Adult-to adult right-lobe living donor liver transplantation (RL-LDLT) has gained widespread acceptance during the last 10 years as an alternative to deceased donor liver transplantation. Although living donor (LD) grafts are always from high-quality donors and are less likely to suffer from reperfusion injury, there is an ongoing concern that the smaller volume of the LD graft in comparison with a full-size deceased donor graft might have a deleterious effect on patient outcomes.
The evidence defining the critical minimum volume required to perform safe living donor liver transplantation (LDLT) is weak. Most transplant centers in North America currently have an arbitrary requirement for a graft to body weight ratio (GBWR) ≥ 0.8 for RL-LDLT or, alternatively, a graft weight (GW)/standard liver volume (SLV) ratio ≥ 40%.1–3 This standard of practice was adopted on the basis of the poor outcomes reported when smaller grafts were transplanted during the early experience with LDLT at Asian centers.4, 5 It is noteworthy, however, that later reports from other Asian centers have failed to show a significant correlation between liver graft volume and postoperative patient outcomes.6–8
We have cautiously transplanted smaller volume LD liver grafts at our center. Here we analyze whether, in our experience, using grafts with a GBWR as low as 0.6 in selected patients had a deleterious effect on patient outcomes in comparison with the use of larger volume LD grafts or the use of full-size deceased donor liver grafts in matched recipients.
ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; CIT, cold ischemia time; eGFR, estimated glomerular filtration rate; GBWR, graft to body weight ratio; GW, graft weight; HCC, hepatocellular carcinoma; INR, international normalized ratio; LD, living donor; LDLT, living donor liver transplantation; MELD, Model for End-Stage Liver Disease; MHV, middle hepatic vein; NA, not available; RL-LDLT, right-lobe living donor liver transplantation; SLV, standard liver volume; Tx, transplantation; WIT, warm ischemia time.
PATIENTS AND METHODS
We used a prospective database of all adult-to-adult LDLT and deceased donor liver transplantation procedures performed at our institution from April 2000 to September 2008 for this analysis. The study was approved by the University Health Network's Research Ethics Board. Living and deceased donation was voluntary and altruistic in all cases.
To be considered for LDLT, recipients had to be listed for deceased donor liver transplantation at our center. For adult-to-adult transplants, we exclusively transplanted right-lobe grafts. All potential donors with a GBWR as low as 0.6 were evaluated as long as the residual donor liver volume was more than 30%. Recipients receiving smaller size grafts (GBWR < 0.8) were informed that smaller size grafts might have a higher rate of failure; the opportunity for LDLT never had to be declined because of recipient concerns about the size of the graft. Other details of our institutional donor evaluation process, surgical techniques, and donor outcomes have been described in previous reports.9, 10
LDLT cases were divided into 2 groups: group A, which consisted of patients with a GBWR < 0.8, and group B, which consisted of patients with a GBWR ≥ 0.8. In addition, patients in groups A and B were compared to a matched group of recipients who received full-size deceased donor liver grafts and for whom the donor age, recipient age, medical Model for End-Stage Liver Disease (MELD) score, presence of hepatitis C, and presence of hepatocellular carcinoma were similar (group C).
Assessment of Transplant Outcomes
Short-term outcomes were assessed by 30-day mortality and graft loss; biochemical markers of hepatocyte injury [aspartate aminotransferase (AST) and alanine aminotransferase (ALT)] and liver function [absolute decrease in the international normalized ratio (INR) and bilirubin within the first 7 days after surgery]; length of hospital stay; and postoperative complications within 3 months, such as the incidence of infections, biliary complications, acute cellular rejection, and renal failure. Postoperative complications were graded according to the Clavien score.11 The estimated glomerular filtration rate (eGFR) was calculated by the Modification of Diet in Renal Disease formula12 and used as a marker of renal function in the first week post-transplantation. Long-term transplant outcome was measured by actuarial graft and patient survival at 1, 3, and 5 years.
