Should Cell Salvage Be Used in Liver Resection and Transplantation? A Systematic Review and Meta-analysis

Abstract

Objective: To evaluate the effect of intraoperative blood cell salvage and autotransfusion (IBSA) use on red blood cell (RBC) transfusion and postoperative outcomes in liver surgery.

Background: Intraoperative RBC transfusions are common in liver surgery and associated with increased morbidity. IBSA can be utilized to minimize allogeneic transfusion. A theoretical risk of cancer dissemination has limited IBSA adoption in oncologic surgery.

Methods: Electronic databases were searched from inception until May 2021. All studies comparing IBSA use with control in liver surgery were included. Screening, data extraction, and risk of bias assessment were conducted independently, in duplicate. The primary outcome was intraoperative allogeneic RBC transfusion (proportion of patients and volume of blood transfused). Core secondary outcomes included: overall survival and disease-free survival, transfusion-related complications, length of hospital stay, and hospitalization costs. Data from transplant and resection studies were analyzed separately. Random effects models were used for meta-analysis.

Results: Twenty-one observational studies were included (16 transplant, 5 resection, n=3433 patients). Seventeen studies incorporated oncologic indications. In transplant, IBSA was associated with decreased allogeneic RBC transfusion [mean difference -1.81, 95% confidence interval (-3.22, -0.40), P =0.01, I 2 =86%, very-low certainty]. Few resection studies reported on transfusion for meta-analysis. No significant difference existed in overall survival or disease-free survival in liver transplant [hazard ratio (HR)=1.12 (0.75, 1.68), P =0.59, I 2 =0%; HR=0.93 (0.57, 1.48), P =0.75, I 2 =0%] and liver resection [HR=0.69 (0.45, 1.05), P =0.08, I 2 =0%; HR=0.93 (0.59, 1.45), P =0.74, I 2 =0%].

Conclusion: IBSA may reduce intraoperative allogeneic RBC transfusion without compromising oncologic outcomes. The current evidence base is limited in size and quality, and high-quality randomized controlled trials are needed.

FIGURE 1

Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA flow diagram of screened, included, and excluded studies.

FIGURE 2

Methodological Index for Nonrandomized Studies (MINORS) risk of bias assessment scores compiled from each included study.

FIGURE 3

A series of forest plot generated from included liver transplant studies for various reported outcomes comparing cell saver (CS) or IBSA use and control (no CS), presented as hazard ratios, with weighted mean difference (IV) and 95% confidence interval (CI). A, Forest plot for all studies that reported volume of intraoperative allogeneic RBC transfused (number of units), with further subgroup analysis of studies including malignancy only indications for transplant (1.2.1) and studies including a mix of benign and malignant indications for transplant (1.2.2). B, Forest plot for all studies that reported overall survival with further subgroup analysis of studies with unadjusted data (2.1.1) and data adjusted with either multivariate analysis or propensity score matching (2.1.2). C, Forest plot for all studies that reported disease-free survival.

FIGURE 4

A series of forest plot generated from included liver transplant studies for various reported outcomes comparing cell saver (CS) or IBSA use and control (no CS), presented as hazard ratios, with weighted mean difference (IV) and 95% confidence interval (CI). A, Forest plot for all studies that reported overall survival with further subgroup analysis of 2 studies with data adjusted with either multivariate analysis or matching (2.2.1) and 1 study with unadjusted data. B, Forest plot for all studies that reported disease-free survival with subgroup analysis of 2 studies with unadjusted data (3.2.1) and 1 study with multivariate adjusted data.