LT should be considered for any child with end-stage liver disease with a predicted prognosis of less than 18 months . Indications for LT are in general derived from adult liver transplant experience, but are modified for children and include :

Extrahepatic biliary atresia (BA) is the most common indication for LT and accounts for 40–50 % of cases listed worldwide. Other common causes include metabolic disorders, tumors, and acute liver failure (ALF). The majority of pediatric recipients under 2 years old have cholestatic diseases, particularly BA, which accounts for 74 % of cases in this age group. Metabolic disorders and ALF are less common indications and account for 9 % each of the overall number [3] .

The timing of LT in children has been based on criteria established in adults and thus is focused on graft and patient survival . Optimal timing was viewed as listing for LT when expected survival was less than 2 years. Children with liver disease may not develop physically, intellectually, and socially at a time of deteriorating liver function, and the timing of transplant needs to take this into account. There is general agreement that KP should be performed for BA and that LT is reserved for those who develop progressive liver disease (apart from rare cases of late presentation > 4 months).

The model for end-stage liver disease (MELD) was introduced in 2002 as a response to increasing waiting list mortality. It provides a means of allocating livers based on likelihood of dying while on the waiting list. PELD was a similar mathematical tool based on data derived from the Studies of Pediatric Liver Transplantation (SPLIT) research group using bilirubin, INR, serum albumin, age > 1 year, and growth [22]. The introduction of MELD (and subsequently PELD) significantly decreased death or removal from the waiting list for being too sick within 2 years for both adults and children [23]. Cowles et al. [24]in reviewing a cohort of 71 children transplanted for BA (61, KP before LT; 10, primary LT) considered that PELD monitoring identified those in need of transplantation. Children with a PELD greater than 12 ( n = 47) had a higher rate of post-LT mortality and retransplantation than those with a PELD of 10 or less. The authors suggested that a PELD score approaching 10 should trigger discussion of LT. PELD is the only scoring system currently used in children and although helpful in advanced liver dysfunction; it is of limited value in the very young (under 1 year of age) and in older recipients, particularly with complications such as recurrent cholangitis, severe portal hypertension, pulmonary hypertension, and hepatopulmonary syndrome [25–29]. Because of these limitations, PELD use has been largely restricted to North America. More research is needed to define optimal timing of transplantation in children to gain most benefit in terms of survival, growth, and intellectual and social development .

Whole liver replacement is relatively uncommon in children under 5 years of age at the present time . Above the age of 5 years, it is more common. The transplant involves excision of the diseased liver, by division of the common bile duct (or Roux loop if there has been previous biliary surgery), hepatic artery, PV, and IVC above and below the liver. Orthotopic liver replacement is accomplished by anastomosis of the corresponding structures with the donor liver and achieving hemostasis; the alternative is the use of the piggyback technique. Management of intraoperative coagulopathy is an essential component of the operation. The technique is very similar to adult LT, but the smaller size of the vascular structures demands a more refined surgical technique, especially for arterial reconstruction. The use of cell salvage has led to bloodless surgery becoming a practical proposition. Closure of the abdomen should only be performed if there is no risk of graft compression .

