Postoperative infectious complications are nowadays a major problem in liver surgery. Better surgical outcomes with a consequent reduction in treatment and hospitalisation costs are a primary objective. The aim of this prospective, randomised study was to evaluate the cytokine response during and after portal clamping in patients undergoing liver resection and continuously fed with enteral nutrition as compared to patients receiving parenteral nutritional support. Forty patients with liver tumours were divided into two groups of 20 on the basis of the presence or absence of chronic liver disease. Furthermore, the latter group of 20 were randomised to two subgroups A and B of 10 patients on the basis of the different perioperative nutrition modalities. Group A patients were fed by so-called uninterrupted enteral nutrition, which means without interruption from the day before surgery with a nutritional solution delivered via a nasojejunal tube. The patients in group B were submitted to hepatic resection with parenteral nutritional support. Liver resection had to consist in resection of at least 30% of the parenchyma in non-cirrhotic patients or in segmental resection in cirrhotic ones. Ten milliliter blood samples were harvested before operation, and 10, 30 and 60 min after declamping and at 24 h. Interleukin 6 and a-tumour necrosis factor values were detected in blood samples. The values of C reactive protein and of prealbumin were recorded at 72 h postoperatively. We also evaluated postoperative complications, resumption of bowel movements, oral intake of nourishment, and patient discharge. Values in blood samples in the two groups showed a statistically significant difference in interleukin 6 values only after 24 h (10 min: group A 121 +/- 25.3, group B 156 +/- 31.4; after 24 h: group A 31.5 +/- 12, group B 105.1 +/- 24.1), while the a-tumour necrosis factor assay showed no significant difference between the two groups. However, there was an appreciably longer hospital stay (group A 10.9 +/- 3.1 days (range: 7-21 days), group B 13.2 +/- 2.7 days (range: 8-19 days) (P < 0.02) and a quicker resumption of bowel movements in group A. The data available show that uninterrupted enteral nutrition produces a modulation of the cytokine response following portal clamping. A lower cytokine activation cascade reduces the impact of the action of cytokines on the hepatic parenchyma with consequent enhancement of the hepatic Kupffer cell component. These factors thus substantially reduce the length of the patient's hospital stay and consequently the cost of medical care.