Revisiting Futility: Normothermic Perfusion Improves Outcomes Across all UK DCD Risk Strata

ABSTRACT

Donation after circulatory death (DCD) liver transplantation is increasingly supported by routine use of normothermic machine perfusion (NMP), yet established risk scores, including the UK DCD score, predate widespread NMP adoption. We conducted a retrospective single‐center cohort study of adult DCD liver transplants at Mayo Clinic Arizona from January 2019 through December 2024 (n = 586), comparing outcomes of NMP versus static cold storage (SCS) across UK DCD low‐, high‐, and futile‐risk strata. Primary outcomes were early allograft dysfunction (EAD), ischemic cholangiopathy (IC), and graft and patient survival censored at two years. Of 586 transplants, 375 were preserved with NMP and 211 with SCS. Despite older donor age and longer warm ischemia times in the NMP cohort, NMP was associated with lower rates of EAD and IC across all strata. Two‐year graft survival was higher with NMP than SCS (91% vs. 80%; p<0.001), and two‐year patient survival was 96% versus 89% (p = 0.002). The graft and patient survival advantage was most pronounced in the high‐risk (graft survival p = 0.004; patient survival p = 0.02) and futile strata (graft survival p = 0.01; patient survival p = 0.01), whereas graft survival was similar between preservation methods in the low‐risk group. These findings support the use of the UK DCD score as a triage tool to guide NMP allocation, with the greatest benefit observed in high‐risk and futile grafts for whom NMP should be considered essential rather than optional.

Abbreviations

BAS

Biliary anastomotic stricture

BMI

Body mass index

CIT

Cold ischemia time

DCD

Donation after circulatory death

dWIT

Donor warm ischemia time

EAD

Early allograft dysfunction

ERCP

Endoscopic retrograde cholangiopancreatography

fWIT

Functional warm ischemia time

HAT

Hepatic artery thrombosis

IC

Ischemic cholangiopathy

MELD

Model for End‐Stage Liver Disease

MRCP

Magnetic resonance cholangiopancreatography

NMP

Normothermic machine perfusion

OCS

Organ Care System

PNF

Primary nonfunction

SCS

Static cold storage

WIT

Warm ischemia time

1. Introduction

Liver transplant remains the definitive treatment for patients with end‐stage liver disease. Although transplant activity has steadily increased over the past decade, organ demand continues to exceed supply, with persistent growth of the waitlist [1]. Donation after circulatory death (DCD) liver transplantation has therefore become an increasingly important strategy to expand the donor pool. In contemporary practice, the adoption of normothermic machine perfusion (NMP) has transformed DCD utilization by mitigating ischemia–reperfusion injury and improving post‐transplant outcomes.

Historically, DCD grafts preserved with static cold storage (SCS) were associated with higher rates of ischemic cholangiopathy (IC), early allograft dysfunction (EAD), primary nonfunction, and inferior graft survival, limiting their widespread use [2, 3, 4]. While careful donor–recipient selection demonstrated that acceptable outcomes were achievable, DCD utilization remained constrained [5, 6, 7, 8]. To guide clinical decision‐making, the UK DCD risk score was introduced, stratifying grafts into low‐risk, high‐risk, and futile categories based on donor and recipient characteristics [9]. Despite this framework, DCD liver nonuse rates remain substantially higher than those for donation after brain death (DBD) grafts, and NMP has emerged as a potential strategy to safely expand utilization of grafts that would otherwise be declined [1].

Importantly, the UK DCD risk score was developed before routine implementation of machine perfusion. Components weighted toward ischemic injury, including prolonged warm ischemia and cold preservation time, may no longer carry the same prognostic weight in the NMP era, when near‐physiologic perfusion attenuates ischemia‐reperfusion injury and enables functional viability assessment prior to implantation. NMP has been shown to reduce biliary complications and improve graft function compared with SCS, particularly in DCD transplantation [10, 11].

As NMP becomes increasingly embedded in routine DCD practice, the relevance and application of established risk stratification tools warrant re‐evaluation. In this study, we compared outcomes of DCD liver transplantation using NMP versus SCS across all UK DCD risk strata in a large single‐center cohort. We hypothesized that NMP would mitigate ischemic complications and improve graft and patient survival across risk strata, and that the UK DCD score could serve as a triage tool to identify which grafts are most likely to benefit from NMP.

2. Methods

2.1. Study Design

We conducted a single‐center, retrospective cohort study of all DCD LT performed at Mayo Clinic Arizona between January 2019 and December 2024. Transplants were stratified by the UK DCD risk score into low‐risk, high‐risk, and futile categories [9]. Within each risk group, outcomes were compared between grafts preserved by SCS and those preserved with NMP. NMP was performed using the TransMedics Organ Care System (OCS; TransMedics Inc., Andover, MA, USA), a portable, normothermic perfusion platform that maintains donor livers in a near‐physiologic, oxygenated, blood‐based perfusion state during ex situ preservation [12]. The protocol was reviewed by the Mayo Clinic Arizona Institutional Review Board and determined to be exempt from ongoing oversight under 45 CFR 46.104(d)(4) (IRB 22‐005178). A waiver of informed consent was granted because the research involved retrospective analysis of existing clinical data recorded in a manner that precluded direct subject identification and posed no more than minimal risk.

