General Session: Tumor, Trauma, and Infection
Presented by: B. Tan - View Audio/Video Presentation (Members Only)
A.S. Zaw(1), B. Tan(1), N. Kumar(1)
(1) National University Health System, Singapore, Singapore
Introduction: Despite advances in surgical techniques for spinal metastases, these procedures are often accompanied by substantial blood loss resulting in patients requiring blood transfusion either during intra or postoperative period. Allogeneic blood transfusion has been no doubt the main replenishment method for lost blood, and saving numerous lives. However, the effects of blood transfusion have been shown to be debatable in various oncological surgeries. We aimed to evaluate the influence of perioperative blood transfusion on the outcomes of patients undergoing spinal tumour surgery.
Methods: This retrospective study included 247 patients who underwent surgery for spinal metastases in our university hospital between 2005 & 2014. Outcome variables for this analysis were survival and postoperative complication rate among transfused and non-transfused patients. Survival was calculated in months from date of surgery till death. Cox regression analysis was performed for survival. Kaplan-Meier survival estimates were performed and log rank test was used to compare the estimates between the subgroups. Logistic regression analysis was exploited to determine the factors influencing postoperative complications.
Results: The overall median blood transfusion was 1 unit(0-10). Overall, 46 patients(19%) experienced at least one postoperative complication, of which 7 patients had more than one complication. The proportion of patients who developed any complication was significantly less for non-transfused patients than transfused patients (21% vs 42%, P=0.01). In univariate analysis, significant variables for developing any complication were blood transfusion units, preoperative ECOG score, preoperative Hb level and number of vertebral metastases. Multivariate analysis revealed that increased amount of blood transfusion units (OR=2.56, 95% CI: 2.01-2.88,P=0.01) was independent predictor of any postoperative complication. Overall median survival was 15 months;95% CI:11-21 months. Patients who received blood transfusion had decreased median survival compared to those who were not transfused (11 vs 21 months). Log rank test revealed that the difference in the survival rates between transfused and non-transfused patients, however, was not significant (P=0.24). Generally, there was a trend towards lower survival rates among transfused compared to non-transfused patients at 6-month, 12-month and 18-month post-operation though the differences were not significant. Univariate Cox regression analyses showed that preoperative ECOG, gender, primary tumour and amount of blood transfusion units have significant influence on survival. In multivariate model, only the primary tumour and preoperative ECOG were significant predictors of survival.
Conclusions: Increased number of blood transfusion units was associated with postoperative complications. Blood transfusion was associated with decreased survival in univariate analysis but not in multivariate analysis. Revisiting blood transfusion trigger in patients with spinal metastases is recommended. Patient blood management strategies including autologous transfusion in these high-risk groups of patients should be explored further for minimizing ABT use and its related potential risks.