Custodiol® HTK (histidine-tryptophan-ketoglutarate; Dr. Franz Köhler Chemie GmbH, Bensheim, Germany) solution preserved kidneys arrived at the transplant center statically cold stored. Grafts were routinely placed on the LifePort kidney transporter (Organ Recovery Systems, Itasca, IL, USA) immediately after arrival. After deemed untransplantable, the organ was taken off the hypothermic machine perfusion (HMP) device and prepared for connection to the NMP circuit. In one case, no HMP was performed due to logistical reasons and the kidney graft was perfused normothermically after routine back table preparation and static cold storage in Custodiol® HTK.

For NMP, the Organ Assist Kidney Assist (Organ Assist BV, Groningen, Netherlands) device was used. The renal artery was cannulated with a 20-Fr (KTX 1, KTX 2) or 16-Fr (KTX 3, KTX 4, KTX 5) straight perfusion cannula (Infusion, Warsaw, Poland). The ureter was cannulated with the provided tubing. A three-way valve was included in the tubing to allow for urine collection and recirculation into the reservoir. The disposable set was adapted by implementing an in-line blood gas analyzer (CDI500, Terumo Medical Corporation, Tokyo, Japan). Perfusate temperature was set at 37°C. Oxygenation of the circuit was facilitated by manual regulation of air (21% oxygen) and CO₂. The perfusion circuit was primed with three units of packed red blood cells (RBCs) of universal donor blood, resuspended in 1000 mL 5% human albumin solution (Albunorm®, Octapharma, Lachen, Switzerland), resulting in a total perfusate volume of approximately 1800 mL. The protocol was adapted from the protocol published by Weissenbacher et al. [7] Before connecting the kidney, the perfusate was supplemented with 750 mg cefuroxime (Sandoz, Basel, Switzerland), 10 mL calcium gluconate 10% (B.Braun, Melsungen, Germany), and 8000 IE enoxaparin (Lovenox®, Sanofi, Paris, France). For pH adjustment prior to initiation of NMP, 10 mL of sodium bicarbonate 8.4% (Fresenius Kabi, Bad Homburg, Germany) were added to the perfusate to achieve a pH level >7.0. Immediately after perfusion start, 5 mL verapamil (Mylan, Vienna, Austria) was administered directly into the arterial line. Kidneys were perfused with a median arterial pressure (MAP) of 90 mmHg. MAP was lowered to 80 mmHg in case flow reached 900 mL/min due to flow restrictions of the device (KTX 1 at 40 h and KTX 2 at 24 h into the perfusion). For glucose and electrolyte monitoring, blood gas analyzer (BGA) measurements (ABL800Flex, Drott Medizintechnik GmbH, Wiener Neudorf, Austria) were performed throughout the perfusion and total parenteral nutrition (Nutriflex® plus, containing 0.15 g/mL glucose) was administered once perfusate glucose levels dropped below 70 mg/dL. Data from the first perfused kidney have already been published as a proof of principle study [8]. To give a better overview on perfusion dynamics and explore the potential of kidney NMP in graft assessment, we included this kidney in this manuscript.

Perfusate and urine samples were obtained throughout the duration of perfusion at hrs 1, 6, 12, 20, 24, 40, and 48. Samples were analyzed upfront by blood gas analyzer and institutional biochemistry laboratory as well as stored after centrifugation at 15,000 G for 15 min at −80°C.

In all kidney grafts, a zero-biopsy was performed and the Remuzzi score was assessed by the on-call pathologist. Follow-up biopsies were taken after 24 h of NMP and upon reaching the endpoint at 48 h. Hemodynamic perfusion parameters were recorded at corresponding timepoints.

Kidney injury markers (Luminex Performance Human Kidney Biomarker Panel [6-Plex], #FCSTM16-06, R&D Systems, Minneapolis, United States) and cytokine (Luminex Performance Human High Sensitivity Cytokine Panel A [12-Plex], #FCSTM09-12, R&D Systems, Minneapolis, United States) levels were measured in perfusate samples stored at −80°C. Sample dilution, processing and analysis were carried out according to manufacturer’s instruction.

Results are expressed as median and interquartile range (IQR) or range. Mann-Whitney test, Kruskal Wallis, Friedman test corrected with Dunn’s multiple comparison test and Wilcoxon matched-pairs signed rank test were used for non-normal distributed data. All tests were two-sided and a p-value of <0.05 was considered statistically significant. Prism GraphPad 9.0 (GraphPad Inc., San Diego, CA, USA) was used for all statistical tests.

Donor age ranged from 41 to 68 (median 62) years (yrs). Four out of five kidney grafts were from DBD donors, only one (KTX 3) was from a DCD donor. Causes of death were all cardiovascular events. All but one donor had a normal ranged serum creatinine level (median 1.06 mg/dL) before organ retrieval. One donor (KTX 4) was on veno-venous extracorporeal membrane oxygenation (vv-ECMO) and renal replacememt therapy (RRT) due to severe aspiration and concomitant sepsis. All donors had urine output before organ retrieval (median 85 mL/h). Two donors had a history of hypertension, none of the donors were diabetic. For details see Table 1.

Cold ischemic times ranged from 11.5 to 28 (median 19.5) hrs. All but one kidney were perfused hypothermically after being transported to our unit on static cold storage (SCS) using the Lifeport® device for a median of 7 h. Reasons for discard were malignancy in the contralateral kidney, poor organ quality and poor perfusion. Remuzzi scores ranged between 0 and 5 (median 3). A detailed overview can be found in Table 2.

