Small for size syndrome following living donor and split liver transplantation

Abstract

The field of liver transplantation is limited by the availability of donor organs. The use of living donor and split cadaveric grafts is one potential method of expanding the donor pool. However, primary graft dysfunction can result from the use of partial livers despite the absence of other causes such as vascular obstruction or sepsis. This increasingly recognised phenomenon is termed “Small-for-size syndrome” (SFSS). Studies in animal models and humans have suggested portal hyperperfusion of the graft combined with poor venous outflow and reduced arterial flow might cause sinusoidal congestion and endothelial dysfunction. Graft related factors such as graft to recipient body weight ratio < 0.8, impaired venous outflow, steatosis > 30% and prolonged warm/cold ischemia time are positively predictive of SFSS. Donor related factors include deranged liver function tests and prolonged intensive care unit stay greater than five days. Child-Pugh grade C recipients are at relatively greater risk of developing SFSS. Surgical approaches to prevent SFSS fall into two categories: those targeting portal hyperperfusion by reducing inflow to the graft, including splenic artery modulation and portacaval shunts; and those aiming to relieve parenchymal congestion. This review aims to examine the controversial diagnosis of SFSS, including current strategies to predict and prevent its occurrence. We will also consider whether such interventions could jeopardize the graft by compromising regeneration.

INTRODUCTION

From the beginning, transplant surgery has always been limited by organ availability. Split cadaveric and living donor grafts have gained popularity in the past twenty years as initial technical hurdles were overcome. From 1988 to 2008 a total of 4103 split cadaveric and 3079 living donor transplantations were carried out in Europe[1] and over 75% of these have taken place in adult-to-adult cases. This practice has generated a phenomenon known as “Small-for-size-syndrome” (SFSS), where a small graft exhibits primary dysfunction. The name is misleading; the graft need not necessarily be small if it is steatotic or if the recipient has adverse risk factors such as existing portal hypertension and Child-Pugh grade C. This review explores the putative mechanism underlying SFSS, risk factors, prevention and treatments.

PATHOPHYSIOLOGY OF “SMALL-FOR-SIZE SYNDROME”

SFSS has become increasingly recognised in the last 20 years since partial liver grafts made the leap from paediatric to adult transplantation. SFSS in orthotopic liver transplantation describes a condition in which a small graft (graft to recipient weight ratio < 0.8) exhibits signs of primary graft dysfunction within the first post operative week in the absence of other diagnoses such as vascular obstruction, biliary leak, sepsis and immune rejection[2]. Coagulopathy, ascites and hyperbilirubinemia are typical manifestations. This definition derives from a survey of 20 expert partial liver transplant surgeons across the world[2]. Whilst all 20 consider SFSS to be a distinct clinical entity, opinion on its underlying pathological basis is very much divided. Most of them consider that portal hyperperfusion forms part of the syndrome but less consensus was obtained about the importance of the role played by outflow obstruction.

Portal hyperperfusion, venous congestion and arterial hypoperfusion, as well as simple insufficiency of liver mass, have all been suggested as contributory mechanisms for pathogenesis. Several case series of living donor liver transplants[3,4] have shown significant elevation of portal venous pressure in small grafts [graft weight to recipient body weight ratio (GWRW) < 0.8]. Differentially elevated portal venous pressure (PVP) in small grafts persisted for as long as fourteen days post operatively compared to non-SFSS grafts[5]. This is clinically significant as PVP > 20 mmHg is correlated with poorer graft survival (38% vs 85%) at six months[6].

Recently, rat, mouse and porcine models have given valuable insight into SFSS under controlled conditions[7-9]. A porcine model where recipients were transplanted with 19.3%-25.3% standard liver volume showed significant increases in portal venous flow, portal pressure and vascular resistance, along with reduced arterial flow. In addition, markers of endothelial and hepatocyte injury were markedly elevated compared to the whole graft group[8]. Another porcine model using a range of liver graft sizes demonstrated typical histopathology of SFSS. Congestion, hemorrhage, sinusoidal/endothelial damage, septal edema and architectural disruption can be seen as soon as five minutes post reperfusion and persist for up to five days post operatively in small 20%-30% grafts. These histological changes were more severe and prolonged in the smaller grafts[9]. These findings correlate with a human study showing sinusoidal endothelial disruption and focal hemorrhage dissecting into connective tissue, along with hepatic artery spasm[10]. However, regeneration rate was also markedly increased in SFSS grafts, a common theme across both human and animal studies, suggesting a degree of portal hyperperfusion may be necessary to induce liver regeneration. An important caveat of animal models in SFSS is that none have taken into account disease status in the recipient, a crucial consideration in human orthotopic liver transplants.

