Four randomized clinical trials comparing cyclophosphamide plus cisplatin (CP) versus cyclophosphamide, doxorubicin, and cisplatin (CAP) individually failed to show a significant survival difference in the treatment of ovarian carcinoma. However, by pooling 1,194 patients from these trials in a meta-analysis, there is a statistically significant survival benefit for CAP (P = .02); in addition, there is a significant advantage for CAP in frequency of negative second-look laparotomy (CAP, 30%; CP, 23%; P = .01). Because the dose intensity of CAP was greater than CP in three of the trials, it remains unresolved to what extent the benefit of CAP is from greater dose intensity and to what extent it is from the doxorubicin itself. Either interpretation suggests directions for improving the chemotherapy of ovarian carcinoma.