Objectives: Despite data showing equivalent outcomes between grafts from marginal versus standard criteria deceased liver donors, elevated donor transaminases constitute a frequent reason to decline potential livers. We assessed the effect of donor transaminase levels and other characteristics on graft survival.
Materials and Methods: We performed a retrospective cohort analysis of adult first deceased donor liver transplant recipients with available transaminase levels registered in the Organ Procurement and Transplantation Network database (2008-2018). We used Cox proportional hazards regression to determine the effects of donor characteristics on graft survival.
Results: Of 53 913 liver transplants, 52 158 were allografts from donors with low transaminases (≤ 500 U/L; group A) and 1755 were from donors with elevated transaminases (> 500 U/L; group B). Group A recipients were more likely to be hospitalized (P = .01) or in intensive care (P < .001) or to have mechanical assistance (P < .001), portal vein thrombosis (P = .01), diabetes mellitus (P = .003), or dialysis the week before liver transplant (P = .004). Multivariable analysis (controlling for recipient characteristics) showed donor risk factors of graft failure included diabetes mellitus (P < .001), donation after cardiac death (P < .001), total bilirubin > 3.5 mg/dL (P < .001), serum creatinine > 1.5 mg/dL (P = .01), and cold ischemia time > 6 hours (P < .001). Regional organ sharing showed lower risk of graft failure (P = .02). Donor transaminases > 500 U/L were not associated with graft failure (relative risk, 1.02; 95% CI, 0.91-1.14; P = .74).
Conclusions: Donor transaminases > 500 U/L should not preclude the use of liver grafts. Instead, donor total bilirubin > 3.5 mg/dL and serum creatinine > 1.5 mg/dL appear to be associated with higher likelihood of graft failure after liver transplant.
Key words : Alanine aminotransferase, Aspartate aminotransferase, Extended criteria donors, Liver enzymes, Marginal donors
Liver transplant (LT) plays a vital role in the armamentarium for the management of end-stage liver disease, acute liver failure, metabolic liver disorders, and liver malignancies.1 However, the growing number of LTs performed is coupled with a rising number of LT candidates added to the wait list, and thus access to this rare treatment source is limited by the shortage of the liver donor pool.2 This scenario leads to approximately 10% wait list mortality, and another 10% of delisting due to illness annually (based on Organ Procurement and Transplantation Network [OPTN] data as of January 15, 2020). To bridge this discrepancy between liver allograft demand and supply, an approach toward accepting extended criteria donors has been initiated,3 and survival outcomes after use of these grafts seem to be equivalent to grafts from standard criteria donors.4 A common reason to preclude the use of liver grafts for LT is elevated donor liver enzyme levels (alanine aminotransferase [ALT] and aspartate aminotransferase [AST]) because increased transaminase levels serve as indicators of hepatocyte injury and necrosis.5-7
Previous studies with large sample sizes from a noncontemporary era suggest that transaminase levels are not accurate predictors of outcome8 and thus were not incorporated in the Donor Risk Index.9 On the other hand, current evidence on LT outcomes with the use of liver allografts from donors with elevated transaminase levels before organ procurement is limited to small, single-center studies.10,11 Donor transaminase levels are considered by transplant surgeons to determine the suitability of a liver allograft for LT; therefore, it is important to determine the effect of donor transaminase levels in LT outcomes.
We investigated the outcomes of LT after the use of liver grafts from donors with low versus elevated transaminase levels at the time of procurement, the trends in use of these grafts over time, and the possible donor risk factors of graft failure.
Materials and Methods
Data from the OPTN collected through June 30, 2018, and released on September 7, 2018, were used in the present analysis. The United Network for Organ Sharing (UNOS, a contractor for the OPTN) supplied these data. The interpretation and reporting of these data are the responsibility of the authors and in no way should be seen as an official policy of or interpretation by the OPTN or the US Government. The OPTN/UNOS database is deidentified and publicly available; therefore, this study was exempt from evaluation by the Institutional Review Board.
All adult (≥ 18 years) first LT recipients registered in the OPTN database from January 1, 2008, to June 30, 2018, were initially identified. We included only deceased donor LT recipients for whom the donor transaminase levels at the time of procurement were available. There were 33 donors for whom donor transaminase levels were not reported, and thus the recipients of these donor livers were excluded from this analysis. Patients were divided into 2 groups based on donor transaminase levels (low [≤ 500 U/L] = group A vs elevated [> 500 U/L] = group B). Kernel smoothing revealed that AST and ALT levels of approximately 500 U/L were associated with an increase in the rate of graft loss, and this was established as the cutoff point (data not shown). Exclusion criteria consisted of prior transplant of other organs, retransplant, multiorgan transplant, split/reduced LT, and pediatric (< 18 years) LT recipients. No age restriction was applied for donors.
