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

Liver surgery, including liver transplantation and resection, is technically challenging due to the complex and variable regional anatomy and rich hepatic blood supply.1 There have been increasing use of restrictive blood transfusion strategies and improvements in surgical techniques over time. Despite this, intraoperative blood loss requiring allogeneic red blood cell (RBC) transfusion remain an important clinical consideration. Rates of perioperative allogeneic RBC transfusion are variable, but remain elevated in patients undergoing liver transplantation (50.5%–62.6%)2,3 or resection (17%–23%).4

Although at times lifesaving, allogeneic RBC transfusions are associated with significant short- and long-term morbidity. Short-term complications are well described and include inadequate dosing, wrong product, and volume overload,5 as well as a small risk of viral or bacterial transmission, and allergic or immune transfusion reactions.6 Long-term complications include higher postoperative infection rates, prolonged hospital stay, higher rates of graft failure, and possibly more rapid cancer recurrence.7–9 RBCs are also limited, altruistically donated, and expensive resources, costing on average USD$761 per unit.10 Consequently, minimizing intraoperative blood loss and reducing allogeneic RBC transfusions are important considerations for hepatobiliary surgeons, and a priority recently highlighted by the World Health Organization (WHO) patient blood management strategy.11

Intraoperative blood cell salvage and autotransfusion (IBSA) has emerged as a cost-effective blood loss management strategy.12 It is increasingly utilized in surgery, and has been shown to decrease allogeneic RBC transfusions by up to 40% in adult elective cardiac and orthopedic surgery.13 IBSA systems collect blood from the operative field; this salvaged blood is then anticoagulated with heparinized saline or citrate, processed, and reinfused, salvaging ~60% to 80% of lost blood from surgery.14,15

IBSA adoption in oncologic surgery has been limited by concerns of reinfusion and dissemination of malignant cells, and potential implications on cancer-specific survival and recurrence.13,16,17 More recently, observational studies have demonstrated that IBSA can be used in oncologic surgery to decrease transfusion requirements without worsening oncologic outcomes.18–21 Furthermore, additional technologies including leukocyte depletion filters (LDF) can be incorporated with IBSA, potentially filtering out tumor cells from the salvaged blood during oncologic operations.22

Currently there is no consensus on the efficacy and safety of IBSA in liver surgery, and concerns about oncologic safety remain. Therefore, the objective of this work was to perform a systematic review and meta-analysis to define the impact of IBSA on intraoperative allogeneic RBC transfusion and postoperative outcomes in patients undergoing either liver resection or transplantation.

METHODS

A protocol was written and registered prospectively with the Prospective Register of Systematic Reviews (PROSPERO) (CRD42021231600). This review is reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines (Table, Supplemental Digital Content 1, https://links.lww.com/SLA/E79).23

Search Strategy and Data Sources

A search strategy was designed in collaboration with a dedicated information specialist with experience developing search strategies for systematic reviews. The following databases were searched from inception until May 2021: EMBASE Classic+, Ovid MEDLINE, The Cochrane Central Register of Controlled Trials, and Transfusion Evidence Library.24 References identified in relevant systematic reviews were screened for additional studies missed in the initial search. There were no language, date, or publication status restrictions. The search strategy is available in Document, Supplemental Digital Content 2, https://links.lww.com/SLA/E80.

Study Eligibility Criteria

The study population of interest included all patients undergoing liver surgery (either resection or transplantation) for any indication, including oncologic and nononcologic operations. Within this population, the implementation of any IBSA device, regardless of the addition of irradiation or LDF, and a corresponding comparator control arm (no IBSA use) was necessary for inclusion. The primary outcome was the proportion of patients receiving intraoperative allogeneic RBC transfusion, and the volume of intraoperative allogeneic RBC transfused. Secondary outcomes were prioritized into “core” secondary outcomes considered relevant to clinicians and patients in perioperative decision-making, emphasizing survival outcomes, perioperative complications, and resource utilization.25,26 These core secondary outcomes included overall survival (OS), disease-free survival (DFS), transfusion-related complications, length of hospital stay, and hospitalization costs. Other secondary outcomes included transfusion of other blood products [fresh frozen plasma (FFP) and platelets], and postoperative hemoglobin. Randomized controlled trials (RCTs) and observational cohort studies reporting at least 1 outcome of interest were eligible for inclusion. Case reports, case series with fewer than ten patients, and studies with no control group were excluded.

