Purpose: From 1979 to 1984, 356 eligible patients with advanced or recurrent endometrial carcinoma no longer amenable to therapy with surgery, radiotherapy, or progestins were treated with doxorubicin alone or doxorubicin in combination with cyclophosphamide. Patients and Methods: Patients were randomized to receive doxorubicin 60 mg/m2 intravenously (IV) with or without cyclophosphamide 500 mg/m2 IV every 3 weeks for eight drug courses. All patients had received prior therapy with progestins with subsequent progression of disease. No patients had received prior therapy with cytotoxic drugs. Of 356 patients, 300 had measurable disease. Results: Among 132 patients treated with doxorubicin alone, there were seven complete responses (5%), 22 partial responses (17%), 73 with stable disease (55%), and 30 with increasing disease within 2 months of study entry (23%). For the 144 patients who received the combination, there were 18 complete responses (13%), 25 partial responses (17%), 75 with stable disease (52%), and 26 with increasing disease (18%). The median progression-free interval for those patients who received doxorubicin alone was 3.2 months, while it was 3.9 months for those who received the combination. The median survival duration for doxorubicin patients was 6.7 months, while it was 7.3 months for the combination patients. None of the unadjusted estimates of treatment differences are statistically significant. Prognostic features that had an impact on outcome included one factor associated with an increased likelihood of response (presence of measurable lung metastases) and four features associated with a poorer survival (poor performance status [PS] of 2 or 3, high pathologic grade, and presence of liver metastases or other intraabdominal disease). If these features are taken into account in multivariate analyses, there is no statistically significant evidence for differences in response rates (relative odds of response, 1.58; P = .06, one- tailed test), and survival duration is slightly longer in the combination regimen (17% reduction in death rate; P = .048). Conclusion: The combination of doxorubicin plus cyclophosphamide thus appears to offer a small advantage over doxorubicin alone in the management of endometrial carcinoma at the expense of more frequent and severe myelosuppression and gastrointestinal toxicity.