Frasco, Peter E. MD; Poterack, Karl A. MD; Hentz, Joseph G. MS; Mulligan, David C. MD
Because of cadaveric donor shortages, the use of living donation is an option at some transplant centers for patients awaiting liver transplantation. The MELD (model for end-stage liver disease) score is a weighted sum of the serum creatinine concentration, prothrombin time (PT), serum bilirubin concentration, and etiology of end-stage liver disease (1). The MELD score was proposed as a tool to predict outcome after transjugular intrahepatic portosystemic shunt performed as decompressive therapy for portal hypertension. It has become a generalized marker for liver disease severity that has replaced the Child-Turcotte-Pugh (CTP) classification and United Network for Organ Sharing (UNOS) urgency statuses 2A, 2B, and 3 for prioritization of orthotopic liver transplantation (OLT) (2–4). Those patients in whom living donation is possible bypass this prioritization procedure and may have lower MELD scores at the time of transplantation.
There have been reports that it is possible to predict intraoperative blood loss and component replacement therapy during OLT by the use of preoperative coagulation tests or indices of liver disease severity such as the CTP score (5–7). The reproducibility of these associations has been challenged in subsequent studies (8–10). In addition, in a study of cadaveric donation OLT, patients with less advanced liver disease, as defined by CTP score, were more likely to undergo OLT without transfusion of packed red blood cells (PRBCs) (11).
The MELD score may be an improvement over the CTP, as it includes both a measure of coagulation function (PT) and etiology of end-stage liver disease. The addition of a variable that considers the etiology of liver disease may be important in predicting blood loss and blood component therapy. Although some studies present conflicting data (9,10,12), patients with end-stage liver disease secondary to viral hepatitis or alcoholic cirrhosis may have increased intraoperative blood loss and transfusion requirements compared with patients having primary biliary cirrhosis or sclerosing cholangitis (7,10,13). The MELD score may, therefore, be a better marker of perioperative blood loss and transfusion requirements than the CTP score. Despite this advantage, inspection of a more complete panel of coagulation tests (e.g., platelet concentration and function, fibrinogen concentration) may also be important because the MELD score only incorporates the PT.
There are relatively few data regarding the use of blood products in living donor OLT (9,14) and the potential relationship of preoperative MELD score and intraoperative blood component therapy (2,8).
We hypothesized that patients undergoing living donation OLT would undergo the OLT procedure at lower MELD scores with preserved coagulation function and require less component therapy transfusion compared with patients undergoing cadaveric donation OLT. In addition, we examined the relationship between MELD score and baseline laboratory data and blood component therapy in these two populations.
After IRB approval, we reviewed the charts for all primary OLTs after the introduction of a living donation OLT program at our facility performed by the senior surgeon (DCM) for a 3-yr period from April 2001 through March 2004. The perioperative and anesthetic management of patients during OLT was similar regardless of donation source.
Standard noninvasive monitors, radial or brachial arterial catheter, continuous cardiac output/mixed venous oximetry pulmonary artery catheter, and bladder catheter were used in all patients. All patients were treated with upper and lower extremity forced-air warming units. Cell salvage and the subsequent transfusion of processed cell salvaged blood was used in all patients except for those patients with known hepatocellular malignancy, increased carcinoembryonic antigen or α-fetoprotein levels, or patients with known or suspected bacterial peritonitis. All patients received an aprotinin (Trasylol®; Bayer Corporation, West Haven, CT) infusion (500,000 kallikrein-inhibiting units [KIU] bolus followed by infusion of 250,000 KIU/h) during the OLT procedure. Fluid therapy with Plasmalyte® (Baxter Healthcare, Deerfield, IL) and 5% albumin solution was guided by central venous pressure, urine output, and clinical condition. Blood samples were obtained at incision, 5 min before the anhepatic interval, 10 min into the anhepatic interval, 5 min before recirculation, 10 min after recirculation, and 70 min after recirculation. Arterial blood gas analysis, hemoglobin concentration, hematocrit, platelet count, thromboelastography, PT with international normalized ratio (INR), partial thromboplastin time, fibrinogen concentration, and d-dimer concentration were measured from these samples. Thromboelastography (TEG®) was performed with a 5000 series Thromboelastograph® (Haemoscope, Skokie, IL) using 360 μL of whole, native, noncitrated blood. Transfusion was guided by an algorithm based on these results (Fig. 1) and clinical condition (e.g., acuity of blood loss, changes in arterial blood pressure, central venous pressure, pulmonary capillary wedge pressure). All transplant procedures were performed without venovenous bypass using the piggyback (vena caval preservation) technique described by Figueras et al. (15). The technique used for donation harvest of the right hepatic lobe and implantation in the living donation transplant group was originally described by Marcos et al. (16).
