Categorical variables were reported using frequencies and percentages. Continuous variables were reported using medians and interquartile ranges. Group differences were assessed using the Chi-square test for categorical variables and the Mann-Whitney U test for continuous variables. Pearson’s correlation coefficient was used to evaluate changes in the percentage of each preservation group and donor age group over time.

Post-LT patient and graft survival were the primary outcomes. Patients were censored at the last follow-up. The Kaplan-Meier method was used to estimate patient and graft survival. Log-rank test was used to assess differences in post-transplant survival across graft types. Further analysis using multivariable Cox regression evaluated survival by preservation group (for parsimony, defined as SCS vs. non-SCS), adjusting for confounding factors. The variables included in the multivariable Cox regression were selected a priori based on biological importance and data availability to avoid the inferential limitations of selecting variables for multivariable models based on univariable comparisons or stepwise procedures [28]; these were preservation group, WIT, donor age group, donor and recipient BMI, recipient age, laboratory Model for End-stage Liver Disease (MELD) score, and recipient transplant indication/diagnosis. Cohort development and statistical analyses were conducted using Stata IC 19.0 (Stata Corp LLC, College Station, Texas, US).

N = 11,648 DCD LTs. Although the annual use of DCD livers, particularly older DCD liver grafts, increased across both eras, this augmentation was exponential in the MP era (Figures 1, 2).

In the early period, 77.2% of DCD livers were from donors aged <50 years, with the remainder aged 50–69. No DCD livers from donors aged ≥70 were used before 2021 (Table 1).

In the late period, annual DCD LT volume growth accelerated (Figure 2), with significant increases in donors aged 50–59 (from 19.3% to 26.1%) and 60–69 (from 3.5% to 24.1%). Donors ≥60 increased from 3.5% to 27.3% – a 7.8-fold rise–, including 266 (3.2%) DCD donors aged ≥70; a USA first. The proportion of younger donors (18–49 years) declined from 77.2% to 55.3%.

During our study period, total LT rates increased by 62.7% (from 8,497 LTs in 2016 to 13,824 in 2025). By the end of 2025, DCD allografts (n = 5,669) accounted for 43.2% of deceased-donor allografts (n = 13,116). In the MP era, donors ≥50 (n = 3,478) accounted for the largest share of DCD donors (61.35%) (Figure 2). DCD SCS dropped from 94.8% to 31.4% in the MP era, while there was a steep rise in NRP (3.0% vs. 16.5%), NMP (2.1% vs. 38.9%), and sequential NRP-NMP deployment (0.1% vs. 10.7%) (p < 0.001) (Figure 3; Table 1).

The MP-era was associated with 44.6% decrease in median waitlist time (112–62 days, p < 0.001) (Table 1). The primary LT indication shifted from hepatocellular carcinoma (27.2%) to alcohol-associated liver disease (35.9%) (p < 0.001). The median donor WIT increased from 23.0 to 27.0 min.

In the pre-MP era, graft survival was higher in donors aged 18–49 (Figure 4C). Early graft survival was comparable across all age groups in the MP era (Figure 4D). No statistically significant difference in patient survival was observed amongst the three donor age groups (Figures 4A,B).

For the 50–59-year cohort, 1- and 3-year graft survival was 87.9% and 78.4% in the pre-MP era (n = 639), and 90.1% and 78.8% in the MP era (n = 2,178), respectively (Table 2). For the 60–69-year group, 1- and 3-year graft survival was 85.0% and 80.6% in the pre-MP era (n = 115), and 90.0% and 71.3% in the MP era (n = 2,007), respectively (Table 2). Utilization of ≥70-year-old DCD livers was reported only in the MP era, with 1-year survival of 87.8% (Table 2).

In multivariable Cox regression, SCS (hazard ratio [HR] = 1.19, 95% confidence interval [95%CI]: 1.00–141, p = 0.045) and increasing recipient age (HR = 1.04, 95% CI: 1.03–1.04, p < 0.001) were associated with increased risk of mortality when adjusted for donor age group, WIT, donor BMI, recipient BMI, diagnosis, and laboratory MELD score (Table 3). SCS (HR = 1.29, 95% CI: 1.13–1.48, p < 0.001), donor age 50–69 years (vs. 18–49 years), and increasing recipient age (HR = 1.01, 95% CI: 1.01–1.02, p < 0.001) were also associated with increased risk of graft loss when adjusted for WIT, donor BMI, recipient BMI, diagnosis, and laboratory MELD score (Table 3).

