Initial hepatic microcirculation correlates with early graft function in human orthotopic liver transplantation

Authors

  • Gero Puhl,

    Corresponding author
    1. Klinik für Allgemein-, Viszeral-, und Transplantationschirurgie, Charité, Campus Virchow-Klinikum, Medizinische Fakultät der Humboldt-Universität zu Berlin, Berlin, Germany
    • Klinik für Allgemein-, Viszeral- und Transplantationschirurgie, Charité Campus Virchow-Klinikum, Medizinische Fakultät der Humboldt-Universität zu Berlin, Augustenburger Platz 1, 13353 Berlin, Germany
    Search for more papers by this author
    • Telephone: 49 30450552393; FAX: 49 30450552900

  • Klaus-D. Schaser,

    1. Klinik für Unfall- und Wiederherstellungschirurgie, Charité, Campus Virchow-Klinikum, Medizinische Fakultät der Humboldt-Universität zu Berlin, Berlin, Germany
    Search for more papers by this author
  • Daniel Pust,

    1. Klinik für Allgemein-, Viszeral-, und Transplantationschirurgie, Charité, Campus Virchow-Klinikum, Medizinische Fakultät der Humboldt-Universität zu Berlin, Berlin, Germany
    Search for more papers by this author
  • Katrin Köhler,

    1. Klinik für Allgemein-, Viszeral-, und Transplantationschirurgie, Charité, Campus Virchow-Klinikum, Medizinische Fakultät der Humboldt-Universität zu Berlin, Berlin, Germany
    Search for more papers by this author
  • Brigitte Vollmar,

    1. Abteilung für Experimentelle Chirurgie, Universität Rostock, Rostock, Germany
    Search for more papers by this author
  • Michael D. Menger,

    1. Institut für Klinisch-Experimentelle Chirurgie, Universität des Saarlandes, Universitätsklinik Homburg/Saar, Homburg/Saar, Germany
    Search for more papers by this author
  • Peter Neuhaus,

    1. Klinik für Allgemein-, Viszeral-, und Transplantationschirurgie, Charité, Campus Virchow-Klinikum, Medizinische Fakultät der Humboldt-Universität zu Berlin, Berlin, Germany
    Search for more papers by this author
  • Utz Settmacher

    1. Klinik für Allgemein-, Viszeral-, und Transplantationschirurgie, Charité, Campus Virchow-Klinikum, Medizinische Fakultät der Humboldt-Universität zu Berlin, Berlin, Germany
    2. Klinik für Allgemeine und Viszerale Chirurgie, Friedrich Schiller Universität Jena, Jena, Germany
    Search for more papers by this author

Abstract

Microcirculatory disturbances are an initial causative determinant in hepatic ischemia/reperfusion injury. The aim of this study was to assess sinusoidal perfusion during human liver transplantation using orthogonal polarization spectral imaging and to evaluate the significance of intraoperative microcirculation for early postoperative graft function. Hepatic microcirculation was measured in 27 recipients undergoing full-size liver transplantation and compared to a group of 32 healthy living-related liver donors. The microvascular parameters were correlated with postoperative aspartate aminotransferase and bilirubin levels. Hepatic perfusion following liver transplantation was found to be significantly decreased when compared with the control group. Volumetric blood flow within the individual sinusoids increased due to sinusoidal dilatation and enhanced flow velocity. Regression analysis of postoperative aspartate aminotransferase and bilirubin with microvascular parameters revealed significant correlations. The extent of volumetric blood flow increased within the first 30 minutes after reperfusion and showed a significant correlation with postoperative aspartate aminotransferase release and bilirubin elimination. In conclusion, postischemic hepatic microvascular perfusion was analyzed in vivo, demonstrating significant microvascular impairment during liver transplantation. Sinusoidal hyperperfusion appears to confer protection against postischemic liver injury, as given by the correlation with aspartate aminotransferase and bilirubin levels. Thus, these findings may have therapeutic importance with respect to mechanisms mediating postischemic reactive hyperemia. (Liver Transpl 2005;11:555–563.)