2.2. UK DCD Risk Score

The UK DCD risk score, developed by Schlegel et al., stratifies donor–recipient combinations into low‐, high‐, and futile‐risk categories based on donor, recipient, and procedural factors. The score assigns points as follows: donor age (≤60 years = 0; >60 = 2), donor BMI (≤25 = 0; >25 = 3), functional donor WIT (≤20 min = 0; >20–30 min = 3; >30 min = 6), CIT (≤6 h = 0; >6 h = 2), recipient age (≤60 years = 0; >60 = 3), recipient MELD (≤25 = 0; >25 = 2), and retransplantation (no = 0; yes = 9). Scores ≤5 indicate low risk (>95% predicted 1‐year graft survival), 6–10 indicate high‐risk (>85% predicted 1‐year survival), and >10 indicate futility (<40% predicted 1‐year survival) [9].

Donor warm ischemia time (dWIT) was defined as the interval from withdrawal of life support to initiation of cold aortic perfusion and was used for UK DCD Risk Score calculation in the primary analysis. The UK DCD Risk Score was originally derived using functional donor warm ischemic time (fWIT), defined as the interval from the donor's systolic blood pressure (sBP) falling below 50 mmHg to the start of cold aortic perfusion. As sBP <50 mmHg was not systematically captured in our prospectively collected dataset, dWIT was used as a pragmatic surrogate. Our center did prospectively record an alternative fWIT using a composite physiological threshold (time from oxygen saturation <80% or mean arterial pressure <60 mmHg to cold perfusion) which captures the same hemodynamic deterioration and represents an established and widely used definition of functional warm ischemia in North American DCD programs. To confirm robustness of the primary findings, a sensitivity analysis was performed in which UK DCD Risk Scores were recalculated using this institutionally available fWIT definition (Table S1). CIT was defined as the interval from initiation of in situ cold perfusion to reperfusion in the recipient. However, given the physiological distinction between SCS and NMP, this definition risks overestimating ischemic exposure in NMP‐preserved grafts, as prolonged preservation on NMP does not represent cold ischemia. To address this, we calculated an adjusted CIT for all NMP grafts, defined as: Adjusted CIT = total preservation time − total NMP perfusion time. Since adjusted CIT was <6 h in all NMP cases, zero points were assigned to the CIT component. Donor macrosteatosis was categorized according to SRTR convention as <10%, mild (<30%), moderate (30%–60%), and severe (>60%) macrosteatosis.

Between 2019 and March 2022, the maximum accepted dWIT was 30 min, except for select low‐risk donor—recipient combinations. This threshold was increased to 40 min in March 2022 and to 45 min in August 2022 for grafts placed on NMP. From January through September 2022, NMP was applied exclusively to high‐risk or futile grafts. Beginning in October 2022, NMP became the standard preservation method for all DCD grafts. During this transition, donor selection criteria were expanded from age <65 years and dWIT <30 min to include donors up to 75 years of age, with dWIT <60 min and fWIT <45 min.

2.3. Donor Definition and Recovery Methods

All donors were Maastricht category III and procured using the super‐rapid recovery technique. Between January and September 2022, NMP was applied selectively to high‐risk or marginal DCD liver allografts, as defined by the UK DCD risk score, which incorporates donor, recipient, and procedural factors predictive of post‐transplant complications. Beginning in October 2022, NMP became the standard preservation method for all DCD livers at our institution. Beginning in June 2024, normothermic regional perfusion (NRP) was routinely implemented by the local organ procurement organization. Cases involving normothermic regional perfusion (NRP) or simultaneous multiple organ transplantation were excluded.

2.4. Normothermic Machine Perfusion and Viability Assessment

All NMP was performed using the OCS with a packed red blood cell—based perfusate, initiated at the donor hospital and maintained at normothermia during transport until reperfusion. Rather than applying a fixed preservation time limit, graft utilization was guided by perfusion performance; in our experience, livers have been safely preserved for up to 45 h on NMP. Grafts were assessed serially using a standardized institutional protocol informed by published viability frameworks, and were considered suitable for transplantation if they demonstrated progressive lactate clearance with a terminal perfusate lactate below 2.5 mmol/L, stable hepatic arterial and portal venous flows within acceptable physiologic ranges without escalating vasoactive support, and absence of concerning macroscopic or histologic features such as extensive necrosis or severe macrosteatosis on procurement or back‐table assessment. Sixteen livers were deemed unsuitable for transplantation on the basis of persistently abnormal lactate trajectories, histopathologic findings demonstrating significant fibrosis or necrosis, or technical and pump‐related complications, and are not included in this analysis.