Arterial flow was stable throughout the perfusion of all kidney grafts (Figure 1A) irrespective of reason for discard. In the first three perfused kidneys, the median arterial flow exhibited higher values, measuring 841, 775, and 721 mL/min, in contrast to the second pair of kidneys, which recorded flow rates of 405 and 406 mL/min, respectively (p < 0.001). Correspondingly, resistance indices (Figure 1B) were lower in the first three kidneys compared to the last two (p < 0.001). Median hemoglobin and hematocrit levels ranged between 9.5 and 7.9 g/dL, and 0.23 and 0.29 L/L (Table 3; Figures 1C, D). Free hemoglobin increased over time (Figure 1E). Median free hemoglobin levels ranged from 74 to 144 mg/dL (Table 3). Anti-FXa activity ranged between 2.7 and 4.1 UI/mL. Levels were slowly decreasing over time and only KTX2 demonstrated a clear clearance during perfusion (Figure 1F).

Median (IQR) venous pO₂ und pCO₂ levels were 113 (125–107) mmHg and 37.8 (39.6–36.1) mmHg for KTX 1, 79.9 (86.9–78.5) mmHg and 34.6 (37.8–31.2) mmHg for KTX 2, 81.9 (107–76.2) mmHg and 36.3 (40.7–29.1) mmHg for KTX 3, 90.3 (97.4–79.7) mmHg and 13.4 (22.2–10.8) mmHg for KTX 4, and 80.3 (89.2–73.8) mmHg and 24.8 (32.5–19.8) mmHg for KTX 5.

With 35–69 mL of summative Nutriflex supplementation, glucose concentrations between 50 and 150 mg/dL were achieved (Figure 2A). Lactate levels ranged between 179 and 89.5 mg/dL. Even tough median lactate levels were similar between groups (p = 0.063), lactate dynamics differed. While lactate was cleared in KTX 1 and 5, it remained stable in KTX 2 and increased over the duration of perfusion in KTX 3 and 4 (Figure 2B). Supplementation of 15–40 mL sodium bicarbonate were necessary to achieve median pH levels between 7.09 and 7.34 (Table 3; Figure 2C). Despite a uniform basic perfusate formulation, kidneys exhibited different electrolyte levels in the perfusate (Table 3; Figure 3). Perfusate sodium and chloride levels increased over the first 24 h of perfusion and stabilized afterwards (Figures 3A, C). Perfusate potassium levels, on the contrast, decreased over the course of the first 24 h in all cases. Thereafter it further decreased, stabilized or increased depending on potassium loss via urine sampling (Figure 3B).

All kidney grafts had urine output throughout the perfusion. Median urine output was 0.43 mL/min at the end of perfusion. Median urine sodium content was significantly lower than in the perfusate in all perfused kidney grafts (Figure 3D, p = 0.031). In contrast, median chloride (Figure 3F, p = 0.031) and potassium (Figure 3E, p < 0.001) levels differed only in KTX 2.

Perfusate samples were analyzed for generic and kidney specific injury markers (Figure 4) as well as for the presence of pro- and anti-inflammatory markers (Figure 5).

Generic injury markers aspartate aminotransferase (AST) and lactate dehydrogenase (LDH) steadily increased over time (Figure 4). Significant differences in enzyme release were noticed with KTX 3 displaying the highest absolute enzyme levels (p < 0.001). Chemokine CXCL10 (IP-10, p < 0.001), trefoil factor 3 (TFF3, p = 0.010) that is found upregulated in chronic kidney disease (CKD) and the pro-inflammatory cytokine granulocyte macrophage colony-stimulating factor (GM-CSF, p < 0.001) all displayed highest levels in KTX 3 and 5 (Figure 4). Neutrophil gelatinase-associated lipocalin (NGAL) is released upon tubular injury into the perfusate. Despite a numerical higher level in KTX 3, similar levels between the five perfused kidneys (Figure 4, p = 0.080) were seen. Similarly, a comparable level of Cystatin C was measured in all groups. Clusterin, a protein released from multiple cell types upon injury with cytoprotective properties, was found in abundance in all groups and displayed a similar pattern over time in all kidney grafts. Osteopontin (OPN), a widely expressed protein during inflammation and arteriosclerosis, was found to be, unlike the other injury markers, highest in KTX 1 and KTX 2.

Highest levels of proinflammatory markers TNFa, IFNy, IL-1b, IL-2, IL-4 IL-5, IL-6 and IL-12p70 were seen in KTX 3 and KTX 4, respectively (Figure 5). In the histological work-up, KTX 4 was found to suffer from fungal contamination. A steep increase of these cytokines especially towards the end of the 48 h perfusion duration might reflect the presence of fungal contamination (Figure 5). Anti-inflammatory IL-10 expression spiked after 6 and 20 h in DBD and DCD organs, respectively. In addition, IL-10 levels increased after fungal contamination in KTX 4.

All kidneys had a good reperfusion. Only the two poorly perfused kidneys, KTX 3 and KTX 4, showed some purple cortical areas (Figure 6). These vanished over the first 6 h of perfusion and further homogenous reperfusion was achieved until the end of the 48 h experiment. Despite favorable perfusion dynamics, macroscopic appearance of KTX 2 was the worst in our series, and KTX 4, despite a favorable appearance, was overall performing poorly.

All kidneys displayed unspecific acute tubular injury that showed moderate progression over the course of the 48 h of perfusion. No glomerular and/or tubular necrosis. In biopsies of kidneys perfused for 48 h, a small fraction (2%) of glomeruli had signs of acute thrombotic microangiopathy (Figure 6B). No evidence of thrombotic events were found at earlier timepoints. Focal oidia of the Candida type were present in tubules and glomeruli of KTX 3.