Ovarian carcinoma is the seventh most common cancer of women in the world and is responsible for the greatest number of deaths from gynaecological malignancy in Europe and North America. Although many studies have explored the use of chemotherapy in this disease, most individual trials have been too small to show clear benefit of one type of chemotherapy over another. The type and intensity of chemotherapy used routinely for women with advanced ovarian cancer has varied because of uncertainty about the effectiveness of the different regimens. The objective of this review was to compare single drugs versus combinations of drugs, platinum versus non-platinum, and carboplatin versus cisplatin-based chemotherapy in women with advanced ovarian cancer. Search strategy: We searched MEDLINE, and CancerLit bibliographic databases and the National Cancer Institute and the UK Co-ordinating Committee on Cancer Research registers of trials. We also handsearched the proceedings of meetings and contacted experts in the field and drug companies. Randomised trials of: (1) single non-platinum versus non-platinum combination chemotherapy (2) single non-platinum versus platinum combination chemotherapy (3) non-platinum regimen versus the same regimen plus cisplatin (4) single platinum versus platinum combination chemotherapy (5) cisplatin versus carboplatin-based chemotherapy in women with advanced ovarian cancer. Individual patient data were obtained from the trial investigators, checked by the reviewers and finally verified by the trial investigator. Main results: 49 trials involving 8763 women were included. The data were combined to calculate hazard ratios (HR) for survival on an intention-to-treat basis. For single non-platinum versus platinum combination chemotherapy the overall HR for survival was 0.93, 95% confidence interval (CI) 0.83-1.05 in favour of platinum-based combination chemotherapy. For non-platinum regimens compared with the same regimen plus cisplatin the survival HR was 0.88, 95% CI 0.79- 0.98 in favour of the addition of platinum to drug regimens. Single platinum compared with platinum combination therapy gave a HR of 0.91, 95% CI 0.79-1.05 in favour of combination chemotherapy. Cisplatin versus carboplatin gave a HR of 1.02, 95% CI 0.93-1.12. Sub-group analyses for age, stage, grade, histology, resection, bulk of residual tumour and performance status were undertaken for cisplatin versus carboplatin only. No difference in effect was found. The available evidence, although not conclusive, suggests that platinum-based chemotherapy is better than non-platinum therapy. There is some evidence that combination therapy improves survival compared with platinum alone. No difference in effect has been shown between cisplatin and carboplatin.