Hypothermic Oxygenated Perfusion Versus Static Cold Storage in Transplantation of Extended Criteria Liver Grafts: A Systematic Review and Meta‐Analysis

ABSTRACT

Background

The use of extended criteria donor livers (ECD) is becoming more routine in many transplant centers. These organs have higher risks for complications; however, hypothermic‐oxygenated perfusion (HOPE) was found to improve outcomes, including graft survival. We aim to assess the effect of HOPE on different types of ECD liver grafts.

Methods

A systematic search was conducted of PubMed, EMBASE, and Cochrane databases to identify studies that compared HOPE versus static cold storage (SCS) for ECD. Subgroup analysis on ECD from brain death (DBD‐ECD) and circulatory death (DCD) donors, and of randomized controlled trials (RCT) were conducted. Primary endpoints were primary non‐function (PNF), early allograft dysfunction (EAD), length of ICU/hospital stay, vascular/biliary complications, retransplantation, and graft survival.

Results

Twelve studies were identified comprising 1833 transplant patients (29% receiving HOPE and 71% SCS). Pooled analysis showed a significant reduction of EAD, 1‐year graft failure rate, retransplantation rate, non‐anastomotic biliary strictures and Clavien‐Dindo Grade ≥ 3 complications favoring HOPE. Subgroup analysis on DBD‐ECD grafts yielded lower EAD and shorter length of the hospital stay with HOPE. Further subgroup analysis on DCD grafts demonstrated lower EAD rates, superior 1‐year graft survival rates, and reduced NAS in the HOPE group. Finally, analysis including RCTs revealed decreased EAD and retransplantation rates in the HOPE group.

Conclusions

Reported outcomes after ECD liver transplantation were significantly improved with HOPE compared to SCS alone. This effect was even more pronounced in DCD grafts.

Abbreviations

AS

anastomotic stricture

CCI

comprehensive complication index

DBD

donor after brain death

DCD

donor after cardiac death

EAD

early allograft dysfunction

ECD

extended criteria donor

HAT

hepatic artery thrombosis

HOPE

hypothermic‐oxygenated perfusion

LT

liver transplantation

NAS

non‐anastomotic stricture

PNF

primary non‐function

PVT

portal vein thrombosis

1. Introduction

To increase the pool of safely usable donors in transplantation, livers from extended criteria donors (ECD) are increasingly used worldwide. ECD livers by definition fall outside benchmark or standard criteria and include grafts from donors after brain death (DBD) with, for example, advanced donor age, prolonged cold ischemia time (>12 h), and/or greater degrees of macrosteatosis [1]. Donation after circulatory death (DCD) donors are always counted as ECD organs [2, 3]. ECD livers are more susceptible to ischemia reperfusion injury (IRI) [4], resulting in higher complication rates, including primary non‐function (PNF), early allograft dysfunction (EAD), higher rates of overall and biliary complications and graft loss [2, 3].

Static cold storage (SCS) has been historical standard method of organ preservation worldwide. However, data so far has associated SCS with greater complications when used for ECD organs compared to machine perfusion techniques, which are now used frequently in both the US and Europe [5, 6, 7]. Two main dynamic preservation techniques are currently used, based primarily on perfusion temperature. First, normothermic perfusion approaches, which do include upfront normothermic regional perfusion (NRP)—a technique used for DCD donors and applied during procurement [8]. The second normothermic technique includes ex situ normothermic machine perfusion (NMP), applied either upfront after short SCS or endischemic in the recipient center. Conversely, the other endischemic technique is hypothermic‐oxygenated perfusion (HOPE) [6, 7, 8], where highly oxygenated artificial fluids are recirculated through the liver. HOPE has also been generally shown to improve outcomes in many large cohort studies and randomized controlled trials (RCT) [4, 9, 10, 11]. Such results are based on a protective effect on mitochondria when oxygen is re‐introduced at colder temperatures compared with the traditional reperfusion under warm conditions either at transplant or, ex situ on a device [12].

Although several meta‐analyses have been done with the HOPE and other perfusion techniques [13, 14, 15, 16], robust data supporting the decision whether to perfuse all livers, including benchmark quality organs or only a selection of different types of ECD grafts, including DCD, remains scarce. This study uniquely focuses, therefore, on the impact of HOPE on outcomes after transplantation of different ECD organs. Moreover, previous studies have reported a variety of outcomes, often underreporting clinically relevant key metrics [17]. Recognition of this issue led to the recent definition of a Core Outcomes Set (COS), a set of outcome parameters most relevant to liver transplantation, along with a defined method of assessing and reporting such outcomes [17]. This study analyses the impact of HOPE on such core outcomes after transplantation of different types of ECD liver grafts.

