A Comparison of Pulsatile Hypothermic and Normothermic Ex Vivo Machine Perfusion in a Porcine Kidney Model : Transplantation

Journal Logo

Original Basic Science—General

A Comparison of Pulsatile Hypothermic and Normothermic Ex Vivo Machine Perfusion in a Porcine Kidney Model

Vallant, Natalie PhD1,2; Wolfhagen, Nienke MD1; Sandhu, Bynvant PhD1,2; Hamaoui, Karim PhD1,2; Cook, Terence FMedSci2; Pusey, Charles MB2; Papalois, Vassilios PhD1

Author Information
doi: 10.1097/TP.0000000000003599
  • Free

Abstract

INTRODUCTION

Kidney transplantation as the treatment of choice for end-stage renal failure is limited by the shortage of donor organs,1 resulting in a high mortality rate on the waiting list.2 In an attempt to overcome the organ shortage, centers are accepting kidney grafts from deceased brain dead extended criteria donors3,4 as well as from deceased donors after circulatory death (DCDs).5,6 Usage of these organs increases the risk for primary nonfunction and delayed graft function,7,8 resulting in decreased long-term graft survival rates.9,10 Strategies to overcome this problem and to improve the quality of marginal grafts are under investigation.

Ex vivo machine perfusion of organs has been the focus of interest, driven by the aim to not only optimize but also preassess graft viability and to repair damaged organs before transplantation. Hypothermic machine perfusion (HMP) and normothermic machine perfusion (NMP) both have been studied intensively and are being introduced into clinical practice.11

HMP has been demonstrated to be superior to cold static storage (CSS), especially in organs coming from extended criteria donors. Its application resulted in a reduction of delayed graft function and improved graft survival in the first year after transplantation. This effect could be demonstrated for kidneys from deceased brain dead donors, as well as for kidneys from DCD donors.12-15 In clinical settings, HMP has been adopted into routine practice in many countries throughout Europe, with all kidneys from DCD donors undergoing HMP in the Netherlands, and the United Kingdom not being far behind, following a recommendation of HMP by the National Institute for Health and Care Excellence.16

NMP aims to keep organs in a functional state by keeping up an aerobic metabolism at physiological temperatures. Clinical experience in ex vivo normothermic kidney perfusion, showing superiority of NMP to CSS, is mainly based on experimental and clinical studies by Hosgood and Nicholson.17-22

Limited data are available on a direct comparison between NMP and HMP. Therefore, the aim of our experiments was to conduct a study looking at the 2 preservation methods, not only during the respective hypothermic and normothermic perfusion periods but also foremost during the reperfusion period with whole blood during which 2 organs are exposed to the same temperature, with the exact same perfusion fluid and perfusion parameters. In this first head-to-head kidney perfusion study, outcomes after NMP and HMP of kidney pairs from respective same donors were directly compared in a porcine-controlled DCD model.

MATERIALS AND METHODS

This research was conducted at the laboratories of Imperial College London. All experiments were performed in accordance with and approval of departmental and institutional guidelines and policies regarding ex vivo research using animal tissue.

Organ Retrieval

Porcine kidneys were retrieved from adult landrace pigs (70–90 kg) following euthanasia at a local abattoir. The warm ischemia time (WIT) was kept at 25 min. Kidneys were rapidly retrieved and flushed with 500 mL of Soltran via the renal artery at a hydrostatic pressure of 100 cm H2O before being placed on ice (in Soltran) and transported to the laboratory at Imperial College London. At the laboratory, kidneys were stored in ice for 24 h before being allocated into their respective study groups.

Study Design and Groups

A total of 34 porcine kidneys were used for this study. Eighteen paired kidneys from 9 donor pigs were retrieved and 1 respective organ was allocated to either undergo NMP (n = 9) or HMP (n = 9) for 4 h, before being normothermically reperfused with autologous whole blood for 2 h using 1 RM3 perfusion machine. Twelve consecutive unpaired kidneys from 12 donor pigs also underwent NMP (n = 6) or HMP (n = 6) for 4 h, followed by 2 h of normothermic reperfusion, as controls for other studies. During simulated reperfusion, perfusion and functional parameters were analyzed and compared for the 18 paired kidneys, as well as for the 12 unpaired kidneys. Furthermore, 4 consecutive single kidneys underwent CSS for 28 h before also undergoing reperfusion with autologous whole blood for 2 h. In summary, this resulted in the following 3 main treatment groups, including “paired” and “unpaired” treatment groups:

  • Group NMP: kidneys after 25 min of WIT and 24 h of CSS, followed by 4 h of NMP and 2 h of simulated reperfusion with autologous whole blood (9 paired + 6 unpaired kidneys).
  • Group HMP: kidneys after 25 min of WIT and 24 h of CSS, followed by 4 h of HMP and 2 h of simulated reperfusion with autologous whole blood (9 paired + 6 unpaired kidneys).
  • Group CSS: kidneys after 25 min of WIT, 28 h of CSS, and 2 h of simulated reperfusion with autologous whole blood (n = 4).

Machine Perfusion—NMP and HMP

All perfusion experiments were conducted using the Waters Medical Systems RM3 perfusion machine (Rochester, MN). It pumps the perfusate through the kidney vasculature in a pulsatile and pressure-controlled way. The systolic perfusion pressure can be varied from 0 to 100 mm Hg. The RM3 continuously measures the Renal Resistive Indices (RRIs; mm Hg/mL/min) as well as the perfusate flow rates (mL/min) and the perfusate temperature (°C).

