Efficacy of liver graft washout as a function of the perfusate, pressure, and temperature


  • The authors have no financial disclosures or any other conflicts of interest to disclose.

Address reprint requests to Thomas M. van Gulik, M.D., Ph.D., Department of Surgery (Surgical Laboratory), Academic Medical Center, University of Amsterdam, Meibergdreef 9, 1105 AZ Amsterdam, the Netherlands. Telephone: +31 20 5665572; FAX: +31 20 6976621; E-mail: t.m.vangulik@amc.uva.nl


Donor graft washout can be impaired by colloids in organ preservation solutions that increase the viscosity and agglutinative propensity of red blood cells (RBCs) and potentially decrease organ function. The colloid-induced agglutinative effects on RBCs and RBC retention after liver washout with Ringer's lactate (RL), histidine tryptophan ketoglutarate solution, University of Wisconsin solution, and Polysol were determined as a function of the washout pressure (15 or 100 mm Hg) and temperature (4 or 37°C) in a rat liver washout model with 99mTc-pertechnetate–labeled RBCs. Colloids (polyethylene glycol in Polysol and hydroxyethyl starch in University of Wisconsin) induced RBC agglutination, regardless of the solution's composition. RL was associated with the lowest degree of 99mTc-pertechnetate–labeled RBC retention after simultaneous arterial and portal washout at 37°C and 100 mm Hg. RL washout was also associated with the shortest washout time. A single portal washout with any of the solutions did not result in differences in the degree of RBC retention, regardless of the temperature or pressure. In conclusion, no differences were found in portal washout efficacy between colloidal solutions, histidine tryptophan ketoglutarate, and RL. Simultaneous arterial and portal washout with RL at 37°C and 100 mm Hg resulted in the least RBC retention and the shortest washout time. Liver Transpl 19:843-851, 2013. © 2013 AASLD.


hydroxyethyl starch


high-power field


polyethylene glycol–free histidine tryptophan ketoglutarate


polyethylene glycol–containing histidine tryptophan ketoglutarate


polyethylene glycol


polyethylene glycol–free Polysol


polyethylene glycol–containing Polysol




red blood cell


Ringer's lactate


region of interest


hydroxyethyl starch–containing University of Wisconsin

The procurement of a liver graft is preceded by organ washout with a preservation solution (perfusate) under hypothermic (4°C) conditions to prepare the organ for storage.[1] The completeness of the washout has been proposed to be an important determinant for liver graft function.[2, 3] During washout, perfusates prevent acidosis by providing a high buffer capacity, and they counteract cell swelling or hypo-osmotic extravasation of fluids because of added impermeants and/or colloids. Histidine tryptophan ketoglutarate solution contains mannitol as an impermeant. Although mannitol may not strictly act as an impermeant in the liver,[4, 5] it yields similar or improved washout results in comparison with hydroxyethyl starch–containing University of Wisconsin (UW+) solution.[6, 7] UW+ and polyethylene glycol–containing Polysol (PS+), which is an experimental perfusate, contain hydroxyethyl starch (HES)[8] and polyethylene glycol (PEG),[1] respectively, as colloids.

The addition of colloids to a perfusate increases the viscosity of the solution, which is exacerbated by the low temperatures at which a perfusate is used to cool an organ.[9] In addition, colloids have a hyperaggregatory effect on red blood cells (RBCs) that results in rouleaux formation.[10] Consequently, the use of colloidal perfusates during a washout procedure may induce endothelial damage and lead to occlusion of the microvasculature and corollary no-reflow phenomena.[11] Furthermore, the washout procedure is regarded as an important part of the organ preservation protocol inasmuch as the primary goals of the initial washout are to clear the organ of blood and to reduce its core temperature when hypothermic storage follows. Higher washout pressures (100 mm Hg) or normothermic temperatures (37°C) might prevent microcirculatory RBC entrapment and ensure proper distribution of the perfusate through the liver.[2, 3, 12-15] However, high pressures are associated with increased shear stress, which can lead to endothelial lining perturbations or even affect cellular energy metabolism.[16, 17] These potential complications during washout have been successfully circumvented by the application of Ringer's lactate (RL) as a washout and short-term preservation solution.[18, 19]

