Living donor liver transplantation (LDLT) is considered a useful means of overcoming the shortage of cadaveric donor organs and is almost the only available option for treating end-stage liver disease in Asia.1-3 The problem of graft size is one of the major factors limiting the expansion of adult-to-adult LDLT, and it often occurs when the graft-to-recipient weight ratio (GRWR) is <0.8%.4, 5 Problems such as small-for-size syndrome (SFSS) often occur in LDLT patients receiving grafts with a GRWR < 0.8% because the grafts cannot meet the functional demands of the recipients.6
A number of reports in the literature concern methods of improving the outcome after LDLT with small-for-size grafts (SFSGs). Kiuchi et al.7 reported that the outcomes of LDLT patients who received SFSGs with a GRWR < 1.0% were poor, and they concluded that the use of SFSGs led to lower graft survival. Generally, the selection of a graft with a GRWR > 0.8% (and preferably > 1.0%) has been recommended to improve graft survival and prevent postoperative graft dysfunction. However, Ben-Haim et al.8 reported that a GRWR as low as 0.6% can be used safely in recipients without cirrhosis or Child class A recipients. Furthermore, they claimed that Child class B or C recipients require a GRWR > 0.85% to avoid SFSS and related complications. Accordingly, the topic of sufficient graft volume remains an issue of controversy in LDLT; in particular, the factors that affect graft survival after LDLT with an SFSG with a GRWR < 0.8% have yet to be identified. Consequently, we compared the outcomes of adult-to-adult LDLT recipients who received right lobe grafts with a GRWR < 0.8% to the outcomes of recipients who received right lobe grafts with a GRWR ≥ 0.8%. Moreover, we attempted to identify the risk factors that affect graft survival after adult-to-adult LDLT with SFSGs.
PATIENTS AND METHODS
Between June 1997 and April 2008, 427 patients underwent LDLT with right lobe grafts at the Department of Surgery of Samsung Medical Center (Sungkyunkwan University School of Medicine). The analysis of risk factors of graft survival included recipient factors [age, sex, Model for End-Stage Liver Disease (MELD) score, Child-Pugh score, and indication for LDLT], donor factors (age, sex, GRWR, and fatty liver), operative factors [cold ischemic time, warm ischemic time, operative time, middle hepatic vein (MHV) drainage, and period of LDLT], and postoperative complications (postoperative bleeding, vascular complications, biliary complications, rejection, and SFSS). The era of LDLT was divided into 2 periods: 1997-2002 and 2003-2008. Information was obtained retrospectively from medical records. Recipients were divided into 2 groups: group A with a GRWR < 0.8% (n = 35) and group B with a GRWR ≥ 0.8% (n = 392).
Intraoperative ultrasonography was performed to confirm adequate hepatic venous anatomy before donor hepatectomy. Donor hepatectomy was carried out with a Cavitron ultrasonic surgical aspirator (Valleylab, Boulder, CO) and by bipolar electrocautery (Codman, Raynham, MA) with the hanging maneuver. After donor hepatectomy, grafts were flushed with 4 L of iced University of Wisconsin solution or histidine tryptophan ketoglutarate solution. Actual graft weights were measured after flushing. MHV reconstruction was performed with a cryopreserved iliac artery or iliac vein on the back table when the size of the MHV branch was >5 mm or when a lot of blood gushed out during flushing with perfusion solution. After total recipient hepatectomy, grafts were transplanted in a piggyback fashion. The orifice of the recipient right hepatic vein was enlarged with a downward incision, and anterior and posterior wall excisions were made to form an oval orifice to obtain sufficient outflow. After right hepatic vein anastomosis, any significant right inferior hepatic vein was anastomosed to the inferior vena cava in an end-to-side fashion. Portal vein anastomosis was performed with 6-0 Prolene continuous sutures (Ethicon, Somerville, NJ) with growth factor. After portal vein anastomosis, arterial reconstruction was performed with 8-0 Ethilon interrupted sutures (Ethilon, Livingston, United Kingdom) under a surgical microscope. Bile duct reconstruction was performed by either duct-to-duct anastomosis (n = 395) or Roux-en-Y hepaticojejunostomy (n = 32).
