- Top of page
The progressive shortage of liver donors has mandated investigation of living-donor transplantation (LDT). Concerns about increasing risk to the donor are evident, but the impact of the degree of parenchymal loss has not been quantified. We analyzed clinical and biological variables in 45 LDT performed by our team over 2years to assess risks faced in adult LDT. All donors are alive and well with complete follow-up through to February 2001. When the three operations were compared, right hepatectomy (RH) was significantly longer in terms of anesthesia time and blood loss compared with left hepatectomy (LH) and left lobectomy (LL). Donor remnant liver was significantly reduced after RH compared with LH and LL. There were significant functional differences as a consequence of the remnant size, measured by an increase in peak prothrombin time after RH. RH for adults represents a markedly different insult from pediatric donations in terms of parenchymal loss and early functional impairment. Left hepatectomy donation offers modest advantage over right lobes but seems to confer substantial technical risk for a small gain in graft size. Unless novel strategies are developed to enhance liver function of the LH graft in the adult recipient, right lobe donation will be necessary for adult LDT.
- Top of page
The exponential growth of the waiting list for liver transplantation in the past decade has mandated investigation of living donors for liver transplantation (LDT). While in 1993, the liver waiting list was equal to the number of transplants performed in that year, 8years later the list exceeds transplant capacity fourfold (1). LDT was introduced for children nearly a decade ago and ethical parameters for evaluating this modality were proposed (2, 3). In subsequent years widespread acceptance of LDT for children occurred (2, 4) along with clear definition of the risk to the donor based on the accumulation of an international experience of over 1000 cases (5–7). Because the growth in the transplant waiting list has occurred primarily among adult candidates, extension of LDT to these patients required modification of the approach to the donor. Early efforts to use small grafts in larger recipients were disappointing, due to the inadequate liver volume transplanted in the standard donor hepatectomy. In 1996 we described the syndrome of cholestatic dysfunction, which occurs when the liver graft is too small (8). Auxiliary LDT, in which a small graft was used supplemented by preservation of the native liver has been disappointing, plagued by a broad spectrum of technical failures (9). The inevitability of increasing the donor hepatectomy led to use of the right or even the extended right lobe, introduced by Fan (10) and Tanaka (11) in the early 1990s and later applied in a large series of recipients by Marcos who reported the first large series of right lobe procurements with promising results in 1998 (12).
While the benefits of transplanting a larger graft are obvious for the recipient, the incremental risks to the donor need to be better defined. In the world experience with left lateral lobe donation two deaths have been reported, corresponding to a rate of 2/1000 or 0.2% (5). This risk seems acceptable though it is an order of magnitude higher than the estimated risk for renal donation (13). As the hepatectomy increases, the donor faces longer anesthetic exposure, a greater risk of injury to biliary and vascular structures, and emerges from the operation with a smaller liver remnant. Despite these concerns the use of right lobe donation has proliferated rapidly. As of the end of 2000, an estimated 600 cases have been performed in Europe and North America (14). Three well-documented deaths of right lobe donors (0.5%) have been presented (14–16) though none have been published to date. Siegler pointed out that this mortality is between five- and tenfold that faced by renal donors (17). Because these deaths all occurred in centers with substantial expertise in partial liver transplantation, the real risk might well be higher as teams with less experience adopt the technique.
There has been little study of the impact of donation as a function of the extent of resection as few centers have large experience with LDT of both adults and children. For example, Kawasaki et al. have limited LDT in adults to left lobe resection, arguing that right lobe donation involves unacceptable risk (18). In contrast, Fan systematically uses an extended right hepatectomy, which includes the middle hepatic vein as the necessary graft for an adult recipient (19). Our experience with all three resections led us to analyze the components of the operation, which might contribute to risk in the donor. In this study, our goal was to better define the correlation between the magnitude of the surgery, the extent of parenchyma resected and the functional consequences upon the remnant liver.