Variables included in our analysis refer to both the donor and the recipient at the time of LT. Donor characteristics consisted of age, sex, height (in m), weight (in kg), race, cause of death (anoxia, trauma, cerebrovascular accident [CVA], other), any type of diabetes mellitus (DM), history of hypertension, type of graft (donation after brain death vs donation after circulatory death [DCD]), transaminase levels (AST and ALT, in U/L), total bilirubin levels (in mg/dL), serum creatinine levels (in mg/dL), organ sharing region (local, regional, national), and cold ischemia time (CIT, in h; that is, the length of time the donor organ is flushed with cold solution until it is removed from ice just prior to anastomosis in the recipient). Recipient characteristics included age, sex, race, height (in m), weight (in kg), cause of liver disease, body mass index (BMI, calculated as weight in kilograms divided by height in meters squared), DM of any type, dialysis within 7 days before LT, albumin levels (in g/dL), Model for End-Stage Liver Disease (MELD) score, patient location (out of hospital, in hospital, in the intensive care unit), mechanical assistance, portal vein thrombosis, previous abdominal surgery, donor-to-recipient body surface area matching ratio (graded as appropriate size, too small, or too large12), time period, and center code. The body surface area values for donors and recipients were calculated by the formula described by Mosteller.13 The main outcome of interest was all-cause graft failure.
We used frequencies and percentages to summarize the categorical variables, and median values and interquartile ranges (IQRs) to summarize the continuous variables. Between-group comparisons were performed using parametric and nonparametric methods, as appropriate. If less than 1% of the categorical values were missing, then the majority value was given. If less than 1% of the continuous values were missing, then the median value was given. We evaluated data for donor AST, ALT, total bilirubin, and serum creatine levels for potential correlations with the Pearson correlation coefficient. We used the Kaplan-Meier method to generate graft survival curves for patients with low versus elevated transaminase levels, and we used the log-rank test for comparisons. We used Cox proportional hazards regression to determine the contributions of donor variables when controlling for the recipient variables on graft survival between the 2 groups. We performed all statistical analyses with the JMP Pro statistical package (version 14.3.0; SAS Institute). Results were considered significant if P < .05.
Recipient and donor characteristics
Of the 53 913 LT recipients included in our study, 52 158 (96.7%) received liver allografts from donors with low transaminase levels (≤ 500 U/L; group A) and 1755 (3.3%) from donors with elevated transaminase levels (> 500 U/L; group B). For this analysis, AST and ALT levels were grouped together as “transaminase levels” because of the strong correlation between the 2 variables (rho, 0.71; P < .001), whereas no significant correlation was identified between either transaminase and either total bilirubin or serum creatinine levels. Detailed donor and recipient characteristics of the whole cohort are depicted in Table 1. Significant differences between the 2 groups were identified in all donor age subgroups, and donors in group B were more likely to be younger. Donors in group A were more likely to be female (40.5% vs 36.3%, respectively; P < .001), Asian (2.5% vs 1.3%, respectively; P < .001), Black (18.2% vs 13.7%, respectively; P < .001), and “other” race (1.8% vs 1.0%, respectively; P = .01), while they were less likely to be White (65.1% vs 72.1%, respectively; P < .001) compared with group B. In group A, the donor cause of death was more likely to be CVA (37.1% vs 8.4%, respectively; P < .001) or trauma (31.1% vs 24.7%, respectively; P < .001) and less likely to be anoxia (29.4% vs 65.0%, respectively; P < .001) compared with group B. Donors in group A were more commonly characteristized by any type of DM (13.3% vs 7.8%, respectively; P < .001), a history of hypertension (38.5% vs 18.8%, respectively; P < .001), and DCD (6.3% vs 3.5%, respectively; P < .001) but less commonly characterized by total bilirubin > 3.5 mg/dL (1.5% vs 3.7%, respectively; P < .001) or serum creatinine > 1.5 mg/dL (27.9% vs 57.2%, respectively; P < .001) compared with donors in group B. Local organ sharing was more commonly seen in group A (70.5% vs 63.8%, respectively; P < .001) than in group B, but national organ sharing was less common in group A compared with group B (4.2% vs 9.3%, respectively; P < .001).
