Anastomotic leakage following bowel resection in radical cytoreductive surgery for ovarian cancer: A subgroup analysis of the prospective AGO-OVAR.OP3/LION.

Abstract

5566 Background: Anastomotic leakage (AL) is an important and severe complication following bowel resection in cytoreductive surgery (CRS) for ovarian cancer (OC). Identifying patients (pts) at risk for AL could improve clinical management and reduce frequency and severity of complications. Methods: AGO-OVAR.OP3/LION intraoperatively randomized 647 pts with advanced OC (FIGO IIB-IV) following complete cytoreduction with unsuspicious lymph nodes to either undergo systematic pelvic and paraaortic lymphadenectomy (LNE) or not. All pts who underwent bowel resection were included in this analysis. Potential prognostic impact of AL was analyzed by Kaplan-Meier method and log-rank tests regarding progression-free (PFS) and overall survival (OS). Risk factors for AL in a subset of clinicopathological parameters were evaluated by calculating odds ratio (OR) with 95% confidence interval (CI) in univariate and multivariate logistic regression models. A stepwise variable selection algorithm using the Akaike Information Criteria (AIC) for identification of the final logistic regression model was applied. P values presented were two-tailed, and P<0.05 was considered as statistically significant. Results: Among all 647 randomized pts, AL was noted in 24 (3.7 %) pts. The AL rate of the 336 pts with bowel surgery during CRS (median age 61.3 years; 316 large bowel, 94 small bowel resections) was 7.1% (24/336 pts). Pts following stoma formation had an AL rate of 5.5% (3/55 pts) compared to 7.5% (21/281 pts) in pts without stoma. Median PFS was 18 months in pts with AL versus 19 months in the no-AL-group (hazard ratio (HR) 0.86; 95% CI 0.5 to 1.4, P=0.53), median OS was 31 months in the AL group and 58 months in the no-AL-group (HR 0.69; 95% CI 0.4 to 1.2, P=0.17). The clinicopathological characteristics “volume of blood loss” (OR 1.05 per 100cc; 95% CI 1.01-1.10) and “LNE vs non-LNE” (OR 4.10; 95% CI 1.60-12.62) were identified as factors potentially predictive for AL in univariate analysis, and both factors (volume of blood loss [OR 1.04 per 100cc, 95% CI 1.0001-1.09], LNE vs non-LNE [OR 3.67, 95% CI 1.41-11.39]) remained a significant independent factor in multivariate analysis. Conclusions: Considering the high surgical complexity in this large prospective surgical trial, the overall rate of AL following bowel surgery was relatively low and had no significant impact on PFS or OS. While protective stoma formation was not identified as protective factor in this cohort, volume of blood loss and LNE procedure were clinical parameters associated with higher risk of AL. Being potential surrogates for extensive surgery, these factors should be considered in perioperative management in pts with advanced OC. Further specific factors predicting AL were not identified. Clinical trial information: NCT00712218 .