OBJECTIVE: The purposes of this study were to assess modified radical hysterectomy including systematic pelvic and para-aortic lymphadenectomy followed by adjuvant chemotherapy in patients with para-aortic lymph node (PAN) metastasis in endometrial carcinoma and to identify the multivariate independent prognostic factors for long-term survival during the past 10 years. METHODS: Between December 1987 and December 2002, we performed modified radical hysterectomy with bilateral salpingo-oophorectomy including systematic pelvic and para-aortic lymphadenectomy and peritoneal cytology in 284 endometrial carcinoma patients according to the classification of the International Federation of Gynecology and Obstetrics (stage IA, n = 66; stage IB, n = 96; stage IC, n = 33; stage IIA, n = 5; stage IIB, n = 20; stage IIIA, n = 28; stage IIIC, n = 28; and stage IV, n = 8) who gave informed consents at our institute. Patients with tumor confined to the uterus (stages IC and II) were treated by 3 courses of cyclophosphamide 750 mg/m2, epirubicin 50 mg/m2, and cisplatin 75 mg/m2 regimen 3 to 4 weeks apart, and patients with extrauterine lesions involving adnexa and/or pelvic lymph node (PLN) were treated by 5 courses. In addition, 10 courses were given to patients with PAN metastasis. Patients with PLN metastasis received adjuvant chemotherapy, and adjuvant radiation was not part of our institutional protocol. For multivariate regression modeling with proportional hazards, the regression model of Cox was used. Survival curves were analyzed by the Kaplan-Meier method, and analysis of the differences was performed by the log-rank test. RESULTS: The overall incidence of retroperitoneal lymph node metastasis assessed by systematic pelvic and para-aortic lymphadenectomy was 12.0% (34/284) in stages I to IV endometrial carcinoma, and incidences of PLN and PAN metastases were 9.2% (26/284) and 7.4% (21/284), respectively. However, PAN metastasis rate is 50% (13/26) in patients with PLN metastasis. Univariate analysis of prognostic factors revealed that International Federation of Gynecology and Obstetrics clinical stage (P < 0.0001), histological finding (P = 0.0292), myometrial invasion (P < 0.0001), adnexal metastasis (P < 0.0001), lymphovascular space invasion (P < 0.0001), tumor diameter (P = 0.0108), peritoneal cytology (P = 0.0001), and retroperitoneal lymph node metastasis (P < 0.0001) were significantly associated with 10-year overall survival. Survival was not associated with age (P = 0.1558) or cervical involvement (P = 0.1828). A multivariate analysis showed that adnexal metastasis (P = 0.0418) and lymphovascular space invasion (P = 0.0214) were significantly associated with 10-year overall survival. The 5- and 10-year overall survival rates in patients with negative PAN were 96% and 93% versus 72% and 62% in patients with positive PAN (P = 0.006). CONCLUSIONS: It is suggested that surgery with systematic pelvic and para-aortic lymphadenectomy followed by adjuvant chemotherapy could improve long-term survival in patients with PAN metastasis, although there are only 21 patients with PAN metastasis.