Risk Factors Affecting Outcomes in Pediatric Liver Transplantation: A Real-World Single-Center Experience

Background

Since the first liver transplant (LT), which was performed by Starzl in 1963, the outcomes of LT have gradually improved due to development in surgical techniques, perioperative management, and immunosuppressive agents. LT is now considered the treatment of choice for acute liver failure and chronic end-stage liver disease in both adults and children [1].

In 1988, our center performed the first LT in Korea, for a 14-year-old girl suffering from Wilson disease; she survived for 30 years after LT [2]. Living donor liver transplantation (LDLT) was first performed in Korea in 1999, in a 1-year-old boy with biliary atresia who is still alive. Since the first LT in 1988, more than 250 pediatric LTs have been performed. Our pediatric LT program has developed into 3 stages [3]. The LT program, which includes adult and pediatric LTs, was set up in the early stage (1988–2007). The improvement in survival after LDLT and the prospective collection of donor data started in the middle stage (2008–2011). In the recent stage (2012–2018), recipient data were prospectively collected and a specialized LT program for children was created; furthermore, there was an increase in the number of split LTs. Under this program, we previously reported that the rate of hepatic artery complications was higher in pediatric recipients with metabolic liver disease than in those with biliary atresia [3]. However, the study had several limitations, including confinement to metabolic liver disease and biliary atresia as disease entities.

This study aimed to evaluate the outcomes of pediatric LT in Korea and identify the risk factors affecting the outcomes, including hepatic artery complications.

Material and Methods

STATISTICAL ANALYSIS:

Results were expressed as mean and standard deviation (SD) for continuous data and as numbers with percentages for categorical data. Overall patient survival and graft survival rates were calculated using the Kaplan-Meier method and compared between groups (male vs female, age <1 vs ≥1 year, body weight <6 vs ≥6 kg, early vs middle vs recent period of LT, LDLT vs whole-liver DDLT vs split DDLT, BA vs others, fulminant vs others, chronic liver disease vs others, pediatric end-stage liver disease [PELD] score <30 vs ≥30, Child-Pugh score A vs B or C, hepatic artery complication yes vs no, portal vein complication yes vs no, hepatic vein complication yes vs no, bile duct complication yes vs no, acute cellular rejection yes vs no, post-transplant lymphoproliferative disorder yes vs no, donor sex male vs female, donor age <18 vs ≥18 years, and relationship to recipient mother vs other than mother) using a log-rank test. Factors independently associated with recipient and graft survival (P<0.05) were then included in multivariate Cox proportional hazards regression models. Univariate analysis of factors associated with hepatic artery complications was also performed. Factors independently associated with hepatic artery complications (P<0.05) were included in a subsequent multivariate logistic regression model using backward selection. A P value <0.05 was considered statistically significant. SPSS (SPSS Inc., Chicago, IL) was used for the statistical analysis.

Results

BASELINE CHARACTERISTICS:

The demographic and clinical characteristics of patients and their donors are summarized in Table 1. Of the 255 patients, 116 were male and 139 were female. The mean age was 59.8 months and the mean body weight was 19.2 kg. There was a significant majority of LDLTs (64.3%); biliary atresia (55.7%) was the most common underlying liver disease, followed by metabolic liver disease (13.3%). The proportion of liver disease with underlying chronic liver disease was 77.6%. The mean Child-Pugh and PELD scores were 8.2 and 13.0, respectively. Of the 255 donors, 139 were male and 116 were female. The mean donor age was 28.9 years and the mean body weight was 60.8 kg. Among the 164 patients who underwent LDLT, the mother (n=79) was the most common donor, followed by the father (n=64).