2. Methodology

2.1. Search Strategy

A systematic search was performed in three databases, PubMed, EMBASE, and Cochrane, in March 2024, with no restriction on publication year. The following search strategy was utilized: (hypothermic oxygenated perfusion OR HOPE OR perfusion OR static cold storage) AND (liver transplantation OR transplant) AND (extended criteria donor). Two authors (V.S. and B.O.T.) independently reviewed the literature, and the disagreement was solved by the third author. The systematic review and meta‐analysis were conducted according to the Cochrane Collaboration and the Preferred Reporting Items for Systematic Reviews and Meta‐Analysis (PRISMA) recommendations [18]. The pre‐specified research protocol has been registered and is available at PROSPERO (CRD42024580024).

2.2. Selection Criteria and Analysis

Studies that met the following criteria were included: (1) Examined clinical outcomes following liver transplant (LT) with HOPE in ECD grafts, (2) Included a static cold storage (SCS) control group, and (3) Reported at least one of the endpoints required in this study. ECD grafts are defined as one of the following: donor age > 65 years; at least 7 days of ICU stay before donation; donor BMI > 30; >30% macrosteatosis on histology; serum Na > 155 mmol/L; AST or ALT > 3× upper limits normal; cardiac arrest before procurement. In case of duplicated data, we included the most complete study and/or most recent. Articles that were not in English, Spanish, or Portuguese were excluded. Case reports, case series, reviews, and editorials were excluded from this study.

2.3. Endpoints

Meta‐analysis was performed on the following endpoints: early allograft dysfunction (EAD), primary non‐function (PNF), hepatic artery thrombosis (HAT), anastomotic and non‐anastomotic stricture (NAS), portal vein thrombosis (PVT), 1‐year graft failure rate, and retransplantation rate, length of the hospital and ICU stay, comprehensive complication Index (CCI), and Clavien–Dindo grade. Complications were reported according to the COS for liver transplantation where possible, noting that not all studies reported all outcomes in the COS [17].

2.4. Quality Assessment

Two independent authors (B.O.T. and V.S.) performed the risk of bias and quality assessment for each study. For RCTs, a quality assessment was performed using the Cochrane tool for assessing bias of RCTs [19], wherein studies are categorized as low risk, moderate risk, or high risk in following five domains: randomization, deviation from intended intervention, missing outcome, measurement of the outcome, and selection of the reported results. For cohort studies, the Cochrane Collaboration tool for assessing the risk of bias in non‐randomized studies (ROBINS‐I) [20] was used. Studies were categorized as low, moderate, serious, or critical risk in the following domains: confounding, selection, classification, deviation from intended interventions, missing data, measurement of the outcomes, and selection of the reported results. Sensitivity analysis was performed using leave‐one‐out analysis.

2.5. Statistical Analysis

Effects for the binary endpoints were compared using pooled odds ratios (OR) with 95% confidence intervals (CI), weighted mean differences (MD) were used for continuous outcomes both using Mantel–Haenszel method. Heterogeneity was calculated using the Cochran Q test and I² statistics; p values inferior to 0.10, and I² > 25% were considered significant for heterogeneity. Fixed‐effect model was used for endpoints with low heterogeneity and a random‐effect model for endpoints with high heterogeneity. Cochrane Collaboration and the Preferred Reporting Items for Systematic Reviews and Meta‐Analysis (PRISMA) statement guidelines were used to perform this meta‐analysis. Review Manager 5.4 (Nordic Cochrane Centre, The Cochrane Collaboration, Copenhagen, Denmark) and the metafor package for RStudio version 4.2.2 (R Foundation for Statistical Computing, Vienna, Austria) were used for statistical analysis. First, pooled analysis was performed in all DCD and DBD ECD. Then we separately analyzed DBD ECD and DCD, and lastly performed an analysis on only randomized controlled studies for DBD ECD and DCDs.

