Etoposide is a schedule-dependent agent with greater activity against small cell lung cancer (SCLC) when a given dose is administered over several days compared with a 1-day administration of the same dose. In an attempt to capitalize on the schedule dependency of etoposide, 22 previously untreated extensive-stage SCLC patients were given cisplatin (100 mg/m2 on day 1) plus 21 days of low-dose, oral etoposide (50 mg/m2/d). Chemotherapy was repeated every 28 days for four cycles. Complete blood counts were monitored weekly, and etoposide was discontinued if either the leukocyte or platelet count dropped below 2000/microliters or 75,000/microliters, respectively. All 22 patients were evaluable for response; 18 had either a complete (9%) or partial response (73%), an overall response rate of 82% (95% confidence interval, 62% to 93%). The median response duration was 7 months, and the median survival was 9.9 months (range, 1 to 17+ months). Sixteen (73%) patients received all planned cycles of etoposide. In Cycle 1 of chemotherapy, the median leukocyte nadir was 2700/microliters (range, 100 to 6300/microliters), and median platelet nadir was 180,000/microliters (range, 51,000 to 397,000/microliters). Life-threatening leukopenia (less than 1000/microliters) was rare (3 of 74 cycles). There were three treatment-related deaths, only one of which was associated with neutropenia. One patient had mild renal insufficiency that resolved after discontinuation of therapy. Alopecia was observed in all patients, but other nonhematologic toxicities were uncommon. A randomized study is necessary to determine if this schedule of cisplatin and etoposide administration is superior to more standard methods. However, these data do not indicate a major survival benefit will be derived from increasing the duration of etoposide administration when used in combination with cisplatin given every 28 days.