Nutrition support has been widely advocated as adjunctive therapy for a variety of underlying illnesses, including surgery and medical oncotherapy (radiation or chemotherapy for cancer). Both parenteral and enteral nutrition have been mistakenly viewed as feeding, when, in fact, they are medical interventions with associated risks and costs. The argument that nutrition support has to be provided to patients to prevent 'starving to death' confuses the difference between dying in a malnourished state and dying as a direct consequence of nutrient deprivation; cancer patients fit into the former category. As is true for any other medical intervention, efficacy is best established by randomized controlled clinical trials. When these forms of nutrition support have been so assessed, they have not usually been found to be any more efficacious than food on a tray or intravenous 5% dextrose solutions. In fact, parenteral nutrition actually caused harm in patients receiving medical oncotherapy (more total and infectious complications and fewer tumor responses). With regard to cancer patients, the only benefit that was demonstrated was the use of preoperative parenteral nutrition in patients undergoing attempted curative surgery for cancer of the upper gastrointestinal tract (esophagus, stomach, or pancreas). As nutrition support has associated complications (infections, mechanical problems with the tubes, and metabolic problems from the infusates) as well as costs, it cannot be recommended for cancer patients with the exception of the preoperative care of those with upper gastrointestinal malignancies and the occasional patient with gastrointestinal tract inadequacy owing to a slow-growing lesion.