3. Results

3.1. Study Selection and Baseline Characteristics

The initial search yielded 513 results; after duplicate removal and exclusion of studies that did not meet inclusion criteria, 19 studies were selected for the full‐text review. The PRISMA flow diagram illustrates the study selection and exclusion process (Figure 1). Twelve studies were ultimately included in the final cohort [21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32]: five randomized controlled and seven non‐randomized studies, comprising 1.833 patients. 29% (531 patients) received livers after hypothermic oxygenated perfusion, and 71% (1.302 patients) received grafts with static cold storage. Baseline characteristics and the endpoints were extracted and reported in (Tables 1 and 2). Recipient median age varied 46–63 years old in the HOPE group and 51–63 years old in SCS group; recipient median MELD score ranged from 11 to 19.5 in the HOPE group and 11.8 to 22 in the SCS group. Machine perfusion parameters are shown in (Table S1). None of the studies in this analysis included DCD livers that underwent procurement with normothermic regional perfusion (NRP) prior to HOPE or SCS.

FIGURE 1.

Prisma flow diagram: 543 articles were identified; 19 studies were selected for full‐text reading; 7 studies were excluded after full‐text review; 12 studies were included for meeting the pre‐specified inclusion criteria.

TABLE 1.

Baseline characteristics of included studies.

Refs.	Country	Design	Center type	ECD type	Follow‐up Months	Patients (#N)	Recpt. Age	Recpt. BMI	Recpt. MELD	Preservation time (min)
[31]	USA	RCT	Multicenter	Both	12 months	HOPE a : 63 SCS: 73	55 56	28.8 29.1	— —	— —
[30]	Poland	RCT	Single	DBD	3 months	HOPE: 26 SCS: 78	46 51	— —	12 14	608 567
[32]	Sweden	Retrospective	Single	DBD	—	HOPE: 30 SCS: 47	60.4 60.4	30.0 28.9	17.6 18.2	— —
[23]	Italy	Retrospective	Single	DBD	22 months:HOPE 43 months:SCS	HOPE: 121 SCS: 723	60.6 57.2	25.8 25.1	13 13	498 439
[22]	Italy	RCT	Single	DBD	12 months	HOPE: 55 SCS: 55	57 60	25.4 25.3	15 14	145 420
[21]	France	Retrospective	Single	DBD	12 months	HOPE: 25 SCS: 69	63 62	26.7 27.4	18.3 18.3	— —
[28]	Netherlands	RCT	Multicenter	DCD	6 months	HOPE: 78 SCS: 78	60 60	— —	14 16	524 409
[25]	Germany	RCT	Multicenter	DBD	12 months	HOPE: 23 SCS: 23	60 63	28 28	13 17	495 502
[27]	Switzerland	Retrospective	Multicenter	DCD	12 months	HOPE: 50 SCS: 50	58 57	— —	11 11.8	360 288
[29]	Netherlands	Prospective	Single	DCD	12 months	HOPE: 10 SCS: 20	57 52	— —	16 22	521 503
[9]	Switzerland	Prospective	Multicenter	DCD	12 months	HOPE: 25 SCS: 50	60 56	— —	13 16	— —
[24]	USA	Retrospective	Single	DBD	12 months	HOPE a : 31 SCS: 30	57.5 58.4	— —	19.5 21.4	— —

Note: Categorical data are presented as absolute numbers. Continuous data are presented as median or mean.

Abbreviations: BMI, body mass index; DBD, donor after brain death; DCD, donor after cardiac death; ECD, extended criteria donor; HOPE, hypothermic‐oxygenated perfusion; LT, liver transplant; MELD Score, Model for End‐Stage Liver Disease Score; RCT, randomized controlled trial; Recipt, recipient; SCS, static cold storage.

^a These two studies, Guarrera et al. and Panayotova et al. performed HOPE without additional oxygenation of the perfusate.

TABLE 2.

Perioperative and outcomes related to HOPE versus SCS for extended criteria liver grafts.

