Persufflation (or Gaseous Oxygen Perfusion) as a Method of Organ Preservation

A. Heart

Although the earliest studies of PSF were focused on the heart, research in heart PSF had subsided for about three decades (1960s-1990s) in favor of research in liver and kidney PSF. More recently, cardiac PSF has been rekindled and several studies have been published in which PSF was used prior to transplant, including the use of PSF to preserve hearts having suffered short periods of warm ischemia. Collectively, these studies have at least established that cardiac PSF is technically possible and that it can preserve heart tissue.

The advent of cardiopulmonary bypass (CPB) in the mid-1950s provided impetus for exploring the use of PSF. In 1959, Sabiston et al at Johns Hopkins explored the use of PSF in conjunction with CPB [76]. In the first set of experiments, hearts from medium-sized dogs were cannulated at the coronary ostia, flushed with pre-oxygenated normal saline, then persufflated with humidified gaseous carbogen (a mixture of 95% O₂ and 5% CO₂). This approach, generally termed anterograde PSF (A-PSF), contrasts with retrograde PSF (R-PSF) which would be tried in the heart [8; 87] and subsequently in the kidney [28; 32]. Hearts maintained at 37°C continued to beat for an average duration of 5.1 hours (2.5–8 hours, range). Cardiac contractility remained strong for the first 2–3 hours and then gradually weakened. In some cases, the electrical activity of the heart continued for periods up to 4 hours following cessation of a visible heartbeat. The second set of experiments examined in situ A-PSF of the heart. Once the heart had been isolated from the systemic circulation, A-PSF was performed for 25–30 minutes. A normal hemodynamic response was restored in 9 of 12 animals and some animals maintained a heartbeat for 48 hours following the reestablishment of native coronary circulation. This study established the use of PSF in cardiac surgery by showing that oxygen gas can be used by the heart and that coronary blood flow may be reestablished after PSF. In 1960, a follow-up study introduced the concept of R-PSF [87]. At the time, retrograde perfusion of oxygenated blood via the coronary sinus was used to maintain a heartbeat and protect the heart from anoxia for short periods of time during open aortic valve procedures [22; 43]. Talbert et al used this knowledge, along with their previous work on A-PSF, to investigate whether R-PSF could be performed successfully in the heart. In their experiments, the coronary sinus was cannulated in 7 canine hearts, flushed with normal saline and started on R-PSF. These organs maintained a beat for an average duration of 3.5 hours (2–4 hours, range). In 3 separate hearts, the anterior cardiac veins were additionally cannulated and persufflated. These organs maintained a visible beat for an average duration of 5.5 hours and up to 7 hours. They noted that cardiac activity remained strong over the first 2 hours of experimental conditions and then gradually became weaker until ventricular fibrillation or complete asystole had occurred. Talbert compared the R-PSF with A-PSF and determined that the heartbeat was visibly weaker and sustained for a shorter period of time using the retrograde approach. Nevertheless, they concluded that oxygen gas can be delivered retrogradely and that the method exhibits some efficacy.

In 1966, the Talbert et al concept of R-PSF was applied during CPB by Camishion et al [8]. They noted that continuous blood perfusion during open aortic valve procedures was cumbersome due to the obstruction of the surgical field caused by cannulation of the coronary vessels from within the aorta. Consequently, they tried to determine whether animals could survive CPB by using R-PSF of the coronary sinus as the main preservation technique. This was investigated by repeating previous work by Talbert et al on in situ R-PSF using a similar canine model [87]. They reported that each of 10 canine hearts maintained a sinus rhythm for at least 31 minutes while being retrogradely persufflated. Following the loss of a sinus rhythm, 8 animals maintained a nodal or ventricular rhythm for up to 7 hours and 2 animals developed and sustained ventricular fibrillation for up to 6 hours. When hearts were persufflated with nitrogen gas, sinus rhythm was maintained for 5 minutes or less and a visible beat was lost in all 10 animals within an average duration of 11 minutes and no longer than 25 minutes. When persufflation was switched to oxygen gas, a 30-fold increase in the tissue partial oxygen tension of the hearts was observed almost immediately. In 2 animals, asystole was converted to a persistent ventricular rhythm. In a second experimental study, the coronary sinus in 20 porcine and 10 canine hearts were cannulated and animals were placed on CPB using R-PSF for 1 hour. During CPB, 25 of 30 animals maintained sinus rhythm for the entire hour. Of the remaining 5 animals, all developed ventricular fibrillation after an average of 30 minutes and one spontaneously reverted to a nodal rhythm after 20 minutes of sustained ventricular fibrillation. Following removal of CPB, 22 of 30 animals remained in sinus rhythm. Four animals with ventricular fibrillation were converted electrically to sinus rhythm, 3 animals developed fibrillation after reperfusion, all of which could be converted electrically to sinus rhythm. The remaining animal exhibiting nodal rhythm converted to normal sinus rhythm spontaneously following cessation of CPB. After re-establishment of native coronary blood flow, mean aortic blood pressure was maintained between 60–120 mmHg and central venous pressures remained 4.4–14.7 mmHg in all animals. Of the 30 experimental animals in the second group, only one exhibited signs of heart failure postoperatively. This animal developed severe pulmonary edema following transfusion of 2500 mL of blood for ongoing hemorrhage. From the vantage point of a contemporary understanding of shock and transfusion medicine, it is conceivable that this animal may have developed a variant of acute respiratory distress syndrome or transfusion-related acute lung injury, which may have been misinterpreted as pulmonary edema from congestive heart failure – though the true pathology will never be known. Nevertheless, these studies illustrated that a heart could be preserved by PSF during CPB and that these organs recovered their function following reperfusion. No evidence of air embolization in the brain or viscera of any experimental animal. The authors commented on the fundamental difference between gas embolization and PSF; gas emboli are small bubbles that may occlude smaller vessels, whereas PSF is characterized by the free flow of a gas within the system. This distinction is still not fully appreciated by the clinical community and services to highlight that this will need to continue to be proven experimentally. Camishion et al also raised the possibility of using this preservation technique for the maintenance of donor hearts in cardiac transplant, even before the first successful heart transplant was performed in South Africa a few years later [3].

Also in 1966, Gabel et al examined the physiology of the persufflated heart by using juvenile feline hearts that were anterogradely persufflated with gaseous carbogen mixture (95% O₂, 5% CO₂) via cannulae secured in the proximal aorta and compared with controls perfused with substrate-free Krebs solution [21]. They found that the heart rate in persufflated hearts declined rapidly over the first hour and then declined more slowly over the next 9 hours, whereas the heart rate in the liquid-perfused heart exhibited a steady decline over the entire experimental period. Contractility measurements under A-PSF showed an initial rise in contractile force during the first 20 minutes and subsequent fall after 4 hours. The persufflated hearts reached 50% of the initial contractile force after 7 hours, whereas liquid-perfused hearts declined to 50% of initial contractile force in only 80 minutes. Metabolic studies revealed that glycogen, carbohydrate, lactate, and pyruvate levels decreased rapidly in the persufflated heart, but when these hearts were treated with pharmacologic agents they responded as expected. In addition, they found that rhythm changes in the A-PSF model reproducibly occurred above a certain threshold PSF pressure. Gabel et al concluded that persufflated hearts exhibited stronger contractile forces, performed more work and were slower to fail than the hearts perfused with liquid. They theorized that oxygen gas allowed an increase in cardiac work, even though oxygen supply is not traditionally considered a major determinant of work capacity. It may be that cardiac oxygen extraction is altered when oxygen is delivered by PSF or that simply more oxygen is delivered using PSF. Another hypothesis emerging from these findings was that, in the case of PSF, active metabolites equilibrate solely between the extracellular fluid and the intracellular space, as opposed to being flushed away by a liquid perfusate.

