The current organ allocation policy for orthotopic liver transplantation (OLT) in the United States is based on disease severity determined by the model for end-stage liver disease (MELD). The MELD score, calculated from 3 laboratory values (total serum bilirubin, serum creatinine, and international normalized ratio of prothrombin time) was originally developed to predict the short-term mortality of patients undergoing transjugular intrahepatic portosystemic shunt procedures for portal hypertension.1 Subsequently the MELD score has been validated as a disease severity index in a variety of liver diseases, both compensated and uncompensated due to a variety of causes.2–5 Further studies have confirmed that patients with high MELD scores have a high risk of dying on the waiting list without OLT.2, 6 In the current MELD system, patients with the highest scores are considered to be most critically ill from advanced liver disease and have the most urgent need for transplantation; therefore they are given the highest priority to receive OLT.
Since the implementation of the MELD system by the United Network for Organ Sharing in February 2002, several significant developments have been observed. The average MELD score of patients undergoing OLT has increased as a direct result of the implementation of the MELD system.7 Patient mortality on the wait list and the overall number of patients waiting for OLT has decreased in the post-MELD era.8 Although the MELD system has been very successful in the above area, a concern that patients with high MELD scores may suffer suboptimal posttransplant outcome remains.7, 9–15 It is unclear what impact the organ allocation policy changes have on the perioperative management. Since patients with high MELD scores are more acutely ill, they may require a higher level of pretransplant care. Furthermore, patients with higher MELD scores may have a more complicated intraoperative course compared to patients with low MELD scores. The ability to use the MELD score as an index for pretransplant characteristics or intraoperative course could be attractive, since such an index would be used in resource anticipation, preparation, and utilization.
The current study was undertaken (1) to identify the pretransplant characteristics of liver transplant recipients that were related to MELD scores and pertinent to intraoperative management, and (2) to determine if patients with high MELD scores have a more complicated intraoperative course during OLT. Our hypothesis was that patients with high MELD scores had more pretransplant comorbidities and a more complicated intraoperative course during OLT.
From January 1, 2004, to December 31, 2004, 183 patients underwent liver transplantation at the University of California Los Angeles Medical Center. After the approval of the institutional review board, medical records of all recipients were reviewed. A total of 59 patients (31 pediatric patients <18 years old, 15 retransplant patients, 3 combined organ transplant patients, and 10 United Network for Organ Sharing status 1 patients) were excluded and the remaining 124 patients were evaluated in the study. The MELD scores immediately prior to transplantation were obtained from our Liver Transplantation Database and were the same scores that were submitted to the United Network for Organ Sharing. The minimum MELD score was 6, and the maximum was 40; scores exceeding 40 by calculation from the MELD formula were capped to 40. Patients were then divided into 2 groups according to their MELD scores: a low MELD group with scores of 30 or lower and a high MELD group with scores greater than 30.
All clinical data were collected prospectively as a part of the University of California Los Angeles transplantation database and the liver transplant anesthesia database except for intraoperative serum potassium (K+) values, which were retrieved retrospectively from the computerized laboratory database of the University of California Los Angeles Medical Center. The pretransplant data were patient demographics, preoperative characteristics and baseline laboratory values. Demographics and pretransplant characteristics were age, gender, the etiology of liver disease (modified from United Network for Organ Sharing diagnosis code), history of chronic systemic hypertension, coronary artery disease, diabetes mellitus, presence of encephalopathy or ascites (>1 liter) immediately before OLT, history of variceal bleeding, pulmonary hypertension, requirements for pretransplant hemodialysis, requirements for pretransplant endotracheal intubation/mechanical ventilation, or use of vasopressors before transplantation. Baseline laboratory data collected were hematocrit (Hct), international normalized ratio of prothrombin time, blood urea nitrogen (BUN), serum creatinine, K+, base excess, platelets, and fibrinogen levels.