GRWR, graft-to-recipient weight ratio; INR, international normalized ratio; LDLT, living donor liver transplantation; MELD, Model for End-Stage Liver Disease; MHV, middle hepatic vein; PT, prothrombin time; SFSD, small-for-size dysfunction; SFSG, small-for-size graft; SFSS, small-for-size syndrome.
Small-for-Size Dysfunction (SFSD)
As suggested by Dahm et al.,6 SFSD was defined as the presence of 2 of the following on 3 consecutive days in the absence of any other cause: a bilirubin level > 100 μmol/L, a prothrombin time international normalized ratio (PT-INR) > 2, and encephalopathy (grade 3 or 4). Other causes of graft dysfunction were categorized as technical (arterial or portal occlusion, outflow congestion, and bile leaks), immunological (rejection), or infection factors (cholangitis and sepsis).6
Data are expressed as means and standard deviations. Analysis was performed with the Student t test, the Mann-Whitney test, the χ2 test, the log-rank test, and Cox's regression test. The Kaplan-Meier method and Cox's regression test were used to calculate graft survival probabilities after LDLT. Graft survival rates in the 2 study groups were compared with the log-rank test. Multivariate analysis of risk factors of graft survival was performed with the Cox proportional hazards model. The Lausen and Schumacher method was used to evaluate the best cutoff value for donor age. Statistical significance was accepted for a P value < 0.05. Statistical analysis was carried out with Predictive Analytics Software version 17.0 (SPSS, Inc., Chicago, IL).
Demographic Findings for the Recipient, Donor, and Operative Factors
Donor, recipient, and operative LDLT data are summarized in Table 1.
Table 1. Clinical Features of Recipients of Grafts with a GRWR < 0.8% and Recipients of Grafts with a GRWR ≥ 0.8%
Total (n = 427)
Group A (n = 35)
Group B (n = 392)
NOTE: For group A, the GRWR was less than 0.8%; for group B, the GRWR was greater than or equal to 0.8%. SFSD was found for group A, and graft dysfunction was found for group B. Significance was defined as P < 0.05.
All recipients were male in group A (P < 0.001), and donors in group A had a higher female sex ratio [15/20 (57.1%) versus 270/122 (31.1%), P = 0.003], which reflected the fact that LDLT with an SFSG occurred more frequently when a female donor graft was used for a male recipient. The mean GRWR was found to be significantly lower in group A (0.74% ± 0.05% versus 1.13% ± 0.22%, P < 0.001). Seven recipients used grafts with a GRWR ≤ 0.7%, and the minimum GRWR was 0.63%. Group microvesicular and macrovesicular fatty changes were not significantly different. In terms of operative factors, no group differences were observed in the cold ischemic time, warm ischemic time, or operative time. SFSGs were used more frequently during the early period versus the later period (12.9% versus 7.0%, P = 0.080). Overall, MHV reconstruction was performed in 260 cases (60.9%), with no difference between group A (62.9%) and group B (60.7%).
Graft Survival Rates and Postoperative Complications
The cumulative 1-, 3-, and 5-year graft survival rates in groups A and B were 87.8%, 83.4%, and 74.1% and 90.7%, 84.5%, and 79.4%, respectively (P = 0.852; Fig. 1). Graft survival according to GRWR was not significant in group A (P = 0.669), group B (P = 0.721), or the whole group (P = 0.871). Postoperative complications after LDLT are shown in Table 1. No intergroup differences were observed in terms of postoperative bleeding, acute cellular rejection, hepatic artery thrombosis, biliary complication, or wound problems. SFSD occurred in 2 of 35 patients (5.7%) in group A and in 14 of 392 patients (3.6%) in group B (P = 0.522). Portal vein complications were more frequently observed in group A (8.6%) versus group B (1.8%, P = 0.011).