Lochner et al subsequently studied the metabolism and function of anterogradely persufflated guinea pig and rat hearts [40]. Hearts were persufflated with carbogen gas mixture (95% O₂, 5% CO₂) at 37°C for 1 hour. With persufflated hearts, the peak systolic pressures and the first derivative of the left ventricular systolic pressure decreased, while exhibiting very little change in the heart rate. This seemed to indicate that the persufflated heart could continue to generate hemodynamic work. In a few of the hearts, A-PSF time was extended to 2 hours, resulting in no additional decreases in heart rate, left ventricular systolic pressure and its first temporal derivative, or isovolumetric work. Isovolumetric work and the first derivative of left ventricular systolic pressure following PSF were calculated to be 16.4% and 18.4% of values characteristic for liquid-perfused hearts, respectively. Measurements of tissue creatinine phosphate and ATP were similar between the PSF and liquid perfusion groups. This led the authors to suggest that the decrease in cardiac work was likely not due to a lack of available cellular energy. They also discovered that work capacity could be enhanced by increasing the PSF pressure or the diastolic filling pressure.

Following these early studies, there was a gap of several decades during which no work was published on the gaseous perfusion of hearts. It was not until the late 1990s that interest in cardiac PSF rekindled, largely as a result of the successful application of PSF in other organs, in particular the kidney and liver. What had been previously referred to as ‘gaseous oxygen perfusion’ was eventually termed ‘persufflation’ by Denecke [10]. After pursuing extensive work with kidneys and livers, Fischer’s group in Cologne explored cold preservation via cardiac PSF in 1998 [35]. Porcine hearts were flushed and stored at 0–1°C using three different methods: 1) SCS using modified Euro-flush solution with glutathione (based on Euro-Collins solution); 2) SCS with University of Wisconsin (UW) solution; and 3) A-PSF via the ascending aorta in combination with SCS using Histidine-Tryptophan-Ketoglutarate solution modified by adding hyaluronidase. The overall mean preservation time was 14.5 hours. Hearts were orthotopically transplanted into recipient pigs of comparable body weight using standard CPB, reperfused on CPB using whole blood for an average of 154 minutes before being weaned off CPB to allow the hearts to take over normal circulation. Following transplantation, hemodynamic parameters were measured to estimate cardiac function and serum creatine kinase values were obtained as an indicator of myocardial damage. Prior to sacrificing the porcine recipient, left ventricular biopsies were taken to estimate myocardial water content and ATP levels. Persufflated hearts exhibited stroke work similar to preoperative values, whereas comparable measurements could not be obtained in static cold-stored organs due to severe arrhythmia and ventricular dyskinesia. Measurements of cardiac output, left ventricular systolic pressure and its first temporal derivative revealed that the persufflated hearts fared significantly better than hearts stored in modified Euro-flush solution alone and had better cardiac output than hearts stored in UW solution alone. Equivalent creatine kinase levels between each group indicated that the degree of cellular damage using A-PSF may have been similar to conventional SCS. However, persufflated hearts exhibited significantly higher ATP levels than UW/Euro-flush solution-stored hearts. Collectively, these data seem to indicate that A-PSF may permit superior recovery of post-transplant heart function compared with SCS using either UW or modified Euro-flush solution. Importantly, myocardial water content measurements indicated that there was significantly less myocardial edema with A-PSF than static UW solution alone. Decreased myocardial edema is a distinct benefit of PSF, as it is known that tissue edema can significantly impair cardiac function [37]. A follow-up study examined the effects of A-PSF on myocardial tissue quality and post-transplant cardiac function by comparing against SCS in Histidine-Tryptophan-Ketoglutarate solution with and without hyaluronidase [36]. The cohort of hearts preserved by coronary A-PSF showed significantly higher left ventricular systolic pressure and its first temporal derivative, and cardiac output compared to hearts preserved by SCS with Histidine-Tryptophan-Ketoglutarate solution, but not when modified with hyaluronidase. Persufflated hearts maintained normal circulatory function for longer when compared to either SCS methodology. Additionally, tissue ATP levels were significantly higher in transplanted hearts following A-PSF than after SCS only. Post-transplant myocardial water content was not elevated in persufflated hearts over controls.

Up to this point, most of the research into cardiac PSF had thus far involved experimental operations with hearts experiencing no “down-time” or conventional warm ischemia as experienced with donation after cardiac death (DCD). The opportunity to resuscitate DCD hearts inspired the group in Cologne to study PSF following warm ischemic damage. Thus, Yotsumoto et al studied the effects of several hypothermic preservation techniques on post-transplant cardiac function following a mean warm ischemia time (WIT) of 16.7 minutes in a porcine autotransplant model [102]. Three hours of A-PSF was compared with SCS, with an additional set of controls not damaged by warm ischemia and also stored in Histidine-Tryptophan-Ketoglutarate solution modified with hyaluronidase. As with previous studies examining the effects of preservation on heart function, a number of physiologic parameters were recorded and samples were taken to assess the metabolic recovery of the cardiac tissue. It was reported that control and persufflated hearts were completely weaned from CPB within 2 hours of transplantation, whereas the static cold-stored hearts exhibited significantly diminished functional recovery. Near the end of the 3-hour reperfusion period – cardiac output, left ventricular contractility, and the relaxation velocity were significantly higher in the A-PSF group as compared to SCS. It appeared that persufflated DCD hearts had functional outcomes similar to hearts procured from heart-beating donors using conventional storage methods. Importantly, Troponin T levels were significantly higher under SCS than for undamaged controls and hearts preserved by A-PSF at 1 hour after reperfusion. These data indicate that A-PSF may limit myocardial injury incurred during WIT. The authors noted that the transplant field is reluctant to adopt PSF as a legitimate cardiac preservation technique largely due to concerns about resulting endothelial damage. More recent studies have shown that the coronary arteries of porcine hearts following 3 hours of oxygen A-PSF had normal functioning endothelium post-transplant [14; 18; 34]. Additionally, hearts transplanted following 14 hours of A-PSF exhibited no topographic signs of endothelial damage, as assessed by scanning electron microscopy [34]. Fischer has recently reviewed work done by his group, and has described the detailed experimental approach in which A-PSF is recommended as the preferred method [16]. Collectively, these works have shown that cardiac PSF has considerable potential as an emerging preservation technique, and Table 2 summarizes the published work on heart PSF presented in this review.

