Telephone: 416-340-5230; FAX: 416-340-5242
Living donor hepatectomy: The importance of the residual liver volume
Article first published online: 29 NOV 2011
Copyright © 2011 American Association for the Study of Liver Diseases
Volume 17, Issue 12, pages 1404–1411, December 2011
How to Cite
Reichman, T. W., Sandroussi, C., Azouz, S. M., Adcock, L., Cattral, M. S., McGilvray, I. D., Greig, P. D., Ghanekar, A., Selzner, M., Levy, G. and Grant, D. R. (2011), Living donor hepatectomy: The importance of the residual liver volume. Liver Transpl, 17: 1404–1411. doi: 10.1002/lt.22420
- Issue published online: 29 NOV 2011
- Article first published online: 29 NOV 2011
- Accepted manuscript online: 17 AUG 2011 01:44PM EST
- Manuscript Accepted: 1 AUG 2011
- Manuscript Received: 31 DEC 2010
Living liver donation is a successful treatment for patients with end-stage liver disease. Most adults are provided with a right lobe graft to ensure a generous recipient liver volume. Some centers are re-exploring the use of smaller left lobe grafts to potentially reduce the donor risk. However, the evidence showing that the donor risk is lower with left lobe donation is inconsistent, and most previous studies have been limited by potential learning curve effects, small sample sizes, or poorly matched comparison groups. To address these deficiencies, we conducted a case-control study. Forty-five consecutive patients who underwent left hepatectomy (LH; n = 4) or left lateral segmentectomy (LLS; n = 41) were compared with matched controls who underwent right hepatectomy (RH) or extended right hepatectomy (ERH). The overall complication rates of the 3 groups were similar (31%-37%). There were no grade 4 or 5 complications. There were more grade 3 complications for the RH patients (13.3%) and the ERH patients (15.6%) versus the LH/LLS patients (2.2%). The extent of the liver resection significantly correlated with the peak international normalized ratio (INR), the days to INR normalization, and the peak bilirubin level. A univariate analysis demonstrated that hepatectomy, the spared volume percentage, and the peak bilirubin level were strongly associated with grade 3 complications. A higher peak bilirubin level, which correlated with a lower residual liver volume, was associated with grade 3 complications in a multivariate analysis (P = 0.005). RH and grade 3 complications were associated with an increased length of stay (>7 days) in a multivariate analysis. In conclusion, this analysis demonstrates a significant correlation between the residual liver volume and liver dysfunction, serious adverse postoperative events, and longer hospital stays. Donor safety should be the first priority of all living liver donor programs. We propose that the surgical procedure removing the smallest amount of the liver required to provide adequate recipient graft function should become the standard of care for living liver donation. Liver Transpl, 2011. © 2011 AASLD.
Living donor liver transplantation (LDLT) is a successful procedure with patient survival rates that are equivalent to those with deceased donor liver transplantation.1–4 In regions in which donor rates are low, LDLT reduces wait-list death rates and improves survival from the time of listing in comparison with deceased donor liver transplantation.5
Currently, right lobe liver grafts are procured for most adult recipients of living donor organs in North America, whereas left lobe grafts or left lateral segment grafts, which typically provide less liver mass, are reserved for infants and children. Recipient outcomes are excellent with right lobe LDLT, but donor morbidity rates range from 20% to 40%.6 The drive to identify safer alternatives has stimulated renewed interest in the use of left lateral segment grafts or left lobe grafts for adults, with inflow modulation of the portal vein often used to prevent small-for-size syndrome after transplantation.7
The renewed interest in using smaller living donor liver grafts in adults is based on the premise that left-sided liver resections are safer than right-sided liver resections. However, previous reports comparing right hepatectomy (RH) and left hepatectomy (LH) have yielded inconsistent results. Recently, a large multicenter study in Japan, for example, examined 3565 right and left lobe liver donations and reported similar morbidity rates after right lobe donation (9.4%) and left lobe donation (8.7%),8 whereas other Asian studies have demonstrated significant differences. Morioka et al.,9 for example, reported an increase in the number of major complications with right lobe grafts, but Chan et al.10 demonstrated a higher incidence with left lobe grafts. Single-center North American studies have generally reported lower complication rates with left lobe or left lateral segment donation versus right lobe donation, but many of these studies have not provided detailed descriptions of complications or documentation showing that the patient groups were similar.11, 12 Another common limitation of previous reports was the failure to control for significant differences in the ages of the comparison groups: left lateral segment donors were generally young and very healthy parents donating to a child, whereas right lobe donors were generally older adults donating to a sibling, friend, or partner.10, 13, 14 Given these conflicting results, we decided to examine the outcomes of living donation at the University of Toronto to assess the impact of the residual donor liver volume on complications. To reduce the risk of patient assignment biases when we were assessing the outcomes of the different donor procedures, we used for the first time a case-control study design to compare right lobe liver donation, left lobe liver donation, and left lateral segmentectomy (LLS).