2.5. Recipient Selection and Outcome Measures

Recipient selection was at the discretion of nine LT surgeons and followed the UNOS match‐run sequence. During the NMP era, primary consideration was made for size matching and logistical feasibility (recipient geography). We did not deliberately exclude candidates on the basis of retransplant status, high MELD score, or complex operative anatomy. Collected data included demographics, underlying liver disease, allocation Model for End‐Stage Liver Disease (MELD) score and retransplant status. Primary outcomes were EAD, defined using Olthoff criteria [13]; IC, diagnosed by a multidisciplinary team (radiology, advanced endoscopy, transplant surgery, hepatology) using standardized criteria after exclusion of HAT, clinically significant hepatic artery stenosis requiring intervention, and other vascular abnormalities, with routine MRCP performed in all patients at 4 months and 1 year regardless of laboratory or clinical findings; PNF, defined as graft failure within 7 days requiring urgent retransplantation without surgical complications; and graft and patient survival.

2.6. Statistical Analysis

Categorical variables were summarized as frequencies and compared using χ² or Fisher's exact tests, as appropriate. Continuous variables were expressed as medians with interquartile ranges and compared using the Kruskal–Wallis test or analysis of variance (ANOVA). One‐year cumulative incidence of first readmission, graft survival, and patient survival were analyzed using the Kaplan–Meier method and compared with the log‐rank test. A two‐sided p‐value ≤0.05 was considered statistically significant. Analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC) and R version 4.2.2 (R Foundation for Statistical Computing, Vienna, Austria).

3. Results

Between January 2019 and December 2024, 734 liver transplants were performed. After exclusion of simultaneous multi‐organ transplants and cases involving NRP, 586 were included in the analysis. Stratification by UK DCD risk score yielded 126 low‐risk cases (70 NMP, 56 SCS), 347 high‐risk cases (224 NMP, 123 SCS), and 113 futile cases (81 NMP, 32 SCS).

3.1. Donor and Recipient Characteristics

Donors in the NMP cohorts were significantly older than those in the SCS cohorts in both the high‐risk (median 54.0 vs. 53.0 years, p = 0.01) and futile (62.0 vs. 51.5 years, p < 0.001) groups. Donor BMI was higher in the NMP group within the high‐risk stratum (31.1 vs. 28.1, p < 0.001), while donor warm ischemia times were longer in NMP cohorts across all three risk categories: low‐risk (19.5 vs. 18.0 min, p = 0.04), high‐risk (24.0 vs. 22.0 min, p < 0.001), and futile (31.0 vs. 27.0 min, p = 0.01). As expected, total preservation times were longer for NMP grafts in each risk group due inclusion of perfusion time. Macrosteatosis distribution did not differ significantly between NMP and SCS in the low‐risk (p = 0.86) or futile (p = 0.41) groups. In the high‐risk group, SCS grafts had significantly lower macrosteatosis than NMP grafts (p<0.001): the majority of SCS grafts had less than 10% macrosteatosis (91.1% vs. 74.1% for NMP), while NMP grafts had higher rates of mild (19.2% vs. 6.5%) and moderate steatosis (6.7% vs. 2.4%), with no severe steatosis in either group. Severe macrosteatosis was rare overall, occurring in only one NMP graft in the low‐risk group (1.4%) and none elsewhere. Recipient demographics, BMI, MELD scores, and retransplant rates were comparable between NMP and SCS cohorts across all strata, with the exception of recipient BMI in the futile group, which was higher in NMP recipients [28.8 vs. 27.3, p = 0.04] (Table 1).

TABLE 1.

Donor and recipient characteristics for DCD liver transplants preserved with static cold storage (SCS) and normothermic machine perfusion (NMP), stratified by UK DCD risk score group (low‐risk, high‐risk, and futile).