Normothermic Machine Perfusion

After 24 h of CSS, kidneys were benched and a steel cannula was tied into the renal artery. For the collection of urine, a Foley catheter was placed into the ureter and the balloon inflated slightly to keep it in place. The circuit was primed with Ringer’s solution and the NMP perfusate was added, which was prepared as described in previous publications23: per 1 unit of porcine autologous leukocyte-depleted blood, 200–400 mL of Ringer’s Lactate, 10–25 mL of 10% mannitol, 2 mL of dexamethasone, 10–40 mL of 8.4% sodium bicarbonate (B. Braun), 5000 IU of heparin, and 0.1 g of creatinine were added. This resulted in 600 mL of NMP perfusate with a physiological pH (7.35–7.53). Autologous whole blood was collected into a flask containing 20.000 IU of heparin and 750 mg of cefuroxime at the time of organ retrieval. Five hundred milliliters were transferred into citrate phosphate dextrose adenine blood collection bags (Macopharma) for normothermic reperfusion and 500 mL were filtered using a leukocyte filter (Macopharma) to produce 1 unit of leukocyte-depleted blood as a basis for the NMP perfusate. All blood bags were stored at 4 °C over night. On the next day, before starting the perfusion experiments, the perfusate for NMP was prepared as described above.

To use the RM3 machine for NMP, an oxygenator (Eos ECMO) and a water-heating pump (Grant) were incorporated into the circuit as shown in Figure 1A. The oxygenator was connected to an oxygen (O2)/carbon dioxide (CO2) mixture of 95% O2 and 5% CO2. The flow rate was kept at 0.2 mL/min. The circuit was primed with Ringer’s lactate solution and thereafter, the NMP perfusate was added. When a temperature of >35 °C was reached, blood gas analyses were performed and acidosis was corrected by titration of 8.4% sodium bicarbonate solution. During 4 h of NMP, the systolic arterial blood pressure was kept between 50 and 60 mm Hg. Temperature, arterial blood pressure, RRI, perfusate flow rate, and urine output volumes were recorded at 0, 0.25, 0.5, 1, 2, 3, and 4 h of perfusion. Also, at these time points, arterial and venous perfusate samples were collected for arterial blood gas analysis and urine samples were collected for biochemical analyses. Prostacyclin (Epoprostenol, Flolan), mannitol 10% solution (Baxter), and glucose 5% solution (Fresenius) were added to the circuit in units of 5 mL/h.23 Urine output was replaced mL per mL with Ringer’s lactate. After 4 h of NMP, kidneys were detached from the machine and flushed with 500 mL of cold Soltran perfusion solution to mimic a clinical situation. Kidneys were then kept on ice in Soltran solution for the duration of changing the circuits (30–40 min) and until the start of the simulated reperfusion period.

F1
FIGURE 1.:
Circuits for NMP (A) and HMP (B) using the Waters RM3 machine. Photographs of the circuits for NMP (A) and HMP (B) are shown using the Waters RM3 machine. For HMP, the machine was used as per user guidance as illustrated in (B), with a kidney attached and undergoing HMP. Photograph (A) shows the setup for a kidney undergoing NMP, with the main difference of having an oxygenator (Eos ECMO; white arrow) included in the circuit. Instead of the ice chamber to cool the perfusion fluid, a water-heating pump (Grant) was attached to the machine. HMP, hypothermic machine perfusion; NMP, normothermic machine perfusion.

Hypothermic Machine Perfusion

For kidneys undergoing HMP, 500 mL of machine perfusion solution (Belzer) were used for the circuit. Kidneys were perfused at a systolic pressure between 40 and 50 mm Hg for 4 h. Perfusate and kidneys were kept at temperatures between 4 °C and 6 °C. Temperature, arterial blood pressure, RRI, perfusate flow rates, and urine production rates were recorded at 0, 0.25, 0.5, 1, 2, 3, and 4 h of perfusion. Perfusate samples were collected at these time points for biochemical analysis. Figure 1B shows the circuit used for HMP. After 4 h of HMP, kidneys were detached from the machine and kept on ice in Soltran solution for the duration of changing the circuits (30–40 min), until the start of the simulated reperfusion period.

Cold Static Storage

Four single kidneys from 4 different donors were stored at 4 °C in a bag filled with Soltran, for a total of 28 h. Thereafter, kidneys were benched, and in preparation for the simulated reperfusion period, a cannula was inserted into the renal artery and a Foley catheter was placed into the ureter as previously described for the other groups.

Machine Perfusion—Simulated Reperfusion With Whole Blood

After 4 h of HMP, NMP, or CSS, kidneys were reperfused with autologous whole blood. This was retrieved at the time of retrieval at the abattoir and stored in the refrigerator at 4 °C until used; 500 mL of the pig’s whole blood were diluted 1:1 with 500 mL of 0.9% saline (Baxter), resulting in 1 L of perfusate. A 0.2 g of creatinine and 5000 IU of heparin were added to that. Ringer’s lactate was used to prime the circuit, and 600 mL of the perfusate were used to start the simulated reperfusion phase; the remaining 400 mL were kept on a stirrer at room temperature and used to replace insensitive losses (eg, urinary output of the kidneys). Simulated reperfusion on the RM3 pulsatile perfusion machine was performed for all kidneys for 2 h at 36–37 °C, with a systolic perfusion pressure of 100 mm Hg. Paired kidneys underwent reperfusion at the same time and on the same machine. Unpaired kidneys underwent reperfusion at different times and separately but were analyzed for the same parameters. Oxygen (95% O2/5% CO2) was added to the perfusate through an oxygenator at a flow rate of 0.2 L/min. Temperature, arterial blood pressure, RRI, perfusate flow rates, and urine production rates were recorded at 0, 0.25, 0.5, 1, and 2 h of reperfusion. Furthermore, at 0, 0.25, 1, and 2 h of simulated reperfusion, arterial and venous blood samples were analyzed using an arterial blood gas machine. Further blood and urine samples were stored for further analysis in the laboratory.

