A preoperative tool for estimating the risk of blood transfusion over an ovarian cancer debulking surgery (113)

Abstract

<b>Objectives:</b> To develop a preoperative tool for estimating the risk for packed red blood cell transfusion during primary debulking surgery (PDS) for ovarian cancer and in the postoperative setting. This will help physicians identify a subgroup of patients that will benefit from surgical planning and blood-saving techniques. <b>Methods:</b> We conducted a single-center ovarian cancer database review of women undergoing PDS between 2001-2019. Exclusion criteria included patients with preoperative hemoglobin <10.0g/dl. Associations between the criteria and PDS outcomes were tested using the Wilcoxon Rank-Sum test for continuous variables and Fisher's exact test for categorical variables. A receiver operating characteristics (ROC) curve was generated to build a model predictive of transfusion; AUC was used as a measure of predictive accuracy. A multivariable regression model was developed to evaluate the odds of transfusion. These criteria were then assigned a risk of transfusion—a "BLood transfusion Over an Ovarian cancer Debulking Surgery (BLOODS)" score—according to their multivariate odds ratios (ORs). The total score of all patients in the cohort was subsequently calculated, and the odds of receiving a transfusion corresponding to each total score were determined. <b>Results:</b> A total of 1289 patients met the eligibility criteria. Median estimated blood loss was 650cc (range: 10-8150cc), with 375 patients experiencing a blood loss ≥1000cc (29%). A total of 1087 (84%) underwent optimal PDS (≤1cm of residual disease), 654 (51%) had complete gross resection. The presence of upper abdominal disease ≥1cm was most heavily associated with the need for transfusion intraoperatively and/or postoperatively (OR: 3.11, 95% CI: 2.47-3.91), followed Table by the presence of ascites (OR: 3.07, 2.44-3.87), albumin ≤3.5 (OR: 2.51, 1.76-3.59), carcinomatosis (OR: 2.27, 1.76-2.93), and CA-125 ≥600 (OR: 2.46, 1.95-3.10). On multivariable analysis, seven criteria were associated with blood transfusion: ASA ≥3 (1.05, 1.00-1.10), albumin ≤3.5 (1.26, 1.19-1.34), ascites (1.06, 0.98-1.15), CA-125 ≥600 (OR: 1.10, 1.04-1.16), carcinomatosis (1.01, 0.93-1.10), creatinine >0.90 (1.14, 1.07-1.22), upper abdominal disease ≥1cm (1.16, 1.07-1.25). A predictive (BLOODS) score was assigned to each of the seven criteria based on their multivariable ORs. A BLOODS score of ≥3 is projected to identify a rate of 85% of primary ovarian cancer cytoreductive cases requiring an intraoperative blood transfusion, 78% requiring a postoperative blood transfusion, and 80% overall. <b>Conclusions:</b> The BLOODS score can be used as a universal preoperative scoring system for patients with ovarian cancer undergoing PDS. The model estimates the risk of requiring a blood transfusion, which is directly proportional to ASA, albumin, ascites, CA-125, carcinomatosis, creatinine, and the presence of upper abdominal disease. This will help physicians with surgical planning and can benefit patient outcomes.