Patients and Methods
A retrospective analysis of 45 living-donor liver transplantation (LDLT), performed in our center by a single team from January 1998 to February 2001 was conducted. Data were obtained from medical record review and follow-up was complete as of March 2001. LDT grafts accounted for 22% and 55% of adults and children, respectively, who were transplanted in our center during this period. Of these 45 LDLT, 22 were donors for pediatric recipients and 23 for adults. The median donor age was 31years (range 19–57) and sex ratio (M/F) 20/25. The donors consisted of 21 parents, 3 siblings, 9 sons, 3 daughter, 1 niece, 1 daughter in-law, 4 spouse and 3 friends (Table 1).
Table 1. Donor demographics
|Age||19–57years (median 31)|
|Sex ratio (male/female)||20/25|
| Spouse|| 4|
| Siblings|| 3|
| Daughter|| 1|
| Niece|| 1|
| Sons|| 9|
| Friend|| 3|
| Daughter in-law|| 1|
| LH|| 9|
Indications for liver transplantation in the pediatric recipients included biliary atresia (n = 15), primary sclerosing cholangitis (n = 1), Alagille's syndrome (n = 2), Tyrosinemia (n = 1), primary familial idiopathic cholesterolemia (n = 2) and parenteral nutrition hepatotoxicity (n = 1). In the adults recipients indications included hepatitis C cirrhosis (n = 8), alcoholic cirrhosis (n = 5), hepatocellular carcinoma (n = 5), autoimmune cirrhosis (n = 2), hepatitis B cirrhosis (n = 1), cryptogenic cirrhosis (n = 1) and primary sclerosing cholangitis (n = 1).
Because liver transplant recipients in our region face extreme donor scarcity, it has become our practice to systematically inform all candidates for liver transplantation of the possibility of LDT. We have previously reported our sequential approach to the donor evaluation (20). Prospective donors underwent systematic medical and psychosocial evaluation as well as routine screening blood work. Arteriography and liver biopsy are used selectively. We have adopted magnetic resonance imaging (MRI) as the single pre-operative analysis of liver. MRI provides three-dimensional imaging of the afferent and efferent vasculature as well as the biliary anatomy and the parenchyma. Intra-operatively, further analysis of the vascular and biliary anatomy was performed using duplex ultrasound (5MHz probe; B-K medical, Copenhagen, Denmark). Intra-operative cholangiogram via the cystic duct remnant was selectively performed.
General endotracheal anesthesia was induced and monitored by a single transplant anesthesia team. Arterial blood pressure, cardiac, and pulse oximetry monitoring were employed. Venous access was obtained most commonly through two peripheral intravenous lines, and rarely via a central venous catheter. Cell saver suctioning was used in all donor hepatectomies. A long mid-line incision was used for left lobectomy (LL) hepatectomy with a standard liver transplant incision for the left and right hepatectomy (RH) hepatectomies. More recently, we have used the mid-line incision for all donations.
Three hepatectomies were defined according to the segmental anatomy of Couinaud (21). LL for resection of segments II and III, left hepatectomy (LH) for segments II, III and IV, and RH for segments V, VI, VII and VIII resection (Figure 1). In all cases preliminary hepatic venous dissection permitted passing a tape behind the liver in the plane of transection, thereby guiding the parenchymal division. There were 21 LL, 15 RH and 9 LH. In all cases of hemihepatectomies the parenchyma was transected to the right of the middle hepatic vein. The liver parenchyma was transected with the ultrasonic dissector (CUSA) and the vascular pedicle was divided after transection of the liver parenchyma. Right hepatectomy was conducted by transection of the short hepatic veins and dissecting the right hepatic vein; cholecystectomy; right hepatic artery discovery; right portal vein discovery; biliary ducts dissection; finally parenchymal transection and pedicle division. Left hepatectomies were conducted with first dissection of left and middle hepatic veins; cholecystectomy; left hepatic artery discovery; left portal vein discovery with mobilization of the caudate lobe portal veins and dividing the bile ducts. For LL, left hepatic vein was dissected; left hepatic artery discovery; left portal vein discovery with mobilization of all branches for segment IV; parenchymal transection was finally performed with repairing the falciform ligament. Donors were not heparinized prior to cross-clamping. Heparin was introduced in the graft perfusate on the back table after removal of the liver from the operative field.