Medical records were reviewed for procedure performed (cadaveric donation OLT or living donation OLT), primary hepatic disease, preoperative MELD score and baseline TEG® (“r” time, maximum amplitude [MA], and α angle), hematology, and lab indices of coagulation status. The formula for the MELD score is 3.8 × ln(bilirubin (mg/dL)) + 11.2 × ln(INR) + 9.6 × ln(creatinine (mg/dL)) + 6.4. Intraoperative blood product use and intraoperative cell salvage (if used) was recorded. Lowest core body temperature, graft ischemic time (time from clamping of donor inflow to reperfusion via portal system), anhepatic time (time from clamping portal vein and infrahepatic and suprahepatic vena cava to reperfusion via the portal system) and duration of surgical procedure (time from incision to closure) were recorded. Total volumes of crystalloid and albumin infusion were recorded. Blood product usage was quantified by counting the number of treatments of components. One unit of PRBC, one unit of fresh-frozen plasma (FFP), one unit of platelets obtained from plasma pheresis from a single donor, or a 10-unit dose of cryoprecipitate were each considered to be a single treatment.
Mean levels of preoperative measures and the number of treatments were compared between the two types of OLT. The statistical significance was calculated by using the two-sample Student’s t-test. The relationships between preoperative measures and measures of blood product usage were calculated by using linear regression. Multiple linear regression and a forward selection strategy were used to determine which factors were most strongly associated with the number of transfusions. The type of transplant, MELD, INR, r time, α angle, MA, platelet count, fibrinogen concentration, and hemoglobin concentration were assessed. The transfusion of blood products was modeled with multiple logistic regression. P < 0.05 was considered significant.
There were 69 patients in the cadaveric donation OLT group and 27 patients in the living donation OLT group during the study period. Diagnosis at time of transplantation is illustrated in Figure 2. There were no important differences in diagnosis at OLT between the groups. Perioperative characteristics are summarized in Table 1. There was no significant difference in the mean anhepatic time or in the duration of the surgical procedure between the groups. The mean lowest core body temperature during the procedure was 0.4°C lower in the cadaveric donation transplant group (P = 0.03). The mean ischemic time in the living donation OLT group was 3.6 h shorter compared with the cadaveric donation OLT group (P < 0.001). There were no significant differences in infused volume of Plasmalyte® or albumin between the two groups.
The mean MELD score was lower in the living donation OLT group compared with the cadaveric donation OLT group (P < 0.001). In addition, the baseline hemoglobin concentration, INR, platelet count, fibrinogen concentration, TEG® MA, and TEG® α angle (TEG® α) were significantly increased (by at least 10%) in the living donation OLT group compared with the cadaveric donation OLT group (Table 1). Patients in the living donation OLT group received 66% fewer total blood component treatments on average compared with patients in the cadaveric donation OLT group (P < 0.001). Patients in the living donation group also received at least 65% fewer units of each category of component therapy (red blood cells [RBC], FFP, platelets, and cryoprecipitate) compared with the patients in the cadaveric group (P ≤ 0.003) (Fig. 3).
There was no significant difference in the amount of cell-salvage blood transfused between the 2 groups (600 ± 320 mL in the living donation OLT group versus 900 ± 670 mL in the cadaveric donation OLT group; P = 0.08). Cell salvage was used in 19 (70%) of 27 patients in the living OLT group versus 50 (72%) of 69 in the cadaveric OLT group. The mean number of component treatments among the 68 patients in whom cell salvage was used was 6.2 ± 5.3 versus 4.5 ± 5.0 among the 28 patients in whom cell salvage was not used. This difference was not significant.
Thirteen (48.1%) patients in the living donation group required no component or PRBC transfusion during the surgical procedure versus 6 (8.9%) cadaveric graft recipients (P ≤ 0.001).
