Risk factors for hepatocellular carcinoma in HCV-cirrhosis: What we know and what is missing

Hepatocellular carcinoma (HCC) is a major health problem, being the fifth most common cancer worldwide with 626,000 new cases in 2002 [[1]Parkin D.M. Bray F. Ferlay J. Pisani P. Global cancer statistics 2002.CA Cancer J Clin. 2005; 55: 74-108Crossref PubMed Scopus (17264) Google Scholar]. The incidence and absolute mortality of HCC is increasing in Europe and the United States [[2]El Serag H.B. Mason A.C. Rising incidence of hepatocellular carcinoma in the United States.N Engl J Med. 1999; 340: 745-750Crossref PubMed Scopus (2713) Google Scholar], and is currently the leading cause of death amongst cirrhotic patients [3Sangiovanni A. Del Ninno E. Fasani P. et al.Increased survival of cirrhotic patients with a hepatocellular carcinoma detected during surveillance.Gastroenterology. 2004; 126: 1005-1014Abstract Full Text Full Text PDF PubMed Scopus (507) Google Scholar, 4Fattovich G. Stroffolini T. Zagni I. Donato F. Hepatocellular carcinoma in cirrhosis: incidence and risk factors.Gastroenterology. 2004; 127: S35-S50Abstract Full Text Full Text PDF PubMed Scopus (1921) Google Scholar]. Improvement of the overall outcome of these patients will come from prevention policies, early detection programs and advancement in the efficacy of therapies. Identification of the at-risk populations is particularly crucial to implement efficient prevention policies, and to engage the adequate populations in cost-effective surveillance programs. Thus, the impact of the knowledge of at-risk populations in the surveillance strategy, and the relevance of these factors for the prevention and treatment strategies should be carefully analyzed. Surveillance programs in well-selected cirrhotic patients enable to identify tumours at early stages, when potentially curative therapies are able to expand life expectancy [[5]Llovet J.M. Burroughs A. Bruix J. Hepatocellular carcinoma.Lancet. 2003; 362: 1907-1917Abstract Full Text Full Text PDF PubMed Scopus (3768) Google Scholar]. However, no reliable data yet supports a clear effect in reducing cancer-related deaths. In fact, the single randomized controlled trial (RCT) comparing surveillance (ultrasonography and alpha-fetoprotein assessment every six months) vs. non-surveillance published so far has been conducted in Chinese patients with prior hepatitis B infection [[6]Chen J.G. Parkin D.M. Chen Q.G. Lu J.H. Shen Q.J. Zhang B.C. Zhu Y.R. Screening for liver cancer: results of a randomised controlled trial in Qidong.China J Med Screen. 2003; 10: 204-209Crossref PubMed Scopus (235) Google Scholar]. All the evidence to justify the application of such programs in HCV-cirrhotic individuals is derived from cohort studies assessing specific screening schedules or from cost-effectiveness analysis using mathematical modeling [3Sangiovanni A. Del Ninno E. Fasani P. et al.Increased survival of cirrhotic patients with a hepatocellular carcinoma detected during surveillance.Gastroenterology. 2004; 126: 1005-1014Abstract Full Text Full Text PDF PubMed Scopus (507) Google Scholar, 7Bruix J. Sherman M. Management of hepatocellular carcinoma.Hepatology. 2005; 42: 1208-1236Crossref PubMed Scopus (5063) Google Scholar, 8Bruix J. Llovet J.M. HCC surveillance: who is the target population?.Hepatology. 2003; 37: 507-509Crossref PubMed Scopus (51) Google Scholar, 9Sarasin F.P. Giostra E. Hadengue A. Cost-effectiveness of screening for detection of small hepatocellular carcinoma in western patients with Child-Pugh class A cirrhosis.Am J Med. 1996; 101: 422-434Abstract Full Text PDF PubMed Scopus (293) Google Scholar]. It has been estimated that surveillance in cirrhotic patients is cost-effective when it increases longevity for at least 3 months with a cost per year of life saved below $50,000 [[10]Laupacis A. Feeny D. Detsky A.S. Tugwell P.X. How attractive does a new technology have to be to warrant adoption and utilization? Tentative guidelines for using clinical and economic evaluations..Can Med Assoc J. 1992; 146: 473-481PubMed Google Scholar]. According to the guidelines for the management of HCC of the American Association for the Study of the Liver (AASLD), such a scenario is achieved in cirrhotic patients of any etiology when the risk of HCC is 1.5%/year or greater [[7]Bruix J. Sherman M. Management of hepatocellular carcinoma.Hepatology. 