Prognosis of adult patients transplanted with liver grafts < 35% of their standard liver volume

Authors


  • See Editorial on Page 1392

Abstract

We have previously reported that a graft volume (GV) > 30% of the recipient's standard liver volume (SLV) can meet the recipient's metabolic demands. Here we report our experience with adult-to-adult living donor liver transplantation using left side grafts < 35% of the recipient's SLV. Of 143 adult living donor liver transplants, 13 auxiliary partial orthotopic liver transplants, 8 right side grafts, and 2 retransplantation cases were excluded. The resulting 120 cases were divided into 2 groups: group S consisted of 33 patients who received liver grafts < 35% of their SLV, and group L consisted of 87 patients who received liver grafts ≥ 35% of their SLV. Patient characteristics, postoperative liver function, duration of hospital stay, and recipient survival rates were compared between the 2 groups. There were no significant differences between groups in recipient or donor background characteristics. The mean GV/SLV ratio of group S was 31.8%, whereas that of group L was 42.5%. There were no significant differences in the postoperative serum total bilirubin levels, prothrombin time international normalized ratio, daily ascites volume, or duration of postoperative hospital stay between the groups. The 1- and 5-year survival rates in group S were 80.7% and 64.2%, respectively, whereas those of group L were 90.8% and 84.9%, respectively, with no significant difference between groups. In conclusion, graft size was not considered to be the only cause of so-called small-for-size graft syndrome, and left side grafting appears to be the procedure of choice for adult-to-adult living donor liver transplantation because of the lower risk to donors in comparison with right lobe grafting. Liver Transpl 15:1622–1630, 2009. © 2009 AASLD.

Living donor liver transplantation (LDLT) was initially used only in pediatric patients to resolve the shortage of available deceased donors. This procedure involves using the lateral segment or the left lobe as a graft, which generally produces good results. In Japan, following the first case of LDLT performed in 1989, deceased donor liver transplantation was not performed again until 1999, and a total of only 45 such procedures have been performed in Japan through the end of 2007. In an attempt to expand the indications of LDLT for the treatment of adult patients suffering from liver failure, we succeeded in performing the first adult-to-adult LDLT in the world in 1993.1 Widespread application of LDLT has considerable potential for increasing the pool of available organs and thus alleviating the problem of organ shortage.

One of the biggest problems in adult-to-adult LDLT is that the graft volume (GV) can be insufficient for the recipient's metabolic demands. Accordingly, we proposed the concept of standard liver volume (SLV) and developed a formula for calculating SLV on the basis of the recipient's body weight and height.2 Using this formula, we previously reported that grafts with a volume ≥ 30% of the recipient's SLV could meet the recipient's metabolic demands. However, small grafts have been reported to cause so-called small-for-size graft syndrome, and some investigators have reported that grafts < 40% of the SLV produce poor results. Currently, right lobe grafts are used in about 60% of adult-to-adult LDLT cases in Japan to obtain a sufficient liver mass for transplantation.3

It is generally accepted that 30% of the total volume in a normally functioning liver meets the patient's metabolic demands during liver resection,4 and a residual liver volume of ≥30% of the total liver volume is deemed safe in right liver graft cases.5 We have previously performed LDLT with a graft volume to recipient standard liver volume (GV/SLV) ratio of >30%. Here we report our experience in adult-to-adult LDLT using a left side graft < 35% of the recipient's SLV.

Abbreviations

BW, body weight; CT, computed tomography; DD, duct-to-duct; FFP, fresh frozen plasma; GRWR, graft-to-recipient weight ratio; GV, graft volume; GV/SLV, graft volume to recipient standard liver volume; HCC, hepatocellular carcinoma; LDLT, living donor liver transplantation; MELD, Model for End-Stage Liver Disease; POD, postoperative day; PT-INR, prothrombin time international normalized ratio; RY, Roux-en Y; SLV, standard liver volume; T. Bil, total bilirubin.

