The chronic toxicity and possible carcinogenicity of the sugar replacer isomalt was studied in Wistar rats and Swiss mice. Groups of 50 animals of each sex were fed 0, 2.5, 5 or 10% isomalt in the diet for nearly 2.5 yr (rats) or 2 yr (mice). Control groups received either basal diet with 10% maize starch or basal diet with 10% sucrose. Additional groups of ten rats/sex were fed the same diets and were killed after 1 yr. Isomalt and sucrose were included in the diet at the expense of maize starch. Administration of isomalt was started, in rats, in utero, and in mice, at weaning age. Feeding isomalt did not affect the appearance or behaviour of rats or mice, nor did it cause diarrhoea. Mortality rate was unaffected. Body weights of rats and mice fed 10% isomalt were generally slightly lower than those of controls. Periodic examinations of rats for haematological criteria, clinical chemistry of the blood, urine composition and kidney function did not reveal any changes of toxicological significance. Periodic haematological examinations of mice were likewise negative. Caecal enlargement was observed in rats and mice of the high-dose group, but the microscopic structure of the caecal wall was unaffected. An increased number of treated male and female rats showed hyperplasia of the urothelium in the renal pelvis accompanied by mineralization, whereas the number of females showing corticomedullary mineralization was decreased in the treated groups. The incidence, type or location of neoplasia provided no evidence of a carcinogenic potential of isomalt. Feeding 10% sucrose did not induce significant differences compared with the controls fed 10% maize starch, whereas isomalt at levels of up to 10% produced some of the changes that are common to rats fed high levels of poorly digestible carbohydrates.