Study Selection and Data Extraction

Titles and abstracts were screened, and the full text of eligible publications were examined both independently and in duplicate by 2 reviewers. Reasons for full-text exclusion were documented. Disagreements were resolved by consensus or in discussion with a senior reviewer (G.M.). Conference abstracts were included as part of the gray literature, in the absence of related full text, to avoid publication bias.27 The snowballing technique was performed with manual review of the reference lists of included papers and relevant systematic reviews.28

Relevant data were extracted from included papers using a data extraction form (Table, Supplemental Digital Content 3, https://links.lww.com/SLA/E81). Data extraction was conducted independently and in duplicate. The data extraction form was piloted on five studies by the 2 reviewers, with modifications made to ensure complete and accurate data extraction and consensus in methodology.

The following study details and population demographic characteristics were extracted: study design, timeline, location, sample size, number in each intervention arm, use of adjustment techniques for baseline patient characteristics, age, presence of cirrhosis, Child-Pugh score, duration of follow-up, and average preoperative hemoglobin. Data on surgical indications, blood loss, cancer-specific variables (including number of lesions, size, tumor grade, differentiation, and vascular invasion), liver resection–specific variables (including number of major and minor resections and operative time), and liver transplant specific variables (including type of transplant, model of end stage liver disease score, and proportion transplanted within Milan criteria) were also extracted. Extracted outcome data included: proportion of patients receiving intraoperative autologous RBC transfusion, volume of autologous RBC transfusion (mL), proportion of patients exposed to allogeneic RBC transfusion, volume of allogeneic RBC transfusion (units or mL), postoperative hemoglobin (g/L), transfusion of other blood products (platelets, FFPin units or mL), postoperative complications, length of stay, hospital cost, OS, DFS, and disease recurrence.

Risk of Bias and Quality Assessment

Risk of bias was performed independently and in duplicate by 2 reviewers using the methodological index for nonrandomized studies (MINORS) criteria.29 on the basis of this instrument, each included study was assessed based on 12 items, and scored as 0 (if not reported), 1 (if reported but inadequate), or 2 (reported and adequate), with a maximum score of 24. Studies with a MINORS score of at least 17 were considered of high quality.29 Disagreements were resolved by consensus or by the senior author (G.M.). No study was excluded based on risk of bias. The Grading of Recommendations Assessment Development and Evaluation (GRADE) tool was used to assess the certainty of the evidence based on the study design, risk of bias, inconsistency of evidence, precision, directness, and overall effect.30

Data Synthesis and Statistical Analysis

The unit of analysis was the individual study participant. Primary and secondary outcomes were pooled for the meta-analysis when appropriate. Both unadjusted and adjusted outcome data were extracted. When both unadjusted and adjusted data were available for a given outcome, adjusted data were preferentially used for meta-analysis. Sensitivity analyses of adjusted data were also performed when possible. Dichotomous variables were summarized using proportions or odds ratios with 95% confidence intervals (95% CI). Continuous variables were represented by the mean, standard deviation, median, range, and interquartile range, as well as mean differences (MDs) with 95% CI. Time-to-event variables were summarized by hazard ratios (HR) and 95% CI. Liver transplants and liver resections were analyzed separately. Only studies with malignant surgical indication were included in analysis for oncologic outcomes (OS and DFS). For all other outcomes, subgroup analysis of benign versus malignant surgical indication was planned a priori. Preplanned sensitivity analyses included analysis of adjusted data and studies at low risk of bias (MINORS score ≥17).

RBC and platelet transfusions reported as a volume (mL) were converted to units by dividing by 300.31,32 FFP units reported as a volume were converted to units by dividing by 250.33 Where appropriate, means and SD were estimated from medians and ranges or interquartile ranges using the method of Wan et al.34 Where HRs were not reported, they were derived from Kaplan-Meier curves using the methods described by Tierney et al.35 All meta-analyses were performed using Review Manager 5 (The Cochrane Collaboration, Oxford, United Kingdom).36 Random effects models were used to account for expected between-study differences in populations, study designs, and transfusion strategies. The I² statistic was used to estimate statistical heterogeneity. Categories of low (0%–25%), moderate (25%–50%), and substantial (50%–100%) were used to interpret the I² statistic. The threshold for interpretation was defined according to the Cochrane Handbook for Systematic Reviews of Interventions.37 A P value of <0.05 was considered statistically significant.

