Resting energy expenditure is abnormal in most patients with cancer and may contribute to cancer cachexia. These metabolic abnormalities may be a direct measure of tumor metabolism, or represent alterations in the size or activity of the body cell mass, or both. To unravel this pathogenesis, we prospectively studied 68 preoperative patients with cancer about to undergo curative resection by measuring resting energy expenditure before and after tumor resection. The preoperative measured resting energy expenditure was compared with expected resting energy expenditure based on Harris-Benedict resting energy expenditure predictions: 10 patients were hypometabolic (less than 90% Harris-Benedict); 35 were normometabolic (90% to 110% Harris-Benedict); and 23 were hypermetabolic (greater than 110% Harris-Benedict). Using each patient as his or her own control, resting energy expenditure normalized or remained normal following curative resection. In contrast, after palliative resection, resting energy expenditure remained hypermetabolic or significantly increased toward hypermetabolism. Tumor induces an abnormal metabolic rate, since tumor removal results in prompt normalization of resting energy expenditure. The abnormal energy expenditure of patients with cancer cannot be solely attributed to abnormal host body composition.