The use of partial grafts was the solution to both organ shortage and size restriction in children. In Europe, more than 10,000 LT have been performed in recipients under 16 years old. Of these, approximately 38 % have been performed with whole organs. Partial grafts account for 80 and 52 % of all LT performed among patients aged 0–2 and 2–15 years old, respectively. Early experience was with reduced-size grafts, either left lobe or left-lateral segment (LLS) which then led onto to split and LDLT with both techniques being incorporated into routine clinical practice from 1991 onwards. Roberts et al. analyzed the data of 6467 LTs performed in patients under 30 years old from the SRTR–Organ Procurement and Transplantation Network (OPTN) database [30]. It was noted that patient and graft survival during the first year after transplant for each donor graft type varied according to the recipient age group. For children of 2 years and under, living donor (LD) grafts had a 51 and 30 % lower relative risk (RR) of graft failure than deceased donor split (DD-S) and deceased donor full (DD-F), respectively. A similar difference in mortality risk in the same group of age favored recipients of LD grafts over DD-S (RR = 0.71, p = 0.08). Recipients in the 0–2-year age group had higher risk of mortality and graft failure with DD-S livers than DD-F livers (RR = 1.31, p = 0.04 for mortality; RR = 1.42, p < 0.001 for graft failure). For patients aged 2–10 years, the RRs of mortality and graft loss were higher after LD than after DD-F (RR = 1.78, p = 0.02 for mortality; RR = 1.53, p = 0.02 for graft loss) but not after DD-S. In the 11–16-year age group, a significantly higher RR of graft failure was observed after LD than after DD-F transplant (RR = 3.63, p = 0.0001) or DD-S transplant (RR = 2.87, p = 0.02), although mortality risks were similar for all three donor graft types. Subsequent publications have confirmed the excellent outcomes with LDLT in young children, but published experience with children over 12 years of age remains limited outside of India and the Far East.

Publications of outcomes of LLS reduced-size LT tend to be from the early 1990s, and there are few direct comparisons with those observed after split LT (SLT). SLT has become an established technique which has successfully addressed organ shortage in children while preserving the pool of liver grafts for adults. A recent study of 251 LTs, which included 138 reduced and 30 split, reported 1-year patient and graft survivals that were comparable at 73 and 67 %, respectively [31]. In addition, no differences in vascular complications were observed. Of note, biliary complications were significantly more common after split when compared with reduced-size grafts (21 vs. 4 %, p < 0.0001). The most common biliary complication after SLT was late stricture, in contrast to reduced-size LT (RLT), which was a cut-surface bile leak. Patients undergoing SLT had a 6.7-fold increased risk of biliary complications compared to those receiving an RLT; however, these complications did not appear to impact on graft or patient survival. A further series from the University of California Los Angeles (UCLA) has compared the outcome of whole and partial grafts in both adult and pediatric recipients [31]. Of 442 LTs, 284 were whole, 109 were split-LLS (SL-LLS), and 49 were LD-LLS. The 10-year patient survival for children was similar for all graft types. Multivariate analysis confirmed that history of previous LT and SL-LLS was independent predictors of reduced survival. Chronic rejection and hepatic artery thrombosis were the most common reasons for graft loss. The largest study from Hong et al. of outcomes after partial liver grafting in children analyzed data from the SPLIT registry and compared the outcome of each variants with that of whole organ transplantation (1183 whole, 261 split, 388 reduced, and 360 live donor (LDLT)) [32]. There was a clear difference in outcomes at 1 year (W, 93 %; R, 82 %; S, 87 %; L, 89 %) and 4 years (whole liver transplantation (WLT), 89 %; RLT, 79 %; SLT, 85 %; LDLT, 85 %) after transplant. However, the groups were not strictly comparable in terms of era of transplant, recipient selection, and center. Children receiving a technical variant waited on average 2.3 months less than those receiving a whole liver and tended to be younger. Complications were significantly higher after partial grafts: At 24 months, the incidence of biliary complications was WLT, 17.3 %; SLT, 28.5 %; RLT, 25.3 %; LDLT, 40.1 % and vascular complications were WLT, 16.5 %; SLT, 23.8 %; RLT, 23.5 %; LDLT, 24.4 % [33].