Liver transplantation (LTX) is currently considered to represent the primary therapy for end-stage liver disease and acute liver failure. For hepatic transplantation the liver is transiently deprived of oxygen and subsequently reperfused, inevitably resulting in ischemia/reperfusion (I/R) injury of varying extents. Additionally, a number of clinical situations, such as those associated with low flow states and diverse surgical procedures, are known to profoundly influence postoperative graft function. Thus, identification of relevant parameters during the early posttransplantation course is prerequisite for rapid induction of therapeutic interventions to inhibit development of graft dysfunction and improve outcome after LTX. Even if a variety of donor- and/or recipient-related factors have been shown to correlate with poor outcomes after liver transplantation,1–3 perception and knowledge of reliable parameters being predictive for the outcome of liver transplantation are still major issues of the liver transplant community.2

Distinct techniques of organ preservation and cold storage solutions have markedly contributed to reduce the ischemic injury. Despite the fact that reestablishment of blood flow to the liver represents a vital requirement for recovery of cellular and organ function, it typically aggravates ischemia-induced tissue damage. Sinusoidal perfusion failure, which is characterized by endothelial cell swelling and dysfunction, sludge formation with microthromboses, and leukocyte stagnation, is a major determinant of I/R injury of the liver. Hypoxic damage of hepatic sinusoidal endothelium is generally considered to be the first pathophysiological change in the cascade of events resulting in I/R injury and manifestation of graft dysfunction.4 Potential mechanisms involve free radical damage, release of inflammatory and procoagulatory mediators, release of vasoactive substances, as well as expression of adhesion molecules.5

Liver transplantation studies in rats demonstrated the pivotal role of the first few minutes after revascularization in the development of hepatic reperfusion injury6; raising the question of initial microvascular dysfunction as a predictive parameter for the early graft function during the clinical course following liver transplantation. Intravital fluorescence microscopy currently represents the standard method for the experimental assessment of microcirculatory impairment; however, intravital fluorescence microscopy cannot be used clinically because of the complex technical setup and the necessity of contrast enhancement by potentially toxic fluorescent dyes. Therefore, numerous non-imaging techniques have been used clinically, including thermodiffusion7 and laser Doppler flowmetry.8 Both methods demonstrated their value in estimating aspects of liver microcirculation and tissue perfusion under clinical conditions.7, 9 However, these techniques do not permit the direct visualization of the individual segments of the human hepatic sinusoidal network in vivo, and interpretation of the acquired data is a controversial discussion. In contrast, the technology of orthogonal polarization spectral (OPS) imaging10 has the potential to non-invasively visualize and quantify human hepatic microvascular perfusion intraoperatively11, 12 comparable to intravital fluorescence microscopy for animal experimental studies.

Based on these previous experimental studies we hypothesized the prognostic value of initial graft microcirculation as a predictor for postoperative hepatocellular injury and early graft function following full-size cadaveric liver transplantation. Therefore, we aimed to intraoperatively visualize and quantify the human hepatic microcirculation by OPS imaging technique and to correlate these microvascular perfusion changes to postoperative aspartate aminotransferase (ASAT) and bilirubin levels.

Abbreviations

LTX, liver transplantation; I/R, ischemia and reperfusion; OPS imaging, orthogonal polarization spectral imaging; ASAT, aspartate aminotransferase; RP5, measurement 5 minutes following reperfusion; RP30, measurement 30 minutes following reperfusion; RP5-RP30, procentual change between the 5 and 30 minutes postreperfusion measurements; RBCV, red blood cell velocity; SD, sinusoidal diameter; CIT, cold ischemia time; FSD, functional sinusoidal density; POD, postoperative day; VBF, volumetric blood flow within the individual sinusoid.

