Lactitol (beta-galactosido-sorbitol) is not absorbed in the small bowel but metabolized by colonic bacteria, and should therefore be as effective in the treatment of portal-systemic encephalopathy as lactulose (beta-galactosido-fructose). This hypothesis was tested in a 61-year-old alcoholic with an end-to-side portacaval anastomosis and chronic portal-systemic encephalopathy. Under controlled conditions he was switched from optimized treatment with lactulose to several regimens with lactitol (40-68 g/day), after which he was maintained on the new treatment for 1 year. On lactitol his condition was at least as good as on lactulose, but lactitol produced no taste aversion because it is less sweet. In addition, the patient no longer had nausea after taking the drug because lactitol can be supplied in a nonhygroscopic, chemically pure, crystalline form and therefore is less hyperosmotic than lactulose, which is supplied with contaminations of galactose and lactose. Obviously the data in single case represent only a feasibility study. Nevertheless, the outcome in this patient, together with the advantages of the new sugar, justify the planning of controlled clinical trials.