Histological Analysis

Two upper pole kidney punch biopsies (4 mm) were taken from each kidney immediately after retrieval and at the end of the simulated reperfusion period. Biopsies were fixed in 4% formalin, dehydrated, and embedded in paraffin wax. Four-micrometer sections were cut and stained with periodic acid-Schiff staining for light microscopy evaluation. Slides were randomized and blinded for 2 independent pathologists, who ranked the slides for markers of tubular injury: tubular brush boarders, tubular dilatation, tubular debris, and epithelial flattening. For each marker, a score of 0–4 was given in 3 fields of 20× magnification, respectively, with 0 describing no abnormalities and 4 describing changes affecting >75% of the sample. Individual scores were then added to result in a score from 0 score points for the best outcome to 48 score points for the worst outcome. Statistical comparison between the groups was performed using a Mann-Whitney U test in Graph Pad Prism.

TUNEL Staining

TUNEL staining was performed on the paraffin-embedded sections for HMP and NMP porcine samples, which were obtained at the end of the simulation of reperfusion. An in situ cell death detection kit (Sigma-Aldrich, Cat. Nr. 11684795910) was used as per manufacturer’s instructions. For each section, 10 fields at a 20× magnification were analyzed. Bright green cells were counted as a positive result for cells in apoptosis, and the mean values were compared between groups. A paired t test was performed for statistical analysis.

Statistical Analysis

Statistical method values are represented as mean and SD. Continuous variables were plotted as levels versus time points for each kidney (eg, RRIs, oxygen consumption, perfusate flow rates) and compared using analysis of variance. Mean levels at specific time points (eg, lactate levels) were compared using a paired t test. Mean values from kidneys after CSS were used for comparison purposes, but because of the smaller sample number (n = 4), these were not included for statistical analyses. Histology and TUNEL staining scores were compared using the Mann-Whitney U test. Data and statistical analyses were performed using GraphPad Prism. All tests were 2-tailed and a P value of ≤0.05 was considered significant.

RESULTS

Direct Comparison of NMP and HMP Using Porcine Kidney Pairs From the Same Donor

Perfusion and Functional Parameters

During 4 h of HMP or NMP, we observed significantly lower RRIs over time in the HMP group compared with the NMP group, with mean values of 0.71 ± 0.27 versus 1.49 ± 0.19 mm Hg/mL/min (P = 0.0075) as shown in Figure 2A. Perfusate flow rates during simulated reperfusion with whole blood in HMP kidneys were significantly higher than in NMP kidneys, with mean values of 46.5 ± 14.9 versus 26.1 ± 7.8 mL/min/100 g (P = 0.042; Figure 2B). Also, oxygen consumption rates of kidneys during that period were significantly higher in kidneys after HMP than in kidneys after NMP, with a mean of 23.7 ± 5.3 versus 11.1 ± 3.2 mL/min/100 g (P = 0.006; Figure 2C).

F2
FIGURE 2.:
Perfusion parameters of paired kidneys during HMP, NMP, and simulated reperfusion. Intrarenal resistance indices (mm Hg/mL/min) during HMP and NMP are visualized in (A), including the changes over time for both groups. Perfusate flow rates in mL/min/100 g during simulated reperfusion for the different treatment groups are shown in (B). Oxygen consumption rates in mL/min/100 g for the different groups at the named time points of measurement during simulated reperfusion are shown in (C), and urinary output rates are shown in mL/min for named time points during the reperfusion period in (D). Statistical analyses to compare NMP and HMP groups were performed using ANOVA; statistical significance: *P < 0.05; **P < 0.005. ANOVA, analysis of variance; CSS, cold static storage (n = 4); HMP, hypothermic machine perfusion (n = 9); NMP, normothermic machine perfusion (n = 9).

Urinary output rates during simulated reperfusion were significantly higher in HMP kidneys than in NMP kidneys, both for continuous measurements with a mean of 4.8 ± 1.7 versus 2.0 ± 1.2 mL/min (P = 0.011; Figure 2D), as well as for the total volume collected over 2 h (120.7 ± 49.5 versus 55.5 ± 13.8 mL; P = 0.041).

Perfusate Analyses

Kidneys after HMP showed a trend toward a higher creatinine clearance at all measured time points during simulated reperfusion, with a mean of 6.87 ± 2.24 versus 2.98 ± 1.43 mL/min for kidneys after NMP (P = 0.26). The mean creatinine clearance for 4 kidneys after CSS was 4.78 ± 0.93 mL/min (Figure 3A). Fractional sodium excretion rates during reperfusion were similar for kidneys after NMP and HMP, with mean values of 0.81% ± 0.18% in the HMP group and 0.81% ± 0.1% in the NMP group (P = 0.99). For kidneys after CSS, the mean fractional sodium excretion rate was 1.01% ± 0.09% (Figure 3B). The mean lactate levels in venous perfusate samples during the reperfusion period were similar, with 6.2 ± 1.23 mmol/L in the HMP group and 6.77 ± 0.87 mmol/L in the NMP group. The mean level in the CSS group during reperfusion was 7.51 ± 1.18 mmol/L. However, lactate levels increased more in the NMP, as well as the CSS group toward the end of reperfusion when compared with HMP. Hence, at time point 120 min after commencement of the reperfusion period, mean lactate levels were highest in the CSS group with mean concentration of 8.62 ± 1.69 mmol/L, followed by the NMP group with a mean concentration of 5.26 ± 2.84 mmol/L and the HMP group with a mean concentration of 3.32 ± 2.11 mmol/L. The difference between the HMP and the NMP group for that time point was significant (P = 0.03; Figure 3C).

