⬍.0001 ⬍.0001

88 (18%) 170 (36%) 139 (29%) 81 (17%) 336 (70%) 25.4 ⫾ 5.0

15 (7%) 38 (17%) 78 (36%) 89 (40%) 123 (56%) 25.6 ⫾ 4.4

22 (30%) 22 (30%) 18 (25%) 11 (15%) 63 (86%) 24.9 ⫾ 4.0

248 (52%) 160 (34%) 70 (15%) 142 (30%)

115 (52%) 74 (34%) 31 (14%) 69 (31%)

47 (64%) 15 (21%) 11 (15%) 20 (27%)

NOTE. N ⫽ 771. and alcohol. bMean ⫾ SD shown. aHCV

⬍.0001 NS

NS

1064

N’KONTCHOU ET AL

CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 4, No. 8

Table 2. Clinical Characteristics of Patients According to Body Mass Index 2

BMI ⬍25 kg/m

Mean time since enrollment, Mean age, ya Number (%) males Cause of cirrhosis Alcohol HCV Mixedb Diabetes

ya

2

2

BMI 25–30 kg/m

BMI ⱖ30 kg/m

N

Mean

N

Mean

N

Mean

410

22.06

249

27.2

112

33.8

4.3 ⫾ 4.1 61.2 ⫾ 10.6 167 (41%)

4.0 ⫾ 3.4 61.6 ⫾ 10.0 68 (27%)

4.5 ⫾ 4.2 61.8 ⫾ 9.9 40 (36%)

248 (60%) 115 (28%) 47 (12%) 104 (25%)

160 (64%) 74 (30%) 15 (6%) 88 (35%)

70 (63%) 31 (28%) 11 (9%) 39 (35%)

P NS NS .002 NS

.012

NOTE. N ⫽ 771. ⫾ SD. bHCV and alcohol. aMean

last abdominal ultrasonography control. A total of 220 patients (158 men and 62 women) developed HCC. HCC diagnosis was established by histology in 130 and by strictly applied noninvasive criteria in 90 patients with nodules or masses more than 2 cm in diameter. Overall, 135 patients received percutaneous ablative therapies (ethanol or radiofrequency ablation), 16 patients were treated by transplantation, 5 patients by hepatic resection, 6 by transcatheter arterial embolization, and 62 were not eligible for HCC treatment. The incidence rate (per 100 person-years) of HCC was 6.7% in surviving patients. The cumulative incidence of HCC was 20% at 5 years in the global population (16% in the alcohol

group, 22% in the HCV group, and 29% in the mixedcause group). During the same period, 224 patients died and death was considered to be related directly to HCC in 81 patients. Predictive Factors for Hepatocellular Carcinoma Occurrence The distribution of patients according to the World Health Organization’s classification of BMI was as follows: BMI of less than 25 kg/m2, 410 patients (53%); BMI of 25–30 kg/m2, 249 patients (32%); and BMI of 30 kg/m2 or more, 112 patients (15%). Statistical dif-

Table 3. Risk Factors for Hepatocellular Carcinoma Occurrence: Univariate Analysis HCC Risk factors Age, y ⬍50 50–60 60–70 ⬎70 Sex Female Male BMI, kg/m2 ⬍25 25–30 ⱖ30 Cirrhosis cause Alcohol HCV Mixedb Diabetes status No Yes aPer

100 person-years. and alcohol.

bHCV

N

(n)

Incidencea

HR (95% CI)

P

125 230 235 181

13 47 76 84

3.3 5.2 75 8.8

1 1.5 (.8–2.8) 2.2 (1.2–3.9) 2.5 (1.4–4.5)

NS .01 .002

275 496

62 220

4.6 8.2

1 1.8 (1.4–2.4)

⬍.0001

410 249 112

83 80 57

4.7 8.1 11.3

1 1.8 (1.3–2.4) 2.4 (1.7–3.3)

⬍.0001 ⬍.0001

478 220 73

108 81 31

5.6 8.0 10.7

1 1.4 (1.1–1.9) 2.0 (1.4–3.0)

.022 .001

540 231

123 97

5.5 9.4

1 1.7 (1.3–2.2)