The relationships between the preoperative and baseline variable and blood product usage for the study population are shown in Table 2. MELD was positively correlated to transfusion of component therapy (r = 0.55) (Fig. 4). On average, an increase of 3 points in MELD score was associated with one additional blood component treatment.
The relationships between baseline coagulation tests and MELD versus blood product usage were weaker in the living donation OLT group compared with the cadaveric donation OLT group. However, the sample size was too small to form strong statistical conclusions for this comparison.
In the multivariable analysis, the factors most strongly associated with the number of transfusions were the MELD score (P < 0.001), fibrinogen concentration (P < 0.001), and hemoglobin concentration (P = 0.02). The MELD score accounted for 19% of the variation in the number of transfusions. Fibrinogen accounted for an additional 13% and hemoglobin accounted for an additional 5%. The multiple logistic regression model defined the following formula: 6.14 + (0.109 × MELD) − (0.462 × hemoglobin concentration) − (0.00575 × fibrinogen concentration). Scores <1.8 were associated with no requirement for intraoperative transfusion whereas scores >1.8 were associated with an intraoperative transfusion requirement of at least one transfusion of component therapy. We could predict patients who did not need transfusion with 83% specificity and 68% sensitivity. In our study population in which 81% of patients required at least one transfusion of component therapy, a score >1.8 would correctly predict “transfusion requirement” 95% of the time. However, a score of <1.8 would correctly predict “no requirement for transfusion” 37% of the time.
We detected significant differences in the preoperative MELD score and the intraoperative use of blood and blood component therapy in patients undergoing living donation OLT compared with patients undergoing cadaveric donation OLT. Patients in the living donation OLT group had lower MELD scores, indicating less advanced liver disease. In addition, baseline measures of coagulation status indicated preservation of coagulation function in the living donation group compared with the cadaveric donation group. Moreover, there were significant differences in the transfusion of RBC, FFP, platelets, and cryoprecipitate between the two groups, with patients in the living donation group receiving fewer transfusions. Finally, more patients in the living donation OLT group completed the surgical procedure without need for transfusion compared with the cadaveric donation OLT group. Although patient and graft survival were not investigated in the present study, avoidance of RBC transfusion may be an important predictor of long-term outcome in OLT (17).
We also detected a significant association between MELD and total number of component transfusions in the population as a whole. Our results differ from a previous study that indicated that resource allocation (requirement for postoperative mechanical ventilation, blood component therapy, and dialysis) is not predicted by preoperative MELD score (2). These authors reported that neither MELD nor CTP score correlated with intraoperative transfusion requirements, although there was a trend towards increased transfusion requirements with increasing MELD (2).
The two study groups also differed in the relationship of preoperative MELD score and baseline hematologic and coagulation laboratory results to the transfusion requirements. In the cadaveric donation OLT group, MELD scores were associated with the total number of transfusions whereas individual coagulation and hematologic tests correlated with their respective component therapy transfusions. We were unable to detect similar relationships within the living donation OLT group. Although speculative, as we did not study changes from baseline in coagulation tests after reperfusion, reperfusion coagulopathy may have been less severe as a result of the shortened cold ischemic time and superior preservation of core body temperature in the living donation group. These patients were transplanted at lower MELD scores, indicating less advanced liver disease, compared with the cadaveric group. Again, although speculative, in the setting of less advanced liver disease, there may have been reduced bleeding during the hepatectomy as a result of less severe portal hypertension and less severe disruption of the coagulation system in this group. In addition, we cannot exclude the possibility of a Type II error; i.e., the sample size may have been too small to detect a significant relationship.
Our findings regarding the transfusion requirements in living donation OLT compared with cadaveric donation OLT conflict with those from earlier reports. Tully et al. (14) failed to show any differences in intraoperative blood component requirements in patients undergoing cadaveric versus living donation liver OLT. The authors did not note whether an algorithm was used to guide transfusion therapy (14). In addition, there was no difference in the CTP score between the living donation OLT and cadaveric donation OLT groups in this study (14), indicating that patients, regardless of donation source, were transplanted at similar stages of hepatic disease. In our investigation, there was a significant difference in the MELD score between the two groups. Similarly, Pirat et al. (9) did not detect a significant difference in intraoperative transfusion between living donation and cadaveric donation recipients. The authors of this study do not report MELD score, UNOS stage, or CTP score of the study patients. In addition, the authors appear to report only RBC transfusion, which represents only part of the transfusion requirements of the typical OLT recipient.