Intraoperative course or management during OLT were also captured by collecting the following data: (1) intraoperative potassium (K+) (evaluated hourly and immediately following reperfusion), administration of intravenous insulin (continuous infusion or bolus for hyperkalemia prophylaxis or treatment), and duration of operation; (2) transfusion requirements for each blood component: packed red blood cells, fresh frozen plasma, cryoprecipitate, and platelets recorded for the duration of operation and for the prereperfusion and postreperfusion periods; and (3) requirement for intraoperative vasopressors. The use of intraoperative vasopressors was defined as either the administration in continuous infusion of phenylephrine, epinephrine, norepinephrine, or dopamine or in boluses of phenylephrine (total intraoperative dose >2 mg), epinephrine (>50 μg), or norepinephrine (>40 μg) during OLT. Administration of antifibrinolytic agents (epsilon-aminocaproic acid, tranexamic acid, or aprotinin) for either prophylaxis or rescue treatment was recorded.
Statistical analyses were performed using the Statistical Package for the Social Sciences 12.0 for Windows (SPSS, Inc., Chicago, IL). Continuous variables were presented as means ± SD. Means of the two groups were compared by Student's t test following conversion of raw data into a logarithmic scale when appropriate. Trends in means across ordered categories were assessed using ANOVA methods with a post hoc weighted linear and higher order trend test. Categorical variables were presented as proportions and analyzed by Pearson chi-square test or chi-square test for trend when there were ordered categories. The nonparametric Spearman correlation was also computed to assess trend for both means and proportions. A 2-tailed P value <0.05 was considered statistically significant.
The mean patient age was 52.4 years (range, 19 to 74 years). The majority of patients were male (67.7%). The 3 main indications for liver transplant were hepatitis C (51.6%), hepatitis B (10.5%), and Laennec's cirrhosis (8.9%). The mean pretransplant MELD score of all patients was 27.86 ± 9.07 (SD). The mean MELD score was 21.78 ± 6.56 for the low MELD group (n = 73) and 36.57 ± 3.02 (n = 51) for the high MELD group. The distribution of the MELD scores is shown in Figure 1.
The patient demographics, preoperative clinical characteristics and baseline laboratory data in the low and high MELD groups are compared in Table 1. As expected, the baseline international normalized ratio and creatinine, 2 variables used to calculate MELD scores, differed significantly between the 2 groups. There was no difference between the 2 groups with respect to age, gender, and history of hypertension, coronary artery disease, or diabetes mellitus. Baseline K+, platelets, and base excess were similar between the 2 groups. A higher proportion of patients (63.0%) in the low MELD group had hepatitis C as an indication for liver transplantation than those in the high MELD group (35.3%). All patients with autoimmune disease (n = 6) and none with primary biliary cirrhosis (n = 4) were in the high MELD group. Presence of hepatic encephalopathy and history of variceal bleeding were similar between the 2 groups, but presence of ascites at operation was more common in the high MELD group. Not surprisingly, all 18 patients (35.3%) who required pretransplant hemodialysis were in the high MELD group.
Table 1. Demographics, Pretransplant Clinical Characteristics, and Baseline Laboratory Data
Low MELD group (n = 73)
High MELD group (n = 51)
NOTE: Data are presented as mean ± SD or n (%).
Abbreviations: INR, international normalized ratio of prothrombin time; NS, not significant.
Data have been converted into a logarithmic scale before statistical analyses.
Several baseline laboratory data not directly related to the MELD were also significantly different between the 2 groups. The patients in the high MELD group had significantly lower baseline Hct and fibrinogen levels compared to patients in the low MELD group (P = 0.010 and 0.016, respectively). Two other general indices of disease severity—pretransplant endotracheal intubation/mechanical ventilation and use of vasopressors for the management of hypotension—were also more common in the high MELD group. A total of 12 patients in the high MELD group compared with only 1 in the low MELD group required pretransplant intubation (P < 0.001). Seven patients in the high MELD group required pretransplant vasopressors compared to none in the low MELD group (P < 0.001).
The mean intraoperative K+ values in the 2 groups showed a similar pattern: slightly increasing over time before reperfusion, reaching the highest levels immediately after reperfusion, and then gradually returning to baseline. The mean intraoperative K+ values in the different time periods and the frequency of hypokalemia or hyperkalemia during the different time periods were not different between the 2 groups. Administration of insulin and duration of operation between the 2 groups were similar (35.3% vs. 28.8% and 5:12 ± 0.55 vs. 5:15 ± 1.14, in hr:min, respectively).