Patients and Methods

The study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki as reflected by the approval of the local Ethics Committee. Informed consent was obtained from all patients. Proof and safety of the OPS imaging system for the clinical use are documented by the EC certificate of conformity no. GOD 00 10 41750 001 according to Annex IV, section 5, of council directive 93/42/EEC concerning medical devices, test report no. DM1V0113201 and DM 1V0113202 released with the aforementioned certificate number by the certification body of TÜV product service.

Patient Population

Tw enty-seven recipients were studied while undergoing orthotopic full-size liver transplantation. The microvascular parameters following full-size liver transplantation were compared to the baseline perfusion parameters of healthy patients undergoing laparotomy for living-related right liver lobe donation, since these data have to be considered to represent physiological values of human hepatic microcirculation. There were no significant differences between mean age, body weight, and gender of the cadaveric graft donors, cadaveric graft recipients (n = 27), and the living donor control group (n = 32) (age 43 ± 22 vs. 41 ± 20 vs. 43 ± 14 years, body weight 64 ± 24 vs. 63 ± 28 vs. 69 ± 12 kg, gender M/F 15/12 vs. 18/9 vs. 15/17). All grafts were ABO compatible and matched for physiological body to liver weight ratio within the recipients.

The cadaveric donor livers were all suitable for transplantation, as estimated by macroscopic appearance, histological examination, and standard liver function tests (Table 1). The donor hepatectomy was carried out according to the standards of multiorgan procurement. Preservation of the grafts was performed using University of Wisconsin solution by pressure-controlled retrograde aortal and antegrade portal vein perfusion. The microvascular analysis of the cadaveric donor by the means of a before and after measurement has been impossible, because in most of the cases the organ procurement occurred outside our transplant center.

Table 1. Donor Age, Stay on ICU, ASAT and Sodium Levels, Evidence of Fatty Changes and Macroscopic Appearance (According to the EUROTRANSPLANT Liver Report), and Cold Ischemia Time (CIT) of the Liver Grafts
Age (years)ICU stay (days)ASAT (U/L)Sodium (mmol/L)MacroscopyFat (%)CIT (hours)
  1. NOTE. Subdivision in two postreperfusion groups with ASAT either less than 1000 U/L or more than 1000 U/L (bold).

  2. Abbreviations: ICU, intensive care unit; ASAT, aspartate aminotransferase.

44318153Good10711
742109150Good20568
47311140Good 713
48411145Good<5353
74323133Acceptable 697
511107138Good<5653
61210162Acceptable30-40939
57513145Good 400
176130152Good 449
54326149Good<5426
7358151Good 564
56218157Good20768
39218153Good20483
39325137Good 283
51226155Good20356
59232143Acceptable30272
153484146Good<5765
65435142Good101010
45513146Acceptable30420
8631149Good 384
21518142Good 413
65625138Good10820
59439142Good20437
17342140Good 330
14374141Good 656
66210143Good10349
8578153Good 765

Recipient Anesthesia

With regard to the reduced metabolic ability of the recipients' livers no oral premedication was administered. After rapid sequence induction with 1 mg cis-atracurium, 3 to 5 mg/kg body weight thiopental, and 1.5 mg/kg body weight succinylcholine, patients were intubated orotracheally and mechanically ventilated with an intermittent positive pressure ventilation mode and positive end expiratory pressure of 5 cm H2O. General anesthesia was maintained with an oxygen/air mixture (FiO2 0.4), desflurane (endexpiratory concentration up to a maximum of 6 vol%) and continuous infusion of fentanyl as required (0.06-0.12 μg/kg body weight/min). Muscle relaxation was maintained by 1 mg/kg body weight cis-atracurium as required with respect to relaxometric parameters. The intraoperative hemodynamic monitoring consisted of central venous pressure measurement, invasive arterial blood pressure analysis, continuous electrocardiography, and capnometry aiming at an end-tidal carbon dioxide tension of 35 to 40 mmHg. Fluid replacement was guided by blood loss, arterial blood pressure, and central venous blood pressure, keeping the hemoglobin concentration between 9 and 12 g/dL, the mean arterial blood pressure above 60 mmHg, and the central venous pressure below 5 mmHg to prevent engorgement of the liver. The application of fentanyl was stopped at the beginning of the anhepatic phase and anesthesia was continued with desflurane only. Twenty minutes prior to expected reperfusion, 250 mg methylprednisolone was administered intravenously. Postoperatively, all patients were extubated in the operating room and transferred to the transplant intensive care unit.

