LIVER TRANSPLANTATION 17:1448-1456, 2011
ORIGINAL ARTICLE
Hepatic Blood Flow Plays an Important Role in Ischemia-Reperfusion Injury Dympna M. Kelly,1 Hiroaki Shiba,1 Shunichi Nakagawa,1 Samuel Irefin,2 Bijan Eghtesad,1 Cristiano Quintini,1 Federico Aucejo,1 Koji Hashimoto,1 John J. Fung,1 and Charles Miller1 Departments of 1General Surgery and 2Anesthesiology, Cleveland Clinic, Cleveland, OH
Severe ischemia/reperfusion (IR) injury is associated with poor hepatic microperfusion. The aim of this study was to investigate the role of hepatic artery flow (HAF) and portal vein flow (PVF) in IR injury. From January 2004 to June 2008, 566 patients underwent orthotopic liver transplantation (OLT). The data were retrospectively reviewed via the transplant database. Patients with hepatic artery (HA) or portal vein (PV) thrombosis and retransplant patients were excluded. Intraoperative PVF and HAF values and graft weights were measured routinely, and the central venous pressure, mean arterial pressure, cardiac output, and cardiac index were recorded with hepatic blood flow measurements. Complete data were available for 312 primary OLT recipients (215 males and 97 females; mean age ¼ 54 6 10 years). The patients’ follow-up ranged from 215 to 1746 days (705 6 408 days). IR injury was defined by the aspartate aminotransferase (AST) level on postoperative day (POD) 2, and the patients were divided into 3 groups: (1) mild IR injury [AST < 500 U/L; n ¼ 160 (51%)], (2) moderate IR injury [AST ¼ 500-1000 U/L; n ¼ 85 (27%)], and (3) severe IR injury [AST > 1000 U/L; n ¼ 67 (21%)]. The demographics and pre-OLT variables (the Model for End-Stage Liver Disease score (MELD), platelet counts, PV thrombosis, transjugular intrahepatic portosystemic shunts, and shunts on computed tomography scans) were similar in all groups. The graft survival rate was 99% in group 1, 95.2% in group 2 (P ¼ 0.02), and 92.3% in group 3 (P ¼ 0.016). The patient survival rates were similar in the 3 groups. The cold ischemia time (CIT) was significantly higher in group 3 versus group 1 (P < 0.007). In the statistical analysis, low HAF, PVF, total liver blood flow (TLBF), and augmented HAF values were associated with a greater likelihood of elevated AST levels on POD 2. The strongest univariate predictors of AST were reduced augmented HAF (mL/minute/100 g) values (P < 0.001) and reduced TLBF (mL/minute/100 g) values (P < 0.001). In a covariate analysis with adjustments for CIT and donor variables, the blood flow parameters remained important predictors of graft function. In conclusion, this report demonstrates for the first time that reduced hepatic blood flow is a significant finding in patients with severe hepatic IR injury. Liver Transpl 17:1448-1456, 2011. V 2011 AASLD. C
Received December 16, 2010; accepted August 14, 2011.