B. Kidney

The initial studies with kidney PSF occurred in the 1960s, shortly after the early development of heart PSF. It was in the kidney that PSF has been most extensively evaluated, likely because the vascular anatomy and associated transplant models are considered to be most straightforward (amongst the major transplantable organs). Following an initial study by Talbert et al at Johns Hopkins in the 1960s [88], most all work on kidney PSF was performed by Fischer, Isselhard and others in Cologne, Germany. Early research efforts were comprehensive in developing the technical aspects of PSF (including optimization of flow pressures, partial oxygen pressures, temperature, and type of approach – whether anterograde or retrograde – used during kidney PSF) by evaluating their effects on the bioenergetic status and function post-reperfusion. The groundwork produced by the researchers in Cologne stimulated interest in the field and by the 1980s a number of other institutions had initiated studies to explore the value of kidney PSF.

The initial study by Talbert et al involved in situ PSF of 7 canine kidneys [88] and showed that A-PSF with gaseous carbogen mixture (95% O₂ and 5% CO₂) could be used to preserve kidney function for 4 hours. A-PSF was performed by feeding a catheter through the left iliac artery and positioning it at the renal artery. Once the catheter was appropriately positioned and the proximal renal artery around the catheter was sealed, the left renal vein was clamped distally and the left gonadal vein was divided and used for drainage. Once the blood was flushed using normal saline, A-PSF was started at a pressure of 120–150 mmHg (to expel the liquid perfusate) and gradually decreased to 80–100 mmHg. The left kidney was persufflated for 2–4 hours, flushed with normal saline until no visual evidence of gas appeared in the venous effluent and then renal blood flow was re-established. This was performed by pulling back on the renal arterial catheter, removing the renal venous clamp and ligating the proximal stump of the left gonadal vein. The animals were then monitored for up to a year. The study included two sets of controls. In the first set, 4 dogs had the left renal artery isolated and clamped for 2 hours, while in the second set of controls the left renal artery was cannulated, flushed and the renal circulation was reestablished after 2 hours of warm ischemia. Compared to controls, persufflated kidneys functioned for some time after the treatment. Renal function was determined primarily through intravenous pyelography and also by assessing left kidney function following contralateral nephrectomy. Furthermore, histologically, most of the persufflated kidneys exhibited some signs of tubular atrophy and scarring, but these findings were considered minimal in comparison with ischemic controls. The authors concluded that PSF has the potential to prevent the harmful consequences of warm ischemia and that the afforded protective effects are likely a result of tissue being able to utilize and survive by consumption of gaseous oxygen. They noted that simply clearing the renal vasculature of blood (to prevent coagulation during the ischemic period) was not enough to preserve organ function. These findings are highly significant in that PSF appeared to keep kidneys alive during 2 hours of WIT.

It would be 10 years before these encouraging observations reported by Talbert et al were pursued further by others. In 1971, Denecke et al developed an in situ canine renal ischemia model [10], similar to the one developed by Talbert et al [88]. This study involved a comparison between hypothermic A-PSF at 100 mmHg and SCS. Kidneys undergoing either treatment were initially flushed clear of blood by perfusion with an unspecified crystalloid solution. Following 4 hours of A-PSF or SCS, contralateral nephrectomy was performed and circulation to the remaining experimental kidney was re-established. It was reported that A-PSF was actually more harmful to the kidneys than SCS alone; 4 of 5 dogs had died within 7 days, while the remaining dog survived but exhibited marked uremia. Of the 3 dogs having their kidneys preserved by SCS, all survived. Additionally, it was determined that persufflated kidneys had difficulty maintaining normal blood flow following the treatment, with perfusion having decreased to roughly one-third of normal. Moreover, despite an increase in tissue levels of ATP during A-PSF, the ATP levels quickly diminished following reperfusion. The authors postulated that as a result of enhanced oxygenation, the renal vasculature had responded reflexively by increasing the resistance to flow, thereby decreasing global reperfusion of the kidney. As far as we are aware, further evidence has not been provided to substantiate this claim. In our opinion, the physiological response invoked to explain these observations does not seem tenable under hypothermic conditions. It is more likely that the decrease in perfusion may have been related to vascular damage caused by hyperoxia and elevated PSF pressures. On a historical note, this was the first time that the term ‘persufflation’ was substituted for gaseous oxygen perfusion.

Follow-up studies resulting from this original report are important for addressing the largely unexpected outcome that A-PSF had a distinctly detrimental outcome as compared with R-PSF. Isselhard and his collaborators published a series of studies that examined the differences between A-PSF and R-PSF [28; 29; 30; 31; 32; 77], which is PSF by delivering the gas in the direction opposite to physiologic flow (starting at the venous end). Historically, the technique of R-PSF also involved the introduction of small, pin-pricks into the surface of the organ – to facilitate gas efflux as illustrated in Figure 1. In these studies, the effects of SCS, A-PSF and R-PSF on the bioenergetic profile of canine kidneys throughout preservation and after reperfusion were explored using their established in situ model. The degradation rate of high-energy phosphates at 37°C, 26°C and 6°C in canine kidneys was studied to better understand the effect of hypothermia on ATP, ADP and AMP levels. Furthermore, they measured the levels of high-energy phosphates and lactate in kidneys undergoing A-PSF and R-PSF using pure gaseous oxygen (100% O₂), 40% oxygen gas (mixed with 55% N₂ and 5% CO₂), and room air (21% O₂) to also study the effects of delivered oxygen concentration. A-PSF was performed at 60 or 100 mmHg and R-PSF at 30 or 60 mmHg. To study the impact of preservation protocol on metabolic status, renal cortical biopsies were taken at various time-points before and during preservation and after blood flow had been reestablished. ATP depletion rate dropped by a factor of 2 and nearly 10 for kidneys preserved at 26°C and 6°C, respectively, as compared with measurements at 37°C. These findings confirmed that hypothermia diminishes the pace of energy utilization during storage. In the same study, Isselhard et al were able to illustrate that the operational pressures of both A-PSF and R-PSF needed optimization for the best outcomes. ATP levels during R-PSF at 26°C and for 8 hours were strongly dependent on the driving pressures, averaging 81% and 98% of control values at 30 and 60 mmHg. They also reported that R-PSF was generally better at the lower pressure (30 mmHg) than A-PSF at either 60 or 100 mmHg, based on these metabolic assays. It also appeared that lowering the PSF pressure from 100 to 60 mmHg during A-PSF had a stronger, negative effect on ATP metabolism than lowering the PSF pressure from 60 to 30 mmHg during R-PSF. The authors specifically stated that pressures below 60 mmHg were unable to sustain adequate gas flow during A-PSF and they also pointed out that the rate at which ATP was degraded increased inversely with gaseous oxygen concentration. Not surprisingly, lactate levels rose as the oxygen concentration decreased (from 100% to 40% and 21%) in both A-PSF (at 60 mmHg) and R-PSF (at 30 mmHg), but more dramatically during A-PSF. A remarkable accomplishment was the demonstration that ATP levels were maintained at 40% and 30% of control values under A-PSF (at 60 mmHg, 66 hours) and R-PSF (at 30 mmHg, 72 hours) at 6°C, respectively. In contrast, during SCS – ATP levels were reduced to negligible levels within minutes. These investigators noted that despite an ability to maintain a healthy bioenergetic status in preserved kidneys by PSF, the energy disparity (between utilization and production) remains during hypothermic storage and is only slowed down.