F3
FIGURE 3.:
Perfusate analyses for paired kidneys during simulated reperfusion after HMP, NMP, and CSS. Creatinine clearance rates in mL/min for the different treatment groups during simulated reperfusion at measured time points are shown in graph (A). Fractional sodium excretion rates in % are illustrated in (B) and lactate concentrations in mmol/L within respective venous perfusate samples at measured time points are shown in (C). The difference between levels in the HMP (n = 9) and NMP (n = 9) group reached statistical significance at t = 120 min (P = 0.026) after starting simulated reperfusion (*). Statistical analyses to compare NMP and HMP groups were performed using ANOVA and paired t test for defined time points. No statistical significance (no asterisk); *P < 0.05; **P < 0.005. ANOVA, analysis of variance; CSS, cold static storage; HMP, hypothermic machine perfusion (n = 9); NMP, normothermic machine perfusion (n = 9).

Histological Features

Looking at tubular injury, we observed less tubular dilatation and tubular epithelial flattening in the HMP samples (Figure 4A, arrows) when compared with samples within the NMP group (Figure 4B, arrows) and the CSS group (Figure 4C, arrows). Statistical analysis revealed no significant difference between the groups (Figure 4D). Both HMP and NMP resulted in slightly better preservation of healthy histological features than kidneys after reperfusion in the CSS group (Figure 4).

F4
FIGURE 4.:
Histological evaluation after simulated reperfusion. A–C, Photographs of PAS staining (×20 magnification) for samples taken after HMP (A), NMP (B), and CSS (C), after simulated reperfusion. All the slides were ranked blindly by 2 individual pathologists, and a Mann-Whitney U test was performed for statistical analysis (excluding the CSS group) (D). The sample size for kidneys after CSS was only n = 3; 1 sample was lost during processing for histology. We observed less tubular dilatation and tubular epithelial flattening in the HMP group than in the NMP and CSS groups (A–C, arrows). Statistical significance: *P < 0.05; **P < 0.005. CSS, cold static storage (n = 3); HMP, hypothermic machine perfusion (n = 9); NMP, normothermic machine perfusion (n = 9); PAS, periodic acid-Schiff.

TUNEL staining showed a significantly higher mean number of apoptotic cells per magnification field (20×) for kidneys within the NMP group than within the HMP group (3.9 ± 6.17 vs 1 ± 3.05; P = 0.027; Figure 5).

F5
FIGURE 5.:
TUNEL staining after simulated reperfusion. Photographs of TUNEL-stained slides (×20 magnification) of samples after NMP (A) or HMP (B) and simulated reperfusion in paired kidneys and kidneys after CSS were taken and green fluorescent spots, which represented cells undergoing apoptosis, were counted and compared. Ten fields of 20× magnification per slide were counted, and a paired t test was used for statistical analysis of the HMP and the NMP groups. The columns represent the mean values of apoptotic cells after respective treatments. A significant difference in favor of HMP with less apoptotic cells in comparison with the NMP group was noted (P = 0.027). Statistical significance: *P < 0.05; **P < 0.005. CSS, cold static storage (n = 3); HMP, hypothermic machine perfusion (n = 9); NMP, normothermic machine perfusion (n = 9).

Direct Comparison of HMP and NMP in Unpaired Porcine Kidneys

Perfusion and Functional Parameters

A similar comparison focusing on physiological parameters during machine perfusion was conducted using 12 unpaired porcine kidneys that underwent either 4 h of HMP (n = 6) or NMP (n = 6) as control groups for other studies. Thereafter, these kidneys also underwent 2 h of simulated reperfusion using autologous porcine whole blood. Parameters from the previously described 4 kidneys after CSS were again used as a control group to compare findings.

Within this cohort, during 4 h of HMP or NMP, kidneys undergoing HMP showed lower RRIs than kidneys undergoing NMP, with mean values of 0.82 ± 0.5 mm Hg/mL/min in the HMP group versus a mean of 1.31 ± 0.26 mm Hg/mL/min in the NMP group (P = 0.0; Figure 6A). Perfusate flow rates during simulated reperfusion were significantly higher in the HMP group than in the NMP group, with mean flow rates of 47.93 ± 10.1 versus 17.44 ± 2.1 mL/min/100 g (P = 0.02; Figure 6B). As a comparison, the mean perfusate flow rate for kidneys after CSS was 45.85 ± 17.5 mL/min/100 g (Figure 6B). Oxygen consumption rates during simulated reperfusion were significantly higher for kidneys after HMP, with a mean of 25.48 ± 10.17 mL/min/100 g, than after NMP, with a mean of 11.64 ± 1.44 mL/min/100 g (P = 0.014; Figure 6C). Furthermore, we observed higher urinary output rates during reperfusion after HMP than after NMP, with means of 5.37 ± 2.74 versus 3.08 ± 1.62 mL/min (P = 0.14; Figure 6D).