The intraoperative transfusion data revealed a significantly increased transfusion requirement for patients in the high MELD group (Table 2). Patients in the high MELD group required an average of 5 more units of red blood cells during OLT than patients in the low MELD group (15.86 ± 9.56 vs. 10.62 ± 8.13): 3 more units in the prereperfusion period (9.85 ± 6.21 vs. 6.85 ± 5.94), and 2 more units in the postreperfusion period (5.83 ± 6.19 vs. 3.55 ± 3.55). Requirement for fresh frozen plasma during OLT showed a similar pattern: Patients in the high MELD group required 7 more units of fresh frozen plasma than the patients in the low MELD group (22.10 ± 14.39 vs. 15.12 ± 9.74); 4 more units in prereperfusion period (14.10 ± 8.34 vs. 9.93 ± 7.28), and 3 more units in the postreperfusion period (7.85 ± 8.28 vs. 4.89 ± 4.24). For patients in the high MELD group, total transfusion requirements for cryoprecipitates or platelets were nearly doubled (Table 2). In addition, antifibrinolytic agents were more frequently needed in high MELD patients than in low MELD patients (39/51 or 76.5% vs. 40/73 or 54.8%, P = 0.014).
Table 2. Requirements of Transfusion in Different Periods During Liver Transplant Surgery (Units)
Low MELD group (n = 73)
High MELD group (n = 51)
NOTE: Data are presented as means ± SD or n (%).
Abbreviations: RBC, packed red blood cells; FFP, fresh frozen plasma; NS, not significant.
6.85 ± 5.94
9.85 ± 6.21
3.55 ± 3.55
5.83 ± 6.19
10.62 ± 8.13
15.86 ± 9.56
9.93 ± 7.28
14.10 ± 8.34
4.89 ± 4.24
7.85 ± 8.28
15.12 ± 9.74
22.10 ± 14.39
0.23 ± 0.49
0.52 ± 0.55
0.61 ± 0.65
0.76 ± 0.64
0.86 ± 0.90
1.33 ± 0.89
0.17 ± 0.41
0.41 ± 0.54
0.35 ± 0.51
0.59 ± 0.72
0.52 ± 0.77
1.06 ± 0.95
Antifibrinolytics (n =)
Patients in the high MELD group required not only more pretransplant but also more intraoperative hemodynamic support with vasopressors (Table 3). The requirement for continuous infusion of vasopressors was significantly higher in patients in the high MELD group than patients in the low MELD group (27/51 or 52.9% vs. 14/73 or 19.2%). The requirement for vasopressors boluses was also significantly increased in the high MELD group compared to the low MELD group (21/51 or 41.2% vs. 12/73 or 16.4%). Analyses performed after exclusion of the 7 patients who required pretransplant vasopressors showed similar results.
Table 3. Requirements for Intraoperative Vasopressors
Low MELD group
High MELD group
NOTE: Data are presented as n (%). Intraoperative vasopressors requirements before and after exclusion of 7 patients who required pretransplant vasopressors.
Infusion or bolus
Infusion or bolus
To study a trend between MELD scores and the requirements for blood transfusion and vasopressors, patients were divided into 5 groups by MELD scores: MELD score of 6 to 20 (n = 25), 21 to 25 (n = 24), 26 to 30 (n = 24), 31 to 35 (n = 23), and 36 to 40 (n = 28). Higher MELD scores were associated with greater intraoperative transfusion requirements (Table 4). Similarly, there was a positive association between intraoperative vasopressors use and MELD scores (Table 5). Spearman correlation coefficients were 0.26 to 0.29 with P values 0.001 to 0.004 for these trends in transfusion requirements and 0.29 to 0.36 with P values <0.001 to 0.002 for the trends in the use of vasopressors. The incremental increase in both transfusion and vasopressor requirements was most dramatic between groups 3 and 4.