Further studies by this group remained focused on the important differences between A-PSF and R-PSF, by assessing their effects on renal bioenergetic status and function after reperfusion [29; 32]. In situ A-PSF (at 90–100 mmHg) was performed on canine kidneys for 4 hours at 6°C. Following the preservation period, a contralateral nephrectomy was carried out and the treated kidneys were reperfused. These persufflated kidneys were compared with healthy control kidneys and kidneys preserved by SCS for 4 hours and at 6°C. In summary, it was shown that A-PSF was better at maintaining ATP levels than SCS alone. However, once blood flow was restored, kidneys preserved by A-PSF fared no better. Sixty minutes after reperfusion, static cold-stored kidneys restored their ATP to levels comparable to those achieved following A-PSF. It was observed that kidneys preserved by A-PSF exhibited healthy levels of ATP during the first 30 minutes of reperfusion, but that deterioration quickly ensued. The authors attributed this fall in ATP to the development of poor intrarenal blood flow following reperfusion and speculated that the cause was damage inflicted on glomerular vessels during A-PSF. In vivo renal function studies yielded findings that supported the belief that filtration had been most affected. Within 8 days, all dogs having a kidney preserved by A-PSF had died, while all dogs in the SCS control group survived. A progressive decline in renal function was documented through failing urine production, uremia and systematic elevation in serum creatinine. The glomerular filtration rate and renal plasma flow had dropped drastically by post-operative day (POD) 2 in all animals that died following A-PSF. Following this study, Isselhard pursued an identical study using R-PSF and, in contrast to A-PSF, R-PSF for 4 hours at 30 mmHg did not result in the same deterioration in kidney function. ATP levels remained similar after 60 minutes of reperfusion, were no different from healthy controls, and better than static cold-stored controls. Furthermore, serum urea and creatinine values were generally lower following R-PSF than with SCS, but remained above baseline at POD 10. Glomerular filtration rate and renal plasma flow were normalized by POD 2 in persufflated kidneys, but static cold-stored kidneys did not fully normalize until POD 21, highlighting the accelerated recovery of renal function following R-PSF. As a direct result of these studies, the Cologne group primarily adopted R-PSF as the most promising of these approaches.

In the mid-to-late 1970s, Fischer’s group contributed some of the most fundamental work on kidney PSF. In 1978, Fischer et al presented a study in which the functional recovery of kidneys was documented after 2 or 30 minutes of WIT and following 24 hours of R-PSF, all of which preceded heterotopic autotransplantation into dogs [17]. For this, the authors developed a unique model where a contralateral nephrectomy was not used to isolate the functional output of each kidney – rather the preserved kidney was by transplanted into the collar region while the contralateral kidney was left in the retroperitoneum. It was demonstrated that 24 hours of R-PSF, in the presence of up to 30 minutes of WIT, was capable of preserving post-transplant renal function. Key measurable parameters of kidney function were glomerular filtration rate and renal plasma flow during a 3-hour period following transplantation. In kidneys subjected to only 2 minutes of WIT, the persufflated and autotransplanted kidney exhibited mean glomerular filtration rate and renal plasma flow that were 46% and 56% of the healthy contralateral kidney, respectively. Similarly, in kidneys undergoing 30 minutes of WIT – the preserved kidney had mean values of glomerular filtration rate and renal plasma flow that were 32% and 49% of the healthy control during the observation period. These results demonstrated that ex vivo, hypothermic R-PSF for 24 hours can preserve renal function in the face of considerable WIT (30 minutes). At the same time, Isselhard suggested that the duration of cold preservation by R-PSF could possibly be extended even further, up to 48 hours [27] – leading to a new line of investigation. In summary, they found that R-PSF maintained ATP levels for up to 120 hours. A more significant finding was that the effect of 30 minutes of WIT did not have as profound an effect on the ability of R-PSF to resuscitate the organ. ATP levels were monitored for 72 hours and were maintained at levels comparable to kidneys not damaged by warm ischemia. As cold ischemia time was prolonged, the capacity for aerobic metabolism measurably decreased but it was apparent that R-PSF may extend the life of kidneys during cold preservation.

At this point, the reported merits of PSF had not been tried clinically. However, in 1975, Flatmark et al described a short study in which they reported their experiences with the accidental gas perfusion of human kidneys during HMP [19]. These kidneys were preserved in SCS (at 4°C, for 4–7.5 hours), transported from the site of procurement and then started on machine perfusion (at 8–10°C, with the perfusate equilibrated to 66% N₂, 33% O₂ and 1% CO₂) once received at their institution in Oslo. At some time after the start of HMP (between 3–12.5 hours), leaks were discovered near the oxygenator, which allowed air to be pulled into the flow circuit. In each of 4 kidneys the leak persisted and these organs were persufflated with air for about 60–120 minutes. Once the leak was identified, liquid perfusion was re-established and continued for the remainder of cold preservation, or 2–18.5 hours. Each kidney was subsequently transplanted, all produced urine immediately and most achieved healthy renal function by 4 weeks post-operatively. One of the 4 patients unfortunately died, but the cause of death was not reported. In conclusion the authors stated that PSF (which they referred to as ‘massive gas embolization’) for up to 2 hours did not adversely affect post-transplant renal function. This was the first time that PSF was performed on human organs, albeit inadvertently.

The significant contributions of the Cologne group provided impetus for other investigators to pursue this field of investigation. In Australia, Ross and Escott explored 24-hour PSF of canine kidneys following 30 minutes of WIT and the effects of PSF on heterotopic autotransplant outcomes [70; 71]. These studies focused on how the composition of the gaseous perfusate affected post-transplant renal function and survival. Also, these investigators tried persufflating via the ureter for the first time. Three interesting observations were reported: Firstly, that R-PSF was better at preserving renal function than A-PSF, as measured by serum creatinine values post-transplant. Secondly, that carbogen (95% O₂, 5% CO₂) may be better than pure oxygen gas (100% O₂) in normalizing renal function post-transplant. Thirdly, that ureteral PSF may work, but that it needs further exploration. One puzzling finding from these studies was the relatively high incidence of intravascular thrombosis in all groups studied. They had postulated that it may have been due to endothelial damage resulting from a gas-drying effect, yet the greatest incidence of thrombosis occurred in the group receiving humidified pure oxygen gas. Traumatic damage to the vasculature may have been an alternative explanation. An explanation not explicitly considered was the possibility of endothelial damage resulting from hyperoxia.

