Ackroyd SA, et al. - Given that blood transfusions in patients with ovarian cancer have been associated with poorer 30-day surgical morbidity and mortality, researchers conducted this retrospective database cohort study to determine risk factors for blood transfusion with packed red blood cell (pRBC) and construct a transfusion risk score to identify patients undergoing surgery for ovarian cancer who are at lowest risk for a blood transfusion. Findings showed a high incidence of blood transfusions observed in the perioperative period in patients undergoing hysterectomy for ovarian cancer. In the transfusion risk score proposed in this study, use of preoperative risk factors and planned surgical procedures could help predict anticipated blood requirements.

Methods

• Researchers used the National Surgical Quality Improvement Program database (NSQIP) from 2014-2016, to analyze data from 3,470 patients who had hysterectomy for ovarian cancer.
• They used logistic regression to model the link between risk factors with respect to 30-day postoperative blood transfusion.
• They developed a risk score to predict blood transfusion.

Results

• A total of 891 (25.7%) patients received a blood transfusion.
• Multivariate analysis revealed an independent association of blood transfusion with the following: age (OR: 1.90, P < 0.01), African American race (OR: 2.30, P < 0.01), ascites (OR: 1.89, P=0.02), preoperative hematocrit < 30% (OR: 10.70, P < 0.01), preoperative platelet count > 400 (x10⁹/L; OR: 1.75, P < 0.01), occurrence of disseminated cancer (OR: 1.71, P < 0.01), open surgical approach (OR: 7.88, P < 0.01), operative time > 3 hours (OR 2.19, P < 0.01) and additional surgical procedures, including large bowel resection (OR: 4.23, P < 0.01), bladder/ureter resection (OR: 1.69, P=0.02), and pelvic exenteration (P=0.02).
• With good discriminatory ability (C-statistic=0.80, P < 0.001, C-statistic=0.69, P < 0.001 for derivation and validation datasets, respectively) and calibration (Hosmer-lemeshow goodness-of-fit P=0.081, P=0.56 for derivation and validation datasets, respectively), a preoperative risk score using age, race, ascites, preoperative hematocrit and platelets, presence of disseminated cancer, planned hysterectomy approach and procedures enabled an accurate prediction of blood transfusion.