Meta-analysis of cisplatin, doxorubicin, and cyclophosphamide versus cisplatin and cyclophosphamide chemotherapy of ovarian carcinoma.

Objective To compare the survival with cisplatin, doxorubicin (Adriamycin), and cyclophosphamide versus that of cisplatin and cyclophosphamide in women with advanced epithelial ovarian cancer, to evaluate the effect of dose intensity, and to evaluate meta-analysis methodology. Methods Meta-analysis was done on 30 studies of 2060 women with stages III and IV epithelial ovarian cancer. All had 3-year survival data, adequate follow-up, no other chemotherapy, no radiation therapy, and had information for various prognostic variables (age, stage, grade, and residual disease). We used four different methods of meta-analysis: pooled published data and modified effect-size analyses of the entire group (30 studies), and pooled published data and effect-size analyses of the subset of five prospective randomized studies. Results Three-year survival for the entire group was 43% for cisplatin, doxorubicin, and cyclophosphamide versus 36% for cisplatin and cyclophosphamide; for the five prospective randomized studies, the rates were 46 and 35%, respectively. The survival advantage of cisplatin, doxorubicin, and cyclophosphamide was statistically significant when analyzed by the pooled published data and modified effect-size meta-analysis of the entire group and the pooled published data meta-analysis of the five prospective randomized studies. The effect-size meta-analysis of the five prospective studies did not reach statistical significance. Total dose intensity and doxorubicin dose intensity were not significantly associated with survival advantage in cisplatin, doxorubicin, and cyclophosphamide use. Conclusions There seems to be a survival advantage to treatment with cisplatin, doxorubicin, and cyclophosphamide versus treatment with cisplatin and cyclophosphamide. We believe this to be due to the properties of multiagent chemotherapy (the addition of doxorubicin) rather than to increased dose intensity. In addition, we believe that physicians need to familiarize themselves with meta-analysis methodology.