EAD Outperformed L-GrAFT7 After NMP Liver Transplant

Key Takeaways

• In deceased donor liver transplantation performed after ex-situ normothermic machine perfusion, early allograft dysfunction occurred in 40% of cases and was associated with lower 6-month graft survival.
• In this small normothermic machine perfusion cohort, L-GrAFT7 and MEAF did not show clear discrimination for early graft loss.
• Higher donor BMI, longer agonal time in donation after circulatory death grafts, longer anastomosis time, post-reperfusion syndrome, impaired glucose metabolism, delayed lactate clearance, greater bicarbonate use, and higher perfusate transaminases were associated with early allograft dysfunction.
• Recipients with early allograft dysfunction had longer hospital stay and more biopsy-proven acute cellular rejection, while several other posttransplant complications did not differ significantly.

Ex-situ normothermic machine perfusion can shift much of ischemia-reperfusion injury onto the pump, leaving transplant teams to judge early postoperative graft dysfunction in a different physiologic setting than the one in which static cold storage-era scores were developed. That creates uncertainty about whether legacy posttransplant dysfunction scores still reflect early graft risk after implantation and how much weight to give pump-derived quality signals during organ use decisions. At a U.S. transplant center using back-to-base perfusion, investigators examined how those score frameworks aligned with outcomes after transplantation.

Investigators conducted a single-center retrospective cohort of deceased donor liver transplants performed after back-to-base ex-situ normothermic machine perfusion (NMP) between August 20, 2022 and May 31, 2024. Of 126 grafts that underwent NMP, 115 were transplanted after NMP, and 107 transplants were analyzed after excluding multi-organ transplants and donation after circulatory death (DCD) grafts that underwent normothermic regional perfusion (NRP) followed by NMP. The analyzed cohort included 44 donation after brain death (DBD) grafts and 63 DCD grafts, reflecting contemporary U.S. expansion of DCD liver use. Early allograft dysfunction (EAD) was assessed within the first 7 postoperative days, the model for early allograft function (MEAF) from postoperative days 1 to 3, and liver graft assessment following transplantation (L-GrAFT7) across postoperative days 1 to 7. Viability concerns on pump included poor flows or failure to clear lactate within 6 hours, bicarbonate was used to maintain pH above 7.2, and bile chemistries were not part of viability assessment during the study period.

In the Shah et al. graft dysfunction score comparison after ex-situ normothermic machine perfusion, EAD occurred in 43 of 107 recipients (40%) and was associated with lower 6-month graft survival than no EAD, at 89% versus 100% (p=0.036). Overall patient survival and overall graft survival were otherwise not significantly different by EAD status, and in this small cohort, high versus low L-GrAFT7 and MEAF groups did not show clear discrimination for early graft loss. EAD also clustered with higher donor BMI, at 31.12 versus 26.23 kg/m2 (p<0.001), and more frequent post-reperfusion syndrome, at 39.5% versus 12.5% (p=0.002), while longer agonal time in DCD grafts, longer anastomosis time, greater bicarbonate use, impaired glucose metabolism, delayed lactate clearance, and higher perfusate transaminases followed the same direction. Recipients with EAD had longer hospital stay and more biopsy-proven acute cellular rejection, whereas biliary complications, renal dysfunction, infections, readmission, and endoscopic retrograde cholangiopancreatography use did not differ significantly as a group. Higher pump transaminases also tracked with higher peak recipient transaminases within the first postoperative week, and investigators reported that 10 grafts with pump alanine aminotransferase above 6000 U/L were transplanted without 6-month graft loss in this cohort.

The analysis was retrospective, single-center, and modest in size, with few hard early events. Only nine graft failures occurred overall and just one within 90 days, limiting assessment of scores designed for early graft failure and precluding multivariable modeling. Data constraints also shaped interpretation: bicarbonate provision data were available for 89.7% of perfused grafts, perfusate transaminases were not measured in the first 20 grafts, and bile production and chemistries were not included in viability assessment. The authors also noted that lower Model for End-Stage Liver Disease-Sodium scores in the EAD group may reflect allocation of marginal grafts to lower-acuity recipients, so observations about using grafts above conventional perfusion thresholds may not generalize to higher-acuity recipients.

According to the authors, liver transplantation after NMP may involve a dysfunction pattern that is not fully captured by static cold storage-era score frameworks. In this small cohort, EAD retained short-term prognostic signal whereas L-GrAFT7 and MEAF did not show clear discrimination for early graft loss, and the combination of impaired glucose metabolism, delayed lactate clearance, increased bicarbonate demand, and elevated perfusate transaminases may represent a distinct pump-era phenotype. They called for prospective multicenter validation before broader adoption of any dynamic perfusion-based framework.

Clinician Questions

How was early allograft dysfunction defined after liver transplantation following ex-situ NMP?

In liver transplantation after ex-situ NMP, EAD was defined as bilirubin 10 mg/dL or higher on postoperative day 7, international normalized ratio 1.6 or higher on postoperative day 7, or aspartate aminotransferase or alanine aminotransferase above 2000 U/L within the first 7 postoperative days.

Which liver transplant recipients were included and excluded from the ex-situ NMP cohort?

This single-center cohort included deceased donor DBD and DCD liver grafts that underwent back-to-base ex-situ NMP and excluded pediatric transplants, living donor transplants, multi-organ transplants, static cold storage transplants, and DCD grafts that underwent NRP or NRP plus NMP.

What perfusion features were used in viability assessment during this ex-situ NMP program?

Viability concerns in this ex-situ NMP program included poor arterial or portal flows and failure to clear lactate to below 2 mmol/L within 6 hours of perfusion, bicarbonate was administered to maintain pH above 7.2, and bile production and bile chemistries were not part of viability assessment during the study period.