PATIENTS AND METHODS

Through October 2007, we performed LDLT 143 times in adult (>18 years old) recipients. All procedures were performed after informed consent from patients and approval from the Ethics Committee of Shinshu University were obtained.

Left side liver transplantation was performed in 134 cases (left lobe graft in 107 cases and left lobe with caudate lobe graft in 27 cases), a right posterior segmental graft was used in 1 case, and 4 domino transplants were performed with right lobe grafts. Use of the original right lobe was limited to the remaining 3 cases who fulfilled our indication criteria: the donor left lobe volume was estimated to be <30% of the recipient SLV, the donor right lobe was estimated to be >30% of the recipient SLV, and the donor's remnant left lobe was estimated to be >35% of the donor's entire liver volume. LDLT was performed in an orthotopic liver transplantation manner in 130 cases and in an auxiliary partial orthotopic liver transplantation manner in 13 cases. These 13 auxiliary partial orthotopic liver transplantation cases, 8 right side liver graft cases, and 2 retransplantation cases were excluded.

The remaining 120 cases were divided into 2 groups: in group S, the implanted liver graft was <35% of the recipient's SLV, whereas in group L, the implanted liver graft was ≥35% of the recipient's SLV.

Donor and Graft Selection

The primary selection criterion for a living donor was voluntary. Acceptance criteria for donors included an age of 20 to 65 years, a relationship within a fourth degree of consanguinity with the recipient or spouse, blood group compatibility, and negative serological test results for hepatitis B, hepatitis C, and human immunodeficiency viruses. All donors had normal liver function test results and no history of liver disease. Electrocardiography and pulmonary function tests were also performed. For the evaluation of donor candidates > 30 years old, gastrointestinal fiberscope endoscopy and fecal occult blood testing were performed to rule out gastrointestinal tract malignancy. For the evaluation of donor candidates > 40 years old, ultrasound echocardiography was routinely performed to rule out asymptomatic heart disease. Any candidates with heart disease or malignancy were excluded from the study.

Eligible donors underwent imaging studies, including chest and abdominal X-rays, 5-mm-slice computed tomography (CT), and magnetic resonance imaging. CT was used to exclude any unrecognized intra-abdominal pathological conditions and for volumetric analysis of graft size matching and delineation of vascular anatomy. Magnetic resonance imaging was used to delineate the bile tree and to detect fatty liver. We accepted a donor-recipient combination if the GV/SLV ratio was ≥30%. Percutaneous liver biopsy was not routinely performed. Recipient SLV was calculated according to the formula developed by Urata et al.2 GV was predicted from the CT volumetric analysis. The type of liver graft used depended on the build of the recipient and the calculated segmental volume of the donor liver.

Donor Surgical Procedure

The surgical procedure has been described in detail elsewhere.6 Briefly, whole left lobectomy, including the middle hepatic vein with or without the caudate lobe, was performed.7 If necessary, hepatic graft venoplasty was conducted with the left and middle hepatic veins on a side table to form a common trunk for the drainage vein of the graft. The graft was perfused with cold University of Wisconsin solution (Viaspan, Dupont, Wilmington, DE). The actual GV was measured at the side table during the procedure. The GV/SLV ratio was used for graft size matching.