PATIENTS AND METHODS
Living donor cases were prospectively entered into the University of Toronto/Toronto General Hospital LDLT database. Cases between June 1996 and May 2010 were studied. LLS cases (segments 2 and 3) and LH cases (segments 2-4) were identified first. Then, these patients were matched by age, sex, body mass index (BMI), and date of surgery to chosen individuals who had undergone RH (segments 5-8 minus the middle hepatic vein) or extended right hepatectomy (ERH; segments 5-8 plus the middle hepatic vein). Approval for this study was obtained from the Toronto General Hospital/University Health Network research ethics board (#10-0184-AE).
The donor workup has been reported elsewhere.6 Donors were routinely evaluated by 2 staff transplant surgeons, a social worker, a transplant hepatologist, a psychiatrist, and an anesthesiologist. Potential donors who were actively smoking cigarettes were required to stop smoking 4 weeks before donation because of the increased risk of thrombosis. Additional consultations were obtained when they were required to investigate other health issues. Liver imaging included ultrasound, computed tomography angiography, and magnetic resonance cholangiopancreatography. Computed tomography cholangiography was performed if the magnetic resonance cholangiopancreatography images were inadequate. Volumetric analyses were performed with Windows-based software (Optical, Spectroscopic, and Infrared Remote Imaging System).
The techniques used for RH and ERH have been described previously.15 Generally, the incision for left-sided liver procurement was shorter than the incision used for an RH donor. All right lobe donor hepatectomies were performed through a right subcostal incision with a midline extension up to the xiphoid. LH and LLS were performed through a midline incision if the body habitus permitted this, but the incision was extended along the right subcostal line if this was necessary for safety reasons.
Donor safety was a priority when we were selecting donors and making decisions about the technical aspects of the donor surgery. Only those donors who would retain at least 30% of their original liver volume were accepted. The decision to include the middle vein in the graft was based on the distribution of the venous drainage of segments 4, 5, and 8, the estimated final graft volume with and without the middle hepatic vein, and the estimated residual donor liver volume percentage with and without the middle hepatic vein. The Cavitron ultrasonic surgical aspirator was initially used to divide the liver parenchyma, but we switched to the Hydro-Jet (Erbe) and used it for more than 90% of the cases (43 of the 45 LH/LLS patients, 45 of the 45 RH patients, and 45 of the 45 ERH patients).
The donors were initially managed in an acute care (step-down) unit. After approximately 48 hours, most of the patients were transferred to a general transplant ward for the remainder of their hospitalization. Outpatient donor follow-up was performed at 1 month, 3 months, 12 months, and yearly thereafter for up to 10 years after donation to monitor for late complications.
Patient data were collected prospectively through the routine recording of the estimated blood loss, the autologous blood transfusion requirements, the graft weight, the predicted graft weight, the spared volume, the spared volume percentage, and the length of the hospital stay. The complications for each group were identified and were reported prospectively with the Clavien 5-tier scoring system.16
Continuous variables were compared with a 1-sided analysis of variance. Categorical values were compared with either a chi-square test or Fisher's exact test according to the sample size. The overall complication rates and the severity of the complications were compared with Fisher's exact test or a t test as appropriate. A bivariate logistical regression analysis was performed to identify statistical predictors of significant postoperative morbidity (Clavien grade 3 complications) and extended hospital stays. Factors thought to be clinically significant or factors that had a P value < 0.1 were included in the multivariate analysis.
