Understanding the physiologic impact of donor extubation and warm ischemia on the DCD heart is fundamental to developing a successful resuscitation strategy (24). After extubation, the DCD heart is forced to function in an increasingly hypoxemic environment while attempting to maintain systemic oxygen delivery (24). Progressive hypoxemia and hypercarbia cause constriction of the pulmonary vasculature and distention of the right ventricle. These changes prompt a catecholamine surge and a transient hyperdynamic circulatory phase; however, myocardial energy stores are rapidly depleted, cardiac output declines, and the donor progresses to circulatory arrest (24). The donor heart remains in a warm, distended, and ischemic state during the ethically mandated standoff period (24); therefore, at the time of organ procurement the DCD heart has withstood exposure to a catecholamine surge and sustained significant ischemic injury (Figures 2B,C).

The ischemic injury sustained by the DCD heart results in the depletion of adenosine triphosphate (ATP) stores and anaerobic metabolism, which cause intracellular acidosis, activation of the sodium–hydrogen exchanger (NHE), and sodium influx into the myocyte (30, 31). The sodium-potassium ATPase normally functions to extrude sodium ions entering the myocyte via the NHE. In the DCD context, however, intracellular acidosis develops concurrently with the depletion of ATP stores. The combined effect of increased NHE activity and inhibition of the sodium–potassium ATPase produces a pathological accumulation of intracellular sodium (Figure 4A) (31, 32). Subsequent reperfusion at the time of organ procurement rapidly normalizes the extracellular pH and creates a large hydrogen ion gradient across the plasma membrane that causes further sodium influx via the NHE (33). This increase in intracellular sodium forces the sodium–calcium exchanger (NCX) to function in reverse mode and import calcium ions across the sarcolemma (Figure 4B). The resultant intracellular calcium overload propagates myocyte death through the development of hypercontracture, generation of reactive oxygen species (ROS), activation of the mitochondrial permeability transition (MPT) pore, and initiation of apoptotic pathways (Figure 5) (30, 31, 33). Limiting the severity of ischemia-reperfusion injury (IRI) at the time of organ procurement represents the cornerstone of DCD heart resuscitation.

With the aim of minimizing IRI and optimizing functional recovery, two methods of DCD heart resuscitation have emerged and have been successfully utilized in the clinical context (Table 4) (35). The first approach has been termed Direct Procurement and Perfusion (DPP). This approach involves delivery of a cardioplegic solution during organ procurement that is designed to promote ischemic post-conditioning and limit the detrimental effects of IRI. With this approach, a rapid cardiectomy is performed, the heart is connected to an ex situ heart perfusion (ESHP) device and preserved in a normothermic and beating state until eventual transplantation (Figure 3A) (22, 36). The second approach is Normothermic Regional Perfusion (NRP) (35). Following declaration of circulatory death, a median sternotomy is performed and the cerebral circulation is isolated (a clamp is placed across the arch vessels). With the brain excluded from the circulation, the donor is placed on veno-arterial extracorporeal membrane oxygenation (ECMO) and reperfused for 60 min. The donor is subsequently weaned from ECMO, which facilitates the assessment of donor heart function in vivo. Viable organs are then arrested with a traditional cardioplegic solution, connected to an ESHP device, and preserved in a normothermic, beating state until transplantation (Figure 3B) (23, 35).

At the time of organ procurement, the DCD heart is energy deplete and vulnerable to the influx of sodium and calcium upon reperfusion. The reactivation of ATP production in the calcium-overloaded myocyte after prolonged ischemia can propagate calcium oscillations and the development of hypercontracture (Figure 5) (37, 38). However, initial reperfusion with a cardioplegic solution inhibits myocardial contraction at the onset of re-oxygenation. This facilitates the repletion of myocardial energy stores and the restoration of intracellular calcium homeostasis before activation of the myofibrillar contractile unit, thereby preventing hypercontracture (39, 40). Previous research has demonstrated that hearts subjected to a period of ischemia exhibit better functional recovery when reperfused with a cardioplegic solution compared to reperfusion with unmodified blood (41–44). Therefore, initial reperfusion of the DCD heart should maintain cardiac arrest to provide an opportunity for the restoration of intracellular ion homeostasis and limit hypercontracture.

