Abnormal Liver Biopsies of Donor Grafts in Pediatric Liver Transplantation: How Do They Fare?

Introduction

The shortage of livers available for transplantation in children on the waiting list remains a critical problem for pediatric patients with end-stage liver disease. The number of pediatric liver transplants performed in the United States has slightly decreased from its peak in 2018, driven primarily by fewer deceased donor liver transplants [1]. In addition, 100 potential pediatric allografts are discarded every year for various causes, such as donation after circulatory death (DCD) status and elevated liver function tests [2]. Traditionally, transplanting surgeons have relied on intraoperative subjective assessments by the procuring surgeon regarding allograft characteristics such as fat percentage [3,4]. As such, there is limited knowledge on what the lower and upper limits are for utilizing donor liver grafts with abnormal features. When there are concerns of usability on visual inspection, a biopsy can be performed to objectively evaluate histological features of the liver. Typically, healthy-looking livers with minimal macrosteatosis are used for pediatric patients.

Donor liver biopsies are a useful tool to characterize the quality of a liver. This can be performed by organ procurement organizations (OPOs), and up to 10% of donor livers are biopsied [5]. A study on the use of donor biopsies in adult and pediatric DCD livers found that 80% of those biopsied were retained and successfully transplanted in comparison to 49% being transplanted in those not biopsied, highlighting the value of histological data in maximizing the utilization of grafts [6].

Use of donor liver biopsies has not been examined in pediatric liver transplantation, and outcomes of pediatric patients transplanted with macrosteatotic, fibrotic, and/or inflamed livers have not been well-studied. The aim of this study was to evaluate the outcomes of adult and pediatric livers transplanted into pediatric recipients that underwent a donor biopsy.

Material and Methods

The UNOS Standard Transplant Analysis and Research (STAR) files were used to identify all pediatric patients who received a deceased donor liver transplant between January 2015 and December 2022 in the United States. Only recipients with liver biopsy data were included in the analysis. A pediatric recipient was defined as one who was transplanted prior to the age of 18 years. Recipients of multi-organ transplants and those recipients who were lost to follow-up were excluded from the analysis.

Biopsies were categorized as follows: 1) livers with ≤5% macrosteatosis, no fibrosis, and no inflammation were classified as normal livers (NL); 2) allografts with either >5% macrosteatosis, any fibrosis, or any inflammation were considered abnormal livers (AL). Biopsy data from the STAR file are determined by pathologists at or around the time of procurement who do not know if the liver will be transplanted. If done on site, biopsies are usually read by a pathologist at the donor hospital.

Demographic data for donors and recipients were examined, and outcomes of interest included length of stay and allograft survival. STATA (v18, StataCorp, College Station, TX), R (v4.2.3, R Foundation for Statistical Computing, Vienna, Austria) and GraphPad Prism (v10.2.2, GraphPad Software, Boston, MA) software were used for statistical analyses. Continuous variables were compared using Wilcoxon rank-sum test, and categorical variables were compared using either the Chi-squared or Fisher’s exact test. Differences were considered significant when p<0.05. Data are expressed as median (IQR). Institutional Review Board (IRB) exemption was obtained for this study from our institution.

Results

GRAFT CHARACTERISTICS OF NL AND AL ALLOGRAFTS:

Of the 213 allografts meeting inclusion criteria, 114 NL and 99 AL were identified. Among the AL, 35.4% (35/99) had >5% macrosteatosis, 64.6% (64/99) had any inflammation, and 18.2% (18/99) had any fibrosis (Figure 1). The average macrosteatosis was 1.2±2.1% in NL (95% CI: 0.82–1.6) and 7.6±11.4% in AL (95% CI: 5.3–9.9) (p<0.001). 20.2% of the AL grafts and 22.8% of the NL grafts were split, partial, or reduced livers (p=0.6) (Table 1).

DONOR CHARACTERISTICS OF NL AND AL ALLOGRAFTS:

The median age of AL graft donors was 22y (IQR: 7.5–36, 95% CI: 20–27), which was significantly older than NL donors at 14.5 y (IQR: 2.3–25, 95% CI: 14–19) (p=0.002). Donor risk index (DRI) (1.8 vs 2.0, p=0.1) and donor BMI (22.6 vs 21.3 p=0.12) were similar in AL vs NL groups. There were no other significant differences in donor gender, ethnicity, comorbidities, or cause of death (Table 1).