Table 4. Requirements of Total Transfusion in 5 Groups (Units)
The present study confirmed that patients with higher MELD scores had more comorbidities prior to OLT. This was suggested by both MELD-related and MELD-unrelated pretransplant characteristics and baseline laboratory data. Some of these pretransplant characteristics have important implications on perioperative management. High MELD patients have lower baseline Hct and fibrinogen levels, and they are more likely to require vasopressor and mechanical ventilation before OLT. On the other hand, MELD score had little impact on other systemic diseases, such as systemic hypertension, coronary artery disease, and diabetes mellitus. These observations and further identification of pretransplant characteristics may allow us to use MELD score in pretransplant decision making, in perioperative resource utilization and in refining future efforts designed to improve patient outcomes.
Blood transfusion during OLT is considered not only as an important component of resource utilization, but also as a risk factor for a long-term morbidity and mortality.16 Previous studies have identified a long list of factors that may have an influence on intraoperative transfusion during OLT.17, 18 Recent studies also suggested that the severity indices of liver diseases, such as MELD score, were associated with increased requirements for transfusion. Massicotte et al.17 found that MELD score was correlated with blood transfusion in a univariate analysis; however, such relationship could not be confirmed in a multivariant model. Using both living-related and cadaveric liver transplantation data, Frasco et al.19 showed a positive association between MELD score and transfusion requirement during OLT. In the present study we also confirmed that patients with high MELD scores had significantly increased transfusion requirements during OLT. However, our study differed from Frasco's in several areas. First, the mean MELD score was 27.9 in our study, substantially higher when compared to 13.4 for living related and 22.8 for cadaveric OLT in Frasco's study. This is probably, in part, because contrary to their study in which pre-MELD data were included, more recent, post-MELD era data were used in our study. Another possibility is that disease severity differs among different transplant centers. Our hospital is a high-volume liver transplant center; such centers are known to have a patient population with higher MELD scores than do low-volume centers.20 Second, the difference in transfusion requirement between the low vs. high MELD groups in our study is higher. There is a 12-unit (5 red blood cell and 7 fresh frozen plasma) difference between our low and high MELD groups vs. a 5-unit difference if using a rule deriving from their study. Although some authors have reported a dramatic decrease in transfusion requirements with an average of 2 to 3 units of red blood cells during OLT,17 blood transfusion requirement in our study remains high during OLT, especially in patients with higher MELD scores. It is possible that the presence of patients with severe liver disease in this medical center and higher MELD scores in our patient population contribute to higher blood transfusion requirements.
Administration of vasopressors is frequently required during OLT and has been associated with negative impact on hepatic blood perfusion.21 To our knowledge, this is the first study to demonstrate that patients with high MELD scores required more vasopressors before and during transplant surgery. Higher vasopressor requirements before and during OLT clearly suggest that these patients have more advanced diseases. The greater amount of blood loss during liver transplant surgery, reflected by increased requirements of blood transfusion, may also exacerbate hypotension, which may further lead to increased intraoperative vasopressor requirements. Understanding the relationship between MELD scores and pretransplant characteristics and the intraoperative course of patients with high MELD may provide insight to develop strategies to reduce the morbidity and mortality of patients undergoing OLT. Further risk stratification and modification in the perioperative management may have a positive impact on a long-term posttransplant outcome.
There are several limitations to this study. Although data were prospectively collected, a retrospective study design using single-center data has well-known limitations. The results of the study may not apply to other centers, since the patient populations and MELD score distribution vary from center to center. Furthermore, since the management of intraoperative variables was not standardized, individual practitioner preferences could have an influence on the results of the study. Nonetheless, our center is an experienced, high-volume transplant center, and these data may be of use to other centers. Lastly, the donor characteristics and their effect on the quality of the graft, which may influence transfusion and vasopressor requirements in the postreperfusion period, were not addressed in the study.
In conclusion, patients with MELD scores greater than 30 have lower pretransplant Hct and fibrinogen levels and increased requirements for pretransplant mechanical ventilation and vasopressors compared to patients with lower MELD scores. Patients with high MELD scores also have more complicated intraoperative courses, with increased requirements for blood component transfusion and vasopressors during liver transplant surgery.
The authors thank Dr. Jeff Gornbein of the University of California Los Angeles Department of Biomathematics for statistical support and helpful discussion, and Dr. Marie Csete of Emory University Department of Anesthesiology in Atlanta, GA, for critical review of the manuscript.