National Trends in Utilization of Normothermic Machine Perfusion in DCD Liver Transplantation

Background. In liver transplantation, advances in ex situ normothermic machine perfusion (NMP) have improved outcomes compared with traditional static cold storage (SCS) in donation after circulatory death (DCD) organs. We aimed to characterize trends in the utilization of NMP versus SCS in DCD liver transplantation in the United States. Methods. This retrospective cohort study used data from the United Network for Organ Sharing database to identify recipient–donor adult liver transplant pairs from DCD donors from January 2016 to June 2022. Utilization of NMP and changes in donor risk index (DRI) and components between NMP and SCS were assessed across transplant year eras (2016–2018, 2019–2020, and 2021–2022). Statistical comparisons were made using the Kruskal-Wallis test or the chi-square test. Results. A total of 3937 SCS and 127 NMP DCD donor transplants were included. Utilization of NMP ranged from ~0.4% to 3.5% from 2016 to 2021 and rose significantly to 11.2% in early 2022. Across transplant eras, median DRI increased significantly for SCS and NMP, but the magnitude of the increase was larger for NMP. With NMP DCDs, there were significant increases in median donor age, national share proportion, and “cold ischemic time” over time. Finally, there was a shift toward including higher DRI donors and higher model for end-stage liver disease score transplant recipients with NMP in later transplant eras. Conclusions. In recent years, NMP utilization has increased and expanded to donors with higher DRI and recipients with higher model for end-stage liver disease score at transplant, suggesting increasing familiarity and risk tolerance with NMP technology. As NMP remains a relatively new technique, ongoing study of patient outcomes, organ allocation practices, and utilization patterns is critical.


National Trends in Utilization of Normothermic Machine Perfusion in DCD Liver Transplantation
Samir Abu-Gazala, MD, 1 Helen Tang , MD, 2 Peter Abt, MD, 1 and Nadim Mahmud , MD, MS, MPH, MSCE 3,4,5,6 Background.In liver transplantation, advances in ex situ normothermic machine perfusion (NMP) have improved outcomes compared with traditional static cold storage (SCS) in donation after circulatory death (DCD) organs.We aimed to characterize trends in the utilization of NMP versus SCS in DCD liver transplantation in the United States.Methods.This retrospective cohort study used data from the United Network for Organ Sharing database to identify recipient-donor adult liver transplant pairs from DCD donors from January 2016 to June 2022.Utilization of NMP and changes in donor risk index (DRI) and components between NMP and SCS were assessed across transplant year eras (2016-2018, 2019-2020, and 2021-2022).Statistical comparisons were made using the Kruskal-Wallis test or the chi-square test.Results.A total of 3937 SCS and 127 NMP DCD donor transplants were included.Utilization of NMP ranged from ~0.4% to 3.5% from 2016 to 2021 and rose significantly to 11.2% in early 2022.Across transplant eras, median DRI increased significantly for SCS and NMP, but the magnitude of the increase was larger for NMP.With NMP DCDs, there were significant increases in median donor age, national share proportion, and "cold ischemic time" over time.Finally, there was a shift toward including higher DRI donors and higher model for end-stage liver disease score transplant recipients with NMP in later transplant eras.Conclusions.In recent years, NMP utilization has increased and expanded to donors with higher DRI and recipients with higher model for end-stage liver disease score at transplant, suggesting increasing familiarity and risk tolerance with NMP technology.As NMP remains a relatively new technique, ongoing study of patient outcomes, organ allocation practices, and utilization patterns is critical.
(Transplantation Direct 2024;10: e1596; doi: 10.1097/TXD.0000000000001596.) L iver transplantation (LT) is a life-saving procedure for eli- gible patients with cirrhosis or liver malignancies, but the demand for organs exceeds availability.Thus, efforts to expand the number and viability of organs in the donor pool are critical. 1mong these, donation after circulatory death (DCD) liver donation has emerged as an important modality.3][4] In particular, a recent study by Ibeabuchi et al 3 demonstrated that utilization of NMP for DCD donation was associated with a reduced hazard of graft failure and a lower incidence of need for retransplantation.As DCD outcomes continue to improve, DCDs are an increasing source of donor livers in the United States and Europe. 5However, to date, it remains unclear to what extent NMP utilization for DCD has expanded and whether the characteristics of recipients have evolved as experience with NMP accumulates.To address these knowledge gaps, this study aimed to characterize trends in the utilization of NMP versus SCS in DCD LT in the United States and to determine whether the use of NMP has expanded to higher-risk donor-recipient pairs over time.