RECIPIENT CHARACTERISTICS OF NL AND AL ALLOGRAFTS:

Recipients of AL grafts had significantly higher PELD scores at time of transplant than NL recipients, with a median of 17 (IQR: 8–30.5, 95% CI: 15–21) and 13 (IQR: 0–21, 95% CI: 9.9–15), respectively (p=0.011). AL graft recipients tended to be older, with a median age of 7 (IQR: 1–14) compared to NL recipients with a median age of 3.5 (IQR: 1–10, p=0.059). The most common primary diagnoses for all recipients were biliary atresia (n=66), metabolic disease (n=33), and acute hepatic necrosis (n=24), but the AL group tended to have more recipients with cirrhosis and acute hepatic necrosis. There was a trend towards more recipients with a prior liver transplant in the AL group (4.4% of NL vs 10% of AL, p=0.10). There were no significant differences between AL vs NL grafts with regard to recipient gender, ethnicity, time on waitlist, or BMI (Table 2).

OUTCOMES OF NL AND AL ALLOGRAFTS:

There was no significant difference in long-term allograft survival between the NL and AL cohorts (HR=1.569 [0.66, 3.70], p=0.30) (Figure 2). Post-transplant length of stay and rejection rates at 6- and 12-months post-transplant were similar between the 2 groups. There were 15 and 9 rejection episodes during hospital admission in the NL and AL groups, respectively (Table 3). There were also no differences when comparing the causes of allograft failure, such as primary non-function, thrombotic events, and recurrent disease. Of the 8 NL grafts with known graft failure causes, 6 were due to hepatic artery, portal vein, or other vascular thrombosis, 1 was due to primary non-function, and 1 was due to chronic rejection (Table 3). Of the 4 AL grafts with known failure causes, 2 were due to primary non-function, 1 was due to recurrent disease, and 1 was due to hepatic artery thrombosis (Table 3). One year allograft survival rates in NL and AL were 95.5% and 93.9%, respectively (p=0.6) (Table 3). There were 7 recipients of donor livers with >30% macrosteatosis, and among these, 1 graft (14%) failed at 2.5 years post-transplantation. The majority of AL in this study also had Ishak fibrosis scores of 0–1 or periportal infiltration scores of 0–1. Of the 4 AL with Ishak fibrosis score of 2, 1 graft (25%) failed, which was the same graft as the failure event in the >30% macrosteatosis. Of the 3 AL with portal infiltration score of 2, there were no graft failure events.

In addition, we analyzed and compared specific subsets of AL by each abnormal variable. A total of 35 AL had >5% macrosteatosis, 64 had inflammation, and 18 had fibrosis. Of these, 83 had only 1 abnormal variable, 14 had a combination of 2 variables, and 2 had all 3 variables (Figure 1). Among the 3 independent variables, inflammation appeared to have the highest impact on allograft survival. Presence of inflammation led to a two-fold greater risk for graft failure compared to NL (HR=2.641 [1.04, 6.71], p=0.04), and neither macrosteatosis nor fibrosis led to a significant difference in graft survival (Figure 3). Of the 12 AL grafts that were lost, inflammation was present in 92% (11/12), >5% macrosteatosis was present in 33% (4/12), and fibrosis was present in 17% (2/12). Chi-square analysis revealed that graft inflammation had a higher propensity for graft loss compared to NL (p=0.06) (Figure 4). Finally, a multivariate Cox proportional hazards model using all 3 abnormal factors as covariates showed that inflammation alone was an independent negative prognostic factor for survival of the grafts (HR=2.36 [1.08, 5.15], p=0.03) (Table 4).

Discussion

Every year, up to 60 children die waiting for a liver transplant, representing 16% of the national pediatric waitlist [7]. The use of “marginal” livers, including those with intrinsic abnormalities, such as higher degree of macrosteatosis, the presence of fibrosis, or inflammatory portal infiltrates, is a potentially effective solution for helping ameliorate the supply and demand imbalance of pediatric livers in the United States [8,9]. While these livers as well as technical variant grafts have shown promising outcomes for expanding the donor pool, there is currently a wide variation in their adoption for pediatric patients, and their use in the Unites States remains uncommon [8].

With the rising prevalence of obesity and non-alcoholic fatty liver disease in the United States, macrosteatosis is an important histological abnormality to consider in liver allografts. Since the 1980s, it has been known that donor liver macrosteatosis can be a risk factor for allograft survival and primary graft dysfunction [10]. However, a recent meta-analysis of steatotic liver transplantation outcomes determined that livers with moderate (30–60%) to severe (>60%) macrosteatosis were suitable for adult transplant recipients, with no difference in graft or patient survival compared to control livers when appropriately matched [11]. Historically, in pediatric liver transplantation, healthier grafts with lower levels of macrosteatosis are preferred. Recent studies have shown that obese donors or liver grafts with >5% macrosteatosis do not adversely impact the outcomes of pediatric transplant recipients, thus we chose 5% macrosteatosis as a threshold for our study [12,13]. Irie et al demonstrated that pediatric patients who received a living donor liver transplant with moderate-to-severe macrosteatosis also had no difference in patient survival and that the degree of macrosteatosis decreased over time [14].

