Fifty-three patients with recurrent and advanced stage (III and IV) non-small-cell lung cancer (NSCLC) were treated with a combination of bleomycin, etoposide (VP-16-213), and cis-diamminedichloroplatinum (BEP). Forty-eight patients were appraisable for response. The response rates were 44% for the entire group, 57% in 30 patients with combined squamous-cell and large-cell carcinoma, and 22% in 18 patients with adenocarcinoma (40%, 50%, and 19%, respectively, if patients not appraisable for response are included as nonresponders). The median survival time of patients with squamous-cell and large-cell carcinoma was slightly longer than that of patients with adenocarcinoma (23 weeks v 19 weeks). Patients with responsive disease survived significantly longer (median, 34 weeks) than did patients with unresponsive disease (median, 16 weeks) (P = .001). In the entire group, the median survival time of patients with an Eastern Cooperative Oncology Group (ECOG) performance status of 0 or 1 was better (23 weeks) than of those with a status of 2 or 3 (15 weeks), but this difference was not seen in the subgroup with squamous-cell and large-cell carcinoma (24 weeks v 23 weeks, respectively). Thus, the performance status was not of prognostic value in the histologic subgroups experiencing the best response rate. There were two treatment-related deaths, but otherwise the toxicity of BEP was acceptable. Only four of the 119 treatment cycles were followed by fever even though there was significant neutropenia (0.5 X 10(9)/L) after 20 of 97 treatment cycles. The majority of patients receiving BEP experienced relief of cough, hemoptysis, pain, and fatigue associated with their disease. There was a good correlation between objective responses and palliation of symptoms. Thus, BEP offers good palliation, particularly for patients with squamous-cell and large-cell lung cancer.