The purposes of this study were to compare the relationships between para-aortic lymph node metastasis and various clinicopathologic factors to evaluate whether para-aortic lymph node dissection is necessary when treating endometrial cancer. A retrospective study was performed on 841 patients with endometrial cancer, who underwent the initial surgery at the Keio University Hospital. Clinicopathologic factors related to para-aortic lymph node metastasis significant on a univariate analysis were analyzed in a multivariate fashion using a logistic model. According to the multivariate analysis, the clinicopathologic factor most strongly related to the existence of para-aortic lymph node metastasis was positive pelvic lymph node metastasis (P < 0.01). Among the 155 patients who underwent pelvic and para-aortic lymph node dissection, the difference of 5-year overall survival by the presence of retroperitoneal lymph node metastasis was examined by Kaplan-Meier method. The prognosis was poor even if para-aortic lymph node dissection was performed in cases of positive para-aortic lymph node metastasis. In conclusion, when deciding whether to perform para-aortic lymph node dissection in patients with endometrial cancer, it is necessary to consider the pelvic lymph nodal status. If there is no pelvic lymph node metastasis, it could not be necessary to perform para-aortic lymph node dissection.