To investigate the prognostic factors in Western patients with hepatocellular carcinoma, 206 patients with condjrmed diagnoses of hepatocellular carcinoma were etadied in terms of survival. All patients were dialplosed between 1983 and 1987. A multivariate survival analysis (Cox regreesion model) using clinical, biochemical, ultrasonographical and pathological data obtained at diagnosis dieclosed that bilirubin (p = O.OOOl), aecitea (p = O.OOOl), toxic syndrome (defined by the preaence of weight loss >lo% premorbid weight, malabe and anorexia) (p = 0.0091, b l d urea nitrogen (p = 0.0261, tumor size (p = 0.001), yglutamyltranmpeptidase (p = O.OOOS), age (p = 0.0006), serum d u m (p = 0.003) and presence of metaetasee (p = 0.002) were independent predictore of survival. According to the contribution of each of these factors to the final model, a pragmatic index was condructed allowing division of patients in different groups according to their relative rielr of death: RRD = EXP (Age x 0.03 + Ascites x 0.8281 +BUN x 0.0137 + Serum sodium x (-0.0638) + y-Glutamyltranspeptidase x 0.0019 + Bilirubin x 0.0734 + Tumor size x 0.33 + Toxic syndrome x 0.4966 + Metaetases x 0.66). These d t s facilitate the stratification of hepatocellular carcinoma patients to design and evaluate future controlled trials. (HEPATOLOGY 19w);12:763-760.)
Hepatocellular carcinoma (HCC) usually occurs in a background of cirrhosis and is considered to have a dismal prognosis (1). Although in the past 10 y-r several studies have defined the prognostic factors that affect survival in patients with compensated (2) or decompensated (3-6) cirrhosis, factors & d i n g prognosis in HCC are less well known. Most of the prognostic studies published up to now in patients with HCC were made to define histological (7,8) or surgical (9-11) factors and/or applied only univariate analysis (1, 12, 13). The only studies applying multivariate analysis including large numbers of patients and wide sets of variables were