2005; 42: 1208-1236Crossref PubMed Scopus (5063) Google Scholar]. The risk of HCC development in HCV-cirrhosis ranges between 7.1 and 3–4%/ year in Japan and in Europe, respectively, being HBV co-infection and alcohol intake clear co-factors increasing the incidence approximately 2- to 4-fold. In genetic hemochromatosis is of ∼3–4% per year, whereas in HBV-cirrhosis ranges between 2 and 4% per year. Finally, the incidence is around ∼1.5 per year in alcoholic-induced cirrhosis and in primary biliary cirrhosis. The other etiologies of cirrhosis bear a lower HCC risk [[4]Fattovich G. Stroffolini T. Zagni I. Donato F. Hepatocellular carcinoma in cirrhosis: incidence and risk factors.Gastroenterology. 2004; 127: S35-S50Abstract Full Text Full Text PDF PubMed Scopus (1921) Google Scholar]. Ikeda et al. reported in this issue of the Journal the risk of HCC development in a cohort of 183 consecutive patients with HCV-related cirrhosis (Child-Pugh class A or B) enrolled between 1974 and 1990 and followed for a median of 9.5 years [[11]Ikeda K. Arase Y. Saitoh S. Kobayashi M. Someya T. Hosaka T. Prediction model of hepatocarcinogenesis for patients with hepatitis C virus-related cirrosis. Validation with internal and external cohort.J Hepatol. 2006; 44: 1089-1097Abstract Full Text Full Text PDF PubMed Scopus (42) Google Scholar]. It is interesting to note that the 5-year cumulative HCC risk of 31.5% reported in the study population is comparable to the 5-year cumulative HCC incidence of 30% reported in a systematic review of longitudinal Japanese studies of untreated patients with HCV-related compensated cirrhosis [[4]Fattovich G. Stroffolini T. Zagni I. Donato F. Hepatocellular carcinoma in cirrhosis: incidence and risk factors.Gastroenterology. 2004; 127: S35-S50Abstract Full Text Full Text PDF PubMed Scopus (1921) Google Scholar]. Among a set of clinical and laboratory variables, age >55 years, male gender, platelet count <100,000 counts/mm [[3]Sangiovanni A. Del Ninno E. Fasani P. et al.Increased survival of cirrhotic patients with a hepatocellular carcinoma detected during surveillance.Gastroenterology. 2004; 126: 1005-1014Abstract Full Text Full Text PDF PubMed Scopus (507) Google Scholar] and alpha-fetoprotein >20 ng/ml were the only independent prognostic factors for HCC by using Cox's proportional hazard analysis. A prognostic model was developed defining a 5-year cumulative HCC risk of 66% in patients with four adverse risk factors and of 10% in patients with none of them. This model was validated using two independent cohorts including more than 500 patients. The following questions arise concerning the results of the present study: Is surveillance recommended in all the subgroups reported? And if so should the specific surveillance schedule vary according to the at-risk population targeted? Are these prognostic factors consistent with the state of the art knowledge? And if so how should the results impact the design of prevention studies? And finally, are there any environmental or genetic backgrounds underlying the differences between the incidence of HCC development reported in Japanese and Western HCV- cohorts? As stated before, surveillance programs are recommended in patients with a risk of HCC development above 1.5% per year, and thus all the subgroups proposed in the study should be engaged in these programs. The high incidence of HCC reported prevents to identify any single subgroup in which surveillance is not cost-effective. It is also clear that the high-risk group (10-year cumulative HCC incidence of 94%) should not be followed by shorter screening intervals, for instance every 3 months, since it has been repeatedly stated that higher HCC risk does not mean faster tumour growth [[7]Bruix J. Sherman M. Management of hepatocellular carcinoma.Hepatology. 2005; 42: 1208-1236Crossref PubMed Scopus (5063) Google Scholar]. This statement might be eventually reconsidered in the future if the identification of at-high risk populations links genetic predisposition with biological behaviour. This would be the case of the identification of germ-line or somatic mutations involving genes in pathways conferring an aggressive phenotype [12Vogelstein B. Kinzler K.W. Cancer genes and the pathways they control.Nat Med. 2004; 10: 789-799Crossref PubMed Scopus (3302) Google Scholar, 13Savage S.A. Chanock S.J. Using germ-line genetic variation to investigate and treat cancer.Drug Discov Today. 