.0002

August 2006

RISK FACTORS FOR HCC WITH ALCOHOLIC OR VIRAL C CIRRHOSIS

1065

ferences were observed in these BMI subgroups according to sex and diabetes status (Table 2). The distribution of age classes taken into account for statistical analysis were as follows: younger than 50 years, 125 patients (16%); 50 – 60 years, 230 patients (29%); 60 –70 years, 235 patients (31%); and older than 70 years, 181 patients (24%). Univariate Analysis For the whole cohort (n ⫽ 771), factors associated with HCC occurrence in univariate analysis were as follows: BMI 25–30 kg/m2 and BMI of 30 kg/m2 or more (vs ⬍25 kg/m2), diabetes, male sex, age older than 50 years (vs ⬍50 y), and HCV cause (vs alcoholic) (Table 3). Multivariate Analysis For the whole cohort, in multivariate analysis, predictive factors were as follows: BMI between 25–30 kg/m2 (hazard ratio [HR], 2.0; 95% confidence interval [CI], 1.4 –2.7), BMI 30 kg/m2 or more (HR, 2.8; 95% CI, 2.0 – 4.0), diabetes (HR, 1.6; 95% CI, 1.2–2.1), age 60 –70 years (HR, 2.4; 95% CI, 1.3– 4.3), age older than 70 years (HR, 3.0; 95% CI, 1.7–5.5), male sex (HR, 2.0; 95% CI, 1.4 –2.7), HCV (HR, 1.6; 95% CI, 1.1–2.2), and mixed cause (HR, 2.6; 95% CI, 1.7– 4.0) (Table 4). A significant positive linear relationship between BMI levels and HCC risk was observed. In patients with alcohol-related cirrhosis: age older than 70 years (HR, 2.6; 95% CI, 1.3–5.3; P ⫽ .008) (vs ⬍50 y), male sex (HR, 2.1; 95% CI, 1.3–3.4; P ⫽ .005), BMI of 25–30 kg/m2 (HR, 2.0; 95% CI, 1.3–3.3; P ⫽ .004), and Table 4. Risk Factors for Hepatocellular Carcinoma Occurrence: Multivariate Analysis Risk factors Age, y ⬍50 50–60 60–70 ⬎70 Sex Female Male Cirrhosis cause Alcohol HCV Mixeda BMI, kg/m2 BMI ⬍25 BMI 25–30 BMI ⱖ30 Diabetes status No Yes aHCV

and alcohol.

HR (95% CI)

P

1 1.6 (.9–2.9) 2.4 (1.3–4.3) 3.0 (1.7–5.5)

NS .005 ⬍.0001

1 2.0 (1.4–2.7)

⬍.0001

1 1.6 (1.2–2.2) 2.6 (1.7–4.0)

.002 ⬍.0001

1 2.0 (1.4–2.7) 2.8 (2.0–4.0)

⬍.0001 ⬍.0001

1 1.6 (1.2–2.1)

.001

Figure 1. Incidence of HCC according to BMI level in the 478 patients with alcoholic cirrhosis (Kaplan–Meier method). P ⬍ .0001.

30 kg/m2 or more (HR, 2.9; 95% CI, 1.7– 4.9; P ⫽ .0001) (vs ⬍25 kg/m2), and diabetes (HR, 1.6; 95% CI, 1.1–2.3; P ⫽ .029) were identified as risk factors. In patients with HCV-related cirrhosis, a BMI of 25–30 kg/m2 (HR, 1.7; 95% CI, 1.0 –2.8; P ⫽ .049), BMI of 30 kg/m2 or more (HR, 2.9; 95% CI, 1.6 –5.5; P ⫽ .001), and male sex (HR, 2.1; 95% CI, 1.3–3.4; P ⫽ .001) were risk factors. The association between diabetes and HCC was not significant (HR, 1.5; 95% CI, .9 –2.4; P ⫽ .10). The analysis in patients with mixedcause was not performed because of the small sample size of this subgroup. The probabilities of HCC occurrence according to BMI classes in alcoholic- and HCV-related cirrhosis are shown in Figures 1 and 2. After multivariate Cox analysis, all possible interactions of variables were tested with the Cox final model. The only statistically significant interaction observed was between the BMI and diabetes. A composite index was constructed by combining classes of BMI and diabetes status as follows: BMI of less than 25 kg/m2 without diabetes, BMI of less than 25 kg/m2 with diabetes, BMI of 25–30 kg/m2 without diabetes, BMI of 25–30 kg/m2 with diabetes, BMI of 30 kg/m2 or more without diabetes,

Figure 2. Incidence of HCC according to BMI level in the 220 patients with viral C cirrhosis (Kaplan–Meier method). P ⬍ .0001.