Because the aforementioned effects of the perfusate temperature and colloid-induced RBC agglutination during liver washout had not been clarified, this study was designed to quantify RBC retention after liver washout at different temperatures and pressures. First, the agglutination of RBCs was assessed microscopically in perfusates in the presence or absence of colloids. Second, a rat liver washout model was employed to determine the washout efficacy of several perfusates through scintigraphic analysis of intrahepatically retained, 99mTc-labeled RBCs after hypothermic or normothermic washout at a low or high washout pressure. This study revealed that the extent of intrahepatic RBC retention was not affected by the applied pressure, solution, colloidal content or type, temperature, or washout time during liver washout via the portal vein. Simultaneous arterial and portal washout with RL at 37°C and 100 mm Hg resulted in the least RBC retention and the shortest washout time.



Two types of perfusates were used: colloid-containing perfusates and colloid-free perfusates. The colloid-free perfusates included RL (Baxter, Deerfield, IL); polyethylene glycol–free histidine tryptophan ketoglutarate (HTK−; Dr F Köhller-Chemie, Bensheim, Germany); and a custom-produced batch of polyethylene glycol–free Polysol (PS−; Organoflush, Amsterdam, the Netherlands), which is an experimental perfusate. The colloid-containing perfusates were polyethylene glycol–containing histidine tryptophan ketoglutarate (HTK+) enriched with PEG-35 (20 g/L; Sigma-Aldrich), UW+ solution (Fresenius Hemocare, Emmer-Compascuum, the Netherlands), and standard PS+ (containing PEG-35).

Assessment of RBC Agglutination With Brightfield Microscopy

The RBCs that were used were prepared from fresh platelet- and leukocyte-poor human RBC concentrates (packed cells; Sanquin, Amsterdam, the Netherlands) suspended in saline-adenine-glucose-mannitol medium in conformity with the blood bank's donor protocol. The packed cells contained a mix of RBCs from at least 2 donors. RBCs were added to solutions approximately 5 minutes before imaging. A 2-μL RBC suspension was prepared 5 minutes before imaging in RL, HTK−, PS+, or UW+ with a hematocrit of 20% (Advia 2120, Siemens, Munich, Germany) and transferred to a microscope slide for imaging. Additionally, RBCs were added to HTK+ and PS− to demonstrate colloid-dependent RBC agglutination. The preparations were visualized through a 100× oil immersion objective on a Leica DMBL microscope (Leica Microsystems, Wetzlar, Germany) equipped with a Leica DC200 charged coupled device camera that was controlled with QWin software (Leica Microsystems).

In Vivo Radioactive Labeling of RBCs

The institute's animal ethics committee approved all animal experiments (BEX102507), and animals were treated in accordance with the Guide for the Care and Use of Laboratory Animals (National Institutes of Health). Male Wistar rats weighing 254 ± 27 g were anesthetized with an intraperitoneal injection (0.27 mL/100 g of body weight) of Hypnorm (10 mg/mL fluanisone and 0.515 mg/mL fentanyl citrate; VetaPharma, Leeds, United Kingdom), Dormicum (5 mg/mL; Actavis Group, Zug, Switzerland), and water in a 1:1:2 ratio.

RBCs were labeled in vivo with the 99mTc-pertechnetate procedure adapted for small laboratory animals.[20] Briefly, 50 μg of freshly prepared pyrophosphate (PYP; Technescan, Covidien, Petten, the Netherlands) in a 0.2-mL saline solution was injected intravenously into the tail vein, and this was followed by 40 MBq of 99mTc-pertechnetate (Ultra-Technekow, Covidien) in a 0.2-mL saline solution after 30 minutes.