2004; 9: 610-618Crossref PubMed Scopus (6) Google Scholar]. The recent identification of a molecular HCC signature that links cell origin (progenitor cells), disrupted signalling pathways (c-fos, jun) and poor prognosis (median survival <10 months) opens a new avenue to the understanding of the pathogenesis of HCC [[14]Lee JS, Heo J, Libbrecht L, Chu IS, Kaposi-Novak P, Calvisi DF et al., A novel prognostic subtype of human hepatocellular carcinoma derived from hepatic progenitor cells. Nat Med; 2006 [on line, March 12].Google Scholar]. Molecular studies assessing genomic profiles and genetic aberrations comparing cirrhotic tissues of patients with and without HCC at the same stage of fibrosis are needed to unravel this dilemma. The risk factors identified (age, sex, platelet count and AFP levels) are consistent with those previously reported [4Fattovich G. Stroffolini T. Zagni I. Donato F. Hepatocellular carcinoma in cirrhosis: incidence and risk factors.Gastroenterology. 2004; 127: S35-S50Abstract Full Text Full Text PDF PubMed Scopus (1921) Google Scholar, 5Llovet J.M. Burroughs A. Bruix J. Hepatocellular carcinoma.Lancet. 2003; 362: 1907-1917Abstract Full Text Full Text PDF PubMed Scopus (3768) Google Scholar, 7Bruix J. Sherman M. Management of hepatocellular carcinoma.Hepatology. 2005; 42: 1208-1236Crossref PubMed Scopus (5063) Google Scholar]. Studies aimed at developing prognostic models for HCC development need to consider at least three determinants of the course of the disease: stage of liver disease, all known co-factors and geographical area. Of particular importance is the stage of hepatic fibrosis and cirrhosis. In Japanese studies, the HCC incidence rate was 7.1 per 100 person years in patients with HCV-related compensated cirrhosis and 1.8 in patients with chronic hepatitis C without cirrhosis at diagnosis [[4]Fattovich G. Stroffolini T. Zagni I. Donato F. Hepatocellular carcinoma in cirrhosis: incidence and risk factors.Gastroenterology. 2004; 127: S35-S50Abstract Full Text Full Text PDF PubMed Scopus (1921) Google Scholar]. It is unknown whether the initially non-cirrhotic patients effectively developed cirrhosis between the time of entry into the study and HCC diagnosis, nevertheless these data suggest that duration and stage of cirrhosis influences the risk of hepatocarcinogenesis. To support this concept, epidemiological data have shown that Child-Pugh class B or C cirrhotic patients have a 3- to 8-fold increased risk of liver cancer than those with Child-Pugh class A cirrhosis [[4]Fattovich G. Stroffolini T. Zagni I. Donato F. Hepatocellular carcinoma in cirrhosis: incidence and risk factors.Gastroenterology. 2004; 127: S35-S50Abstract Full Text Full Text PDF PubMed Scopus (1921) Google Scholar]. Overall these data indicates that a homogeneous study population is required to obtain precise prognostic factors. Other epidemiological co-factors need to be specifically assessed in prognostic studies. There is evidence that alcohol intake of 60 g/day or more interacts with HCV infection increasing the HCC risk (2-fold relative to HCV infection alone) [[15]Donato F. Tagger A. Gelatti U. Parrinello G. Boffetta P. Albertini A. et al.Alcohol and hepatocellular carcinoma: the effect of lifetime intake and hepatitis virus infection in men and women.Am J Epidemiol. 2002; 155: 323-331Crossref PubMed Scopus (504) Google Scholar]. A fundamental issue, however, is the dose–effect relationship between alcohol intake and the HCC risk. A systematic review of epidemiological studies has reported that the pattern of HCC risk due to alcohol intake shows a continuous dose–effect curve without a definite threshold, although for HCV infected subjects a significant increase in HCC risk was shown at a level of 40 g/day, suggesting that even a low alcohol intake can not be regarded as safe in this population [[16]Donato F, Gelatti U, Limina MR, Fattovich G. Southern Europe as an example of interaction between various environmental factors. A systematic review of the epidemiologic evidence. Oncogene; 2006 [in press].Google Scholar]. While in the study by Ikeda et al. cirrhotic patients with alcohol abuse were excluded, individuals drinking up to 70 g/day could still be enrolled and this confounding factor may influence the probability of HCC appearance. Thus the dose–effect instead of all-or-none relationship between alcohol intake and HCC risk should be explored in prognostic studies of HCV-related cirrhosis. From a methodologically point of view the study of Ikeda et al. suffers from some weaknesses regarding evaluation of prognostic variables. The use of cut-off points for quantitative variables and the arbitrary dichotomization according to the median value may lead to loss of prognostic information in these variables. The choice of arbitrary thresholds may cause a variable to become significantly more or less strongly/strictly associated with the occurrence of a well-defined outcome. It is methodologically fully correct to utilize the full value spectrum of quantitative variables, although sometimes it may be necessary to transform the variable, e.g. using logarithmic transformation, to fulfil the model assumption [[17]Christensen E. Multivariate survival analysis using Cox's regression model.Hepatology. 