1066

N’KONTCHOU ET AL

CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 4, No. 8

Figure 3. Incidence of HCC according to composite index in the 771 patients with alcoholic and/or viral C cirrhosis (Kaplan–Meier method). The composite index was defined as follows: group 1: BMI of less than 25 kg/m2 and no diabetes; group 2: BMI of ⬍ 25 kg/m2 and diabetes; group 3: BMI of 25–30 kg/m2 and no diabetes; group 4: BMI of 25–30 kg/m2 and diabetes; group 5: BMI of 30 kg/m2 or more and no diabetes; group 6: BMI of 30 kg/m2 or more and diabetes. P ⬍ .0001 from group 1 to group 6.

and BMI of 30 kg/m2 or more with diabetes. The relation between the composite index and the occurrence of HCC was statistically significant (P ⬍ .00001). The probability of HCC occurrence according to the composite index in the whole population is shown in Figure 3. The HRs of each index level before and after adjusting for other risk factors, age, sex, and cause of cirrhosis were similar (Table 5).

Discussion Our results show that BMI and diabetes are predictive factors of the occurrence of HCC in patients with HCV or alcohol-related cirrhosis. Because the pathogenesis of HCC might differ according to the cause of cirrhosis, we chose to study these factors in homogenous groups of patients. We focused our study in patients with cirrhosis caused by alcohol and HCV, which are the main causes of HCC in the Western world. In addition,

a small number of patients with cirrhosis related to other causes, such as hepatitis B virus infection and genetic hemochromatosis, precluded a specific analysis of these subgroups. Even if HCC were not histologically proven in all cases, the noninvasive criteria used have been shown to be fairly specific in the screening situation if not sensitive enough. In addition, the diagnosis of HCC was confirmed by evolution in most of the cases, making the risk of false-positive diagnosis extremely low. Not unexpectedly, the incidence of HCC was higher in the subgroup of patients with mixed-cause cirrhosis, a finding that confirms previous reports suggesting an additional role of ethanol and HCV in hepatocarcinogenesis.21–23 In accordance with previous reports, the incidence of HCC also was greater in patients with HCV cirrhosis compared with those with alcoholic cirrhosis.23 HCC was observed in 20% of the total population at 5 years in our study. This is in the upper range of the annual incidence recorded in several screening studies (2%– 6.5%).23–26 Some prospective studies of patients with cirrhosis have shown a lower incidence, but these discrepancies could be expected considering differences in selection criteria and potential recruitment bias. In our study, patients younger than 40 years old were not included and 30% of patients were 70 years and older. The absence of the upper limit of age for inclusion was justified by the fact that new curative options for HCC such as transcutaneous treatments could be applied even in very old patients. It still is unclear whether age is by itself a risk factor of cancer or a risk factor of progression of the underlying liver disease, or the hallmark of the duration of this disease. From a practical point of view, because age is a main risk factor of HCC in this study as in others, the mean age of the studied population obviously could influence HCC incidence. In addition, in our study we included patients with Child B cirrhosis without ascites and patients with severe thrombocytopenia who required a transveinous liver biopsy examination. These groups of patients carrying a high risk of HCC often are not taken into account in screening studies.