There were no differences between the 2 groups in terms of recipient age, sex, and race. Acute hepatic necrosis was more commonly reported in group A than in group B (4.4% vs 3%, respectively; P = .002), whereas there was no other difference between the 2 groups in terms of the cause of liver disease. Recipients in group A were more likely to be hospitalized (18.6% vs 16.4%, respectively; P = .01) or in the intensive care unit (13.2% vs 9.7%, respectively; P < .001) or to have mechanical assistance (7.6% vs 5.1%, respectively; P < .001), portal vein thrombosis (12.3% vs 10.3%, respectively; P = .01), DM (26.7% vs 23.5%, respectively; P = .003), or dialysis within 7 days prior to LT (9.7% vs 7.6%, respectively; P = .004). Statistically significant differences between the 2 comparison arms were also detected in the subgroup with BMI 18.6 to 30 (60.9% vs 64.5%; P = .002) and all MELD subgroups, as recipients in group B were more likely to have lower MELD scores. Detailed comparisons of group A versus group B are summarized in Table 2.
Trends over time
The trends in the use of grafts from donors with low versus elevated transaminase levels over the years are summarized in Table 3. The percentage of grafts that were used from donors with elevated transaminase levels over the years are depicted in Figure 1. In 2008, 2.9% of the grafts accepted for transplant were from donors with elevated transaminase levels, versus 4% in 2018.
Cumulative 1-, 3-, 5-, and 8-year graft survival rates are shown in Table 4. Unadjusted graft survival was superior in the recipients of grafts from donors with elevated transaminase levels at the time of procurement compared with recipients of grafts from donors with low transaminase levels at the time of procurement (P = .02; Figure 2). In the multivariable Cox proportional hazard model (Table 5), after controlling for the recipient variables, AST/ALT > 500 U/L was not associated with the likelihood of graft failure (relative risk [RR], 1.02; 95% CI, 0.91-1.14; P = .74). The donor age groups of 31 to 45 years (RR, 1.12; 95% CI, 1.06-1.18; P < .001), 46 to 60 years (RR, 1.24; 95% CI, 1.18-1.31; P < .001), and > 61 years (RR, 1.45; 95% CI, 1.36-1.55; P < .001) were associated with an increased likelihood of graft failure. In terms of donor cause of death, only CVA was associated with higher likelihood of graft failure (RR, 1.13; 95% CI, 1.08-1.18; P < .001). The following donor characteristics were also associated with higher likelihood of graft failure: DM (RR, 1.13; 95% CI, 1.07-1.19; P < .001), DCD (RR, 1.56; 95% CI, 1.45-1.67; P < .001), total bilirubin > 3.5 mg/dL (RR, 1.25; 95% CI, 1.10-1.44; P < .001), serum creatinine > 1.5 mg/dL (RR, 1.05; 95% CI, 1.01-1.10; P = .01), and CIT groups > 6 to 8 hours (RR, 1.09; 95% CI, 1.05-1.14; P < .001), > 8 to 12 hours (RR, 1.19; 95% CI, 1.14-1.25; P < .001), and > 12 hours (RR, 1.41; 95% CI, 1.26-1.58; P < .001). National organ sharing was not found to be associated with graft failure, whereas regional organ sharing was associated with lower likelihood of graft failure (RR, 0.95; 95% CI, 0.91-0.99; P = .02).
In an attempt to meet the growing demand for liver grafts and to expand the liver donor pool, several transplant centers have initiated the use of extended criteria deceased donors,3 including donors with elevated transaminase levels at the time of procurement.14 In this large national cohort of 53 913 LT recipients, our multivariable analysis revealed that recipients of liver allografts from donors with transaminase levels > 500 U/L at time of procurement did not exhibit inferior graft survival compared with recipients of liver allografts from donors with transaminase levels ≤ 500 U/L. Instead, donor total bilirubin > 3.5 mg/dL and donor serum creatinine > 1.5 mg/dL were associated with a higher likelihood of graft failure. Although previous reports have shown that national organ sharing is a risk factor of graft failure,9,15 as transplant surgeons exhibit greater acceptance of extended criteria donor livers and as lengths of CIT increase, our results indicate that procurement of organs from a national-level organ sharing region is not significantly associated with graft failure. Indeed, as described by Feng and colleagues9 when the variables “organ sharing region” and “CIT” were excluded from the multivariable model, results did not change, thus emphasizing that serum creatinine > 1.5 mg/dL can serve as a better indicator of graft failure than organ sharing region.
