Widespread application of cadaveric split or living donor liver transplantation bears considerable potential to increase the pool of available organs and thus alleviate the problem of organ shortage. Although splitting of a cadaveric liver into two grafts for adult recipients can be performed successfully, sufficient function of undersized grafts is a major concern. To minimize the risk for living donors, transplant surgeons aim at procuring the least necessary liver volume, also leading to potentially small grafts. When small partial grafts are unable to meet the functional demands, the recipients can develop a so-called small-for-size syndrome (SFSS). There is currently limited data on the pathogenesis of SFSS, with clinical studies mainly focusing on portal hyperperfusion. Additional aspects include graft-related factors such as functional and regenerative capacity, as well as recipient-related factors, such as overall health status and severity of cirrhosis. However, there is currently no consensus on the definition of SFSS. We propose a novel definition, based on simple clinical criteria, which divides the syndrome into either nonfunction or dysfunction of a small graft after the exclusion of other causes. This definition should ease comparability of future clinical trials, and thus improve understanding of the pathogenesis of SFSS.
The liver can sustain vital functions even with a substantially reduced parenchymal mass. The techniques developed in the 1980s in reduced and split liver transplantation for children paved the way for partial orthotopic liver transplantation (OLT) in adults. Partial OLT currently includes the splitting of a cadaveric graft for two recipients or partial liver transplantation from a living donor. These strategies are gaining increasing acceptance worldwide to overcome the shortage of organs and mortality on the waiting list. However, a graft size of about 50% of standard liver volume (SLV) is necessary to secure a favorable outcome after transplantation. Such a graft volume is rarely available from a cadaveric donor to transplant two adult recipients, and procuring such graft from a living donor usually necessitates a right hemi-hepatectomy, which subjects a healthy individual to a significant operative risk.
The successful use of smaller grafts would be a revolution in transplantation due to the potential to dramatically alleviate mortality on the waiting list, as small partial grafts obtained in cadaveric donors could be routinely used for two or more adult recipients and living donor liver transplantation (LDLT) could be performed using, for example, only a bi-segmentectomy 2 and 3 graft. Such an operation is associated with minimal risk, possibly in the range of kidney donation, and might therefore be widely accepted by potential donors. Thus, a focus on small grafts is very timely.
A partial liver graft unable to meet the functional demands of the recipient results in liver failure including coagulopathy, ascites, prolonged cholestasis and encephalopathy, often associated with pulmonary and renal failure, and frequently leads to death of the recipient in the absence of re-transplantation (1). This ill-defined clinical picture is considered to be primarily linked to an insufficient graft size, and hence termed ‘small-for-size-syndrome’ (SFSS).
Deleterious effects of graft size mismatching might be one of the reasons for inferior graft survival in adult living donor transplantation compared to a matched population of cadaveric organ recipients (2). But factors in addition to graft size per se are obviously critical for the outcome of partial OLT. The goal of this minireview will be to propose a unifying definition of SFSS and to critically assess our current knowledge on the pathogenesis of this entity.
Definition of SFSS
The term SFSS has been used to describe a variety of clinical presentations ranging from mild hepatic dysfunction with isolated hyperbilirubinemia to irreversible graft failure leading to death of the patient in the absence of an available organ for re-transplantation. The differentiation between SFSS and other etiologies of graft dysfunction is not straightforward. Therefore, other causes of hepatic dysfunction after partial OLT might unintentionally be described as SFSS, making it difficult to compare the results of different studies. We suggest SFSS to be seen as a distinct and well-defined entity leading to liver dysfunction after partial OLT. Currently, there is no consensus about the definition and pathogenesis of SFSS. To this end, we conducted a survey of 20 expert surgeons in the field of partial liver transplantation from Europe (12), South America (2), North America (2), the Near East (2) and Asia (2). While SFSS was agreed to be a distinct disease entity, there was considerable disagreement on other aspects of SFSS (Table 1). For example, opinions were divided whether outflow obstruction or arterial hypoperfusion are a part of SFSS. There were also widely differing responses concerning the signs necessary to diagnose SFSS. While most surgeons considered prolonged hyperbilirubinemia as a prerequisite to diagnose SFSS, responses regarding other signs such as encephalopathy or elevated transaminases were conflicting (Figure 1). Consequently, a consensus is needed to aid the comparability of future studies, thereby improving our understanding of the pathogenesis of SFSS. There is no prospective study allowing an evidence-based definition of SFSS, and no accepted definition of hepatic dysfunction after resection or transplantation (3,4). Therefore, cutoff values have to be arbitrarily chosen. We aim to provide a working definition of SFSS, which needs to be evaluated and possibly modified in the future. On the basis of the current knowledge, we propose a definition of SFSS, which relies on simple and objective functional criteria (Table 2). In analogy to whole liver transplantation, we propose to divide SFSS into small-for-size dysfunction (SFSD) and small-for-size nonfunction (SFSNF). Before a diagnosis of SFSS can be made, other reasons for graft dysfunction, particularly technical problems such as impaired outflow, must be excluded.