SURVIVAL OUTCOMES AND RISK FACTORS:

The 1-, 5-, and 10-year overall survival rates were 90.5%, 88.4%, and 87.8%, respectively (Figure 1A), while the 1-. 5-, and 10-year graft survival rates were 89.7%, 88.5%, and 87.2%, respectively (Figure 2A). Graft failure was recorded when the patient underwent retransplantation or when the graft was no longer functioning at the time of death. The most common cause of graft failure was primary nonfunction, followed by hepatic artery complications and acute cellular rejection (Table 2). Univariate analysis showed that factors that were significantly associated with patient survival were body weight (<6 vs ≥6 kg), type of LT (LDLT vs whole-liver DDLT vs split DDLT), underlying liver disease (fulminant vs others, with vs without chronic liver disease), PELD score (<30 vs ≥30) and postoperative hepatic artery complications (Table 3). Multivariate analysis showed that liver disease without underlying chronic liver disease (hazard ratio [HR]: 2.69, confidence interval [CI]: 1.14–6.38, P=0.024) (Figure 1B) and PELD score (HR: 2.81, CI: 1.07–7.37, P=0.036) (Figure 1C) were the only factors that were independently associated with overall survival.

Factors affecting graft survival were body weight <6 kg (P=0.035), type of LT (LDLT vs whole-liver DDLT vs split DDLT) (P=0.043), liver disease other than biliary atresia (P=0.015), fulminant liver failure (P=0.001), liver disease without underlying chronic liver disease (P=0.004), and postoperative hepatic artery complications (P < 0.001) according to the univariate analysis (Table 4). Multivariate analysis showed that body weight <6 kg (HR: 2.91, CI: 1.00–8.45, P=0.050) (Figure 2B), whole-liver DDLT compared to LDLT (HR: 3.93, CI: 1.71–9.03, P=0.001) (Figure 2C), fulminant liver failure (HR: 3.13, CI 1.35–7.27, P=0.008) (Figure 2D), and postoperative hepatic artery complications (HR: 6.1, CI: 2.42–15.36, P<0.001) (Figure 2E) were the only factors that were independently associated with graft survival.

RISK FACTORS AND OUTCOMES OF HEPATIC ARTERY COMPLICATIONS:

Hepatic artery complications are surgical complications that are critical (ie, they directly result in patient or graft death); therefore, another univariate analysis was performed to identify the risk factors of postoperative hepatic artery complications.

According to the univariate analysis, liver disease other than biliary atresia (P=0.036), fulminant liver failure (P=0.013) and liver disease without underlying chronic liver disease (P=0.004) were significantly associated with postoperative hepatic artery complications (Table 5). Multivariate analysis identified that liver disease without underlying chronic liver disease was the only factor independently associated with hepatic artery complications (HR: 5.22, CI: 1.73–15.76, P=0.003).

Among 14 patients who experienced postoperative hepatic artery complications (HAT: n=11, HAS: n=3), 4 patients (28.6%) eventually died. Six patients underwent retransplantation; 4 of them underwent thrombectomy and hepatic artery reconstruction before retransplantation. Furthermore, 6 patients underwent hepatic artery reconstruction without retransplantation. Consequently, 2 patients eventually died due to graft failure, while the others survived. One patient who underwent thrombolysis survived without retransplantation while another patient who only underwent arteriography without further intervention died. The details of patients who experienced hepatic artery complications are summarized in Table 6.

Discussion

The overall patient and graft survival rates observed in this study were excellent and consistent with previous studies conducted worldwide on LDLT and DDLT [9–11]. The risk factors for overall and graft survival rates from the present study also matched the results of previous studies [9–11].

Children with higher PELD scores were more likely to be hospitalized on parenteral nutrition and under intensive care unit management at the time of LT, which reflects a worse pre-LT condition. The PELD score has been used as a predictor of mortality in children with chronic liver diseases listed for LT as well as in children with acute liver failure [12]. Previous studies showed that patients with PELD scores >20 do not show significantly poorer survival compared to that of those with a PELD score <20; another study showed that the cutoff PELD score was 33 as there was a significant survival difference at that point [3,12]. Similarly, the present study showed no significant survival difference when the PELD cutoff value was 15 or 20, whereas a significant difference in survival was seen when the PELD cutoff value was 30.