	Low‐risk				High‐risk				Futile
Characteristics	NMP (n = 70)	SCS (n = 56)	Total (n = 126)	p‐value	NMP (n = 224)	SCS (n = 123)	Total (n = 347)	p‐value	NMP (n = 81)	SCS (n = 32)	Total (n = 113)	p‐value
Donor age in years, median [IQR]	54.0 [39.2–60.8]	51.5 [38.0–55.0]	52.0 [39.0–57.0]	0.14	54.0 [47.0–59.0]	53.0 [42.0–58.0]	53.0 [45.0–59.0]	0.01	62.0 [56.0–‐64.0]	51.5 [43.8–57.2]	60.0 [51.0–63.0]	<0.001
Donor BMI, median [IQR]	24.1 [22.3–28.0]	25.5 [22.6–30.2]	24.3 [22.3–29.8]	0.23	31.1 [26.3–37.3]	28.1 [25.3–31.4]	29.5 [25.8–35.0]	<0.001	30.3 [27.8–33.6]	30.2 [27.3–34.1]	30.2 [27.5–33.8]	0.25
Donor WIT in minutes, median [IQR]	19.5 [17.0–23.0]	18.0 [16.0–20.0]	19.0 [17.0–21.0]	0.04	24.0 [21.0–27.0]	22.0 [19.0–26.0]	23.0 [20.0–27.0]	<0.001	31.0 [25.0–36.0]	27.0 [25.0–29.2]	29.0 [25.0–35.0]	0.01
Total preservation time in hours, median [IQR]	15.1 [11.6–19.2]	6.0 [5.2–6.8]	9.3 [6.1–15.9]	<0.001	16.8 [13.7–20.8]	5.8 [5.0–7.0]	13.3 [6.7–19.0]	<0.001	17.6 [14.4–20.9]	7.0 [6.2–7.6]	15.1 [8.3–19.0]	<0.001
Macrosteatosis				0.86				<0.001				0.41
<10%, No [%]	55 [78.6]	47 [83.9]	102 [81.0]		166 [74.1]	112 [91.1]	278 [80.1]		60 [74.1]	27 [84.4]	87 [77.0]
Mild, No [%]	11 [15.7]	8 [14.3]	19 [15.1]		43 [19.2]	8 [6.5]	51 [14.7]		17 [21.0]	5 [15.6]	22 [19.5]
Moderate, No [%]	3 [4.3]	1 [1.8]	4 [3.2]		15 [6.7]	3 [2.4]	18 [5.2]		4 [4.9]	0 [0.0]	4 [3.5]
Severe, No [%]	1 [1.4]	0 [0.0]	1 [0.8]		0 [0.0]	0 [0.0]	0 [0.0]		0 [0.0]	0 [0.0]	0 [0.0]
Recipient age in years, median [IQR]	54.5 [45.2–58.8]	55.0 [48.0–58.0]	55.0 [47.0–58.0]	0.70	62.0 [54.0–67.0]	63.0 [57.0–67.0]	63.0 [55.0–67.0]	0.47	64.0 [61.0–68.0]	64.5 [61.0–69.0]	64.0 [61.0–69.0]	0.89
Recipient sex				0.52				0.46				0.96
Female, No [%]	34 [48.6]	24 [42.9]	58 [46.0]		76 [33.9]	37 [30.1]	113 [32.6]		30 [37.0]	12 [37.5]	42 [37.2]
Male, No [%]	36 [51.4]	32 [57.1]	68 [54.0]		148 [66.1]	86 [69.9]	234 [67.4]		51 [63.0]	20 [62.5]	71 [62.8]
Recipient BMI, median [IQR]	27.7 [23.9–31.8]	26.2 [23.2–30.6]	27.2 [23.4–31.1]	0.47	27.8 [25.0–32.6]	27.9 [25.0–31.8]	27.9 [25.0–31.7]	0.91	28.8 [25.0–34.4]	27.3 [24.5–28.4]	27.7 [24.5–32.9]	0.04
MELD, median [IQR]	17.0 [11.0–21.8]	18.0 [15.0–21.2]	17.5 [12.2–21.8]	0.26	17.0 [12.0–21.0]	17.0 [11.5–20.0]	17.0 [12.0–21.0]	0.50	16.0 [12.0–23.0]	15.0 [10.0–23.2]	16.0 [11.0–23.0]	0.68
Re‐transplant, No [%]	0 [0]	0 [0]	0 [0]	NA	2 [0.9]	0 [0]	2 [0.6]	0.54	7 [8.6]	2 [6.2]	9 [8.0]	1.00

Abbreviations: BMI, body mass index; IQR, interquartile range; MELD, Model for End‐Stage Liver Disease; NMP, normothermic machine perfusion; SCS, static cold storage.

3.2. Primary Outcomes

Rates of EAD were significantly lower with NMP than SCS across all risk strata: low‐risk (21.4% vs. 44.6%; p = 0.01), high‐risk (22.8% vs. 55.3%; p<0.001), and futile (21.0% vs. 59.4%; p<0.001). PNF was rare in all groups and did not differ significantly between NMP and SCS: no events occurred in the low‐risk group, rates were 0.9% versus 0.8% in the high‐risk group (p = 1.00), and 1.2% versus 0% in the futile group (p = 1.00). Early HAT was similarly uncommon across strata, occurring in 1.4% versus 0% in the low‐risk group (p = 1.00), 0.4% versus 0.8% in the high‐risk group (p = 1.00), and 0% versus 3.1% in the futile group (p = 0.28). Late HAT was rare and did not differ significantly between groups in any stratum (low‐risk: 5.7% vs. 0%, p = 0.13; high‐risk: 1.8% vs. 0%, p = 0.30; futile: 3.7% vs. 0%, p = 0.56) (Figure 1).

FIGURE 1.

Incidence of early allograft dysfunction (EAD), primary non‐function (PNF), early hepatic artery thrombosis (HAT), and late HAT for DCD liver transplants preserved with normothermic machine perfusion (NMP) and static cold storage (SCS), stratified by UK DCD risk score group (low‐risk, high‐risk, and futile). NMP was associated with significantly lower EAD rates across all three risk strata, while PNF and HAT were rare and did not differ significantly between preservation groups.

3.3. Biliary Complications

BAS rates were similar between NMP and SCS across all risk strata (Low Risk: 34.3% vs. 35.7%, p = 0.87; High Risk: 42.4% vs. 43.1%, p = 0.90; Futile: 32.1% vs. 34.4%, p = 0.82). Time to diagnosis, rates of MRCP and ERCP performed, and type of intervention required did not differ between groups in any stratum, with combined dilation and stenting representing the predominant treatment in both NMP and SCS recipients (Table 2).

TABLE 2.

Biliary anastomotic stricture and ischemic cholangiopathy: Incidence, timing and management for livers preserved with static cold storage (SCS) and normothermic machine perfusion (NMP), stratified by UK DCD risk score group (low‐risk, high‐risk, and futile).