Liver Washout

After 10 minutes of 99mTc-pertechnetate circulation, washout was performed according to a previously described procedure.[21] After mobilization of the liver and intestines, a blood sample was obtained from the inferior caval vein (radioactive RBC reference) and weighed on a precision scale (AB204, Mettler Toledo, Greifensee, Switzerland). Next, heparin (1000 U/animal; Leo Pharma, Breda, the Netherlands) was administered via the tail vein. After 5 minutes, the aorta was ligated cranially of the celiac trunk and at the level of the iliacal bifurcation, and cannulated with a 4-Fr enteral feeding tube (Vygon, Ecouen, France) in case of arterial washout. All groups underwent a portal washout procedure for which the portal vein was ligated proximally, dissected, and cannulated with a 6-Fr enteral feeding tube (Vygon). The inferior caval vein was ligated just above the renal vein and thereafter cut at the level of the iliac bifurcation and suprahepatically to prevent congestion. Subsequently, the liver was washed out with 50 mL of PS+, UW+, HTK−, or RL (n = 6 per solution) at 4 or 37°C (n = 6 per temperature) and at a pressure of 15 (standard) or 100 mm Hg (n = 6 per pressure)[2] through aortic and portal catheters while washout times were recorded with a stopwatch (Fig. 1). The washout pressure was calibrated before each experiment with a pressure transducer used in intravenous lines (Edward Lifesciences, Irvine, CA) that was positioned proximally to the portal vein and/or hepatic artery.

Figure 1.

Diagram of the experimental setup and group sizes.

After the 50-mL washout, the liver was removed and weighed. In the control group (no washout), the suprahepatic and infrahepatic caval vein, portal vein, and hepatic artery were simultaneously ligated, which trapped all the blood in the liver before removal. Rats were sacrificed by heart dissection during the washout or excision of the liver.

RBC Retention After Washout

Analysis of radioactive RBCs in blood samples and livers was performed with a γ-counter (Auto-Gamma 5530, Packard Instruments, Downers Grove, IL). The intrahepatic blood content was calculated as the weight percentage of radioactive RBCs in the liver by the total liver γ-count divided by the liver weight and corrected for the reference blood sample (blood γ-count/blood weight) as follows:

display math

The amount of residual radioactive RBCs was expressed as a percentage difference from control livers.

99mTc-Pertechnetate Distribution in Rats

On the basis of the high residual RBC retention after portal vein washout (discussed later), the RBC labeling dynamics of 99mTc-pertechnetate/PYP were verified with a dynamic γ-scan in four rats. Two rats underwent standard in vivo RBC labeling (as discussed previously), whereas PYP was omitted for the other two rats and was replaced with a physiological saline solution that does not interact with 99mTc-pertechnetate.[22] Images were acquired on a dual-head γ-camera [E.cam (180 frames/5 seconds, 128 × 128 pixel resolution), Siemens, Erlangen, Germany] to study the distribution and excretion of 99mTc-pertechnetate and 99mTc-pertechnetate/PYP. Furthermore, a pinhole collimator (3-mm insert, 128 × 128 pixel resolution)–equipped γ-camera (ADAC ARC 3000, Philips, Amsterdam, the Netherlands) was used to magnify the duodenal region to assess the possibility of radioactive duodenal excretion via the biliary system in the absence of a gallbladder in rats. The scintigraphic images were processed with Hermes image software (Hermes Medical Solutions, Stockholm, Sweden).


Livers after washout were sectioned into 3 parts and placed in a 10% (vol/vol) buffered formalin solution at 4°C for 5 days to allow radioactive decay. Subsequently, the parts were embedded in paraffin, cut into 5-μm sections, deparaffinized, and stained with hematoxylin and eosin. An experienced liver pathologist who was blinded to the experimental procedure assessed the liver sections for washout-induced injury according to the scoring method presented in Table 1. Images were acquired on an Olympus BX41 microscope equipped with a UC30 charged coupled device camera (Olympus, Tokyo, Japan).

Table 1. Histological Scoring System
Portal inflammation0Absent
Sinusoidal dilatation0Absent
Intralobular foci of inflammation0Not present
1<2 foci per 10× objective field
22-4 foci per 10× objective field
35-10 foci per 10× objective field
4>10 foci per 10× objective field
Endothelial inflammation0Not present
1<2 foci per 10× objective field
22-4 foci per 10× objective field
35-10 foci per 10× objective field
4>10 foci per 10× objective field
Portal edema0Not present
1<25% of portal tracts involved
225%-50% of portal tracts involved
351%-75% of portal tracts involved
4>75% of portal tracts involved
Confluent parenchymal necrosis0Not present
1Affecting <25% of the parenchyma
2Affecting 25%-50% of the parenchyma
3Affecting 51%-75% of the parenchyma
4Affecting >75% of the parenchyma
Hepatocellular mitosis0Not present
1<2 foci per 8 HPFs (40× objective)
22-4 foci per 8 HPFs (40× objective)
35-10 foci per 8 HPFs (40× objective)
4>10 foci per 8 HPFs (40× objective)
Cytosegresome formations (Councilman bodies)0Not present
1<2 foci per 10× objective field
22-4 foci per 10× objective field
35-10 foci per 10× objective field
4>10 foci per 10× objective field