RESULTS

A total of 2365 citations were screened. One study by Kwon et al24 shared an overlapping patient population with Han et al,38 and included only patients with advanced hepatocellular carcinoma (HCC) beyond transplant criteria. The former was excluded, whereas the latter was included. After full-text screening, 22 studies met eligibility criteria and were included in the review for data extraction (Fig. 1). Two studies, Fujimoto et al39 and Hirano et al40 included the same patient population, with the former reporting intraoperative blood loss, allogeneic and autologous transfusion, whereas the latter reported long-term survival outcomes. These results were collated and considered as a single study (Fujimoto/Hirano). This led to 21 unique patient populations for analysis.

F1FIGURE 1:

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

Study Characteristics

A total of 3433 patients were included, with 8 studies having fewer than 100 patients.41–48Table 1 provides the characteristics of the included studies. All included studies were observational cohorts, with all except one being retrospective in nature.45 Studies were conducted in heterogeneous patient populations. Eleven studies adjusted for confounding using multivariate analysis (n=6)16,44,49,53,54,56 propensity score matching (n=3)17,38,48 or simple matching (n=2).39,45 Sixteen studies included patients who underwent liver transplantation (2667 patients, 77.7%),16,17,38,41–47,49–54 whereas 5 studies included patients undergoing hepatic resection (766 patients, 22.3%).39,48,55–57

TABLE 1 - Study Characteristics of Liver Transplant and Liver Resection Populations References Design Adjustment Surgical Indication Intervention (% Autologous) Control (% Autologous) Cell Salvage Technology IBSA Indications Transplant Akbulut et al41 Retrospective None HCC Implied 100 0 Fresenius CATS NS Araujo et al49 Retrospective MV 1 degree hepatic malignancy Implied 100 0 Fresenius CATS+LDF Surgery/ anesthesia decision Brajtbord et al42 Retrospective None All 100 0 Haemonetics Cell Saver 3 NS Dos Reis et al50 Retrospective None All 100 0 Cell Saver All, excluding malignancy, sepsis Foltys et al16 Retrospective MV HCC 100 0 Haemonetics Cell Saver 5 All, excluding malignancy, unless major blood loss/ incidental Gupta et al43 Retrospective None HCC Implied 100 0 Autotransfusion + LDF NS Han et al38 Retrospective PSM HCC 100 0 Haemonetics Cell Saver 5+LDF All, autologous transfusion based on hemoglobin Ivanics et al17 Retrospective PSM Cirrhosis Proportion NS 0 Haemonetics Cell Saver 5 All, excluding malignancy; incl. incidental HCC Kim et al51 Retrospective None HCC 100 0 Haemonetics Cell Saver 5+LDF One center all IBSA (vs Control other center no IBSA) Lai et al44 Retrospective MV HCC NS 0 NS NS Massicotte et al52 Retrospective None All Proportion NS 0 Fresenius CATS All in period 2 (vs historic controls) Muscari et al45 Retrospective Matched HCC 100 0 Haemonetics Cell Saver 5 NS Pereira et al53 Retrospective MV All Proportion NS 0 NS NS Pinto et al54 Retrospective MV HCC Proportion NS 0 Haemonetics Cell Saver Elite Autotransfusion Blood loss >1000 mL or hemodynamic instability Sankarankutty et al46 Retrospective None All 100 0 Haemonetics Cell Saver 5 All, excluding HCC and sepsis Weller et al47 Retrospective None HCC 100 0 Haemonetics Cell Saver 5, CATS±irradiation Anesthesiology decision, tumor-free /treated HCC (5 cases) Resection Fujimoto/ Hirano et al39,40 Retrospective cohort Matched HCC 100* 0 Haemonetics Cell Saver All between 1988–1989 (vs Control 1986–1987) Jia et al55 Retrospective cohort None Benign disease 100 0; all pringle Haemonetics blood apheresis system <1 L blood loss, tumor >1 cm from large vessel; no malignancy Kang et al56 Retrospective cohort MV CRC metastases Proportion NS autologous only 0* Cell Saver ≥200 mL salvaged blood Perlmutter et al57 Retrospective cohort None CRC metastases 100 0 NS NS Zacharias et al48 Retrospective cohort PSM All Implied 100 0 Dideco Electa Essential Concept Cell Saver+LDF All after January 2013 (vs Control 2007–2012)

*Preoperative phlebotomy and intraoperative cell salvage blood transfused.

