As a result of donation after circulatory death liver grafts’ poor tolerance to cold storage, there has been increasing research interest in normothermic machine perfusion. This study aims to systematically review the current literature comparing normothermic perfusion to cold storage in donation after circulatory death liver grafts and complete a meta-analysis of published large animal and human studies. A total of nine porcine studies comparing cold storage to normothermic machine perfusion for donation after circulatory death grafts were included for analysis. There was a significant reduction in AST (mean difference −2291 U/L, CI (−3019, −1563);
Liver transplant remains the only definitive therapy for end stage liver disease. However the shortage of quality organs remains significant in the United States with 1673 patients dying while on the waitlist and a further 1227 removed, too sick to undergo transplant during 2015 [
DCD grafts represent an important source of organs to expand the donor pool. The number of DCD grafts used continues to increase; however there is also a rise in the percentage of DCD grafts recovered but not transplanted [
Ex vivo perfusion is now being studied as a method of increasing use of DCD grafts. Studies using hypothermic and subnormothermic perfusion have shown promising results in both large animal [
The aim of this paper is to systematically review the current literature comparing NMP to SCS in DCD liver grafts in large animal (pig) and human studies. The secondary aim is to complete a meta-analysis of NMP versus SCS livers in published DCD porcine liver perfusions.
Searches were conducted in Ovid MEDLINE, OVID EMBASE, EBSCO CINAHL, WOS, SCOPUS, Proquest Dissertations and Theses, and PROSPERO by an expert librarian (SC) in June 2017 and updated in July 2017. Searches employed both controlled vocabularies (e.g., MeSH, EMTREE) and key words such as (DCD livers) and (ex vivo perfusion or normothermic perfusion). Search strategies were adapted for each database. Search strategies are available in the supporting information (
All full text, porcine, and human trials comparing NMP to SCS for the preservation of DCD livers were included for analysis. Studies that did not include DCD livers and those that focused only on hypothermic or subnormothermic machine perfusion were excluded.
Titles and abstracts from the primary search were reviewed independently by two authors (JN, DS) for studies that met inclusion criteria. When this was not clear from the titles and abstracts, full text articles were reviewed to determine inclusion.
Primary outcomes in ex vivo perfusion studies included assessment of alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels as markers of hepatocellular damage, as well as bile production and lactate clearance as markers of liver function. Secondary outcomes were histological preservation and hemodynamic stability indicated by hepatic arterial flow. Primary outcomes in orthotopic pig liver transplant studies included posttransplant peak AST, bile production, and graft survival. Secondary outcomes included histologic preservation. Where there was missing data for quantitative analysis, this information was requested via email from the publication corresponding authors. We received two responses but no further data for inclusion. Where possible, this data was estimated from published figures using Adobe Acrobat Reader DC software.
Articles were assessed by two authors (JN, DS) using the Systematic Review Centre for Laboratory Animal Experimentation (SYRCLE) risk of bias assessment tool [
A trained statistician performed statistical analysis. Outcomes assessed in the meta-analysis included AST, ALT, total bile production, and hepatic artery flow for perfusion studies, as well as peak AST for transplant studies. They are all continuous variables expressed as mean ± standard deviation (SD). The mean difference (MD) was used as a summary measure of efficacy between groups treated by NMP and SCS. When no SD was provided, a pooled SD was estimated as previously described [
Three hundred and eighty-six titles were identified through our primary search, with 228 remaining for screening after the removal of duplicates. Of these, 201 titles were excluded for the following reasons: published abstract with no complete full text article, comparison of hypothermic or subnormothermic perfusion without NMP, and studies without DCD grafts. Nine articles that directly compared cold storage to NMP for DCD grafts were included for analysis (Figure
Summary of pig liver perfusion study results.