F6
FIGURE 6.:
Functional parameters of unpaired kidneys during HMP, NMP, and simulated reperfusion. Intrarenal resistance indices (mm Hg/mL/min) for kidneys during HMP (n = 6) and NMP (n = 6) are shown in (A). Perfusate flow rates in mL/min/100 g during simulated reperfusion for the different treatment groups including CSS are shown in (B). Oxygen consumption rates during reperfusion in mL/min/100 g for the same cohort and time point are demonstrated in (C). Urine output rates for the different groups are shown in (D). Graphs from the 4 kidneys on CSS were used as control but not included in statistical analyses. Statistical analyses were performed using ANOVA; statistical significance: **P < 0.05; **P < 0.005. ANOVA, analysis of variance; CSS, cold static storage (n = 4); HMP, hypothermic machine perfusion (n = 6); NMP, normothermic machine perfusion (n = 6); RRI, Renal Resistance Index.

Perfusate Analyses

Comparison of perfusate parameters in unpaired kidneys showed similar rates of creatinine clearance for all 3 groups, with a mean of 4.57 ± 0.75 mL/min in the HMP group versus a mean of 4.24 ± 1.02 mL/min in the NMP group (P = 0.98) and for comparison purposes, a mean of 4.78 ± 0.93 mL/min in the CSS group (Figure 7A). The mean fractional sodium excretion rates measured 0.77% ± 0.24% in the HMP group versus 0.92% ± 0.07% in the NMP group (P = 0.3) and 1.01% ± 0.09% in the CSS group during simulated reperfusion (Figure 7B). Mean lactate levels were higher in the NMP group throughout the reperfusion period than in the HMP group, with mean levels of 9.07 ± 0.8 versus 5.47 ± 1.53 mmol/L (P = 0.08). Mean lactate perfusate level in the CSS group during reperfusion was 7.51 ± 1.18 mmol/L. Also, within this group, at time points 60 and 120 min after the start of simulated reperfusion, lactate levels after HMP were significantly lower than after NMP with mean levels of 4.05 ± 2.11 versus 9.88 ± 1.31 mmol/L (P = 0.003) and 3.58 ± 1.57 versus 9.35 ± 2.47 mmol/L (P = 0.007), respectively. The respective mean levels after CSS were 8.43 ± 1.85 mmol/L at 60 min and 8.62 ± 1.69 mmol/L at 120 min after the start of reperfusion (Figure 7C).

F7
FIGURE 7.:
Perfusate analyses for unpaired kidneys during simulated reperfusion after HMP, NMP, and CSS. Creatinine clearance rates in mL/min for the different treatment groups during simulated reperfusion at measured time points are shown in graph (A). Fractional sodium excretion rates in % are illustrated in (B) and lactate concentrations in mmol/L within respective venous perfusate samples at measured time points are shown in (C). The difference between levels in the HMP (n = 6) and NMP (n = 6) group reached statistical significance at t = 60 min (P = 0.003) and t = 120 min (P = 0.007) after starting simulated reperfusion (**). Statistical analyses to compare NMP and HMP groups were performed using ANOVA and paired t test for defined time points. No statistical significance (no asterisk); *P < 0.05; **P < 0.005. ANOVA, analysis of variance; CSS, cold static storage; HMP, hypothermic machine perfusion (n = 9); NMP, normothermic machine perfusion (n = 9).

Direct Comparison of HMP and NMP Combining Paired and Unpaired Kidneys for Each Group

Perfusion and Functional Parameters

Statistical analysis of data from 18 paired and 12 unpaired kidneys did not change the overall results described above but increased the statistical significance for some of them. With 15 porcine kidneys per group, the previously observed lower renal resistance indices during the HMP or NMP phase became more significant, with mean values of 0.47 ± 0.35 mm Hg/mL/min for HMP versus 1.43 ± 0.23 mm Hg/mL/min for NMP kidneys (P = 0.0008; Figure 8A). Perfusate flow rates were higher within the HMP group during simulated reperfusion when compared with the NMP group (mean of 46.24 ± 12.49 versus 26.16 ± 4.57 mL/min/100 g; P = 0.0051; Figure 8B). Mean oxygen consumption rates were 22.71 ± 6.27 mL/min/100 g in the HMP group versus 11.83 ± 1.29 mL/min/100 g in the NMP group (P = 0.0016; Figure 8C). As a comparison, for the 4 kidneys being kept on CSS, the mean oxygen consumption rate was 17.93 ± 8.32 mL/min/100 g (Figure 8C). Urinary output rates during reperfusion for this cohort were 5.31 ± 2.07 mL/min in the HMP group versus 2.45 ± 1.19 mL/min (P = 0.002; Figure 8D). As a comparison, the mean urinary output rates for kidneys after CSS were 2.35 ± 1.25 mL/min.

F8
FIGURE 8.:
Functional parameters in combined paired + unpaired kidneys during HMP, NMP, and simulated reperfusion (n = 15). Intrarenal resistance indices (mm Hg/mL/min) for kidneys during HMP (n = 15) and NMP (n = 15) are shown in (A). Perfusate flow rates in mL/min/100 g during simulated reperfusion for the different treatment groups including CSS (n = 4) are shown in (B). Oxygen consumption rates during reperfusion in mL/min/100 g for the same cohort and time are demonstrated in (C). Urine output rates for the different groups are shown in (D). Graphs from 4 kidneys after CSS were used as control but not included in statistical analyses. Statistical analyses were performed using ANOVA, statistical significance: *P < 0.05; **P < 0.005. ANOVA, analysis of variance; CSS, cold static storage (n = 4); HMP, hypothermic machine perfusion (n = 15); NMP, normothermic machine perfusion (n = 15); RRI, Renal Resistance Index.