Refs.	Preservation	PNF	EAD	ICU Stay	LOS	HAT	PVT	TBC	NAS	AS	BL	CDG ≥3	1‐Year graft failure	rLT
[23]	HOPE: 121 SCS: 723	— —	3 39	3 3	11 12	2 22	— —	25 111	5 35	23 94	— —	21 180	31 105	— —
[24]	HOPE a : 31 SCS: 30	1 2	6 9	— —	13.6 20.1	1 2	0 2	4 13	— —	— —	1 3	— —	6 6	— —
[22]	HOPE: 55 SCS: 55	0 2	7 19	4 4	18 17	— —	— —	— —	— —	— —	— —	— —	1 7	0 6
[21]	HOPE: 25 SCS: 69	2 2	7 29	3 5	15 20	— —	— —	2 8	0 0	— —	— —	6 31	3 10	— —
[25]	HOPE: 23 SCS: 23	1 1	4 8	5 8	20 36	0 2	— —	4 6	— —	— —	— —	10 17	2 5	1 2
[9]	HOPE: 25 SCS: 50	0 3	5 22	3 3	20 18	1 3	— —	5 23	— —	— —	— —	— —	3 16	0 9
[27]	HOPE: 50 SCS: 50	0 2	— —	— —	— —	— —	— —	— —	4 11	12 9	1 1	— —	7 18	— —
[28]	HOPE: 10 SCS: 20	0 0	— —	2 2	22 23	0 2	— —	— —	1 7	2 3	— —	— —	— —	0 5
[29]	HOPE: 78 SCS: 78	0 1	20 31	2 2	15 15	2 2	0 2	— —	5 14	23 22	6 8	— —	— —	3 6
[30]	HOPE: 26 SCS: 78	0 3	7 26	5 4.5	— —	— —	— —	— —	— —	— —	— —	8 36	3mo‐ 1 3mo‐ 6	— —
[31]	HOPE a : 63 SCS: 73	0 3	7 12	3 3	9 9	1 1	0 1	8 19	0 4	7 12	1 3	— —	4 2	1 4
[32]	HOPE: 30 SCS: 47	— —	4 9	— —	— —	— —	— —	— —	— —	— —	— —	— —	— —	— —

Note: Categorical data are presented as absolute numbers. Continuous data are presented as median (range) or mean (± standard deviation).

Abbreviations: AS, anastomotic stricture; BL, biliary leak; CDG, Clavien Dindo Grade; HBV, hepatitis B viral infection/HCV, hepatitis C viral infection; LT, liver transplant/MELD Score, Model for End‐Stage Liver Disease Score; rLT, retransplantation; TBC, total biliary complication.

^a These two studies, Guarrera et al. and Panayotova et al. performed HOPE without additional oxygenation of the perfusate.

3.2. Pooled Analysis

Ten studies reported EAD, and pooled analysis demonstrated a significant reduction in the HOPE group compared to SCS (RR 0.59; CI 0.46–0.76; p < 0.001) (Figure 2a). Sensitivity analysis showed no differences in the heterogeneity after omitting each study (Figure 2b). Seven studies reported 1‐year graft failure rate with a significantly lower failure rate in the HOPE group (RR 0.56; CI 0.33–0.94; p = 0.02) (Figure 3a). Sensitivity analysis showed no difference after omitting each study (Figure 3b).

FIGURE 2.

EAD: (a) Pooled EAD analysis including DCD and DBD, showing significant reduction of EAD in the HOPE group (p < 0.001); (b) result remains significant after omitting each study during sensitivity analysis.

FIGURE 3.

(a) Pooled 1‐year graft failure analysis of DCD and DBD favoring a significant reduction in the HOPE group (p = 0.029); (b) high heterogeneity found in the pooled 1‐year graft failure sensitivity analysis; (c) 1‐year failure analysis of DCD grafts only, showing statistically significant differences favoring the HOPE group (p = 0.004).

Six studies reported retransplantation rate with a significant reduction in retransplantation in the HOPE group (RR 0.30; CI 0.12–0.71; p = 0.007) (Figure 4a), and no difference was seen in the heterogeneity after omitting each study (Figure 4b). Six studies reported NAS, again demonstrating a lower rate in the HOPE group (RR 0.46; CI 0.27–0.78; p = 0.004) (Figure S1a). Finally, Clavien Dindo Grade >3 was also lower in the HOPE group (RR 0.64; CI 0.49–0.84; p = 0.001) (Figure S1b).

FIGURE 4.

(a) pooled retransplantation rate analysis yielded lower rates in the HOPE group (RR 0.30; p = 0.007); (b) results remain statistically significant after omitting each study in a sensitivity analysis—lower heterogeneity; (c) subgroup analysis of RCTs also showing significant reductions of retransplantation rate in the HOPE group compared to SCS group.

No significant differences were seen in primary non‐function (RR 0.54; CI 0.22–1.32; p = 0.17), hepatic artery thrombosis (RR 0.59; CI 0.26–1.34; p = 0.21), portal vein thrombosis (RR 0.74; CI 0.10–5.13; p = 0.76), anastomotic biliary stricture (RR 1.20; CI 0.91–1.58; p = 0.18), bile leak (RR 0.62; CI 0.27–1.41; p = 0.26), and total biliary complication (RR 0.61; CI 0.34–1.10; p = 0.10) (Figure S1c–i) respectively.