Recipient Surgical Procedure

Recipient procedures have been described elsewhere.6 Recipient total hepatectomy was performed without venovenous bypass with preservation of the inferior vena cava. To completely avoid intraoperative portal venous congestion, we used a temporary shunt between the portal venous branch and inferior vena cava that we previously developed for adult patients with noncirrhotic liver diseases such as familial amyloidosis, citrullinemia, and fulminant hepatic failure.8 Recipient left and middle hepatic veins were used for venoplasty. Following venoplasty, end-to-end anastomosis of the hepatic veins and reconstruction of the portal vein between the donor left portal vein and the recipient left portal vein were performed. Immediately after graft reperfusion, the temporary shunt (if used) was closed. The donor left hepatic artery was anastomosed to either the recipient right hepatic artery or the left hepatic artery, depending on the size match, under direct observation with a surgical microscope. When the graft involved multiple hepatic arteries, the largest one was reconstructed first. The remaining graft arterial branches were not reconstructed when pulsatile bleeding was observed from the nonanastomosed stump and good intrahepatic arterial flow was confirmed in each subsegment by intraoperative Doppler ultrasound.9 Bile duct reconstruction was performed in a Roux-en Y fashion before 2006 and in a duct-to-duct fashion after 2006. In the case of Roux-en Y reconstruction, a jejunostomy tube was placed in the afferent limb for intrabowel decompression. In the case of duct-to-duct reconstruction, a tube was inserted into the remaining cystic duct, and its end was placed in the anastomosis as a stent, ensuring biliary drainage. The tube was anchored to the anastomosis with an absorbable stitch. In both cases, the tube was removed 3 months after LDLT. In cases with a platelet count < 30,000/mm3, splenic arterial ligation or splenectomy was performed.

Postoperative Management

The initial immunosuppressive regimen consisted of tacrolimus and steroids. As postoperative anticoagulation therapy, low-dose low-molecular-weight heparin (for 2 weeks), antithrombin III (for 2 weeks), protease inhibitor (for 2 weeks), and prostaglandin E1 (for 1 week) were administered intravenously. Fresh frozen plasma (FFP) was also administered in order to supply anticoagulants. Patient hematocrit was maintained between 20% and 30% to minimize blood viscosity. Color Doppler ultrasonography was performed on each postoperative day for at least 1 month in all recipients to confirm the patency of intrahepatic arteries. Postoperative ascitic fluid was drained through indwelling catheters. On the basis of the protein level in the ascitic fluid, FFP was infused in order to maintain the serum protein level within the normal range.

Analysis of the Recipients

The serum total bilirubin (T. Bil) levels and ascites volume of the recipient, which are 2 of the parameters for determination of small-for-size graft syndrome, were compared between groups on postoperative days 1, 3, 5, 7, 10, 14, 21, and 28. The prothrombin time international normalized ratio (PT-INR) was also compared as a marker of protein synthesis. The volume of FFP administered postoperatively was compared as it might affect PT-INR. The cause of death of the recipients and the 1-, 3-, and 5-year survival probabilities of the grafts and recipients were compared between the 2 groups.

Evaluation of GV at 1 Month After LDLT

To compare graft regeneration between the 2 groups, the GV in the recipients who underwent CT at approximately 1 month (3-5 weeks) after LDLT was calculated with CT-based volumetry. Because our hospital adopted multidetector CT with computerized memory storage in 2002, the subjects of this study included 41 recipients who underwent multidetector CT at approximately 1 month after LDLT (13 recipients in group S and 28 in group L).

Analysis of Donor Safety

The intraoperative blood loss and postoperative complications in donors were examined in order to evaluate donor safety. Furthermore, 2 domino donors (the first recipient of domino liver transplantation) were excluded from this study.

Statistical Analysis

Univariate analysis was performed for categorical variables with the use of the chi-square test. We analyzed continuous variables with a 2-tailed unpaired t test. On univariate analysis, P < 0.05 was considered significant. Kaplan-Meier estimates were used to calculate graft survival curves. Differences in survival curves were compared by log-rank analysis.

RESULTS

Recipient and Donor Characteristics

Characteristics of the recipients and donors of both groups are listed in Table 1. Thirty-three cases were assigned to group S, and 87 cases were assigned to group L. There were no significant differences between group S and group L with respect to recipient age, recipient gender, Model for End-Stage Liver Disease (MELD) score, etiology, intraoperative blood loss, donor age, donor gender, body weight, blood type incompatibility, or relationship to the recipient. The average MELD score was relatively low because the population of patients with metabolic disease was larger. The mean MELD score of the patients with metabolic disease was 9.1 in group S and 8.7 in group L. On the other hand, the mean MELD score of the patients with non-metabolic disease was 18.0 in group S and 18.6 in group L. These differences were not statistically significant. The mean GV/SLV ratio of group S was 31.8%, whereas that of group L was 42.5%. The mean recipient body weight in group S was greater than that in group L (58.4 versus 52.6 kg, P = 0.0087), whereas the mean explanted graft weight of group S was lower than that of group L (363.3 versus 454.6 g, P < 0.0001).

