Etoposide is a highly schedule-dependent agent. We previously reported that a 21-day schedule of oral etoposide had good activity in small cell lung cancer (SCLC). The current phase II study was designed to test the combination of 21-day oral etoposide with cisplatin in hopes of capitalizing on etoposide's schedule dependency. Sixteen extensive stage SCLC patients were treated with cisplatin 100 mg/m2 day 1 plus 21 days of low-dose oral etoposide 50 mg/m2/day. Chemotherapy was repeated every 28 days for 4 cycles. Blood counts were monitored weekly, and etoposide was discontinued if the leukocyte or platelet count dropped below 2.0 x 10(9)/l or 75 x 10(9)/l, respectively. Fifteen of 16 patients were evaluable for response; 13 achieved either a complete (13%) or partial response (73%), for an overall response rate of 86% (95% confidence interval, 62-93%). Median response duration was approximately 7 months; median survival was not reached. Thirteen patients (81%) received all the planned cycles of chemotherapy. In cycle 1 of chemotherapy, the median leukocyte nadir was 2.8 x 10(9)/l (range, 0.1-6.3 x 10(9)/l; median platelet nadir was 180 x 10(9)/l (range, 51-397 x 10(9)/l). Life-threatening leukopenia (less than 1.0 x 10(9)/l) was unusual (2 of 58 cycles). There was 1 treatment-related death. One patient developed mild renal insufficiency that resolved after therapy. Nonhematologic toxicities were uncommon, but alopecia occurred in all patients. These data do not suggest that a major survival benefit will be derived for patients with extensive stage SCLC by increasing the duration of etoposide administration when used in combination with cisplatin. A randomized study is needed to determine if this long-term schedule of etoposide plus cisplatin is superior to the standard schedule of etoposide plus cisplatin.