Statistical Analysis

Data were processed in GraphPad (GraphPad Software, La Jolla, CA), and arterial liver washout times were analyzed with the Kruskal-Wallis test with Dunn's post hoc test. All other liver washout and histological scoring data were tested for intragroup and intergroup differences via a 2-way analysis of variance with the Bonferroni post hoc test. Results are expressed as means and standard deviations, and a P-value < 0.05 was considered statistically significant.


Agglutination of RBCs

Brightfield microscopy revealed extensive agglutination of RBCs suspended in colloid-containing perfusates; this was evidenced by rouleaux formation and morphological alterations of cells in HTK+, PS+, and UW+ (Fig. 2C,D,F). RBC agglutination was not observed in solutions not containing colloids; as shown for RL, HTK−, and PS− (Fig. 2A,B,E).

Figure 2.

Brightfield micrographs of RBC-containing suspensions diluted with different solutions to a hematocrit of 20%: (A) RL; illustrating blebbing of RBCs without agglutinative effects, (B) HTK−; illustrating no agglutination, (C) HTK+; illustrating extensive rouleaux formation, (D) PS+; illustrating extensive rouleaux formation, (E) PS−; illustrating no agglutinative effects, and (F) UW+; illustrating agglutinative effects microscopically comparable to those shown in panels C and D.

HTK+ and PS− confirmed that the RBC agglutination was colloid-induced and not attributable to other perfusate constituents. Furthermore, impermeant- and colloid-free RL led to extensive RBC blebbing (shedding of bullae from the cell surface), as shown in Fig. 2A.

Liver Washout

When RBC retention was averaged over all perfusate groups, portal washout was found to have reduced the mean RBC retention to 44.5% ± 1.7% (P < 0.05) of the liver weight (8.1 ± 2.5 g), whereas the retention in the control group (no washout) was 91.2% ± 17.5%. In contrast to the controls, simultaneous arterial and portal washout at 37°C and 100 mm Hg further reduced the retention (averaged over all perfusates) to 29.3% ± 7.9% (P < 0.001; Fig. 3A).

Figure 3.

(A) Residual 99mTc-pertechnetate–labeled RBC content in the liver categorized by the washout pressure group. The 4°C group is shown in blue, the 37°C group is shown in red, and the control group (no washout) is shown in black. Simultaneous arterial and portal washout with RL at 37°C and 100 mm Hg showed significantly less retention than the other washout modalities (P < 0.001). (B) Times needed for 50-mL washout via the portal vein are depicted in blue (4°C) and red (37°C), with black used for the arterial washout time of simultaneous arterial and portal washout at 37°C and 100 mm Hg. The use of noncolloidal solutions or the application of a washout pressure of 100 mm Hg significantly reduced the washout times (P < 0.001).

A perfusate-dependent reduction in RBC retention was achieved when RL was used for simultaneous arterial and portal washout at 37°C and 100 mm Hg (18.7% ± 9.4%) versus portal washout with RL at 4°C and 15 mm Hg (45.6% ± 5.1%, P < 0.01) and at 4°C and 100 mm Hg (50.5% ± 24.1%, P < 0.001). Also, portal washout with RL at 37°C and 15 mm Hg resulted in higher RBC retention (45.7% ± 30.5%, P < 0.01) than simultaneous arterial and portal washout with RL at 37°C and 100 mm Hg. However, simultaneous arterial and portal washout with RL at 37°C and 100 mm Hg was significantly different only from simultaneous arterial and portal washout with UW+ (37.4% ± 6.6%, P < 0.01).