Trying to understand and balance the increased opportunity of being transplanted against the higher incidence of complications and potential graft loss associated with SLT is difficult. Merion et al. tried to address this issue by comparing the predicted lifetimes for SLT for an adult and a child recipient with WLT for an adult to determine the best use of this limited supply of organs [34]. They analyzed mortality risk for 48,888 patients on the waiting list: 907 SLT and 21,913 WLT recipients (between 1995 and 2002). Of 23,996 donor livers used for transplantation, 533 were split. Donors aged 10–39 years comprised 81.6 % of split livers and 48.5 % of livers used for WLT. Only 11.8 % of livers that were split were from donors older than 40 years. They analyzed years gained per SLT performed against the waiting list death ratio and concluded that “the potential annual net gain in life years could be as high as 169 patient-years in the first 2 post-transplant years, if all livers meeting accepted criteria were used for splitting.” For every 100 donor livers, an extra 11 years of life were predicted over the first 2 years of follow-up if organs were split rather than transplanted whole. In addition, they identified a significant survival benefit for pediatric recipients of SLT compared with children continuing on the waiting list and concluded that it could provide enough organs to satisfy the entire current demand for pediatric donor livers.

LDLT has become an important source of grafts for children worldwide. Following on from the first description by Raia et al. in 1988 [35], the first long-term survivor from Strong et al. in 1989 [36], and the early series published by Broelsch et al. [37], the technique has become established with excellent short- and long-term survival. The LLS is the most commonly used graft [38]. Full left or right lobe grafts are used less commonly and tend to be in young adults. Donation is most commonly from parents, although other family members and altruistic donation have been well reported. The ethics and understanding of the desire to donate are easily understood, and the risks appear to be low, but not negligible. Donor mortality for LSS grafts is of the order of 1:1500. The incidence of other significant donor complications which include bile leak (1–2 %), bleeding and the need for transfusion (1–2 %), deep vein thrombosis and pulmonary embolus, and incisional hernia (5 %) are considered acceptable. Risks of donor mortality are similar for left liver and rise to 0.5 % for right lobe donation. The risks of bile leak are also higher for the donor, particularly with the right lobe grafts where the incidence may be as high as 5 %. Donor age above 55 years has become generally accepted as a contraindication (especially for the right lobe) due to the slower regeneration and the increased risk in this population .

The LLS segment graft accounts for approximately 20–25 % of the adult liver, but provides a full-sized “liver” for the child. Assessment of the donor is designed to ensure that the liver segment is anatomically and functionally suitable for transplantation and to identify additional risk factors such as procoagulant abnormalities, smoking, and steatosis within the liver. Many centers continue to follow the two-stage assessment and consent model reported by Brolesch et al. which includes psychiatric/psychological assessment of the donor and family [37]. Depending on the availability of cadaveric transplantation, suitability for living donation will vary from 50 to 90 % when assessing families. The techniques of surgery have been standardized and outcomes are excellent for all liver diseases. The advantages of living related LT (LRLT) include the ability to perform surgery electively with an excellent quality graft, dry cut surface (minimal blood loss), and excellent outcome. The incidence of early rejection is similar to that of cadaveric transplantation, but long term it is considered, although the supporting evidence is limited, that these recipients may be tolerant of their grafts and that immunosuppression (IS) withdrawal may be more likely .

Morioka et al. reported the long-term outcome of LRLT in 46 children who underwent LDLT for metabolic disorders [39]. Mean age at diagnosis and LDLT was 48.6 (0–196) and 86.5 (1.4–199) months, respectively with survival rates of 86.9 and 81.2 % at 1 year and 5 and 10 years post-transplant. Patient survival was significantly better in children with liver-centered disease which included Wilson disease, ornithine transcarbamylase deficiency, tyrosinemia type 1, Crigler–Najjar syndrome type 1, and bile acid synthetic defect than in those with non-liver centered disease (glycogen storage disease, propionic acidemia, methylmalonic acidemia, and erythropoietic protoporphyria; p = 0.003). Statistical analysis showed that cumulative survival of patients with normal or slightly delayed physical growth at the time of LDLT was significantly better than for those with delayed physical growth ( p = 0.012).

LDLT is an excellent option in the management of ALF in children. The workup can be performed rapidly if the unit is regularly performing elective LDLT. Ethical concern has been expressed regarding the ability of the donor to appreciate the risk in these circumstances [40]. The lower risk of LLS donation has led to the use of LDLT as an accepted therapy. Greater debate surrounds the use of right lobe living donation for ALF particularly in countries with effective cadaveric organ donation .