During this same era, Pegg and his group in Cambridge, England published a series of studies exploring the utility of PSF in kidney preservation. This group developed a fairly elaborate canine autotransplantation model that they used to examine the effects of varying lengths of WIT (30 and 60 minutes) in combination with R-PSF for 24 or 48 hours duration [64; 68; 69]. They also explored the differences between pure oxygen, air, nitrogen and helium PSF on post-transplant renal function and survival (for up to 3 months). It was reported that no kidney transplanted after 60 minutes of WIT and 24 hours of SCS was able to sustain recipient survival. On the contrary, R-PSF with pure oxygen gas was capable of sustaining long-term renal function and survival in most recipients. When air was substituted for oxygen, kidneys remained functioning, but the survival rates decreased. R-PSF with nitrogen and helium gases generated results similar to kidneys preserved by SCS. Kidneys that had undergone 30 minutes of WIT and 48 hours of cold ischemia time performed significantly better post-transplant, particularly if they had been persufflated; most animals (80% of group) survived R-PSF, while only a single animal (20% of group) survived following transplantation of a control kidney. Another noteworthy finding was that they were not able to establish any differences in total adenine nucleotide content between persufflated and control kidneys, or between kidneys persufflated for 24 versus 48 hours or with different gases. One explanation is that the authors do not report ATP, ADP or AMP levels, but rather total adenine nucleotide content, which is the sum of these three; it is possible that the individual ATP and ADP fluctuations are masked by their summation. It is also important to note that the R-PSF pressures used in their study were particularly low, which may have resulted in inadequate PSF of the entire organ. This may be corroborated by the better outcomes using pure oxygen gas rather than air. Since the total adenine nucleotide content measurements are derived from tissue processed following biopsy, it is possible that the sampled regions of the kidney were poorly persufflated or that these samples do not reflect real-time total adenine nucleotide content levels because the tissue processing involves several steps and takes some time. Nevertheless, these data are compelling because the authors raise very reasonable questions regarding the interpretation of ATP and ADP measurements and their use as accurate predictors of metabolic quality in organ preservation. It has been widely appreciated that high energy phosphates recycle rapidly and that a single measurement is only a snapshot into the tissue metabolic status that does not necessarily provide information regarding the ability of tissue to recover from an insult [86].

The Cambridge group also published a follow-up pilot study in which a number of DCD human kidneys were persufflated prior to transplantation and compared with static cold-stored kidneys post-transplant in a paired fashion (meaning that each donor provided one kidney for R-PSF and another for SCS) [69]. The average WIT was 55 minutes, whereas the average cold ischemia time was 21.5 hours. The persufflated organs performed better post-transplant, exhibiting a mean onset of function at 8.4 ± 2.6 days versus 13.9 ± 1.4 days in the paired controls. Furthermore, the reported mean serum creatinine levels at POD 15 were 457 μM and 826 μM for the R-PSF and SCS groups, respectively. It was also explained that cyclosporine was largely omitted from the immunosuppressive regimen in order to limit the impact of calcineurin inhibitor-associated nephropathy on the study results. Even so, of the 6 recipients that received cyclosporine, 4 of them received a kidney that was preserved by R-PSF. This first clinical study exploring R-PSF in preserving DCD kidneys illustrated that: 1) R-PSF can be executed within the current clinical infrastructure, including timeframe; and 2) R-PSF also exhibits the ability to resuscitate organs that have suffered from significant warm ischemia. Collectively, these two observations suggest that R-PSF may be easily implemented and could make more DCD organs suitable for transplant.

In an attempt to further elucidate the mechanism of preservation by PSF, Pegg et al carried out an interesting study in 1989 in which they compared 24 hours of SCS to 24 hours of R-PSF following 60 minutes of WIT in an ex vivo rabbit kidney model [64]. Their hypothesis was that gaseous oxygen PSF enables continued aerobic respiration and that hypothermic conditions only slow the rates governing energy turnover. They studied ATP levels following R-PSF in conjunction with pharmacologic treatment with ouabain or cyanide/iodoacetate. Concomitant administration of ouabain, a potent sodium pump inhibitor, led to elevated ATP levels over R-PSF controls, which exhibited significantly higher ATP levels if compared with kidneys preserved by SCS. Conversely, treatment with cyanide/iodoacetate resulted in a fall of ATP, below levels present in static cold-stored kidneys. These measurements confirmed that if oxygen is available, even as a gas at low temperatures, it will be consumed by viable tissue. If pumps that consume ATP, like the Na⁺/K⁺-ATPase, are actively blocked, ATP levels will transiently rise, suggesting that ‘new’ ATP has been produced. On the contrary, if the electron transport chain is inhibited, then the capacity to generate ATP drops and the available oxygen supply will not be consumed. These findings provided insight into the mechanistic principles governing oxygen consumption. Histology and electron microscopy revealed that renal tissue after R-PSF exhibited cellular damage over smaller foci and were largely surrounded by healthy tissue. Intriguingly, electron micrographs were capable of discriminating between the “orthodox” (or energized) mitochondrial configuration present in persufflated samples and the “condensed” (de-energized) mitochondria of the cold-stored tissues [64]. As a whole, these findings support the belief that R-PSF can maintain renal tissue viability and prevent irreversible injury during prolonged cold storage.

During this time, Stowe et al at the Cleveland Clinic pursued ex vivo R-PSF in conjunction with a canine autotransplantation model [84]. After resecting the left kidney and flushing immediately with chilled lactated Ringers solution, R-PSF was performed for 48 hours (at 4–5°C and 7–10 mmHg). Two days following the start of cold preservation, a contralateral nephrectomy was performed and the preserved kidney was transplanted heterotopically. Only 25% of animals that were autotransplanted with kidneys preserved by SCS or R-PSF survived to POD 22. In addition to comparing R-PSF with SCS, 3 animals received an intramuscular dose (500 mg) of deferoxamine mesylate. All 3 animals receiving a kidney preserved by R-PSF and treated with intramuscular deferoxamine mesylate survived to POD 22. In general, persufflated kidneys exhibited better overall renal function than static cold-stored kidneys. Though the total number of transplants was low, it appeared that deferoxamine mesylate contributed to an elevated survival rate post-transplant. The authors postulated that this was possibly due to free radical-scavenging activity in deferoxamine mesylate.

Several years later, Kootstra’s group at the University Hospital in Maastricht addressed the question of whether the presence of adenosine benefits the quality of warm ischemically-damaged (30 minutes of WIT) and preserved rat kidneys during both R-PSF and SCS [101]. Yin et al used UW solution containing exogenous adenosine and UW solution with no adenosine. All kidneys, whether preserved by 24 hours of R-PSF or SCS, were flushed with either of these UW solutions [101]. In brief, they determined that regeneration of ATP was not affected by the presence of adenosine in UW solution. The authors postulated that since the levels of hypoxanthine (degradation product of adenosine) were significantly higher in renal tissue preserved with exogenous adenosine than in tissue without, that most of the additionally available adenosine was degraded in the tissue and not used in the direct replenishment of ATP. However, hypoxanthine levels were significantly lower in persufflated kidneys than in static cold-stored kidneys, which were also flushed with adenosine-containing UW solution. Alternatively, it may be that adenosine found in the preservation solutions may not be able to cross the plasma membranes of viable cells. The authors also went further by transplanting 20 kidneys, 10 from each of the R-PSF and SCS groups. They reported that none of the rats transplanted with static cold-stored kidneys survived, whereas 3 of the rats transplanted with retrogradely persufflated kidneys survived an observation period of 2 weeks. Because the survival rates between the two groups were not significantly different, they concluded that singular measurements of ATP during the preservation period may not predict survival outcomes. These conclusions were similar to those generated at Cambridge [68]. However, a point of caution must be raised regarding the choice in animal model. In general, PSF (or HMP, for that matter) will impart its greatest benefit if an organ is large and cannot obtain adequate oxygen by passive diffusion from the surface alone. In other words, success in a rat model may not be comparable to the human clinical situation because of significant size disparity. This scaling problem has already been encountered when comparing the rat and porcine pancreata during SCS [63]. Rat data must be interpreted with appreciation for oxygen transport limitations, especially when trying to translate outcomes from animal models into the clinical arena.

