Background: In East Asia, S-1 + CDDP (SP) has been employed as first-line therapy for advanced gastric cancer (AGC) from the results of SPIRITS trial. Patients who were resistant to chemotherapy with S-1 in the first-line treatment were widely treated with taxane alone or CPT-11 alone as the second-line treatment. On the other hand, the response rate of combination therapy with S-1 is higher than that of CPT-11 alone. Then, we hypothesized that S-1 + CPT-11 prolongs survival in the second-line treatment comparing with CPT-11 alone after failure in the first-line treatment with S-1. (NCT00639327). Methods: Patients with AGC who confirmed disease progression by imaging after the first-line therapy with SP, S-1 + cocetaxel or S-1 alone except S-1 + CPT-11 were allocated into S-1 plus CPT-11 group (Group A) or CPT-11 alone group (Group B) as second-line chemotherapy. Patients who were relapsed to adjuvant chemotherapy with S-1 were not enrolled. Primary endpoint was overall survival, and secondary endpoints were progression free survival, response rate and adverse events. Results: From March 2008 to June 2011, 304 patients were enrolled, and 294 were eligible for analysis. The overall survival was 8.8 months (M) in the Group A and 9.4M in the Group B. There is no statistically significant difference in both groups (P=0.9156). The progression free survival was 4.8M in the Group A and 4.9M in the Group B (P=0.1568). The response rate was 7.6% in the Group A and 7.4% in the Group B. Grade 3 or higher leukopenia, neutropenia and febrile neutropenia were observed more frequently in the Group A than in the Group B. Conclusions: From our results, we do not recommend consecutive use of S-1 as second-line treatment in patients who are refractory to S-1 in first-line chemotherapy. Clinical trial information: 00639327.