Predictors of Blood Product Use in Orthotopic Liver Transplantation Using the Piggyback Hepatectomy Technique

Predictors of Blood Product Use in Orthotopic Liver Transplantation Using the Piggyback Hepatectomy Technique R.S. Mangus, S.B. Kinsella, M.M. Nobari, J.A. Fridell, R.M. Vianna, E.S. Ward, R. Nobari, and A.J. Tector ABSTRACT Orthotopic liver transplantation (OLT) has historically been associated with massive blood loss and hemodynamic instability related to the coexistence of varices, coagulopathy, thrombocytopenia, and portal hypertension. Piggyback hepatectomy (PGB) is a technique increasingly utilized in OLT to avoid veno-venous bypass and vena cava clamping. This study evaluated the factors associated with blood loss and blood product requirement in PGB. Methods. This study is a retrospective review of the anesthesia preoperative and operative notes and computerized lab values for all adult cadaveric liver transplants over a 42-month period. These data were combined with the liver transplant database for analysis. Approximately 98% of the transplants were performed using a standard piggyback approach with no use of veno-venous bypass. Results. Data were included for all 526 transplants performed during this time period. Estimated blood loss (EBL) was 1000 cc. Median transfusion requirement was 3 units packed red blood cells, 7 units fresh frozen plasma, and 6 units platelets. Multivariate linear regression demonstrated that predictors of EBL were age, MELD score, preoperative hemoglobin, initial fibrinogen, initial central venous pressure, and total anesthesia time. Predictors of PRBC useage were age, MELD score, preoperative hemoglobin, initial fibrinogen, and anesthesia time. Postoperatively increased transfusion requirement was associated with increased length of hospital stay and lower 90-day and 1-year graft and patient survivals. Conclusion. These results demonstrate that PGB can be safely accomplished in nearly all liver transplant patients without venovenous bypass or vena cava clamping and with less warm ischemia, which may ultimately be associated with less perioperative morbidity and improved outcomes.

O

RTHOTOPIC liver transplantation has been associated with massive blood loss related to the complexity of the procedure and the coexistence of portal hypertension and varices, coagulopathy, and thrombocytopenia. Blood loss, and the associated baseline vasodilatory status of patients in liver failure, can be associated with hemodynamic instability. As the liver transplant procedure has matured, intraoperative blood loss has lessened.1–7 Use of adjunct therapies such as red cell dilution, aprotinin, and Factor VII have been suggested as means of controlling blood loss.8 –13 Technical expertise has improved as surgeons and trainees accumulate more experience in this difficult operation. Several clinical studies have been published that outline important predictors of intraoperative blood loss such as an elevated preoperative international

From the Department of Surgery (R.S.M., J.A.F., R.M.V., A.J.T.), Transplantation Section, Indiana University School of Medicine, Indianapolis, Indiana; Department of Anesthesia (S.B.K.), Indiana University School of Medicine, Indianapolis, Indiana; and Indiana University (M.M.N., E.S.W., R.N.), School of Medicine, Indianapolis, Indiana. This work was supported by the Health Outcomes Research Feasibility Fund, General Clinical Research Center #1367, Indiana University, School of Medicine. Address reprint requests to Richard S. Mangus, MD MS, Transplantation Section, Department of Surgery, Indiana University School of Medicine, 550 N University Blvd, Room 4601, Indianapolis, Indiana 46202-5250. E-mail: [email protected]

© 2007 by Elsevier Inc. All rights reserved. 360 Park Avenue South, New York, NY 10010-1710

0041-1345/07/$–see front matter doi:10.1016/j.transproceed.2007.09.029

Transplantation Proceedings, 39, 3207–3213 (2007)