Table 1. Patient Characteristics
 Group SGroup LP Value
  1. Abbreviations: BW, body weight; DD, duct-to-duct; GV/SLV, graft volume to recipient standard liver volume; HCC, hepatocellular carcinoma; MELD, Model for End-Stage Liver Disease; RY, Roux-en Y.

Recipients   
 Age (years)46.6 ± 14.746.9 ± 13.80.9243
 BW (kg)58.4 ± 8.952.6 ± 10.40.0087
 Gender (male/female)15/1832/550.5095
 MELD16.1 ± 9.916.1 ± 7.70.9942
 MELD > 257120.4752
 Child-Pugh classification   
  A9190.6990
  B10290.9218
  C14390.9754
 Etiology   
  Cholestatic cirrhosis9310.5153
  Noncholestatic cirrhosis12270.7351
  Acute hepatic failure560.2960
  Metabolic disease7220.8205
  Others01>0.9999
 With/without cirrhosis21/1259/280.9227
 With/without HCC10/2319/680.4664
Donors   
 Age (years)34.4 ± 12.138.6 ± 11.60.0857
 Gender (male/female)24/965/22>0.9999
 Blood type compatibility (identical/nonidentical)27/663/240.4087
 Relationship   
  Spouse3160.3340
  Parent4140.7967
  Child17400.7355
  Sibling8130.3533
  Domino11>0.9999
  Other consanguinity030.6704
Graft   
 GV/SLV (%)31.8 ± 2.442.5 ± 6.0<0.0001
 Graft weight (g)363.3 ± 36.1454.6 ± 63.8<0.0001
 Left lobe/left lobe with caudate lobe24/970/170.5029
Recipient operation   
 Reconstruction of bile duct (DD/RY)5/286/810.2960
 Blood loss (mL)2826 ± 25263033 ± 28480.7201

Postoperative Graft Function

The majority of liver grafts were immediately functional and showed progressive normalization. Serial changes in the postoperative serum T. Bil levels, PT-INR, ascitic fluid volume drained through indwelling catheters, and administered FFP volume in the 2 groups are shown in Fig. 1. Although the mean preoperative PT-INR in group S was significantly higher than that in group L (P = 0.0307), no significant differences were observed in the mean recipient postoperative serum T. Bil levels, PT-INR, ascitic fluid volume, and FFP volume in the 2 groups. In 36 recipients, 13 in group S (39.4%) and 23 in group L (26.4%), the T. Bil levels on day 14 after LDLT did not return to less than 3.0 mg/dL, and the elevation persisted until day 28 after LDLT in 6 group S patients (9.1%) and 13 group L patients (14.9%). The proportion of recipients showing elevated T. Bil levels did not significantly differ between the 2 groups. Among the recipients with prolonged cholestasis (more than 28 days), 4 recipients in group S and 3 in group L died within 1 postoperative year. The daily ascitic fluid volume increased postoperatively and peaked on postoperative days 5 to 7. Thereafter, the volume gradually decreased. The daily ascitic fluid volume did not significantly differ between the 2 groups. The volume on day 14 exceeded 1000 mL in 6 group S patients and 9 group L patients. Persistent massive ascites on day 28 was recognized in 1 group S recipient and 2 group L recipients. No significant difference was noted in the incidence of persistent massive ascites between the 2 groups.

Figure 1.