The role of arterial hypoperfusion in SFSS is less well studied as it is secondary to portal hyperperfusion. Low hepatic artery flow seen in post-transplant grafts was formerly thought to be related to diversion of blood through the splenic artery and was called “splenic artery steal syndrome” but is now considered to be due to a normal homeostatic mechanism termed hepatic arterial buffer response (HABR)[11]. The role of HABR is to maintain constant total blood flow to the liver and it is mediated by adenosine washout. Portal blood flow removes adenosine which has a local vasodilator effect on the arterial system[12,13]. However, in states of extreme portal hyperperfusion as seen in small-for-size grafts, an exaggerated HABR may contribute to ischemic injury[10,14]. In one porcine small-for-size model, an infusion of adenosine in 20% standard liver size grafts was able to inhibit HABR and significantly reduce graft injury as determined by histology[15].

PREDICTING SFSS

Is it possible to identify cases at greater risk of developing SFSS? Commonly cited pre-operative risk factors include GWRW < 0.8[16] or graft weight ratio less than 30%-40%[17], with small grafts at significantly greater risk of prolonged bilirubinemia and coagulopathy. The first study is notable as 88% of patients were pediatric recipients undergoing transplant post-Kasai procedure, not representative of the usual adult liver transplant population. However, more recent studies have suggested small graft size alone is insufficient to account for SFSS. One retrospective study of 107 patients[18] undergoing live donor (n = 76) and split cadaveric transplants (n = 31) found no significant difference in either incidence of SFSS or graft survival at one year between the GWRW < 0.8% group (n = 22) and > 0.8% (n = 85) group, although the author reported a significantly greater number of SFSS cases in the 0.8%-1.0% region. In another study on a series of 75 patients[19], no difference was observed in development of SFSS between those receiving grafts less than 40% standard liver volume (n = 26) compared to those that received more than 40% (n = 73). This discrepancy can be accounted for in several ways. Firstly, there is increasing recognition that factors such as graft steatosis, pre-existing portal hypertension, recipient Child-Pugh grade and venous congestion also contribute to SFSS. Secondly, retrospective studies may lack power to detect a significant difference as a much smaller proportion of patients receive small-for-size grafts. Thirdly, in the later studies, patients thought at greater risk of developing SFSS often receive prophylactic measures such as splenic artery ligation[18].

When small-for-size grafts are stratified by disease severity in the recipient, it has been shown that SFSS is more likely to occur in patients with Child-Pugh score B and C. One-year graft survival is also significantly lower if Child B and C patients receive grafts GWRW < 0.8% and this is in addition to the poor prognosis conferred by pre-operative disease severity[20]. However, Child-Pugh A recipients can safely receive grafts as small as GWRW < 0.6%. It has been suggested that pre-existing portal hypertension may exacerbate the hyperperfusion seen in SFSS.

STRATEGIES TO PREVENT SFSS: MODULATING INFLOW

Since portal hyperperfusion is thought to be central to the pathogenesis of SFSS, the most popular strategies for prevention have focused on modulating inflow to the liver via inputs to the portal system. These include splenic artery modulation (ligation/embolisation), portacaval shunts and less commonly, splenectomy. To date, there are no studies directly comparing outcome from these techniques.

Splenic artery modulation techniques

Splenic artery modulation (SAM) was originally performed for portal hypertension secondary to cirrhosis. It was shown to be an effective treatment for post transplant patients exhibiting signs of SFSS, probably because of the reduction in portal pressure gradient[21]. In patients with portal hypertension, occluding the splenic artery reduces portal flow by 52% on average[22]. More recently, there is increasing tendency to perform SAM as a prophylactic procedure either based on algorithms predicting high risk of SFSS pre-operatively or based on intraoperative detection of high portal flow.

Gruttadauria et al[23] performed splenic artery embolisation (SAE) in six patients who developed SFSS after transplantation of GWRW < 0.8% grafts. Rapid resolution of symptoms occurred post SAE. However, one patient suffered massive colliquation of the spleen necessitating re-laparotomy and another suffered septic shock with consequent re-transplantation.