Recipient Operation

Liver transplantation closely followed the technique introduced by Starzl. The recipient hepatectomy required the interruption of recipients' vena cava blood flow with the need of a porto-femoro-axial bypass in all patients. The implantation was performed by end-to-end anastomosis of the donor to the recipients' supra- and infrahepatic vena cava. The grafts were perfused simultaneously after subsequent reconstruction of the portal vein and hepatic artery. Biliary reconstruction was performed by side-to-side anastomosis with introduction of a T-tube.

OPS Imaging of the Liver Microcirculation

The entire system (Cytoscan, Cytometrics Inc., Philadelphia, PA) consists of an optical probe homing the video microscope and the base processing the images. For the illumination of the tissue light is collected within the OPS imaging probe and passes a spectral filter to isolate the wavelength of 548 nm (isobestic point of hemoglobin). Light that is remitted from the target tissue is collected to form an image of the illuminated region within the target tissue upon a charge-coupled device videocamera. Since polarization is preserved in reflection, photons scattered in the tissue contribute to the images, forming a virtual light source within the tissue by back illumination. Hence, the contrast is obtained from the absorption of the light by hemoglobin, and hemoglobin-carrying structures appear in a negative contrast. In fact, the microcirculation can be visualized using OPS imaging similar to epi-illumination intravital microscopy.10 The technology enables for the measurement of the vessel diameter, the length of the perfused vessels per observation area, the distance between the perfused vessels, and the red blood cell velocity within the vessel.11, 12

Using a standard video recorder (S-VHS, AG 7350-E, Panasonic, Matsushita Electric Ind., Osaka, Japan) and video screen, online imaging as well as recording of images was performed with a final magnification of 465-fold and a depth of focus of 200 to 500 μm. For the intraoperative measurements the probe including the entire cable system was enpacked into sterile foil (OpMi Drape, Zeiss, Oberkochen, Germany).

For the measurements the OPS probe was gently positioned under physiological saline immersion by hand, allowing stable video recordings. Using this technique, in vivo analysis of the hepatic microvasculature was performed in four regions of interest on the upper and lower liver surfaces of the liver graft and recorded for 30 to 60 seconds in each region. A total of 27 liver grafts were examined. With the concept of simultaneous portal vein and hepatic artery reperfusion, the microcirculation was measured 5 (RP5) and 30 (RP30) minutes after reestablishment of blood flow.

Within the control group (n = 32), the baseline conditions of the microcirculation were recorded directly after laparotomy and prior to mobilization of the liver from the ligaments and preparation of the hepatic hilus structures.

Microcirculatory Analysis

Quantification of the microhemodynamic parameters was performed offline by single frame and frame-to-frame analysis of the videotaped images using a computer-assisted image analysis system (CapImage, Zeintl, Heidelberg, Germany). The analysis included the determination of sinusoidal diameter (SD) and the red blood cell velocity (RBCV) within the sinusoids. The volumetric blood flow within the sinusoids (VBF) was calculated from the sinusoidal diameter and the red blood cell velocity [VBF = π/4 × D2 × RBCV], assuming cylindrical geometry of the sinusoids. Functional sinusoidal density (FSD) has been introduced previously,11 and is defined as the length of all red blood cell-perfused sinusoids per observation area (200 μm × 200 μm).

Parameters of Liver Injury and Function

For the estimation of liver injury and function, ASAT and bilirubin were measured by standard assays daily until postoperative day (POD) 5. The timeframe of observation has been limited to the first five PODs to focus on I/R-related changes and to exclude influences such as viral reinfection or acute rejection. General infectious diseases, such as peritonitis or pneumonia, have been excluded by routine clinical and laboratory examinations as well as x-rays. Vascular complications have been excluded by routinely performed daily ultrasounds.

