Essential Updates 2024/2025: History, Pathogenesis, Definition, Prevention, and Management of Small‐for‐Size Syndrome in Living‐Donor Liver Transplantation

ABSTRACT

Although living‐donor liver transplantation (LDLT) has become the standard treatment for end‐stage liver disease, one of its major challenges is small‐for‐size syndrome (SFSS). SFSS is characterized by severe icterus and intractable ascites, although the severity of the condition can vary. Some cases are managed with medical treatment alone, others require interventions such as splenic embolization, and some may result in graft loss or necessitate re‐transplantation. A recent area of interest in this field is the new grading system introduced through collaboration between the International Liver Transplantation Society, the International Living‐Donor Liver Transplantation Study Group, and the Liver Transplant Society of India in 2003. This grading system has helped define SFSS in the context of LDLT. Recent trends also include right lobe graft selection with V5/V8 reconstruction and optimal outflow to manage the high venous pressure associated with end‐stage liver disease. To develop effective strategies for transplanting small‐for‐size grafts and preventing SFSS, it is crucial to have a comprehensive understanding of how to evaluate graft quality and volume, alongside portal pressure management, during LDLT. We review the latest literature on the pathogenesis of SFSS and the strategies for overcoming it after LDLT.

Short abstract

Small for size syndrome is among the serious issues after living‐donor liver transplantation. In this review article, History, pathogenesis, definition, prevention, and management of small‐for‐size syndrome in living‐donor liver transplantation.

Abbreviations

CT

computed tomography

EAD

early allograft dysfunction

GRWR

graft‐to‐recipient weight ratio

GV

graft volume

INR

international normalized ratio

LDLT

living‐donor liver transplantation

MELD

Model for End‐Stage Liver Disease

PGD

primary graft dysfunction

SAL

splenic artery ligation

SFSG

small‐for‐size graft

SFSS

small‐for‐size syndrome

SLV

standard liver volume

1. Introduction

Living‐donor liver transplantation (LDLT) was introduced for pediatric patients in Brazil in 1989 [1], and the positive outcomes gradually established LDLT as a treatment of choice for pediatric patients with end‐stage liver disease [2]. In 1993, Kawasaki et al. [3] reported in their pediatric LDLT series that a graft volume‐to‐standard liver volume ratio (GV/SLV) of 46% was within the safe limit. They were the first to use the term small‐for‐size graft (SFSG), referring to grafts with a GV/SLV of < 1.0. In 1994, the same group performed the first adult LDLT using a left lobe graft with a GV/SLV of 45%, transplanted into a small female recipient with primary biliary cirrhosis [4].

Emond et al. [5] reviewed their pediatric LDLTs using lateral segment grafts (n = 23) and adult LDLTs using left lobe grafts (n = 2) and found that five recipients who received an SFSG (defined as < 50% of the expected total liver size) showed delayed functional recovery. Additionally, a patient who received a graft comprising only 23% of the expected liver volume developed graft nonfunction, representing a case of small‐for‐size syndrome (SFSS). The authors reported that SFSGs comprising < 50% of the expected total liver volume exhibited characteristic pathological features—ischemia, cholestasis, and regeneration—after LDLT. In response to these findings, Fan et al. [6] advocated for the use of right lobe grafts with the middle hepatic vein in adult LDLT to prevent SFSS, demonstrating the safety of harvesting extended right lobe grafts from donors.

2. Pathogenesis of SFSS

LDLT initially evolved from procedures involving adult donors and pediatric recipients, primarily to avoid graft size mismatch. However, as the indication expanded to include adult‐to‐adult transplantation, various adverse clinical outcomes emerged, including SFSS [5]. The basic pathogenesis of SFSS stems from a temporal gap between hepatocyte regeneration and the regeneration of endothelial cells and bile ducts. In a murine model [7] developed nearly 30 years ago, bile duct and endothelial regeneration lagged several days behind hepatocyte regeneration. As a result, bile produced by the regenerating hepatocytes cannot be adequately excreted into the intestine. Concurrently, the increased portal flow following LDLT contributes to the development of increased or intractable ascites due to a combination of multiple factors (Figure 1). Interestingly, restoration of bile duct and endothelial regeneration has been shown to ameliorate the clinical symptoms of SFSS.