	Low‐risk				High‐risk				Futile
	NMP (n = 70)	SCS (n = 56)	Total (n = 126)	p‐value	NMP (n = 224)	SCS (n = 123)	Total (n = 347)	p‐value	NMP (n = 81)	SCS (n = 32)	Total (n = 113)	p‐value
Biliary Anastomotic Stricture
BAS, No [%]	24 [34.3]	20 [35.7]	44 [34.9]	0.87	95 [42.4]	53 [43.1]	148 [42.7]	0.90	26 [32.1]	11 [34.4]	37 [32.7]	0.82
Time to BAS diagnosis, days, median [IQR]	56.0 [14.2–132.0]	45.5 [23.0–110.0]	52.5 [22.0–128.8]	0.85	56 [21.5–129.0]	38.0 [19.0–102.0]	48.0 [19.8–126.0]	0.22	61.5 [28.0–115.5]	70.0 [42.5–146.5]	66.0 [34.0–120.0]	0.29
MRCP performed BAS, No. [%]	21 [87.5]	18 [90.0]	39 [88.6]	1.00	90 [94.7]	46 [86.8]	136 [91.9]	0.12	26 [100.0]	11 [100.0]	37 [100.0]	NA
ERCP performed for BAS, No [%]	24 [100.0]	20 [100.0]	44 [100.0]	NA	92 [96.8]	53 [100.0]	145 [98.0]	0.55	25 [96.2]	11 [100.0]	36 [97.3]	1.00
Intervention required for BAS, No [%]				0.20				0.75				1.00
Stent only	0 [0.0]	0 [0.0]	0 [0.0]		2 [2.1]	1 [1.9]	3 [2.0]		0 [0.0]	0 [0.0]	0 [0.0]
Dilatation only	0 [0.0]	2 [10.0]	2 [4.5]		3 [3.2]	3 [5.7]	6 [4.1]		1 [3.8]	0 [0.0]	1 [2.7]
Dilation + stent	24 [100.0]	18 [90.0]	42 [95.5]		87 [91.6]	46 [86.8]	133 [89.9]		24 [92.3]	11 [100.0]	35 [94.6]
None	0 [0.0]	0 [0.0]	0 [0.0]		3 [3.2]	3 [5.7]	6 [4.1]		1 [3.8]	0 [0.0]	1 [2.7]
Ischemic Cholangiopathy
IC present, No [%]	5 [7.1]	13 [23.2]	18 [14.3]	0.02	26 [11.6]	32 [26.0]	58 [16.7]	<0.001	4 [4.9]	12 [37.5]	16 [14.2]	<0.001
Time to IC diagnosis, days, median [IQR]	122.0 [84.0–125.0]	106.0 [49.0–131.0]	109.5 [51.5–130.5]	0.66	138.5 [57.0–188.0]	96.5 [57.8–177.0]	117.5 [57.2–177.0]	0.28	63.0 [52.5–87.2]	118.5 [58.8–160.8]	105.0 [56.5–157.0]	0.25
MRCP performed for IC, No [%]	4 [80.0]	10 [76.9]	14 [77.8]	1.00	23 [88.5]	28 [87.5]	51 [87.9]	1.00	4 [100.0]	12 [100.0]	16 [100.0]	NA
Intervention required for IC, No [%]				1.00				0.26				NA
Stent only	0 [0.0]	0 [0.0]	0 [0.0]		1 [3.8]	0 [0.0]	1 [1.7]		0 [0.0]	0 [0.0]	0 [0.0]
Dilation + stent	5 [100.0]	12 [92.3]	17 [94.4]		24 [92.3]	28 [87.5]	52 [89.7]		4 [100.0]	12 [100.0]	16 [100.0]
None	0 [0.0]	1 [7.7]	1 [5.6]		1 [3.8]	4 [12.5]	5 [8.6]		0 [0.0]	0 [0.0]	0 [0.0]
Re‐transplantation due to IC, No. [%]	0 [0.0]	0 [0.0]	0 [0.0]	NA	4 [15.4]	7 [21.9]	11 [19.0]	0.74	0 [0.0]	4 [33.3]	4 [25.0]	0.52

Abbreviations: BAS: biliary anastomotic stricture; ERCP: endoscopic retrograde cholangiopancreatography; IC: ischemic cholangiopathy; IQR: interquartile range; MRCP: magnetic resonance cholangiopancreatography; NMP: normothermic machine perfusion; SCS: static cold storage.

IC was lower in the NMP cohorts: low‐risk (7.1% vs. 23.2%, p = 0.02), high‐risk (11.6% vs. 26.0%, p < 0.001), and futile (4.9% vs. 37.5%, p < 0.001). The number of days to diagnosis, the rate of MRCP performed, or the number of interventions required did not differ between NMP and SCS or across risk groups. Re‐transplantation due to IC was rare overall and did not differ significantly between NMP and SCS across risk groups. In the futile stratum, however, no NMP recipients required re‐transplantation for IC compared with 33.3% of SCS recipients, a clinically notable difference that did not reach statistical significance (p = 0.52), likely reflecting the limited sample size in this subgroup (Table 2).