Despite the high demand for more livers, a significant number of potential liver allografts are subject to discardment for various reasons. According to McCormack and colleagues, abnormal donor biochemistry, including transaminase levels, constitutes the reason for organ discardment in 2.5% of the cases.16 Data from the Scientific Registry of Transplant Recipients database showed that approximately 69% of the donor livers with elevated transaminase levels (≥ 1000 U/L) were discarded versus around 22% of the donor livers with low transaminase levels (< 1000 U/L).10 The increase in transaminase levels has traditionally been an indicator of ischemic liver injury and consequently is considered to be a factor for evaluation of liver allografts10; however, precise quantitative correlation of hepatic necrosis and necroinflammatory activity with biopsy findings is not always possible.17 Hence, the use of donor transaminase levels to determine the likelihood of graft failure remains a questionable tool. The highest concentrations of AST are found in liver, skeletal muscles, heart, kidney, red blood cells, and brain, whereas ALT is primarily found in liver and kidney and in lower concentrations in the skeletal muscle and heart; additionally, it is noteworthy that elevated transaminase levels can also be increased secondary to nonhepatic causes (ie, myocardial infarction, muscle injury, hemolysis).10 Although such conditions are not uncommon in deceased donors (ie, donors who die secondary to trauma), trauma was most commonly reported as a cause of death in donors with low versus elevated transaminase levels in our patient cohort. Feng and colleagues9 described the concept of a Donor Risk Index in 2006 in an attempt to more precisely quantify the risk of graft failure and to define parameters toward the acceptance or discardment of liver grafts. The investigators did not identify AST levels as a donor parameter known to be significantly associated with liver graft failure. Similar efforts were made again in the US (Survival Outcomes Following Liver Transplantation [SOFT] score)15 and in Europe (Eurotransplant Donor Risk Index)18 to describe risk indexes, and transaminase levels were again not determined to be a significant parameter on multivariate analysis. Cuende and colleagues8 showed that liver enzyme levels above 200 U/L are not independent risk factors for graft failure.
Fakhar and colleagues compared recipients of high (> 500 U/L; n = 24) versus low transaminase liver donors (≤ 500 U/L; n = 834) and reported no difference between the 2 groups in terms of posttransplant complications and survival outcomes.11 Radunz and colleagues19 demonstrated acceptable outcomes in a small cohort of 15 patients who received donor livers with increased transaminase levels. Mangus and colleagues20 evaluated transaminase levels in 1348 donor livers (normal/mildly increased [0-499 U/L], moderately increased [500-999 U/L], severely increased [≥ 1000 U/L]), and they did not identify a significant difference among the groups when controlled for donor age and recipient MELD score. In another recently published single-center study, Martins and colleagues showed similar graft survival rates between recipients of livers with high (> 1000 U/L; n = 14) versus low (< 1000 U/L; n = 224) peak transaminase levels.10
Compared with previously published studies, we studied a large sample of national data that ranged over a contemporary period of time (2008-2018) to further validate the findings of older database analyses and recent single-center studies. We suggest that the use of grafts from donors with elevated transaminase levels has only increased by 1.1% from 2008 to 2018. We also found that the cumulative graft survival rates in the recipients of grafts from donors with elevated transaminase levels were superior compared with recipients of grafts from donors with low transaminase levels; however, this could be attributed to the possibility of overall greater severity of illness in the first group. Indeed, multivariable analysis showed no association between transaminase levels and graft survival. Therefore, our findings indicate that discardment of marginal livers on the sole basis of transaminase levels is not justified, because elevated transaminase levels were not associated with increased graft failure. Instead, based on our results, donor serum creatinine > 1.5 mg/dL was found to be independently associated with higher likelihood of graft failure. The SOFT score also determined donor serum creatinine > 1.5 mg/dL to be an independent risk factor of graft failure,15 whereas the Donor Risk Index score9 and the Eurotransplant Donor Risk Index18 score both failed to determine the importance of donor serum creatinine levels. We also found donor total bilirubin > 3.5 mg/dL to be independently associated with a higher likelihood of graft failure, whereas all 3 of the above-mentioned indexes do not include donor total bilirubin levels.