FIGURE 1.

Multiple factors, including a small graft and various negative clinical insults, can contribute to early allograft dysfunction. In the context of living‐donor liver transplantation, the use of a small graft often leads to what is referred to as “small‐for‐size graft syndrome.”

3. Consequences of SFSS

The problems related to the use of SFSGs and the development of SFSS were first reported by Emond et al. [5], with further clarification provided in subsequent studies [8, 9, 10, 11, 12]. SFSS can arise in the context of various adverse clinical factors, including an extra‐small graft, poor recipient condition, transplant‐related infection, and inadequate surgical technique. Among these, a small graft relative to the recipient size remains the most critical factor in the onset of SFSS. Clinically, the syndrome is marked by severe icterus and increased ascites, typically manifesting about 1 week after LDLT—around the time the transplanted liver enters the regeneration phase [12]. The hallmark histopathologic findings include hepatocyte ballooning and cholestasis in the central venous zone, often with a distinct demarcation line that may reflect microcirculatory disturbances. Ikegami et al. [13] found that more severe SFSS, indicated by a higher serum bilirubin peak, correlated with a shift of this boundary from the central venous to the periportal zone.

Although the exact minimum residual liver volume necessary for survival after hepatic resection remains unclear, clinical analyses suggest that volumes below 25%–27% are associated with significant morbidity and mortality. Nonetheless, there are several reports of successful SFSG use without additional surgical intervention: a GRWR of 0.59% by Kiuchi et al. [12], a GV/SLV of 25% by Lo et al. [14], and 26% by Nishizaki et al. [15]. Other cases have demonstrated successful transplantation of extra‐small grafts when accompanied by surgical interventions, such as a 20% GV/SLV with a portocaval shunt by Masetti et al. [16], and a 34% GRWR with splenectomy by Lauro et al. [17].

These findings suggest that an SFSG with a 25%–30% GV/SLV may be feasible with meticulous pre‐, intra‐, and postoperative care, while grafts between 20% and 25% may require adjunctive surgical strategies. However, outcomes are not determined by GV alone; graft quality, recipient condition, and technical factors also play critical roles. For this reason, defining a strict minimum liver volume threshold must account for these multiple variables. Still, identifying “pure” SFSS—excluding other causes—remains challenging. Patients with SFSS are at higher risk of septic complications and renal and/or pulmonary insufficiency.

4. Definition of SFSS

Since the first observations of SFSS following LDLT around 30 years ago, its main clinical features have included temporary moderate icterus and intractable ascites, often associated with acute graft failure. These symptoms reflect the balance between the graft's functional reserve and the cumulative insults it endures [12, 13]. Over the past three decades, numerous authors have proposed varying definitions of SFSS. The earliest definition was suggested by Dahm et al. [18]: 3 consecutive days of a bilirubin level of > 6.7 mg/dL, international normalized ratio (INR) of > 2, and grade 3 or 4 encephalopathy. However, this definition was not based on mortality data, and the clinical picture closely resembled primary nonfunction immediately after deceased‐donor liver transplantation. A clearer, though still non‐data‐driven, definition was later developed during an expert meeting. SFSS was divided into small‐for‐size dysfunction, defined as graft dysfunction in the first postoperative week after exclusion of other causes and involving a “small” partial liver graft (GRWR of < 0.8%), and small‐for‐size nonfunction reported under similar conditions. Graft dysfunction was defined according to previously described criteria. The first objective definition of SFSS was proposed by Soejima et al. [19] as a total bilirubin level of > 5 mg/dL and daily ascites output of > 1 L on postoperative Day 14. This was later revised to define SFSS as prolonged functional cholestasis with a total bilirubin level of > 10 mg/dL and daily ascites output of > 1 L on postoperative Day 14, in cases where the graft survival rate remained at 90%. Hill et al. [11] defined SFSS as significant cholestasis (bilirubin level of > 10 mg/dL after postoperative Day 7), coagulopathy (INR of > 1.5), and ascites output of > 2 L/day in the absence of technical issues. Subsequent studies recognized that GV alone does not fully explain post‐LDLT functional cholestasis. Other important factors include graft quality, donor age, severity of the recipient's disease, and surgical insults, all of which influence graft function after LDLT.