3.3. Subgroup Analyses

3.3.1. Donation After Brain Death (DBD)

First, subgroup analysis was performed using seven studies that included only ECD grafts from donors after brain death (DBD). EAD was reported by all seven studies showing statistically significant reduction of EAD in the HOPE group (RR 0.59; CI 0.43–0.82; p = 0.002) (Figure S2a). Similarly, the length of the hospital stay, reported by five studies, was significantly reduced in the HOPE group (Mean Difference—MD –3.92; –7.64 to –0.19; p = 0.04) (Figure S2b).

Five studies reported PNF with no difference between groups (RR 0.85; CI 0.28–2.57; p = 0.78) (Figure S2c). One‐year graft failure rate was also reported by five studies with comparable results between the groups (RR 0.82; CI 0.38–1.80; p = 0.63) (Figure S2d). Similarly, HAT was reported in three studies and was similar between the groups (RR 0.46; CI 0.15–1.42; p = 0.18) (Figure S2e). Retransplantation rate was also comparable (RR 0.23; CI 0.04–1.54; p = 0.13) (Figure S2f).

3.3.2. Donation After Circulatory Death (DCD)

We next performed a subgroup analysis using four studies that included only DCD grafts. Two studies in this subgroup reported EAD with significant reduction of EAD in the HOPE group (RR 0.59; CI 0.40–0.89; p = 0.01) (Figure S3a). Similarly, two studies reported 1‐year graft failure again favoring reduced incidence in the HOPE group (RR 0.38; CI 0.20–0.73; p = 0.004) (Figure S3c). The rate of NAS was significantly lower in the HOPE group (RR 0.35; CI 0.18–0.69; p = 0.002) (Figure S3b).

Four studies reported PNF, and three studies each reported retransplantation and HAT. There was no difference in PNF (RR 0.26; CI 0.05–1.52; p = 0.13) (Figure S3c), retransplantation rate (RR 0.33; CI 0.11–1.01; p = 0.05) (Figure S3d), and HAT (RR 0.72; CI 0.19–2.65; p = 0.62) (Figure S3e). The length of the hospital stay, and anastomotic stricture were also comparable between the groups (MD 0.63; CI –1.04 to 2.30; p = 0.46) (Figure S3f) and (RR 1.13; CI 0.76–1.70; p = 0.54) (Figure S3g), respectively.

3.3.3. Randomized‐Controlled Trials

And last, we performed subgroup analysis of randomized studies including DCD and DBD grafts. Five studies reported EAD again with reduced EAD in the HOPE group (RR 0.61; CI 0.45–0.83; p = 0.002) (Figure S4a). The retransplantation rate was again lower in the HOPE group (RR 0.36; CI 0.14–0.95; p = 0.04) (Figure 4c) and was reported by four studies. The rate of NAS was significantly lower in the HOPE group (RR 0.32; CI 0.13–0.81; p = 0.02) (Figure S4b).

Five studies reported PNF, four reported hospital stay and three reported HAT. The rate of PNF, HAT, length of hospital stay and anastomotic stricture were comparable between the groups (RR 0.36; CI 0.10–1.33; p = 0.12), (RR 0.73; CI 0.18–2.95; p = 0.66), (MD –1.71; CI –5.93 to 2.5; p = 0.43), and (RR 0.94; CI 0.61–1.44; p = 0.78) (Figure S4c‐f), respectively.

3.4. Quality Assessment

Figure S5a,b summarize the individual appraisal of each study included in this meta‐analysis using the Rob‐2 and ROBINS‐I quality assessment tools, respectively. In four RCTs, the risk of bias assessment demonstrated insufficient personnel and blinding participants resulting in high risk of randomization process. One RCT has a risk for unblinded outcome assessment. For the non‐RCTs, moderate risk of bias due to confounding was seen in most of the studies, in addition, one study had the risk of bias in the measurement of outcomes.

4. Discussion

This systematic review and meta‐analysis assessed the effect of HOPE for ECD donor livers compared to SCS and showed the following main findings: First, HOPE significantly improves certain posttransplant outcomes including EAD, graft failure, retransplantation, non‐anastomotic biliary strictures (NAS), and overall major complications graded by Clavien Dindo. Second, HOPE had greater impact on improving 1‐year graft survival in the DCD subgroup compared to ECD‐DBD livers. Similarly, we observe the reduction of NAS in the DCD subgroup. Third, sensitivity analysis was performed limited to RCT studies, confirming the HOPE technique demonstrated lower rates of EAD, retransplantation and NAS. Finally, we find that several studies did not report biliary complications with enough detail, which highlights the need for standardized outcomes reporting across machine perfusion studies and in general [17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33]. Overall, our findings support the broad use of HOPE for dynamic preservation of ECD liver grafts compared to the SCS.