Finally, more recently, Treckmann built upon the work of his predecessors and published studies that highlight the promising prospects of clinical PSF [96; 97]. In 2006, Treckmann et al used a porcine autotransplantation model to show that R-PSF of kidneys damaged by significant WIT, when compared with conventional SCS, may yield improved survival [97]. Left kidneys from porcine donors were clamped off in situ for 60–120 minutes of WIT, resected, flushed anterogradely, preserved by R-PSF or SCS for 4 hours and autotransplanted. They demonstrated that all animals transplanted with persufflated kidneys after 60 minutes of WIT survived the 7-day observation period, with robust renal function. Animals receiving a similarly-treated kidney but stored by SCS survived in 57.1% of cases. This paired comparison was also performed for both 90 and 120 minutes of WIT, but did not reveal strong differences between R-PSF and SCS, at least in terms of post-transplant survival. It is important to note that only the persufflated kidneys received a pre-treatment of superoxide dismutase (SOD), which had been determined from their earlier liver work, to help protect against the oxidative damage caused by hyperoxia and reperfusion [49]. A follow-up study compared R-PSF and HMP, using the same porcine autotransplant model; thirteen kidneys were resected following 60 minutes of WIT and preserved by SCS, HMP or R-PSF following pre-treatment with SOD [96]. Results showed that all indicators of renal function were significantly better in persufflated kidneys versus HMP and static cold-stored kidneys at POD 7. Recipient survival after 7 days was 57%, 60% and 100% after transplantation of kidneys preserved by SCS, HMP and R-PSF, respectively. These results in the larger animal model suggest that R-PSF is a promising way to maintain organ quality in DCD kidneys, particularly because R-PSF was directly compared against the already clinically-accepted HMP and performed comparably if not better. The collective work performed on kidney PSF has been summarized in Table 3.

C. Liver

PSF was tested for the first time in rat livers around 1980 by Fischer’s group at the University of Cologne [15], but was not studied in depth (by the same group) until the 1990s. Initial work revolved around establishing that livers damaged by prolonged WIT could be resuscitated using R-PSF. Each liver in the R-PSF cohort was flushed with cold preservation solution and R-PSF was performed by administering gaseous oxygen at 18–30 mmHg via a hepatic vein, while providing an escape route for the gas by introducing small, pin-sized holes on the liver surface. They reported that 24 hours of R-PSF resulted in no detectable lactate accumulation, but a substantial decrease in glycogen content.

It was during this period that anti-oxidants were identified and used to provide additional value in preservation using R-PSF, by conferring improved hepatocellular integrity and function following reperfusion [49]. In the 1990s, the same group extended some of their in vitro studies into small animal (rat) and large animal (pig) liver transplant models. They showed that poorer quality livers could be revived to function successfully post-transplant. As recently as 2008, the same researchers showed that clinical PSF is possible; having a cohort of 5 patients transplanted with persufflated livers and showing no adverse effects and strong graft function [94]. Another interesting niche for PSF has been in the conditioning of a liver near the end of the preservation period, in so-called end- or post-ischemic conditioning [56]. As it stands, liver PSF is actively being explored and the preceding three decades of work highlights some of the reasons why.

In 1993, Minor et al studied reperfusion injury in the rat liver following an ischemic period during preservation [49]. They procured a number of rat livers, flushed sequentially with lactated Ringers and Euro-Collins solutions via the hepatic vein, and bathed the organ in Krebs-Henseleit solution at 37°C for 60 minutes. After the controlled duration of WIT, the liver was submerged in Euro-Collins solution at 4°C and stored for another 60 minutes. Following static storage alone, the organs underwent normothermic (37°C) R-PSF for 30 minutes and were subsequently flushed with lactated Ringers solution. Persufflated organs received some combination of anti-oxidant pre-treatment, either a bolus injection of allopurinol prior to ischemia and/or the addition of allopurinol or SOD into the flushing solution. At the end of the 2.5-hour treatment, the liver was flushed with lactated Ringers solution via the infrahepatic caval vein. The effluent was collected and analyzed for ATP and total adenine nucleotide content. The authors also determined the amount of malondialdehyde accumulated (via free radical-induced lipid peroxidation) and the amounts of liver enzymes released by damaged hepatic parenchyma. It was reported that R-PSF of the liver was capable of partially reversing ATP levels after 120 minutes of combined warm and cold ischemia. Additionally, it was shown that pre-treatment with anti-oxidants decreased the degree of lipid peroxidation and improved ATP recovery with R-PSF. Samples from persufflated livers pre-treated with allopurinol/SOD revealed that gaseous PSF alone can harm the liver due to oxidative damage, but also highlighted the potential of anti-oxidant administration as an adjunctive therapy.

In a follow-up study, the same group showed that early administration of R-PSF reduces lipid peroxidation and may actually suppress the adverse effects of free-radical damage [46]. The authors speculated that immediate PSF may prevent damage by enabling the preservation of free-radical scavenging activity, which itself can require energy. Two years later this concept was studied in more depth using a rat liver reperfusion model [47; 48]. The effect of R-PSF in suppressing ischemia reperfusion injury was compared to preservation with 48 hours of SCS in UW solution at 4°C and followed by a 30 minute period of re-warming with normal saline at 25°C. Reperfusion consisted of pre-oxygenated (95% O₂, 5% CO₂), re-circulating Krebs-Henseleit solution delivered through the portal vein for up to 45 minutes. The effluent was analyzed and it was reported that endothelial and hepatic parenchymal damage was lowered and that activation of Kupffer cells was reduced in livers that had been persufflated. These findings suggested that R-PSF may avoid disturbances in perfusion (like higher portal venous pressures) that result from damage to vessels. In addition, ATP concentrations were higher in persufflated livers than in static cold-stored livers and, at different times throughout the reperfusion, comparable to fresh control livers that were not subjected to cold storage. It was interesting to note that livers preserved by R-PSF exhibited significantly greater ATP levels as compared to reference values obtained from native rat liver [47].