Because many LDLT centers now use right lobe grafts in adult recipients and set 0.8 as the lower limit of the GRWR, GV has ceased to be the main determinant of graft dysfunction with mortality in the 21st century [20]. As a result, the term “SFSS,” with its emphasis on GV, may no longer adequately represent the underlying disease process. In such contexts, the term primary graft dysfunction (PGD) has been proposed, as advocated by Ikegami et al. [13], to describe nonsurgical, multifactorial graft dysfunction following LDLT. They showed that factors such as a high Model for End‐Stage Liver Disease (MELD) score, presence of a portosystemic shunt, older donor age, elevated portal venous pressure, and operative blood loss were more closely associated with PGD than GV. Olthoff et al. [21] introduced the concept of early allograft dysfunction (EAD) in LDLT, defined by a bilirubin level of > 10 mg/dL or INR of > 1.6 on postoperative Day 7, in the absence of technical complications. The subsequent Adult‐to‐Adult Living‐Donor Liver Transplantation Cohort Study (A2ALL) showed that EAD—associated with smaller GV, higher preoperative bilirubin, elevated portal venous pressure, older donor age, and higher donor body mass index—was linked to a five‐fold increase in short‐term mortality [22]. The Kyoto group later re‐evaluated the EAD criteria [23], demonstrating that graft survival was significantly lower when both total a bilirubin level of > 10 mg/dL and INR of > 1.6 were present on Day 7 after LDLT, despite the fact that a GRWR of < 0.8 had no negative effect on graft survival. While SFSS remains a widely used term because of the inherently small grafts in LDLT, EAD more comprehensively encompasses the spectrum of SFSS‐related disease processes.

For determining the need for re‐transplantation after LDLT, the early graft mortality rate associated with each definition of SFSS should be taken into account. Reported early graft mortality rates were 26% for the revised SFSS definition by Soejima et al. [19], 9% for the SFSS definition by Hill et al. [11], 59% for PGD by Ikegami et al. [13], 24% for EAD by the A2ALL study [23], and 59% for EAD by Okamura et al. [23] (Table 1). The two definitions with the highest mortality rates—PGD and EAD by Okamura et al. [23], both at 59%—likely reflect similar levels of graft deterioration, typically beginning with moderate hyperbilirubinemia and coagulopathy by postoperative Day 7, followed by progression to severe hyperbilirubinemia. Notably, PGD combined with coagulopathy (INR of > 2) was associated with a graft mortality rate of 77% [13]. In 2003, the consensus conference of the International Liver Transplantation Society, International Living‐Donor Liver Transplantation Study Group, and Liver Transplant Society of India brought together a majority of LDLT experts, including surgeons, hepatologists, radiologists, pathologists, and social workers. Their discussions focused on graft outcomes (Table 2) [24, 26]. According to their grading, Grade A SFSS had a graft mortality rate of 9%, Grade B SFSS (portal hypertensive phase) had a graft mortality rate of 9%–26%, and Grade C SFSS (liver failure phase) had a graft mortality rate of 59%–77%. However, a key limitation of this grading system is that classification occurs 1–2 weeks after LDLT. Consequently, if a graft progresses to Grade C SFSS, the window for arranging and realizing re‐transplantation—whether from a deceased or living donor—is critically short. Ideally, as in deceased‐donor liver transplantation, decision‐making parameters for re‐transplantation should be established within the first 3 postoperative days. The treatment of SFSS is primarily medical and includes interventions such as albumin‐infused fluid administration to manage massive ascites, transfusions to support renal perfusion, and plasma exchange in cases of extreme hyperbilirubinemia (> 30 mg/dL).

TABLE 1.

Various definitions for SFSS.