Figure 1. Three hepatectomies are depicted: right hepatectomy, left hepatectomy, and left lobectomy. In right hepatectomy and left hepatectomy, the liver is divided to the right of the middle hepatic vein. During left lobectomy, the parenchyma is divided to the right of the falciform ligament.
Download figure to PowerPoint
Post-operatively, donors were in observation in the intensive care unit (ICU) overnight and then transferred to the ward. Post-operative analgesia was managed with intravenous (i.v.) administration of narcotics generally dosed by patient-control analgesia (PCA). Crystalloid fluids were given in the perioperative period without the need for exogenous blood products. Diets were reintroduced with return of normal gastro-intestinal function. Donors were discharged once incisional pain was adequately controlled, once ambulating without assistance and tolerating a regular diet.
Volume studies were performed using pre-operative MR images in the first 32 donors. On a MRI workstation (GE Medical Systems, Milwaukee, WI, USA), the volume of each graft was calculated by the method described by Heymsfield et al. in 1979 (22). The edge of liver parenchyma and the limits of the grafts (LL, LH, RH) were traced with the cursor on serial MR images in both axial and coronal planes. The surface area of each section was calculated by the computer and used to estimate volume based on slice thickness. Total volume was calculated as a summation of individual section volumes. Grafts were drawn and volumes were calculated as described for total liver volume.
To validate the MR measurements, total measured liver volumes were compared with the standard liver volumes calculated with the Urata formula (23) (Table 2). Overall, the total volume of the liver measured by axial computation was highly correlated to the predicted liver volume (r = 0.750, p = 0.0001) though the relationship was not perfect (Figure 2). Furthermore, no significant differences were observed, in each group, between the measured volume and the predicted liver volume. The RH graft volume was significantly higher than LH or LL as well as LH compared with LL. More importantly for the donor, the mean residual volume after RH was estimated to 41%, 71% after LH and 82% after LL. The smallest remnant observed after RH was 32%, accounting for 0.6% of body weight. The largest remnants of the RH group accounted for 0.9% of body weight. In contrast, all remnants for the LH and LL were greater than 1% of body weight. Thus, increasing the hepatectomy from LL to LH only added 11% to the mass of the graft. Highly significant differences in parenchymal resection were observed when RH was compared with either of the other resections.
Table 2. MRI measured volumes compared with calculated volumes (Urata) in each group of hepatectomy: anovat-test
|Variable||RH Mean ±SD||LH Mean ±SD||LL Mean ±SD||RH/LL p||RH/LH p||LH/LL p|
|Calculated total volume||1349 ± 198||1390 ± 132||1208 ± 136||0.016*||0.63||0.06|
|Measured total volume||1420 ± 492||1457 ± 311||1264 ± 353||0.35||0.85||0.35|
|Graft volume 1* (total volume)|| 850 ± 339 (60%)|| 429 ± 124 (29%)|| 222 ± 81 (18%)||< 0.0001*||0.0001*||0.02*|
|Residual volume|| 571 ± 186||1028 ± 196||1042 ± 289||0.0001*||0.0014*||0.9|
|(% total volume)||(40%)||(71%)||(82%)|| || || |
Figure 2. Regression analysis of the correlation between predicted liver volume based on the Urata equation, and the total liver volume estimated by axial computation in our donors (n = 34).
Download figure to PowerPoint
Operative anatomy is summarized in Table 3.