3.4. Causes of Graft Loss

Among grafts that were lost, the distribution of causes did not differ significantly between NMP and SCS in any risk stratum (low‐risk p = 0.30; high‐risk p = 0.54; futile p = 0.07). Death with a functioning graft was the most common cause of graft loss overall and in most subgroups, though hepatic artery complications predominated among the small number of NMP graft losses in the low‐risk and futile strata. Biliary complications as a cause of graft loss were numerically more frequent with SCS than NMP in the high‐risk (30.3% vs. 18.2%) and futile (21.4% vs. 0%) groups, consistent with the lower IC rates observed with NMP, though these differences did not reach statistical significance given the limited number of graft loss events in each stratum (Table 3).

TABLE 3.

Causes of graft loss for livers preserved with static cold storage (SCS) and normothermic machine perfusion (NMP), stratified by UK DCD risk score group (low‐risk, high‐risk, and futile).

	Low‐risk				n = High‐risk				Futile
Characteristics	NMP (n = 70)	SCS (n = 56)	Total (n = 126)	p‐value	NMP (224)	SCS (n = 123)	Total (n = 347)	p‐value	NMP (n = 81)	SCS (n = 32)	Total (n = 113)	p‐value
Graft loss, No [%]	5 [7.0]	9 [16.1]	14 [11.1]	0.15	22 [9.8]	33 [26.8]	55 [15.9]	<0.001	9 [11.1]	14 [43.8]	23 [20.4]	<0.001
Cause of graft loss				0.30				0.54				0.07
Acute/Chronic rejection, No [%]	0 [0.0]	1 [11.1]	1 [7.1]		1 [4.5]	1 [3.0]	2 [3.6]		1 [11.1]	1 [7.1]	2 [8.7]
Biliary complications, No [%]	0 [0.0]	1 [11.1]	1 [7.1]		4 [18.2]	10 [30.3]	14 [25.5]		0 [0.0]	3 [21.4]	3 [13.0]
Hepatic artery complications, No [%]	3 [60.0]	1 [11.1]	4 [28.6]		4 [18.2]	3 [9.1]	7 [12.7]		3 [33.3]	2 [14.3]	5 [21.7]
Primary non function, No [%]	0 [0.0]	0 [0.0]	0 [0.0]		1 [4.5]	1 [3.0]	2 [3.6]		1 [11.1]	0 [0.0]	1 [4.3]
Venous complications or disorders, No [%]	0 [0.0]	0 [0.0]	0 [0.0]		1 [4.5]	0 [0.0]	1 [1.8]		0 [0.0]	0 [0.0]	0 [0.0]
Others, No [%]	0 [0.0]	0 [0.0]	0 [0.0]		1 [4.5]	0 [0.0]	1 [1.8]		2 [22.2]	0 [0.0]	2 [8.7]
Death with functioning graft, No [%]	2 [40.0]	6 [66.7]	8 [57.1]		10 [45.5]	18 [54.5]	28 [50.9]		2 [22.2]	8 [57.1]	10 [43.5]

NMP: normothermic machine perfusion; SCS: static cold storage.

3.5. Perfusion Parameters

Perfusion parameters did not differ significantly across the low‐risk, high‐risk, and futile groups, including time to NMP initiation, hepatic artery and portal vein hemodynamics, starting and terminal lactate, and bile production (Table 4 ).

TABLE 4.

Perfusion parameters for livers preserved with normothermic machine perfusion (NMP), stratified by UK DCD risk score group (low‐risk, high‐risk, and futile).

Characteristics	Low Risk (n = 70)	High Risk (n = 224)	Futile (n = 81)	Total (n = 375)	p‐value
Cross clamp to start of NMP in hours, median [IQR]	2.4 [2.1–2.9]	2.6 [2.2–3.0]	2.6 [2.2–3.1]	2.6 [2.2–3.0]	0.30
Final HA flow in L/min, median [IQR]	0.6 [0.5–0.6]	0.6 [0.5–0.7]	0.6 [0.5–0.6]	0.6 [0.5–0.6]	0.46
Final HA pressure in mmHg, median [IQR]	47.0 [42.2–52.8]	46.0 [42.0–51.0]	46.0 [43.0–50.0]	46.0 [42.0–51.0]	0.61
Final PV flow in L/min, median [IQR]	1.4 [1.3–1.4]	1.4 [1.3–1.4]	1.4 [1.3–1.4]	1.4 [1.3–1.4]	0.71
Starting lactate in mmol/L, median [IQR]	7.3 [5.4–9.7]	7.9 [6.0–9.8]	7.6 [6.3–9.6]	7.7 [6.0–9.7]	0.33
End lactate in mmol/L, median [IQR]	0.4 [0.3–0.9]	0.6 [0.3–1.1]	0.7 [0.3–1.1]	0.6 [0.3–1.1]	0.11
Bile volume in mL, median [IQR]	50.0 [25.0–75.0]	50.0 [35.0–85.0]	50.0 [40.0–75.0]	50.0 [35.0–80.0]	0.62
Bile flow in mL per hour, median [IQR]	4.6 [3.7–6.2]	4.5 [3.4–6.2]	4.1 [2.9–5.2]	4.4 [3.3–6.1]	0.24

Abbreviations: HA: hepatic artery; IQR: interquartile range; NMP: normothermic machine perfusion; PV: portal vein.

