If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
The intelligent practice of living donor liver transplantation (LDLT) has evolved from measures that strike a balance between maximum donor safety and the handling of manageable sequelae from the use of small-for-size grafts (SFSGs). SFSGs should be common and favored in adult LDLT on the premise of donor safety. Whenever possible, the left liver graft should be used. Nevertheless, for a recipient of large body size with respect to the donor, even the right liver graft can be small for size.
Any graft smaller than the standard liver volume (SLV) of the recipient could be called an SFSG. In clinical practice, however, a graft is small for size only if the size affects the outcome and requires a change in management. An SFSG is a prerequisite for the development of small-for-size syndrome (SFSS) in LDLT, a common mode of early treatment failure. SFSS is a clinical entity attributable to a relative shortage of functional liver mass and is manifested by cholestasis, ascites, pleural effusion, pulmonary edema, gastrointestinal bleeding, renal impairment, and encephalopathy. The electrolyte imbalance predisposes patients to the development of central pontine myelinolysis. Vulnerability to infection is the main reason for mortality. Although SFSS develops from SFSGs, circumstances that predispose patients to this process remain elusive. The histopathology of SFSS includes hepatocyte ballooning, steatosis, centrilobular necrosis, and parenchymal cholestasis.1 Transient portal hypertension results in irreversible sinusoidal damage in rats.2 Such damage is ultrastructurally demonstrable by electron microscopy in humans.3 Thus, an early increase in the portal pressure may play a key role in SFSS.4 On the contrary, it has been demonstrated that partial grafts with portal hyperperfusion have worse outcomes.5
In our center, the importance of free venous outflow is addressed systemically by the inclusion of the middle hepatic vein (MHV). A priori, the MHV is included in the right liver graft in LDLT to optimize hepatic venous outflow. The interrelationship of sufficient graft size, graft quality, adequate venous outflow, and tolerable portal hypertension or inflow has been conceptualized.6, 7 Nevertheless, whether the influence of portal hypertension or portal hyperperfusion induces SFSS requires clarification. This study was designed to delineate the relationship between graft size, portal pressure and flow, and recipient outcomes in right liver adult LDLT including the MHV. We employed continuous electronic manometry of the portal system and ultrasonic flowmetry of graft inflow to study the graft's ability to relieve portal hypertension and accommodate portal hyperperfusion. A possible correlation of the findings from such observations with surgical outcomes was investigated.
G/SLV, graft to standard liver volume; INR, international normalized ratio; LDLT, living donor liver transplantation; MELD, Model for End-Stage Liver Disease; MHV, middle hepatic vein; SFSG, small-for-size graft; SFSS, small-for-size syndrome; SLV, standard liver volume.
PATIENTS AND METHODS
Forty-six patients undergoing right liver LDLT (Department of Surgery, The University of Hong Kong, Queen Mary Hospital, Hong Kong) were recruited over a 2-year period. All the donor and recipient operations were performed by the same team of surgeons throughout this period. Cases with arterial anastomotic abnormalities, difficulties in cannulation of the inferior mesenteric vein due to a thickened retroperitoneum, or missing data were excluded from the analysis. The study was approved by the institutional review board.
Donor right hepatectomy has been described.8 Portal inflows of the right portal vein were recorded with a transit-time ultrasound flowmeter (Transonic Systems, Ithaca, NY) once they could be isolated in the donor during the donor operation. The alignment of the flow probe was adjusted to avoid twisting of or pressure on the vessel. All the right liver grafts included the MHV. The right hepatic vein and the MHV were joined together to form a single cuff by venoplasty.9 To measure the size of the right liver graft, it was weighed after being flushed with histidine-tryptophan-ketoglutarate solution on the back-table.10 This was expressed as a percentage of the SLV calculated with the Urata formula.11
In the recipient, the flow of the main portal vein prior to total hepatectomy was measured. After graft right portal vein anastomosis, the portal flow was measured. The continuous portal pressure was monitored by cannulation of the inferior mesenteric vein. A 20-cm-long, 16-gauge polyurethane catheter (Arrow International, Reading, PA) was placed before total hepatectomy of the native liver and was removed after completion of bile duct anastomosis. The pressure was read with a manometric pressure transducer and recorded.3 The readings of the portal pressure and flow were taken at 3 time points: before total hepatectomy, during the anhepatic phase, and after all the vascular anastomoses and prior to closure of the wound. The single-cuffed MHV and right hepatic vein were anastomosed with a venotomy of corresponding dimensions to the inferior vena cava.12 After temporal clamping of the graft right portal vein, the clamps on the inferior vena cava were released, and this restored venous return from the inferior vena cava.10 Venovenous bypass was not used to avoid the negative effects of this technique (eg, hypothermia and platelet consumption).13
Liver graft biopsy was performed prior to closure of the abdomen, and the sample was interpreted by a pathologist without previous knowledge of the graft size or details of the donor and recipient.
