Ibrahim Sultan1, Valentino Bianco2, Edgar Aranda-Michel2, Arman Kilic3, Derek Serna-Gallegos3, Forozan Navid3, Yisi Wang4, Thomas G Gleason3. 1. Division of Cardiac Surgery, Department of Cardiothoracic Surgery, University of Pittsburgh, Pittsburgh, Pa; Heart and Vascular Institute, University of Pittsburgh Medical Center, Pittsburgh, Pa. Electronic address: sultani@upmc.edu. 2. Division of Cardiac Surgery, Department of Cardiothoracic Surgery, University of Pittsburgh, Pittsburgh, Pa. 3. Division of Cardiac Surgery, Department of Cardiothoracic Surgery, University of Pittsburgh, Pittsburgh, Pa; Heart and Vascular Institute, University of Pittsburgh Medical Center, Pittsburgh, Pa. 4. Heart and Vascular Institute, University of Pittsburgh Medical Center, Pittsburgh, Pa.
Abstract
OBJECTIVE: To report long-term outcomes after deep hypothermic circulatory arrest (DHCA) with or without perioperative blood or blood products. METHODS: All patients who underwent proximal aortic surgery with DHCA from 2011 to 2018 were propensity matched according to baseline characteristics. Primary outcomes included short- and long-term mortality. Stratified Cox regression analysis was performed for significant associations with survival. RESULTS: A total of 824 patients underwent aortic replacement requiring circulatory arrest. After matching, there were 224 patients in each arm (transfusion and no transfusion). All baseline characteristics were well matched, with a standardized mean difference (SMD) <0.1. Preoperative hematocrit (41.0 vs 40.6; SMD = 0.05) and ejection fraction (57.5% vs 57.0%; SMD = 0.08) were similar between the no transfusion and blood product transfusion cohorts. Rate of aortic dissection (42.9% vs 45.1%; SMD = 0.05), hemiarch replacement (70.1% vs 70.1%; SMD = 0.00), and total arch replacement (21.9% vs 23.2%; SMD = 0.03) were not statistically different. Cardiopulmonary bypass and cross-clamp time were higher in the blood product transfusion cohort (P < .001). Operative mortality (9.4% vs 2.7%; P = .003), stroke (7.6% vs 1.3%; P = .001), reoperation rate, pneumonia, prolonged ventilation, and dialysis requirements were significantly higher in the transfusion cohort (P < .001). In stratified Cox regression, transfusion was an independent predictor of mortality (hazard ratio, 2.62 [confidence interval, 1.47-4.67]; P = .001). One- and 5-year survival were significantly reduced for the transfusion cohort (P < .001). CONCLUSIONS: In patients who underwent aortic surgery with DHCA, perioperative transfusions were associated with poor outcomes despite matching for preoperative baseline characteristics.
OBJECTIVE: To report long-term outcomes after deep hypothermic circulatory arrest (DHCA) with or without perioperative blood or blood products. METHODS: All patients who underwent proximal aortic surgery with DHCA from 2011 to 2018 were propensity matched according to baseline characteristics. Primary outcomes included short- and long-term mortality. Stratified Cox regression analysis was performed for significant associations with survival. RESULTS: A total of 824 patients underwent aortic replacement requiring circulatory arrest. After matching, there were 224 patients in each arm (transfusion and no transfusion). All baseline characteristics were well matched, with a standardized mean difference (SMD) <0.1. Preoperative hematocrit (41.0 vs 40.6; SMD = 0.05) and ejection fraction (57.5% vs 57.0%; SMD = 0.08) were similar between the no transfusion and blood product transfusion cohorts. Rate of aortic dissection (42.9% vs 45.1%; SMD = 0.05), hemiarch replacement (70.1% vs 70.1%; SMD = 0.00), and total arch replacement (21.9% vs 23.2%; SMD = 0.03) were not statistically different. Cardiopulmonary bypass and cross-clamp time were higher in the blood product transfusion cohort (P < .001). Operative mortality (9.4% vs 2.7%; P = .003), stroke (7.6% vs 1.3%; P = .001), reoperation rate, pneumonia, prolonged ventilation, and dialysis requirements were significantly higher in the transfusion cohort (P < .001). In stratified Cox regression, transfusion was an independent predictor of mortality (hazard ratio, 2.62 [confidence interval, 1.47-4.67]; P = .001). One- and 5-year survival were significantly reduced for the transfusion cohort (P < .001). CONCLUSIONS: In patients who underwent aortic surgery with DHCA, perioperative transfusions were associated with poor outcomes despite matching for preoperative baseline characteristics.