3.6. Survival Outcomes

Graft survival at 2 years was higher with NMP than with SCS (91% vs. 80%; HR, 2.52 [95% CI, 1.59–4.01]; p<0.001). By UK DCD stratum, graft survival was similar in low risk (93% vs. 89%; HR, 1.52 [95% CI, 0.47–4.93]; p = 0.49) but superior in NMP high‐risk (91% vs. 79%; HR, 2.52 [95% CI, 1.41–4.51]; p = 0.004) and futile (91% vs. 66%; HR, 4.40 [95% CI, 1.72–11.26]; p = 0.01) groups. (Figure 2) Patient survival at 2 years was also higher with NMP compared to SCS (96% vs. 89%; HR, 2.72 [95% CI, 1.44–5.12]; p = 0.002), with a significant advantage in the high‐risk group (95% vs. 88%; HR, 2.56 [95% CI, 1.17–5.57]; p = 0.02) and futile groups (98% vs. 81%; HR, 7.55 [95% CI, 1.52–37.48]; p = 0.01), but not in low‐risk group (96% vs. 95%; HR 1.25 [95% CI, 0.25–6.18]; p = 0.79) (Figure 3).

FIGURE 2.

Kaplan‐Meier curves for two‐year graft survival in DCD liver transplants preserved with normothermic machine perfusion (NMP) and static cold storage (SCS), stratified by UK DCD risk score group: low‐risk (A), high‐risk (B), and futile (C). Two‐year graft survival was 93% versus 89% in the low‐risk group (p = 0.49), 91% versus 79% in the high‐risk group (p = 0.004), and 91% versus 66% in the futile group (p = 0.01).

FIGURE 3.

Kaplan‐Meier curves for two‐year patient survival in DCD liver transplants preserved with normothermic machine perfusion (NMP) and static cold storage (SCS), stratified by UK DCD risk score group: low‐risk (A), high‐risk (B), and futile (C). Two‐year patient survival was 96% versus 95% in the low‐risk group (p = 0.79), 95% versus 88% in the high‐risk group (p = 0.02), and 98% versus 81% in the futile group (p = 0.01).

4. Discussion

In this analysis of 586 DCD transplants from a single high‐volume center, NMP was associated with lower rates of EAD and IC across all UK DCD risk strata and with superior two‐year graft and patient survival in the high‐risk and futile groups. Graft survival was similar between NMP and SCS in the low‐risk stratum (93% vs. 89%), but significantly better with NMP in the high‐risk (91% vs. 79%, p = 0.004) and futile groups (91% vs. 66%, p = 0.01), with the same pattern observed for patient survival. Most notably, in the futile stratum, where the UK DCD Risk Score originally proposed that grafts should not be transplanted without graft treatment, NMP was associated with markedly better two‐year outcomes. These associations were observed despite older donor age, longer warm ischemia times, and higher donor BMI in the NMP cohort across multiple risk strata, underscoring that the benefit of NMP is not explained by more favorable donor selection.

The maximum clinical benefit in UK DCD high‐risk and futile categories is consistent with a risk‐modifying rather than uniform effect with NMP. When baseline ischemic burden is greater, attenuation of ischemia—reperfusion injury is more likely to translate into superior graft and patient survival. By contrast, among low‐risk grafts, survival was similar between groups, although NMP was associated with fewer biliary complications compared to its SCS counterpart. The pattern is biologically plausible and points to biliary protection as a key pathway. Normothermic perfusion restores microcirculatory flow, limits endothelial and mitochondrial injury, and sustains cholangiocyte metabolism via arterialized perfusion of the peribiliary plexus, which together reduces the ductal vulnerability that characterizes DCD and other extended‐criteria grafts [14, 15, 16, 17]. These mechanisms would be expected to confer the greatest risk mitigation where donor WIT, age, BMI, or macrosteatosis push grafts toward a cholangiopathy‐prone phenotype, while yielding a more attenuated clinical benefit in low‐risk settings [7]. The selective reduction in IC, even when survival differences are small, also raises the possibility that survival curves may diverge further beyond two years, with IC acting as a mediator of longer‐term graft loss. Longer follow‐up in a larger cohort will clarify whether fewer biliary complications ultimately translate into a patient survival advantage in the low‐risk group, and whether the survival benefit already observed in the high‐risk and futile strata is sustained over time.