3207

3208

normalized ratio (INR) thrombocytopenia, preoperative anemia, and previous surgery.14 –20 Additionally, the transjugular intrahepatic portosystemic shunt (TIPS) procedure has become ubiquitous and serves to decompress the portal venous system in patients with severe portal hypertension.21 Piggyback hepatectomy (PGB) is a surgical technique increasingly utilized in both cadaveric and living-donor liver transplantation. This technique was first described by Calne as a means of transplanting small-for-size grafts.22 PGB gained more widespread acceptance and usage when it was described again and popularized by Tzakis.23 The conventional method for liver transplantation requires clamping of both portal flow from the viscera and caval flow from the lower body. PGB is also called the caval preservation technique as it avoids clamping of the vena cava while maintaining flow from the lower body back to the heart throughout the transplant. Preservation of cardiac preload serves to maintain hemodynamic stability and avoid large infusions of fluid volume or vasopressors that may be necessary to maintain blood pressure during the anhepatic phase. With this advantage, PGB also reduces or eliminates the need for venovenous bypass as previously established collaterals with flow through the inferior vena cava maintain forward flow (and portal systemic decompression) throughout the transplant. PGB significantly reduces warm ischemia time because it requires one less anastomosis prior to reperfusion compared to the conventional technique. With a shorter warm ischemia time, it is possible to avoid altogether redirecting portal flow simply by proceeding with the transplant at a rapid pace and decompressing portal flow through the transplanted liver. PGB is technically more demanding and time consuming than the conventional approach. Additional time is required to dissect the liver off the native vena cava and to ligate the short hepatic veins, which increases the risk of heavy venous bleeding from the vena cava. Additionally, a large caudate lobe or an intrahepatic vena cava increases the complexity and risk of this technique. Our center has used PGB almost exclusively over the past 5 years. This study was conducted to determine the predictors of blood loss using PGB at a high volume center and to calculate blood product usage with this increasingly utilized technique for liver transplantation. Blood product usage was then studied as a predictor for overall patient outcomes, including hospital length of stay and graft and patient survival.

MANGUS, KINSELLA, NOBARI ET AL liver transplant or a simultaneous cadaveric liver and kidney transplant. Graft and patient survival data were collected from the transplant database. In patients receiving retransplantation within 30 days of the original transplant, the analysis included only patient and graft survival data for the first transplant. For this cohort, there were no ABO mismatches. While patients receiving combined liver and kidney transplantation were included in the analysis, those receiving other multiorgan transplants were excluded. All recipients were listed for transplantation according to standard procedures and protocols as established by the United Network for Organ Sharing (UNOS). Mean MELD score at transplant was 18 years (range 6 – 48). MELD was developed as a predictor of the risk of death in patients with end-stage liver disease awaiting transplantation. The MELD score is constructed using the patient’s age, serum total bilirubin level and creatinine, and INR. Certain variables could not be reliably ascertained from the available data including the presence of preoperative bleeding or varices and the presence or extent of portal vein thrombosis. Donor livers were recovered using standard procurement techniques including aortic and portal vein flushing and cold storage. The median cold ischemia time was 7 hours with median warm ischemia time of 34 minutes. Ninety-eight percent of the transplants during the study period were performed using a standardized piggyback hepatectomy approach that was equivalent for all participating surgeons. Briefly, the hepatectomy was approached through a standard bilateral subcostal incision with upper midline extension. In no case was venovenous bypass utilized. The liver was mobilized off the inferior vena cava after ligation and transection of the common bile duct, hepatic artery, and portal vein in the liver hilum. The right hepatic vein was clamped and transected separately from the middle and left hepatic veins. After removal of the liver, the transplant liver was anastomosed either to the common cuff of the middle and left hepatic veins or to the joined right, middle, and left veins, depending on the size of the donor vena cava. The hepatic vein clamp may partially occlude the vena cava flow to the heart. After the hepatic vein anastomosis was completed, the clamp was moved from a side-bite of the vena cava to a position solely on the hepatic veins, thus restoring full vena cava flow as early as possible. The preservation solution was flushed from the donor liver prior to reperfusion with a 3000 mL flush of 5% albumin. Cases in which PGB could not be performed included patients with tumor abutting the vena cava, Budd-Chiari syndrome, and rare cases in which the vena cava was very friable and received considerable iatrogenic damage during the hepatectomy. Median anesthesia time was 4 hours 40 minutes. A temporary portal-cava shunt can be constructed in the PGB technique, but this was not done in any of the included cases. There were no cases in which either red blood cell salvage or Factor VII were utilized. Aprotinin was given routinely as a continuous infusion throughout the transplant procedure at a rate of 70 mg/hour.

Statistics MATERIALS AND METHODS The medical records of all orthotopic liver transplants performed between January 1, 2003 and June 30, 2006 were reviewed (42 months). Extracted data came from the comprehensive transplant recipient registry at our center and individual written and electronic medical records. The original anesthesia preoperative and operative records were reviewed to determine usage of packed red blood cells (PRBC), fresh frozen plasma (FFP), platelets, and cryoprecipitate. Recipient inclusion criteria included all transplant recipients age 18 and older receiving either an isolated orthotopic

Primary outcomes included intraoperative blood loss, as recorded by the anesthesiologist, and use of PRBCs, FFP, and platelets. Patient and graft survival at 3 months and 1 year were analyzed using a statified comparison by intraoperative use of blood products. All occurrences of graft loss or patient death were included in the final analysis regardless of etiology, comorbidities, or timing. Specifically, there were no exclusions for perioperative mortality or graft loss or for nonliver-related deaths. Subgroup analysis was performed using chi-square testing to individually identify those factors predictive of intraoperative use of blood products. The