1987; 7: 1346-1358Crossref PubMed Scopus (301) Google Scholar]. Alternatively, the most informative cut-off can be identified by Receiver Operating Characteristics (ROC) curves. A number of similar analysis have been performed in patients with HCV-related cirrhosis confirming age and gender as independent factors associated with HCC, but showing that other variables such as bilirubin, albumin, oesophageal varices, small and large cell dysplasia and increased labelling index for proliferating cell nuclear antigen hold additional prognostic information [4Fattovich G. Stroffolini T. Zagni I. Donato F. Hepatocellular carcinoma in cirrhosis: incidence and risk factors.Gastroenterology. 2004; 127: S35-S50Abstract Full Text Full Text PDF PubMed Scopus (1921) Google Scholar, 18Mazzella G. Accogli E. Sottili S. Festi D. Orsini M. Salzetta A. et al.Alpha interferon treatment may prevent hepatocellular carcinoma in HCV-related liver cirrhosis.J Hepatol. 1996; 24: 141-147Abstract Full Text PDF PubMed Scopus (314) Google Scholar, 19Fattovich G. Pantalena M. Zagni I. Realdi G. Schalm S.W. Christensen E. Effect of hepatitis B and C infections on the natural history of compensated cirrhosis: a cohort study of 297 patients.Am J Gastroenterol. 2002; 97: 2886-2895Crossref PubMed Google Scholar, 20Degos F. Christidis C. Ganne-Carrie N. Farmachidi J.P. Degott C. Guettier C. et al.Hepatitis C virus related cirrhosis: time to occurrence of hepatocellular carcinoma and death.Gut. 2000; 47: 131-136Crossref PubMed Scopus (330) Google Scholar, 21Donato M.F. Arosio E. Del Ninno E. Ronchi G. Lampertico P. Morabito A. Balestrieri M.R. Colombo M.High rates of hepatocellular carcinoma in cirrhotic patients with high liver cell proliferative activity.Hepatology. 2001; 34: 523-528Crossref PubMed Scopus (101) Google Scholar]. Finally, although the variables identified in the study by Ikeda et al. are handled in the modelling as having equal importance, the prognostic influence of age and alphafetoprotein seems to be greater than that of sex and platelet count as indicated by the hazard ratio reported. Previous studies reported that the HCC incidence in HCV-related cirrhosis appears to be about 2-fold higher in Japan than in the West (5-year cumulative incidence of 30 and 17%, respectively) [[4]Fattovich G. Stroffolini T. Zagni I. Donato F. Hepatocellular carcinoma in cirrhosis: incidence and risk factors.Gastroenterology. 2004; 127: S35-S50Abstract Full Text Full Text PDF PubMed Scopus (1921) Google Scholar]. This high incidence of HCC in HCV-cirrhotic patients in Japan is confirmed herein, and indicates the need to launch studies to unravel whether the interaction of genetic and/or environmental factors are responsible of this 2-fold increase. Classical studies in cancer epidemiology have highlighted marked differences in the incidence of a particular type of cancer according to geographical regions. For instance, the incidence of prostate cancer in Afro-American males is 70 times higher than in Chinese males [[13]Savage S.A. Chanock S.J. Using germ-line genetic variation to investigate and treat cancer.Drug Discov Today. 2004; 9: 610-618Crossref PubMed Scopus (6) Google Scholar]. In a recent study from Japan, single nucleotide polymorphisms (SNPs) of three genes (SCYB14, GFRA1, and CRHR2) were significantly and independently associated with HCC development [[22]Kato N. Ji G. Wang Y. Baba M. Hoshida Y. Otsuka M. Large-scale search of single nucleotide polymorphisms for hepatocellular carcinoma susceptibility genes in patients with hepatitis C.Hepatology. 2005; 42: 846-853Crossref PubMed Scopus (54) Google Scholar]. Interestingly, SCYB14 encodes secreted proteins involved in immunoregulatory and inflammatory processes. The relevance of these SNPs should be further validated in population-based studies, and may represent genetic markers that are in linkage disequilibrium with other causative variations. These results link with recently published population-based studies, which have identified findings in germ-line SNPs associated with a moderate to small effect (1.5- to 2.0-fold) in cancer [[23]Altshuler D. Brooks L.D. Chakravarti A. Collins F.S. Daly M.J. Donnelly P. International HapMap Consortium.A haplotype map of the human genome.Nature. 