Machine perfusion (MP) has emerged as a new standard of care for liver transplantation in extended-criteria donors (ECDs), such as those with steatotic grafts, aged grafts, donation after circulatory death (DCD), and grafts with viral infections. The application of various MP techniques has significantly expanded the donor pool, particularly for DCD livers, which now account for up to 40% of transplants in some countries.1 Several pilot and randomized controlled trials (RCTs) have demonstrated the efficacy of MP in this setting. The first multicenter RCT using normothermic MP (NMP) for ECD demonstrated that NMP significantly reduced early graft injury, organ discard rate, and prolonged preservation time but did not benefit patient or graft survival.2 In parallel, another European multicenter RCT explored dual hypothermic oxygenated MP (D-HOPE), reporting superior outcomes in DCD liver transplantation, including the reduced rates of early allograft dysfunction (EAD), postreperfusion syndrome (PRS), and biliary complications.3 As high-quality RCTs continue to be conducted globally, MP indications are expanding, together with the development of combined or sequential perfusion protocols.4,5 Nevertheless, evidence remains scarce regarding which MP approach is best suited for specific ECD subtypes. In the current issue, Stefano et al6 presented new data addressing this question. In this retrospective single-center study, they evaluated multiple MP strategies—including normothermic regional perfusion (NRP) followed by either D-HOPE or NMP, as well as D-HOPE alone—applied to various subtypes of ECDs, such as DCD and ECDs after brain death (ECD-DBD). Strategically, the team used NRP followed by MP for DCD grafts: 50 donors received by D-HOPE, whereas the other 11 underwent NMP. The median NRP duration was 240 min for all DCD grafts. For ECD-DBD grafts, pure D-HOPE was used with a median perfusion time of 137 min, significantly shorter than the MP time in the NRP-MP group. The DBD grafts preserved by static cold storage served as controls. Following propensity score matching, demographic characteristics were comparable across groups, with the exception of significantly higher warm ischemic time (44 min) in the DCD group and older donor age in the ECD-DBD group. Despite these differences, their study reported similar long-term patient and graft survival across donor subtypes, managed with tailored MP protocols. One-year survival rates ranged from 91.8% to 94.2%, and 3-y survival rates varied between 88% and 93.4%, consistent with benchmark studies. Notably, the incidence of ischemic cholangiopathy was only 3.3% in the MP group and 4.9% in the static cold storage group. Other early complications, including EAD, PRS, and acute kidney injury, were also comparable among the 3 groups. The sequential application of NRP followed by NMP was associated with improved outcomes and feasibility for long-distance organ transport. Although a relatively higher incidence of severe PRS (41.7%) was observed in this subgroup, the small sample size necessitates further validation through multicenter prospective studies to better assess its association with clinical meaningful outcomes. Notably, the DCD cohort in this study faced additional challenges due to Italian legislation requiring 20 min of absent cardiac electrical activity to declare death, prolonging warm ischemic time by an extra 10–15 min. Despite this, the use of tailored MP protocols still yielded excellent outcomes, thereby underscoring the role of personalized medicine in graft preservation and transplantation. Current RCT or cohort studies also underscore the potential for innovative MP applications beyond preservation. MP can serve as a platform for targeted drug delivery, graft bioengineering, and quality assessment.7 Several preclinical studies have already demonstrated this concept, using MP to deliver siRNA, defatting cocktails, novel oxygen carriers, and other agents. An automated MP device was designed for simultaneous management of physiological homeostasis and assessment of graft quality, enabling injured human livers to be transplanted after 7 d of ex vivo perfusion.8 The same team also applied defatting strategy to severely steatotic liver during MP for 12 d, increasing the organ viability for transplantation.9 Another team used MP in animal models to deliver siRNA into grafts, achieving efficient gene editing within a short perfusion window and highlighting the potential of MP-facilitated genetic modification.10 In summary, MP has established its role in improving outcomes for ECD livers, with modalities such as NMP and HOPE demonstrating significant benefits in graft survival and complication reduction. Looking ahead, the field must now move toward defining personalized perfusion protocols, matching specific MP techniques to distinct donor graft subtypes and recipient underlying diseases to optimize results. Furthermore, through close integration of basic science research and clinical validation, the future of MP promises to unlock innovative applications as a dynamic therapeutic platform, such as ex vivo graft repair, targeted drug delivery, and genetic modification, ultimately rescuing untransplantable organs and further expanding the donor pool (Figure 1).FIGURE 1.: Personalized machine perfusion: key factors and its future directions as a diagnostic and therapeutic platform. D-HOPE, dual hypothermic oxygenated machine perfusion; HMP, hypothermic machine perfusion; NMP, normothermic machine perfusion; NRP, normothermic regional perfusion. Modified using BioRender.
Liu et al. (Wed,) studied this question.