One hundred patients with advanced carcinoma undergoing highly cytotoxic chemotherapy were enrolled in a prospective randomized trial comparing subcutaneous G-CSF, thymopentin, a combination of the two, and placebo as preventive treatment of febrile leukopenia. Data from this study show that G-CSF was very active in reducing the incidence of chemotherapy-related fever and leukopenia as compared to placebo (22% versus 64%). This difference was statistically highly significant (P < 0.001). Thymopentin was associated with a reduction in febrile episodes as compared to placebo (52% versus 64%), but this difference did not reach statistical significance. Moreover, the addition of thymopentin to G-CSF did not result in a statistically significant improvement of results obtained with G-CSF alone. Similar results were achieved for fungal infections. Tolerance to thymopentin was excellent, while less than 9% of patients on G-CSF treatment complained of mild nausea and generalized bone pain.