In accordance to the definition of severe I/R injury as set by our clinical center,13 the patients were divided in two groups by peak ASAT levels of less and more than 1000 U/L. Coagulation parameters were not analyzed for this study, since the recipients received different amounts of fresh frozen plasma and coagulation factors. External bile flow into the T-tube has not been considered as a valuable parameter, as bile drained outside may not be regarded as total bile flow.

Statistics

The results are expressed as mean ± SD. The assumption of normality and homogeneity of variance was tested using the Kolmogorov-Smirnov test. For comparison of transplantation and control group the ANOVA for repeated measurements was performed. Within the transplantation group, Student t test was performed. Differences were considered significant for P values less than 0.05. The dependency between microcirculatory parameters, donor age, and intensive care unit stay of the donors before organ procurement, preservation period, anhepatic period, postoperative enzyme release, and bilirubin excretion was evaluated by the Pearson Product Moment Correlation. A correlation to the histological findings was not performed, as far as the histological examination was not performed in all cases.

Results

Sinusoidal Perfusion

Upon simultaneous reperfusion, mean RBCV was 702 ± 273 μm/s following 5 minutes and 723 ± 243 μm/s 30 minutes thereafter. The RBCV within the control group averaged 828 ± 185 μm/s, which was 15 and 13% more when compared to the reperfusion. Between 5 and 30 minutes after reperfusion, RBCV increased by 6%. Analysis of SD in the control group revealed an average value of 8.4 ± 0.5 μm. Reperfusion caused dilatation of hepatic sinusoids by 17% at RP5 and 15% at RP30. In parallel to the RBCV and SD, mean VBF within the individual sinusoids elevated by 22% (RP5) and 26% (RP30) when compared to the control group. After reperfusion, VBF between RP5 and RP30 increased by 4%. In contrast to that, FSD declined by 28% (RP5) and 27% (RP30) cm per cm2.

The Pearson product moment correlation between VBF and FSD revealed an inverse correlation (P < 0.05) both after 5 (r = −0.744) and 30 minutes (r = −0.662) of reperfusion (Fig. 1A). Despite the fact that the CIT did not correlate with FSD or VBF, the procentual increase of VBF between RP5 and RP30 showed a weak but significant relationship (r = −0.402, P < 0.05) (Fig. 1B). Postischemic microvascular parameters did not significantly correlate with donor age, stay on intensive care unit, recipient age, and the length of the anhepatic period, as well as the macroscopic appearance estimated by the explant surgeons (according to the EUROTRANSPLANT liver report) (data not shown).

Figure 1.

The Pearson product moment correlation between VBF and FSD revealed an inverse correlation (P< 0.05) both after 5 (r= −0.744) (A) and 30 minutes (r= −0.662) of reperfusion, as well as between the procentual increase of VBF between RP5 and RP30 and the length of the CIT (r= −0.402) (B). Straight lines represent regression; dotted gray lines represent confidence intervals. Data include the 20 patients with ASAT levels <1000 U/L.

Parameters of Liver Injury and Function

Peak levels of postoperative ASAT were found at the first day after transplantation (910 ± 1273 U/L). More detailed analysis showed that 20 patients developed a peak postoperative ASAT level of less than 1000 U/L (292± 267 U/L), whereas the remaining seven patients showed ASAT levels exceeding the 1000 U/L threshold (2675 ± 1373 U/L). Within the first five PODs the daily ASAT levels significantly differed between the two groups. In contrast, the course of bilirubin displayed a significant difference between these groups (ASAT <1000 U/L vs. >1000 U/L) only on the second and third PODs.