Before MP widespread adoption, the standard organ preservation approach was SCS, which involved flushing the organ with preservation solution at 4 °C and transporting it under cold conditions (Table 4). Older DCD liver grafts were associated with high rates of ischemic biliary complications, EAD, post-reperfusion syndrome, and graft loss [14, 30, 31]. Donors ≥40 years and extended CIT appeared to further increase this risk [3]. Post-reperfusion syndrome, more common in DCD LT recipients, was associated with higher rates of acute kidney injury and decreased 5-year graft survival [31].

These early reports defined the DCD practice in the USA for the following decade (2010–2020). Transplant centers adopted strict donor selection criteria, often excluding liver grafts from donors >50⁵. Most of these organs were more likely to be discarded rather than transplanted [14, 19, 32]. During the same period, older DCD liver grafts were used in Europe, albeit selectively, with acceptable outcomes [7, 32, 33]. The UK has historically had the largest cohort of cold-stored DCD LTs without MP, with favorable outcomes [6, 34, 35].

The use of older DCD donors reached a pivotal juncture in 2021, which coincided with the FDA approval of the TransMedics ® Organ Care System (OCS) platform (September 2021), followed by the FDA approval of OrganOx metra® System (December 2021), and the first publicized reports of NRP LTs in the USA (May 2022) [27, 29, 36]. By 2025, DCD livers accounted for 41.0% of LTs, a remarkable 5.5-fold increase over a decade (Figure 1). This sharp rise indicates a major shift in transplant practice and the acceptance of older DCD liver grafts in the USA. Similar trends have been observed globally, making older DCD donors the latest Frontier in the organ donor pool [13].

NMP enhances the use of older DCD liver grafts by enabling real-time viability assessments and reducing organ discard rates (7.25% vs. 30.52% for SCS) [37, 38]. NMP also reduces the incidence of EAD [31, 39–41]. However, according to the European experience, its effect on long-term graft survival and biliary complications is less pronounced than that of HOPE [42] (Table 4).

Following the first landmark USA NRP cohort publications [20, 22], NRP utilization increased from 3% to 21.7%, and to nearly 41.5% when including sequential NRP. NRP has been associated with significant reductions in NAS, PNF, and EAD, as well as improved 1-year graft survival [20, 22, 31, 42]. Although it is increasingly adopted as the preferred approach in DCD procurement across the organ procurement organizations (OPOs), its logistical complexity often limits its use [31, 42] (Table 4).

HOPE improves 1-year graft survival and reduces retransplantation rates [43] (Table 4). Although long-term survival data are still emerging, large real-world cohorts of HOPE-treated DCD liver grafts, including those from older donors, report a 5-year death-censored graft survival rate of 81% and low rates of graft loss due to PNF or NAC, regardless of donor age or risk profile [23].

The latest US-based RCT (PILOT trial) on HOPE used a portable HOPE device under FDA and institutional review board oversight [44]. The trial confirmed safety and demonstrated improved early clinical function with HOPE, including in patients with DCD livers from donors >65⁴⁴. These findings are consistent with broader evidence that HOPE reduces EAD and biliary complications in DCD livers, including those from older donors, compared to SCS [12, 44, 45]. This technology has only recently received FDA approval for standard-of-care use in LT in the USA [24].

Sequential NRP, i.e., NRP followed by ex-situ NMP or HOPE, has shown a lower rate of NAS compared to SCS and single-modality perfusion, even in grafts from donors ≥60 years [19, 42] (Table 4). Meta-analyses confirm that HOPE-based strategies, including sequential protocols, are associated with improved 1-year graft survival and reduced graft loss, as well as a lower risk of biliary complications compared to single-modality NMP or SCS [46, 47].

In conclusion, the widespread adoption of MP has lifted prior reservations about the broader use of older DCD livers. Although not explicitly associated with the introduction of MP—given that multiple simultaneous changes occurred during this period—nearly 50% of current DCD donors are aged 50 or older, including those aged 70 or above. This demographic constitutes the fastest-growing donor group in the US [29].

NMP requires a perfusion system, blood products, and specialized personnel, making it most suitable for centers that can absorb the higher acquisition costs and accommodate complex logistics [40, 41, 48]. NMP may be initiated at the donor hospital or applied at the recipient center. The latter (“end-ischemic” or “back-to-base” NMP) is less effective in terms of its ischemic bilioprotective effect due to the inevitable intervening cold ischemic period during SCS [14, 41].