Risk Factors of Graft Survival After LDLT with SFSGs
Factors related to graft survival with SFSGs were further analyzed. Patient deaths unrelated to graft function [pneumonia (1), sudden cardiac arrest (1), and hepatocellular carcinoma recurrence (2)] were excluded from analysis. Graft failure occurred in 6 patients [hepatic artery thrombosis (1) at 1 month, hepatitis B virus recurrence (1) at 8 months, biliary complications (1) at 52 months, and graft dysfunction (3) at 0-26 months]. Variables found to be associated with long-term graft survival by univariate analysis are shown in Table 2. Donor age was a significant risk factor for graft survival by univariate analysis in group A (P = 0.042). Furthermore, the 5-year graft survival rates of recipients with MHV tributary drainage were significantly higher than those of recipients without MHV tributary drainage in group A (95.5% versus 53.8%, P = 0.038; Fig. 2).The graft survival rates of recipients without SFSD tended to be higher than those of recipients with SFSD in group A (85.3% versus 50.0%, P = 0.074). Patients in the late period (2003-2008) tended to have better graft survival than those in the early period (5-year graft survival, 90.8% versus 61.4%, P = 0.182). These risk factors were analyzed for multivariate analysis, and only donor age was found to be a significant factor (P = 0.033, hazard ratio = 1.102, confidence interval = 1.008-1.205; Table 3). A Kaplan-Meier analysis of graft survival by donor age showed significant differences at all ages between 37 and 45 years (37-40 years old, P = 0.031; 41-42 years old, P = 0.02; 43 years old, P = 0.012; 44 years old, P = 0.005; and 45 years old, P = 0.045). Although we failed to obtain a statistically significant cutoff value after correction by the Lausen and Schumacher method, the best cutoff value was found at the age of 44 years (P = 0.0565). Three-year graft survival was 92.2% with grafts from donors younger than 44 years versus 53.6% with grafts from donors 44 years old or older (Fig. 3).
Table 2. Risk Factors of Graft Survival After LDLT with SFSGs: Results of Univariate Analysis
NOTE: For group A, the GRWR was less than 0.8%. Graft survival rates were estimated by the Kaplan-Meier method, and the log-rank test was used to compare categorical variables. Analysis for continuous variables was performed with Cox's regression model; significance was accepted for P values < 0.05.
MELD score (<30/≥30)
Child-Pugh score (A/B&C)
Cold ischemic time (minutes)
Warm ischemic time (minutes)
Operative time (minutes)
MHV tributary drainage (yes/no)
Period of LDLT (1997-2002/2003-2008)
Table 3. Risk Factors of Graft Survival After LDLT with SFGSs: Results of Multivariate Analysis
Risk Ratio of Death
NOTE: The P value was obtained with Cox's regression model. Initially, 4 variables (donor age, MHV tributary drainage, period of LDLT, and SFSD) were included in this model, and finally only donor age remained as a significant variable for group A.
Donor age (years)
We compared the outcomes of adult-to-adult LDLT recipients who received grafts with GRWR values < 0.8% and those who received grafts with GRWR values ≥ 0.8%. The cumulative graft survival rates 1, 3, and 5 years after the operation were 87.8%, 83.4%, and 74.1%, respectively, for group A and 90.7%, 84.5%, and 79.4%, respectively, for group B (Fig. 1.). Some authors in Western countries have reported that rates of survival after LDLT with a right liver graft range from 83% to 88%.9, 10 These rates are similar to our findings with SFSGs with a GRWR < 0.8%, and this indicates that SFSGs with a GRWR < 0.8% can be safely used for LDLT.
Ben-Haim et al.8 reported that a small graft could be used safely in patients in good condition before liver transplantation, such as patients without cirrhosis or Child class A patients. However, in the present study, patients with SFSGs had a relatively high MELD score of 24 (Table 1), and patients with a MELD score ≥ 30 had a survival rate similar to that of patients with a MELD score < 30 even with a GRWR < 0.8% (Table 2). We therefore think that a high MELD score should not be a criterion for discarding SFSGs. These findings have been reported previously. Yi et al.11 reported that a high MELD score (> 25) did not predict the 1-year survival rate of hepatitis B virus–infected recipients after LDLT, and it was also not an important predictor of the 1-year survival rate in cases with SFSGs.
