Beta adrenoreceptor blocking drugs have been used for the prevention of haemorrhage from oesophageal varices. However, it is possible that these agents, by virtue of their effects on hepatic blood flow, may impair liver function and precipitate hepatic encephalopathy. We have therefore studied the effect of the beta blocking drug propranolol on hepatic encephalopathy in patients with cirrhosis and portal hypertension. Twenty patients were randomly assigned to receive 4 weeks treatment with propranolol or an identical-looking placebo, the former given in a dose sufficient to reduce resting pulse rate by greater than or equal to 25%. Before and after treatment patients were assessed for the severity of liver disease and the presence of encephalopathy. EEG mean cycle frequency and fasting arterial ammonia concentrations were also measured, and in order to detect latent hepatic encephalopathy, each patient underwent a battery of psychometric tests. Patients were blinded as to their treatment, as were those assessing their responses. Neither propranolol nor placebo had any significant effect on the parameters measured. On propranolol median EEG mean cycle frequency fell from 9.08 ct s-1 (range 8.63-11.0 ct s-1) to 8.73 ct s-1 (range 8.27-11.44 ct s-1), and median fasting arterial ammonia concentration fell from 66 mumol litre-1 (range 40-329 mumol litre-1) to 49 mumol litre-1 (range 37-188 mumol litre-1). Psychometric test values, while initially abnormal and suggestive of latent hepatic encephalopathy in the majority of patients, did not change significantly during the study.