When response rates to antineoplastic therapy are good and the associated nutritional morbidity is high, TPN is recommended when the gastrointestinal tract is unavailable for use. Patients should not be denied adequate cancer treatment because of the fear of complications secondary to malnutrition. On the other hand, prolongation of pain and suffering for the terminally ill cancer patient is obviously not an indication for TPN. Appropriate clinical judgment is required in proper selection of the TPN candidate. Malnutrition is almost always a prerequisite, as is failure of the gastrointestinal tract. Therapies designed to direct nutritional support to preferentially benefit the host and "starve" the tumor have had only limited short-term success in animal models. The use of diets that are glutamine- and arginine-supplemented are only now being examined in animal experiments. Since TPN allows total and complete control of exogenous nutrients, it should be used to explore alternative therapies in the form of nutrient composition. For example, if tumors are preferential glutamine utilizers, then TPN-enriched solutions might be employed to increase the number of cells in a phase of mitosis that makes them more susceptible to attack by cell-cycle specific chemotherapeutic agents. Stimulation of tumor growth by TPN in patients is not scientifically documented. Only anecdotal reports appear in the literature. Thousands of cancer patients worldwide have received TPN. Were stimulation of tumor growth to be a major problem with TPN solutions as currently formulated, it should be known by now.