Impact of donor age on liver transplant outcomes in patients with hepatocellular carcinoma: Analysis of the SRTR dataset

Background: Donor age is an important predictor for liver transplant recipients. Studies have not fully explored its impact on transplant outcomes in hepatocellular carcinoma (HCC) patients, and its involvement in tumor recurrence has not yet been fully explored. Methods: HCC recipients who received liver transplants from 2010–2017 from the Scientic Registry of Transplant Recipients database were included. The recipients were divided into four groups on the basis of donor age: I (18–34 years), II (35–49 years), III (50–64 years), and IV ( ≥ 65 years). Transplant outcomes, including the overall survival (OS), tumor recurrence, and risks, were evaluated. Results: A total of 13,276 HCC recipients were included in this study. The OS was signicantly different among four groups. The best 5-year survival was 76.0% in group I, followed by 73.5% in group II, 72.8% in group III, and 69.2% in group IV (P < 0.001). Donor age, recipient age, recipient race, recipient serum creatinine, sum of tumor diameter, and α-fetoprotein (AFP) were found to be independent predictors of OS. In subgroup analysis, OS was signicantly different in hepatitis C virus (HCV) recipients, but there was no signicant difference between recipients with hepatitis B virus (HBV), alcoholic diseases and nonalcoholic steatohepatitis (NASH). The cumulative tumor recurrence rates and the mean time to recurrence post-transplantation did not differ among the four groups. Conclusions: Older donor age was associated with decreased OS in HCC recipients, but it was not correlated with post-transplant tumor recurrence. Older donor age also had different effects in patients with different underlying liver diseases

studies have recently shown the safety of use of old donor livers, even using livers from octogenarian donors [7].
Despite this, the effect of donor age on liver transplantation speci cally in HCC has not yet been fully elucidated. Tumor recurrence after liver transplantation is an important concern in such patients. We have already established recipient selection criteria so that donors are matched with the most feasible patients to obtain favorable transplant outcomes [8,9]. However, currently, practical experience with regard to how donor variables such as donor age could affect transplant outcomes, including tumor recurrence, in HCC patients is scarce. Although previous studies have demonstrated the association of donation after cardiac death (DCD) and post-transplant mortality in HCC recipients, they have not focused on tumor recurrence [10]. Orci et al. previously evaluated the effect of donor characteristics on tumor recurrence after liver transplantation based on the Scienti c Registry of Transplant Recipients (SRTR) [11]. Their study included patients from 2004-2011. Policies surrounding treatment of HCC have changed during this time, with improvements in liver transplantation and increased use of marginal donors. Therefore, it is important to re-evaluate donor characteristics on the transplant outcomes in HCC patients in this setting.
Here, we use renewed data from the SRTR to evaluate the effect of donor age on liver transplant outcomes, Research Institute (HHRI) as the contractor for the Scienti c Registry of Transplant Recipients (SRTR). The interpretation and reporting of these data are the responsibility of the author(s) and in no way should be seen as an o cial policy of or interpretation by the SRTR or the U.S. Government [12]. Organs from executed prisoners were not used in this study. The protocol for the present study was in accordance to Declaration of Helsinki and was approved by the Ethics Committee of the First A liated Hospital, College of Medicine, Zhejiang University, China (approval number 2019-1020).
We included patients with HCC who received liver transplantation from through January 1, 2010, to December 31, 2017. The inclusion criteria included recipients ≥18 years of age, with a primary diagnosis of "hepatocellular carcinoma" or "hepatoma" at transplant. Patients with a previous liver transplantation, those who received transplant for benign liver disease or liver tumor other than HCC, and those <18 years of age were excluded from participation in this study. Among all the liver transplant recipients during this time period, a total of 13,276 recipients were eligible based on the inclusion and exclusion criteria. Figure 1 illustrates the patient selection owchart.
To assess how donor age affects transplant outcomes, we divided recipients into four groups according to donor age: group I, donor age =<34 years (N = 4,723); group II, donor age 35-49 years (N = 3,572); group III, donor age 50-64 years (N = 3,743); group IV, donor age >=65 years (N = 1,238). The donor and recipient data as well as tumor characteristics were compared among the four groups.
For long-term outcomes, we rstly compared post-transplant OS among four groups, which was evaluated from the time of LT and de ned recipient death as the endpoint. OS was also assessed according to recipient underlying liver diseases including hepatitis C virus (HCV), hepatitis B virus (HBV), alcoholic liver diseases and nonalcoholic steatohepatitis (NASH). The post-transplant HCC recurrence was then compared among the four different groups, and the de nition of post-transplant HCC recurrence was in accordance with the report by Samoylova et al. [13]. The detailed identi cation procedure was performed in accordance with that described in the study by Orci et al. [11].