In 1986, Murry et al. first demonstrated that repetitive periods of brief ischemia could protect the myocardium from a subsequent period of prolonged ischemia and established the concept of ischemic pre-conditioning (45). In 2003, Zhao et al. demonstrated that repeated cycles of ischemia and reperfusion following a prolonged index ischemic insult (ischemic post-conditioning) could attenuate reperfusion-induced myocardial injury to the same extent achieved with ischemic pre-conditioning (46). Subsequent research has demonstrated that common molecular pathways are involved in these two processes (47). These landmark discoveries have tremendous implications for DCD heart transplantation because the delivery of therapeutics to the donor prior to the declaration of death (ischemic pre-conditioning) is ethically prohibited in the majority of jurisdictions. However, the activation of ischemic post-conditioning pathways at the time of organ procurement provides an opportunity to mitigate IRI and resuscitate the DCD heart. The therapeutic window is narrow and optimal success is realized in the first minutes of organ reperfusion (48); therefore, enhancing the composition of the initial reperfusion cardioplegia to activate ischemic post-conditioning pathways and inhibit mediators of IRI is the primary focus of DCD heart resuscitation.

The activation of prosurvival kinases at the time of reperfusion confers powerful cardioprotection against IRI through inhibition of the MPT pore, and activation of the reperfusion injury salvage kinase (RISK) and survivor activating factor enhancement (SAFE) pathways (47, 49, 50). The delivery of erythropoietin, adenosine, and insulin during reperfusion has been shown to activate the RISK pathway and mitigate myocardial IRI (49, 51, 52). Erythropoietin-supplemented cardioplegia has also been shown to activate the SAFE pathway and inhibit mitochondrial permeability pore transition (53–55). Additionally, insulin causes vasodilation, inhibits apoptosis, limits the inflammatory response, and reduces ROS generation (56). Adenosine inhibits apoptosis through upregulation of the antiapoptotic protein Bcl-2 and has anti-inflammatory properties that attenuate neutrophil infiltration into endothelial cells and ROS production (57–63). The nitric oxide donor glyceryl-trinitrate has been shown to activate prosurvival kinases and Bcl-2 (64). Overall, a variety of pharmacologic activators of the RISK and SAFE pathways confer protection against IRI and hold great promise in optimizing DCD heart resuscitation.

The influx of sodium and calcium into the myocyte play a central role in the pathogenesis of IRI. Reperfusion at the time of organ procurement normalizes the extracellular pH, and creates a large hydrogen ion gradient across the plasma membrane that causes sodium influx via the NHE and calcium influx via reverse NCX (Figure 4B). Inhibiting these ionic fluxes early in the reperfusion period represents an important opportunity to limit hypercontracture and activation of the MPT pore (Figure 5) (49).

The cardioprotective effects of NHE inhibition in IRI are well documented. Cariporide has been shown to limit myocardial IRI in clinical trials (65–67) and translational animal models of DCD heart transplantation (32, 68–70). Zoniporide is a recently developed NHE inhibitor that possesses greater potency and selectivity toward the cardiac NHE (71), and has potent cardioprotective effects through activation of prosurvival kinase pathways and inhibition of apoptosis (72). Unfortunately, NHE inhibitors are no longer available for clinical use and enthusiasm for further development has been limited by the results of the Expedition trial (67). However, drug delivery in the context of DCD heart resuscitation is isolated to the heart in a deceased donor and the potential detrimental impact on other organ systems is irrelevant. Therefore, NHE inhibitors may still have an important role in DCD hearts resuscitation. Alternatively, delaying pH normalization at the onset of reperfusion by delivering an acidic cardioplegic solution inhibits NHE, and limits calcium overload and IRI (Figures 4 and 5) (73–75). We have recently shown that initial reperfusion with a moderately acidic solution optimizes the functional recovery of DCD hearts (34). Therefore, modifying the cardioplegic solution composition may provide an avenue to realize the clinical benefit of NHE inhibition without relying on pharmacologic inhibitors that are not currently available for clinical use.