Hyperbilirubinemia and the presence of ascites were used to identify patients with prolonged graft dysfunction. Graft dysfunction was defined as a bilirubin level ≥ 60 μmol/L and the presence of moderate (defined as clinically detectable) ascites despite diuretic therapy 4 weeks after LDLT. Patients with any identifiable cause of graft dysfunction, such as biliary or vascular complications, rejection, or early hepatitis C recurrence, were excluded. Patients fulfilling these criteria were labeled as presenting with a small-for-size–like syndrome independent of the actual GBWR.
Statistical analysis was performed with SPSS for Windows 11.0 (SPSS, Chicago, IL). A chi-square test or Fisher's exact test was used for categorical variables, whereas analysis of variance was performed for continuous parametric variables with a post hoc Bonferroni correction for multiple comparisons. Graft survival and patient survival were calculated with the Kaplan-Meier survival analysis and compared with the log rank test. Potential predictors of graft survival were determined with a Cox regression analysis. A P value below 0.05 was considered significant.
Donor Demographics and Outcomes
Seven hundred fifty-four potential donors were evaluated to perform 271 LDLT procedures. Fifty-one percent of donors were male with a median donor age of 37 years (18–60 years) and a mean donor body mass index of 25 (range = 17–38). The mean duration of the LD operation was 7.3 ± 0.8 hours with a mean blood loss of 946 ± 471 cc. Only 7 patients (2.6%) required allogeneic blood transfusions perioperatively. The median postoperative length of hospital stay was 7 days (range = 4–17 days). The 3-month perioperative complication rate was 24%, and no deaths occurred in the donor population; grading the severity of the complications with the Clavien score, we found that all donor complications were less than or equal to grade 3, and grade 4 or 5 complications did not occur. Per the study design, the graft weight was significantly lower in the small LD group (group A) versus the large LD group (group B; 573 ± 298 versus 932 ± 150 g, P = 0.0001), but the other donor characteristics were similar (Table 1).
Table 1. Donor Characteristics for Living Donor Liver Transplantation Cases with a Small GBWR or a Large GBWR and Matched Deceased Donor Liver Transplantation Cases
Matched Deceased Donors
GBWR < 0.8
GBWR ≥ 0.8
Abbreviations: BMI, body mass index; CIT, cold ischemia time; GBWR, graft to body weight ratio; GW, graft weight; MHV, middle hepatic vein; NA, not available; SLV, standard liver volume; WIT, warm ischemia time.
0.72 ± 0.63
1.2 ± 0.3
36 ± 4.7
59 ± 14.1
Donor age (years)
39 ± 15
36 ± 11
39 ± 14
Donor BMI (kg/m2)
25 ± 3.9
26 ± 5.9
26 ± 3.9
Donor gender: male
>2 bile ducts
MHV with graft
94 ± 43
96 ± 57
453 ± 152
56 ± 12
54 ± 16
54 ± 18
Comparison of GBWR and GW/SLV
Both GBWR and GW/SLV13 have been used to estimate the required graft volume for LDLT. There was a strong correlation (Pearson correlation = 0.98, P < 0.0001) between these ratios, and thus we elected to use only GBWR for the analyses.
Between April 2000 and September 2008, 271 adult-to-adult RL-LDLT procedures and 614 cadaveric liver transplants were performed. Twenty-two cases (group A) had a GBWR < 0.8, whereas in 249 LD transplants (group B), the GBWR was ≥0.8. Within group A, 9 patients had a GBWR between 0.75 and 0.79; in 9 patients, the GBWR was ≥0.65 and <0.75; and in 4 cases, the GBWR was <0.65 with a minimal GBWR of 0.59. The rate of splenectomy during LDLT or within the first postoperative week as a measure for reducing portal flow was higher in the recipients with smaller grafts [2/22 (9%) for group A] versus the recipients with larger grafts [6/249 (2.4%) for group B, P = 0.11]. No splenectomy was performed in the 66 matched patients receiving full-size deceased donor grafts (group C).
