1. The aim of this study was to investigate the pharmacokinetics and the beta-adrenoceptor blocking activity according to time of conventional (C) and long acting (LA) propranolol in cirrhotic patients. Twenty-four patients with alcoholic cirrhosis and oesophageal varices were randomly assigned to receive either 160 mg C propranolol, 160 mg LA propranolol or placebo acutely and then following repeated administration (acute and chronic phases). Thereafter propranolol concentrations and beta-adrenoceptor blockade (resting and exercise heart rates) were measured at different intervals. 2. The Cmax was significantly higher with C propranolol in both phases. The time of Cmax was significantly later with LA propranolol in both phases. The AUCs were significantly higher after chronic administration with both formulations of propranolol. 3. The exercise peaks of beta-adrenoceptor blockade were similar between the two formulations and between the two phases of administration of propranolol. The duration of effective beta-adrenoceptor blockade was significantly longer in the chronic phase and seemed to be longer with LA than with C propranolol although this was not significant (72 +/- 31 vs 48 +/- 18 h, respectively). 4. There was a significant correlation between the log propranolol concentration and exercise heart rate but not with resting heart rate. No correlation could be demonstrated between pharmacological data and the Child Pugh score. 5. We conclude that in cirrhotic patients exercise testing was a reliable method in the assessment of beta-adrenoceptor blockade. Pharmacology of propranolol was found to be different according to the formulation or to the phase of administration.