Between April 1985 and May 1988, we conducted a randomized study comparing two standard chemotherapy regimens with the same regimens given on an alternating basis in patients with small-cell lung cancer. The patients were randomly assigned to receive cyclophosphamide at a dose of 800 mg/m2 intravenously (IV) on day 1, doxorubicin at 50 mg/m2 IV on day 1, and vincristine at 1.4 mg/m2 IV on day 1 (CAV); cisplatin at 80 mg/m2 IV on day 1 and etoposide at 100 mg/m2 IV on days 1, 3, and 5 (PE); or CAV alternating with PE (CAV/PE). Each regimen was repeated every 3-4 weeks. Three hundred patients were entered in the study, and 288 of them were eligible for analysis (97 for CAV, 97 for PE, and 94 for CAV/PE). The response rates for PE (78%) and CAV/PE (76%) were significantly higher than the rate for CAV (55%), while the complete response rates were similar (14%, 16%, and 15%, respectively). Nine (23%) of 39 patients who failed to respond to the initial CAV regimen responded to PE when they were crossed over. In contrast, only one (8%) of 13 patients responded to CAV after failing to respond to the PE regimen, suggesting that these two regimens were partially non-cross-resistant. The response duration on CAV/PE was significantly longer than that with CAV (P = .004). The survival time with CAV/PE (11.8 months) was superior to that with CAV (9.9 months) (P = .027) or that with PE (9.9 months) (P = .056). In patients with limited disease, the survival in the alternating arm was significantly superior to the survival in the CAV arm (P = .014) or the survival in the PE arm (P = .023). The toxic effects were acceptable in all three chemotherapy regimens. These results favor the alternating chemotherapy over either standard chemotherapy, such as CAV and PE, although the differences are not dramatic.