BLOOD PRODUCT USE IN PIGGYBACK LIVER TRANSPLANT

3209

(Table 1). Eighty-three recipients required no PRBCS intraoperatively (17.5%), and 44 (9.3%) required more than 10 units. The primary predictors of increased blood loss on bivariate analysis included higher recipient MELD score and history of major abdominal surgery. Laboratory values associated with increasing estimated blood loss include: lower starting hemoglobin, higher starting INR, lower starting platelet count, higher preoperative creatinine, higher initial central venous pressure (CVP) and initial fibrin degradation products and a lower initial intraoperative fibrinogen level. Blood loss also increased with increasing total anesthesia time (Table 2). All of these same factors were similarly significant predictors of need for PRBC transfusion. Use of FFP increased with male gender, higher BMI, higher MELD score, and previous major abdominal surgery. FFP and platelet use was increased in a manner similar to blood loss and PRBC useage for laboratory values, initial CVP, and anesthesia time. MELD score was a consistent predictor of blood loss and need for PRBCs, FFP, and platelets.

linear regression model was constructed using a direct entry method with covariates removed from the model for P ⬎ .10. Statistical testing was performed on SPSS software (SPSS 14.0 for Windows, SPSS, Chicago, Ill, USA). This study was reviewed and approved by the institutional review board of the Indiana University School of Medicine.

RESULTS Demographics

There were 526 liver transplants performed during the 42-month study period, and all data from these subjects were included in the analysis. Recipient demographics are listed in Table 1 Sixty-five percent of recipients were male with a history of major abdominal surgery in 10.1% and TIPS in 10.8%. Mean recipient age was 52.3, body mass index (BMI) 28.1, and MELD 18.1. Blood Loss and Transfusion Predictors

Median blood loss was 1000 cc. Median transfusion requirement was 3 units PRBC, 7 units FFP, and 6 units platelets

Table 1. Demographic Data and Blood Product Usage for 526 Liver Transplants from 2001 to 2006 Using Primarily Piggyback Hepatectomy With Continuous Aprotinin Infusion Characteristic

Overall Mean; median; range Recipient gender Male Female Recipient age (years) Less than 40 40 to 49 50 to 59 60 and older Recipient body mass index Less than 25.0 25.0 to 29.9 30.0 to 34.9 35.0 and higher MELD score Less than 15 15 to 24 25 and higher Primary diagnosis Viral hepatitis Alcoholic liver disease (ALD) Viral hepatitis and ALD Cholestatic liver disease Previous abdominal surgery None Minor Major Previous TIPS procedure No Yes

N (%) (mean, median)

Blood Loss (cc) (mean, median)

P

PRBCs (mean, median)

P

FFP (mean, median)

P

PLTS (mean, median)

P

526 (100%) 1396, 1000, 150 to 18,500

4.7; 3; 0 to 74 NS

340 (64.5%) 187 (35.5%) 52.6, 52 50 (9.5%) 142 (26.9%) 204 (38.7%) 131 (24.9%) 28.1, 28.0 156 (29.6%) 187 (35.5%) 129 (24.5%) 55 (10.4%) 18.1, 17 156 (29.6%) 187 (35.5%) 55 (10.4%)

1465, 1000 1264, 1000

164 (31.1%) 62 (11.8%) 77 (14.6%) 224 (42.5%)

1280, 800 1622, 1000 1431, 1000 1404, 1000

366 (69.4%) 108 (20.5%) 53 (10.1%)

1334, 1000 1217, 1000 2163, 1000

470 (89.2%) 57 (10.8%)