These studies were extended to evaluate whether R-PSF in combination with SOD pre-conditioning would resuscitate the DCD rat liver, harvested after 30 minutes of WIT [50]. Livers are known to be poorly resistant to warm ischemic damage [93]. In fact, post-transplant outcomes following DCD transplant have been strongly tied to the extent of warm ischemic injury [66; 67; 93]. In this study, DCD livers preserved for 24 hours with R-PSF appeared to be healthier when compared with livers preserved by SCS in UW solution. The extent of lipid peroxidation was found to be lower in the R-PSF cohort. Interestingly, bile production and ATP levels were higher, while endothelial damage and portal perfusion pressures were lower than even fresh controls following 45 minutes of reperfusion with pre-oxygenated Krebs-Henseleit solution. It appeared that the putative, harmful effects of high oxygen concentrations during R-PSF could be prevented with the help of anti-oxidant treatment and resuscitation of livers from DCD was possible using R-PSF.

In 1997, Minor et al introduced a novel rat liver transplant model [52] in which rat livers were preserved with SCS alone or R-PSF. The hypotheses they tested was the proposal that R-PSF limited proteolytic degradation of the liver, believed to contribute to hepatocellular injury during SCS, and that R-PSF would result in improved post-transplant indicators such as decreased plasma levels of malondialdehyde, decreased alanine aminotransferase, increased bile production and increased hepatic tissue perfusion. Proteolysis was estimated by a measured tissue content of free L-alanine. Twenty-four hours of SCS in UW solution resulted in significantly higher concentrations of free L-alanine than fresh (non-stored) controls, while R-PSF seemed to prevent some proteolysis. Following in vivo reperfusion, Minor et al reported that both the static cold-stored only and persufflated livers experienced decreased hepatic perfusion initially, but that the persufflated organs exhibited an overall better recovery. Furthermore, plasma levels of malondialdehyde and alanine aminotransferase were significantly lower in the livers having undergone R-PSF, but still elevated in both, while hepatic bile production was significantly increased in persufflated livers and comparable to fresh controls. It was concluded that R-PSF may help the ischemic liver obviate some of the ill-effects of hypoxia (like activation of cytosolic proteases and autolysis) by maintaining a more favorable metabolic condition. The authors reiterated that protecting tissue from oxidative damage necessitates sufficient energetic support, fueled by an adequate oxygen supply.

In the same year, Minor et al examined the effect of lower pressure (9 mmHg) versus higher pressure (18 mmHg) R-PSF and the use of both pure gaseous oxygen and air [53]. The effectiveness and homogeneity of the R-PSF was studied by detecting autofluorescence of nicotinamide adenine dinucleotide [54; 55; 59], accumulated primarily in anoxic tissue. The reason why R-PSF was traditionally performed at 18 mmHg of pressure was arbitrary; it had been determined that it was the pressure required for visual detection of bubbles escaping from the surface of the perforated liver. Nonetheless, interrogation at 1 and 24 hours after the start of cold preservation revealed that R-PSF at 9 and 18 mmHg resulted in a comparable and significant decrease in nicotinamide adenine dinucleotide over static cold-stored controls. On the other hand, R-PSF with air at 18 mmHg did not decrease detected nicotinamide adenine dinucleotide levels below those detected in livers on SCS, suggesting that air may not provide adequate oxygen during liver PSF.

Following the extensive studies in rat livers, translation of PSF into a larger animal model was the next step. In 1998, Minor et al described work in which recipients were allotransplanted with DCD porcine livers following one of two preservation protocols [57]. All livers were harvested after 45 minutes of in situ WIT in a non-heparinized donor. Using the first protocol, the donor livers were flushed with heparinized normal saline and UW solution via the portal vein and stored on UW solution at 4°C for 4–5 hours. With the second protocol, the donor livers were also flushed with heparinized normal saline and UW solution, but the last 100 mL of UW solution used to flush the liver was spiked with SOD. Following the flush, R-PSF was initiated via the inferior vena cava for 4–5 hours. After the period of cold preservation, the stored livers were orthotopically transplanted into recipients. Shortly following transplantation, it was determined that SCS alone was not able to adequately preserve DCD livers; all 5 recipients died in the early post-transplant course. On the other hand, all livers stored by R-PSF were capable of normalizing ammonia levels by POD 1 and aspartate aminotransferase plasma levels by POD 7.

In a follow-up study, the same investigators again compared SCS and R-PSF/SOD preservation by measuring plasma levels of aspartate aminotransferase, alanine aminotransferase, lactate dehydrogenase and clotting times at 1-hour post-transplant, with the primary endpoint of recipient survival at POD 5 [75]. DCD livers underwent 60 minutes of WIT prior to 4 hours of SCS with UW solution or R-PSF and pre-treatment with SOD. All DCD transplants were compared with control livers resected and transplanted immediately following cardiac arrest. They reported that UW-stored livers fared poorly post-transplant, accounting for significantly higher plasma hepatic enzyme levels, lactate dehydrogenase and prolongated partial thromboplastin time, suggesting that the liver had incurred significant hepatocellular injury during the preservation. Additionally, all of the animals receiving such livers died within 3 hours of reperfusion. On the contrary, animals receiving livers preserved by R-PSF/SOD survived the observation period. Impressively, it was stated that persufflated porcine DCD livers performed comparably to fresh livers during the post-transplant course. Since very few DCD livers are transplanted nowadays, making more DCD livers available for transplant would reduce the numbers of patients on liver transplant waiting lists.

Around the same time PSF was being used in the resuscitation of DCD livers, Minor et al explored an interesting application of PSF in pre-conditioning the long-term, static cold-stored liver for reperfusion [56]. In his first study on this topic, rat livers were subjected to either 48 hours of SCS on UW solution or 47 hours of SCS on UW solution followed by R-PSF for 60 minutes. Each of the organs were then warmed and reperfused in vitro for 45 minutes. Following reperfusion, livers conditioned (with R-PSF) fared significantly better – exhibiting decreased parenchymal enzyme levels and portal venous pressures, and improved bioenergetic status. It is believed that the delivery of gaseous oxygen prior to reperfusion may prevent edema formation due to the improvement in metabolic condition of the organ. Additionally, it may be that free-radical scavenging activity has been depressed during SCS and that pre-conditioning for reperfusion using PSF helps provide the oxygen necessary for replenishment of such activity [56; 58].

The concept of using PSF to condition for reperfusion was studied more than a decade later for the preservation of fatty rat livers [58]. In these most recent studies, R-PSF for 90 minutes following 20 hours of SCS decreased hepatic parenchymal enzyme release, lipid peroxidation, cellular apoptosis and autophagy, and improved the functional clearance of ammonia, microscopic morphology and overall metabolic status of these livers as compared with unconditioned livers. These preliminary studies illustrate that post-ischemic conditioning of preserved organs prior to transplant is an area meriting further study.