Forty-five consecutive left-sided donors [LLS (41) and LH (4)] were identified from June 1996 to May 2010 (Table 1). These patients were matched on a 1:1 basis by age, sex, operation date, and BMI with selected patients who had undergone RH or ERH (Table 1). There were no significant differences between the 3 groups with respect to age, sex, or BMI, and this confirmed that the cohorts were well matched. The majority of the donors were biologically related to the recipients (87%). A history of smoking indicates donors who were actively smoking cigarettes at the time of their initial evaluation. This includes patients who, knowing the requirements of the program, quit smoking weeks before their initial evaluation. The program requires that all patients quit smoking 4 weeks before donation because of the increased risk of thrombosis. There was no statistical difference between the 3 groups with respect to the number of patients with a history of smoking. The mean follow-up time was 56.6 months (range = 3.4-168 months).
|Age (years)*||34.2 ± 8.6 (19.5-58.1)||35.3 ±10.3 (18.7-58.3)||34.6 ± 8.6 (20.1-56.4)||0.12†|
|Sex: male/female (n/n)||14/31||14/31||14/31||1‡|
|BMI (kg/m2)*||24.4 ± 4.1 (18-36)||24.7 ± 3.2 (19-31)||25.2 ± 3.4 (18-33)||0.53†|
|History of smoking [n/N (%)]||4/45 (9)||11/45 (22)||14/45 (31)||0.03‡|
|History of previous surgery [n/N (%)]||25/45 (56)||31/45 (69)||27/45 (58)||0.42‡|
|Relationship with recipient|
|Related [n/N (%)]||37/45 (82)||37/45 (87)||42/45 (93)||0.28‡|
|Unrelated [n/N (%)]||8/45 (18)||6/45 (13)||3/45 (7)|
Donor Liver Dysfunction/Injury
The predicted volumes correlated well with the measured graft weights (Table 2). The mean spared volume percentage with LH/LLS was 76.0%, which was more than twice the residual volume percentage after RH or ERH. The right-sided donor operations (RH and ERH) were associated with higher estimated blood losses, which were almost twice the losses with LH/LLS (P < 0.001). A longer length of stay was noted for the RH and ERH groups versus the LH/LLS group (P < 0.003).
|Clinical Outcome||LH/LLS||RH||ERH||P Value|
|Estimated blood loss (mL)||457 ± 190||893 ± 530||857 ± 486||<0.001*|
|Transfusion requirement [n/N (%)]||1/45 (2.2)||4/45 (9)||2/45 (4)||0.35†|
|Graft weight (g)||315.7 ± 88||828 ± 142||936 ± 115||<0.001*|
|Predicted weight (g)||335 ± 79||910 ± 204||953 ± 168||<0.001*|
|Spared volume (g)||1075 ± 205||550 ± 128||525 ± 127||<0.001*|
|Spared volume (%)||76.0 ± 4.5||37.7 ± 6.5||35.5 ± 5.4||<0.001*|
|Hospital stay (days)||5.8 ± 1.1||6.8 ± 1.4||7.5 ± 1.7||0.003†|
Liver biochemistry findings after living liver donation are detailed in Table 3. The peak aspartate aminotransferase (AST) and alanine aminotransferase (ALT) levels were similar for the 3 groups. However, there was a greater degree of hepatocellular dysfunction with the larger, right-sided hepatectomies. The median peak international normalized ratio (INR) was higher and the median time to an INR within the normal range was longer for the RH and ERH groups versus the LH/LLS group (P < 0.001 and P < 0.001 respectively). The median peak bilirubin levels were also significantly higher in the RH and ERH groups versus the LH/LLS group.
