This study compares two cytotoxic regimens comprising the same dose and schedule of cisplatin (CP) plus vinorelbine (VNR) or gemcitabine (GEM) administered under the same schedule to patients with advanced non-small cell lung cancers (NSCLC). From April 1998 to February 2003, 285 patients were randomised to receive either VNR 25 mg/m(2) on days 1 and 8 as an intravenous (i.v.) bolus plus CP 75 mg/m(2) on day 1 (regimen A) or GEM 1200 mg/m(2) on days 1 and 8 as an i.v. 30-min infusion plus CP 75 mg/m(2) on day 1 (regimen B). Both treatments were recycled every 21 days. If no progression had occurred after six cycles, the patients continued to receive VNR or GEM monochemotherapy weekly. Cross-over of the two single agents was considered if disease progression occurred. Objective response (OR), time to progression (TTP) and overall survival (OS) were analysed according to the intention-to-treat principle. 272 patients were ultimately eligible (137 on A and 135 on B). Their main characteristics were: male/female ratio 214/58; median age 63 (range 32-77) years; median Karnofsky Performance Status (PS) 80 (range 70-100); stage IIIB 34%, stage IV 61%, recurrent disease 5%; histology - epidermoid 29%, adenocarcinoma 53%, other NSCLC 18%. The characteristics of the patients in the two arms were well matched. The following response rates were observed in regimens A and B, respectively: complete response (CR) 0.7% and 3.7%, partial response (PR) 31.9% and 22.2% (P = 0.321). Median CR+PR duration was 8 months in both arms. Clinical benefit represented by an improvement in symptoms was evident in 25.7% and 28.1%, respectively. Median TTP was 5 months in both arms and median OS 11 months in both arms. Grade III-IV neutropenia occurred in 30.7% and 17.7% of the patients in arms A and B, respectively (P = 0.017); thrombocytopenia occurred in 0% and 9.3% (P = 0.004), respectively. No difference in the incidence of anaemia was observed. Non-haematological toxicity was generally mild: a higher incidence of grade 1-2 peripheral neurotoxicity and grade 1-2 local toxicity with regimen A and grade 1-2 liver toxicity with regimen B was reported. A pharmaco-economic comparison showed a difference between the two doublets, principally due to the different costs of VNR and GEM. Under the study conditions the combination of VNR or GEM with the same dose and schedule of CP produced similar OR, clinical benefits, TTP and OS in advanced NSCLC, and only mild toxicological differences were observed. Pharmaco-economic evaluation favoured the CP + VNR doublet.