1373, 1000 1583, 1200

7.7; 7; 0 to 62 NS

5.0, 3 4.2, 3 NS

1739, 1000 1313, 1000 1305, 1000 1495, 850

NS

NS

NS

6.6, 6 7.5, 6 8.4, 8 9.9, 10 ⬍.001

3.1, 3 4.8, 3 10.2, 7 NS

⬍.001

NS

⬍.01

NS

⬍.01

NS

NS 9.7, 6 7.9, 6 10.0, 10 7.1, 0

⬍.01 7.5, 6 7.4, 7 10.3, 8

NS 4.6, 3 5.7, 5

⬍.01 6.9, 0 8.3, 6 13.6, 8

7.6, 7 7.8, 7 7.6, 8 7.7, 6

4.4, 3 4.6, 4 7.4, 4

NS 8.2, 6 8.5, 6 7.3, 0 11.0, 10

6.9, 7 7.6, 6 11.2, 10

4.4, 3 5.2, 3 3.6, 3 5.2, 4

NS 10.8, 0 7.8, 6 9.2, 6 6.7, 0

⬍.01

NS

⬍.001

NS 8.0, 6 9.0, 6

7.1, 5 8.5, 8 7.7, 7 7.0, 6

4.7, 3 4.4, 3 5.0, 3 5.1, 3

1056, 800 1433, 1000 2441, 1500

0.01 8.2, 7 6.7, 6

6.1, 3 4.0, 3 4.4, 3 5.5, 4

1422, 1000 1299, 825 1352, 1000 1786, 1000

8.4; 6; 0 to 80

NS 8.4, 5 8.0, 6 8.7, 4

NS 7.6, 6 8.5, 8

NS 8.2, 6 9.6, 6

P values calculated using one-way ANOVA. P ⬎ .10 listed as NS. PRBCs: packed red blood cells, FFP: fresh frozen plasma, PLTS: platelets, MELD: model for end-stage liver disease score, TIPS: transjugular intrahepatic portosystemic shunt.

3210

MANGUS, KINSELLA, NOBARI ET AL

Table 2. Intraoperative Data for Entire Sample, and Comparison of Intraoperative Coagulation Factors and Operative and Graft Ischemia Times, Stratified by Blood Transfusion Requirement Characteristic

Overall Starting hemoglobin (g/L) Less than 10.0 10.0 to 12.9 13.0 and higher Starting INR value Less than 1.3 1.3 to 1.6 1.6 and higher Starting platelet count 50,000 51,000 to 99,000 100,000 and higher Starting CVP (mm Hg) 8 9 to 14 15 and higher Preoperative creatinine Less than 1.0 1.0 to 1.4 1.5 and higher First fibrinogen level Less than 150 150 to 250 250 and higher First FDP Less than 0.8 0.81 to 3.0 Higher than 3.0 Total anesthesia time (minutes) Less than 4:00 hours 4:01 hours to 5:59 6:00 hours or more

N (%) Blood Loss (cc) (mean, median) (mean, median)

526 11.8, 12.0 92 (17.5%) 283 (53.7%) 152 (28.8%) 1.7, 1.4 174 (33.0%) 250 (47.4%) 103 (19.5%) 99, 78 88 (16.7%) 277 (52.6%) 162 (30.7%) 12, 12 105 (19.9%) 300 (56.9%) 122 (23.1%) 1.2, 1.0 235 (44.6%) 209 (39.7%) 83 (15.7%) 209, 191 135 (25.6%) 279 (52.9%) 113 (25.6%) 2.1, 1.4 141 (26.8%) 247 (46.9%) 139 (26.4%) 4:57, 4:40 105 (19.9%) 337 (63.9%) 85 (16.1%)

P

1396, 1000

PRBCs (mean, median)

P

4.7, 3 ⬍.001

1871, 1200 1451, 1000 1010, 700 ⬍.01

⬍.001

⬍.01

.02 1225, 800 1298, 1000 1751, 1000

12.0, 10 7.6, 0 4.6, 0 .02

7.3, 7 7.2, 6 8.9, 8 ⬍.001

2.5, 2 4.1, 3 9.5, 6

⬍.001

.02

⬍.001

⬍.001

.03 7.7, 6 7.5, 6 11.8, 6

9.0, 8 7.3, 6 6.9, 7

3.1, 2 4.6, 3 6.5, 5

869, 600 1202, 1000 2727, 1775

.01

⬍.01

⬍.01

NS 6.7, 0 8.7, 6 9.0, 6

7.0, 6 7.8, 6 9.4, 8

6.5, 5 4.2, 3 3.8, 3

1007, 700 1503, 1000 1620, 1000

.05

⬍.001

⬍.01

⬍.001 15.3, 10 9.0, 6 2.8, 0

6.7, 6 7.6, 6 8.7, 7

3.4, 3 5.3, 4 7.4, 5

1807, 1000 1291, 1000 1130, 800

⬍.01

.02

⬍.001 1121, 800 1505, 1000 1965, 1400

.08 6.4, 0 9.5, 6 8.9, 10

9.5, 8 7.7, 7 6.7, 6

4.2, 2 4.3, 3 6.1, 4

.02 11.1, 6 8.5, 6 6.2, 0

⬍.001

⬍.01

NS 7.8, 6 7.6, 0 10.0, 6

⬍.001 4.8, 5 7.5, 7 11.8, 10

P

8.4, 6

6.4, 5 7.9, 7 9.4, 8

6.6, 3 4.7, 4 3.8, 3

PLTS (mean, median)

