OBJECTIVE: To assess the association between patient-disease characteristics and overall survival (OS) after second-line intraperitoneal (IP) treatment of ovarian cancer. METHODS: Data were aggregated from 7 Gynecologic Oncology Group (GOG) phase 2 studies conducted between 1988 and 1995 to evaluate IP therapy for partially responsive or recurrent disease but no tumor masses greater than 0.5 cm. Factors evaluated include age, performance status, extent of residual disease, tumor grade, and histologic diagnosis. RESULTS: A total of 432 eligible women were treated on 1 of the 7 second-line phase 2 IP trials. The median OS was 2.4 years (range of individual study medians, 1.9-2.9 years). Relative to women with grade 1 cancers, those with grade 2 or 3 cancers experienced 1.82 (95% confidence interval [CI], 1.21-2.74) and 2.02 (95% CI, 1.35-3.03) times greater instantaneous death rates, respectively, and those with clear cell adenocarcinoma experienced death rates 6.00 (95% CI, 3.27-10.9) times greater. The extent of residual disease, surgically assessed before starting study treatment, was also associated with OS. Relative to those who had no evidence of gross disease, those with gross disease that was completely resected, or with unresectable disease (not larger than 0.5 cm), experienced death rates 1.74 (95% CI, 1.25-2.42) and 2.26 (95% CI, 1.67-3.05) times greater, respectively. CONCLUSIONS: There are patient and disease characteristics strongly associated with survival after second-line IP treatments. These factors are relevant to clinicians considering IP therapy outside the investigative setting and for the development of future studies in this area.