Serial changes in (A) the postoperative serum T. Bil levels, (B) PT-INR, (C) ascitic fluid volume drained through indwelling catheters, and (D) administered FFP volume in (▪) group S and (○) group L. Abbreviations: FFP, fresh frozen plasma; PT-INR, prothrombin time international normalized ratio; T. Bil, total bilirubin.

Graft Regeneration

The GV increased rapidly after LDLT in all recipients. The regeneration rate of the graft at 1 month (GV at 1 month after LDLT/harvested liver volume) in group S (2.24 ± 0.36) was significantly greater (P = 0.014) than that in group L (1.90 ± 0.40), whereas the GV/SLV ratio at 1 month did not significantly differ between group S patients (71.1% ± 11.1%) and group L patients (80.4% ± 18.1%).

Postoperative Hospital Stay

The mean duration of the postoperative hospital stay of the group S cases (89.0 ± 43.2 days) was shorter than that of the group L cases (103.9 ± 91.8 days); however, this difference was not significant.

Postoperative Complications in the Recipients

As shown in Table 2, the incidence of rejection, vascular complications, biliary complications, sepsis, massive ascites, and prolonged cholestasis did not significantly differ between the 2 groups.

Table 2. Incidence of Postoperative Complications in Recipients
 Group S (n = 33)Group L (n = 87)P Value
  • Abbreviation: POD, postoperative day.

  • *

    As defined by Dahm et al.15

  • Ascites volume > 1000 mL.

  • Total bilirubin > 3.0 mg/dL.

Rejection   
 Acute rejection7340.1037
 Chronic rejection00
Vascular complications   
 Hepatic arterial thrombosis24>0.9999
 Portal venous thrombosis360.9845
Biliary complications   
 Bile leakage030.6704
 Biliary stenosis10170.3097
Sepsis450.4263
Small-for-size graft syndrome*01>0.9999
Massive ascites   
 POD 14690.3953
 POD 2812>0.9999
Prolonged cholestasis   
 POD 1413220.1960
 POD 286130.8776

Causes of Death

A total of 14 patients died within 1 year after LDLT. Furthermore, the causes of the 6 deaths that occurred in group S were sepsis (in 2 patients) and multiple organ failure related to the side effects of the anticancer agents used preoperatively for the treatment of malignant lymphoma, hepatic arterial thrombosis, hepatitis C virus–related fibrosing cholestatic hepatitis, and Pneumocystis carinii pneumonia (in 1 patient each). The causes of the 8 deaths in group L were hepatic artery thrombosis, simultaneous hepatic artery and portal vein thrombosis, veno-occlusive disease, hepatic failure, hepatitis C virus–related fibrosing cholestatic hepatitis, aspergillosis, thrombotic thrombocytopenic purpura, and intracranial hemorrhage (in 1 patient each).

Between 1 and 5 years after LDLT, 7 patients (3 in group S and 4 in group L) died. The causes of death were cerebral infarction, recurrence of hepatocellular carcinoma, and uterine cancer in 1 patient each in group S and recurrence of hepatocellular carcinoma, intracranial hemorrhage, recurrence of colon cancer, and sepsis in 1 patient each in group L. These deaths occurred in patients who had good liver function after they had resumed their normal lives. In both groups, the causes of death were not related to insufficient graft size.

Graft and Recipient 1-, 3-, and 5-Year Survival Rates

The smallest GV/SLV ratio in this series was 26.7%, and the patient transplanted with this graft is currently doing well. The overall 1-, 3-, and 5-year patient and graft survival rates were 87.5%, 85.2%, and 80.1%, respectively. In group S, the 1-, 3-, and 5-year patient and graft survival rates were 80.7%, 77.0%, and 64.2%, respectively, and those in group L were 90.8%, 88.1%, and 84.9%, respectively. There were no significant differences in survival rates between the 2 groups.