All transplanted livers sustain ischemia/reperfusion (IR) injury, and cold ischemia is an important component of this injury. In up to 10% of cases, severe IR injury results in graft failure, and less severe forms manifest with various degrees of graft dysfunction, elevated aminotransferase levels, cholestasis, and coagulopathy.1,2 In clinical practice, IR injury is
defined by postoperative AST levels. A significant correlation has been reported between a high peak AST value within the first 72 hours and lower graft and patient survival rates.1,2 Although there has been an abundance of research in the laboratory, clinical research is sparse. Sinusoidal injury has been implicated in the pathogenesis of hepatic IR injury, and
Abbreviations: ACR, acute cellular rejection; ALT, alanine aminotransferase; AST, aspartate aminotransferase; CIT, cold ischemia time; CT, computed tomography; HA, hepatic artery; HABR, hepatic artery buffer response; HAF, hepatic artery flow; ICU, intensive care unit; INR, international normalized ratio; IR, ischemia/reperfusion; MELD, Model for End-Stage Liver Disease; OLT, orthotopic liver transplant; POD, postoperative day; PV, portal vein; PVF, portal vein flow; SLV, standard liver volume; TIPS, transjugular intrahepatic portosystemic shunt; TLBF, total liver blood flow. This study was supported by an American Society of Transplant Surgeons/Wyeth midlevel faculty award (2008-2010). Address reprint requests to Dympna M. Kelly, M.D., M.Ch., F.R.C.S.I., Department of General Surgery, Cleveland Clinic, 9500 Euclid Avenue, A80, Cleveland, OH 44195. Telephone: 216-444-1888; FAX: 216-445-2153;; E-mail:
[email protected] DOI 10.1002/lt.22424 View this article online at wileyonlinelibrary.com. LIVER TRANSPLANTATION.DOI 10.1002/lt. Published on behalf of the American Association for the Study of Liver Diseases
C 2011 American Association for the Study of Liver Diseases. V
LIVER TRANSPLANTATION, Vol. 17, No. 12, 2011
nonparenchymal cells, endothelial cells, Kupffer cells, and Ito cells are more susceptible to cold IR injury than hepatocytes.3 In cases of moderate to severe IR injury, extensive changes occur in the endothelial cells: nuclear membrane vacuolization, mitochondrial swelling, plasma membrane breakdown, cytoplasmic rarefaction, ball-like rounding, and nuclear condensation.4 This results in a complete loss of sinusoidal endothelial cell viability, denudation of the sinusoids, and disruption of the microcirculatory blood flow.4 Using orthogonal polarization spectral imaging, Puhl et al.5 demonstrated that an early alteration of hepatic microperfusion was associated with IR injury in humans. They found that the quality of microvascular perfusion within the first 30 minutes after reperfusion correlated inversely with postoperative AST and bilirubin levels. Klar et al.6 reported a significant negative correlation between intraoperative hepatic microperfusion ( 1000 U/L).
Donor and Recipient Characteristics The donor data included the age, sex, cause of death, occurrence of cardiac arrest before procurement, serum sodium level, bilirubin level, AST level, international normalized ratio (INR), and preservation solution. The preoperative recipient characteristics included the age, sex, body mass index, etiology of liver disease, and calculated Model for End-Stage Liver Disease (MELD) score. In addition, we reviewed preoperative computed tomography (CT) scans for evidence of PV thrombosis and splenorenal shunts (ie, venous connections between the splenic vein and the left renal vein that decompressed the PV system; any splenorenal shunt that was visualized on a CT scan was included). Either of these factors could influence PVF after reperfusion. We also reviewed patient records to identify all patients undergoing a transjugular intrahepatic portosystemic shunt (TIPS) procedure before OLT. Postoperative details included the following: the HA resistive index from duplex ultrasound on POD 1; the serum AST, ALT, and bilirubin levels and INR on POD 2; the incidence of biopsy-proven acute cellular rejection (ACR); complications; and graft and patient survival.
PATIENTS AND METHODS From January 2004 to June 2008, 566 patients underwent orthotopic liver transplantation (OLT) at Cleveland Clinic. The data were retrospectively reviewed via the transplant database and operative and anesthesia records. The study was approved by the Institutional Review Board (IRB 05-172). Liver transplantation was performed with standard orthotopic or piggyback techniques with or without venovenous bypass.7 Before 2005, donor livers were preserved with University of Wisconsin solution; thereafter, histidine tryptophan ketoglutarate solution was used. The basic immunosuppression consisted of basiliximab induction, tacrolimus, and steroids. Mycophenolate mofetil was used as a renal-sparing agent to minimize the side effects of tacrolimus. Patients with HA or PV thrombosis in the first postoperative week, retransplant patients, and recipients of living donor grafts were excluded from the study. Donation after cardiac death donors during the study period were included in the analysis. Intraoperative measurements of PVF and HAF were performed routinely, although data were not available for all patients. Blood flow data and graft weight measurements were available for 312 primary OLT patients. Vascular patency was confirmed with duplex ultrasound on postoperative day (POD) 1 in all cases The severity of IR injury was defined clinically by the serum AST level on POD 2.1,2 Patients were divided into 3 groups:
Intraoperative Data Liver grafts were weighed after back-table preparation of the liver, and the weights were recorded in grams. Intraoperative HAF and PVF values were measured after the completion of all vascular anastomoses and before biliary anastomoses with a transit time ultrasound flowmeter (VeriQ system, software version 2.0, MediStim, Oslo, Norway). HAF and PVF were expressed as milliliters per minute and as milliliters per minute per 100 g of liver tissue. Augmented HAF (ie, HAF when PVF is zero) was also measured. Augmented HAF is used as an indicator of the hepatic artery buffer response (HABR) because it provides an objective measure of the degree of change in HAF when PVF is reduced to zero. PVF was reduced to zero by temporary PV clamping with a vascular clamp. Systemic hemodynamic parameters (ie, the directly transduced mean arterial pressure (MAP), cardiac output (CO), and central venous pressure (CVP)) were recorded at the time of the blood flow measurements, and the cardiac index (CI) was calculated. Additional reviewed data included the transfusions of blood products, the cold ischemia time (CIT), and the operating room time
Recipient Complications All complications were recorded and were divided into broad groups as follows: (1) biliary complications