Pharmacologic inhibitors of reverse-mode NCX have also been investigated as a means of minimizing calcium influx into the myocyte during reperfusion (Figure 4). NCX inhibitors minimize calcium overload, hypercontracture, infarct size, and contractile dysfunction following ischemia in animal models (76–80), and may be more cardioprotective than NHE inhibitors (81). While there are no clinically approved NCX inhibitors at present, calcium influx during reperfusion can be limited if the initial reperfusion solution is rendered hypocalcemic. This serves to minimize the calcium gradient that favors reverse NCX activity during early reperfusion, thereby limiting calcium overload (82–86). We have recently demonstrated that initial hypocalcemic reperfusion optimized the functional recovery of DCD hearts (34). Therefore, utilizing a hypocalcemic cardioplegic solution may provide a simple avenue achieve NCX inhibition and minimize IRI.

Conventional cardioplegic solutions rely on hyperkalemia to depolarize the membrane potential and achieve diastolic arrest; however, membrane depolarization is associated with an increase in intracellular sodium via non-activating sodium currents that may exacerbate calcium overload during reperfusion (72). Alternatively, normokalemic adenosine-lidocaine cardioplegic solutions have been proposed, in which lidocaine blockade of sodium fast channels causes a diastolic arrest and adenosine maintains a polarized membrane potential (87–89). Ischemic rat hearts reperfused with an adenosine-lidocaine cardioplegia exhibit improved myocardial function compared to traditional hyperkalemic cardioplegia (88). Similarly, polarized arrests using potassium channel openers minimize calcium overload and improve myocardial function (90–92). Mohri et al. have applied this concept in a large animal model of DCD heart transplantation and demonstrated improved posttransplant outcomes in hearts treated with a potassium channel opener (93). Further, we have utilized normokalemic adenosine-lidocaine based solutions in our translational models of DCD heart transplantation (34, 94, 95). Further research is required to determine the role of such alternative cardioplegic solutions in clinical DCD heart transplantation.

A wide variety of strategies have demonstrated efficacy in mitigating IRI; however, it is likely that the delivery of a cardioplegic solution containing a cocktail of complementary pharmacologic post-conditioning agents that target a variety of pathways involved in the pathogenesis of IRI will further improve DCD heart resuscitation. For example, the synergistic beneficial effects of glyceryl-trinitrate, erythropoietin, and zoniporide have been demonstrated in animal models of DCD transplantation (96, 97). We have utilized a cardioplegic solution containing the ischemic post-conditioning agents, adenosine and insulin, and rendered the solution acidic and hypocalcemic to minimize calcium overload during reperfusion (34). ROS scavengers have been shown to limit IRI in DCD hearts (98); however, ROS signaling in the early reperfusion period is essential for the activation of post-conditioning pathways and the role of ROS scavengers in DCD heart resuscitation requires further research (99–101). Calpain inhibitors and MPT pore inhibitors may represent other pharmacologic strategies to further optimize DCD heart resuscitation (102, 103). Further research in this area will undoubtedly contribute to improved posttransplant outcomes in the future.

Optimizing the conditions of the initial reperfusion may further improve DCD heart resuscitation. Initial cardioplegic reperfusion provides an opportunity to restore myocardial energy stores and intracellular ion homeostasis prior to myocardial contraction; however, hypothermia markedly lowers the activity of the ion pumps that restore intracellular ionic homeostasis (104–106). The delivery of hypothermic cardioplegia at the time of organ procurement is standard practice when procuring hearts from DBD donors, to minimize metabolic demands during the subsequent period of cold-static storage (Figure 2A). In the context of DCD; however, the benefit of inducing a profoundly hypothermic state in the short period of time between procurement and initiation of normothermic ESHP may be outweighed by the negative impact of hypothermia on the reparative processes essential for resuscitation of the DCD heart (95).

Previous work has demonstrated that delivery of warm cardioplegia following ischemia improves functional recovery (44, 107–109). These results have been confirmed in clinical trials of patients undergoing cardiac surgery, which demonstrated that terminal warm cardioplegia limits myocardial injury and improves function postoperatively (110, 111). Since DCD hearts have sustained significant ischemic injury, application of the terminal warm cardioplegia concept to the initial reperfusion of DCD hearts may optimize their resuscitation (41). Translational experiments in animal models of DCD have advocated for the avoidance of profound hypothermia during the initial reperfusion (41, 94, 98, 112–114), and we recently confirmed that the avoidance of profound hypothermia during initial reperfusion minimizes injury and improves the functional recovery (95).