Prognosis of Patients with MELD Scores > 25

Of the patients included in our study, 7 (21%) in group S and 12 (14%) in group L showed MELD scores > 25, and the proportions of these patients in the 2 groups were comparable. The etiologies leading to such high MELD scores were as follows:

5 fulminant hepatic failure and 2 cholestatic cirrhosis in group S, and 3 fulminant hepatic failure, 5 cholestatic cirrhosis, 3 non-cholestatic cirrhosis, and 1 metabolic disease in group L.

Among them, 3 group S patients and 2 group L patients died within 1 year. The 1-year survival rates of the patients and grafts were 57.1% in group S and 83.3% in group L; the difference between the 2 groups was not significant.

Donor Outcomes

The intraoperative blood loss in donor groups S and L was 515 ± 203 and 733 ± 477 mL, respectively. The mean postoperative hospital stay in groups S and L was 23.2 ± 9.2 and 22.0 ± 10.4 days, respectively. One donor in group L required an allo-blood transfusion. The major complications in the donors are listed in Table 3. Gastric volvulus was the most common complication in our series, and it was corrected endoscopically.10 All donors could resume their normal lives after discharge. In our series, relaparotomy was performed perioperatively in 1 donor in group S for hernioplasty of an incisional hernia.

Table 3. Postoperative Complications in Donors
 Group S (n = 32)Group L (n = 86)
Gastric volvulus49
Peripheral nerve palsy04
Incisional hernia11
Sepsis01

DISCUSSION

Donor safety has been considered to be the most crucial problem associated with LDLT since its inception in 1989.11 To minimize the risk to living donors, many transplant surgeons aim at procuring the smallest liver volume required, and this leads to grafts that are potentially too small. Small grafts that are unable to meet the recipient's metabolic demands can result in liver failure, including coagulopathy, ascites, prolonged cholestasis, and encephalopathy; they are often associated with pulmonary and/or renal failure; and they frequently lead to death of the recipient in the absence of retransplantation.12-14 This ill-defined clinical picture is considered to be primarily linked to insufficient graft size and is often called small-for-size graft syndrome. However, the precise definition and underlying pathogenesis of this syndrome remain controversial.

Dahm et al.15 proposed a definition of small-for-size graft syndrome, as summarized in Table 4. In our present series, 80 patients received a graft that weighed < 0.8% of their body weight. However, none of these patients fulfilled the criteria for small-for-size dysfunction or small-for-size nonfunction. Although the number of patients in our series was small, the incidence of small-for-size graft syndrome was very low within our supposed small-for-size population.

Table 4. Definition of Small-for-Size Graft Syndrome
  • NOTE: This table is based on the work of Dahm et al.15

  • Abbreviation: GRWR, graft-to-recipient weight ratio.

  • *

    Graft dysfunction is defined as the presence of 2 of the following on 3 consecutive days: bilirubin > 100 μmol/L, international normalized ratio > 2, and encephalopathy grade = 3 or 4.

  • Graft failure is defined as retransplantation or death of the recipient.

  • Exclusion criteria include technical (eg, arterial or portal occlusion, outflow congestion, or bile leak), immunological (eg, rejection), and infectious (eg, cholangitis and sepsis) criteria.

Small-for-Size Dysfunction
Dysfunction* of a small partial liver graft (GRWR < 0.8%) during the first postoperative week after the exclusion of other causes
Small-for-Size Nonfunction
Failure of a small partial liver graft (GRWR < 0.8%) during the first postoperative week after the exclusion of other causes

In Japan, left lobe graft donation for adult recipients has been almost abandoned as sufficient data about the safety and efficacy of this procedure are lacking; in 2005, 66% of adult-to-adult LDLT procedures were performed with whole right lobe grafts.3, 16 However, there is a potential difference in morbidity, mortality, extent of surgery, and subsequent recovery that exists between right hepatic lobectomy and left hepatic lobectomy. Some differences with respect to rates and severity of postoperative morbidity have previously been reported.17-19 Accordingly, with respect to minimizing hepatectomy-associated risks in the donor, LDLT with the left lobe may be ideal. However, LDLT performed with left lobe grafts may be disadvantageous for the recipient in comparison with right lobe grafting. In the patients from whom left-sided grafts were obtained, no life-threatening complications occurred, and all donors resumed their normal lives in our cases. The morbidity of the living donors after left hemihepatectomy was low or minimal.

