Liver transplantation (LT) is a definitive therapy for end-stage liver disease with excellent long-term outcomes. Approximately 6500 deceased donor LTs are performed annually in the United States, with over 17 000 patients on the transplant waitlist.1,2 Various approaches, such as deceased donor and split liver transplantation, have been used to expand the existing donor pool, but technical complexities and ethical concerns limit their widespread use to optimize the existing donor pool.3-5 An alternate approach to expand the available donor pool is to use grafts with extended criteria donors, such as donors with prolonged cold ischemia time (>12 hours), higher donor age (>60 years), hypernatremia (Na >155), positive hepatitis serology, elevated transaminases, elevated bilirubin, donor steatosis (>30%) and non–heart-beating donors (NHBD) or donors after cardiac death.6-12
Of the 65% of the US adults who report drinking alcohol at some point in their lifetime, about 8% to 10% admit to heavy drinking (HD), defined with more than 2 drinks per day.13 There are formal guidelines regarding the use of alcohol in living donors.14 The Organ Procurement Transplant Network/United Network for Organ Sharing (UNOS) AD Hoc Living Donor Committee advises all potential living donors to be evaluated for “active substance abuse,” with requirement of potential living liver donors to be “free from active alcoholism or frequent heavy alcohol use.”15 In contrast, there is a paucity of guidelines on alcohol use history among deceased liver donors.16 We performed this study to test our hypothesis that history of HD among deceased donors affects graft and patient survival among LT recipients.
Study Design and Population
Using the UNOS registry, we examined adult (>18 years of age) recipients of LT using deceased donors between January 2002 and December 2014 (Figure 1). This database has clinical information on all the recipients at the time of transplant listing, donor information, and follow-up information on the recipient. Liver transplant recipients for common indications, such as hepatitis C virus (HCV), nonalcoholic steatohepatitis, alcoholic liver disease, alcoholic liver disease with HCV, cryptogenic cirrhosis, chronic cholestatic liver disease, primary biliary cholangitis, primary sclerosing cholangitis, metabolic liver disease, hepatitis B, hepatocellular carcinoma, and alcoholic hepatitis were included. Exclusion criteria were: recipients younger than 18 years, grafts from live donors, and missing data on donors with alcohol use history.
We extracted recipient and donor data on our study population for demographic characteristics (eg, body mass index (BMI), age, sex, etc.), liver disease etiology, alcohol use history, model for end-stage disease (MELD) score, alanine aminotransferase (ALT), aspartate aminotransferase (AST), serum bilirubin, graft and patient survival and complications. Donor risk index (DRI) was obtained using the donor variables (age, race, height, cause of donor death, NHBD, cold ischemia time, split/partial graft or not, local/regional/national source of donor) as well described in literature.8 The institutional review board at the University of Alabama at Birmingham approved the study.
The study population of LT recipients was stratified based on the drinking history of the deceased donor into HD donors (>2 drinks per day) and non-HD (ND) donors with reported alcohol use of 2 or less drinks per day. Donors with available information on alcohol use were included in the final analysis. Deceased donor file was merged with the recipient file using the donor ID after excluding donors with missing or unknown information on alcohol use. Baseline recipient and donor characteristics were compared among donors with and without history of heavy alcohol use.
Propensity Score Modeling for Graft From HD Donor
We performed propensity score analysis to minimize selection bias in the study. The propensity score for being a HD graft recipient was generated using a logistic regression model. Recipient characteristics including age, MELD score, BMI, creatinine, transplant year, sex, race, etiology of liver disease, diabetes mellitus, UNOS transplant region, ABO blood type, need for dialysis, and whether on ventilator or not were included in the model. For each HD donor used, we then performed greedy matching to select 3 LT recipients using ND donors. For the greedy 1:3 matching, a HD donor graft recipient is first randomly selected. Then 3 ND donor graft recipients, whose propensity score is closest to that of this randomly selected HD donor graft recipient, are chosen for matching to the HD donor graft recipient. This process is repeated for each HD donor graft recipient.17
In the matched cohort, primary transplant outcomes were 5-year graft and patient survival. Time to patient survival was defined as the time from transplant date to date of death. Graft failure in the UNOS database is defined in case of (a) liver transplant recipient death, (b) need for retransplant, and (c) transplant-related complications, making the graft nonfunctional in the opinion of transplant hepatologist taking care of the patient. Time to graft survival was defined as the time to graft failure from the date of transplantation.