Groups A, B, and C had similar characteristics, including recipient age (49 ± 9 for A, 51 ± 10 for B, and 49 ± 9 for C, P = 0.4), MELD score (22 ± 11 for A, 19 ± 8 for B, and 22± 19 for C, P = 0.1), warm ischemia time (56 ± 19 minutes for A, 54 ± 18 minutes for B, and 54 ± 16 minutes for C, P = 0.8), presence of hepatitis C (46% for A, 42% for B, and 48% for C, P = 0.4), and presence of hepatocellular carcinoma (31% for A, 26% for B, and 28% for C, P = 0.78) (Table 2). LD recipients had a significantly shorter cold ischemia time (94 ± 43 minutes for A, 96 ± 57 minutes for B, and 453 ± 152 minutes for C, P = 0.0001). The rate of implantation of the middle hepatic vein was similar in groups A (53%) and B (58%), as was the rate of the requirement for 2 or more biliary anastomoses (31% for A and 32% for B; Table 1).
Table 2. Recipient Characteristics for Living Donor Liver Transplantation Cases with a Small GBWR or a Large GBWR and Matched Deceased Donor Liver Transplantation Cases
Matched Deceased Donors
GBWR < 0.8
Abbreviations: GBWR, graft to body weight ratio; HCC, hepatocellular carcinoma; MELD, Model for End-Stage Liver Disease.
49 ± 9.9
51 ± 10
49 ± 9.8
Recipient gender male
22 ± 11
19 ± 8
22 ± 9
Impact of GBWR on the Short-Term Outcomes After RL-LDLT
Thirty-day graft loss rates (4.5% for A versus 3.6% for B, P = 0.57) and patient death rates (4.5% for A versus 3.2% B, P = 0.5) were similar in LD recipients with small and large GBWRs. Maximum AST (481 ± 295 U/L for A versus 545 ± 482 U/L for B, P = 0.55) and ALT (371 ± 263 U/L for A versus 445 ± 398 U/L for B, P = 0.4) within 48 hours post-transplantation as absolute values were similar between LD recipients with small and large GBWRs. Then, we determined graft function by evaluating the changes in INR and bilirubin within the first week after transplantation. LD recipients with a small GBWR versus LD recipients with a large GBWR tended to have a higher pretransplant bilirubin level (127 ± 130 μmol/L for A versus 80 ± 88 μmol/L for B, P = 0.06) and INR (1.85 ± 1.23 μmol/L for A versus 1.6 ± 1 μmol/L for B, P = 0.48). At day 7, the bilirubin decrease was −34 ± 188 μmol/L in small-GBWR LD recipients versus −7 ± 120 μmol/L in large-GBWR LD recipients (P = 0.34). No difference was observed in the decrease in INR between small-GBWR LD recipients and large-GBWR LD recipients at day 7 (−0.77 ± 1.2 for A versus −0.52 ± 1.08 for B, P = 0.3; Table 3).
Table 3. Postoperative Biochemical and Clinical Parameters of Outcome
Matched Deceased Donors
GBWR < 0.8
NOTE: Infection, bleeding, renal failure, and Clavien score refer to the 3-month postoperative period.
Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; eGFR, estimated glomerular filtration rate; GBWR, graft to body weight ratio; INR, international normalized ratio; Tx, transplantation.
481 ± 295
595 ± 482
1698 ± 1781
371 ± 263
445 ± 398
951 ± 806
Bilirubin before Tx (μmol/L)
127 ± 137
80 ± 88
120 ± 164
Change in bilirubin in the first week (mmol/L)
−34 ± 188
−7 ± 126
−57 ± 164
INR before Tx
1.85 ± 1.23
1.6 ± 1
2.1 ± 1.9
Change in INR in the first week
−0.77 ± 1.2
−0.52 ± 1.08
−0.9 ± 1.9
eGFR before Tx
83 ± 50
92 ± 48
93 ± 53
Change in eGFR in the first week
12 ± 48
−2 ± 43
−15 ± 54
Length of stay [median (range)]
0.95 ± 1.4
1.2 ± 1.4
1.24 ± 1.2
30-day graft loss
The pretransplant eGFR was not different in the small-GBWR and large-GBWR groups (83 ± 50 versus 92 ± 48 mL/minute/1.73 m2, P = 0.45). The change in eGFR in the first postoperative week was similar in LD recipients with small and large grafts (+12 ± 48 versus −2 ± 43 mL/minute/1.73 m2, P = 0.1).