Year	Authors	Term	Definition of SFSS	Mortality
2005	Dahm et al.	SFSS	Two of the following three conditions: T.Bil > 5.9 mg/dL, INR > 2, or grade 3 or 4 encephalopathy	Not available
2006	Soejima et al.	SFSS	T.Bil > 10 mg/dL on Day 14, ascites > 1 L on Day 14	26%
2009	Hill et al.	SFSS	T.Bil > 10 mg/dL after Day 7, INR > 1.5 or ascites 2 L	9%
2012	Ikegami et al.	PGD	T.Bil 20 mg/dL after Day 7. INR > 2	77%
2016	Pomposelli et al.	EAD	T.Bil 10 mg/dL, or INR > 1.6 on Day 7	24%
2018	Okamura et al.	EAD	T.Bil 10 mg/dL, and INR > 1.6 on Day 7	59%

Abbreviations: EAD, early allograft dysfunction; INR, international normalized ratio; PGD, primary graft dysfunction; SFSS, small‐for‐size syndrome; T.Bil, total bilirubin.

TABLE 2.

Grading system of SFSS after LDLT.

Severity grades	Day 7	Day 14	Graft loss (%)	Recommended treatment
Grade A (Pre‐SFSS)	T.Bil 5 mg/dL	T.Bil 5 mg/dL or ascites 1 L/day	< 9%	Medical care
Grade B (Portal hypertensive phase)	T.Bil > 10 mg/dL or INR > 1.6	T.Bil 10 mg/dL and ascites 1 L/day	9%–26%	Medica care/Inflow modulation if not done yet
Grade C (Liver failure phase)	T.Bil 10 mg/dL and INR > 1.6	T.Bil > 20 mg/dL	59%–77%	Inflow modulation Possible re‐transplantation

Abbreviations: LDLT, living‐donor liver transplantation; SFSS, small‐for‐size syndrome; T.Bill, total bilirubin [24, 25].

5. Graft Shrinkage and SFSS

The fundamental approach to preventing SFSS is to ensure appropriate graft size relative to the recipient. In general, a GRWR of > 0.8 combined with reconstruction of the major middle hepatic venous branches is widely accepted as the standard for graft selection to minimize the risk of SFSS. However, multiple factors can influence the clinical course of the graft, making this decision complex. When considering GV, it is important to recognize the potential discrepancy between expected and actual GV. Urata et al. [25] reported that the conversion ratio between liver weight (g) and liver volume (cm³) is not 1:1 but rather 1.12 g/mL, introducing a possible 12% margin of error. Hiroshige et al. [27], using a small‐animal model, found that grafts began losing weight during preservation—approximately 4% within 15 min—due to the high osmolarity of the preservation solution. Radtke et al. [28] observed that liver volume increased by up to 7% during the venous phase of contrast‐enhanced computed tomography (CT) scans compared with the plain phase. Separately, volume was found to shrink by as much as 11% during procurement [29]. Because three‐dimensional CT volumetry is typically performed using venous‐phase CT, the combined discrepancy between expected and actual volume can reach up to 18%. To address this, Yoneyama et al. [30] introduced coefficient factors of 0.84 for right lobe grafts and 0.86 for left lobe grafts to estimate actual GV more accurately from the expected GV. Interestingly, they found no discrepancy between actual and explanted cirrhotic liver volumes, with a coefficient factor of 1.01. This suggests that soft, healthy liver grafts with high elasticity are more prone to differences between predicted and actual weights/volumes. Kayashima et al. [29] further noted that grafts from younger donors showed greater shrinkage during preservation (14.0%) than did those from older donors (4.4%). This highlights the need for readers to recognize that many surgical reports reference actual GV when describing good graft function, even if the graft size is relatively small. Therefore, while the graft selection process is based on expected GV, the actual GV may be 15%–20% smaller by the time of transplantation. In practice, a transplant surgeon may find that a graft initially assessed to have adequate predicted GV presents with significantly reduced actual GV on the back table. As such, the definition of an SFSG should ideally be based on an expected GRWR of 0.8 or expected GV/SLV of 40%, rather than on actual values.