A significant portal washout time reduction (P < 0.001) was achieved through washout with UW+ and PS+ at 4°C and 100 mm Hg (18.3 ± 8.0 and 15.1 ± 1.8 seconds, respectively) versus washout at 4°C and 15 mm Hg (84.4 ± 23.8 and 71.4 ± 35.3 seconds, respectively), as shown in Fig. 3B. Also, portal washout with HTK−, UW+, and PS+ at 37°C and 100 mm Hg significantly reduced washout times (7.0 ± 2.0, 13.3 ± 2.7, and 22.6 ± 28.3 seconds, respectively) in comparison with washout at 37°C and 15 mm Hg (49.7± 38.2, 106.0 ± 20.2, and 79.3 ± 6.5 seconds, respectively; P < 0.001). The influence of pressure instead of temperature on the washout time was evidenced by the significant reduction (P < 0.001) in the washout times at 4°C and 100 mm Hg with HTK−, UW+, and PS+ (10.5 ± 8.1, 18.3 ± 8.0, and 15.1 ± 1.8 seconds, respectively) versus the washout times at 37°C and 15 mm Hg. The arterial washout time was reduced with HTK− (29.7 ± 3.8 seconds, P < 0.05) and RL (21.6 ± 7.7 seconds, P < 0.001) at 37°C and 100 mm Hg versus UW+ at 37°C and 100 mm Hg (96.3 ± 7.2 seconds).

99mTc-Pertechnetate Distribution in Rats

Intravenously injected PYP, bound to the β-chain of hemoglobin, chemically reduces intracellular pertechnetate and causes 99mTc to become trapped in RBCs after intravenous administration. Renal accumulation of radioactivity and subsequent excretion into the bladder were seen in rats that received PYP and 99mTc-pertechnetate (Fig. 4A,C). The thyroid and salivary glands could be distinguished with low-level radioactivity because of the small fraction of 99mTc-pertechnetate unbound to PYP. In rats receiving 99mTc-pertechnetate without PYP, the accumulation of 99mTc-pertechnetate in the thyroid and salivary glands was slightly higher than that in PYP-receiving rats, and excretion was predominantly visualized in the gastric mucous membrane in the absence of renal excretion (Fig. 4B,D). With intrahepatic RBC retention after arterial and portal washout with RL at 37°C and 100 mm Hg, questions arose about whether the RBCs were hepatically cleared. Low-level radioactive biliary excretion into the duodenal loop was observed after approximately 20 minutes of circulation of 99mTc-pertechnetate/PYP, and this potentially accounted for or contributed to the residual radioactivity (18.7%).

Figure 4.

Biodistribution of 99mTc-pertechnetate in rats. (A) In the presence of PYP, 99mTc-pertechnetate/PYP was excreted via the urinary tract, whereas (B) in the absence of PYP, 99mTc-pertechnetate was taken up by the salivary and thyroid glands as well as the gastric mucosa. (C,D) The percentages of the administered radioactive dose that accumulated or excreted over time per ROI are depicted for conditions A and B, respectively. The ROIs were located over (1) the heart, (2) liver, (3) right kidney, (4) left kidney, (5) urinary bladder, (6) stomach, (7) abdominal background, (8) peripheral background, (9) thyroid gland, (10) parotid gland, and (11) submandibular gland.


A histological assessment of the livers did not show differences between the washout groups or perfusates. Furthermore, injury scores were very low with a mean of 6.5 ± 0.5 out of a 40-point maximum score. The obtained histological scores were 6.3 ± 1.1 for 37°C and 15 mm Hg, 6.5 ± 0.2 for 37°C and 100 mm Hg, 6.6 ± 0.8 for 4°C and 15 mm Hg, 6.6 ± 0.6 for 4°C and 100 mm Hg, and 6.6 ± 0.6 for combined washout. Individual scores and images of observed histological injuries can be found in Supporting Figs. 1 and 2, respectively.


This study has shown that PEG and HES cause agglutination of RBCs, regardless of the preservation solution used, whereas simultaneous arterial and portal washout with RL at 37°C and 100 mm Hg significantly reduces RBC retention and washout times. In light of this colloid-induced RBC agglutination, the need for colloids in washout solutions has been questioned.[23, 24] Moreover, liver washout with solely RL has been shown to yield good functional results, even though RL is not suitable for preservation that encompasses prolonged storage.[19, 25] Our study has confirmed that RL is best at reducing RBC retention after simultaneous arterial and portal washout at 37°C and 100 mm Hg. Although RL is equally as viscous as HTK− at 37°C,[9] HTK− did not lead to reduced RBC retention. Any potential coagulation-related side effects of RL during intravenous infusions, as reported previously,[26] can be overcome by full heparinization.

