Prospective randomly controlled trials have failed to demonstrate the clinical efficacy of providing nutritional support to most cancer patients in terms of morbidity, mortality, and duration of hospitalization. Serious shortcomings in study design have limited the possibility of drawing definitive conclusions from the data. Thus, nutritional intervention needs to be seen as a method of support, with the aim of maintaining nutritional and functional status during the stress of the oncology treatment to prevent or attenuate cachexia. There is no disease during which the patient benefits from prolonged wasting. Pretreatment weight loss is quoted as a major indicator of poor survival and response to therapy of cancer patients. As a consequence, an early and serial assessment of nutritional status, perhaps followed by an immediate intervention with nutritional support is strongly recommended. There are other specific reasons for using the gut rather than the intravenous route for nutrient administration besides the often reported disadvantage of significant cost. Local intestinal stimulation prevents the mucosal atrophy and bacterial translocation that can be triggered by several precipitating factors, as frequently seen in oncologic patients. These include endotoxin, radiation therapy, cytotoxic and immunosuppressive drugs, cytokines, bowel and biliary obstruction, broad-spectrum antibiotics, and the tumour itself, as well as parenteral nutrition (PN). As the enteral route of nutritional support has been found to be as good as or preferable to PN in terms of maintenance of nutritional status or immune function, prevention of bacterial translocation, maintenance of normal gut flora, transit and histology, and prevention of hypercatabolic responses to stressful events, it is always preferable in terms of physiological response, local and systemic competence, quality of life and cost, and should be the method of choice for the nutritional support of cancer patients. Although retrospective studies of PN suggest a benefit for patients with cancer who are undergoing surgery, radiation, or chemotherapy, carefully designed, prospective studies report less conclusive findings. The failure of conventional PN to improve clinical outcomes in patients with cancer may be related to the fact that standard formulations do not address or reverse abnormalities of intermediate metabolism that result in cancer cachexia. Supplemental substances have been proposed in an attempt to improve the efficacy of PN, including insulin, growth hormone and branched chain amino acids. The difficult task is to identify those patients who are at risk for malnutrition and at the same time identify the subset of patients who will benefit clinically from parenteral nutritional repletion. Severe malnutrition in patients requiring surgery, bone marrow transplantation in patients unable to tolerate enteral supplementation and postoperative complications necessitating nutritional support are specific indications. Routine use of PN should be discouraged.