This multicenter retrospective study is based on 192 patients with advanced ovarian cancer in pathological complete response at second-look surgery. Ninety-four (48.9%) patients developed recurrent disease after a median time of 18 months (range, 4-89 months) from surgical reassessment. The recurrence involved the pelvis in 45 (47.9%) cases, the abdomen in 42 (44.7%), the retroperitoneal lymph nodes in 13 (13.8%), and distant sites in 20 (21.2%). On the whole series, 5- and 7-year disease-free survival rates after negative second-look were 47.4 and 44.5%, respectively. By log-rank test the disease-free survival rate was related to FIGO stage (P = 0.008), tumor grade (P = 0.0021), size of residual disease after initial surgery (P = 0.0038), and type of second-look (laparoscopy vs laparotomy, P = 0.0061), but not to histological type and first-line chemotherapy. Cox proportional hazard model showed that tumor grade, size of residual disease, and type of second-look were independent prognostic variables for disease-free survival. The risk ratio of relapse was 2.386 (95% CI, 1.140-4.990) for grade 2 and 3.118 (95% CI, 1.515-6.416) for grade 3 compared to grade 1 disease. For patients with residual disease 1-2 cm and > 2 cm the risk ratio was, respectively, 1.877 (95% CI, 1.117-3.156) and 2.156 (95% CI, 1.324-3.511) compared to patients with residual disease < 1 cm. The risk ratio was 1.826 (95% CI, 1.121-2.973) for patients who were submitted to a laparoscopic second-look compared to those who underwent a laparotomic reassessment. Poorly differentiated grade and large residual disease after initial surgery are the strongest prognostic variables for recurrence after a negative second-look.