Two case-control series have tested the effectiveness of prophylactic splenic artery modulation[24,25]. Both studies found a significant reduction in portal flow following SAM and Umeda et al[25] were able to show a significant reduction in incidence of SFSS. Remarkably, neither group reported any cases of splenic infarction and this fortunate low complication rate was not replicated in other smaller case series[26]. This discrepancy in complication rates suggest splenic infarction may be reduced in experienced hands but remains a formidable problem.

Portacaval shunts

Portacaval shunting has gained favor in recent years due to its potential reversibility and to avoid possible splenic infarction[27,28]. Three case series where hemiportacaval shunts were constructed by anastomosing the right portal vein to the inferior vena cava have reported reasonable success in improving outcome of small for size grafts[29-31]. Portacaval shunts were able to reduce portal blood flow and pressure and lessen the likelihood of deranged liver function tests (LFTs) and international normalised ratio post operatively. One study[30] reported increase in graft survival from 20% in the control group to 75% in the shunt group.

In small-for-size rat models, however, concerns have been raised over the safety of a long term shunt[32] where a group of rat liver transplants with large portacaval shunts showed significantly worse graft survival rates at one year compared to the small shunt or no shunt groups. A case report from Japan where a portacaval shunt was constructed for small-for-size living donor liver transplantation (LDLT) noted progressive graft atrophy and a decision was made to close the shunt at 11 months. Fortunately this resulted in regeneration of the graft[33].

The natural history of a portacaval shunt is to occlude with time. In one study[31], 55% of shunts remained patent at six months and only 20% were patent at one year. In the shunts that remain patent, it is possible that persistent diversion of blood flow to the liver will compromise long term viability of the graft through mechanisms such as chronic ischemia. Hence shunts are likely to improve graft survival in the weeks immediately post operatively by reducing incidence of SFSS but may become a liability in the long term. Is there an optimum time, then, for electively closing the shunt?

MIDDLE HEPATIC VEIN CONTROVERSY: DONOR SAFETY VS GRAFT CONGESTION

The middle hepatic vein (MHV) is considered “dominant” in drainage of the hemiliver in 27% of cases[34]. A right hepatectomy without the MHV or reconstruction can induce congestion of the paramedian segments V and VIII, reducing functional capacity of the graft. Harvesting the MHV in extended hepatectomy increases the risk of complications in the donor. The questions are therefore threefold: should we harvest or reconstruct the MHV? If so, which recipients would benefit most? What is the risk to the donor?

Lee et al[35] reported grafts without the MHV exhibited congestion of the right median sector leading to ascites and severe LFT derangement. Kamei et al[36] introduced the concept of non-congestive GRWR (ncGRWR) as a better measure of graft function than size ratio alone and showed patients with ncGRWR < 0.65 developed prolonged cholestasis, one of the features of SFSS.

Other studies found no significant difference in graft survival with or without harvest of the MHV as long as a vein interpositional graft was used for anastomosis[37,38]. It is important to note that most studies on MHV included grafts of all sizes. If we stratify grafts by size, the importance of venous congestion emerges[39,40]. In a series of 120 patients where 67% had reconstruction of MHV, there was no significant benefit in venous reconstruction for a large graft. For small-for-size grafts (GWRW < 1), however, patients who did not have venous reconstruction had deranged LFTs for significantly longer periods of time and also had slower regeneration of the graft (95% vs 80% at one month). In the medium term, grafts with reconstruction of the MHV had higher rates of survival at six months[41].

In the early days of living related and split cadaveric grafts, the decision to harvest the MHV was set by institutional policy. More recently, several centers have tried to rationalise harvesting the MHV by designing algorithms to predict recipients most likely to suffer small-for-size syndrome. One of the earliest algorithms incorporated donor-recipient weight ratio, right lobe-to-recipient standard liver volume estimate and donor hepatic vein anatomy, including diameter and number of tributaries[42]. This split the patients into two cohorts with comparable baseline demographics and they were able to obtain similarly low complication rates regardless of whether the MHV formed part of the graft. Later algorithms have incorporated hepatic vein dominance measured by 3D CT and congestion volumes[43]. This is potentially better representative of the relative importance of the MHV.