Next, we determined the frequency of postoperative complications within 3 months between small-GBWR and large-GBWR LD recipients. No difference was observed with respect to the incidence of renal failure (defined as the need for posttransplant dialysis or a creatinine level > 300 μmol/L; 9% for A versus 5% for B, P = 0.32), postoperative bleeding (defined as relaparotomy for hemostasis; 0% for A versus 2% for B, P = 0.54), or the overall infection rate (22% for A versus 47% for B, P = 0.06). The median length of hospital stay was comparable in the 2 groups [14 days (range = 1–109 days) for A versus 15 days (range = 3–206 days) for B]. Grading the severity of the complications by the required intervention (Clavien score) did not reveal a difference between small-GBWR and large-GBWR LD recipients (0.95 ± 1.4 for A versus 1.2 ± 1.4 for B, P = 0.29; Table 3).
Then, we investigated the outcome of LD recipients with a complication > grade 3. Complications > grade 3 occurred in the same frequency in recipients of small LD grafts and recipients of large LD grafts (14% versus 22%, P = 0.52). Similarly, the 1-month patient survival (98% versus 95%, P = 0.7) and the median length of hospital stay [21 days (range = 7–45 days) versus 19 days (range = 2–171 days), P = 0.6] were not different in small-graft recipients (group A) and large-graft recipients (group B) with a complication > grade 3.
Small-for-size–like syndrome was diagnosed in 8 of 271 LD recipients (3%). In the small-GBWR group (A), 2 patients (9%) developed small-for-size syndrome (GBWRs of 0.79 and 0.7), whereas 6 recipients (2.5%) of larger LD grafts (B) were diagnosed with small-for-size syndrome (GBWRs of 0.84, 0.86, 0.97, 1.28, 1.3, and 2.3, P = 0.13). GBWR as a continuous and categorical variable (≥0.8 versus <0.8), MELD score, cold and warm ischemia times, donor age, and recipient age were investigated as risk factors for small-for-size syndrome in univariate analysis. Only MELD score was a significant risk factor for small-for-size syndrome, with an average MELD score of 26 versus 19 in small-for-size patients versus non–small-for-size patients (P = 0.006). GBWR as a continuous variable (P = 0.55) or categorical variable (0.13) did not predict the occurrence of small-for-size syndrome. Overall, patients with small-for-size–like syndrome had a 1-year mortality rate of 50% in contrast to a rate of 10% in LD recipients without a small-for-size condition (P = 0.008).
Comparison of the Results of Small-GBWR LD Graft and Full-Size Deceased Donor Graft Transplantation
Small-GBWR LD recipients (group A) and full-size deceased donor graft recipients (group C) had similar 1-month graft loss (4.5% versus 3%, P = 0.58) and mortality rates (4.5% versus 0%, P = 0.25). AST (481 ± 295 U/L for A versus 1698 ± 1781 U/L for C, P = 0.003) and ALT (371 ± 263 U/L for A versus 951 ± 806 U/L for C, P = 0.002) as markers of hepatocyte injury were significantly higher in the deceased donor group. Pretransplant bilirubin (127 ± 130 μmol/L for A versus 120 ± 164 μmol/L for C, P = 0.49) and INR (1.8 ± 1.2 for A versus 2.1 ± 1.9 for C, P = 0.88) were similar between small-GBWR LD and deceased donor recipients. The small-GBWR LD recipients and deceased donor recipients had similar decreases in bilirubin levels in the first postoperative week (−34 ± 188 μmol/L for A versus −57 ± 164 μmol/L for C, P = 0.59). There was no difference with respect to the decrease in INR in the first postoperative week between small-graft and large-graft LD recipients (−0.77 ± 1.2 versus −0.9 ± 1.9, P = 0.72).