6. Prevention and Management of SFSS

The basic concept of EAD, including SFSS, is characterized by the absence of technical complications, as well as by previously described inflow and outflow issues [31, 32]. Major donor and recipient factors associated with EAD after LDLT have been well documented in the literature and include a GRWR of < 0.8, donor age of > 45 years, MELD score of > 20, recipient sarcopenia (e.g., inability to walk independently), post‐transplant portal pressure of > 15 mmHg, and post‐LDLT complications such as sepsis [33]. Given the challenges in managing recipients with end‐stage liver disease, the only viable preventive strategies for SFSS are careful donor and graft selection and the use of inflow modulation to control portal pressure.

Among inflow modulation techniques, splenic artery ligation is widely recognized as one of the safest and most effective options [33, 34, 35]. Its benefit lies in increasing hepatic arterial inflow through splenic artery obstruction and reducing portal inflow via decreased splenic venous return. The main risks associated with this procedure are potential bleeding and pancreatic injury because the splenic artery is located deep beneath the pancreas. We have performed simultaneous splenectomy during LDLT using vessel‐sealing and endo‐stapling devices, enabling tie‐less and bloodless splenectomy with greater portal decompression than splenic artery ligation [36, 37, 38]. The indications for splenectomy vary by center. Its rationale lies in the spleen's role as a major source of vasoconstrictive molecules (including endothelin), which contribute to portal hypertension. Splenectomy, by removing this source, leads to hepatic vasodilation [36, 37, 38, 39]. Ikegami et al. [40] generally perform splenic artery ligation first, prior to splenectomy, to allow easier and safer handling of the enlarged spleen in patients with end‐stage liver disease. However, splenectomy has not gained global acceptance as a standard inflow modulation procedure in LDLT because of risks such as bleeding, portal or splenic vein thrombosis, and post‐splenectomy sepsis [41]. Post‐transplant sepsis can be multifactorial but is mainly associated with graft dysfunction (Figure 1). At Kyushu University, splenectomy has been shown to significantly improve graft survival without increasing the rate of sepsis, likely because of good graft function [37]. To reduce the risk of overwhelming post‐splenectomy sepsis, especially from pneumococcal infections, pneumococcal vaccinations are administered every few years as a preventive measure. At Tokyo University, splenectomy is generally avoided [42] because of the high bleeding risk (possibly related to the lack of tie‐less techniques), which has led to increased re‐laparotomy rates and greater surgical stress on recipients. Poorly performed splenectomy may also result in pancreatic leakage, and post‐splenectomy sepsis may stem from such increased surgical insults. In Japan, three major LDLT centers differ in their indications and cutoff values for portal venous pressure when considering splenectomy. These differences are summarized in Table 3.

TABLE 3.

Indications for splenectomy at Tokyo, Kyoto, and Kyushu Universities.

Institute	Indication for splenectomy	Cutoff portal venous pressure	Institute
Tokyo	No splenectomy	Not applicable	Tokyo
Kyoto	Graft/spleen volume ratio < 0.7	15 mmHg	Kyoto
Kyushu	Left lobe: all cases Right lobe: > 20 mmHg	Left lobe: all cases Right lobe: > 20 mmHg	Kyushu

A recent study from the Asan Medical Center focused on the use of splenic devascularization in LDLT, demonstrating adequate portal decompression and significantly fewer complications compared with splenectomy [43]. However, because of the unstable portal inflow during liver regeneration after LDLT and the portal steal phenomenon, which can lead to graft mortality, no Japanese institutions have recently performed this procedure [40]. The creation of a portosystemic shunt for portal decompression was explored in the early to mid‐2000s, with considerable success in LDLT using extremely small grafts [41, 42, 43, 44, 45, 46, 47]. We reported a case in which we used an extra‐small graft with 36% of the expected GV/SLV and 23% of the actual GV/SLV. The graft was procured from a 20‐year‐old donor and transplanted into a patient with hepatocellular carcinoma. To modulate excessive portal inflow to the small graft, we performed an end‐to‐side portocaval shunt after splenectomy [45]. Four days post‐surgery, the graft showed hepatofugal portal flow, necessitating immediate re‐laparotomy for shunt ligation, which resulted in a favorable postoperative course. We later realized that this young, soft graft, with 36% of expected GV/SLV, could not accommodate enough portal flow because of the reduction to 23% actual GV/SLV, which was likely exacerbated by the patient's temporary dehydration.