.01 9.2, 8 7.7, 6 6.8, 6

3.6, 3 4.5, 3 7.0, 5

1938, 1000 1392, 1000 1113, 800

P

7.7, 7 ⬍.001

7.5, 5 4.9, 4 2.8, 1

1120, 800 1406, 1000 1831, 1225

FFP (mean, median)

⬍.001 3.7, 0 8.4, 6 13.9, 10

P values calculated using one-way ANOVA. P ⬎ .10 listed as NS. PRBCs: packed red blood cells, FFP: fresh frozen plasma, PLTS: platelets, INR: international normalized ratio, CVP: central venous pressure, FDP: fibrin degradation products.

Two linear regression models were constructed to determine those covariates predictive of blood loss and PRBCs useage (Table 3). The most important predictors of blood loss were preoperative hemoglobin, MELD score, and initial CVP. For PRBC useage, the most important predictors included preoperative hemoglobin and MELD score. Patient Outcomes

The median hospital length of stay overall was 11 days, and it was found to increase incrementally with increasing intraoperative blood loss and increasing requirement for PRBCs (Table 4). Three-month and 1-year graft and patient survivals decreased with increasing intraoperative blood loss and increasing requirement for PRBCs. DISCUSSION

These results demonstrate that blood product use for the PGB technique in liver transplantation is similar to that

published in analyses of liver transplantation using the conventional hepatectomy method. However, PGB can be safely accomplished without venovenous bypass or clamping of the inferior vena cava and with a shorter warm ischemia time. These results support the argument that the PGB is a superior approach to liver transplantation and does not incur a penalty related to intraoperative blood loss or requirement for blood products. PGB may, in fact, be the preferred method in high-risk patients such as the elderly or those with poor physiologic reserve. PGB tends to preserve hemodynamic and physiologic stability throughout the transplant, which may be associated with less perioperative morbidity and mortality. Avoidance of venovenous bypass removes the risk of catheter-related injuries to the axilla and groin, as well as access site infections. PGB and the conventional method have been compared previously for multiple surgical outcomes. One of the first prospective studies, in 2000, included 90 patients who were

BLOOD PRODUCT USE IN PIGGYBACK LIVER TRANSPLANT

3211

Table 3. Linear Regression Analysis for Calculation of Estimated Blood Loss and Intraoperative Requirement for Packed Red Blood Cells Model 1 (EBL)

Beta

Standard Error

Significance

Mean

Constant Age at transplant MELD score Preoperative Hemoglobin Initial fibrinogen in OR Initial CVP in OR OR anesthesia time (min)

⫺1354.696 16.413 27.192 ⫺84.384 ⫺1.757 28.297 0.135

716.77 6.895 9.987 32.91 0.735 14.368 0.013

0.059 0.018 0.007 0.011 0.017 0.05 0.0001

52.57 18.07 11.815 208.75 12.17 297

Model 1

R

R2

Adjusted R2

Standard Error

0.276

0.266

1410.47

ANOVA Model 1

0.525 Significance ⬍0.0001

Model 2 (PRBCs)

Beta

Standard Error

Significance

Mean

Constant Age at transplant MELD score Preoperative Hemoglobin Initial fibrinogen in OR OR anesthesia time (min)

⫺2.2112 0.071 0.164 ⫺0.599 ⫺0.007 0.001

2.351 0.025 0.036 0.117 0.003 0.001

0.37 0.004 ⬍0.001 ⬍0.001 0.009 ⬍0.001

52.57 18.07 11.815 208.75 297

Model 2

R

R2

Adjusted R2

Standard Error

0.575 Significance ⬍0.001

0.33

0.323

5.103

ANOVA Model 2

EBL: estimated blood loss, PRBC: packed red blood cells, MELD: model for end-stage liver disease, OR: operating room; CVP: central venous pressure.

between piggyback and conventional techniques in posttransplant pulmonary function but did note an increased risk for postoperative infiltrates related to piggyback hepatectomy.25 Moreno-Gonzalez et al reported on 50 patients assigned to three groups: (1) PGB, (2) conventional approach with bypass, and (3) conventional approach without bypass. The three groups did not differ in postoperative complications, including renal failure. PGB did require a longer operative time but was actually associated with a decreased requirement for PRBCs (1 ⫽ 12 units, 2 ⫽ 18

assigned to PGB or conventional approach with venovenous bypass. Use of PGB significantly shortened the anhepatic phase (41 min vs 121 min), total operative time (425 min vs 486 min), and lowered the transfusion requirements (PRBCs: 8.9 vs 12.1 units; FFP: 5.0 vs 7.4 units; platelets: 32.9 vs 84.8 units), as well as the total infused volume of fluid (12.3 L vs 15.4 L). These results translated into a shorter ICU and hospital stay and in lower overall hospital charges.24 Isern and colleagues conducted a randomized trial of 67 patients and found no differences