Pretransplant eGFR was similar in small-GBWR LD and deceased donor recipients (83 ± 53 mL/minute/1.73 m2 for A versus 93 ± 53 mL/minute/1.73 m2 for C, P = 0.45). In the first postoperative week, eGFR increased in small-graft LD recipients (+11 ± 44 mL/minute/1.73 m2) and decreased in the deceased donor group (−15 ± 54 mL/minute/1.73 m2, P = 0.03).
Rates of renal failure (9% for A versus 15% for C, P = 0.37), postoperative bleeding (0% for A versus 7.6% for C, P = 0.22), and overall infection (27% for A versus 33% for C, P = 0.4) were similar in the small-GBWR group (A) and full-graft group (C). The median length of hospital stay was 14 days (range = 1–109 days) in the small-GBWR group versus 18 days (range = 7–223 days) for the deceased donor recipients (P = 0.31). The 2 groups had complications of similar severity as graded by the Clavien score (0.95 ± 1.4 for A versus 1.24 ± 1.2 for C, P = 0.38).
Impact of GBWR on Graft and Patient Survival
Graft survival and patient survival were analyzed after 1, 3, and 5 years in LDLT recipients with a GBWR < 0.8 (A) or ≥ 0.8 (B) and in matched deceased donor liver transplantation recipients (C). No difference was observed with respect to graft survival between the small LD recipients (A), the large LD recipients (B), and the deceased donor group (C) after 1 year (89% for A, 91% for B, and 92% for C), 3 years (81% for A, 86% for B, and 82% for C), and 5 years (81% for A, 79% for B, and 80% for C, P = 0.92; Fig. 1). Similarly, patient survival was identical between small LD recipients (A), large LD recipients (B), and deceased donor recipients (C) at 1 year (89% for A, 91% for B, and 93% for C), 3 years (81% for A, 87% for B, and 89% for C), and 5 years (81% for A, 83% for B, and 87% for C, P = 0.63; Fig. 2).
A univariate Cox proportional regression analysis was performed to determine risk factors for patient survival. GBWR as a continuous variable, GBWR < 0.8 versus GBWR ≥ 0.8 as a categorical variable, donor and recipient age, cold and warm ischemia times, recipient medical MELD score, donor body mass index, and presence of hepatitis C or hepatocellular carcinoma were investigated. Only the presence of hepatitis C was significantly associated with decreased patient survival (P = 0.014, hazard ratio = 2.6, confidence interval = 1.2–5.6). Neither GBWR as a continuous variable (P = 0.2, hazard ratio = 1.9, confidence interval = 0.7–5.3) nor GBWR < 0.8 versus GBWR ≥ 0.8 as a categorical variable (P = 0.83, hazard ratio = 1.1, confidence interval = 0.2–4.8) were associated with patient survival.
This is the first Western series investigating the impact of graft volume on the outcome of adult LDLT. Two important observations were made. First, when graft size was analyzed as a continuous and discontinuous variable, we found no evidence of inferior outcomes with smaller size grafts versus larger size LD grafts or full-size deceased donor grafts. Thus, when a recipient has several potential LDs available, as long as the graft size is within the accepted clinical range, graft size should not influence donor selection, nor should it be a factor influencing the choice between RL-LDLT and deceased donor liver transplantation. Second, even when GBWR was below the accepted limit of 0.8, a low donor liver volume did not have a deleterious effect on the outcomes of LDLT. Taken together, these data suggest that further studies are warranted to cautiously explore the safe lower limits of GBWR for LDLT.
The current practice of arbitrarily requiring a GBWR ≥ 0.8 for RL-LDLT is based on the experience in early Asian studies.4, 8, 14 The present study suggests that it may be inappropriate to use the data from these studies to define current policy for transplantation in North America. First, the relationship between liver volume and body weight may be different in Asian and Western populations, as evidenced by the different formulas that have been proposed to calculate the SLV by Western13 and Asian15, 16 authors. Second, the landmark Asian studies used to establish the current limits for LD were based on the early surgical experience with LDLT, and these outcomes may not apply to the modern era.4–6, 8, 17, 18 Third, the inclusion of both adult and pediatric recipients as well as left-lobe and right-lobe grafts may have confounded the analyses of these early experiences. Fourth, LDLT was often the only option for transplantation in these early Asian experiences, whereas in Europe and North America, patients needing a liver transplant also have the option of waiting for a full-size deceased liver graft.