1450 KELLY ET AL.
LIVER TRANSPLANTATION, December 2011
TABLE 1. Donor Characteristics
Characteristic Age (years) Sex: male/female (%/%) Cause of death [n/N (%)] Anoxia Cerebrovascular accident/trauma Trauma (no head injury) Other Sodium (mmol/L) Bilirubin (mg/dL) AST (U/L)
Statistical Analysis: P Values
Group 1: AST < 500
Group 2: AST ¼ 500-1000
Group 3: AST > 1000
Group 1
Group 1
U/L [n ¼ 160
U/L [n ¼ 85
U/L [n ¼ 67
Versus
Versus
Versus
(51%)]
(27%)]
(21%)]
3 Groups
Group 2
Group 3
Group 3
43 6 18 56/44
48 6 15 63/37
44 6 17 62/38
0.1 0.48
0.041 0.31
0.7 0.35
0.13 0.99
0.06
0.02
0.22
0.57
0.048 0.9 0.35
0.89 0.68 0.27
0.019 0.96 0.66
0.039 0.68 0.15
31/158 (20) 65/158 (41)
6/84 (7) 47/84 (56)
6/66 (9) 33/66 (50)
60/158 (38)
29/84 (35)
27/66 (41)
2/158 (1) 147 6 7 1.0 6 0.8 73 6 81
2/84 (2) 147 6 8 0.9 6 0.5 87 6 174
0 150 6 8 0.8 6 0.5 79 6 108
Group 2
NOTE: The data are presented as means and standard deviations unless otherwise noted. For categorical variables, 3-group and 2-group P values are based on chi-square or Fisher’s exact tests according to the observed event counts. For continuous variables, 3-group and 2-group P values are based on Kruskal-Wallis and Wilcoxon rank-sum tests, respectively. Specific P values are reported when the 3-group P value is 0.10 or less.
(common bile duct leaks or strictures and intrahepatic biliary strictures); (2) postoperative bleeding requiring surgery; (3) sepsis requiring treatment with antibacterial, antiviral, or antifungal agents; (4) cytomegalovirus infections, viremia, or disease; (5) neurological complications; (6) cardiac complications; (7) thromboembolisms; (8) acute abdominal complications requiring surgery; (9) graft-versus-host disease; (10) wound complications; (11) respiratory complications; and (12) late vascular complications (HA thrombosis/stenosis, PV thrombosis/stenosis, and vena cava stenosis).
Analyses were performed with R version 2.8.1 (R Project for Statistical Computing).
RESULTS There were 312 patients in this study: 215 were male, and 97 were female (mean age ¼ 54 6 10 years). The patients’ follow-up ranged from 215 to 1746 days (705 6 408 days). The distribution of the patients in the individual groups was as follows: 160 patients (51%) in group 1 (mild IR injury), 85 patients (27%) in group 2 (moderate IR injury), and 67 patients (21%) in group 3 (severe IR injury).