There are a number of recent reports that have analyzed preoperative clinical and laboratory variables in relation to intraoperative transfusion requirements (8–10,12). Massicotte et al. (8) reported that only baseline INR and platelet count along with duration of surgery were related to the number of units of RBCs transfused during primary OLT. Neither MELD nor CTP score were correlated with the volume of RBC transfusion. In another study, Steib et al. (12) examined the variables associated with increased blood loss and transfusion of RBC during OLT. The authors reported that, although CTP score was a univariate predictor of increased blood loss, only baseline hemoglobin and fibrin degradation product levels, along with previous upper abdominal surgery, were independent variables associated with increased blood loss (12). In the largest study of preoperative variables related to transfusion during OLT, Findlay and Rettke (10) reported that only age, creatinine, pseudocholinesterase, and bilirubin concentration were independent predictors of RBC transfusion.
The results from our study (cadaveric and living donation, combined) may have differed from these previous studies for a number of reasons. Of note, TEG® variables were not reported in these studies (8–10,12). We report data from a single center during a limited period of time (approximately 3 years) during which no changes were made in the perioperative care of these patients. Previous reports are either multicenter studies or report data from a period of time in which surgical and anesthetic technique may have changed (10,12,18). The previous reports also presented data derived from a mixture of surgical techniques (venovenous bypass, vena cava replacement, piggyback technique) (10–12). In the present study, all patients were transplanted using the piggyback technique for caval preservation. This technique may be associated with a reduction in RBC transfusion (11) compared with caval interruption techniques. In addition, all patients in the present study received an aprotinin infusion during the OLT procedure. Aprotinin infusion has been associated with a reduction in transfusion of RBC and FFP in primary OLT surgery (19–21). The previous studies either did not report use of aprotinin or other antifibrinolytic therapy (9–11) or used aprotinin in an unconventional bolus dosage (12). Finally, we used a comprehensive algorithm that combined routine quantitative coagulation tests and TEG® as a guide for transfusion. There are reports from the cardiac surgery and OLT literature that point-of-care testing and algorithm-driven transfusion practice results in reduced use of blood products (22–24). The use of point-of-care testing and algorithm-driven transfusion combined with the relatively short duration of surgical procedure resulted in a relatively small use of PRBC and component therapy compared with other studies.
There are also several significant limitations of our methodology that merit discussion. First, this was a retrospective review from a single center. Our results, which are merely associations, may not be easily generalized to other institutions and different paradigms of care. Second, intraoperative blood loss was not reported. It is not our practice to record intraoperative blood loss during the procedure. There may have been significant differences in intraoperative blood loss between the two groups that could have influenced transfusion requirements. Third, we examined only baseline data. Blood loss during OLT occurs during the hepatectomy phase as the result of dissection of dilated collateral vessels in the presence of portal hypertension and coagulopathy, during the anhepatic phase because of the cessation of production of coagulation factors and inhibitor, and after recirculation because of anastomotic bleeding, reperfusion coagulopathy, and fibrinolysis. Tests of coagulation function and TEG® often change dramatically from baseline during these phases of OLT. Fourth, there are a number of variables that can contribute to intraoperative blood loss that we did not investigate. The etiology of end-stage liver disease, the presence and severity of portal hypertension, previous abdominal surgery, the size of the diseased liver, and the size of the patient are potentially important contributors to blood loss during the preanhepatic phase. Although we did not detect any differences in the etiology of liver disease between the groups, there may have been differences in these other variables that may have confounded our findings. Finally, we limited our review to the surgical cases performed by one of three transplant surgeons. We did so because the other two surgeons do not perform living donation OLT. In limiting our review to this one surgeon, we may have introduced some selection bias; however, we have presented a more homogeneous population with regard to consistency in surgical technique and ability.
In conclusion, we detected a number of significant differences between patients undergoing living donation OLT compared with patients undergoing cadaveric donation OLT. The living donation OLT patients were transplanted at lower MELD scores, had more preserved baseline coagulation function, and received fewer transfusions. We detected a significant association between MELD and total number of component transfusions and a significant association between hematologic and coagulation laboratory results and component therapy transfused in the population as a whole.