In contrast to macrosteatosis, other donor liver allograft abnormalities have not been well-studied. For example, a donor liver biopsy also often provides the degree of fibrosis. A study from D’Errico et al found that patients who received a liver allograft with histologically-proven portal fibrosis had a significantly lower patient and graft survival in comparison to patients without fibrosis [15]. Wadhera et al found no difference in graft survival when using liver allografts containing early-stage fibrosis [16]. However, these studies were not specific for the pediatric population. Thus, the effect of donor allograft fibrosis in pediatric patients has yet to be assessed. Portal inflammation is another variable that can be evaluated using donor liver biopsy. Liver allografts with portal infiltrates on biopsy are higher risk for worse outcomes in long-term graft survival in adult recipients [17]. In children transplanted from living donors, portal inflammation did not affect outcomes, but patients with higher PELD scores had a trend of lower graft and recipient survival rates [18].

In this study, we assessed the impact of multiple types of histological abnormalities on outcomes in pediatric liver transplant recipients. We categorized the donor livers into normal vs abnormal using donor liver biopsy data and found that there were no significant differences in outcomes between groups. Abnormal livers yielded excellent short- and long-term graft survival outcomes. Furthermore, the percentage of patients who underwent re-transplantation with an AL graft was twice that of NL grafts, yet they still had comparable outcomes. This suggests that abnormal grafts may be suitable even for appropriately selected higher-risk patients, such as those with higher PELD scores or who had a prior liver transplant. When comparing outcomes of AL by each abnormal variable, we further found that AL with >5% macrosteatosis were no different in terms of overall graft survival compared to NL. Our results correspond with the evolving paradigm that livers with increased levels of macrosteatosis may still be suitable for pediatric transplants [10–14].

In our cohort, the presence of inflammation in donor liver biopsies led to a lower overall graft survival and was found in 90% of graft loss events. AL with fibrosis seemed to have no difference in outcomes when compared to NL. Identifying the extent of fibrosis and inflammation in donor livers could be of particular interest, given that post-transplant fibrosis and inflammation are fairly common in pediatric patients [19–21]. Although this graft damage is often subclinical, one multicenter study discovered from protocol surveillance biopsies that up to 53% of liver grafts had chronic hepatitis and 79% had fibrosis 10 years after transplantation [19]. Given that pediatric transplant recipients are now living longer, the prior existence of inflammation or fibrosis in the donor liver allograft could have serious implications for long-term graft outcomes. It would be interesting to see if findings from donor liver biopsies could offer any information on the development of progressive post-transplant fibrosis or inflammation, which would help guide early identification and management of these patients.

The limitations of this study include small sample size, given the low incidence of donor liver biopsies for pediatric recipients. Certain potential high-risk characteristics of interest such as macrosteatosis ≥30%, Ishak score >1, and moderate portal inflammation score >1 were also uncommon in our sample (out of 10 grafts, only 1 failed at 2.5 years post-transplant). Also, the threshold of marginal organ acceptance from pediatric transplant institutions is unclear. For example, the threshold percentage of liver macrosteatosis has not been well-defined in pediatric liver transplantation as it has been for adults and offers an opportunity for further investigation. There is likely underutilization of abnormal grafts for pediatric recipients as many of these allografts are transplanted into appropriately sized adult recipients. Because of our size limitations, it is especially important to validate our findings with larger cohorts, either from multicenter studies or dedicated prospective cohorts designed to study abnormal donor biopsy outcomes in pediatric transplantation patients. Our data was limited in the granularity of each abnormal finding, and further studies could help define more accurate lower and upper limits of biopsy results for donor grafts.

Conclusions

Our results demonstrate that donor livers with abnormal biopsies can be safely transplanted to help ameliorate the organ shortage in pediatric transplantation. Careful consideration is still warranted for increased utilization of abnormal livers in pediatric recipients, but our study suggests that livers with >5% macrosteatosis, fibrosis, or inflammation may be suitable. Even when considering donor livers with macrosteatosis ranging from 5–80% and any kind of inflammation or fibrosis, the percentage of grafts lost was comparable between both groups, and there was no increased risk in using the abnormal livers for pediatric transplantation. The use of liver biopsies for donor grafts in pediatric transplantation could be an effective way to characterize allografts and facilitate research to further define and more broadly expand the acceptance criteria.