|Postoperative Biochemistry||LH/LLS||RH||ERH||P Value|
|Peak AST (IU/L)||325 ± 206||305 ± 185||236 ± 171||0.35|
|Peak ALT (IU/L)||345 ± 243||302 ± 190||266 ± 155||0.18|
|Peak INR||1.43 ± 0.2||1.8 ±0.3||1.8 ± 0.3||<0.001|
|Days to normal INR||2.8 ±1.1||5.0 ±1.3||6.6 ± 3.6||<0.001*|
|Peak bilirubin (μmol/L)||22.6 ± 11.8||47.4 ± 26.0||51.0 ± 30.0||<0.001|
Donor Clinical Outcomes
Clinical outcomes after LH/LLS, RH, and ERH were compared (Fig. 1 and Table 2). There was no statistically significant difference in the overall complication rates of the LH/LLS, RH, and ERH groups (range = 31%-37%), and there were no grade 4 or 5 complications. The rates of grade 1 and 2 complications were similar. No grade 3 complications were noted in the LH patients (n = 4), although this group was small. Specific complications for each group are listed in Table 4.
|Complication||LH/LLS (n)||RH (n)||ERH (n)|
|Urinary tract infection||1||1||3|
|Fever requiring antibiotics||1||0||1|
|Pleural effusion requiring medical management||0||1||1|
|Abdominal wall spasm||1||0||0|
|Hemorrhage requiring transfusion||0||0||1|
|Hemorrhage requiring surgery||0||3||2|
|Abscess (intra-abdominal) requiring percutaneous drainage||0||1||1|
|Wound infection requiring debridement||1||0||0|
|Pleural effusion requiring thoracentesis||0||0||1|
|Incisional hernia requiring operative repair||0||1||0|
|Bile duct injury||0||0||1|
|Prolonged elevated bilirubin requiring endoscopic retrograde cholangiography||0||0||1|
Univariate factors that were associated with serious complications within the first 30 days after donation (grade 3) included RH (or ERH), the spared liver volume percentage, and the peak bilirubin level (Table 5). Factors thought to be clinically significant or factors with a P value < 0.1 were included in a multivariate analysis. Only a peak bilirubin level > 50 μmol/L was associated with postoperative morbidity in the multivariate model (relative risk = 0.139, P = 0.005).
|Predictor of Complications||No Grade 3 Complication||Grade 3 Complications||Univariate P Value||Multivariate P Value||Relative Risk (Confidence Interval)|
|History of smoking (n/N)||27/121||2/14||0.73*|
|Spared volume (g)||715.8||573.3||0.06*||NS|
|Spared volume (%)||49.8||38.9||0.009*||NS|
|Graft weight (g)||799.0||861.0||0.21*||—|
|Blood loss (mL)||749.4||827.3||0.61*||—|
|Peak AST (IU/L)||304.3||260.4||0.44*||—|
|Peak ALT (IU/L)||312.9||233.9||0.06*||NS|
|Peak bilirubin (μmol/L)||36.6||74.1||0.001†||0.005‡||0.139 (0.035-0.547)‡|
|Days to normal INR||4.6||6.1||0.12*||—|
|Blood transfusion (n)||5||2||0.15†||—|
Factors associated with an increased length of stay (>7 days) were also determined (Table 6). Patients were placed into 2 categories: (1) 7 days or less and (2) more than 7 days. A univariate analysis was performed, and RH, an increased peak INR, any complications, and severe complications (grade 3) were found to be significant. After a multivariate analysis, only RH (relative risk = 14.3, P = 0.01) and grade 3 complications (relative risk = 4.1, P = 0.02) were found to be independent predictors of an increased hospital stay.