Contemporary data are consistent with these findings. Yamamoto et al. analyzed adult deceased‐donor LTs preserved with a portable NMP platform and found substantially lower ischemic injury. The odds of EAD were markedly reduced (OR, 0.11; 95% CI, 0.03–0.46) and reperfusion syndrome was uncommon with NMP (OR, 0.04; 95% CI, 0.01–0.25). The effect was most pronounced in higher‐risk grafts, including DCD and other extended‐criteria donors [18]. In a recent U.S. registry analysis, Okumura et al. reported that NMP was associated with significantly lower rates of graft failure and mortality among DCD recipients, reinforcing its impact at a national level [19]. The VITTAL prospective viability‐testing study perfused livers declined by all UK centers and transplanted the majority after meeting prespecified functional criteria; 71% were used, with 100% 90‐day graft and patient survival and favorable 1‐year outcomes. Notably, many were extended‐criteria, including DCD grafts, underscoring that physiological assessment on NMP can rescue organs that traditional, pre‐perfusion risk frameworks would exclude. Importantly, five‐year follow‐up confirmed durable benefit [20]. These studies underscore that NMP not only reduces ischemia‐related injury but also enables physiological assessment that can safely expand utilization of grafts previously considered marginal or futile.

Our data extend these observations while also reaffirming the framework within which machine perfusion was originally envisioned. The UK DCD Risk Score was designed to identify which donor‐recipient combinations require machine perfusion for safe transplantation, and to ensure that futile‐group grafts are not used without graft treatment. Our findings confirm this principle at scale. When NMP is systematically applied to high‐risk and futile‐group grafts, outcomes are substantially improved relative to SCS. The attenuated negative effects observed in the futile group are unlikely to be explained by dWIT or fWIT alone, as markedly prolonged WIT (e.g., >45–60 min) remained uncommon; rather, this discordance more plausibly reflects the combined protective effect of NMP in mitigating ischemic injury alongside evolving donor and recipient selection practices. The score also applies an additive penalty for cold ischemia time despite the negligible cold preservation associated with NMP, a technical limitation worth acknowledging in future score iterations. Nevertheless, the clinical message is consistent: NMP modifies the risk profile across all strata, attenuating biliary complications even in low‐risk grafts and conferring a meaningful survival advantage in high‐risk and futile categories compared with SCS. The benefit is not uniform. It is greatest where predicted risk is highest, and it is in precisely these cases that NMP should be prioritized to yield the greatest absolute gain.

Translating these observations into practice will require risk prediction models that are built for the perfusion era. Contemporary models should explicitly incorporate whether a graft undergoes dynamic preservation, including ex situ NMP or, when applicable, NRP, and integrate quantitative viability parameters such as lactate clearance kinetics, bile volume and composition, hepatic arterial and portal venous hemodynamics, and procurement histology [21, 22, 23]. Such frameworks are most likely to improve risk discrimination within the high‐risk and futile strata, where clinical decision‐making is most consequential and where our data demonstrate that responsible expansion of donor utilization is achievable when NMP is applied with standardized protocols and viability thresholds. This shift extends beyond preservation strategy alone. With the integration of NRP, we believe total dWIT may potentially be extended up to 120 min in selected cases, provided fWIT remains within 45–60 min and perfusion‐based viability assessment remains favorable. These developments reflect a broader and necessary shift away from fixed ischemia time cutoffs toward comprehensive physiologic graft assessment, precisely the framework that next‐generation risk prediction models must seek to capture.

This study has several limitations. First, as a retrospective, single‐center analysis from a program with longstanding DCD experience and high utilization, generalizability is limited; practice patterns, procurement logistics, and team expertise at lower‐volume centers may differ. The observed median NMP preservation times of 15–18 h in this cohort reflect an institutional practice that depends on infrastructure, team expertise, and logistical capacity that may not be reproducible at all centers. Second, risk score calculation involved two methodological approximations: adjusted CIT (total preservation minus perfusion time) was used for NMP grafts, which consistently yielded values below 6 h and thus zero CIT points, potentially underestimating risk relative to the original score; and dWIT rather than fWIT was used as the primary warm ischemia measure, as sBP <50 mmHg was not systematically captured. Center‐specific variation in fWIT definitions is well recognized and may affect risk group assignment and between‐study comparability. Third, UK DCD classification was often driven by one or two heavily weighted variables, including retransplantation (9 points), functional WIT (up to 6 points), and in some instances donor BMI, rather than the full composite, a concern in the perfusion era when ischemia‐related risk may be attenuated. Finally, offer‐level discard data, costs, resource utilization, and surgeon‐level variation were not analyzed, limiting conclusions about national discard reduction and cost‐effectiveness. Prospective multicenter studies with standardized warm ischemia definitions, transparent handling of CIT during perfusion, systematic capture of perfusion biomarkers, and longer follow‐up are needed to validate these findings and refine risk prediction in the perfusion era.

5. Conclusion

NMP substantially reduces ischemic injury across all UK DCD risk strata and confers superior graft and patient survival in high‐risk and futile groups, where the benefit is greatest and most clinically consequential. These findings support a fundamental reorientation of DCD risk assessment: from fixed thresholds that exclude grafts to physiologic frameworks that guide treatment and expand utilization.

Disclosure

The authors have nothing to report.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting Information: ctr70669‐sup‐001‐Table S1.docx

Ohara S. Y., Stevens V., Madisetty A., et al. “Revisiting Futility: Normothermic Perfusion Improves Outcomes Across all UK DCD Risk Strata.” Clinical Transplantation 40, no. 9 (2026): e70669. 10.1111/ctr.70669

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.