The present study demonstrated that the small-for-size graft syndrome parameters (postoperative PT-INR, serum T. Bil levels, and daily ascites volume) did not significantly differ between recipients with a GV/SLV ratio > 35% and those with a GV/SLV ratio < 35%. Furthermore, there were no significant differences in recipient and graft survival rates between the 2 groups. The 1-, 3-, and 5-year survival rates in group S were worse than those in group L. However, the graft sizes in the 2 groups were almost equal at approximately 1 month after LDLT, and the causes of death in group S patients after 1 postoperative year were not related to insufficient graft size. Although postoperative sepsis might be related to small-sized grafts, its incidence was not different in the 2 groups. The present results suggest that in adult-to-adult LDLT, liver grafts with volumes < 35% of the SLV can function as well as those with volumes > 35% of the SLV. However, careful consideration is required when small grafts are transplanted into severely ill recipients with MELD scores > 25 as the prognosis in such patients can be worse.

The hospital stay of group S patients was shorter than that of group L patients without significant differences. However, the hospital stay in both groups was considerably long. The reason for the patients' prolonged hospital stay may be as follows. In our hospital, bile duct reconstruction during LDLT is performed in the Roux-en Y or duct-to-duct fashion. In both cases, the inserted drainage tube is removed 3 months after LDLT, and the patients are hospitalized until the drainage tube is removed. This procedure might have prolonged the patients' hospital stay.

Overall, the survival rate of patients who received grafts with GV/SLV < 35% was acceptable in comparison with that of patients who received grafts with GV/SLV > 35%. The acceptable survival rate of recipients of small-sized grafts might be attributable to the following. First, we infused FFP to replenish the proteins lost because of ascitic fluid drainage. The FFP volume that we used was relatively large. Second, postoperative complications could be diagnosed and treated immediately, even if they occurred at a relatively late phase after LDLT because of the patients' substantially longer hospital stays.

In the United States, LDLT experienced rapid growth that peaked in 2001; however, the number of donor hepatectomies performed each year has steadily decreased since that time. LDLT currently accounts for only 5% of liver transplants in the United States. One probable reason for this decrease is a single highly publicized donor death in the United States that was reported in early 2002.20 Interestingly, although a single LDLT donor also died in Japan in 2002, the number of donor hepatectomies did not decrease as a result. However, it is readily conceivable that the next donor death that occurs in Japan may lead to a subsequent increase in hesitancy from both Japanese donors and medical centers as occurred in the United States and may result in a decrease in the number of LDLT procedures performed in Japan. LDLT currently comprises approximately 99% of liver transplantation procedures in Japan; thus, many patients with hepatic failure could potentially lose the chance to be transplanted in Japan if such a decrease in the number of procedures occurred.