Baseline donor and recipient characteristics were compared using χ2 and t tests for categorical and continuous variables, respectively. Cox proportional hazard regression analysis models were built to examine independent effect of HD in donors on graft survival and patient survival after LT. Both models were adjusted for recipient liver disease etiology, recipient age, sex, race, history of diabetes, DRI, MELD, and history of alcohol use in donor. Cox models were repeated with components of DRI instead of DRI in the model. Interaction of alcohol use status in the donor and DRI or components of DRI was also examined. If there were any significant interactions, then subgroup analyses were performed. Kaplan Meier curve comparing ND and HD donors were generated for 5-year graft and patient survival, with Log rank test used for statistical comparison. P-values <0.05 were considered significant. SAS version 9.4 (SAS Institute, Cary, NC) was used for statistical analyses.
A total of 273 052 adult liver transplants were performed in the United States until December 2014, of which 81 478 have been performed between 2002 and 2014. Of these, 71 193 were performed in adults using deceased donors, and of which 56 182 were performed for the common indications as outlined in the methods section (Figure 1). Of these, 47 882 (85%) with available information on alcohol use were included in the analysis to test the hypothesis on impact of HD on graft and patient survival among LT recipients. Comparison of 8300 donors without available alcohol use history with 47 882 donors with available alcohol history is described in Table S1, SDC (http://links.lww.com/TP/B563).
Disposition of Donor Liver Graft
Of 197 274 deceased donors, information on liver organ disposition was available in 106 850 donors (18 382 with HD), with missing information in 88 742 and unknown information on disposition of organ in 1682 donors. Of donors with available information on alcohol use, proportion of graft used for LT was lower among HD donors (63% vs 78%), which was mainly due to either organ not recovered (24% vs 11%), or recovered but not transplanted (13% vs 11%), P < 0.001 (Figures 2A and B).
Data on regional distribution on liver organ disposition were available for 180 503 deceased donors, and showed heterogeneity in the use of organ for LT. For example, proportion of liver grafts used for LT was lowest in region 1, and highest in regions 3, 8, and 11 (66% vs 76%, P < 0.001). This was mainly due to either organ not recovered (23% vs 14-18%) or recovered but not transplanted (11% vs 7-10%), P < 0.001 for all (Figure S1, SDC, http://links.lww.com/TP/B563). Data on liver biopsy on deceased donors were available for 104 631 donors (17 515 with HD), with higher frequency in HD compared to ND donors (46% vs 33%, P < 0.001). Further, recovered liver grafts that underwent liver biopsy evaluation were more often used for LT (N = 66 716) compared with 36 800 donor livers that did not undergo liver biopsy evaluation (80% vs 74%, P < 0.0001). Grafts undergoing liver biopsy assessment compared to those grafts not assessed with liver biopsy were more likely to be not recovered (12% vs 8%) or not used for transplantation after recovery (18% vs 8%), P < 0.0001 for both analyses.
Frequency of HD Among Deceased Donors
Of 47 882 LT recipients with available donor information on alcohol use, 7298 (15%) received liver grafts from HD donors. The proportion of HD donors ranged from 11% to 16% during 2002 to 2014, with no significant change in trend over time, Armitage trend test P = 0.75 (Figure S2A, SDC,http://links.lww.com/TP/B563). Similarly, proportion of HD donors was not different for recipient liver disease etiology, χ2 test P = 0.42 (Figure S2B, SDC, http://links.lww.com/TP/B563).
Baseline Characteristics: HD Versus Nondrinking Donors
ND donors (N = 40 584) compared with HD donors (N = 7298) differed for mean age (41 ± 17 years vs 45 ± 13 years), BMI (27.5 ± 6.3 vs 26 ± 5), male gender (57% vs 74%), white race (67% vs 73%), mean bilirubin (0.93 ± 1.0 vs 0.86 ± 0.84), and donor risk index (DRI) (1.6 ± 0.4 vs 1.58 ± 0.34); P < 0.0001 for all (Table S2, SDC,http://links.lww.com/TP/B563). HD donors compared with ND donors were more often used for HCV and hepatocellular carcinoma as liver disease etiology, and less often for nonalcoholic steatohepatitis and for cryptogenic cirrhosis, P = 0.008 (Figure S3A, SDC,http://links.lww.com/TP/B563). Similarly, HD donors were more often used in UNOS regions 1 and 6, and less often in regions 9 and 11, P < 0.0001 (Figure S3B, SDC,http://links.lww.com/TP/B563). Comparison of LT from HD versus ND donors based on recipient characteristics including gender, race, MELD score, BMI, and diabetes is described in Table S2, SDC,http://links.lww.com/TP/B563.