Interventions to limit IRI must be administered at the onset of reperfusion to be effective; however, the duration of the cardioplegic reperfusion is another variable that impacts DCD heart resuscitation. The initial reperfusion must be of sufficient duration to facilitate repletion of myocardial energy stores, restore ionic homeostasis, and activate ischemic post-conditioning pathways. Cohen et al. have demonstrated that initial reperfusion with an acidic solution for 1 min was not protective; however, when the reperfusion was extended to 2 min it afforded protection against IRI equivalent to that achieved with post-conditioning protocols (115). In contrast, studies have demonstrated that initial reperfusion with an inhibitor of GSK-3β must be extended over 15 min to be effective (116) and initial reperfusion with adenosine required an infusion extended over 40 min to achieve a post-conditioning effect (117). Osaki et al. found that posttransplant outcomes were optimized when initial reperfusion of DCD hearts with a blood-based cardioplegia at 20°C was continued for 20 min (39). These results suggest that the optimal initial reperfusion duration may depend on the composition of the reperfusion solution, the conditions of its delivery, and the means by which activation of the ischemic post-conditioning pathways occur.

Early investigators sought develop techniques for the resuscitation, preservation, and transplantation of hearts from donors that had suffered an anoxic arrest, in an era when little was known about IRI, immunosuppression, or extracorporeal perfusion (168–171). Christiaan Barnard confirmed the clinical impact of such research, by resuscitating hearts from human DCD donors on cardiopulmonary bypass and subsequently transplanting them (11, 12). However, enthusiasm for research regarding DCD heart resuscitation and transplantation waned over the next 20 years, following the acceptance DBD transplantation (7–9). Heart transplantation soon became limited by a shortage of suitable donor organs and the number of transplants performed annually plateaued (1), which prompted a renewed interest DCD. A number of investigators subsequently published studies describing the successful transplantation of hearts subjected to 17–60 min of warm ischemia following donor exsanguination (70, 172–179). In 2009, however, it was demonstrated that an exsanguination model of donor warm ischemia significantly reduced the severity of myocardial injury sustained by the heart compared to the more clinically relevant model of donor extubation (112). In this context, the clinical translation of studies that employed an exsanguination model of donor warm ischemia is difficult.

Gundry et al. utilized a more clinically relevant model of hypoxemic cardiac arrest in lambs and demonstrated that donor pretreatment with dextrose, methylprednisolone, prostaglandin E₁, and nifedipine could facilitate successful transplantation following 40 min of warm ischemia (180). They employed the same protocol in baboons and reported posttransplant survival ranging from 1 to 34 days (181). Donor pretreatment with an endothelin-A receptor antagonist and an ATP-sensitive potassium channel opener have also been shown to facilitate successful DCD transplantation in dogs (93, 182). While these studies provided experimental evidence that hearts sustaining significant periods of warm ischemia following donor extubation could be transplanted, donor pretreatment is ethically prohibited in most jurisdictions.

In 2006, Osaki et al. utilized a clinically relevant model of DCD (donor extubation and no pretreatment) and demonstrated that controlled reperfusion with a tepid blood cardioplegia could facilitate successful transplantation in pigs (39). To minimize donor heart exposure to additional ischemia following procurement, the hearts were continuously perfused in a beating state during preservation and implantation. The same group demonstrated that the addition of a ROS scavenger to the initial reperfusion solution minimized lipid peroxidation and improved posttransplant function (98). Iyer et al. reported successful transplantation of DCD pig hearts exposed to 30-min of warm ischemia following donor extubation (140). These hearts were resuscitated with Celsior solution supplemented with erythropoietin, glyceryl-trinitrate, and zoniporide, and then preserved on the Transmedics OCS. We have also demonstrated the importance of a cardioprotective resuscitation strategy and ESHP to minimize incremental ischemic injury during preservation and transplantation (94). Finally, Ali et al. demonstrated that DCD hearts resuscitated using NRP could be successfully transplanted, with outcomes comparable to hearts from DBD donors (183). These translational studies provided evidence to suggest that DCD heart transplantation could be successfully performed when an approach to donor heart resuscitation, preservation, and transplantation is tailored to the DCD context (Figures 2C and 3).