In these discussions, the generally recommended strategy for graft selection and inflow modulation would be to use a modified right lobe graft with GRWR of > 0.8 and full venous drainage, along with splenic artery ligation. However, special techniques in LDLT should be reserved for experienced centers. Looking ahead, the use of smaller grafts, including the left lobe or posterior segment, and larger grafts such as dual grafts or extended right lobe grafts, should also be considered. Additionally, technically demanding portal inflow modulations, such as tie‐less splenectomy and the creation of portosystemic shunts, warrant further investigation. In recent years, there has been growing interest in LDLT centers in Western countries regarding the use of left lobe grafts [46, 47, 48, 49, 50]. Professor Hashimoto, the liver transplant director at the Cleveland Clinic, one of the largest centers for left lobe grafts in the United States, is originally from Kyushu University. He is well‐versed in the indications for left lobe grafts and the appropriate inflow modulation techniques needed to decompress portal pressure. For patients with a larger spleen/left lobe volume ratio, splenectomy is often indicated because this condition associated with increased portal venous pressure. A key tip for managing inflow in such cases is to remove the spleen before graft implantation, which can help prevent overwhelming portal inflow immediately after reperfusion. Recently, the Asan group reported splenic devascularization in LDLT, with enough portal decompression and many fewer complications, compared with splenectomy [43]. However, because of the unstable portal inflow because of liver regeneration in LDLT and portal steal phenomena leading to graft mortality, no Japanese institutes have performed this procedure recently [40]. Creation of a portosystemic shunt for portal decompression was reported in the early to middle 2000s with considerable outcomes in LDLT using extremely small grafts [40, 41, 42, 43, 44, 45, 46]. We reported a patient in whom we used an extra‐small graft with 36% for the expected GV/SLV and 23% for the actual GV/SLV. The graft was procured from a 20‐year‐old donor and transplanted into a patient with hepatocellular carcinoma. We created an end‐to‐side portocaval shut after splenectomy for modulating excessive portal inflow to the small graft [45]. The graft started to show hepatofugal portal flow 4 days after surgery followed by immediate re‐laparotomy for the shunt ligation, resulting in a good post‐operative course. We should have recognized this young and soft graft had 36% for the expected GV/SLV for accommodating enough portal flow, and 23% for GV/SLV that resulted from the patient presenting with (temporary) dehydration.

In these discussions, the generally recommended strategies in graft selection and inflow modulation might be to use a modified right lobe graft with GRWR > 0.8 and full venous drainage, with splenic artery ligation. However, special techniques in LDLT could be performed in experienced LDLT centers. With regard to future development, the use of smaller grafts including the left lobe or posterior segment, the use of much larger grafts including dual grafts or an extended right lobe graft should be considered. Furthermore, technically demanding portal inflow modulations including tie‐less splenectomy and creation of portosystemic shunts should be investigated.

Recently, there is an increasing interest in LDLT centers in western countries toward the use of left lobe grafts [47, 48, 49, 50]. Liver transplant director of the Cleveland clinic, most major left lobe use institute in US is Professor Hashimoto. He is originated from Kyushu University and know indication of left lobe grafts and the appropriate way of inflow modulation to decompress portal pressure. Performing splenectomy in case of larger spleen/left lobe volume ratio is associated with larger portal venous pressure. The tip of splenic inflow modulation the use of left lobe graft is taking the spleen out before graft implantation. This small tip might prevent overwhelming inflow immediate after repercussion [51].

Author Contributions

Toru Ikegami: conceptualization, data curation, funding acquisition, methodology, writing – original draft. Masashi Tsunematsu: methodology, software. Shinji Onda: methodology, data curation, investigation. Kenei Furukawa: conceptualization, methodology, investigation. Koichiro Haruki: conceptualization, investigation, data curation. Michinori Matsumoto: supervision, validation, conceptualization. Yoshihiro Shirai: resources, validation, investigation.

Conflicts of Interest

The authors declare no conflicts of interest.