Parity and time interval since last birth have been found to be associated with the risk of endometrial cancer, but few previous studies have examined whether these reproductive factors have prognostic impact. We examined this issue among 740 nulliparous and 3355 parous endometrial cancer patients diagnosed in Norway during the period 1961-1999. The mean age at diagnosis was 55.7 years (range, 25-74 years). Hazard ratios (HRs) with 95% confidence intervals were calculated in Cox proportional regression models. Parous women had significantly (P < 0.001) better prognosis than nulliparous women. In analyses adjusted for age at diagnosis, clinical stage, and histological type, the HR for parous versus nulliparous women was 0.68 (95% confidence interval, 0.57-0.82). The prognostic impact of parity was restricted to patients with endometrioid tumors (P = 0.014, test for interaction) and appeared to be most pronounced in women without metastases (P = 0.14, test for interaction). Moreover, the improved prognosis was strongest for women with the shortest time interval since last childbirth. The HRs (nulliparous women as reference) were 0.51, 0.60, and 0.80 for women less than 15, 15 to 24, and 25 years or more since birth (P < 0.001). The observed beneficial effect of pregnancies may be related to a strong exposure to progesterone during pregnancy. However, it is possible that tumors developing in nulliparous and parous women have different biological features. Further knowledge on this issue may provide valuable information that can be used for individualized treatment.