Statistical analysis
Donor and recipient data and tumor characteristics were compared using one-way ANOVA analysis of variance for continuous variables and the chi-square test for binomial variables. OS was assessed by the Kaplan-Meier method and log-rank test was used to compare differences among the groups. Univariable analysis was used to identify potential predictors for OS, and those with P < 0.05 were further analyzed using the Cox proportional hazard ratios (HRs) model to adjust for potential confounders. Time-dependent effects were evaluated based on Schoenfeld's residuals, and cubic spline functions were also introduced in the model [14]. The cumulative tumor recurrence rates were evaluated using a competing risk model. A two-tailed P value

Baseline characteristics
The median follow-up time was 36 months (interquartile range, 18-60 months) for the entire study population. Donor characteristics, including donor height, weight, race, ABO blood type, sex, cause of death, deceased donor type (DCD or not) signi cantly differed among the four age groups. For recipient characteristics, the recipient age, race, ABO blood type, presence of underlying liver disease, height, weight, cold ischemia time, BMI, calculated laboratory model for end-stage liver disease (MELD) score, pretransplant laboratory albumin, bilirubin, international normalized ratio (INR), and creatinine were all statistically different among the four groups, while the donor warm ischemia time and pretransplant laboratory sodium level were comparable among all four groups. Analysis of tumor characteristics revealed that the pre-transplant treatment, including transarterial chemoembolization (TACE), radiofrequency ablation (RFA), chemotherapy, cryoablation, surgery, number of tumors, largest tumor diameter, sum of tumor diameters, tumor staging, and pre-transplant α-fetoprotein (AFP) level were all similar among four groups. Table 1 presents a summary of the data in detail.  Figure 2). We also compared the survival between two age groups at a time, and observed that the survival of group I recipients was signi cantly better than that of the other three groups (group I vs. group II, P = 0.049; group I vs. group III, P = 0.002; group I vs. group IV, P < 0.001). The survival of group II recipients was comparable to that of group III (P = 0.308) but signi cantly better than that of group IV recipients (P = 0.004). The survival of group III recipients was also better than that of group IV (P = 0.038).
Next, we analyzed the survival according to underlying liver diseases in the recipient. The OS signi cantly differed among the recipients with HCV, with the 1-,  Figure 3d).

Univariable analysis for OS
We then performed univariable analyses to identify potential risk factors for recipient OS. Donor characteristics, including donor age and cause of death; recipient characteristics, including recipient age, race, underlying liver diseases, pre-transplant laboratory MELD score, albumin, bilirubin, INR, creatinine, sodium; and tumor characteristics such as the number of tumors, largest tumor diameter, sum of tumor diameters, tumor staging, and pre-transplant AFP level, were all found to be signi cantly associated with the OS. Table 2 presents this information in greater detail.

Multivariable analysis for OS
The multivariable Cox regression analysis showed that donor age, recipient age, recipient race, underlying liver diseases, pre-transplant MELD score, creatinine, sum of tumor diameters, and AFP level were all independent predictors of OS. Table 3 presents this information in greater detail.  Figure 1).
As no difference was observed in terms of recurrence rates among the groups, additional Cox regression model and competing regression model were not used. However, we further investigated the time-dependent effect of donor age on tumor recurrence in a sensitive analysis. The effect of donor age also remained generally stable in the early transplant years, and showed a decreasing trend over the long-term follow-up period (Supplemental Figure 2).