Baseline recipient characteristics comparing 7166 HD donors and propensity score matched 21 498 ND donors are similar as expected (Table 1). However, HD donors compared with ND donors were more likely to be older men, with lower BMI, DRI, and serum bilirubin (Table 1).
Outcomes: HD Versus Nondrinking Donors
A total of 7541 grafts from the propensity score matched cohort were lost over 5 years, with similar 5-year survival comparing ND and HD donors (73.7% vs 73.7%, Log Rank P = 0.98) (Figure 3A). Similarly, 6468 patients died over 5 years, with similar 5-year survival comparing ND and HD donors (77.6% vs 77%, P = 0.36) (Figure 3B).
On a Cox regression model, history of HD in deceased donors did not affect the liver graft (1.02 [0.97-1.08]) or patient survival (1.03 [0.97-1.09]) at 5 years after LT after controlling for other donor characteristics including gender, BMI, NHBD, DRI, bilirubin, AST and ALT (Table 2). Cox regression analysis examining the interaction between history of HD in deceased donors and individual DRI components yielded similar hazard ratios as the abovementioned regression model. Because there were no significant interaction between drinking status of the donor (HD or ND) and DRI or with its various components, subgroup analyses were not performed based on DRI or based on components of DRI.
Causes of Patient Mortality
A total of 7302 patients died with 1099 recipients of liver graft from HD donors. Main causes of patient mortality were graft failure, multiorgan failure, cardiovascular, infection, with similar proportions comparing recipients of liver graft from HD versus ND donors (Table 3). About 5% of patients, the cause of death remained unknown in both the groups.
The main finding of our study is that there is no impact of HD in deceased donors on the post-LT graft and patient survival outcomes. We also found a 15% prevalence of HD among deceased donors without any variation over study duration or for liver disease indication for LT. Furthermore, we identified a higher discard rate of liver grafts among potential donors with history of HD, and this rate varied significantly across the UNOS regions.
The HD prevalence rate of 15% among deceased donors in this UNOS registry-based analysis is similar to the 12% rate observed in a retrospective analysis of 122 deceased liver donors from a single center.18 In contrast, another retrospective analysis reported a rate of alcoholism of 45%. However, the study examined deceased donors which were discarded and not all the donors. Further, in this study, the alcohol consumption was not quantified.19 About 47% of discarded livers in this study were due to refusal by the family to use the organ for LT.18 Another potential reason for organ discard is the transplant surgeon's assessment based on the gross and histological assessment as deemed necessary by the organ recovery team. This is supported by the higher liver biopsy rates among HD donors compared to ND donors. It should be noted that about 80% to 90% of recovered organs in the current analysis were used for transplantation. Based on reperfusion liver biopsy, Mangus et al20 have reported prevalence of only 20% for high risk steatosis and 8% for fibrosis in these donor livers with HD, which is likely reflective of the selection practice in most centers. The lack of guidelines and protocol to make decisions on use of livers among HD donors is also reflected by the significant regional variations in the organ discard rate.
Several donor- and recipient-related factors have been known to impact prognosis after liver transplantation. Donor factors associated with diminished graft and patient survival include prolonged cold ischemic time (>12 hours),3,8 donor age (usually more than 60 years),3,7,8,21 female sex,3,21 high serum sodium (Na > 155),10 donor steatosis (more than 60%),22,23 elevated transaminases, elevated bilirubin,21 and NHBD. It is also known that LT recipients with a MELD score of 20 or higher tend to have worse survival outcomes when receiving extended criteria donor grafts, particularly if the grafts have more than 30% steatosis.24
In our study, the 5-year liver graft and patient survival outcomes were similar among HD and ND donors, as reported earlier in the literature.3,20 In the background of these data and the increasing demand for deceased donor livers, there is a need for liberalization of the graft acceptance criteria to increase utilization of the available liver grafts among donors with history of HD.3,11,25
Large sample size using the national UNOS database is the strength of our study. We also assessed DRI in this analysis, and specifically none of the variables included in the DRI were significant to impact the main finding of the study of no effect of history of HD in the donor on the posttransplant patient and graft survival. However, our study has some limitations. For example, information was missing on stratification to HD and ND in about 15% of the deceased donors. Further, data were unavailable on quantification of alcohol consumption and stratification based on sex, limiting examining of the impact of light or moderate drinking, and of donor gender.
