The prognosis of patients with resectable hepatocellular carcinoma depends mainly on the anatomical extent of the tumour and on the general condition of the patient. Given the growing evidence that proliferation indices may be of prognostic significance in hepatocellular carcinomas and that parameters of cell loss (usually, but not exclusively, due to programmed cell death) are biologically relevant, the identification and quantitation of proliferative capacity and apoptosis may be of prognostic importance. In this study four different methods have been used to assess proliferation in a series of 193 curatively (R0) resected hepatocellular carcinomas: mitotic count, immunohistochemical assessment of MIB-1 (Ki-67), proliferating cell nuclear antigen (PCNA), and silver-stained nucleolar organizer regions (AgNORs). Apoptosis was assessed using the in situ-end labelling (ISEL) technique in combination with morphological criteria. Patients who received liver transplantation were excluded. The results obtained were compared with histopathological stage (according to UICC), Edmondson grade, several other histopathological factors, and survival rate. Significant statistical correlations were seen between the mitotic index, the rate of nuclear positivity for MIB-1 and PCNA, and the number of AgNOR dots. In univariate survival analysis, tumour stage and Edmondson grade, mitotic index, MIB-1 and PCNA index, and mean AgNOR number were significant factors influencing patients' survival. On multivariate Cox survival analysis, mitotix index, concomitant cirrhosis, Edmondson grade, and patient age were the only significant independent prognostic factors. Apoptosis was not related to prognosis or to other parameters examined. These results indicated that mitotic index is an additional prognostic parameter which could provide auxiliary information for patients' outcome. MIB-1 and PCNA immunostaining and AgNORs showed a good correlation among themselves. Apoptosis did not predict prognosis in hepatocellular carcinoma.