RESUMO
INTRODUCTION: Colorectal cancer liver metastasis (CRLM) occurs in upto 50% of cases and drives patient outcomes. Up-front liver resection is the treatment of choice in resectable cases. There is no consensus yet established as to the safety of intraoperative autotransfusion in liver resection for CRLM. METHODS: Patients undergoing curative-intent hepatectomy for CRLM at a single quaternary-care institution from 1999 to 2016 were included. Demographics, surgical variables, Fong Clinical Risk Score (FCRS), use of intraoperative auto and/or allotransfusion, and survival data were analyzed. Propensity score matching (PSM) was performed accounting for allotransfusion, extent of hepatectomy, FCRS, and systemic treatment regimens. RESULTS: Three-hundred sixteen patients were included. The median follow-up was 10.4 years (7.8-14.1 years). The median recurrence-free survival (RFS) and overall survival (OS) in all patients were 1.6 years (interquartile range: 0.63-6.6 years) and 4.4 years (2.1-8.7), respectively. Before PSM, there was a significantly reduced RFS in the autotransfusion group (0.96 vs. 1.73 years, p = 0.20). There was no difference in OS (4.11 vs. 4.44 years, p = 0.118). Patients in groups of FCRS 0-2 and 3-5 both had reduced RFS when autotransfusion was used (p = 0.005). This reduction in RFS was further found when comparing autotransfusion versus no autotransfusion within the FCRS 0-2 group and within the FCRS 3-5 group (p = 0.027). On Cox-regression analysis, autotransfusion (hazard ratio = 1.423, 1.028-2.182, p = 0.015) remained predictive of RFS. After PSM, there were no differences in FCRS (p = 0.601), preoperative hemoglobin (p = 0.880), allotransfusion (p = 0.130), adjuvant chemotherapy (p = 1.000), immunotherapy (p = 0.172), tumor grade (p = 1.000), use of platinum-based chemotherapy (p = 0.548), or type of hepatic resection (p = 0.967). After matching, there was a higher rate of recurrence with autotransfusion (69.0% vs. 47.6%, p = 0.046). There was also a reduced time to recurrence in the autotransfusion group compared with the group without (p = 0.006). There was no difference in OS after PSM (p = 0.262). CONCLUSION: Autotransfusion may adversely affect recurrence in liver resection for CRLM. Until further studies clarify this risk profile, the use of intraoperative autotransfusion should be critically assessed on a case-by-case basis only when other resuscitation options are not available.