Study Design and Cohort Creation
This retrospective cohort study used de-identified data from the United Network for Organ Sharing (UNOS) transplant database.The University of Pennsylvania Institutional Review Board approved the research.Adult recipient-donor LT pairs from DCD donors were identified between January 2016 and June 2022.Patients receiving multiorgan transplants were excluded.Data on recipient demographic (age, sex, and race), causes of liver disease, transplant model for end-stage liver disease (MELD), transplant year, and donor characteristics comprising the donor risk index (DRI) were collected.DRI components included donor age, cause of death (anoxia, cerebrovascular accident, and other), donor race, regional versus national share, and cold ischemic time.Given that cold ischemia time as a component of DRI does not accurately reflect the perfused organ state during NMP, we also calculated the DRI excluding cold ischemia time.From the donor file, the classification of NMP versus SCS was ascertained following prior methods based on coded utilization of machine perfusion. 3Notably, we did not include patients who received hypothermic machine perfusion given that very few cases had been performed in the UNOS data set within the study window among adult DCD recipients (n = 5).Finally, centerand region-level data included distinct center codes, UNOS regions, and distances between donor and recipient hospitals.

Statistical Analysis
Trends in utilization of NMP for DCD transplantation were plotted over time in terms of percent NMP utilization relative to all DCDs.For these data, the transplant year was divided into Q1/2 and Q3/4 for each year.Significant changes in secular trends were assessed using linear regression, where an era interaction variable was introduced on the basis of visual inspection.Beta coefficients and 95% confidence intervals were provided where relevant.Descriptive statistics for NMP versus SCS groups were presented as medians and interquartile ranges and frequencies and percentages for continuous and categorical variables, respectively.Statistical comparisons were made using the Wilcoxon rank-sum test and the chi-square or Fisher exact test, where indicated.Changes in DRI scores between NMP and SCS were assessed across transplant year eras (2016-2018, 2019-2020, and 2021-2022) using box plots, and medians were compared using Kruskal-Wallis tests.As a sensitivity analysis, this comparison was repeated using DRI, excluding cold ischemia time.A similar analytic approach was used to evaluate potential changes in distance between donor and recipient hospitals over transplant year eras, stratified by NMP versus SCS.Next, differences in individual DRI components for NMP DCDs across transplant year eras were evaluated using inferential statistics methods described previously.An alpha threshold of 5% for all tests was used to determine statistical significance.Finally, to visualize changes in donor-recipient "risk" over time, we plotted DRI against transplant MELD in each transplant era, separately for SCS and NMP.Differences were assessed qualitatively in these plots as an exploratory analysis.All analyses were performed using STATA version 17.0/BE (College Station, TX).