In conclusion, history of HD in deceased donors does not appear to impact the graft and patient survival after LT. Prospective studies are needed to identify HD donors with potentially high-risk grafts, as basis for developing recommendations and guidelines for selecting these donors for LT.
1. Starzl TE, Iwatsuki S, Van Thiel DH, et al. Evolution of liver transplantation. Hepatology
2. Singal AK, Guturu P, Hmoud B, et al. Evolving frequency and outcomes of liver transplantation based on etiology of liver disease. Transplantation
3. Tector AJ, Mangus RS, Chestovich P, et al. Use of extended criteria livers decreases wait time for liver transplantation without adversely impacting posttransplant survival. Ann Surg
4. Busuttil RW, Goss JA. Split liver transplantation. Ann Surg
5. Miller CM. Regulation and oversight of adult living donor liver transplantation. Liver Transpl
. 2003;9(10 Suppl 2):S69–72.
6. Busuttil RW, Tanaka K. The utility of marginal donors in liver transplantation. Liver Transpl
7. Cuende N, Grande L, Sanjuan F, et al. Liver transplant with organs from elderly donors: Spanish experience with more than 300 liver donors over 70 years of age. Transplantation
8. Feng S, Goodrich NP, Bragg-Gresham JL, et al. Characteristics associated with liver graft failure: the concept of a donor risk index. Am J Transplant
9. Foley DP, Fernandez LA, Leverson G, et al. Donation after cardiac death: the University of Wisconsin experience with liver transplantation. Ann Surg
10. Totsuka E, Dodson F, Urakami A, et al. Influence of high donor serum sodium levels on early postoperative graft function in human liver transplantation: effect of correction of donor hypernatremia. Liver Transpl Surg
11. Renz JF, Kin C, Kinkhabwala M, et al. Utilization of extended donor criteria liver allografts maximizes donor use and patient access to liver transplantation. Ann Surg
. 2005;242:556–63; discussion 63–5.
12. Marino IR, Doyle HR, Aldrighetti L, et al. Effect of donor age and sex on the outcome of liver transplantation. Hepatology
13. National Institute on Alcohol Abuse and Alcoholism: Alcohol use in the United States. Available at https://www.niaaa.nih.gov/alcohol-health/overview-alcohol-consumption/alcohol-facts-and-statistics
Accessed 2 February 2017.
14. Bramstedt KA. Alcohol abstinence criteria for living liver donors and their organ recipients. Curr Opin Organ Transplant
15. Organ Procurement and Transplantation Network/United Network for Organ Sharing Ad Hoc Living Donor Committee: Living liver donor evaluation guidelines. 2003. https://hods.org/pdf/content/Selection of Donors for live donation.pdf
. Accessed 18 December 2016.
16. Organ Procurement and Transplantation Network. Available at: www.optn.org
. Accessed January 1, 2017.
17. Austin PC. An introduction to propensity score methods for reducing the effects of confounding in observational studies. Multivariate Behav Res
18. Boin IF, Kajikawa P, Palmiero HO, et al. Profile of cadaveric liver donors of the OPO-UNICAMP from 2002 to 2006. Transplant Proc
19. Boin ID, Palmiero HD, Kajikawa P, et al. Reviewing the causes for 1432 discharged liver donors: can donations be increased? Transplant Proc
20. Mangus RS, Kubal CA, Fridell JA, et al. Alcohol abuse in deceased liver donors: impact on post-transplant outcomes. Liver Int
21. Moore DE, Speroff T, Gorden DL, et al. Impact of donor, technical, and recipient risk factors on survival and quality of life after liver transplantation. Arch Surg
22. D'Alessandro AM, Kalayoglu M, Sollinger HW, et al. The predictive value of donor liver biopsies for the development of primary nonfunction after orthotopic liver transplantation. Transplantation
23. Marsman WA, Rodriguez L, Batts KP, et al. Use of fatty donor liver is associated with diminished early patient and graft survival. Transplantation
24. Briceño J, Rufián S, Solórzano G, et al. Assignment of steatotic livers by the Mayo model for end-stage liver disease. Transpl Int
25. Mangus RS, Fridell JA, Vianna RM, et al. No difference in clinical transplant outcomes for local and imported liver allografts. Liver Transpl