Donor- and Graft-Specific Parameters

Given the mean cold ischemia time (CIT) of 555 ± 207 minutes, there was no difference between grafts with an ASAT level of less than 1000 U/L compared with those with a peak level of more than 1000 U/L (522 ± 192 minutes vs. 648 ± 235 minutes, P = 0.264). Additionally, there was no significant difference of donor and recipient age or stay on intensive care unit, as well as length of the anhepatic period between the ASAT <1000 U/L and the ASAT >1000 U/L group as summarized in Table 2. The histological evidence of fatty changes among the donor livers was not adduced in all grafts. Given the macroscopic appearance of the donor livers, no differences between the ASAT <1000 U/L and >1000 U/L group were seen.

Table 2. Donor Age, Stay on ICU, Cold Ischemia Time, and Anhepatic Period in the Total of 27 Liver Graft Recipients
 Total n = 27ASAT <1000 U/L n = 20ASAT >1000 U/L n = 7
  1. NOTE. Patients were further subdivided in two groups with ASAT either less than 1000 U/L or more than 1000 U/L. The anhepatic period is defined as the time from dissection of the portal vein from the recipient's liver (start of veno-venous bypass) to restoration of portal vein inflow. The cold ischemia time (CIT) is defined as the time from the start of cold perfusion in the donor to revascularization in the recipient. ICU, intensive care unit. Blood samples of POD 1 were taken earliest 12 hours following transplantation and daily thereafter. Values are mean ± SD.

Donor age (years)43 ± 2246 ± 2136 ± 26
Stay on ICU (hours)140 ± 207129 ± 216174 ± 187
Cold ischemia time (min)555 ± 207531 ± 196624 ± 237
Anhepatic period (min)82 ± 2981 ± 2984 ± 31

Influence of CIT on Liver Injury and Function

There was no correlation between ASAT levels of PODs 1 and 2 with length of the CIT. However, CIT and the course of postoperative ASAT revealed a positive correlation at PODs 3 to 5 (P < 0.05) within the whole collective of patients, as well as when divided into the group of patients with ASAT <1000 U/L and >1000 U/L (Fig. 2A-B). The dynamics of postoperative bilirubin revealed a positive correlation (P < 0.05) with the CIT on POD 4 (r = 0.414).

Figure 2.

Regression analysis as illustrated by the correlation coefficients r between postoperative ASAT levels and the length of CIT in the total of 27 patients (A) and the subgroups (B); i.e., the group of patients with ASAT less than 1000 U/L (open bars) and more than 1000 U/L (gray bars). *Statistically significant correlation.

Influence of Hepatic Microcirculation on Liver Injury and Function

The Pearson product moment correlation between postoperative ASAT release and FSD revealed an inverse correlation both 5 minutes and 30 minutes following reperfusion. Although there was no correlation to the ASAT peak level on POD 1, regression analysis of ASAT and FSD RP5 values revealed a significant correlation at PODs 3 to 5 (P < 0.05). FSD RP30 values and postoperative ASAT levels showed a significant correlation on PODs 4 and 5. Considering the ASAT threshold of <1000 U/L and >1000 U/L, the correlation between initial FSD and postoperative ASAT course was found to be significant (Fig. 3A-B), whereas the difference between FSD levels in the ASAT <1000 U/L did not significantly differ from ASAT >1000 U/L group (FSD RP5: 321 ± 63 vs. 266 ± 35; FSD RP30: 325 ± 60 vs. 278 ± 43).

Figure 3.

Regression analysis between postoperative ASAT levels and intraoperative FSD in the total of 27 patients (A) and the subgroups (B); i.e., the group of patients with ASAT less than 1000 U/L (open bars) and more than 1000U/L (gray bars). *Statistically significant correlation.

A similar correlation was found for the course of postoperative bilirubin levels and FSD. The higher the FSD RP5 and RP30 values, the lower the bilirubin levels in the follow-up at POD 2 (r = −0.514 and −0.564, P < 0.05), POD 3 (r = −0.494 and −0.528, P < 0.05), POD 4 (r =−0.496 and −0.504, P < 0.05), and POD 5 (r = −0.538 and −0.578, P < 0.05).