Since safety of the donor is paramount in transplant surgery, it is important to quantify the risk. In one series (n = 105) where the MHV was not harvested, 13.3% of donors experienced major complications with eight patients requiring invasive paracentesis and three requiring further surgery. Does harvesting the MHV confer additional risk to the donor? Contrary to common belief, there is in fact remarkably little solid evidence to support this. Dayangac et al[44] found right hepatectomy with MHV harvest does not affect donor liver function or increase donor morbidity compared to control unless the liver remnant is < 30%. Congestion in segment IV of the donor is a common observation after MHV harvest but its long term significance is debatable. In donors who experienced mild, moderate and severe congestion, LFTs were significantly increased in the severe group at seven days post operation but by the end of the first month after transplantation all three groups had normal LFTs[45]. Congestion has been reported to reduce rate of regeneration in segment IV in the donor; however, compensatory growth in segments II and III was able to make up the shortfall[46].

REGENERATION PARADOX

The molecular basis underlying liver regeneration is highly complex and beyond the scope of this review. Here, we wish to briefly address the interesting observation that size of partial liver grafts is negatively correlated to rate of regeneration and the clinical implications of this finding.

It has been noted that liver size is related to the functional demands placed upon it by the body. Hence small-for-size grafts undergo compensatory growth whereas large-for-size grafts shed cells by apoptosis[47]. Smaller grafts have been shown to have a higher rate of regeneration despite showing signs of endothelial injury and sinusoidal congestion[48,49]. However, fast regeneration is not necessarily predictive of good outcome, as a porcine small-for-size model showed that proliferative activity in non-surviving grafts peaked earlier and higher whereas surviving grafts demonstrated a slower but maintained rise[8].

A body of evidence suggests liver graft regeneration is related to velocity and volume of portal flow. Park et al[50] have demonstrated a correlation between portal venous flow or velocity to graft weight ratio with short term regeneration in LDLTs. Regenerative rates have been shown to be proportional to spleen volume and portal inflow[51]. Cirrhotics generally experience faster regeneration compared to those undergoing transplants for other reasons, a phenomenon which is correlated to a persistent hyperdynamic portal venous circulation[52]. However, confounding factors cannot be ruled out when comparing cirrhotics to other transplant recipients. Portal hyperflow induced shear stress and nitric oxide release have been singled out as possible mediators in stimulating liver regeneration in the setting of partial hepatectomy[53,54]. The above findings have profound implications for splenic artery ligation and portacaval shunting techniques as these may compromise compensatory regrowth.

What is the optimal portal flow that will stimulate regeneration without damaging the graft? We need studies that quantitatively correlate portal flow to both rates of regeneration and severity of graft injury. One recent study tentatively suggests a threshold of portal venous flow to graft weight of 300 mL/minutes per 100 g on post operative days one to three, based on 18 LDLTs. Above this level, LFTs were significantly more deranged[55]. Larger, better controlled studies are needed to clarify this threshold which will have key therapeutic implications.

Portal hyperperfusion is thought to cause liver injury and defective regeneration via interleukin-6 and tumor necrosis factor (TNF)-α signalling. Tian et al[56] report a fascinating mouse model where TNF pathways were interrupted by receptor knockout, treatment and gadolinium chloride and pentoxifylline (PTX). These mice underwent 30% liver transplantation. In the groups with TNF pathway intervention better portal flow and sinusoid perfusion was seen with reduced leukocyte adherence. Graft survival was dramatically increased: 14% in controls, 57% in TNF receptor knockout, 43% in gadolinium chloride and 86% in PTX. This elegant demonstration shows it is possible to prevent the effects of hyperperfusion injury without physically reducing flow.

CONCLUSION

Small-for-size syndrome has become an increasingly well recognised condition with the rise in popularity of LDLT and split cadaveric grafts. Better understanding of its pathogenesis, risk factors and strategies for prevention will improve both donor and recipient outcomes and expand the potential organ pool.

Although many significant advances have been made in understanding and managing small-for-size syndrome, much work remains to be done. Of prime importance is an internationally agreed set of diagnostic criteria for SFSS which would help us to clarify the scale of the problem and enable future studies to be standardised.

Portal hyperperfusion appears to be the most important underlying mechanism for SFSS; however, we must remember that regeneration relies on an adequate blood supply and interventions to reduce SFSS should strike a delicate balance between avoidance of hyperflow injury and stimulation of regeneration. Independent contribution of poor hepatic arterial flow to graft dysfunction remains to be clarified. A gold standard measurement of portal hyperperfusion, whether portal venous pressure or flow, should be agreed. Crucially, we need to establish the threshold level of hyperperfusion that does more harm than good.

Splenic artery modulation and portacaval shunting have both shown promise in prophylaxis and treatment of SFSS in multiple case series. Evidence from randomised control trials have so far been lacking but will perhaps become feasible with more LDLT and split cadaveric grafts being performed in the future.