Moreover, in our study, we showed by multivariable analysis that recipients of livers from donors older than 31 years of age, donors with DM, donors who died from CVA, DCD, and donors with a CIT > 6 hours were more likely to experience graft failure. These results are in accordance with previously published studies that identified increased donor age,8,9,15,18,21 death from CVA,9,15,21 DCD,9,18,21 and CIT > 6 hours15,21 as independent risk factors of graft failure. Notably, we determined donor DM to be independently associated with a higher risk of graft failure, which was suspected in previous small studies22,23 but was not included in previously published risk indexes.9,15,18 Diabetes mellitus can serve as a surrogate marker of donor liver graft steatosis and vascular disease (ie, hyaline arteriolar sclerosis) and therefore should be considered when deciding on the use of liver allografts.21
The present study has several strengths: (1) large sample size (n = 53 913), representative of the US population of LT recipients, (2) contemporary era (2008-2018), (3) rigorous and systematic process during the collection and analysis of the data, and (4) graphical representation of the trends to use grafts from donors with elevated transaminases. Nevertheless, the OPTN/UNOS data pose certain limitations to our study, as follows: (1) retrospective nature; (2) additional recipient or donor variables not captured in the OPTN database (ie, pumping of the donor graft, timing of procurement, etc.) may affect graft survival but could not be evaluated; (3) only transaminase levels at the time of procurement were available in the database, whereas peak values or value trends were not available; and (4) the results of kernel smoothing, as well as the fact that the number of patients with transaminase levels above 1000 U/L (severely elevated transaminase levels) represented less than 1% of the total cohort, compelled us to set the cutoff point for our study at 500 U/L (moderately/severely elevated transaminase levels).
Our findings emphasize that use of liver allografts from donors with moderately/severely elevated transaminase levels (> 500 U/L) at the time of procurement should be encouraged because donor transaminase levels were not determined to be independently associated with the likelihood of graft failure. On the other hand, donor serum creatinine > 1.5 mg/dL and donor total bilirubin > 3.5 mg/dL were associated with a higher risk of graft failure and thus should be incorporated in future donor risk indexes. Our results suggest that other donor parameters to be considered significant risk factors for post-LT graft failure include age > 31 years, CVA as the cause of death, donor DM, DCD, and CIT > 6 hours. Presently, there is no perfect model to precisely guide the decision process on liver allograft use versus discardment, but our finding suggest that elevated transaminase levels should not preclude transplant surgeons from using an extended criteria donor liver graft in appropriately selected LT candidates.
DOI : 10.6002/ect.2020.0023
From the 1Department of Surgery, Division of Hepatobiliary Surgery and Liver
Transplantation, Vanderbilt University Medical Center, Nashville, Tennessee,
USA; the 2Department of Surgery, Division of Organ Transplantation, University
of Massachusetts Memorial Medical Center, University of Massachusetts,
Worcester, Massachusetts, USA; the 3Department of Surgery, Division of
Transplant Surgery, University of Washington, Seattle, Washington, USA; and the
4Clinical and Bio-Analytics Transplant Laboratory (CBATL), University of
Washington, Seattle, Washington, USA
Acknowledgements: The authors have not received any funding or grants in support of the presented research or for the preparation of this work and have no declarations of potential interest. Parts of the findings have been presented at the Virtual American Transplant Congress, May 30 to June 3, 2020.
Corresponding author: Martin I. Montenovo, Department of Surgery, Division of Hepatobiliary Surgery and Liver Transplantation, Vanderbilt University Medical Center, 1313 21st Avenue South, Nashville, TN 37232, USA
Phone: +615 875 8995
Table 1. Donor and Recipient Characteristics for the Whole Cohort (n = 53 913)
Table 2. Comparison of Donor and Recipient Characteristics Between Patients With Low Versus Elevated Transaminase Levels
Table 3. Trends in the Use of Grafts From Donors With Low Versus Elevated Transaminase Levels Over Time
Table 4. Point Estimates and Standard Errors of Unadjusted Cumulative Graft Survival
Table 5. Cox Proportional Hazard Regression for Graft Failure
Figure 1. Graphical Representation of the Percentages of Grafts From Donors With Elevated Transaminase Levels Between 2008 and 2018
Figure 2. Kaplan-Meier Curves of Unadjusted Graft Survival in Recipients of Grafts From Donors With Low Versus Elevated Transaminase Levels at the Time of Procurement