Although the VBF did not directly correlate with postoperative enzyme release or bilirubin levels, a significant relationship was found between the increase in percent change of VBF between the RP5 and RP30 measurements. The Pearson product moment correlation between ASAT and VBF RP5-RP30 revealed an inverse correlation at PODs 3 to 5 (P < 0.05). ASAT <1000 U/L and >1000 U/L did significantly correlate as well (Fig. 4A-B). Correspondingly, bilirubin levels and VBF RP5-RP30 displayed a significant correlation on POD 2 (r = −0.489), POD 3 (r = −0.471), POD 4 (r = −0.501), and POD 5 (r = −0.429).

Figure 4.

Regression analysis between postoperative ASAT levels and VBF RP5-RP30 in the total of 27 patients (A) and the subgroups (B); i.e., the group of patients with ASAT less than 1000 U/L (open bars) and more than 1000 U/L (gray bars). *Statistically significant correlation.

Of utmost interest, the percent change of VBF between the RP5 and RP30 measurements significantly differed between the ASAT <1000 U/L and ASAT >1000 U/L group (8.6% ± 8.4% vs. −8.2% ± 23.2%, P < 0.05) (Fig 5).

Figure 5.

Percent change of VBF between the RP5 and RP30 measurements in the total of 27 patients (A) and the subgroups (B); i.e., the group of patients with ASAT less than 1000 U/L (open bars) and more than 1000 U/L (gray bars). Values are mean ± SD; *P< 0.05 vs. ASAT<1000 U/L.

Discussion

The present study presents the first assessment of human hepatic microcirculation in response to I/R injury during full-size cadaveric liver transplantation by direct imaging. Postischemic microvascular architecture and perfusion were directly visualized and quantitatively analyzed in vivo, focusing on microcirculatory impairment during simultaneous graft reperfusion. The present findings demonstrate that quality of microvascular perfusion within the first 30 minutes following reperfusion inversely correlates with postoperative ASAT release and bilirubin elimination and therefore reflects early graft integrity and function. The data further suggest the important role of increased sinusoidal VBF as an indicator of postischemic reactive hyperemia, which appears to be an important compensatory mechanism aiming at the maintenance of adequate sinusoidal perfusion. Furthermore, the extent of reactive hyperemia during the initial reperfusion period seems to determine the recovery of ASAT and bilirubin following LTX.

Warm hepatic I/R is associated with perfusion failure of sinusoids due to significant hemoconcentration, reduced perfusion pressure, pressure related sinusoidal leukostasis, as well as sinusoidal narrowing caused by hypoxia-induced endothelial cell swelling.14 In animal experimental studies it has been demonstrated that a reduction of up to 25% of sinusoidal diameter15 and sinusoidal perfusion rate,16, 17 and a decrease of up to 40% in red blood cell velocity within the sinusoids18 results in hepatic microcirculatory perfusion failure following 20 to 90 minutes of warm ischemia. In contrast, 4°C hypothermia during ischemia protects from hepatic microcirculatory perfusion failure after 90 minutes of lobular ischemia,17 and following one hour preservation with subsequent liver transplantation.18 Further prolongation of CIT results in moderate (6 hours)19 and severe (24 hours)20, 21 impairment of microcirculation. In the present study a moderate microcirculatory deterioration was encountered. After a mean storage period in University of Wisconsin solution of 9.2 ± 3.4 hours, the FSD was reduced to approximately 30% when compared to the control group, which is well in line with the aforementioned experimental studies.