Despite the significance of both fulminant liver failure and liver disease without underlying chronic liver disease as risk factors for overall survival in the univariate analysis, fulminant liver failure was not a significant risk factor in the multivariate analysis. Liver disease not accompanied by chronic liver diseases such as fulminant liver failure, hepatoblastoma, and immature teratoma, and other metabolic liver diseases (glycogen storage disease, tyrosinemia type 1, primary hyperoxaluria, urea cycle defect, and factor H deficiency) were more predictive of the survival outcome than fulminant liver failure alone. On the contrary, isolated fulminant liver failure and not liver disease without chronic liver disease was a significant independent risk factor for graft survival. This can be explained by the fact that 5 patients without chronic liver disease died of causes other than graft failure, and only one of these patients had fulminant liver failure. In other words, a patient with liver disease other than fulminant liver failure without underlying chronic liver disease is likely to die of causes other than graft failure.

Whole-liver DDLT compared to LDLT was another significant independent risk factor for graft survival (P=0.001), with similar patient overall survival (P=0.471) according to multivariate analysis as a result of an increased retransplantation rate in the whole-liver DDLT group. This finding was similar to that of another study, which reported that graft survival, especially in patients ≤6 years old, was better for LDLT compared to DDLT [13]. Further analysis was performed to identify any difference in the distribution of DDLT and LDLT according to the underlying liver disease. No significant difference in the distribution of DDLT and LDLT was observed in patients with biliary atresia (P=0.482). However, more patients with fulminant liver failure underwent LDLT than DDLT (14.6% vs 5.5%, P=0.028), showing that poorer graft survival in whole-liver DDLT is not due to the effect of the underlying liver disease. Although patients with fulminant liver failure have a Korean Network for Organ Sharing status 1 priority, similar to United Network for Organ Sharing status 1, most of them undergo emergent LDLT instead of DDLT in Korea, an LDLT-dominant country, due to the organ shortage. Among 53 patients who underwent whole-liver DDLT, there were 11 cases of graft failure. The most common cause of graft failure in these cases was unknown (n=7), followed by primary nonfunction (n=3) and hepatic artery complications (n=1). The fact that most cases were due to unknown cause or primary nonfunction and not to other technical issues, suggests potential hidden problems in young, deceased donors.

A bodyweight <6 kg and postoperative hepatic artery complications were also associated with increased graft loss in the multivariate analysis. Young age and/or low body weight are well-known risk factors for high mortality and graft loss after pediatric LT [14–18]. Although assessment of body weight itself may often be misleading because of the effects of organomegaly and ascites, body weight is still a useful variable because it is easy to measure. Various surgical innovations are required to improve the outcomes of pediatric LT, especially in patients with body weight <6 kg and those with a liver disease without underlying chronic liver disease [4,19–24]. Therefore, increased care is required as the number of small children undergoing LT and split LT has recently increased in Korea [25].

Hepatic artery complications, including HAT and HAS, are strong risk factors for graft loss. Hepatic artery complications are well known as one of the most severe complications leading to increased morbidity, graft loss, and, eventually, patient death [26–31]. The incidence of HAT has been reported to be 3–18% after pediatric LT [25,32–34]. In our study, among 14 cases, 2 underwent non-surgical treatment, 10 underwent surgery, and 6 eventually underwent retransplantation. Owing to the organ shortage in Korea, as in other Asian countries, patients with hepatic artery complications cannot get a deceased organ at the appropriate time despite having a 1A priority status on the waiting list. Among the 8 patients who could not get retransplanted, 5 survived via an aggressive treatment strategy.

Dividing the disease category by biliary atresia vs others, fulminant liver failure vs others, and with vs without chronic liver disease may lead to some overlapping. Biliary atresia, which rapidly progresses to biliary cirrhosis and results in complications related to portal hypertension, would be included in the other than fulminant liver failure group while also being categorized in the liver disease with underlying chronic liver disease group. Considering the confounding effect and the results of multivariate analysis, liver disease without underlying chronic liver disease, rather than biliary atresia or fulminant liver failure itself, is more strongly associated with hepatic artery complications.