Statistical Analysis Continuous variables were summarized as means and standard deviations. Group comparisons for categorical variables were performed with chi-square tests or with Fisher’s exact test if the contingency table cell counts were not suitable for chi-square tests. Threegroup comparisons for continuous variables were performed with the Kruskal-Wallis test. Two-group comparisons were performed with the Wilcoxon ranksum test. Three-group comparisons were considered significant when P was less than or equal to 0.05, whereas 2-group comparisons were significant when P was less than 0.017 according to Bonferroni correction tests. Patient survival was estimated with the Kaplan-Meier method, and group comparisons for patient survival were performed with log-rank tests. Multivariate linear regression modeling was used to assess the relationships between blood flow parameters and AST levels on POD 2, and adjustments were made for CIT and donor characteristics as covariates.
Donor and Recipient Characteristics There were 5 donation after cardiac death donors in the series: 1 in group 1, 2 in group 2, and 2 in group 3. The donor characteristics were similar between the groups, although the donor sodium levels were significantly different in the 3-group comparison (P ¼ 0.048) and approached significance for group 3 versus group 1 (P ¼ 0.019; Table 1). The preoperative recipient characteristics were similar across the groups (Table 2) with the exception of the body mass index, which was significantly higher in group 3 (32 6 8 kg/m2) versus group 1 (29 6 6 kg/m2, P ¼ 0.008). As for the severity of illness and indicators of portal hypertension, the calculated MELD scores and the pretransplant platelet counts were similar in the 3 groups. Preoperative CT scans were reviewed for abnormalities in the portal circulation; there were no differences between the groups with respect to PV thrombosis, spontaneous splenorenal
LIVER TRANSPLANTATION, Vol. 17, No. 12, 2011
KELLY ET AL. 1451
TABLE 2. Recipient Characteristics
Group 2: AST ¼ 500-1000
Group 3: AST > 1000
Group 1
Group 1
U/L [n ¼ 160
U/L [n ¼ 85
U/L [n ¼ 67
Versus
Versus
Versus
(51%)]
(27%)]
(21%)]
3 Groups
Group 2
Group 3
Group 3
53 6 10 109/51
55 6 10 63/22
55 6 9 42/24
0.31 0.45
0.19 0.54
0.23 0.40
0.96 0.20
29 6 6
29 6 6
32 6 8
0.025
0.80
0.008
0.034
18.7 6 9 97 6 69 16 (10)
17.7 6 7 108 6 70 8 (9)
17 6 8 104 6 75 7 (11)
0.37 0.15 0.95
0.16 0.06 0.87
0.53 0.51 0.85
0.56 0.24 0.76
33 (21)
16 (19)
13 (20)
0.94
0.72
0.90
0.86
16 (10)
9 (11)
4 (6)
0.60
0.90
0.36
0.34
Characteristic Age (years) Sex: male/female (n/n) Body mass index (kg/m2) Calculated MELD score Pre-OLT platelets K/uL Pre-OLT PV thrombus: CT scan [n (%)] Pre-OLT splenorenal shunt: CT scan [n (%)] TIPS [n (%)]
Statistical Analysis: P Values
Group 1: AST < 500
Group 2
NOTE: The data are presented as means and standard deviations unless otherwise noted. For categorical variables, 3group and 2-group P values are based on chi-square or Fisher’s exact tests according to the observed event counts. For continuous variables, 3-group and 2-group P values are based on Kruskal-Wallis and Wilcoxon rank-sum tests, respectively.
TABLE 3. Etiologies of Liver Disease
Etiology of Liver Disease Hepatocellular carcinoma Alcoholic cirrhosis with or without hepatitis C virus Autoimmune/primary biliary cirrhosis/primary sclerosing cholangitis Hepatitis C virus with or without hepatitis B virus Nonalcoholic steatohepatitis Cryptogenic Miscellaneous
Group 1: AST < 500 U/L
Group 2: AST ¼ 500-1000 U/L
Group 3: AST > 1000 U/L
Statistical
[n ¼ 160 (51%)]
[n ¼ 85 (27%)]
[n ¼ 67 (21%)]
Analysis
40 (25) 37 (23)
29 (34) 16 (19)
24 (36) 3 (5)
Not significant
21 (13)
16 (19)
9 (14)
24 (15)
8 (9)
8 (12)
10 (6) 13 (8) 15 (9)
6 (7) 2 (2) 8 (9)
8 (12) 8 (12) 6 (9)
NOTE: The data are presented as numbers and percentages.
shunts, or TIPS placement before transplantation (Table 2). The liver disease etiologies were similar across the groups and are shown in Table 3.
