A randomized double blind clinical comparison of neomycin and lactulose was performed in 33 cirrhotic patients with chronic portal-systemic encephalopathy (PSE) at seven cooperating hospitals. In order to maintain double blindness, sorbitol syrup was used as a control solution along with neomycin and was compared with lactulose syrup and placebo tablets in a double drug protocol. Twenty-nine patients were studied in a crossover investigation in which each received both therapeutic regimens preceded and followed by control periods. Four additional patients received one or the other agent, but did not receive both. Serial, semiquantitative assessments were made in all patients of mental status, asterixis, and the trailmaking test (TMT) and electroencephalograms (EEG) and arterial ammonia levels. Both neomycin-sorbitol and lactulose were effective in the majority of patients (83 and 90%, respectively). Each of these parameters (mental state, asterixis, TMT, EEG, and NH3) was improved significantly by neomycin-sorbitol and lactulose. The post-treatment levels for each of these measures were similar in the neomycin and lactulose-treated groups. Mean stool pH was reduced by neomycinsorbitol to 6.1 and by lactulose to 5.5. This difference was highly significant statistically. Bowel activity was similar in the two groups. Both drugs were free of toxicity. These investigations demonstrate that both lactulose and neomycin-sorbitol are effective in the treatment of chronic portal-systemic encephalopathy.