**Excluded all autologous and allogeneic blood transfusion.

CATS indicates continuous autologous transfusion system; CRC, colorectal cancer; MV, multivariate; NS, not specified; PSM, propensity score matching.

All studies reported using IBSA technology for blood salvage. Twelve studies reported some return of the salvaged blood back to patients (Table 1).16,17,38,42,45–47,50–54 Of these, 5 studies reported blood was returned back in only a proportion of the patients (32%–65%).17,52–54,56 Four studies implied that all patients in the IBSA group were transfused with the salvaged blood, by receiving the continuous autotransfusion system or autotransfusion with LDF.41,43,48,49 There was some variability in the intervention and control groups between studies; 1 study combined preoperative phlebotomy with IBSA in the intervention group.39 In another study, the Pringle maneuver was utilized as an alternative blood conservation strategy in the control group.55 Another study excluded all patients who received perioperative allogeneic RBC transfusion.56

The indication for IBSA use was recorded in fourteen studies (Table 1).16,17,38,39,46–52,54–56 Two studies specifically excluded transplant patients with malignancy or sepsis from IBSA use.46,50 Another 2 studies included patients with incidentally found HCC on explant pathology16,17 or those who had known HCC, but excessive intraoperative blood loss.16 One study included 5/39 (12.8%) patients in the IBSA group that were presumed to be tumor-free or who had treated HCC.47 Four studies had center or contemporary time period differences between intervention groups.39,48,51,52 In addition, 5 studies utilized LDF before autotransfusion.38,43,48,49,51 One study also had 2 subgroups of patients with or without utilization of irradiation for salvaged blood, which were combined into the intervention cell salvage group for data extraction and analysis purposes.47

Given the limited availability of studies reporting primary and secondary outcomes, subgroup analysis was only possible for surgical indication (malignant vs benign) and risk of bias (high vs low) for certain outcomes. Other prespecified subgroup analyses were not possible.

Liver Transplant Population Characteristics

Of the sixteen studies of patients undergoing liver transplantation, 7 reported the proportion of patients with cirrhosis (68%–100% IBSA vs 64%–100% control)16,17,41,45,46,49,52 (Table 2). Only 1 study reported a significant difference between the intervention groups, with a greater proportion of cirrhotic patients in the control group.49 Two studies reported a significant difference in Child-Pugh scores, with greater proportion of Child-Pugh class C patients in the IBSA group.16,51 Of the 6 studies that reported the proportion of patients within Milan criteria for transplantation, there was no significant differences between the intervention and control groups.16,38,41,45,49,51 Similarly, there was no significant differences between the 2 groups in model of end stage liver disease scores.17,38,41,49–52,54

TABLE 2 - Transplant Patient Demographics References Intervention N Cirrhosis (%) Child Score A/B/C (%) MELD Within Milan (%) Oncologic (%) Operative Time (min) Preoperative Hgb Blood Loss (L) Akbulut et al41 IBSA 24 100 25/46/29 14.5±0.9 33 100 Control 59 100 31/49/20 13.6±0.8 37 100 Overall 83 100 29/48/23 36 100 Araujo et al49 IBSA 122 85 10.5 [9–17] 100 100 Control 36 97 9 [8–13.5] 100 100 Overall 158 88 10 [8–15] 100 100 Brajtbord et al42 IBSA 22 Control 22 Overall 44 Dos Reis et al50 IBSA 345 31.3 Control 325 27.9 Overall 670 29.6±9.4 31.6 Foltys et al16 IBSA 40 93 30/18/52 60 100 Control 96 86 61/25/14 60 100 Overall 136 88 43/25/32* 60 100 Gupta et al43 IBSA 43 100 Control 51 100 Overall 94 100 Han et al38 IBSA 283 13±6 60 100 Control 114 12±7 62 100 Overall 397 61 100 Ivanics et al17 IBSA 76 71 1/33/66 21 [17–25] 100 483 [420–570]

留言 (0)

沒有登入
gif