Data were collected prospectively and were expressed as medians and ranges. The Child-Pugh score and Model for End-Stage Liver Disease (MELD) score were included as continuous variables. The portal flow was expressed as millimeters per minute and millimeters per 100 g per minute. The latter was derived from the former; it was divided by the graft weight and multiplied by 100. Correlations of clinical parameters were made with linear regression analysis. Statistical analyses were performed with SPSS for Windows, version 16.0 (SPSS, Chicago, IL). A P value < 0.05 was considered statistically significant.
The median ages of the donors and recipients for these 46 LDLT procedures were 31 (range, 18-54 years) and 50 years (range, 19-66 years), respectively. There were more women among the donors than the recipients, with male-to-female ratios of 17:29 and 40:6 for the donors and recipients, respectively. The MELD score was 23 (range, 7-50). The graft weight was 585 g (range, 390-880 g), and the graft to standard liver volume (G/SLV) ratio was 47.4% (range, 32.4%-69.0%; Table 1).
Table 1. Donor and Recipient Characteristics (n = 46)
Model for End-Stage Liver Disease score
Graft weight (g)
Graft to standard liver volume (%)
The portal pressure dropped by 8 mm Hg (range, −7 to 19 mm Hg) from 23 (range, 8-37 mm Hg) to 14 mm Hg (range, 10-26 mm Hg) after graft implantation. The portal pressure gradient (portal pressure minus central venous pressure) before total hepatectomy of the native liver was 16 mm Hg (range, 5-27 mm Hg), and after graft implantation, it was 6 mm Hg (range, −1 to 11 mm Hg). Only 1 recipient had a portal pressure gradient greater than 10 mm Hg after transplantation. The portal inflow of the right liver increased from 438 mL/minute (200-1220 mL/minute) before donor right hepatectomy to 1892 mL/minute (568-3908 mL/minute). When expressed with respect to graft weight, it increased from 81 mL/minute/100 g (range, 35-210 mL/minute/100 g) before donor right hepatectomy to 318 mL/minute/100 g (range, 102-754 mL/minute/100 g) after graft implantation (Table 2).
Table 2. Portal Pressure and Flow in Donors and Recipients
After Graft Implantation
Portal pressure (mm Hg)
Portal flow (mL/minute)
Portal flow/100 g of graft weight (mL/100 g/minute)
Clearly, the portal inflow to the right liver graft had a positive correlation with the portal pressure before native liver hepatectomy (R2 = 0.305, P = 0.001; Fig. 1A). The correlation of the portal inflow to the portal pressure gradient was, however, poor (R2 = 0.116). A similar and weaker correlation with the portal pressure in the anhepatic phase (R2 = 0.106, P = 0.027) was seen (Fig. 1B). No such correlation with the portal pressure after implantation (R2 = 0.031, P = 0.245) was demonstrable (Fig. 1C). There was no correlation between the graft portal inflow and the graft weight (R2 = 0.026, P = 0.288; Fig. 2A) and G/SLV ratio (R2 = 0.054, P = 0.120; Fig. 2B).
The graft portal inflow, when expressed as the volume per 100 g of graft weight per minute, similarly had a linear correlation with the recipient portal pressure before hepatectomy of the native liver (R2 = 0.261, P = 0.001; Fig. 3A) and was weaker in the anhepatic phase (R2 = 0.098, P = 0.034; Fig. 3B) but not after graft implantation (R2 = 0.029, P = 0.260; Fig. 3C). A negative correlation was found between the graft portal inflow per 100 g per minute and the graft weight (R2 = 0.227, P = 0.001; Fig. 4A) and G/SLV ratio (R2 = 0.247, P = 0.001; Fig. 4B). A significant correlation of the graft portal inflow per 100 g of graft weight per minute was also shown with a quotient derived from the recipient portal pressure before total hepatectomy of the native liver divided by the G/SLV ratio (R2 = 0.381, P = 0.001; Fig. 5).
