IP235. Transfusion Timing and Postoperative Myocardial Infarction and Death in Patients Undergoing Common Vascular Procedures

Abstract

Perioperative blood transfusion, specifically of packed red blood cells (pRBCs), after vascular surgery procedures is one modifiable risk factor that is associated with increased cardiovascular events and 30-day mortality. The aim of this study was to evaluate the effect of transfusion timing (intraoperative vs postoperative) on the rate of postoperative myocardial infarction (POMI) and death. Seven surgical and endovascular modules within the Vascular Quality Initiative from 2013 to 2017 were reviewed at a single institution. Transfusion data on elective and urgent cases were abstracted, and all inpatients underwent routine postoperative troponin and electrocardiography testing. The primary end point was POMI using the universal definition of MI. The secondary end point was 30-day mortality. Multivariable logistic regression analysis was used to evaluate the risk of transfusions on POMI and death. We identified 1154 cases for analysis (299 endovascular aneurysm repairs; 117 infrainguinal bypasses; 127 open abdominal aortic aneurysm repairs; 41 suprainguinal bypasses; 168 thoracic endovascular aortic repairs; and 402 peripheral vascular interventions, of which 90 had femoral endarterectomy). Overall, the POMI rate was 2% and mortality 1.4%. POMI and death rates did not differ by type of procedure (P = .053 for MI and P = .423 for mortality). The mean number of intraoperative pRBC and postoperative pRBC transfusions was higher for patients with POMI (intraoperative, 0.3 vs 1.3; postoperative, 0.4 vs 1.8; both P < .01) and death (intraoperative, 0.3 vs 1.4; postoperative, 0.4 vs 2.5; both P < .01). In addition, age and coronary artery disease was associated with MI on univariable analysis. On multivariable analysis for POMI, coronary artery disease (odds ratio [OR], 5.42; 95% confidence interval [CI], 2.09-14.06; P < .001) and postoperative transfusion (OR, 4.9; 95% CI, 1.92-12.49; P < .001) were both significant; however, intraoperative transfusion was not (OR, 1.37; 95% CI, 0.49-3.83; P = .55). On multivariable analysis, postoperative death was associated with increasing age of the patient (OR, 1.08; 95% CI, 1.01-1.15; P = .017), and statin use was highly protective (OR, 0.26; 95% CI, 0.095-0.74; P = .012), but intraoperative or postoperative transfusion was not associated with death after adjustment. In our series with routine postoperative troponin screening, the use of postoperative but not intraoperative pRBC transfusion was associated with POMI. We did not identify an association of transfusion with postoperative death. These data suggest that the perioperative setting of transfusions is important in its impact on postoperative outcomes and needs to be accounted for in evaluating transfusion outcomes and indications.