Table 1. Responses of senior transplant surgeons (n = 20) to the survey on SFSS*
*Some respondents did not answer every question.
SFSS = small-for-size syndrome.
Do you consider SFSS a distinct disease entity?
Is portal hyperperfusion a part of SFSS?
Is outflow obstruction a part of SFSS?
Is arterial hypoperfusion a part of SFSS?
Would you accept a diagnosis of SFSS if the patient recovers spontaneously?
Can SFSS occur later than 1 week postoperatively?
Table 2. Proposed definition of small-for-size syndrome (SFSS)
*Graft dysfunction = the presence of two of the following on three consecutive days: bilirubin >100 μmol/l, INR > 2, encephalopathy grade 3 or 4.
**Graft failure = re-transplantation or death of recipient.
***Exclusion criteria: technical (e.g. arterial or portal occlusion, outflow congestion, bile leak), immunological (e.g. rejection), infectious (e.g. cholangitis, sepsis).
Small-for-size dysfunction (SFSD)
Dysfunction* of a ‘small’ partial liver graft (GRWR < 0.8%) during the first postoperative week after the exclusion of other causes***.
Small-for-size non-function (SFSNF)
Failure** of a ‘small’ partial liver graft (GRWR < 0.8%) during the first postoperative week after the exclusion of other causes***.
How to study SFSS
Most knowledge about SFSS is derived from analyses of patient series. Human studies are hampered by the relatively low number of cases available, the heterogeneous patient population and by ethical limitations concerning possible interventions. These problems can be circumvented by the use of animal models. The relationship between graft size and outcome was first described in dogs (5). Today, models of partial liver transplantation have been used to study SFSS in pigs (6), rats (7) and mice (8). The mouse model of partial transplantation, although technically demanding, has important advantages: first, it is physiologically more comparable to humans with its hepatic arterial dependency after transplantation (9), and secondly, a large selection of genetically altered mouse strains is available allowing more detailed mechanistic studies. An important factor never incorporated into any animal model is the recipient's disease state. While most patients receiving partial liver grafts suffer from end-stage liver disease (ESLD) with all accompanying changes, until now, all experimental studies have utilized healthy recipient animals. This somewhat restricts transferability of experimental results to the clinical setting. Therefore, it might be of interest to develop models for particular recipient-related aspects such as portal hypertension.
The discussion of pathogenesis will be divided into graft-related and recipient-related factors, covering both clinical as well as experimental data (Table 3).
Liver size and function: In the absence of additional stressors, the healthy liver has a large functional reserve. Although the exact limit of viability is difficult to assess in humans, a residual liver volume of less than 27% leads to an increased incidence of severe postoperative hepatic dysfunction after resections as defined by biochemical and clinical markers (4). In rats, a remnant liver of 10% can be enough for survival (10). In the setting of transplantation, however, a higher liver volume is required. Living donor grafts of less than 40–50% of SLV, corresponding to a graft to recipient weight ratio (GRWR) of 0.8–1.0%, are associated with worse outcome (11). These thresholds are less well-defined for cadaveric split liver transplantation, but presumably, an even larger graft volume is necessary due to additional factors such as brain death and longer preservation injury. However, the issue of sufficient graft volume is still a matter of debate.
An additional factor that influences the functional capacity of the partial graft is the presence of underlying liver disease, such as steatosis. Hepatic steatosis is an increasingly prevalent condition and it is a risk factor for complications after major hepatectomy. This must be kept in mind when considering living donor hepatectomy. Most centers avoid splitting or living donation of steatotic livers with varying cutoffs regarding the degree of steatosis. It appears that living donor grafts with up to 30% of steatosis do not have a worse outcome, and steatosis is decreased in protocol biopsies 10 days after transplantation (12). Nonetheless, steatosis is thought to reduce the effective liver mass for transplantation. In rats this is reflected by the larger partial grafts needed for survival (13). Although the precise influence of steatosis on graft functional capacity is not known, it is prudent to aim for a larger volume when using such a graft.