Instead, donor age affected graft survival when a graft with a GRWR < 0.8% was used. The advantages of having a young donor have been previously mentioned.12, 13 Furthermore, some have reported that partial grafts from elderly donors show delayed regenerative activity,14, 15 and others have found better outcomes with younger donors in deceased donor liver transplantation, especially for recipients with hepatitis C–related liver disease.16, 17 The liver is capable of completely regenerating itself in response to injury. Stocker and Heine18 reported that in young rats after two-thirds partial hepatectomy, over 99% of hepatocytes were involved in proliferation in the remnant liver, whereas in older rats, only 30% of hepatocytes were involved. In the present study, donor age was a significant risk factor for graft survival by univariate analysis within group A (P = 0.042) and was found to be the only significant risk factor for poor graft survival by multivariate analysis (P = 0.033, hazard ratio = 1.102, confidence interval = 1.008-1.205). Although we failed to obtain a cutoff value with statistical significance after correction by the Lausen and Schumacher method, the cutoff at the age of 44 years old showed the best value (P = 0.0565). Patients receiving grafts from donors younger than 44 years had a 3-year survival rate of 92.2% versus 53.6% for patients receiving grafts from donors 44 years old or older (P = 0.005; Fig. 3). We hope that further studies will clarify the best cutoff value of donor age for SFSGs.
Hepatic venous reconstruction is critical for SFSGs because a congested segment may hamper the regenerative process and lead to graft loss.19-21 Fan et al.22 reported that postoperative hepatic function in recipients with an occluded or absent MHV was poorer than that in recipients with a patent MHV, despite similar GRWR values. Kamei et al.23 recently reported that the noncongested volume of a right liver graft is associated with early graft function and noncongestive GRWR (<0.65%) and that it can play a key role in preoperative decision making regarding the need for additional vein reconstruction for MHV tributaries. At our center, reconstruction of the MHV was based on the size of the MHV branch (>5 mm) or the gushing of blood during flushing with perfusion solution without reference to the congested area. However, because patients who did not meet the criteria and had no MHV reconstruction still had poorer outcomes, we believe that more aggressive reconstruction of MHV tributaries, such as the use of the extended right liver, may yield a better outcome, especially when a recipient receives a graft with a GRWR < 0.8%.
Some have described the utility of larger graft sizes in patients with high MELD scores.24, 25 Yoshizumi et al.15 reported that a high MELD score did not affect early graft function (represented by the total bilirubin level, volume of ascites, or PT-INR). However, graft and patient survival tended to be poorer in recipients with high MELD scores. In the present study, no graft survival difference was found between patients with high MELD scores (≥30) and patients with low MELD scores in group A, but we attributed this to the small sample size (n = 35).
Portal vein complications have been shown to be a risk in SFSGs. Kyoden et al.26 reported that portal vein complications were identified in 9% of 310 patients with adult-to-adult LDLT. In comparison with the general population of LDLT recipients, smaller graft size, preexisting portal vein thrombosis, and use of either jump or interposition grafts for portal vein reconstruction were associated with a higher risk of portal vein complications.26-28 Our results support the previous findings. We speculate that SFSGs could hamper the outflow of the portal vein and result in decreased portal blood flow, and anastomosing the portal vein to a smaller graft often results in a longer portal vein and a smaller portal vein diameter, which may facilitate the generation of strictures. We have not evaluated the portal vein pressure or the portal vein flow rate in our study and cannot draw any conclusion; further study will be needed to elucidate these findings.
In conclusion, graft survival after LDLT using a right graft with a GRWR < 0.8% was excellent. Graft survival with SFSGs appeared to depend on both the donor age and the MHV tributary drainage, but only the donor age was significant in multivariate analysis. We conclude that adult-to-adult LDLT using a right lobe with a GRWR < 0.8% can be performed safely when grafts from young donors less than 44 years old are used.
The authors thank Soo-Young Woo, M.S. (Biostatics Unit, Samsung Biomedical Research Institute), for his help with the statistical analysis.