Table 4. Liver Transplant Outcomes for All Patients and Comparison of Outcomes for Recipients Based on Intraoperative Blood Loss and Packed Red Blood Cell Transfusion Requirement 3-Month Survival Length of Hospital Stay (mean; median; range)

Overall Estimated blood loss Less than 500cc 500cc to 1500cc 1500cc to 2500cc More than 2500cc PRBCs transfused Less than 3 3 to 6 units 7 to 10 units More than 10 units

15.1; 11; 0–250 15.2, 9 14.6, 11 13.6, 12 19.5, 12 13.7, 10 14.8, 12 18.0, 13 19.7, 12

1-Year Survival

Graft

Patient

Graft

Patient

90.3% NS 91.9% 91.7% 87.5% 82.5% NS 93.2% 89.6% 90.4% 81.4%

91.3% NS 94.6% 92.0% 89.3% 84.2% 0.06 94.2% 90.5% 92.3% 81.4%

83.3% NS 82.1% 84.8% 86.7% 70.7% NS 86.7% 82.8% 81.1% 72.7%

84.7% NS 85.7% 85.5% 88.9% 73.2% NS 88.0% 84.2% 83.8% 72.7%

Statistical comparison made using the chi-square test with P ⬍ .05 significant. LOS: length of stay, EBL: estimated blood loss, PRBCs: packed red blood cells.

3212

units, 3 ⫽ 24 units); crystalloids (1 ⫽ 3.1 L, 2 ⫽ 6.8 L, 3 ⫽ 9.1 L); and vasoactive drugs (1 ⫽ 18%, 2 ⫽ 24%, 3 ⫽ 38%).26 Cabezuelo et al conducted a similar retrospective study of 184 transplants with comparison of the same three study groups and found the conventional technique to have higher odds of postoperative renal failure compared with the PGB method (OR 3.1, P ⫽ .01).27 Zieniewicz and colleagues reported significantly lower blood loss (3.4 L PGB vs 8.4 L conventional) and shorter warm ischemia time (58 min PGB vs 78 conventional) for PGB with no difference in operative time.28 Complications reported to be associated with PGB include venous outflow obstruction, acute Budd-Chiari syndrome, and postoperative ascites. Etiologies for blood loss associated with OLT can be categorized as either surgical or medical in nature. Surgical aspects include the technical difficulty of the procedure and anatomic variations or changes that may predispose to blood loss such as adhesions, infection, varices, or portal vein thrombus. Examples of medical factors include the cause and the extent of liver disease, preoperative hemoglobin level, thrombocytopenia, coagulopathy related to liver disease, and concomitant renal failure. Presumably, both PGB and conventional approaches will be subjected to similar medical etiologies, though the length of each procedure may exacerbate some of the medical factors through consumption of clotting factors and platelets and prolongation of hypothermia. Review of the surgical factors contributing to blood loss would seem to favor a conventional approach as predictive of lower blood loss given its relative technical ease and less requirement for fine dissection on the vena cava. PGB is technically more demanding and time consuming than the conventional approach. Additional time is required to dissect the liver off the native vena cava, and this increases the risk of heavy venous bleeding from the caudate branches of the liver. Additionally, a very large caudate lobe or an intrahepatic vena cava adds an additional measure of risk and complexity. In this case series, there were very few transplants (⬍2%) in which PGB could not be accomplished. It has been suggested that the impact of portal hypertension on blood loss in PGB may be lessened with the use of a temporary portal cava shunt.29 Alternatively, the consequences of a clamped portal system can be lessened by decreasing the length of time that the portal flow is interrupted. All of the transplants in this series were completed without use of a portal-cava shunt. The portal flow was interrupted on average for 60 to 90 minutes. There was no clinical impact noted for portal clamping within this time range. Other mechanisms for intraoperative blood loss minimization include blood volume dilution by anesthesia, permissive hypotension, and red blood cell salvage. Our center does not practice these techniques, though anesthesia does attempt to maintain central venous pressure less than 10 mm Hg throughout. The goal level for hemoglobin during the transplant is 10 g/dL. Our center does use aprotinin infusion throughout the transplant. Clinically, the use of aprotinin results in decreased blood loss and expe-