According to our previous study that compared the outcomes of pediatric patients with biliary atresia and metabolic liver disease undergoing LT, the rate of hepatic artery complications was higher in patients with metabolic liver disease than in those with biliary atresia [3]. In accordance with the present study, the results indicate that the rate of hepatic artery complications may increase in children without chronic liver disease not characterized by portal hypertension. Several studies reported that smaller or younger children have a higher incidence of hepatic artery complications [26,33–35]. Furthermore, children with chronic liver disease were significantly younger (80.9 vs 53.7 months; P=0.010) and tended to have lower body weight (22.8 vs 18.1 kg; P=0.053) than children who had liver disease without underlying chronic liver disease. This indicates that the disease category (chronic liver disease vs without chronic liver disease) was more strongly associated with HAT or HAS in our study sample rather the recipients’ age or weight. There are several reports describing hepatic artery enlargement in biliary atresia [36–38]. A decrease in portal vein flow in chronic liver disease or cirrhosis is compensated for by an increase in hepatic arterial flow to improve the blood supply [37,38]. The hepatic artery enlargement did not lead to a decrease in its quality in children. The intimal hepatic artery changes in patients with portal hypertension progressed gradually over time [38]. Thus, the artery is less likely to be damaged in a pediatric patient than in an adult patient. Moreover, owing to the thickening of the medial layer, the hepatic artery is not friable in patients with chronic liver disease [38]. Taken together, enlargement as a result of compensation, less intimal changes in children, and altered durability owing to medial thickening allows for easier arterial reconstruction.

It is well known that biliary complications occur in up to 50% of patients after HAT [39,40]. The present study also shows a significantly higher frequency of biliary complications in children with hepatic artery complications compared to those without hepatic artery complications (10.0% vs 35.7%; P=0.013). Among the 14 patients who suffered from hepatic artery complications, 5 patients experienced bile duct complications requiring percutaneous transhepatic biliary drainage at postoperative days 63, 1031, 381, 324, and 171.

Subgroup analysis was performed to identify any specific disease associated with poor prognosis in patients with and without chronic liver disease. Disease entities with <5 patients were grouped to strengthen the statistical power. There was no statistically significant in-group disease entity associated with poor overall survival both in patients with and without chronic liver disease (P=0.152 and P=0.840). However, there was a significant difference in graft survival among patients with different disease entities in the chronic liver disease group (P=0.009). Patients with Alagille syndrome had poorer graft survival among patients with chronic liver disease (P<0.001). Alagille syndrome is a multisystem autosomal dominant disorder [41]. Several studies reported poor outcomes after LT given the multisystemic nature of this condition [41,42]. Our subgroup analysis also showed poorer graft survival in Alagille syndrome compared with other chronic liver diseases. Among the 9 patients with Alagille syndrome in the present study, 2 patients with primary nonfunction (0.2 months and 2.4 months after LT) and one patient with portal vein complication (13.7 months after LT) eventually died. One patient who experienced recurrent cholangitis immediately after LT the graft was affected by the same disease entity as his mother, which led to the need for retransplantation. The poor outcome of these patients is caused by the multisystemic nature of the disease rather than by the low body weight or hepatic artery complications.

There are some limitations to our study. This was a retrospective study that depended on the completeness of the medical records. Since this study was performed in a single center which mostly performs LDLT, the results may not be directly generalizable to other centers that frequently perform DDLT. To draw a definitive and general conclusion, a multicenter study including all patients who are considered for LT would be needed. The Korean Organ Transplantation Registry is an ongoing prospective multicenter registration database that has been in operation since 2014. Further studies using this database after adequate long-term follow-up are needed to validate the results.

Conclusions

Greater caution is recommended in pediatric patients with a high PELD score and/or without chronic liver disease undergoing LT to improve patient survival. To improve graft survival, greater caution is recommended for patients with low body weight and/or patients with liver disease not accompanied by underlying chronic liver disease. Hepatic artery complications were the only surgical complications affecting graft survival. Therefore, further technical innovations and careful management are required to deal with hepatic artery reconstruction, especially in patients without chronic liver disease.