Variant liver oestrogen receptor transcripts in hepatocellular carcinoma are associated with aggressive clinical course and unresponsiveness to tamoxifen. To evaluate the impact on survival and on tumour growth of megestrol (progestin drug acting at post-receptorial level) we enrolled 45 patients with HCC characterized by variant liver oestrogen receptors in a prospective, randomized study with megestrol vs. placebo. Presence of variant oestrogen receptors was determined by RT/PCR. 24 patients were randomized to no treatment and 21 to therapy with megestrol 160 mg day−1. Results were analysed by Kaplan-Meier and Cox methods. Survival of hepatocellular carcinoma characterized by variant oestrogen receptors was extremely poor (median survival 7 months); megestrol significantly improved survival (18 months) (P = 0.0090). Tumour growth at one year was significantly slowed down in megestrol-treated patients (P = 0.0212). Bilirubin levels, presence of portal thrombosis, HBV aetiology and treatment were identified at univariate analysis as factors significantly associated with survival; at multivariate analysis, only megestrol therapy (P = 0.0003), presence of HBV infection (P = 0.0009) and presence of portal vein thrombosis (P = 0.0051) were factors independently related with survival. (1) Megestrol slows down the aggressive tumour growth of patients with hepatocellular carcinoma characterized by variant estrogen receptors and (2) is also able to favourably influence the course of disease, more than doubling median survival. © 2001 Cancer Research Campaignhttp://www.bjcancer.com