Although the incidence of postoperative complications did not vary between the small-sized graft recipients and large-sized graft recipients, a number of patients in our series developed intractable ascites or prolonged cholestasis postoperatively. These complications could represent the so-called small-for-size graft syndrome. The mechanism of small-for-size syndrome remains unclear; however, our results suggest that the small GV/SLV ratio itself is not the cause of small-for-size graft syndrome. Recently, some researchers have reported the relationship between small-for-size graft syndrome and portal venous flow. Shimamura et al.21 speculated that excessive portal venous flow may be a cause of small-for-size graft syndrome, and Konishi et al.22 reported that excessive portal venous flow causes massive ascites. Asakura et al.23 experimentally showed that excessive portal inflow after reperfusion plays a role in the development of this syndrome by aggravating the sinusoidal microcirculatory injury to the graft. Several techniques for the resolution of this problem have been recently reported, such as splenectomy24 and splenic artery ligation.25 In our study, splenectomy or splenic artery ligation was performed, if possible, in patients with platelet counts < 30,000/mm3; however, the volume of postoperative ascites did not significantly differ with or without splenectomy or splenic artery ligation (data not shown). The portocaval shunt is reported to be useful for the prevention of small-for-size graft syndrome.26-28 An increase in portal pressure is known to elevate the hydrostatic pressure within the liver sinusoids, and this may lead to the leakage of ascitic fluid. Therefore, a method that controls the portal venous flow may resolve one of the causes of small-for-size graft syndrome or at least postoperative intractable ascites. Furthermore, a “steal” of the portal flow via the portocaval shunt may occur, possibly causing graft dysfunction.29 In our series, a temporary shunt between the portal venous branch and the inferior vena cava was prepared to minimize the risk of intraoperative hypotension, postoperative renal failure, and complications associated with portal venous congestion in patients who have no established portosystemic collaterals. We ligated the shunt following the recognition of graft reperfusion.8 Therefore, the temporary shunt used in our patients may not prevent small-for-size graft syndrome. Whether the portocaval shunt should be constructed is controversial, and prospective randomized trials may be required to resolve this issue.

Experience gained in hepatic surgery has shown that an extended hepatectomy of up to 70% to 75% of the whole liver can be tolerated in patients without cirrhosis,5, 30 and donor right hepatic lobectomy of up to 70% of the donor whole liver is frequently performed. It is therefore unclear why surgeons require a larger liver volume in the setting of liver transplantation, especially when we consider that the reported minimum GV/SLV ratio for successful adult-to-adult LDLT is 20%.31

For LDLT, Lo et al.32 recommend a GV/SLV ratio of >40% and Kiuchi et al.33 recommended a graft-to-recipient weight ratio (GRWR) of ≥0.8% to achieve good graft and recipient survival rates. In our series, the GV/SLV ratio of group S was below 35%, and the GRWR in group S did not exceed 0.8% (median, 0.65%; range, 0.49%-0.76%); however, the survival rate was fairly good. Therefore, small graft size does not appear to be the only cause of small-for-size graft syndrome, and a GV/SLV ratio of >40% or a GRWR of >0.8% does not seem necessary to achieve good results.

The Kyusyu group reported consecutively about the graft size and the result of LDLT. In 2001, the group reported 33 cases of left liver grafts, among which 5 grafts were <30% of SLV, and they described satisfactory outcome with small grafts, that is, 26% to 29% of SLV with no graft loss.34 In 2008, the group studied 119 cases of left liver LDLT and suggested that aside from the graft size, the age of the donor and the recipient's MELD score played important roles.35 More recently, the group showed a formula to predict graft function and short-term prognosis, and in the formula, a GV/SLV ratio was the only parameter positive for outcome [predictive score = 0.011 × graft weight (% of the standard liver weight of a recipient) − 0.016 × donor age − 0.008 × MELD score − 0.15 × shunt (if present) + 1.757; r2 = 0.497, P = 0.01]36; this indicates that the graft size affects the liver function and prognosis after LDLT. However, on the basis of the results of the present series, we feel that a preoperatively evaluated GV/SLV ratio of 30% to 35% is also acceptable for LDLT donors, although it should be noted that the number of patients in this series was relatively small. Our sole criterion for selecting the graft type was based on the preoperatively calculated GV/SLV ratio. Left lobe grafts were used when the preoperatively predicted GV/SLV ratio was ≤30%. Right lobe grafts may be necessary only when the predicted GV/SLV ratio of left lobe grafting is <30%.

In conclusion, the prognosis of recipients with liver grafts < 35% of their SLV is comparable to that of recipients with larger grafts. Graft size itself is not the only cause of small-for-size graft syndrome, and we feel that left side grafts should be used more frequently in adult-to-adult LDLT because of the lower risk to donors in comparison with right lobe grafts.

Ancillary