|Factor||≤7 Days||>7 Days||Univariate P Value||Multivariate P Value||Relative Risk (Confidence Interval)|
|Age (years)||34.5 ± 9.1||35.2 ± 9.3||0.69*||—|
|History of smoking [n/N (%)]||25/104 (24)||4/31 (13)||0.19*||—|
|BMI (kg/m2)||24.8 ± 3.6||24.7 ± 3.7||0.91*||—|
|Hepatectomy (n)||<0.001†||0.01||14.3 (1.8-111.4)|
|Spared volume (%)||51.8 ± 20.0||37.7 ± 9.3||<0.001*||NS|
|Blood transfusion [n/N (%)]||4/104 (3.8)||3/31 (9.7)||0.20†||—|
|Peak ALT (IU/L)||306.9 ± 206.1||296.9 ± 182.9||0.81*||—|
|Peak AST (IU/L)||296.8 ± 182.6||309.7 ± 201.9||0.74*||—|
|Peak INR||1.64 ± 0.29||1.76 ± 0.35||0.05*||NS|
|Peak bilirubin (μmol/L)||38.9 ± 26.3||45.9 ± 28.6||0.21*||—|
|Any complications [n/N (%)]||31/104 (30)||17/31 (55)||0.011†||NS|
|Severe complications: to grade 3 [n/N (%)]||6/98 (6.1)||8/31 (25.8)||0.001†||0.02||4.1 (1.2-13.6)|
The present analysis is the first to use a case-control method to compare the outcomes of living donation after RH and LH in the modern era. A comparison of these matched cohorts (based on age, BMI, sex, and the time of surgery) revealed that the total complication rates were similar, but the rates of hepatocellular dysfunction and serious complications (grade 3) significantly increased as the volume of the removed liver increased (RH/ERH versus LH/LLS, P = 0.04). Moreover, the multivariate analysis showed that a peak postoperative bilirubin level > 50 μmol/L significantly correlated with serious complications. Together, these data suggest that donor programs should strive to remove the smallest possible liver volume needed to provide adequate function in the recipient and thereby minimize donor liver dysfunction, prevent serious complications, and maximize donor safety.
The largest reported series8 included 3565 right and left lobe liver donor procedures performed in Japan, and an analysis of the outcomes found no material differences in the overall complication rates. This finding was not unexpected because complications are known to be frequently underreported in registry data. In contrast, another multi-institutional study from Asia by Lo17 assigned a higher risk of complications to right lobe donors versus left lobe donors. In single-center reports from Asia, serious complication rates with living liver donation have ranged from 0% to 18.8%, and complications have usually been reported to be less frequent with left-sided surgical procedures.9, 13, 14 Morioka et al.,9 for example, described an experience with 335 consecutive living donor cases and reported an overall complication rate of 39.7%. A logistic regression analysis revealed that case experience (>100) and the type of hepatectomy were independent predictors of postoperative morbidity. Chan et al.10 reported a higher incidence of perioperative complications after LH.
North American single-center studies have also generally reported lower complication rates with left lobe donation. Salamé et al.11 reported that right liver donation was associated with longer anesthesia times, greater blood loss, and greater postoperative biochemical liver abnormalities, but they did not report complication rates or factors associated with severe complications. Salvalaggio et al.12 compared donors who underwent RH and donors who underwent LH or LLS, and they reported a higher rate of complications and more severe complications with right lobe donation; however, the factors potentially associated with severe complications were not analyzed.
A limited experience with the outcomes of adult recipients of small left lobe grafts in North America has been reported.7 In this study, the majority of the left and left lateral segment grafts were transplanted into pediatric recipients, whereas the majority of the right lobe grafts were transplanted into adults. Recipient outcomes were not the focus of this study, but our initial experience indicates that the outcomes with small left lobe grafts in adult recipients are similar to the outcomes with small right-sided liver grafts with respect to the good long-term overall survival rate; however, these recipients tend to have a more prolonged and complicated postoperative course. Because donor satisfaction is linked to recipient outcomes, it is possible that a significant increase in morbidity with smaller grafts would not be viewed favorably by either the donors or the recipients.18 These concerns are less relevant to anonymous donation; in this case, it is usually possible to select a recipient for whom the graft will provide a more than adequate volume.19 More studies are needed to (1) accurately define the overall recipient morbidity and long-term survival rates with left lobe grafts versus right lobe grafts in North American recipients and (2) define the acceptable level of increased recipient risk with a small graft from the perspective of the living donor program for both donors and recipients. Notwithstanding these important considerations, however, we propose that donor safety should take precedence even if the use of lower graft-to-recipient weight ratio grafts results in more morbidity for the recipients.