Over the past 30 years, surgical techniques have become standardized worldwide. LLS and left and right liver LT are used to overcome size discrepancy and to engraft the majority of children. Split and LRLT have been key to sustaining the reduction in deaths on the waiting list. Auxiliary LT offers scope for native liver regeneration in children presenting with ALF and is able to withdraw from IS in the majority. Hepatocyte transplants remain experimental and have been used to bridge young children with metabolic liver disease through the liver replacement. The current challenge is to minimize technical complications that impact on graft survival, such as hepatic artery thrombosis and biliary strictures, and continue to improve outcomes. New technologies are emerging that are currently under evaluation and will impact on practice, and normothermic organ perfusion is perhaps the most prominent.

Over the 30 years, pediatric LT has seen serial improvements in morbidity and mortality. These improvements are related to IS, donor selection and maintenance, surgical technique, and complex team working. The focus of IS has been on the prevention of acute rejection and graft loss; however, in children, morbidity and mortality rates from infections exceed those from rejection and can also impair growth and renal function and increase the risk of some cancers. Getting the balance of IS right in the short and long term is key to a successful LT program. Evidence is also accumulating that children may be more likely to develop graft tolerance and to wean from IS in the long term [50].

Early post-transplant IS has been associated with increasing use of induction therapy (26 %), particularly with interleukin (IL)-2 receptor antibodies either to supplement IS or as a renal sparing regimen. For maintenance IS, calcineurin inhibitors (CNI) remain the cornerstone; but cyclosporin has been largely replaced by tacrolimus. Children often receive an anti-proliferative agent such as azathioprine or mycophenolate mofetil to supplement IS or for CNI/renal sparing. Tacrolimus trough levels of 8–10 ng/l in the first 3 months and 5–8 ng/l thereafter are usually sufficient. Longer-term tacrolimus levels can be further weaned to low levels, if graft function is normal with no rejection episodes. Moderate-to-severe acute rejection is treated with steroid boluses and if unresponsive antithymocyte globulin can be given.

Over-IS is associated with long-term side effects including renal impairment, hypertension, lymphoproliferative disease, and cancer and may hinder the process of immune engagement and the development of immune tolerance. There has been a move toward progressive minimization of IS, including steroid withdrawal which is associated with a growth advantage without any significant rejection-related complications [51, 52]. Many units discontinue steroids during the first or second post-transplant year and more than 50 % of children are on monotherapy tacrolimus by 18 months post LT [53]. It has been suggested that early steroid withdrawal may be associated with a higher risk of acute and chronic rejection and the development of de novo AIH [54]. The pathogenic mechanisms behind de novo AIH remain unclear, but it appears to be related to under IS and affects 2–5 % of pediatric LT recipients [55, 56] .

CNI share a number of side effects in common including nephrotoxicity, neurotoxicity, and hyperlipidemia [57]. There are, however, some differences between tacrolimus and cyclosporin. Cyclosporin is associated with hirsutism and gum hyperplasia while tacrolimus is associated with diabetes and hair loss. For teenagers, especially in females, tacrolimus is often the CNI of choice because of the lower incidence of gum hyperplasia and hirsutism. Tacrolimus has greater water solubility and less dependence on bile salt absorption, thus resulting in improved bioavailability over cyclosporin. However, it has larger interindividual variations. The introduction of once-daily preparations of CNI may help reduce variations in drug levels and improve adherence to medication [58]. Tacrolimus may be superior to cationic steroid antibiotic (CsA) with regard to steroid withdrawal and the incidence of acute and chronic graft rejection. CNI sparing or substitution with mammalian target of rapamycin (mTOR) inhibitors such as sirolimus or everolimus is used for patients with nephrotoxicity, but their efficacy requires validation in long-term studies in large cohorts.