Previous studies on the effects of machine perfusion, including NMP and HOPE demonstrated superior short‐term outcomes favoring machine perfusion over static cold storage [6, 34, 35]. Such findings have even been echoed in a previous meta‐analysis on the impact of HOPE in liver transplantation which demonstrated a reduction of EAD with HOPE [36]. Another meta‐analysis focused on DCD livers also demonstrated a reduction of EAD [15]. There were certain limitations to these analyses which prompted this work. Two of the included studies did, however, include the same proportion of patients. Moreover, reported outcomes were pooled for both DCD and DBD grafts which might not be helpful in assessing the need for perfusion in livers with different risk profiles. This study provides a relevant update including the most recent literature and is the first focusing on the impact of HOPE on different ECD graft types. We notably find a benefit of HOPE versus SCS with respect to several outcome measures, including NAS‐rates and graft survival. Liver transplants with ECD donors are known for increased susceptibility to biliary complications, with previous studies demonstrating clinically that HOPE can mitigate such risk [37]. For example, Guarrera et al. were the first to report an impact on biliary complication with hypothermic machine perfusion [24, 38]. The re‐introduction of oxygen at hypothermic temperatures through HOPE, has been linked to a protective effect on mitochondrial complex 1&2, reducing release of reactive oxygen species and IRI, which in turn reduces cell death and inflammation. Such reduced IRI‐associated inflammation was linked to a better recovery from the hit during transplantation with less biliary structuring and less NAS [4, 27, 39, 40, 41]. This meta‐analysis supports previous studies showing robust protection from NAS, perhaps emphasizing how dynamic preservation can help extend utilization of donor grafts. This might, in turn, help further accelerate improvement in waitlist outcomes and transplant access [42].

Most of the reported studies with MP simply report biliary complications, without distinguishing between anastomotic, non‐anastomotic strictures, or biliary leaks [22, 24, 25, 26]. These three categories of biliary complication have different etiologies and recipients may have a completely different clinical post‐transplant course [43]. Our study separately analyzed NAS from AS and leaks, finding reduced NAS‐rates and no difference in AS or bile leak. This fits mechanistically with the known multiple confounders impacting on AS rates [27, 40, 41]. However, this also highlights an opportunity for improvement in research methodology which was addressed in the recent proposal for adoption of COS in liver transplantation. Ensuring consistency in reporting outcomes in liver transplantation is essential for accurately assessing new interventions and perfusion technologies. To achieve this, outcome reporting should be thorough, standardized and replicable, and should focus on clinical relevance. A recently proposed COS in liver transplantation addressed this need and included, survival results, biliary, vascular, and overall complications. We advocate for a routine use of these metrics, developed based on prior international benchmarks analyses [17, 33, 44].

Regarding overall complications, Czigany et al. reported a significant reduction in cumulative CCI in 90‐day and 6 months favoring the HOPE group, and we do find a reduction in major complications according to the Clavien scale with HOPE. However, we also note a surprising lack of uniformity in the reporting timeframe between the included studies, with studies presenting CCI at hospital discharge time, 90‐day, 6 months, and after 1‐year [22, 23, 25, 27]. We did not perform analysis of CCI specifically for such reasons. However, CCI has been linked to costs and is the most holistic and generalizable outcome making the standardized reporting of such outcomes critical in future studies [17].

The benefit of HOPE in ECD donation after DBD versus DCD remains an area that may benefit from further evaluation. Rayar et al. [21], found for example significantly lower ALT, lactate, and creatinine levels in the early postoperative period. However, there was no difference between HOPE and SCS in early allograft dysfunction (EAD) and primary non‐function. Similarly, others describe no difference in EAD‐rates between study groups [24, 31]. In contrast, one RCT [22] demonstrated a significant reduction in graft dysfunction with HOPE. In our pooled analysis, the odds of EAD were 57% lower in the HOPE group compared to SCS. Similarly, we also found decreased odds for EAD with HOPE in all subgroups including ECD‐DBD and RCTs. However, we also note that EAD has many inherent limitations, including its’ lack of validation after dynamic machine preservation, and the limited correlation with long‐term graft outcomes. In its most basic sense, EAD is a measure of early enzyme levels, which are unsurprisingly altered with hours of dynamic liver preservation with enzyme release into perfusates, followed by discard of that very perfusate [17, 33, 45]. The metric EAD describes liver injury and should ideally be considered in combination with other core outcome set metrics for liver transplantation.