This is the first study comparing donor outcomes after different surgical procedures in well-matched cohorts of patients who were treated in a similar fashion and were followed prospectively with their complications graded according to the Clavien classification system. Similarly to the authors of the large series from Japan,8 we found that overall complication rates were similar with the different types of living donor liver resections. However, serious complications—the ones that were most important to avoid—were significantly more likely to occur when a larger portion of the liver was removed. The postoperative peak bilirubin level was shown to be a strong predictor of severe postoperative morbidity in both univariate and multivariate analyses.
The preoperative factors that were predictive of postoperative complications and longer hospital stays were the spared liver volume percentage and the type of hepatectomy. The postoperative peak bilirubin level was the strongest predictor of severe postoperative morbidity in both univariate and multivariate analyses. In this context, bilirubin likely serves as a global marker for many complications that might affect liver function, such as poor hepatocyte reserves, biliary stasis or obstruction, and hemolysis after hemorrhaging or blood transfusions. The peak bilirubin level typically occurred within the first 48 hours after donation. High bilirubin levels almost always preceded the diagnosis of the complication and, therefore, served as a good predictor of postoperative complications.
Our data suggest that the incidence of serious adverse outcomes is a continuous function of the proportion of the liver that is resected, and this supports the recommendation that left lobe donation be considered first for any potential donor if the volumes permit this. However, despite our efforts to control for key donor factors with a matched cohort analysis, the present study still shares several weaknesses with other reports. First, the treatment was not randomized. Second, the LH group was heterogeneous: 41 patients underwent LLS, and only 4 underwent full left lobe resection. Although LH is a more complex surgery then LLS (ie, a larger parenchymal transection and a greater opportunity for blood loss), postoperative liver dysfunction is often similar because the segment 4 liver remnant after LLS is often left devascularized on account of the division of the segment 4 portal vein and arterial branches. The middle hepatic vein is also taken with LH, but our experience with right lobe grafts suggests that the removal of the vein in and of itself does not substantially affect donor outcomes.15 Thus, it is not certain whether the significant differences in the degree of hepatocellular dysfunction and serious complications identified in this study would translate to a cohort of LH donors with an intermediate residual liver volume, but this merits exploration. Third, we did not perform a detailed analysis of the recipient outcomes. The lower limit for the graft-to-recipient weight ratio with portal flow modulation that will result in minimal additional morbidity in recipients has yet to be determined. Finally, our sample size was relatively small, and the analysis was retrospective. The latter limitations were mitigated by the use of uniform treatment protocols, the careful and consistent prospective collection of donor complications, and 100% follow-up.
This is a good time for reexamining the advantages and disadvantages of right and left lobe living liver donation because the surgical practices for recipients are currently undergoing an evolution. To date, most adult-to-adult LDLT procedures at our center and elsewhere across North America have preferentially used the donor's right lobe for liver transplantation in adults. This practice is based on a desire to ensure that the recipient receives as large a liver segment as possible because this has been thought to be a prerequisite for meeting the metabolic demands of the recipient.20 This assumption is now being challenged. We have reported that as long as the transplanted living liver segment has good venous outflow, a living donor graft with a graft-to-recipient weight ratio < 0.8 (median = 0.72, minimum = 0.59) can be transplanted with excellent outcomes.21 Indeed, other groups worldwide have reported that even smaller grafts can be successfully transplanted when portal inflow modification (eg, portocaval shunts or splenectomy) is used to reduce excessive portal pressures.7, 22 We believe that donor safety must remain the most important priority of living liver donor programs. Our data show that serious complications are much more likely to occur when the donor has a small residual liver volume. Thus, we propose that removing the smallest piece of liver likely to provide adequate function in the transplant recipient should become the standard of care.