The relation between splenectomy and survival time after curative total gastrectomy for gastric cancer was reviewed retrospectively on 492 patients treated at nine hospitals between 1989 and 1993. Altogether 260 patients underwent splenectomy, and 232 patients did not. A univariate analysis revealed that the survival time of patients with splenectomy was significantly less than those without splenectomy (p = 0.0265). In a subgroup of our patients stratified to adjust for the stage of disease, there was no significant difference between the survival rates. Splenectomy remained insignificant according to the multivariate analysis using Cox's proportional-hazard regression. The splenectomy group was associated with more risk factors (e.g., T3/T4 tumors, positive nodes, stage greater than III, large tumor size) that are powerful predictors of death due to gastric cancer. In a separate multivariate analysis after eliminating those who had a T4 tumor invasion or a N2 nodal positivity from the analysis (or both), splenectomy again remained insignificant. In conclusion, we could not find any beneficial effect of splenectomy in gastric cancer patients in this retrospective multivariate analysis. We can presume that splenectomy cannot increase the survival rate so long as the splenectomy group has more risk factors than the nonsplenectomy group. Therefore randomized prospective clinical trials using more precise criteria to indicate the need for splenectomy are needed to assess whether splenectomy is beneficial.