Perfusate Analyses

With the combined number of kidneys (n = 15/HMP or NMP groups), the creatinine clearance rates in the HMP group were similar when compared with the NMP group, with mean values of 5.72 ± 1.22 versus 3.62 ± 0.74 mL/min (P = 0.52). The mean value for kidneys after CSS was 4.78 ± 0.93 mL/min (Figure 9A). Sodium excretion rates were similar between both groups, with mean values of 0.81% ± 0.1% in the HMP group versus 0.81% ± 0.2% in the NMP group (P = 0.84) and a mean value of 1.01% ± 0.09% for kidneys after CSS (Figure 9B). Overall, lactate concentrations throughout the simulated reperfusion phase were highest in perfusate samples after NMP, with mean level of 7.78 ± 0.46 mmol/L, followed by perfusate samples from kidneys after CSS with a mean of 7.51 ± 1.17 mmol/L, and HMP with a mean of 5.88 ± 1.32 mmol/L. The difference between lactate perfusate levels during kidneys after HMP and NMP were not significantly different when analyzed as mean levels for the whole perfusion period, with P = 0.09. However, at time point 120 min after commencement of simulated reperfusion, perfusate samples from kidneys after CSS measured the highest lactate levels, with mean concentration of 8.62 ± 1.69 mmol/L, followed by kidneys after NMP with a mean of 5.02 ± 2.19 mmol/L and kidneys after HMP, with a mean of 3.99 ± 1.82 mmol/L. The difference between HMP and NMP at that time point for this cohort was again statistically significant (P = 0.03; Figure 9C).

F9
FIGURE 9.:
Perfusate analyses for combined paired + unpaired kidneys during simulated reperfusion after HMP, NMP, and CSS. Creatinine clearance rates in mL/min at measured time points for combined paired + unpaired kidneys after HMP (n = 15) and NMP (n = 15), and for comparison purposes after CSS (n = 4) during simulated reperfusion are shown in graph (A). Fractional sodium excretion rates for the same groups in % are illustrated in (B). Lactate concentrations in mmol/L within respective venous perfusate samples at measured time points are shown in (C). The difference between levels in the HMP (n = 15) and NMP (n = 15) group reached statistical significance at t = 120 min (P = 0.03). Statistical analyses performed using ANOVA and paired t test for defined time points. No statistical significance (no asterisk); *P < 0.05; **P < 0.005. ANOVA, analysis of variance; CSS, cold static storage; HMP, hypothermic machine perfusion (n = 15); NMP, normothermic machine perfusion (n = 15).

DISCUSSION

Research groups around the world are working on devices and protocols that allow for continuous periods of ex vivo organ perfusion with an aim to preassess, precondition, and repair organs. Many groups focus on either NMP or HMP and conduct studies with kidneys kept on CSS as a respective control group. So far, only 6 studies have aimed to directly compare HMP and NMP. Only in 1 of those studies, a comparison of kidney pairs from the same porcine donor was performed.24 The outcomes were not very conclusive but pointed toward superiority of NMP. However, experimental numbers were small, with 6 kidneys per study group and different perfusion devices, with a centrifugal pump for NMP and a pulsatile pump for HMP, used. Kidneys were perfused after cold ischemia time of 2 h for up to 16 h, which could pose a problem in terms of fluid and electrolyte replacement during NMP.25,26 Another study in a porcine autotransplant model showed superiority of NMP over HMP with respect to early kidney function; however, the study was performed using single kidneys from different donors and on different perfusion devices for the different modalities.27 Three further studies compared different modalities of organ preservation with each other. One highlighted the beneficial effect of NMP as an end-ischemic treatment after prolonged times of CSS,28 and another study demonstrated similar positive effects of a short period of NMP after HMP versus a period of HMP alone.29 Contrary to those findings, Darius et al30 described a continuous oxygenated period of HMP from time of retrieval to be superior to all end-ischemic preservation strategies, including 2 h of NMP, when looking at early renal function after transplantation. Furthermore, a study performed by Blum et al31 described comparable functional outcomes during a period of simulated reperfusion for kidneys after an 8-h period of HMP or NMP, with greater loss of tubular integrity after NMP, which is concordant with our findings. However, in 3 of the mentioned studies,28-30 perfusion periods of the different preconditioning methods varied between the groups, and in none of them24,27-31 was the reperfusion phase performed with both kidneys on the same device and under the exact same circumstances, as shown in our study. This was possible due to the option of perfusing 2 kidneys on one RM3 machine. Therefore, our study fills a gap, and our results suggest that porcine DCD kidneys after a prolonged period of CSS benefit more from a 4-h period of HMP than a 4-h period of NMP when looking at physiological parameters during 2 h of hemoreperfusion.