The bilirubin level on day 14 was 32 μmol/L (range, 11-483 μmol/L), and the international normalized ratio (INR) on day 14 was 1.1 (range, 1.0-1.3). The bilirubin level of the recipients with a postimplantation portal pressure ≤ 14 mm Hg on day 14 was comparable to that of the recipients with a postimplantation portal pressure > 14 mm Hg (34 versus 31 μmol/L, P = 0.412). The INRs on day 14 were also comparable (1.1 versus 1.1, P = 0.322). In fact, the bilirubin level of the recipients with a portal pressure gradient of 5 mm Hg or less on day 14 was 34 μmol/L, whereas that of the recipients with a portal pressure gradient > 5 mm Hg was 32 μmol/L (P = 0.798). Their INRs on day 14 were 1.1 and 1.1 (P = 0.763).
Only 1 biopsy of the liver graft after perfusion showed moderate sinusoidal dilatation. The hospital mortality rate was 4.4% since 2 recipients died from sepsis. Complications of grade 2 or higher according to Clavien's system14 occurred in 12 recipients (26%) and were not related to the graft portal flow (P = 0.900), graft portal flow per 100 g of graft weight (P = 0.881), or portal pressure before (P = 0.792) or after graft implantation (P = 0.696). It was also not related to the graft weight (P = 0.608) or G/SLV ratio (P = 0.881; Table 3).
Table 3. Clavien Grade 2+ Complications and Portal Flow, Portal Pressure, and Graft Size
Graft portal flow (mL/minute)
Graft portal flow per 100 g of graft weight (mL/minute)
Portal pressure before native liver implantation (mm Hg)
Portal pressure after graft implantation (mm Hg)
Graft weight (g)
Graft to standard liver volume ratio (%)
Two recipients died in the hospital. One of them was a 57-year-old male recipient who received a right liver graft with a G/SLV ratio of 50%. His MELD score was 42. He had hepatorenal syndrome before LDLT. The graft portal inflow was 440 mL/100 g/minute, and the portal pressure after graft implantation was 12 mm Hg. His liver function remained poor. He developed sepsis and succumbed to a subarachnoid hemorrhage 4 months after LDLT. The second recipient who died was a 62-year-old woman who received a right liver graft with a G/SLV ratio of 42%. Her MELD score was 38. She had a Klebsiella urinary tract infection and hepatorenal syndrome before LDLT. The graft portal inflow was 423 mL/100 g/minute, and the portal pressure after graft implantation was 17 mm Hg. She had refractory renal failure and systemic candidiasis and succumbed to multiorgan failure 2 weeks after LDLT.
The graft portal inflow was proportional to the portal pressure of the recipient before recipient hepatectomy. Such relations, though existing, were weaker in the anhepatic phase. These relations were lost after right liver graft implantation. Similarly, the graft portal inflow per 100 g of graft weight was positively correlated with the portal pressure before recipient hepatectomy and in the anhepatic phase but not after graft implantation. As shown in Figs. 1C and 3C, this positive correlation between the portal inflow and portal pressure was lost after graft implantation. This was worth special attention from us because it showed that even with a high portal inflow, the portal pressure after graft implantation was not high.
The graft size (both the graft weight and the G/SLV ratio) had no relation to the graft portal inflow. This also suggested that even SFSGs were effective in accommodating high portal flow. However, when the graft portal inflow was expressed per 100 g of graft weight, it showed a significantly inverse correlation with the graft weight and G/SLV ratio. Small grafts had a high portal inflow per 100 g of graft weight. However, one must be very careful with interpreting graft portal inflow expressed per 100 g of graft weight as the size of each graft confounded the graft inflow expressed in this way. Previous studies suggested that a high portal inflow per 100 g of graft weight was associated with complications. A high portal inflow per 100 g of graft weight existed in the SFSG. Furthermore, portal pressure data were not available from these studies.5, 15 Our data also showed that the graft size in the G/SLV ratio had no correlation with the portal pressure after graft implantation (R2 = 0.013, P = 0.451).
Our previous study indicated that a determinant of recipient outcome was a graft size that was 35% or less of the estimated standard liver weight.16 Our current graft design and technique can cope even with G/SLV ratios slightly lower than 35% and in the face of recipient portal hypertension. In this study, only 2 recipients had a graft size less than 35%, and this is a limitation of this series in representing cases with grafts that are much smaller.