Although perfusate compositions with respect to colloidal solutions versus non-colloidal solutions are the basis of the ongoing washout debate, the perfusates in this study did not differ in RBC retention after portal washout. Thus, colloid-induced agglutination does not appear to cause solution- and corollary viscosity–dependent changes in washout efficacy in rat livers as previously suggested.[10] The viscosities of the solutions applied here were determined previously by our group and were found to be equal with respect to HES- and PEG-containing solutions.[9] However, colloid-lacking solutions were 2.5-fold less viscous, regardless of the temperature. Despite these differences in viscosity, no acute histological damage that might affect liver function was observed in livers washed out with colloid-containing solutions, despite the pressure of 100 mm Hg and the relatively high shear stresses imparted by these solutions.[27]

A potential difference in perfusates when they are employed under hypothermic and normothermic conditions is the pH. Higher temperatures will result in a slightly lower pH.[28, 29] However, adjusting the pH with sodium hydroxide to correct for temperature-induced differences would alter the osmolarity and Na2+/K+ ratio of perfusates. Because an acidic milieu may exert a protective effect on parenchymal cells[28, 29] and we did not anticipate significant effects from minor differences in the pH on liver washout outcomes, we chose not to alter the pH. This consideration was also made in light of the short exposure times during normothermic washout.

Interestingly, a remnant RBC fraction of 18.7% was present after simultaneous arterial and portal washout with RL at 37°C and 100 mm Hg, and this led to the question whether RBCs were cleared hepatically. Despite the comparable 99mTc-pertechnetate kinetics in rats and humans, more rapid elimination in the rat biliary system was found. Radioactive biliary excretion in the gallbladder or ileal loop after RBC labeling in humans is a rare finding, but it has been observed 24 hours after the administration of 99mTc-pertechnetate.[30-34] Although rats do not have a gallbladder, a pinhole collimator–equipped camera revealed radioactive bile being excreted into the duodenum after approximately 20 minutes in our experiments. With a mean bile flow of 0.6 μL/g of liver/minute in rats,[35] accumulation in the intrahepatic biliary tract occurred before visualization by the pinhole camera; as a result, higher residual radioactivity could be found in the liver after washout.

The livers were washed out via the portal vein (as in the rat model described by 't Hart[2]) or simultaneously via the portal vein and hepatic artery. If a colloidal solution were used for arterial washout during hypothermia, a time up to 10 times the time of portal vein washout would be needed. In that scenario, the times needed for portal washout and arterial washout would differ greatly, and this would result in a nonperfused portal vein for the extra duration of the arterial washout.

Prolonged washout times were predominantly observed with colloidal solutions, but in the case of portal washout only, they were shortened by the application of 100 mm Hg washout pressure. However, a high pressure is associated with increased shear stress and has previously been reported to induce endothelial damage.[16] Endothelial damage, in turn, may lead to graft dysfunction as a result of perfusion defects.[36, 37] Contrastingly, increased washout pressure and thus shear stress have been shown to result in improved early organ function.[2, 16, 17] Increased shear stress with higher washout pressures actually reduced endothelial damage by 20% in kidney grafts in comparison with standard-pressure washout with colloidal and noncolloidal perfusates.[2] In the latter study by 't Hart, the presence of a colloid in the perfusate did result in small and large venule injury, whereas large venule injury was observed only in the noncolloidal group. The histological analysis performed in this study did not reveal any differences between the perfusate groups with respect to these types of vascular injury.

In the final analysis, RL at 37°C and 100 mm Hg appeared to be the most optimal washout solution in an organ retrieval setting.[38-40] Accordingly, this study evinced that simultaneous arterial and portal vein washout with RL at 37°C and 100 mm Hg resulted in the least RBC retention in the liver. These results suggest that our long-held concept of cold washout with an organ preservation solution as the first step in effective organ storage should be reconsidered. However, further combined analysis of improved washout, the contributory role of arterial washout,[41] different solutions,[42, 43] and subsequent graft outcomes after transplantation should first confirm these findings.


The authors are grateful to their colleagues at the Departments of Pathology and Nuclear Medicine for their time and practical support during this study. Furthermore, they thank the Sanquin Blood Bank and especially Herbert Korsten for their support and generosity.