In their RCT, Ravaioli et al. [22] reported superior 1‐year graft survival after HOPE for DBD‐ECD grafts. In contrast, Czigany et al. [25], did not report a different graft survival. Combining all reported DBD‐ECD studies, our results found comparable 1‐year graft survival rates between HOPE and SCS, possibly influenced by the overall very heterogenic group of DBD‐ECD grafts lacking one clear definition in the field. However, HOPE treatment achieved better graft survival in both the pooled analysis and the DCD subgroup. Based on the known higher injury, this improvement in DCD transplantation is not surprising and support previous reports of DCD grafts benefiting most significantly from HOPE [9, 11].

DCD liver grafts are also associated with higher biliary complication with the incidence reaching around 30% [46], NAS being the most common biliary complication in these patients. NAS can significantly impact graft survival and the need to redo liver transplantation. A recent meta‐analysis reported a reduction of total biliary complication and NAS in the HOPE group [15]. However, this study did not distinguish between the different graft subtypes. Our pooled and subgroup (DCD and RCT) analysis support such results, though, we did not find a difference in the overall biliary complication among the groups. We hypothesize that ischemic biliary injury is also mitigated with HOPE in DBD grafts, despite not finding significant differences in this cohort, mainly secondary to NAS‐rates in DBD‐ECDs being considerably lower, preventing robust subgroup analysis. Most of the included studies in this meta‐analysis primarily defined NAS based on radiological findings. The RCT by van Rijn et al., explored the effect of HOPE on symptomatic NAS and radiological diagnosed NAS. The study reported the reduction of NAS in symptomatic NAS, but no difference was found when considering all cases that met radiological criteria [29].

Hospital stay is most frequently seen as both an outcome and a process measure and cannot be linked exclusively to surgical factors. Based on the noisy parameter with many confounders, the impact of HOPE on hospital stay is mixed and often recipient hospital stays are comparable [22, 23]. In contrast, some authors report a significant reduction in hospital stay with HOPE [21, 24, 25]. In this meta‐analysis, we found a significant reduction in hospital stay favoring the HOPE, particularly in the subgroup analysis on DBD‐ECD, a finding possibly based on a reduction in early IRI‐associated inflammation that could lead to quicker recipient recovery. Shorter hospital stays may further help to reduce transplant costs with use of HOPE [47]. Overall, the impact of HOPE in reducing ischemia reperfusion injury and immune response modulation [48, 49] greatly benefits ECD grafts, as they are inherently vulnerable to severe complications and impaired posttransplant outcomes. In addition, a recently published network meta‐analysis investigating the impact of HOPE, NMP, and SCS in extended criteria donor, demonstrated superiority of HOPE compared to NMP and SCS in reducing complications after liver transplant [50]. Therefore, we recommend the use of HOPE for ECD as a strategy of mitigating graft dysfunction and optimizing quality and improve outcomes.

This meta‐analysis has several limitations. First, seven of the 12 studies included in this meta‐analysis are non‐randomized studies and are susceptible to retrospective data collection bias. Second, the modality of HOPE differs between the studies. In addition, there was no uniformity in reporting biliary complications as the etiologies of the complications are distinct. There is not a strict and well‐accepted definition for extended criteria donors, with relevant differences between the countries and center. There is an additional lack of standardized definitions for outcomes such as biliary complications including NAS. In addition, outcomes are reported at different time‐points during posttransplant follow up. To prevent inappropriate comparison, we removed studies from consideration if outcomes did not match general definitions and equivalent timepoints, however this did reduce the sample size for certain analyses. Furthermore, we performed sensitivity analyses counting for the outcomes with high heterogeneity. We do also highlight that our study intentionally did not compare HOPE versus NMP, and thus no conclusions can be made based on this study about the effect of NMP on transplant outcomes.