Regeneration: Partial liver grafts and remnant donor livers undergo a rapid regenerative response with the largest changes in liver volume occurring during the first week (14). Three months after the procedure, the liver volume slightly exceeds 100% of the SLV in recipients but reaches only about 80% of the SLV in donors (15). The mechanism of this effect is unclear, but some insight can be gained from the comparison of different recipient groups: cirrhotic patients receiving a partial graft show faster regeneration than noncirrhotics, which could be due to a higher portal flow (16), higher portal pressure (17) or a more pro-proliferogenic environment in patients with chronic hepatic insufficiency. The inter-dependence of portal flow and regeneration is illustrated by impaired regeneration after portosystemic shunting in rats (18).
Whether the immediate functional capacity of a partial liver or the ability to rapidly regenerate is more important to sustain the recipient is unknown. It is commonly claimed that hepatocytes can replicate and fulfill their functions at the same time, as almost all hepatocytes enter into the replicative cell cycle after major hepatectomy. But there is a lower limit of remnant liver where regeneration fails, an effect supported by more recent work in small-for-size mouse liver transplantation (8). Furthermore, warm ischemia (19) as well as cold ischemia (7) have been shown to impair regeneration after partial liver transplantation by interfering with the production of interleukin-6 (IL-6) and tumor necrosis factor α (TNFα) and/or defective progression from the IL-6 signaling pathway to cellular proliferation (20). This could be important for split grafts from cadaveric donors where longer cold ischemia occurs.
Vascular inflow: Portal venous blood flow undergoes very little autoregulation within the liver, while arterial flow does. Changes in portal flow induce reciprocal effects on arterial flow by the so-called hepatic arterial buffer response (21). This buffer response is active even after human liver transplantation, when supra-physiologic levels of portal flow and velocity occur, causing a compensatory decrease in arterial inflow (22). Vice versa, a temporary reduction of portal venous inflow to human liver grafts has been shown to increase arterial flow (23). The relative portal hyperperfusion, even in the setting of complete liver grafts, is only partially due to hyperdynamic splanchnic circulation in ESLD (22). Extensive hepatectomy in dogs leads to higher portal pressure and higher sinusoidal resistance as the portal flow is directed through a small remnant liver (24). Similarly, after partial liver transplantation the graft is subjected to the portal flow destined for a whole liver, so post-transplant hyperperfusion is more pronounced than in whole liver grafts. Absolute and relative portal venous flow increase after human and porcine partial liver transplantation, while arterial flow decreases (25,26). Small grafts have a higher portal resistance, leading to higher portal pressure in patients (27). In this setting of partial transplantation, recipient hemodynamics are also important, as cirrhotic recipients demonstrate higher portal flow than noncirrhotics (16).
There is evidence implicating portal hyperperfusion in the pathogenesis of graft injury in small-for-size transplantation. As early as 5 min after reperfusion sinusoidal congestion and hemorrhage are evident in 20% partial liver grafts in pigs (6). The severity of these changes is inversely related to graft size, being less pronounced in 30% and 60% and absent in full liver grafts. Electron microscopy of 30% grafts in rats shows the same morphological hallmarks, accompanied by gaps in the sinusoidal endothelial lining and mitochondrial swelling as well as vacuolar changes in hepatocytes (28). After 24 h, the space of Dissé is collapsed and the endothelial lining disrupted (28).
Most clinical studies have focused on strategies to reduce portal venous inflow. In a retrospective analysis of 13 LDLT patients, splenic artery ligation, combined with portal banding and portosystemic shunting in some cases, were effective in reducing portal flow, and possibly led to improved post-transplant outcome (29). Reduction of portal venous inflow was accompanied by an increase in hepatic arterial flow. The same group recently reported improved patient and graft survival by using a portocaval shunt in small living donor grafts with a GRWR less than 0.8% (30). A possible efficacy of portal venous inflow reduction is underscored by the use of porto-mesenteric disconnection and meso-caval shunting in porcine small-for-size partial liver transplantation, which was associated with an increase in animal survival (31). We would speculate that a beneficial effect of portal venous inflow reduction stems from the increase in arterial flow. However, it is important to bear in mind that portal flow can only be reduced to a certain degree, as regeneration is also dependent on portal flow (18).
Other aspects: Most analyses of partial liver transplantation have focused on technical aspects, such as vascular in- and outflow. A small number of descriptive molecular studies have been performed in the field. Analyses of 30% liver transplants in rats showed the relative changes in the expression of genes involved in the regulation of hemodynamics (endothelin, endothelin receptor, iNOS), regeneration (IL-6, TNFα) and stress response (HO-1, Hsp-70) (32) without a clear picture emerging yet. A striking finding in small-for-size transplants in mice was the development of massive microvesicular changes in hepatocytes after 2 days (8). These changes, along with the failure to regenerate, indicate severe metabolic derangement and energy depletion in these grafts. Another factor to consider is the immunogenicity of partial regenerating grafts. Some experimental data from rat allotransplantation point to a stronger immune response against partial grafts than whole livers (33), while rejection rates in patients series are often lower in living donor transplantation, presumably due to a high percentage of relatives as donors.
