|
|
Back to Annual Meeting Program
Perioperative Blood Transfusion Reduces Survival in Patients With Pancreatic Adenocarcinoma: a Multi-Institutional Study of 698 Patients
Jeffrey M. Sutton*1, David Kooby2, Gregory C. Wilson1, Dennis J. Hanseman1, Shishir K. Maithel2, David J. Bentrem3,4, Sharon M. Weber5, Clifford S. Cho5, Emily Winslow5, Charles R. Scoggins6, Robert C. Martin6, Hong Jin Kim7, Nipun Merchant8, Alex Parikh8, Daniel E. Abbott1, Michael J. Edwards1, Syed a. Ahmad1
1Surgery, University of Cincinnati College of Medicine, Cincinnati, OH; 2Surgery, Emory University School of Medicine, Atlanta, GA; 3Surgery, Northwestern University Feinberg School of Medicine, Chicago, IL; 4Surgery, Jesse Brown Veterans Affairs Medical Center, Chicago, IL; 5Surgery, University of Wisconsin School of Medicine and Public Health, Madison, WI; 6Surgery, University of Louisville School of Medicine, Louisville, KY; 7Surgery, University of North Carolina School of Medicine, Chapel Hill, NC; 8Surgery, Vanderbilt University School of Medicine, Nashville, TN
INTRODUCTION: In this multi-institutional study of patients undergoing pancreaticoduodenectomy (PD) for pancreatic adenocarcinoma, we sought to identify factors associated with a perioperative blood transfusion requirement. In addition, we investigated the hypothesis that receiving blood transfusion reduces long-term survival in this patient population.
METHODS: A retrospective chart review was performed across six high-volume institutions to identify patients who underwent PD between 2005 and 2010. Data collection included patient demographics, perioperative factors, transfusion status, and survival data. For statistical analysis, patients were then grouped according to whether they received 0, 1-2, or >2 units of packed red blood cells (pRBCs).
RESULTS: Among 698 patients identified, 168 (24%) required blood transfusion. 105 (15%) received 1-2 units and 63 (9%) received >2 units (range 0-25 units). Patient demographics associated with an increased transfusion requirement included age, smoking status, and heart disease (all p < 0.03). Operative variables associated with an increased transfusion requirement included operative time, estimated blood loss, tumor size, and R1/R2 margin status (all p < 0.03). Postoperative complications were not associated with transfusion requirement. However, those patients who received transfusions experienced a longer length of stay (p = 0.0009) as well as increased rate of readmission within 90 days (p = 0.002). The median survival of patients who received >2 units of pRBCs was significantly less than those who received either 0 or 1-2 units (10.2 months vs. 18.4 or 18.9 months, p = 0.0002). A multivariate model including margin status, nodal involvement, tumor size, and transfusion status identified the transfusion of >2 units of pRBCs as an independent predictor of reduced survival (HR 1.56, p = 0.03).
CONCLUSIONS: This multi-institutional study represents the largest series to date analyzing the effects of pRBC transfusion in patients undergoing PD for pancreatic adenocarcinoma. The transfusion rate in this series is less than what has been previously reported. Our data confirm that blood transfusion confers a negative impact on long-term survival in this patient population. These results can be utilized as a benchmark for future studies.
Back to Annual Meeting Program
|