The TransMedics RCT, a multicenter study evaluating the TransMedics® OCS for NMP, reported that OCS enabled the safe utilization of higher-risk and marginal donor livers, including a significantly higher proportion of DCD grafts, compared to SCS [49]. OCS recipients experienced fewer cases of reperfusion syndrome, less blood loss and transfusions, shorter surgeries, and shorter ICU and hospital stays. However, OCS organ acquisition costs were substantially higher. The median OCS organ acquisition cost was $135,930, compared to $50,940 for SCS, representing a nearly threefold increase [49]. Main cost factors include service, disposable perfusion components, and logistics of remote perfusion initiation at the donor hospital. Notably, these costs were not offset by reductions in other perioperative expenses during the index hospitalization, as overall hospitalization costs remained higher for OCS cases ($256,810 vs. $209,144 for SCS) [49] (Table 5).

The back-to-base NMP model, particularly with devices such as OrganOx metra®, although it increases organ acquisition and preservation costs, is associated with fewer perioperative complications and does not increase overall short-term healthcare costs in DCD LT [48] (Table 5).

MP applications also enabled LT to transition from emergency to semi-elective daytime surgery, alleviating strain on the operating apparatus and expediting avoidance of costlier, out-of-hours complex procedures [52].

NRP use has increased over the recent yrs, with more OPOs standardizing its use in DCD procurements. Implementation involves extracorporeal membrane oxygenation equipment, specialized training, and coordination with the OPOs and across procurement teams, making it resource-intensive. However, it is cost-effective because it increases organ yield and reduces procurement costs compared with super-rapid recovery (SRR) with ex-situ NMP [51]. The estimated NRP procurement cost in the USA is $9,463.22 per donor. Conservative estimates indicate that approximately 31 donor allografts can be procured using NRP at a cost equivalent to that of a single allograft procured via SRR with ex-situ perfusion (Table 5) [51]. This difference stems from improved resource utilization and fewer discards (Table 5).

European economic analyses indicate that HOPE is cost-effective, with 1-year post-transplant costs lower than those of SCS, primarily due to reduced post-transplant intervention costs resulting from lower biliary complications. In a multicenter RCT in the Netherlands, the mean total cost per patient was €110,794 for HOPE versus €126,221 for SCS, with the greatest savings in intensive care unit and nonsurgical interventions [50]. Cost-effectiveness was achieved with as few as 1–30 procedures per yr, depending on whether personnel and facility costs were included, making HOPE feasible for both high- and moderate-volume centers [50] (Table 5). HOPE logistics include a perfusion device, oxygen supply, and trained staff. The protocol is straightforward and can be performed at the transplant center [50]. On January 20, 2026, the FDA granted de novo clearance to the Bridge to Life™ VitaSmart™ HOPE system for routine clinical use in LT in the USA [24].

OPTN does not reliably record perfusion modality or detailed DCD agonal-phase metrics, nor does it distinguish SCS from MP. It has limited data on DCD-specific morbidity, such as biliary complications. The MP era has been shorter than the pre-MP era, potentially introducing follow-up bias into the analysis. NRP identification was based on surrogate coding and may have been subject to misclassification. The cohort was also subject to temporal misclassification bias, as era definitions based on OCS FDA approval (2021) predated the first reported NRP LTs (2022), potentially exaggerating the impact of technologies that have been broadly available the longest. Another significant limitation is the aggregation of MP modalities, despite evidence indicating that each modality has distinct mechanisms and outcomes; therefore, it remains uncertain which modality genuinely contributes to the observed improvements, aside from temporal misclassification. This latter misclassification may either amplify or obscure the impact of each modality on the outcomes, depending on the cohort’s exposure to the respective technology. Other limitations stem from the study’s retrospective design, including limited variables, residual confounding, data inaccuracies, and the inability to track changes. Additional issues include selection bias, a smaller sample size in the older group, and a higher risk of type II errors.

A significant challenge lies in the rising costs of organ acquisition and the uncertainty surrounding insurance coverage for these modalities [49].

Large, multicenter RCTs may be required to determine the individual benefits of different perfusion techniques in DCD LT, including older DCD donors. These studies should focus on long-term survival, quality of life, and cost-effectiveness [23, 43]. DCD LT benchmarking should be conducted prospectively, with donor–recipient risk matching, comparing risk and outcomes with the best possible results from SCS [53]. Research should also aim to standardize viability assessment criteria and perfusion protocols [23, 54, 55].

Future clinical use will likely include therapeutic interventions during MP, such as defatting steatotic livers, delivering gene therapies, or administering regenerative agents, to further enhance the quality and utilization of older DCD liver grafts [56–60]. The integration of artificial intelligence and advanced analytics to interpret perfusion data and provide decision support is also anticipated. Ongoing and future studies should also evaluate the impact of these technologies on organ allocation, waitlist mortality, and health system resource use [42, 61].