Pre-transplantation status of the recipient is a critical determinant of outcome. Patients suffering from ESLD have a reduced functional reserve in different organ systems. Portal hypertension and hyperdynamic splanchnic circulation are typical hallmarks in these patients, which could potentially aggravate portal hyperperfusion after partial liver transplantation. Survival after transplantation of a small partial graft (GRWR < 0.85%) has been shown to be lower for Child B and C recipients compared to Child A, while the difference was less pronounced with larger grafts (34). In another study, SFSS developed much more frequently in cirrhotic recipients of small grafts (mean GRWR 0.79%) than in noncirrhotics (35). Importantly, the safe threshold for liver volume was higher in cirrhotics. The safety of transplanting a small partial graft seems to be critically dependent on severity of the recipient's disease. Therefore, higher thresholds of critical liver volume must be considered for sicker patients, although the precise cutoffs are not known.
The reconstruction of the vascular outflow tract of partial liver grafts has received considerable attention in the past, especially in the setting of right liver grafts with undrained segments. If no primary reconstruction of the middle hepatic vein or veins from segments 5 and 8 is performed, most right liver grafts show morphological signs of congestion (36), as well as impaired liver regeneration in the affected segments (37). There is an ongoing debate about the optimal technique to prevent outflow obstruction in partial liver grafts, which is beyond the scope of this minireview (38). Although congestion of anterior segments is a problem specifically related to right partial liver transplantation and is particularly poorly tolerated by small grafts, it is in essence a technical problem. Therefore, outflow obstruction per se should be considered as a separate entity, and needs to be excluded before the diagnosis of SFSS can be made (Table 2).
Interventions and therapeutic strategies
Few approaches to improve outcome after partial liver grafting have been tested. As detailed above, most available clinical trials have focused on reduction of portal venous pressure and flow. Three studies in rats have targeted vascular regulation in partial grafts: treatment with the nitric oxide donor FK409 was able to ameliorate graft injury (39), while an endothelin-receptor-A antagonist had effects on microcirculation, biochemistry and histology, but did not improve survival (40). Treatment with prostaglandin E1 improved survival in partial steatotic grafts in rats (13). An approach to improve the intra-graft stress response was chosen by adenoviral overexpression of heme oxygenase (HO-1), which effectively improved survival in allografts but not in isografts, speaking against a crucial role of HO-1 in protecting a small graft in the early post-reperfusion period (41). Regeneration after rat partial liver transplantation could be improved by overexpression of redox factor 1 (42), by treatment with hepatocyte growth factor (43), or by inhibition of nuclear factor kappa B activation (44). Data from our group show pentoxifylline to improve sinusoidal perfusion and survival in a model of mouse small-for-size transplantation.
In conclusion, SFSS represents a distinct disease entity related to partial liver transplantation. Many clinical studies have been undertaken in the field of small partial liver grafts, but comparisons and interpretation of results are hampered by differing uses of the term SFSS. To approach this problem, we propose a novel definition of SFSS (Table 2). Although retrospective in nature, the clinical studies performed to date have allowed us to recognize the disease entity of SFSS and to formulate hypotheses regarding the pathogenetic mechanisms underlying critical graft size, function, injury, regeneration and recovery. Most clinical studies have focused on the reduction of portal venous inflow to small grafts, thereby strongly implicating portal hyperperfusion as a cause of SFSS.
More recently, partial liver transplantation has been the focus of experimental studies in animals, allowing a first glimpse into the underlying molecular mechanisms. The use of large and small animal models allows detailed molecular and mechanistic studies. Furthermore, different interventions can be tested under controlled conditions. An exciting venue for future experimental models might be the study of pathogenetically important mechanisms in isolation. For example, portal hyperperfusion is implicated in SFSS, yet little is known about the role of hyperperfusion alone. This could be studied in ex vivo models such as the isolated perfused rat liver.
The holy grail of SFSS is the development of strategies to protect small grafts. This would allow smaller graft volumes to be transplanted, expanding the safety limits of cadaveric split liver transplantation and rendering the living donor procedure safer. This could lead to a wider application of split and LDLT. One could speculate that it may solve the problem of organ shortage and death on the waiting list.