Table 5. Impact of Composite Index on Hepatocellular Carcinoma Occurrence Univarite analysis Risk factors BMI BMI BMI BMI BMI BMI

⬍25 and no diabetes ⬍25 and diabetes 25–30 and no diabetes ⱖ30 and no diabetes 25–30 and diabetes ⱖ30 and diabetes

Multivariate analysis

N

HCC

Incidencea

HR

P

Adjusted HRa

P

306 104 161 73 88 37

48 35 47 28 33 29

3.8 6.9 7.8 7.8 8.6 19.6

1 1.7 (1.1–2.7) 2.1 (1.4–3.1) 2.0 (1.2–3.1) 2.3 (1.5–3.6) 5.2 (3.3–8.2)

.017 .0001 .005 .0001 .0001

1 1.4 (.9–2.2) 2.1 (1.4–3.2) 2.1 (1.3–3.4) 2.6 (1.6–4.1) 6.0 (3.7–9.7)

NS .0001 .002 .001 .0001

HR, hazard ratio. aAdjusted for age, sex, and cause per 100 person-years.

August 2006

RISK FACTORS FOR HCC WITH ALCOHOLIC OR VIRAL C CIRRHOSIS

Although this potential recruitment bias might influence the incidence of HCC, it does not preclude studying risk factors. In our study, the influence of BMI and diabetes on the incidence of HCC has been assessed in the whole population and in etiologic subgroups. In our study, diabetes was an independent risk factor in the overall population and in the alcohol subgroup, but did not reach statistical significance in the HCV subgroup. Because the sample size was smaller in these subgroups, we could not exclude a type II error. We secondly constructed a composite index combining BMI and diabetes status to reflect the metabolic syndrome and it is interesting to note that obese patients (BMI ⱖ30 kg/m2) without diabetes had a lower risk for HCC than overweight patients (BMI 25–30 kg/m2) with diabetes. The impact of overweight and obesity was observed in all subgroups, suggesting a common mechanism independent of the cause of liver disease. These findings were not unexpected because the incidence of many types of epithelial cancer is increased in overweight or obese patients, particularly those with insulin-resistance syndrome.5,6 Unfortunately, in our study, dyslipidemia and arterial hypertension could not be taken into account because they were not available in many patients. As a whole, these facts may lead to practical and theoretic consequences. From a practical point of view, BMI level and diabetes diagnosis could help to identify patients with cirrhosis who have a high risk for HCC occurrence. Indexes already have been defined that allow splitting patients among groups with different risk levels. The fact that these indexes are not used widely can be explained by a lack of external validation but also by this relative complexity. Among the main risk factors of HCC, it seems appropriate to select the most simple to record and those with the

1067

strongest predictive value. In addition to age and sex, which appeared predictive in the majority of prospective studies, BMI could take place among these factors usable for bedside assessment. At least for Western patients with HCV and/or alcoholic cirrhosis, a very simple index could be built taking into account sex, age, and BMI. A blood test could allow further discrimination but even without any biological data an accurate prediction could be made because any fatty male patient older than 60 years could be considered at high risk for HCC occurrence. In our study, the risk for HCC occurrence increased consistently with the increasing level of BMI and was increased further by the presence of diabetes, suggesting a role of insulin resistance. From this finding, we speculated that general and hepatic consequences of overweight and diabetes act as cofactors of carcinogenesis in cirrhosis. Overweight and diabetes could favor HCC development through different mechanisms. Increased oxidative stress is well documented in obese and diabetic patients.27,28 In addition, insulin resistance could lead to steatosis and to lipid peroxidation, and particularly to overproduction of malondialdehyde, a mutagenic and carcinogenic compound.29 –33 Insulin acts also as a mitogenic and antiapoptotic factor on hepatocytes.34 Experimentally, HCC was observed after transplantation of pancreatic islet cells in the liver in rats with streptozotocin-induced diabetes.35,36 In addition, insulin resistance is associated with high liver iron content, which seems to favor carcinogenesis.37 Lastly, increased serum levels of hepatocyte growth factor and free sex hormones reported in this condition also could play a role in hepatocarcinogenesis.38,39 These different pathways have to be explored in patients and animal models. Prospective studies that take insulin resistance, for instance by HOMA test, directly into account need to be conducted to confirm

Figure 4. Relationship between BMI level and risk for HCC in the 771 patients with alcoholic and/or viral C cirrhosis.

1068

N’KONTCHOU ET AL

and further precise our findings. The impact of these factors has to be tested in a large cohort of patients with cirrhosis from other causes, in other parts of the world. To what extent a reduction of these risk factors could lead to a lower incidence of HCC also remains to be explored (Figure 4).