1. Malinchoc M, Kamath P, Gordon F, et al. A model to predict poor survival in patients undergoing transjugular intrahepatic portosystemic shunts. Hepatology 2000;31:864–71.
2. Brown RJ, Kumar K, Russo M, et al. Model for end-stage liver disease and Child-Turcotte-Pugh score as predictors of pretransplantation disease severity, posttransplantation outcome and resource utilization in United Network for Organ Sharing status 2A patients. Liver Transpl 2002;8:278–84.
3. Freeman R, Wiesner B, Harper A, et al. The new liver allocation system: moving toward evidence-based transplantation policy. Liver Transpl 2002;8:851–8.
4. Wiesner R, Edwards E, Freeman R, et al. Model for end-stage liver disease (MELD) and allocation of donor livers. Gastroenterology 2003;124:91–6.
5. Ozier Y, Pessione F, Samain E, et al. Institutional variability in transfusion practice for liver transplantation. Anesth Analg 2003;97:671–9.
6. Motschman T, Taswell H, Breecher M, et al. Intraoperative blood loss and patient and graft survival in orthotopic liver transplantation: their relationship to clinical and laboratory data. Mayo Clinic Proc 1989;64:346–55.
7. Butler P, Israel L, Jenkins D, Strazl T. Blood transfusion in liver transplantation. Transfusion 1985;25:120–3.
8. Massicotte L, Sassine M-P, Lenis S, Roy A. Transfusion predictors in liver transplant. Anesth Analg 2004;98:1245–51.
9. Pirat A, Sargin D, Torgay A, Arsian G. Identification of preoperative predictors of intraoperative blood transfusion requirement in orthotopic liver transplantation. Transplant Proc 2002;34:2153–5.
10. Findlay J, Rettke S. Poor prediction of blood transfusion requirements in adult liver transplantations from preoperative variables. J Clin Anesth 2000;12:319–23.
11. Cacciarelli T, Keefe E, Moore D, et al. Primary liver transplantation without transfusion of red blood cells. Surgery 1996;120:698–704.
12. Steib A, Freys G, Lehmann C, et al. Intraoperative blood losses and transfusion requirements during adult liver transplantation remain difficult to predict. Can J Anaesth 2001;48:1075–9.
13. Gerlach H, Slama K, Bechstein W, et al. Retrospective statistical analysis of coagulation parameters after 250 liver transplantations. Semin Thromb Hemost 1993;19:223–32.
14. 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 2002;8:C1.
15. Figueras J, Sabate A, Fabregat T, et al. Hemodynamics during the anhepatic phase in orthotopic liver transplantation with vena cava preservation: a comparative study. Transplant Proc 1993;25:2588–9.
16. Marcos A, Fisher R, Ham J, et al. Right lobe living donor liver transplantation. Transplantation 1999;68:798–803.
17. Cacciarelli T, Keefe E, Moore D, et al. Effect of intraoperative blood transfusion on patient outcome in hepatic transplantation. Arch Surg 1999;134:25–9.
18. Desai N, Mange K, Crawford M, et al. Predicting outcome after liver transplantation: utility of the Model for End-stage Liver Disease and a newly derived discrimination function. Transplantation 2004;77:99–106.
19. Porte R, Molennar I, Begliomini B, et al. Aprotinin and transfusion requirements in orthotopic liver transplantation: a multicenter randomized double-blind study. Lancet 2000;355:1303–9.
20. Scudamore C, Randall T, Jewesson P, et al. Aprotinin reduces the need for bllod products during liver transplantation. Am J Surg 1995;169:546–9.
21. Marcel R, Stegall W, Sult C et al. Continuous small-dose aprotinin controls fibrinolysis during orthotopic liver transplantation. Anesth Analg 1996;82:1122–5.
22. Despotis G, Skubas N, Goodnough L. Optimal management of bleeding and transfusion in patients undergoing cardiac surgery. Semin Thorac Cardiovasc Surg 1999;11:84–104.
23. Gerlach H, Rossaint R, Bechstein W, et al. Goal-directed transfusion management leads to a distinct reduction of fluid requirement in liver transplantation. Semin Thromb Hemost 1993;19:282–5.
24. Goodnough L, Despotis G. Establishing practice guidelines for surgical blood management. Am J Surg 1995;170:16S–20.