MANGUS, KINSELLA, NOBARI ET AL

dites the transplant with a dryer operative field.12,13,30 The rate of hepatic artery thrombosis and portal vein thrombosis in this sample was ⬍1%, and, therefore, the thrombotic risk of aprotinin appears to be minimal in our population. Also, the rate of aprotinin-related renal insufficiency was minimal, and aprotinin was used in the combined liver and kidney transplants without a clinically notable negative effect on renal allograft function. Our results for blood loss and transfused PRBCs, FFP, and platelets are similar to, or less than, amounts reported by other centers. Sixty-four percent of recipients received 4 units PRBC or less. As expected, the highest requirements were in patients with less favorable anatomy (high BMI, previous surgery, or transjugular intrahepatic portosystemic shunt), worse disease (high MELD score), coagulopathic (higher INR), preoperative anemia, and elevated preoperative creatinine. All of these factors are known to impact use of blood products, and these data support the representative value of this sample of recipients. A shorter duration of surgery was associated with less need for PRBCs. Fifty-two percent of these transplants were completed with less than 5 hours total intraoperative anesthesia time. The use of FFP and platelets was found to vary widely depending on the anesthesia team, and subgroup analysis of these blood products was not conducted for this sample. Likewise, the recording of estimated blood loss was taken from the anesthesia records alone and was not confirmed by surgeon input or by independent measure. Therefore, we feel that the best indicator for blood loss is the use of PRBCs, as this is a defined, reproducible number. Unfortunately, even this measure is inadequate as it fails to account for preoperative anemia, which will lead to a larger transfusion requirement independent of the operative procedure. Previous authors have reported an association between intraoperative blood loss and perioperative outcomes.14,16,18 –20 Results of this study support that finding. A high intraoperative transfusion requirement was associated with a longer hospital stay and lower 3-month and 1-year posttransplant survivals. Though this finding has been frequently reported, it is unlikely that blood loss and transfusion requirement directly impact perioperative outcomes. A more plausible explanation is that high-risk patients with more comorbidities and a higher MELD score have a longer hospital stay and worse survival. As seen in Table 2, these are the same patients with high intraoperative transfusion requirements. Intraoperative blood loss and transfusion requirement are, in many cases, simply markers for the high-risk liver transplant recipient. In conclusion, the PGB method for liver transplantation can be safely and effectively performed in the majority of liver transplant recipients. Blood loss and blood product requirements with PGB are similar to, or better than, those for the conventional technique. These results support previously published studies demonstrating PGB potentially to be a superior technique given its benefits of avoiding veno-venous bypass, maintaining hemodynamic and physi-

BLOOD PRODUCT USE IN PIGGYBACK LIVER TRANSPLANT

ologic stability, and decreasing warm ischemia time. This data set represents a large population of transplant recipients, with 98% undergoing liver transplantation by essentially the same PGB technique, without use of bypass or shunting. Preoperative and intraoperative variables and recipient outcomes associated with transfusion requirement are equivalent to those previously reported in studies of transplantation with the conventional method. REFERENCES 1. Yuasa T, Niwa N, Kimura S, et al: Intraoperative blood loss during living donor liver transplantation: an analysis of 635 recipients at a single center. Transfusion 45:879, 2005 2. Frasco PE, Poterack KA, Hentz JG, et al: A comparison of transfusion requirements between living donation and cadaveric donation liver transplantation: relationship to model of end-stage liver disease score and baseline coagulation status. Anesth Analg 101:30, 2005 3. Li Pi Shan W, Barkun J, Metrakos P, et al: Blood product use during orthotopic liver transplantation. Can J Anesth 5110:1045, 2004 4. Massicotte L, Sassine MP, Lenis S, et al: Transfusion predictors in liver transplant. Anesth Analg 98:1245, 2004 5. Ozier Y, Pessione F, Samain E, et al: Institutional variability in transfusion practice for liver transplantation. Anesth Analg 97:671, 2003 6. Tully M, Burkle C, Plevak D, et al: Pilot study to determine blood and blood component transfusion differences between patients receiving orthotopic cadaveric versus living related donor liver transplant. Liver Transpl 8:C1, 2002 7. Steib A, Freys G, Lehmann C, et al: Intraoperative blood losses and transfusion requirements during adult liver transplantation remain difficult to predict. Can J Anesth 4811:1075, 2001 8. Lentschener C, Roche K, Ozier Y: A review of aprotinin in orthotopic liver transplantation: Can its harmful effects offset its beneficial effects? Anesth Analg 100:1248, 2005 9. De Gasperi A, Baudo F, De Carlis L: Recombinant FVII in orthotopic liver transplantation (OLT): a preliminary single center experience. Intensive Care Med 31:315, 2005 10. Schroeder RA, Collins BH, Tuttle-Newhall E, et al: Intraoperative fluid management during orthotopic liver transplantation. J Cardiothoracic Vasc Anesth 184:438, 2004 11. Porte RJ, Hendriks HGD, Slooff MJH: Blood conservation in liver transplantation: The role of aprotinin. J Cardiothoracic Vasc Anesth 184:31S, 2004 12. Findlay JY, Rettke SR, Ereth MH, et al: Aprotinin reduces red blood cell transfusion in orthotopic liver transplantation: a prospective, randomized, double-blind study. Liver Transpl 7:802, 2001 13. Porte RJ, Molenaar IQ, Begliomini B, et al: Aprotinin and transfusion requirements in orthotopic liver transplantation: a multicentre ranomised double-blind study. EMSALT study group. Lancet 355:1303, 2000 14. Ramos E, Dalmau A, Sabate A, et al: Intraoperative red blood cell transfusion in liver transplantation: influence on patient