The first report of clinical DCD heart transplantation in the modern era was published by Boucek et al., describing three pediatric transplants performed between 2004 and 2007 at Denver Children’s Hospital (184). Each donor had suffered birth asphyxia and a decision to WLST was made based on the futility of ongoing care. The donor was monitored for progression to cardiorespiratory arrest and mechanical asystole. A 3-min standoff period was observed in the first donor; however, this was shorted to 1.25 min in the subsequent donors based on the recommendations of the ethics committee. The mean time from extubation to declaration of death was 18 ± 8 min. Cold preservation fluid was then infused though a balloon catheter placed in the ascending aorta and venous exsanguination was undertaken. Posttransplant inotropic support, rejection episodes, ventricular function, and 6-month survival were comparable to control infants who underwent DBD heart transplantation during the same period. This report prompted vigorous ethical debate in the transplant community regarding the definition of circulatory death, whether cardiac transplantation from DCD donors violates the “dead donor rule,” and the minimum standoff period duration that must be observed before death can be declared (185–187).

Ali et al. reported on a DCD donor with a 23-min WIT that was resuscitated using normothermic cardiopulmonary bypass, following exclusion of the cerebral circulation (188). After 5 min of reperfusion the heart spontaneously reverted to sinus rhythm. The patient was re-intubated and 190 min later was weaned from cardiopulmonary bypass. Following insertion of a pulmonary artery catheter, a CI of 2.4 L/min/m² was measured with a pulmonary artery capillary wedge pressure of 13 mmHg. The authors suggested that hearts from Maastricht category III donors might be suitable for use in clinical transplantation. This report again prompted ethical debate surrounding the postmortem restoration of mechanical cardiac function using cardiopulmonary bypass in a donor that had been declared dead based on cardiorespiratory criteria (189). However, the technique of NRP has become an accepted means of facilitating multi-organ DCD retrievals in some countries (190).

Iyer et al. utilized the DPP approach to DCD heart resuscitation in a donor that had suffered a 32-min WIT (191). The heart was reperfused with a resuscitative cardioplegia and then transferred to a modified Transmedics OCS device for ESHP and functional assessment in a physiologic working mode. The heart produced a cardiac output of >5 L/min with a left atrial pressure of 14 mmHg. Rosenfeldt et al. also reported on a DCD donor heart that was reperfused with a resuscitative cardioplegia after 32 min of warm ischemia, and preserved using hypothermic ESHP during transport (160). The heart subsequently underwent ESHP and functional evaluation over a 12-h period. Messer et al. utilized the NRP approach in three CD donors with a mean time WIT of 28 min, and demonstrated that in vivo measurements of myocardial function correlated with assessments conducted ex situ on a modified Transmedics OCS device (192). Finally, Osaki et al. compared the myocardial function of four DCD hearts with five DBD hearts that had been declined for clinical transplantation due to the presence of coronary artery disease, advanced age, and donor social history (193). The DCD hearts sustained a WIT of 34 ± 3 min, were reperfused with a standard cardioplegic solution, and then underwent cold-static storage for 152 ± 55 min before evaluation in an ESHP device. The DBD hearts were arrested with the same cardioplegic solution and underwent cold-static storage for 211 ± 31 min before ex situ evaluation. Despite the additional cold ischemic insult sustained by the DCD hearts, the recovery of myocardial function was not significantly different compared to the DBD hearts. These reports provided early clinical evidence that hearts from DCD donors could be successfully resuscitated, and formed a foundation on which clinical transplant programs could be developed.

