Skeletal muscle protein loss occurs during marrow transplantation despite total parenteral nutrition. To determine if muscle atrophy could be minimized with exercise therapy, 30 patients undergoing marrow transplantation for acute leukemia completed a prospective randomized trial to receive: (1) no therapy (controls), (2) physical therapy thrice weekly (PT3), or (3) physical therapy five times weekly (PT5). Patients were studied through 35 days posttransplant. Muscle protein status and turnover was assessed by weekly nitrogen balance, and creatinine and 3-methylhistidine excretion. Results favored a muscle protein-sparing effect of exercise, as a significant decrease in creatinine excretion in controls only suggested muscle protein loss associated with inactivity. Changes in arm muscle area correlated with energy, but not protein intake. Large individual variation, inadequate nutritional support and differences in admission arm muscle area may have clouded these results.