Background Gemcitabine plus cisplatin is a standard treatment for stages ⅢB and Ⅳ nonsmall cell lung cancer (NSCLC). This randomized phase Ⅱ study evaluated a 3-week versus a 4-week schedule of gemcitabine-cisplatin as first line treatment for Chinese patients with advanced NSCLC.Methods Patients were randomized to receive cisplatin 75 mg/m2 on day 1 plus either gemcitabine 1250 mg/m2 on days 1 and 8 of a 21-day cycle (3-week group) or gemcitabine 1000 mg/m2 on days 1,8 and 15 of a 28-day cycle (4-week group).Results One hundred patients were enrolled in this study. The response rate was 24% (12/51 patients) in the 3-week group and 27% (13/49 patients) in the 4-week group. There were no statistically significant differences between the two treatment groups in survival (hazard ratio: 1.19; 95% CI: 0.68-2.09) with a median survival of 12.1 months and 13.8 months in the 3-week group and the 4-week group respectively. The rate of grade 3/4 toxicity in the 3-week group was 55% compared with 86% in the 4-week group (P=0.001). The difference in the incidence of grade 3/4 haematological toxicities did not reach statistical significance (3-week: 37%, 4-week: 57%), however grade 3/4 drug related neutropenia (3-week: 27%, 4-week: 51%) and thrombocytopenia (3-week: 8%, 4-week: 31%) were significantly lower in the 3-week group. Grade 3/4 nonhaematological toxicities were less in the 3-week group (33% cf 63%; P=0.005).Conclusions The differences in the efficacy endpoints were all in favour of the 4-week schedule of gemcitabine plus cisplatin, however these differences did not reach statistical significance. Fewer grade 3/4 toxicities were observed in the 3-week group compared with the 4-week group.