Dhital et al. reported the first clinical adult DCD heart transplants in the modern era (22). These three transplants represented direct clinical translation of the groups’ research in pharmacologic post-conditioning in large animal models (32, 53, 64, 72, 96, 97, 140). Based on these studies, the authors considered Maastricht category III donors <40 years of age with a WIT <30 min (97). WLST occurred in the intensive care unit, an anesthetic bay, or an adjacent operating room, with standoff periods ranging from 2 to 5 min depending on the jurisdiction of donation within Australia. The WITs ranged from 22 to 28 min, with FWITs between 11 and 21 min (Table 2). Donor hearts were resuscitated using the DPP approach and St. Thomas’ cardioplegia supplemented with erythropoietin and glyceryl-trinitrate (97). Unfortunately, zoniporide is not approved for clinical use and could not be included in the cardioplegia. Donor hearts were preserved on the Transmedics OCS for 245–260 min prior to transplantation. A perfusate lactate concentration <5 mmol/L and evidence of myocardial lactate extraction were used as evidence of myocardial viability. Two recipients required temporary mechanical circulatory support posttransplant; however, all patients were weaned from support and discharged from hospital after 21–28 days. Two patients experienced a rejection episode, but all patients demonstrated normal biventricular function at follow-up (Table 3). A fourth donor failed to progress in the pre-determined time frame and demonstrated a rising lactate concentration during preservation on the OCS; therefore, this heart was not considered for transplantation. This report represented the first clinical evidence that hearts from DCD donors could be resuscitated, preserved using ESHP while being transported from a distant site, and transplanted.

García-Saez et al. described two adult clinical DCD heart transplants into recipients who were bridged to transplant using a durable left-ventricular assist device (194). The authors considered Maastricht category III donors <50 years of age. The WITs ranged from 14 to 36 min, with FWITs between 13 and 21 min (Table 2). Donor hearts were resuscitated using the DPP approach and Custodial cardioplegia supplemented with erythropoietin and glyceryl-trinitrate. Donor hearts were preserved on the Transmedics OCS for 307–360 min prior to transplantation. Neither recipient required mechanical circulatory support postoperatively. At hospital discharge both recipients displayed normal left-ventricular function and moderate right ventricular dysfunction (Table 3). This report demonstrated that hearts from DCD donors could be successfully transplanted into recipients bridged to transplant with an implantable left-ventricular assist device.

Finally, Messer et al. described nine adult clinical DCD heart transplants performed using the NRP approach (23), clinical translation of previous research in large animal models (183). The authors considered Maastricht category III donors <50 years of age; however, donors with WITs extending beyond 30 min were considered. The WITs ranged from 17 to 146 min, with FWITs of 12–25 min (Table 3). Hearts were resuscitated using veno-arterial ECMO and weaned from support after 27–190 min. A CI ≥ 2.5 L/min/m², with a CVP ≤ 12 mmHg, PCWP ≤ 12 mmHg, and an EF ≥ 50% were used as evidence of viability. Donor hearts were then preserved on the Transmedics OCS for 0–428 min prior to transplantation (one local donor heart underwent a brief period of cold-static storage prior to transplant and was not preserved on the OCS). Hearts preserved on the OCS underwent continuous perfusion during transplantation to further minimize exposure to ischemic injury.

Two recipients required mechanical circulatory support posttransplant; however, all patients were weaned from support and discharged from hospital after 15–80 days (Table 3). Interestingly, three of the hearts displayed lactate profiles during preservation on the OCS that suggested non-viability (22, 137); however, these hearts were deemed viable based on in vivo functional assessments and transplanted (one heart required intra-aortic balloon pump support for 1 day). Conversely, the heart that required veno-arterial ECMO following transplant displayed an acceptable lactate profile during preservation on the OCS. These results highlight the need for further research regarding assessments of organ viability prior to transplant. Importantly, this study also demonstrates that DCD hearts with extended WITs may still be considered for transplantation, provided the FWIT remains <30 min and acceptable myocardial function is demonstrated. Finally, the authors utilized continuous myocardial perfusion during transplantation. In the PROCEED II trial, donor hearts were exposed to 83 min of additional cold ischemia during implantation (137). Therefore, utilizing continuous perfusion during implantation may significantly improve posttransplant outcomes.

These recent clinical reports have demonstrated encouraging short-term outcomes following DCD heart transplantation; however, the impact of the warm ischemic insult sustained following WLST on long-term outcomes is unknown. Ischemic injury occurring during DBD heart transplantation has been linked to the development of graft vasculopathy and myocardial fibrosis (195–197), and further research is required to determine if similar sequelae will be observed with DCD hearts.