References 1. Grundy SM. Metabolic complications of obesity. Endocrine 2000; 13:155–165. 2. Kopelman PG. Obesity as a medical problem. Nature 2000;404: 635– 643. 3. Meigs JB. Epidemiology of the insulin resistance syndrome. Curr Diab Rep 2003;3:73–79. 4. Calle EE, Rodriguez C, Walker-Thurmond K, et al. Overweight, obesity, and mortality from cancer in a prospectively studied cohort of U.S. adults. N Engl J Med 2003;348:1625–1638. 5. Moore MA, Park CB, Tsuda H. Implications of the hyperinsulinaemia-diabetes-cancer link for preventive efforts. Eur J Cancer Prev 1998;7:89 –107. 6. Nilsen TI, Vatten LJ. Prospective study of colorectal cancer risk and physical activity, diabetes, blood glucose and BMI: exploring the hyperinsulinaemia hypothesis. Br J Cancer 2001;84:417– 422. 7. Trevisan M, Liu J, Muti P, et al. Markers of insulin resistance and colorectal cancer mortality. Cancer Epidemiol Biomarkers Prev 2001;10:937–941. 8. Wolk A, Gridley G, Svensson M, et al. A prospective study of obesity and cancer risk (Sweden). Cancer Causes Control 2001; 12:13–21. 9. Marchesini G, Bugianesi E, Forlani G, et al. Nonalcoholic fatty liver, steatohepatitis, and the metabolic syndrome. Hepatology 2003;37:917–923. 10. Poonawala A, Nair SP, Thuluvath PJ. Prevalence of obesity and diabetes in patients with cryptogenic cirrhosis: a case-control study. Hepatology 2000;32:689 – 692. 11. Adinolfi LE, Gambardella M, Andreana A, et al. Steatosis accelerates the progression of liver damage of chronic hepatitis C patients and correlates with specific HCV genotype and visceral obesity. Hepatology 2001;33:1358 –1364. 12. Friedenberg F, Pungpapong S, Zaeri N, et al. The impact of diabetes and obesity on liver histology in patients with hepatitis C. Diabetes Obes Metab 2003;5:150 –155. 13. Raynard B, Balian A, Fallik D, et al. Risk factors of fibrosis in alcohol-induced liver disease. Hepatology 2002;35:635– 638. 14. Braga C, La Vecchia C, Negri E, et al. Attributable risks for hepatocellular carcinoma in northern Italy. Eur J Cancer 1997; 33:629 – 634. 15. el-Serag HB. Epidemiology of hepatocellular carcinoma. Clin Liver Dis 2001;5:87–107, vi. 16. El-Serag HB, Richardson PA, Everhart JE. The role of diabetes in hepatocellular carcinoma: a case-control study among United States Veterans. Am J Gastroenterol 2001;96:2462–2467. 17. La Vecchia C, Negri E, Decarli A, et al. Diabetes mellitus and the risk of primary liver cancer. Int J Cancer 1997;73:204 –207. 18. Tazawa J, Maeda M, Nakagawa M, et al. Diabetes mellitus may be associated with hepatocarcinogenesis in patients with chronic hepatitis C. Dig Dis Sci 2002;47:710 –715. 19. Nair S, Mason A, Eason J, et al. Is obesity an independent risk factor for hepatocellular carcinoma in cirrhosis? Hepatology 2002;36:150 –155. 20. Hassan MM, Hwang LY, Hatten CJ, et al. Risk factors for hepatocellular carcinoma: synergism of alcohol with viral hepatitis and diabetes mellitus. Hepatology 2002;36:1206 –1213. 21. Bruix J, Sherman M, Llovet JM, et al. Clinical management of hepatocellular carcinoma. Conclusions of the Barcelona-2000

CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 4, No. 8

22.

23.

24.

25.

26.

27. 28.

29.

30.

31.

32.

33.

34.

35.

36.

37. 38.

39.

