970 OCCURRENCE AND CLINICOPATHOLOGICAL RELEVANCE OF PROGNOSTIC MARKERS IN HEPATOCELLULAR CARCINOMA

The threshold annual incidence for efficacy of HCC surveillance in non-HBV liver cirrhosis has been established in 1.5%. Although alcoholic cirrhosis is clearly a risk factor for HCC, the incidence of the tumour in this population is not accurately known. Aim: To determine the annual incidence of HCC in a cohort of patients with alcoholic cirrhosis followed prospectively and to define groups with different risk of tumour development. Methods: We evaluated 450 consecutive patients with alcoholic liver cirrhosis (369 men, mean age 53.9±7.4 years) included in a surveillance program for HCC detection (mean follow-up 59±49 months). All patients were Child–Pugh class A/B. The diagnosis of cirrhosis was supported by liver biopsy in 140 patients and was based on clinical-biological criteria in the rest. Alcoholic etiology was established by history of alcohol consumption and the exclusion of other etiologies. The first 5 years after inclusion in the program were considered to define potential risk factors. Nineteen demographic, clinical and laboratory variables were analysed. Results: During the overall follow-up, 62 patients developed HCC. The annual incidence was 2.6% and the 10-years cumulative incidence was 25.2%. In the first 5 years after inclusion in the program, 43 patients developed HCC. In univariate analysis, age > 58 years, platelets 2mg/dl, albumin 58 (OR 2.91, 95%CI 1.29– 6.57) and platelets <125×103/mm3 (OR 2.40, 95%CI 1.30–4.46) were independently associated with an increased risk of HCC. These two variables allowed us to define 3 different risk groups. The annual incidence of HCC in the group without any of these two factors (n = 109), with one of them (n=249) or with both (n =87) was 0.6%, 2.8% and 5.3% respectively (p < 0.0001). Conclusions: The annual incidence of HCC in patients with alcoholic cirrhosis Child–Pugh class A/B is about 2.5%. Two simple variables, age and platelets, can be used to distribute the patients in different risk groups for HCC development over the next 5 years. A surveillance program for early detection of HCC is probably not justifiable in the low-risk group.