None of the previously mentioned studies attributed to the fact that mostly, HMP runs on a pulsatile mode, whereas most NMP setups use centrifugal nonpulsatile perfusion pumps. At least for HMP, pulsatility was described as a beneficial factor in terms of reducing inflammation within the endothelium and functional outcomes of perfused porcine kidney grafts in a study performed by Gallinat et al.32 Also, in our study, the RM3 machine was used for both perfusion modalities, even though that system was not designed for NMP in the first place; however, we did not observe hemolysis during NMP, which was confirmed by unchanged hemoglobin values throughout. Furthermore, we acknowledge that there was a difference in systolic perfusion pressures during the preconditioning period, with pressures between 40 and 50 mm Hg during HMP and 50 and 60 mm Hg during NMP. However, recommended pressures for the 2 modalities in clinical applications are different per se, and by keeping systolic pressures during NMP at the lowest acceptable levels, we aimed to keep these differences to a minimum. Hence, our study represents the first preclinical experimental comparison between HMP and NMP after a significant WIT as well as a prolonged period of cold ischemia, in which external influencing factors were kept to a minimum. We aimed to expose kidneys coming from young, healthy, slim donor pigs to as much ischemia injury as possible to assess the influence of both perfusion modalities better. Therefore, we chose a WIT of 25 min and a cold ischemia time of 24 h, which is at the limit of acceptable timings in clinical transplantation. Four single kidneys were kept on CSS for 28 h and were then reperfused separately for 2 h, respectively. Reperfusion parameters for those 4 kidneys were very similar every time and were used for the purpose of comparison of outcomes after HMP and NMP with the current gold standard in clinical transplantation. This study, however, aimed to compare HMP and NMP directly and therefore, findings after CSS were not included in statistical analyses.

Perfusion parameters during simulated reperfusion were significantly better in kidneys during and after HMP than in kidneys undergoing NMP. Especially the RRIs after HMP were lower than after NMP. Eprostenol was applied in hourly doses, as recommended by previously published NMP protocols,23 to support the microcirculation within the kidney. The red cell–based perfusate was fully heparinized to prevent formation of blood clots or microembolism and yet, some disturbance to the microcirculation might have occurred, resulting in higher values for the RRI than in kidneys undergoing HMP. Contrary to most perfusates used for NMP, perfusates commonly used for HMP do not contain any third party–derived red blood cells. This might have immunological advantages with regards to potential HLA exposure of kidney transplant recipients in the setting of clinical transplantation. Also, with regards to clinical transplantation, the cost efficiency of HMP when compared with NMP might have an influence on which method will be more commonly used in the future. HMP poses a straightforward method of organ preservation, which can easily be performed by junior medical personnel. HMP solutions are commercially available and ready to use after storage in the fridge. Contrary to this, for NMP, no specific perfusion machine nor commercially available perfusates are available for kidney preservation, and therefore, it requires a sophisticated technical setup as well as highly trained medical and engineering personnel to perform this form of kidney preservation.

At the end of the simulated reperfusion period, lactate levels in the perfusate were highest in kidneys after CSS, followed by kidneys after NMP and lowest for kidneys after HMP, suggesting the lowest degree of hypoxic damage for the latter group. Creatinine clearance, as well as fractional sodium excretion, was similar between both groups, suggesting equal tissue viability. To further clarify and validate our results, longer reperfusion periods, or more ideally, in vivo reperfusion studies are necessary. In this ex vivo model with limited amounts of autologous whole blood for reperfusion, a longer period was not possible.

By adding unpaired kidneys into our study (n = 6/NMP and HMP group), we increased the numbers within the treatment groups to further validate our data. Unpaired kidneys were kidneys from different donor pigs, respectively, which underwent the same treatment as per respective treatment arm (HMP or NMP) but at different times. Treatment arms with 6 kidneys per group were compared in the same way as for the paired kidneys. The main interest herein was whether findings were similar for organs coming from different donors. Indeed, we could confirm findings, and adding unpaired kidneys within respective treatment arms to the pool of paired kidneys increased the significance of findings, especially looking at functional parameters: RRIs during the preconditioning period were significantly lower in kidneys undergoing HMP, and during simulated reperfusion, differences in perfusate flow, oxygen consumption, and urinary output rates were more significant, supporting initial findings in the smaller (paired) cohort.

Histologically, kidneys after HMP showed slightly better results after reperfusion, with good preservation of tubular structures, followed by kidneys after NMP and CSS. This could point toward a beneficial effect of both perfusion modalities in comparison with leaving the organs on ice for a prolonged period. Most likely, however, histological changes will only become clearer after longer times of simulated reperfusion. TUNEL staining, which is probably more informative at that early stage, showed significantly higher apoptosis rates within kidneys after NMP than after HMP, underlining the better histological results for the HMP group.

Limitations of this study are the fact that this was an ex vivo preclinical and experimental study. We acknowledge that 2 h of simulated reperfusion is relatively short to explore long-term effects of either perfusion modality and that one of the 2 modalities might not exclusively be the way forward. An in vivo study followed by a study investigating the effects of different machine perfusion modalities in human kidneys would definitely be the next step to elucidate our knowledge further. It might become necessary in the future to combine the 2 methods depending on an organ’s background and “injury” to aim for an organ-tailored approach.