3213 outcome, prediction requirements, and measures to reduce them. Liver Transpl 9:1320, 2003 15. Pirat A, Sargin D, Torgay A, et al: Identification of preoperative predictors of intraoperative blood transfusion requirement in orthotopic liver transplantation. Transplant Proc 34:2153, 2002 16. Bennett-Guerrero E, Feierman DE, Barclay GR, et al: Preoperative and intraoperative predictors of postoperative morbidity, poor graft function, and early rejection in 190 patients undergoing liver transplantation. Arch Surg 136:1177, 2001 17. Findlay JY, Rettke SR: Poor prediction of blood transfusion requirements in adult liver transplantations from preoperative variables. J Clin Anesth 12:319, 2000 18. Cacciarelli T, Keefe E, Moore D, et al: Effect of intraoperative blood transfusion on patient outcome in hepatic transplantation. Arch Surg 134:25, 1999 19. Palomo-Sanchez JC, Jimenez C, Moreno-Gonzalez E, et al: Effects of intraoperative blood trnsfusion on postoperative complications and survival after orthotopic liver transplantation. Hepatogastroenterology 45:1026, 1998 20. Mor E, Jennings L, Gonwa TA, et al: The impact of operative bleeding on outcome in transplantation of the liver. Surg Gynecol Obstet 176:219, 1993 21. Boyer TD, Haskal ZJ: AASLD practice guidelines: The role of transjugular intrahepatic portosystemic shunt creation in the management of portal hypertension. J Vasc Interv Radiol 16:615, 2005 22. Calne RY, Williams R: Liver transplantation in man. I. Observations on technique and organization in five cases. Br Med J 4630:535, 1968 23. Tzakis A, Todo S, Starzl TE: Orthotopic liver transplantation with preservation of the inferior vena cava. Ann Surg 2105:649, 1989 24. Shokouh-Amiri MH, Osama Gaber A, Bagous WA, et al: Choice of surgical technique influences perioperative outcomes in liver transplantation. Ann Surg 2316:814, 2000 25. Isern MRM, Massarollo PCB, de Carvalho EM, et al: Randomized trial comparing pulmonary alterations after conventional with venovenous bypass versus piggyback liver transplantation. Liver Transpl 103:425, 2004 26. Moreno-Gonzalez E, Meneu-Diaz JG, Fundora Y, et al: Advantages of the piggyback technique on intraoperative transfusion, fluid compensation, and vasoactive drug requirements in liver transplantation: A comparative study. Transplant Proc 35:1918, 2003 27. Cabezuelo JB, Ramirez P, Acosta F, et al: Does the standard vs piggyback surgical technique affect the development of early acute renal failure after orthotopic liver transplantation? Transplant Proc 35:1913, 2003 28. Zieniewicz K, Krawczyk M, Nyckowski P, et al: Liver transplantation: Comparison of the classical orthotopic and piggyback techniques. Transplant Proc 34:625, 2002 29. Tzakis AG, Reyes J, Nour B, et al: Temporary end-to-side portocaval shunt in orthotopic hepatic transplantation in humans. Surg Gynecol Obstet 176:180, 1993 30. Marcel RJ, Stegall WC, Suit T, et al: Continuous small-dose aprotinin controls fibrinolysis during orthotopic liver transplantation. Anesth Analg 82:1122, 1996