Intestinal barrier function was prospectively examined in the course of a clinical trial evaluating the efficacy and safety of lisofylline for reducing cytotoxic therapy-induced intestinal epithelial damage-related infectious morbidity in patients receiving standard remission-induction therapy for acute myeloid leukaemia. The absorption and permeation of oral D-Xylose, lactulose and mannitol were measured weekly from baseline until marrow recovery in adult recipients of idarubicin plus cytarabine for untreated acute myeloid leukaemia. These studies were correlated with non-haematologic chemotherapy-related toxicities reflecting mucosal damage, including nausea, vomiting, stomatitis, diarrhoea, abdominal pain and systemic infection. D-xylose absorption decreased and lactulose:mannitol ratio reflecting intestinal permeability increased from baseline until the second and third week after the beginning of the treatment followed by recovery. These measures correlated with infection rates, nausea, vomiting, diarrhoea and increased blood product utilization. Lisofylline was associated with increased intestinal permeability, nausea, vomiting and infection-related morbidity despite a reduction in the duration of neutropaenia. These surrogates of intestinal barrier function correlated well with clinically important outcomes despite the failure to demonstrate reduced morbidity with lisofylline and represent useful objective outcome measurements for future clinical trials of products for the amelioration of the effects of cytotoxic therapy on the intestinal mucosa.