In our series, the outcomes with smaller GBWR LD transplants were not only comparable to the large-GBWR LD group; they were also comparable to the outcomes with full-graft deceased donor grafts. There are many reasons why a partial LD liver graft might perform as well or better than a whole organ deceased donor liver graft. First, LDs must be healthy with no liver disease and no major comorbidities. Second, LD grafts have a brief storage time and appear to be less likely to suffer preservation damage; in the current study, the average storage time for a deceased donor liver was almost 4 times longer than that for an LD liver graft. Third, deceased donors are subjected to many stresses that can have a deleterious effect on liver function, such as brain death, a prolonged intensive care unit stay, treatment with high-dose vasopressors, hypotensive episodes, and severe electrolyte shifts.
Several aspects of our protocol for RL-LDLT may have contributed to the excellent function of the lower GBWR grafts transplanted at our center. Regardless of the GBWR, we take steps to ensure that every LD graft has good outflow venous drainage of segments 5 and 8: Thus, in half of the right-lobe grafts transplanted at our center, we retrieved and reconstructed the graft's middle hepatic vein. All of our RL-LDLT recipients receive antibody induction therapy in order to reduce the risk of early graft injury due to rejection while the LD graft is regenerating. Other factors that may have contributed to the favorable outcome of small grafts in this series, which are difficult to measure, include the experience acquired with the relatively high volumes of our LDLT program and unidentified recipient selection biases that did not become apparent in our analysis.
It must be emphasized that we are not advocating the indiscriminant use of small-size LD grafts. Small-for-size syndrome is a very real, life-threatening entity characterized by prolonged postoperative hyperbilirubinemia, coagulopathy, and ascites.14, 19, 20 Several possible mechanisms for the development of small-for-size syndrome have been proposed, including portal hyperperfusion due to a mismatch of portal flow and graft volume, venous outflow congestion, the presence of portosystemic shunting, donor age, and parenchymal factors such a steatosis.19, 21–23 It is possible that the balance between portal vein inflow, hepatic vein outflow, and functional liver mass determines the development of small-for-size syndrome, rather than liver volume itself, and thus the term small-for-size might be misleading. In a mouse model, even a 90% hepatectomy is still associated with only moderate liver injury and represents a survival condition if the hepatic outflow is preserved during surgery.24 However, if a 90% hepatectomy is combined with partial occlusion of the venous outflow, most of the mice will die of small-for-size liver failure syndrome. In our series, the MELD score but not GBWR was a predictive factor for small-for-size–like syndrome. Interestingly, 6 of 8 LD recipients developing small-for-size syndrome in our series had a GBWR > 0.8. These data support the notion that factors other than graft size significantly contribute to the development of small-for-size syndrome.
A splenectomy25 or permanent26–29 or temporary30 portocaval shunt has been proposed to decrease portal flow in patients who are thought to be at risk of developing small-for-size syndrome. Only a small proportion of the LD recipients at our center underwent splenectomy for graft congestion; no portal shunts were performed. Portal shunts do have the potential risk of shunting blood away from the graft; Oura et al.31 reported shrinkage of the LD graft after a portocaval shunt was used that was cured by closure of the shunt.
We conclude that our understanding of the minimum graft size required for safe LDLT is both flawed and incomplete. Our data suggest that a GBWR ≥ 0.8 is not always necessary to achieve excellent outcomes. Graft size alone should not affect clinical decisions about donor selection or whether it is preferable to use LD grafts or deceased donor grafts. In selected patients, it appears reasonable to use donor livers with a GBWR slightly below the standard range as we strive to better understand all the factors contributing to the success or failure of smaller LD liver grafts.