Perfusion studies | Perfusate | Preservation time (hr) | Simulated transplant phase (hr) | WIT (min) | N-NMP | N- SCS | AST (U/L) NMP | AST (U/L) SCS | ALT (U/L) NMP | ALT (U/L) SCS | NMP total bile (ml) | SCS total bile (ml) | NMP HA flow (ml/min) | SCS HA flow (ml/min) |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Boehnert et al. 2013 | Steen | 4 SCS + 8 NMP vs. 12 SCS | 12 | 60 | 6 | 6 | __ | __ | | | __ | __ | | |
| ||||||||||||||
Liu et al. 2014 | Whole blood | 10 | 24 | 60 | 5 | 5 | | | | | | | | |
| ||||||||||||||
Banan et al. 2015 | Saline + whole blood | 6 | 2 | 40 | 3 | 3 | | | | | __ | __ | | |
| ||||||||||||||
Nassar et al. 2015 | Acellular solutions + whole blood | 10 | 24 | 60 | 15 | 5 | | | | | | | | 57 ± 14 ml/min/100 g |
| ||||||||||||||
Liu et al. 2016 | Steen + RBC | 10 | 24 | 60 | 5 | 5 | | | | | | | __ | __ |
| ||||||||||||||
Nassar et al. 2016 | Whole blood | 10 | 24 | 60 | 5 | 5 | | | | | | | __ | __ |
| ||||||||||||||
St Peter et al. 2002 | Whole blood | 24 | 24 | 60 | 4 | 4 | 259 | 3810 | | | __ | __ | 1400 ml/min | 440 ml/min |
Summary of pig orthotopic liver transplant studies.
Pig transplant studies | Preservation time (hr) | Duration of posttransplant monitoring | NMP | SCS | NMP peak AST (U/L) | SCS peak AST(U/L) |
---|---|---|---|---|---|---|
Schön et al. 2001 | 4 | 7 days | 6 | 6 | | |
Fondevila et al. 2011 | 4 | 5 days | 6 | 6 | | |
Boehnert et al. 2013 | 12 | 8 hours | 6 | 6 | | |
Study selection.
Pooled data showed a significant reduction in AST at the end of the simulated transplant phase in the NMP group relative to SCS (MD = −2291 U/L, CI (−3019, −1563);
Forest plots showing pooled AST, ALT, and bile production data from porcine liver perfusion studies.
Total bile production following the simulated transplant phase was significantly higher in the NMP group (MD = 174 ml, CI (155, 193);
There was insufficient data available to perform meta-analysis for lactate clearance.
Limited data was available for hepatic arterial flow. The NMP group did demonstrate higher flows, although this did not reach statistical significance (
Different histological scoring systems were used by different centers and thus were not suitable for meta-analysis. All perfusion studies showed less necrosis and improved architectural preservation in the NMP group relative to SCS [
Posttransplant peak AST was lower in the NMP group (MD = −1019, CI (−1276, −762);
Forest plots showing pooled peak AST data from porcine orthotopic liver transplant studies.
NMP groups demonstrated less necrosis, sinusoidal swelling, and improved overall architectural preservation relative to SCS groups [
The allocation process of animals was unclear in several studies [
The results of this review and meta-analysis must be interpreted with caution, as heterogeneity was high within the perfusion studies limiting the strength of conclusions that can be drawn. Experimental design for the included perfusion studies varied in several fundamental parameters. Major differences included surgical model, duration of preservation and reperfusion, and ex vivo circuit design.
Pigs used as liver donors were 30–40 kg and included landrace [
Boehnert et al. work was the only perfusion study to compare SCS to NMP following a period of SCS [
NMP was then carried out for either 6 [
One study included a dialysis circuit as part of the perfusion setup [
In porcine liver transplant models there was also significant study heterogeneity. The posttransplant observation period ranged from eight hours to seven days. Fondevila et al. [
NMP has shown capacity to recover function in discarded DCD human liver studies [
There are phase I clinical trials comparing NMP to SCS [
There was a large amount of heterogeneity amongst the small number of studies as outlined above. These significant differences in experimental design limit the strength of conclusions that could be drawn from meta-analysis. Furthermore, multiple data points included for meta-analysis were estimated from published figures which may differ slightly from the measured values.
Meta-analysis of published porcine perfusion studies demonstrates that NMP is superior to SCS regarding the preservation of liver architecture and function in DCD grafts. Given significant differences between studies, these results are to be taken with caution. Further study is still required in order to optimize and standardize perfusate composition and to evaluate NMP’s role in preservation following periods of cold storage. Clinical studies involving more DCD grafts will help bring this technology closer to clinical implementation. Economic factors need to be considered in subsequent studies to ensure feasibility within current healthcare systems.
The authors declare that there are no conflicts of interest regarding the publication of this article.
The authors would like to thank Chunhong Tian for statistical assistance.
Supplementary materials include the literature search strategy for this systematic review and meta-analysis (S1).