DISCUSSION
In the United States, DCD donation has expanded the donor pool for LT.Still, DCD organs remain generally less desirable than donation after brain death organs. 6NMP technology has improved outcomes associated with DCD donation and now represents an increasing proportion of DCD donation, 4,7 although to date, only about one-quarter of DCD-performing centers have used NMP technology.In this retrospective UNOS study of DCD LT from 2016 to 2022, we found that the utilization of NMP technology has increased over time, from 0.5% of all DCD transplants in 2016 to 11.2% in 2022; there was a particularly sharp increase in utilization moving from 2021 to 2022.In the early transplant era, NMP was used in younger donors and never with national sharing.However, over time, the median donor age with NMP increased by nearly 2 decades, with a significantly longer distance between donor and recipient hospitals, and a substantial proportion was shared across UNOS regions or nationally.This is reflected in increased "cold ischemic time" with NMP transplants in the most recent transplant era (2021-2022).These variables are all drivers of the significantly increased DRI over time that was observed with NMP utilization.Additionally, in the 2021-2022 transplant era, NMP DCD organs were transplanted for the first time into recipients with MELD scores >30.These findings are strong signals that transplant centers have become more familiar with NMP technology and that they are increasingly willing to use NMP with higher-risk donors and sicker recipients.Taken together with prior work demonstrating a potential reduction in posttransplant graft failure with NMP DCD transplantation, 3 the adoption of machine perfusion technology is likely to continue to expand.Based on the observed trends, it will be important for future research to evaluate potential differences in posttransplant outcomes among sicker recipients who have received NMP DCD transplantation.
This study has several limitations that we acknowledge in this study.First, an important limitation of the UNOS data set is its definition of cold ischemic time, which is associated with poor graft outcomes. 8,9However, cold ischemic time traditionally refers to the time an organ is outside of the donor body before transplantation and, therefore, includes NMP time, which causes less physiological stress on the graft than the equivalent period for SCS livers.As such, increased reported cold ischemic time could be driving observed increases in DRI associated with NMP without accurately reflecting true donor organ risk (although, as noted previously, median donor age also increased).Nonetheless, increased "cold ischemic time" with higher proportions of regional and national sharing over time suggests that NMP enables the viability of transporting DCD organs over longer distances.As utilization of NMP continues to grow over time, it is imperative that national databases, including UNOS, include variables that capture elements specific to NMP protocols.Second, there is also possible misclassification of exposures with regard to NMP utilization; however, there are specified fields in the UNOS database to classify the use of machine perfusion technology with DCD transplantation, and thus, the impact of this potential bias is likely minimal.Third, we did not analyze hypothermic machine perfusion, given that very few cases had been documented in the UNOS data set in the study window.There are ongoing active clinical trials that will make this an important area of future study.Finally, due to the limited granularity of data in the UNOS data set, we were not able to reliably differentiate normothermic regional perfusion (in situ) from ex situ NMP, and thus, both were classified as "NMP."As noted previously, this is another area where additional variables are needed in the UNOS registry, given that the process variables are fundamentally distinct between different machine perfusion techniques.For example, normothermic regional perfusion involves rapid in situ machine perfusion followed by cross-clamp and cold ischemia time, whereas ex situ NMP entails brief periods of cold ischemia time on either end of an extended machine perfusion time.
In conclusion, our study reports an increased utilization of NMP in recent years, and NMP has been used in donors with higher DRI and recipients with higher MELD at transplant over time.This suggests increasing familiarity and growing risk tolerance with NMP technology.However, because NMP remains a relatively new technique, ongoing review of patient outcomes, organ allocation practices, and utilization patterns are critical areas for future study.

FIGURE 1 .
FIGURE 1. Trends in utilization of normothermic machine perfusion in DCD liver transplantation.DCD, donation after circulatory death.

FIGURE 2 .
FIGURE 2. Donor risk index by era and MELD at transplant for static cold storage and normothermic machine perfusion.Box plots of DRI distribution (A and B).Over time, scatterplots of donor risk index and MELD at transplant were stratified by preservation method (C and D).Vertical dashed lines in panels C and D correspond to median values for the transplant era.MELD, model for end-stage liver disease.

TABLE 1 .
Cohort characteristics for DCD liver transplantation, stratified by preservation method, and differences in donor risk index components DCD, donation after circulatory death; IQR, interquartile range; MELD, model for end-stage liver disease; NASH, nonalcoholic steatohepatitis.