2005; 437: 1299-1320Crossref PubMed Scopus (4754) Google Scholar]. However, other epidemiological and environmental factors may account for this difference. There is increasing evidence that additional environmental factors, such as tobacco smoking, diabetes and obesity may favour the development of HCC in patients with HCV infection and these variables may hold additional prognostic information. Some cohort studies from East Asia provided evidence of a synergistic interaction between tobacco smoking and HCV infection on HCC risk [24Mori M. Hara M. Wada I. Hara T. Yamamoto K. Honda M. Naramoto J. Prospective study of hepatitis B and C viral infections, cigarette smoking, alcohol consumption, and other factors associated with hepatocellular carcinoma risk in Japan.Am J Epidemiol. 2000; 151: 131-139Crossref PubMed Scopus (126) Google Scholar, 25Sun C.A. Wu D.M. Lin C.C. Lu S.N. You S.L. Wang L.Y. Wu M.H. Chen C.J. Incidence and cofactors of hepatitis C virus-related hepatocellular carcinoma: a prospective study of 12,008 men in Taiwan.Am J Epidemiol. 2003; 157: 674-682Crossref PubMed Scopus (198) Google Scholar]. Case-control studies found that diabetes can increase the HCC risk about 2- to 4- fold regardless of the presence of major risk factor of the disease (HCV, HBV, alcoholic liver disease) [26Yuan J.M. Govindarajan S. Arakawa K. Yu M.C. Synergism of alcohol, diabetes, and viral hepatitis on the risk of hepatocellular carcinoma in blacks and whites in the US.Cancer. 2004; 101: 1009-10017Crossref PubMed Scopus (245) Google Scholar, 27Davila J.A. Morgan R.O. Shaib Y. McGlynn K.A. El-Serag H.B. Diabetes increases the risk of hepatocellular carcinoma in the United States: a population based case control study.Gut. 2005; 54: 533-539Crossref PubMed Scopus (552) Google Scholar] and one of them found a significant positive interaction between HCV and diabetes for HCC risk [[27]Davila J.A. Morgan R.O. Shaib Y. McGlynn K.A. El-Serag H.B. Diabetes increases the risk of hepatocellular carcinoma in the United States: a population based case control study.Gut. 2005; 54: 533-539Crossref PubMed Scopus (552) Google Scholar]. A French longitudinal cohort study on patients with HCV-related cirrhosis found an association between body mass index (BMI) and HCC with a dose-effect relation: the hazard ratio for HCC was 1.7 for a BMI of 25–30 and 2.9 for a BMI >30 kg/m [2El Serag H.B. Mason A.C. Rising incidence of hepatocellular carcinoma in the United States.N Engl J Med. 1999; 340: 745-750Crossref PubMed Scopus (2713) Google Scholar, 28N'Kontchou G, Paries J, Tin Tin Htar M, Ganne-Carrie N, Costentin L, Grando-Lemaire V, et al. Overweight and diabetes as risk factors for hepatocellular carcinonoma in patients with alcoholic or viral C cirrhosis. Clin Gastroenterol Hepatol; 2006 [in press].Google Scholar]. Overall these data indicate that a number of determinants of HCC risk in HCV-related cirrhosis may not be available or even identified and it is biologically plausible that many factors may interact in a complex way in causing the disease in the individual patient. The descriptive variables used in the model herein analyzed and others recently published [[29]Velazquez R.F. Rodriguez M. Navascues C.A. Linares A. Perez R. Sotorrios N.G. et al.Prospective analysis of risk factors for hepatocellular carcinoma in patients with liver cirrhosis.Hepatology. 2003; 37: 520-527Crossref PubMed Scopus (372) Google Scholar] may explain only part of the observed variations in HCC risk among patients with HCV-related cirrhosis. A better knowledge of the molecular pathogenesis of the disease, and identification of genetic predispositions might provide more comprehensive prognostic information. Although prognostic models are not precise enough to confer a complete predictability, the risk variables should be taken into account when designing preventive studies in HCV-cirrhotic patients in order to discard base-line unbalances of critical independent predictors of risk. Similar recommendations are endorsed for the design of RCT comparing different screening strategies, where stratification for these key variables prior randomization should be mandatory.