REFERENCES

1. Rabbat CG, Thorpe KE, Russell JD, et al. Comparison of mortality risk for dialysis patients and cadaveric first renal transplant recipients in Ontario, Canada. J Am Soc Nephrol. 2000;11:917–922.
2. Brat A, Pol RA, Leuvenink HGD. Novel preservation methods to increase the quality of older kidneys. Curr Opin Organ Transplant. 2015;20:438–443.
3. Rao PS, Ojo A. The alphabet soup of kidney transplantation: SCD, DCD, ECD—fundamentals for the practicing nephrologist. Clin J Am Soc Nephrol. 2009;4:1827–1831.
4. Metzger RA, Delmonico FL, Feng S, et al. Expanded criteria donors for kidney transplantation. Am J Transplant. 2003;3(suppl 4):114–125.
5. Ojo AO, Hanson JA, Meier-Kriesche H, et al. Survival in recipients of marginal cadaveric donor kidneys compared with other recipients and wait-listed transplant candidates. J Am Soc Nephrol. 2001;12:589–597.
6. Morrissey PE, Monaco AP. Donation after circulatory death: current practices, ongoing challenges, and potential improvements. Transplantation. 2014;97:258–264.
7. Summers DM, Johnson RJ, Hudson A, et al. Effect of donor age and cold storage time on outcome in recipients of kidneys donated after circulatory death in the UK: a cohort study. Lancet. 2013;381:727–734.
8. Weber M, Dindo D, Demartines N, et al. Kidney transplantation from donors without a heartbeat. N Engl J Med. 2002;347:248–255.
9. Saidi RF, Elias N, Kawai T, et al. Outcome of kidney transplantation using expanded criteria donors and donation after cardiac death kidneys: realities and costs. Am J Transplant. 2007;7:2769–2774.
10. Pascual M, Theruvath T, Kawai T, et al. Strategies to improve long-term outcomes after renal transplantation. N Engl J Med. 2002;346:580–590.
11. Jochmans I, Akhtar MZ, Nasralla D, et al. Past, present, and future of dynamic kidney and liver preservation and resuscitation. Am J Transplant. 2016;16:2545–2555.
12. Moers C, Smits JM, Maathuis MHJ, et al. Machine perfusion or cold storage in deceased-donor kidney transplantation. N Engl J Med. 2009;360:7–19.
13. Deng R, Gu G, Wang D, et al. Machine perfusion versus cold storage of kidneys derived from donation after cardiac death: a meta-analysis. PLoS One. 2013;8:e56368.
14. O’Callaghan JM, Morgan RD, Knight SR, et al. Systematic review and meta-analysis of hypothermic machine perfusion versus static cold storage of kidney allografts on transplant outcomes. Br J Surg. 2013;100:991–1001.
15. Jochmans I, Moers C, Smits JM, et al. Machine perfusion versus cold storage for the preservation of kidneys donated after cardiac death: a multicenter, randomized, controlled trial. Ann Surg. 2010;252:756–764.
16. National Institute for Health and Care Excellence. Machine Perfusion Systems and Cold Static Storage of Kidneys from Deceased Donors. National Institute for Health and Care Excellence. 2009. Available at https://www.nice.org.uk/guidance/ta165/chapter/1-Guidance. Accessed April 13, 2020.
17. Nicholson ML, Hosgood SA. Renal transplantation after ex vivo normothermic perfusion: the first clinical study. Am J Transplant. 2013;13:1246–1252.
18. Hosgood SA, Nicholson ML. The first clinical case of intermediate ex vivo normothermic perfusion in renal transplantation. Am J Transplant. 2014;14:1690–1692.
19. Hosgood SA, Nicholson ML. First in man renal transplantation after ex vivo normothermic perfusion. Transplantation. 2011;92:735–738.
20. Hosgood SA, Nicholson ML. The evolution of donation after circulatory death donor kidney repair in the United Kingdom. Curr Opin Organ Transplant. 2018;23:130–135.
21. Hosgood SA, Thompson E, Moore T, et al. Normothermic machine perfusion for the assessment and transplantation of declined human kidneys from donation after circulatory death donors. Br J Surg. 2018;105:388–394.
22. Hosgood SA, Nicholson ML. Ex vivo normothermic perfusion of declined human kidneys after inadequate in situ perfusion. Am J Transplant. 2014;14:490–491.
23. Hosgood SA, van Heurn E, Nicholson ML. Normothermic machine perfusion of the kidney: better conditioning and repair? Transpl Int. 2015;28:657–664.
24. Metcalfe MS, Waller JR, Hosgood SA, et al. A paired study comparing the efficacy of renal preservation by normothermic autologous blood perfusion and hypothermic pulsatile perfusion. Transplant Proc. 2002;34:1473–1474.
25. Weissenbacher A, Lo Faro L, Boubriak O, et al. Twenty-four-hour normothermic perfusion of discarded human kidneys with urine recirculation. Am J Transplant. 2019;19:178–192.
26. Weissenbacher A, Voyce D, Ceresa CDL, et al. Urine recirculation improves hemodynamics and enhances function in normothermic kidney perfusion. Transplant Direct. 2020;6:e541.
27. Urbanellis P, Hamar M, Kaths JM, et al. Normothermic ex vivo kidney perfusion improves early DCD graft function compared with hypothermic machine perfusion and static cold storage. Transplantation. 2020;104:947–955.
28. Bagul A, Hosgood SA, Kaushik M, et al. Experimental renal preservation by normothermic resuscitation perfusion with autologous blood. Br J Surg. 2008;95:111–118.
29. Hosgood SA, Barlow AD, Yates PJ, et al. A pilot study assessing the feasibility of a short period of normothermic preservation in an experimental model of non heart beating donor kidneys. J Surg Res. 2011;171:283–290.
30. Darius T, Gianello P, Vergauwen M, et al. The effect on early renal function of various dynamic preservation strategies in a preclinical pig ischemia-reperfusion autotransplant model. Am J Transplant. 2019;19:752–762.
31. Blum MF, Liu Q, Soliman B, et al. Comparison of normothermic and hypothermic perfusion in porcine kidneys donated after cardiac death. J Surg Res. 2017;216:35–45.
32. Gallinat A, Fox M, Lüer B, et al. Role of pulsatility in hypothermic reconditioning of porcine kidney grafts by machine perfusion after cold storage. Transplantation. 2013;96:538–542.
Copyright © 2021 Wolters Kluwer Health, Inc. All rights reserved.