The authors conducted a prospective, longitudinal, multicentre study in patients with liver cirrhosis; a total of 192 patients were on follow-up, of this number 125 died during the study. Mean age, 56 years, was markedly more advanced in the dead; men prevailed among the patients. Duration of complaints and the interval since the diagnosis had been established was shorter in the dead than in those alive; a positive HBs antigen was likewise more often present in the former. Clinical and laboratory parameters in this group were assessed with a view to the differences between the dead and those alive and with a view to the time course of changes in these differences. Multifactorial discriminative analysis of these data for various intervals before death was also performed. The most pronounced differences were noted in clinical manifestations and laboratory signs of hepatic insufficiency, especially in the presence and degree of ascites, jaundice, encephalopathy, and in the values of bilirubin, Quick' test and blood albumin levels; thus the degree of liver insufficiency appears to be the most important prognostic marker in patients with liver cirrhosis. Moreover, a significant increase in urea and leucocyte levels was found in the terminal stages of the disease. Ascites, encephalopathy and a progressive increase in serum bilirubin levels seem to be the best markers for discriminative analysis of prognosis for the last six months of life. Results of this study, while confirming some findings by other authors, offer also new insights into the whole issue.