In summary, this meta‐analysis included nine studies analyzing the impact of HOPE on outcomes after transplantation with different ECD organs. Pooled analysis demonstrated significantly better posttransplant outcomes in the HOPE group compared to the traditional SCS. Despite positive results with HOPE in RCTs including only ECD‐DBD grafts, subgroup analysis showed even better results for HOPE in DCD livers compared to the heterogenic cohort of DBD‐ECD grafts. This underscores the potential of HOPE as a preferred preservation approach for ECD grafts. However, future studies could provide further details on the impact of HOPE on specific ECD‐DBD subgroups. Exploring the impact of a combination of different perfusion strategies on overall and biliary outcomes is of clear interest. Sequential perfusion techniques, including NRP followed by HOPE or NMP or controlled oxygenated rewarming could be compared with isolated perfusion modalities in future studies. The publication of two RCTs with endischemic portal venous HOPE in ECD‐DBD livers are currently awaited and may support the findings of this study further.

Author Contributions

The study was conceptualized and conducted under the direction of Dr. Andrea Schlegel and Dr. Chase J. Wehrle. Data collection and analysis was performed by Valberto Sanha, Bruna Oliveira Trindade, and Chase J. Wehrle. Manuscript drafting was performed by Valberto Sanha, Chase J. Wehrle, Koji Hashimoto, and Andrea Schlegel. Critical manuscript review was performed by these Valberto Sanha, Bruna Oliveira Trindade, Sangeeta Satish, Laura Batista de Oliveira, Omer Faruk Karakaya, Chunbao Jiao, Keyue Sun Muhammad Ahmad Nadeem, Charles Miller, authors.

Conflicts of Interest

Dr. Schlegel is a paid consultant for Bridge to Life Ltd. and Organox Ltd and Dr. Hashimoto is a paid consultant for Organox Ltd. The remaining authors have no conflict of interest.

Supporting information

Table 1: Perfusion parameters

Figure 1a: Pooled NAS: HOPE vs. SCS

Figure 1b: Pooled Clavien Dindo Grade >3: HOPE vs. SCS

Figure 1c: Pooled PNF: HOPE vs SCS

Figure 1d: Pooled HAT: HOPE vs SCS

Figure 1e: Pooled PVT: HOPE vs SCS

Figure 1f: Pooled Anastomotic Stricture: HOPE vs SCS

Figure 1g: Pooled Biliary Leak: HOPE vs SCS

Figure 1h: Pooled Total Biliary Complication: HOPE vs SCS

Figure 1i: Pooled Length of the Hospital Stay: HOPE vs SCS

Figure 2a: DBD Only EAD: HOPE vs. SCS

Figure 2b: DBD Only ‐ Length of Hospital Stay: HOPE vs. SCS

Figure 2c: DBD Only ‐ Primary Non‐Function: HOPE vs. SCS

Figure 2d: DBD only ‐ One‐Year Graft Failure Rate: HOPE vs. SCS

Figure 2e: DBD only ‐ HAT: HOPE vs. SCS

Figure 2f: DBD only ‐ Retransplantation: HOPE vs. SCS

Figure 3a: DCD Only – EAD: HOPE vs. SCS

Figure 3b: DCD Only – NAS: HOPE vs. SCS

Figure 3c: DCD Only ‐ PNF: HOPE vs. SCS

Figure 3d: DCD Only ‐ Retransplantation Rate: HOPE vs SCS

Figure 3e: DCD Only ‐ HAT: HOPE vs SCS

Figure 3f: DCD Hospital Stay: HOPE vs. SCS

Figure 3g: DCD Anastomotic stricture: HOPE vs. SCS

Figure 4a: RCT only (DCD + DBD) ‐ EAD: HOPE vs SCS

Figure 4b: RCT only (DCD + DBD) ‐ NAS: HOPE vs SCS

Figure 4c: RCT only (DCD + DBD) ‐ PNF: HOPE vs SCS

Figure 4d: RCT only (DCD + DBD) ‐ HAT: HOPE vs SCS

Figure 4e: RCT only (DCD + DBD) ‐ Length of Hospital Stay: HOPE vs SCS

Figure 4f: RCT only (DCD+DBD) ‐ Anastomotic Stricture: HOPE vs. SCS

Figure 5a: Risk of bias summary for non‐randomized studies (ROBINS‐I)

Figure 5b: Risk of bias summary for randomized studies (RoB)

Sanha V., Trindade B. O., Satish S., et al. “Hypothermic Oxygenated Perfusion Versus Static Cold Storage in Transplantation of Extended Criteria Liver Grafts: A Systematic Review and Meta‐Analysis.” Clinical Transplantation 39, no. 9 (2025): 39, e70291. 10.1111/ctr.70291

Data Availability Statement

The data that support the findings of this study are publicly available, they are part of published studies. These data were derived from the following resources available in the public domain: [PubMed, EMBASE, Cochrane]