Treatment of patients affected by metastatic gastric cancer with low performance status (PS) is a very hard choice. It is mandatory to define a very well-tolerated schedule to be employed in these subgroup of patients. From June 1999 to December 2001, 21 patients (pts) affected by metastatic gastric cancer with low performance status (> or = 2 ECOG) were treated with bimonthly "de Gramont" schedule. Treatment was planned to perform 6 courses of chemotherapy for each patient plus other 2-4 if a response had been documented. A total of 161 courses of de Gramont schedule was administered to the 21 pts enrolled. We observed 8 PD (38%), 8 SD (38%), 5 objective responses (24%--2 MR, 3 PR). Duration of objective responses (OR) was 5 months, 3 months for 3 PRs and 2 and 1 months for two MRs respectively. At time of observation (June 2002) median overall survival (OS) was 14 months, median survival from the starting de Gramont schedule was 8 months. Toxicity was very mild: grade 3 leukopenia in 1 pt, grade 1-2 anemia and piastrinopenia in 3 pts, grade 1-2 nausea vomiting in 5 pts, grade 1 diarrhea in 4 pts, grade 3 mucositis in 2 pts. No other side effect was renowned. PS ameliorated in 12 (57%) pts, even if a major response was not noted. de Gramont schedule can be safely and effectively employed in metastatic gastric cancer pts with very low performance status.