This study has been undertaken to investigate if the intravenous (i.v.) infusion of fat emulsions may be associated with impairment of some immunological functions thus increasing the risk of septic complications. Fifteen malnourished patients with advanced gastric or esophageal cancer received for 2 weeks preoperatively and 1 week after surgery an isocaloric and isonitrogenous TPN treatment with Intralipid (group A: n=8) or glucose alone (group B: n=7) as energy substrate. Cluster analysis of 11 nutritional parameters and some tests of the humoral and cellular immunity (IgG, IgM, C3c, Factor B; polymorphonuclear (PMN) cells, total lymphocytes, T and B lymphocyte counts; 'in vitro' PMN chemotaxis, adherence to nylon fibers, phagocytosis of latex particles) were sequentially determined. The incidence and severity of post-operative infections were investigated and a 'sepsis score' was calculated for each patient. Pre- and postoperative TPN were not associated with an improvement of the nutritional status. The humoral and cellular immune parameters showed the same behaviour in patients receiving Intralipid and in controls. The chemotactic activity of PMN cells was constantly normal, granulocyte adherence fluctuated below the normality range in controls, whereas phagocytosis of latex was similar in both groups. Post-operative infectious episodes were less severe in patients receiving Intralipid. Our results do not confirm that Intralipid adversely affects some aspects of the humoral and cellular immune response.