With anti-oxidant therapy having already been explored for use with PSF [49; 50; 57; 75], it became apparent that continued success using PSF in preservation may require protection against reperfusion injury. In 2003, Lauschke et al studied the effect of administering taurine or SOD prior to R-PSF [38]. These studies employed a DCD rat model in which the livers were resected following 60 minutes of WIT and preserved either by SCS in UW solution, by R-PSF/taurine, or by R-PSF/SOD for 24 hours. Following the preservation period, livers were reperfused in vitro using Krebs-Henseleit solution maintained at 37°C for 45 minutes. Analysis of the effluent at the end of the reperfusion period revealed that livers treated with anti-oxidant exhibited decreased enzyme release and portal vascular resistance, and increased bile production. Interestingly, taurine and SOD appeared to have a similar protective effect for DCD livers in the face of ischemia-reperfusion. Future work to enhance the successful application of PSF for all organs of interest may require the identification of the appropriate anti-oxidant, dose(s) and schedule.

Most recently, Treckmann and colleagues have taken a major step, by translating R-PSF into the human clinical setting. In 2008, they reported results from a pilot study using R-PSF to resuscitate 5 DCD human livers between April 2004 and March 2005 [94]. These donor cadaveric livers were estimated to have undergone anywhere from 20–60 minutes of WIT and the donors had expired after failed attempts at cardiopulmonary resuscitation or having experienced prolonged periods of hypoperfusion (<60 mmHg). It is also important to note that these donor livers were rejected for transplant by at least three different centers. The livers were procured off-site, perfused with UW solution and Histidine-Tryptophan-Ketoglutarate solution, and then shipped to the transplant center. After establishing recipient consent, the livers were additionally flushed with UW solution containing N-acetylcysteine and were retrogradely persufflated at 18 mmHg for 70–200 minutes prior to orthotopic transplantation. The results of the transplants were encouraging in all cases. All patients survived and none of the patients required a re-transplant; all patients were alive, with strong graft function, at a minimum of 2 years follow-up. Histologic evaluation was performed on biopsies taken immediately prior to and after R-PSF and directly following reperfusion. Analysis revealed that R-PSF did not appear to cause any vascular damage to the liver. Additionally, it was shown that PSF had recovered ATP levels by 2–5 times the pre-PSF measurements. These data are highly encouraging, especially considering the potential impact that PSF could have in expanding DCD liver transplantation. Table 4 summarizes the published work on liver PSF presented in this review.

D. Small bowel and pancreas

To date, the pancreas and small intestine have not been studied extensively as targets of PSF. In 1997, Minor et al published the only known work on luminal gas oxygenation of the small bowel [51]. Rat jejunal segments (15-cm in length) along with the vascular pedicle were harvested and stored in UW solution at 4°C for 18 hours and half of the experimental organs underwent low-pressure, luminal gas oxygenation. Following the experimental storage period, intestinal absorption was estimated by introducing galactose to the lumen and measuring concentrations in portal venous effluent. Collecting the total luminal effluent and subtracting the known inflow volume was used to measure the net influx of water into the intestinal lumen. Results showed significantly increased accumulation of hypoxanthine in the small bowel segments with SCS alone. ATP, creatine phosphate and total adenine nucleotide content were significantly higher in the group undergoing luminal gas oxygenation versus SCS alone and resembled values from rat intestine in vivo. However, intestinal carbohydrate absorption was found to be severely impaired in both static cold-stored and gas oxygenated jejunal segments, even though gas oxygenation significantly improved post-ischemic absorption when compared to SCS. The net secretion of water into the gut lumen was significantly lower following gas oxygenation than SCS, reflecting less damage to intestinal villi. The authors noted that luminal gas oxygenation of the intestine could be improved by the introduction of supplements to the cold preservation solution, such as glutamine, an important substrate for intestinal mucosal cells. Although unique opportunities exist for the preservation of the small intestine using intraluminal gas oxygenation (a variant to intravascular PSF), this area of research has remained largely unexplored.

In contrast to small bowel intraluminal gas oxygenation, pancreas PSF has started to attract more interest in recent years. Currently, it is widely regarded that improvements in organ preservation may have a positive impact on pancreatic islet isolation and transplant outcomes [33]. For some time, the two-layer method (TLM) was considered the state-of-the-art for pancreas preservation before islet isolation. This has recently been challenged by studies showing that islet isolation outcomes are equivalent when comparing TLM and conventional SCS [1; 7; 78]. It is likely that the inefficiency of oxygen delivery by passive diffusion from the organ surface alone is responsible for the limited efficacy of TLM in preserving larger organs [63].

Over the last several years, PSF has been identified as a possible improvement to the current pancreas preservation protocol, particularly before islet isolation. Our research group is currently studying PSF of the pancreas to parallel our concurrent interests in pancreas HMP [89; 91] and have recently published several works on A-PSF of porcine pancreata for the purposes of improving whole organ and islet quality [78; 79]. In one study [79], human and porcine pancreata were preserved using TLM or A-PSF at 4°C. A-PSF was performed via the superior mesenteric artery and either the splenic artery (human) or the celiac trunk (pig). Following procurement, the organs were imaged by conventional MRI and ATP levels and ATP-to-inorganic phosphate ratios were estimated using ³¹P-NMR spectroscopy. MRI revealed well-distributed areas of negative contrast throughout all persufflated pancreata, indicating the homogeneous presence of gas within the organ. Rat pancreata preserved by TLM showed relatively high ATP levels, though ATP levels were nearly undetectable in porcine pancreata preserved with TLM. In contrast, human pancreata preserved by A-PSF exhibited ATP-to-inorganic phosphate ratios similar to those observed in the rat pancreata on the TLM. Additionally, when A-PSF was stopped, ATP-to-inorganic phosphate ratios quickly declined to undetectable levels, similar to porcine organs preserved by TLM. When A-PSF was re-started, ATP levels rose again. In another study, DCD porcine pancreata were procured and the splenic lobe was separated from the connecting and duodenal lobes [78] – the anatomy being described previously [12]. The duodenal lobe was isolated following 1.5–2 hours of SCS and served as a first control, while the connecting lobe was stored on TLM for 24 hours and at 4°C to serve as a second control. Splenic lobes were submerged in cold preservation solution and preserved by A-PSF using a custom-designed, portable electrochemical oxygen concentrator (Giner Inc, Newton, MA) via the celiac trunk and superior mesenteric artery for 24 hours and at 4°C. Biopsies from organs preserved by A-PSF showed distended capillaries and less autolysis and necrosis when compared to organs preserved by TLM. In contrast, TLM-stored pancreata showed frequent pyknotic nuclei, indicating possible irreversible cellular damage. A follow-up study extended the comparison to porcine pancreatic isolation, having shown that 24 hours of A-PSF was better than the TLM in preserving islet morphology, viability and post-culture recovery (unpublished results). Collectively, these results illustrate the potential of A-PSF in improving tissue and islet quality in larger pancreata when compared with conventional preservation methods.

These early and promising results have established the need for more extensive studies since pancreas PSF represents a significant opportunity to improve both solid organ and islet preservation over the current preservation techniques used in the field. Additionally, the portable electrochemical oxygen concentrator used in our studies may enable more widespread use of PSF, not just of the pancreas, since it provides a safe method of delivering gas (or gas mixtures) during transportation and especially during air travel. Table 5 summarizes the published work on small intestine and pancreas PSF presented in this review.

A schematic of the key milestone events during the historical development and use of PSF is presented in Figure 2.