Discussion
In this study, we demonstrated that the OS signi cantly differed in HCC liver transplant recipients categorized according to different donor ages, whereas donor age had no effect on post-transplant tumor recurrence. With regard to underlying liver diseases, there were differences in OS in HCV recipients, but not in recipients with HBV, alcoholic liver diseases and NASH.
The impact of donor age on transplant outcomes has been evaluated for decades, in an attempt by clinicians to expand the donor pool by using organ grafts from older donors. However, there are controversies surrounding the safety and feasibility of grafts from older donors compared with grafts from younger donors. Several studies have shown that older donor age is associated with an increased risk of post-transplant graft loss and poor patient survival. One study investigated the correlation of donor-related variables with posttransplant outcomes, and derived the DRI model [6]. In this model, donor age was reported to be an independent predictor of post-transplant survival, with a 1-year survival of 85-86% in recipients whose donors were <40 years of age versus 61-76% in recipients whose donors were over 70 years old. This DRI model was also validated in a European study [15]. Other modi ed models based on DRI have also been proposed [16,17].
Studies have shown that the impact of donor age is especially important in recipients with underlying HCV infection, leading to more advanced hepatic brosis after transplantation as well as decreased survival [18][19][20]. While these studies almost always showed the negative association of older donor age with survival, several centers have reported early experiences of successful use of septuagenarian and even octogenarian donors, whose transplant outcomes were comparable with those of younger donors. Clinicians who believed that adequate selection of donor liver and donor-recipient matching are the keys to successful use of liver grafts from older donors have endeavored to push the age limits of eligible liver graft donors [4,7]. Other studies have pointed out that the main difference in transplant outcomes between grafts from older and young donors mainly presents in the rst transplant year, and this difference may stabilize in the subsequent years after the transplant [21,22].
While the impact of donor age on transplant outcomes has been widely investigated, little is known about how donor age affects tumor recurrence in recipients with HCC. Liver grafts are likely to undergo senescence (both morphological, synthetic, and functional) with age, seen as shrinking in size and a decrease in metabolic and immune cell functions [23]. Therefore, pathophysiological changes during IRI in liver grafts from old donors differ considerably from those in liver grafts from young patients. In fact, studies have shown that liver grafts from older donors would be more vulnerable to IRI, with deregulation in liver cell mitochondrial alteration, in ammatory response, and autophagic ux [5]. The study by Reese and Gao showed that special attention should be paid to cold ischemia time in cases where the age of the graft donor is greater than 70 years [24,25]. Increased IRI was also found to be associated with tumor recurrence. Man et al. demonstrated that more severe early-phase IRI in small-for-size liver grafts has a higher probability of late-phase tumor recurrence, which might be due to post-transplant mobilization of endothelial progenitor cells and regulatory T cells [26,27]. Thus, there is a potential that older donor age can lead to higher tumor recurrence rate after transplantation. In fact, in a previous study by Orci et al., which also evaluated the association of donor-related variables and HCC recurrence using SRTR data observed that older donor age was associated with higher HCC recurrence [11]. However, their data included recipients who received a liver transplant from 2004 to 2011, and the transplant policies for HCC patients have been modi ed in the past decade, with different indications for liver diseases, new transplant techniques, and improved post-transplant management. Indeed, signi cant progress has been made in the eld of transplantation over the past decades, with improvements in patients, graft survival, and fewer disparities in recipients who received grafts from donors >60 years compared with recipients of grafts from donors below 60 years of age [25]. Considerable progress has been made in treating HCC, with the development of new targeted therapies and immune therapies [3]. Therefore, it is worth reinvestigating the impact of donor age on HCC patients in the current transplant era.
We found that although there were signi cant differences in the OS among transplant recipients from donors of different age groups, no difference was observed with regard to tumor recurrence. We also evaluated the time-dependent HR of donor age during transplant follow-up for both OS and tumor recurrence. The HR for OS remained relatively stable over time and showed an increasing trend in the early transplant months and a decreasing trend with longer transplant years. This might indicate that donor age remained to be the independent predictor for recipient OS although its effects might lessen with longer transplant follow-up time.
The HR for post-transplant tumor recurrence also remained stable during the transplant follow-up years, indicating that the probability of tumor recurrence would not increase with longer follow-up time. As the cumulative tumor recurrence risk did not differ among the four age groups, we only performed sensitivity analysis and did not perform the competing risk model to assess the risk of tumor recurrence. This might re ect improvements in the management of HCC in transplant recipients.
We observed the differential effects of donor age on recipients with different underlying liver diseases. In patients with HCV, older donor age was associated with decreased OS, while in patients with HBV, alcoholic diseases and NASH, donor age did not affect post-transplant survival. This result is in accordance with that reported by Lake et al. [28], who reported that donor age did not have any effect on post-transplant outcomes in recipients with HBV. Their study also showed that transplants from donors >60 years were associated with decreased survival in patients with underlying liver diseases other than HBV and HCV. However, they did not sub-classify these diseases, which we think is important in the current transplant status, as the number of patients with NASH has shown an increasing trend in the past few years and it has become one of the leading indications of liver transplantation in current clinical practice [29,30]. Therefore, our preliminary results on the impact of donor age on liver transplant recipients with HCC with different underlying liver diseases would bene t from validation and further exploration in a larger population so as to study the different causes of liver diseases in transplant recipients.
Recipient age was an independent predictor for OS in HCC recipients, which is in accordance with that reported previously, where older recipient age was associated with inferior long-term survival, particularly in those with higher MELD scores [31][32][33][34]. Patients with HCC are generally older than other waiting list candidates [30,33,35].
Factors associated with older age, including cardiovascular diseases, functional status such as frailty as well as higher extrahepatic tumor risks were all the negative predictors associated with older age, which lead to adverse outcomes in HCC recipients. Our study also observed that a higher MELD score was an independent predictor of OS. Further studies could evaluate the association between recipient age and HCC recurrence after transplantation.
Our study also found that sum of tumor diameters and pre-transplant AFP level were independent predictors of OS, re ecting the importance of the candidate selection process in HCC patients, which we should not only evaluate the general status of the recipient but also study the tumor characteristics using HCC selection criteria, such as the Milan Criteria and etc [8,36]. However, in our previous study we mainly focused on the tumor recurrence based on HCC selection criteria. Our study showed that in addition to tumor recurrence,

Conclusions
Our study demonstrated that older donor age was associated with inferior OS in transplant recipients with HCC, but donor age did not affect post-transplant tumor recurrence in these patients. Older donor age also had different effects in recipients with different underlying liver diseases, with inferior OS in recipients with HCV but with no in uence in recipients with HBV, alcoholic diseases and NASH. These ndings may be useful for clinicians in decision-making with regards to marginal donor allocation and recipient selection to achieve  Flow chart of patient selection process.

Figure 2
Overall survival of HCC recipients categorized by different donor age groups.

Figure 3
Overall survival of HCC recipients according to underlying liver diseases: a) HCV; b) HBV; c) alcoholic liver diseases; d) NASH.