Although standard supportive care for bone marrow transplant (BMT) recipients includes total parenteral nutrition (TPN), it has not been shown that this is the most appropriate method of nutritional support. To determine whether current BMT recipients require TPN during the early recovery period, we conducted a prospective, randomized clinical trial comparing TPN and an individualized enteral feeding program (counseling, high protein snacks and/or tube feeding). Nutritional assessment included measurement of serum proteins, anthropometry, and body composition analysis. For the latter, total body water and extracellular fluid were measured by standard radioisotope dilution techniques and used to quantitate body cell mass and body fat plus extracellular solids (FAT + ECS). In 27 TPN patients, body composition 28 days after BMT, expressed as a percentage of baseline, was body cell mass, 100%, extracellular fluid, 108%, FAT + ECS, 108%, and in 30 enteral feeding program patients, was body cell mass, 93%, extracellular fluid, 104%, and FAT + ECS, 94%. Only the difference in FAT + ECS was statistically significant (p less than 0.01). Compared to the enteral feeding program, TPN was associated with more days of diuretic use, more frequent hyperglycemia, and more frequent catheter removal (prompted by catheter-related complications), but less frequent hypomagnesemia. There were no significant differences in the rate of hematopoietic recovery, length of hospitalization, or survival, but nutrition-related costs were 2.3 times greater in the TPN group. We conclude that TPN is not clearly superior to individualized enteral feeding and recommend that TPN be reserved for BMT patients who demonstrate intolerance to enteral feeding.