As shown in this series, a high portal inflow was a common phenomenon and was well accommodated. Despite the high portal inflow after implantation, sinusoidal dilatation was found in only 1 graft by light microscopy examination of liver graft biopsy samples. The graft and recipient outcomes were unremarkable. The loss of this correlation with the portal pressure after graft implantation reflected efficacious relief of the portal hypertension by the graft. Our data showed no correlation of complications with the high flow rate and portal pressures (before and after implantation). With the exclusion of those with Clavien grade 1 complications, 12 of the 46 recipients had complications of grade 2 or higher. Two recipients who died did not have extreme portal flow or graft size. Nevertheless, they had high MELD scores and hepatorenal syndrome before LDLT. In many other centers, they would not have been accepted as LDLT candidates.
High portal inflow being called portal hyperperfusion is common in SFSGs. The modulation of portal flow has thus been advocated. Such maneuvers include splenic artery ligation with or without splenectomy. Guidance for deciding inflow modulation is based on the G/SLV ratio and the observation of portal hyperperfusion. The authors have recommended modulation of the portal inflow by splenic artery ligation if the portal flow is >250 mL/minute/100 g of graft weight.5 Our decision was made with reference to persistent portal hypertension instead. In a recipient with primary nonfunction of a small-for-size right liver graft, a portal pressure of 27 mm Hg was reduced to 13 mm Hg after splenic artery ligation on the first postoperative day, and this was followed by improvement of liver function.17 Elevated portal pressure in the early postoperative phase was associated with poor recipient survival with SFSGs.18 Inflow modification was not performed in any of these recipients according to the protocol. Only 3 recipients had a portal pressure > 20 mm Hg (21, 22, and 26 mm Hg). All 3 recipients survived. Portal vein shunting to reduce portal inflow also lowered portal pressure. The portal vein shunting procedure included a full or hemiportocaval shunt and a mesocaval shunt. In a pig model, a portocaval shunt in the face of an SFSG maintained sinusoidal structures.19 A mesocaval shunt with downstream ligation of the superior mesenteric vein for an SFSG diverted and reduced portal flow.20 Graft size was not a significant factor for sinusoidal injury in the current study. The inherent risk of shunts is portal hypoperfusion. This chance is minimized by careful intraoperative and frequent postoperative Doppler ultrasonography because a substantial number of the shunts are taken down or ligated during native liver hepatectomy.
High portal inflow (but not elevated portal pressure) is a reflection of excellent outflow capacity and higher portal pressure prior to implantation. In our series, therefore, potential complications of shunting such as portal hypoperfusion and the inconvenience of second laparotomy were eliminated. Evidence from another study also showed that a high portal flow did not jeopardize graft function. In fact, an earlier normalization of the prothrombin time was observed. In a series of 22 right liver LDLT procedures not including the MHV, the median graft portal flow was 243.7 mL/minute/100 g (range, 211.1-290.4 mL/minute/100 g).21 The elevated portal pressure in recipients of SFSGs in many studies may be a result of less than desired outflow capacities. The result of residual portal hypertension is probably a manifestation of insufficient venous outflow and graft congestion and damage. This study has not addressed hepatic arterial perfusion with respect to portal inflow and portal pressure. Graft loss or graft insufficiencies could be a result of arterial complications, which are influenced by the hyperdynamic portal system.22
The inclusion of the MHV in the right liver graft can be substantiated only if donor surgery is safe. Remnant left lobe venous drainage is ensured by preservation of the segment 4b hepatic vein, even when it drains into the root of the MHV.23 The incorporation of the MHV into the right liver graft coupled with a direct anastomosis to the inferior vena cava has 2 effects on the interpretation of flowmetry data. An excellent outflow capacity, regardless of the portal pressure and graft size, increases blood flow through the graft. The efficacious relief of venous pressure downstream also relieves the pressure within the graft, so there is a higher threshold for the development of sinusoidal injury. Because graft and recipient outcomes are not influenced by high portal inflow, portal hyperperfusion should be observed as a normal phenomenon. What is unknown is the minimum volume of liver parenchyma that can accommodate portal inflow before graft damage is seen. Further studies of the subgroup with a larger number of cases with G/SLV ratios less than 35% are needed. The excellent venous outflow capacity of this graft design may provide an opportunity for determining the minimum volume of liver parenchymal cells for graft and recipient survival.