Since publication of the results of the Gynecologic Oncology Group (GOG) III study, the combination of cisplatin and paclitaxel (Taxol; Bristol-Myers Squibb Company, Princeton, NJ) has been adopted widely as the new standard for treating advanced ovarian cancer. Further attempts to optimize first-line chemotherapy with platinum and taxanes include substituting carboplatin for cisplatin, individualizing the carboplatin dose by calculating it according to the area under the concentration-time curve, and reducing the length of the paclitaxel infusion. Attempts to optimize platinum/paclitaxel combinations have led to the initiation of several small phase I/II trials evaluating the carboplatin/paclitaxel combination. The promising results of these studies have prompted the initiation of three phase III trials comparing carboplatin/paclitaxel with the standard combination of cisplatin/paclitaxel. An interim analysis after 1 year's accrual to the prospectively randomized German Arbeitsgemeinschaft Gynäkologische Onkologie (AGO) study is presented. Treatment consists of paclitaxel 185 mg/m2 infused over 3 hours on day 1 followed directly by either cisplatin 75 mg/m2 (arm B) or carboplatin dosed to an area under the curve of 6 (arm A). Treatment is repeated every 3 weeks for six courses. Eligibility criteria are epithelial ovarian cancer International Federation of Gynecology and Obstetrics stage IIB through IV, age of consent, written informed consent, Eastern Cooperative Oncology Group performance status < or =2, life expectancy of more than 12 weeks, adequate bone marrow function defined as neutrophil count 1.5 x 10(9)/L and platelet count > or =100 x 10(9)/L, adequate renal function defined as glomerular filtration rate (GFR) > or =60 mL/min, and adequate liver function defined as serum bilirubin levels within 1.25 x upper limit of normal. From October 1995 to December 1996, 442 of 660 planned patients were recruited to the AGO study. The interim analysis is based on data from 353 patients who were enrolled within the first study year. These preliminary data indicate that hematologic toxicity occurred more frequently in arm A (carboplatin/paclitaxel), while nonhematologic toxicity occurred slightly more frequently in arm B. Dose-intensity analysis did not reveal cumulative dose reductions or increasing use of colony-stimulating factors over subsequent courses in either arm. In all, 44 patients with measurable disease following surgery completed chemotherapy and were evaluable for response. The data remain blinded at this time, and results are reported for the group as a whole. So far, there have been 18 (41%) complete responses and 15 (34%) partial responses, for an overall response rate of 75%. Retrospective comparison with the GOG results reveals no significant difference in response rates between patients in the cisplatin/paclitaxel arm of GOG III and those in the AGO study: the GOG study reported a 73% response rate, compared with a preliminary 75% response rate in the AGO study, resulting in a relative risk of 1.03 (95% confidence interval, 0.83 to 1.27). Overall, this interim analysis did not reveal any reason to terminate this study early. Accrual is ongoing and is expected to be completed in 1997.