Intraoperative Data The operating room times were similar for the 3 groups (Table 4). There were no differences between the groups with respect to the requirements for transfusions of packed red blood cells or fresh frozen plasma (Table 4). CIT was significantly longer in group 3 versus group 1 (P ¼ 0.007; Table 4), and the difference was significant in the 3-group comparison (P ¼ 0.022). Duct-to-duct biliary reconstruction was performed in 279 patients (94%), and choledochojeju-
nostomy was performed in 16 patients (5%). There was no difference between the groups. Conventional caval anastomoses were performed in 106 patients (36%), and piggyback anastomoses were performed in 189 (64%).
Graft Characteristics The differences in the graft weights were significant in the 3-group comparison (P ¼ 0.007; Table 4) and particularly for group 3 versus group 1 (P ¼ 0.004). The graft weight differences between groups 2 and 1 did not reach statistical significance. The graft-torecipient weight ratios were similar across the groups (Table 4). The standard liver volume (SLV) was
1452 KELLY ET AL.
LIVER TRANSPLANTATION, December 2011
TABLE 4. Operative Details Group 2: AST ¼ 500-
Group 3:
AST < 500 U/L [n ¼ 160
1000 U/L [n ¼ 85
AST > 1000 U/L [n ¼ 67
3
Group 1 Versus
Group 1 Versus
Group 2 Versus
(51%)]
(27%)]
(21%)]
Groups
Group 2
Group 3
Group 3
488 6 145
506 6 127
497 6 151
0.29
0.09
0.84
0.44
457 6 125 6.7 6 6.1
487 6 128 7.2 6 6.1
511 6 104 6.5 6 7.1
0.022 0.42
0.17 0.47
0.007 0.37
0.17 0.22
7.6 6 10
5.9 6 5.8
6.1 6 7.2
0.69
0.7
0.39
0.69
1565 6 315 1.89 6 0.51
1653 6 300 1.96 6 0.54
1698 6 361 1.90 6 0.53
0.007 0.51
0.046 0.28
0.004 0.44
0.22 0.88
1436.66 6 200.35 1.10 6 0.23
1455.03 6 181.29 1.15 6 0.23
1496 6 241 1.15 6 0.25
0.26 0.08
0.32 0.13
0.13 0.043
0.46 0.40
Factor Operating room time (minutes) CIT (minutes) Packed red blood cells (U) Fresh frozen plasma (U) Graft weight (g) Graft-to-recipient weight ratio (%) Estimated SLV (g) Graft weight/ estimated SLV
Statistical Analysis: P Values
Group 1:
NOTE: The data are presented as means and standard deviations. The 3-group and 2-group P values are based on KruskalWallis and Wilcoxon rank-sum tests, respectively.
TABLE 5. Hepatic and Systemic Hemodynamics
Factor PVF (mL/minute) PVF (mL/minute/100 g) HAF (mL/minute) HAF (mL/minute/100 g) TLBF (mL/minute) TLBF (mL/minute/100 g) PVF (% of TLBF) Augmented HAF (mL/minute/100 g) Cardiac output (L/minute) Cardiac index (L/minute/m2)
Group 1:
Group 2:
Statistical Analysis: P Values
Group 3:
AST < 500
AST ¼ 500-1000
AST > 1000
Group 1
Group 1
Group 2
U/L [n ¼ 160 (51%)]
U/L [n ¼ 85 (27%)]
U/L [n ¼ 67 (22%)]
3 Groups
Versus Group 2
Versus Group 3
Versus Group 3
1799 6 702 119 6 53 294 6 181 20 6 14 2093 6 717 139 6 57 85 6 9 30 6 16
1710 6 777 107 6 53 274 6 148 17 6 9 1984 6 780 124 6 54 84 6 9 25 6 12
1538 6 648 96 6 62 236 6 124 15 6 8 1775 6 642 111 6 63 85 6 8 21 6 10
0.030 0.001 0.14 0.031 0.008