EASL conference. European Association for the Study of the Liver. J Hepatol 2001;35:421– 430. Donato F, Tagger A, Gelatti U, et al. Alcohol and hepatocellular carcinoma: the effect of lifetime intake and hepatitis virus infections in men and women. Am J Epidemiol 2002;155:323–331. Ganne-Carrie N, Chastang C, Chapel F, et al. Predictive score for the development of hepatocellular carcinoma and additional value of liver large cell dysplasia in Western patients with cirrhosis. Hepatology 1996;23:1112–1118. Ikeda K, Saitoh S, Koida I, et al. A multivariate analysis of risk factors for hepatocellular carcinogenesis: a prospective observation of 795 patients with viral and alcoholic cirrhosis. Hepatology 1993;18:47–53. Velazquez RF, Rodriguez M, Navascues CA, et al. Prospective analysis of risk factors for hepatocellular carcinoma in patients with liver cirrhosis. Hepatology 2003;37:520 –527. Degos F, Christidis C, Ganne-Carrie N, et al. Hepatitis C virus related cirrhosis: time to occurrence of hepatocellular carcinoma and death. Gut 2000;47:131–136. Yang S, Zhu H, Li Y, et al. Mitochondrial adaptations to obesityrelated oxidant stress. Arch Biochem Biophys 2000;378:259 –268. Keaney JF Jr, Larson MG, Vasan RS, et al. Obesity and systemic oxidative stress: clinical correlates of oxidative stress in the Framingham Study. Arterioscler Thromb Vasc Biol 2003;23:434 – 439. Abel S, Smuts CM, de Villiers C, et al. Changes in essential fatty acid patterns associated with normal liver regeneration and the progression of hepatocyte nodules in rat hepatocarcinogenesis. Carcinogenesis 2001;22:795– 804. Gochee PA, Jonsson JR, Clouston AD, et al. Steatosis in chronic hepatitis C: association with increased messenger RNA expression of collagen I, tumor necrosis factor-alpha and cytochrome P450 2E1. J Gastroenterol Hepatol 2003;18:386 –392. Ohata K, Hamasaki K, Toriyama K, et al. Hepatic steatosis is a risk factor for hepatocellular carcinoma in patients with chronic hepatitis C virus infection. Cancer 2003;97:3036 –3043. Niedernhofer LJ, Daniels JS, Rouzer CA, et al. Malondialdehyde, a product of lipid peroxidation, is mutagenic in human cells. J Biol Chem 2003;278:31426 –31433. Yang S, Lin HZ, Hwang J, et al. Hepatic hyperplasia in noncirrhotic fatty livers: is obesity-related hepatic steatosis a premalignant condition? Cancer Res 2001;61:5016 –5023. Khamzina L, Gruppuso PA, Wands JR. Insulin signaling through insulin receptor substrate 1 and 2 in normal liver development. Gastroenterology 2003;125:572–585. Scharf JG, Ramadori G, Dombrowski F. Analysis of the IGF axis in preneoplastic hepatic foci and hepatocellular neoplasms developing after low-number pancreatic islet transplantation into the livers of streptozotocin diabetic rats. Lab Invest 2000;80:1399 –1411. Evert M, Sun J, Pichler S, et al. Insulin receptor, insulin receptor substrate-1, Raf-1, and Mek-1 during hormonal hepatocarcinogenesis by intrahepatic pancreatic islet transplantation in diabetic rats. Cancer Res 2004;64:8093– 8100. Mendler MH, Turlin B, Moirand R, et al. Insulin resistance-associated hepatic iron overload. Gastroenterology 1999;117:1155–1163. Rehman J, Considine RV, Bovenkerk JE, et al. Obesity is associated with increased levels of circulating hepatocyte growth factor. J Am Coll Cardiol 2003;41:1408 –1413. Tamimi R, Mucci LA, Spanos E, et al. Testosterone and oestradiol in relation to tobacco smoking, body mass index, energy consumption and nutrient intake among adult men. Eur J Cancer Prev 2001;10:275–280.

Address requests for reprints to: Gisèle N’Kontchou, MD, Service d’Hépato-Gastroentérologie, Hôpital Jean Verdier, 93143 Bondy Cedex, France. e-mail: [email protected]; fax: (33) 1-48-02-62-02.