Sinusoidal VBF is calculated as the product of RBCV, and the square of the SD according to the equation of Gross and Aroesty has been described as a useful microvascular parameter.22 Its value for the determination of postischemic microvascular changes has been confirmed in our previous study on clinical living donor liver transplantation.12 VBF within the individual sinusoids revealed an increase compared with the control group, which may be best interpreted as postischemic reactive hyperemia. Postischemic increase in VBF was accompanied by an increase in SD of 17%. We further demonstrated that due to a continuous increase of RBCV the VBF increased by about 10% within the observation period between RP5 and RP30, whereas the FSD and the SD remained unchanged. Improved sinusoidal perfusion pressure due to simultaneous portalvenous and hepatic artery reperfusion could play a major role in this finding, because experimental studies emphasized the significance of simultaneous graft reperfusion for the improved quality of initial microcirculation.6 Other factors directly influencing the width of the perfused sinusoids have to be considered as well; e.g., the regulation of the endothelin A (ETA)/endothelin B (ETB) receptor ratio under I/R conditions which results in a ETB-mediated NO release and sinusoidal relaxation, although this is not analyzed within this study. Finally, studies utilizing FITC-dextran fluorescence measurements23, 24 or laser Doppler flowmetry9 confirm the aforementioned hyperemia on the liver surface following cold ischemia and subsequent reperfusion.

ASAT activity levels represent a well-established parameter for the estimation of hepatic injury following organ procurement, preservation, and transplantation. Nonetheless, ASAT levels alone are not predictive of the occurrence of initial non- or poor graft function,25 although studies implicate ASAT >2000 U/L1 and ASAT >5000 U/L25 as determinants of higher patient mortality. In the present study we have characterized the effects of initial hepatic microvascular dysfunction on peak posttransplantation ASAT levels, as well as on the dynamics of ASAT and bilirubin elimination within the first 5 days following transplantation reflecting the extent of graft integrity and function. For the analysis, we differentiated between patients with peak ASAT levels of less and more then 1000 U/L, in accordance to our own center experience.13 Although clinical studies have demonstrated the influence of CIT on graft injury,26 peak ASAT levels did not correlate with the length of CIT.27 In line with these previous findings,27 peak ASAT levels in our study failed to correlate with CIT. However, course of ASAT three to five days following liver transplantation revealed a significant correlation with the length of CIT, both in the ASAT <1000 U/L and ASAT >1000 U/L groups. In rat liver grafts it has been shown that the degree of microvascular dysfunction largely depends on the duration of CIT.28 We now confirm and extend this information for the human situation by an inverse correlation of VBF increase within RP30 and CIT in this study.

Experimental studies have demonstrated that ASAT levels and bile flow positively correlate with the number of nonperfused sinusoids and the decrease in erythrocyte flux, indicating the dependency of hepatocellular integrity and excretory function on the quality of microvascular perfusion and tissue oxygenation following ischemia.14 While intravital fluorescence microscopy represent the gold standard in experimental microcirculatory research, the estimation of human hepatic microcirculation has been restricted to laser Doppler flowmetry,9 thermodiffusion,7 and hepatic oxygenation.29 By using thermodiffusion, Klar et al.7 demonstrated a correlation between hepatic perfusion and clinical parameters with grafts being divided in those with initial good and those with initial non-function. Good graft function was characterized by a hepatic perfusion above 53 mL/100 g/min. Furthermore, ASAT and ALAT peak levels inversely correlated with intraoperative perfusion values below 53 mL/100 g/min. Whether this perfusion failure was the result of direct sinusoidal constriction, lack of sufficient sinusoidal perfusion pressure, or a reduction of the FSD cannot be answered by this analysis.

In this study the increase of VBF during the initial reperfusion was the result of sinusoidal dilatation and flow acceleration. In accordance with the course of the clinical parameters, the present study highlights the reactive hyperemia as a possible mechanism for graft recovery following I/R injury.

Taken together, OPS imaging has been shown to be a valuable tool for intraoperative analysis of I/R-induced deterioration of the hepatic microcirculation. With the knowledge of the significant correlation of clinical and microvascular parameters, VBF has been proved of being a potential key factor limiting postischemic liver injury. This finding should have implications with regard to the development of new therapeutic strategies that target sinusoidal diameters,30, 31 and flow velocity as the underlying components of reactive hyperemia.

Ancillary