Table 3. Operative anatomy findings in each group of hepatectomy (n = number of patients)
|Variables||RH (15)||LH (9)||LL (21)|
| Two RHA||1||–||–|
| Aberrant PRBD-LBD||–||2||–|
| Two RBD||4||–||–|
| Three RBD||1||–||–|
| Two LBD||–||1||1|
Arterial. The most common anatomical variation was a left hepatic artery arising from the left gastric artery in seven patients who underwent left-sided hepatectomy. This artery was taken as the arterial supply of the graft. In one patient undergoing RH, early bifurcation of the right hepatic artery existed. The smaller of the two vessels appeared functionally insignificant and was ligated.
Portal vein. The most notable portal vein anatomical variation identified at the time of hepatectomy was an early bifurcation of the right portal vein. This variation was seen in four patients undergoing right lobectomy. In our initial encounters with this variant we attempted to obtain a single orifice for the right portal vein leading to narrowing of the left portal vein in the donor in two cases. In both patients, the problem was identified intra-operatively and repaired with transverse venoplasty. Most recently, in a fourth patient, the bifurcated right veins were clamped individually avoiding this problem all together.
Hepatic veins. Thirteen patients had accessory right hepatic veins. Accessory right hepatic veins were clinically relevant in nine RH donors, and were used in addition to the main right hepatic vein for the graft outflow in seven patients. Two of these seven had anterior accessory branches draining into the middle hepatic vein. These anterior veins draining medially can be identified by intra-operative ultrasound and spared for re-implantation.
Biliary ducts. Biliary tract variations were mainly problematic in the left-sided hepatectomies. The most significant variant, a posterior right bile duct draining segments VI + VII joining the left bile duct. This was identified in two patients who had LH in which the posterior right bile duct was divided and biliary reconstruction was performed. Other biliary variations included two bile ducts draining either the left liver or the left lateral segment (two patients) and a left bile duct running inferior to the left portal vein (one patient). In five patients, undergoing RH, two right bile ducts were identified in four patients and three bile ducts in one other. Both bile ducts were divided in the donors and used for anastomosis in recipients.
Statistical analysis was performed using StatView 5 software (Abacus Concepts, Cary, NC, USA). Data were presented as mean ± SD. Means were compared using anova with post-hoc intergroup comparisons. Correlations between continuous data were identified with linear regression analysis.
- Top of page
This series provides a comparison within a single surgical series between the three resection types for the analysis of the operative risk of donor hepatectomy. Admittedly, the sample size is too small to address mortality risk; however, based on the morbidity we observed, it is possible to speculate about the hazards faced by the potential donor. Our data make it clear that the extent of hepatectomy has a highly significant impact on functional impairment after donation. Both serum bilirubin and coagulation profile reflect significant disturbances in hepatic function early after donation, most marked in the right lobe donors. Despite this important difference, the consequent impact on clinical outcome, such as the duration of hospitalization and the occurrence of surgical complications was seemed to be minor. Left hepatectomy, which is a much less severe resection in terms of hepatic parenchymal loss and duration of surgery was associated with the most severe intra-operative complications, though with no sequelae to date. Overall, we were disappointed with the results of left lobe donation because the gain in hepatic parenchyma was modest (+11% over left lateral donation), and the technical hazards associated with the biliary dissection may exceed those of the right hepatectomy. Finally, our results confirm the remarkable safety of left lateral donation (the standard for pediatric transplantation), an operation accomplished quickly, with the least blood loss and minimal functional impact on the donor's liver function.
Each of the hepatectomies must have a place in the treatment of patients, choosing the one least likely to cause morbidity while producing an acceptable graft for transplantation. To date our efforts to avoid increasing the hepatectomy, including the use of small LL grafts as auxiliary transplants in adults, have not achieved success in adult recipients. For the time being total left or total right grafts are mandatory if LDT is to be performed in adults. The most extreme resection, the right trisegmental graft is difficult to justify though Fan et al. (10) have routinely extended the right hepatectomy to the left of the middle hepatic vein to optimize the venous outflow of the right lobe for the recipient. The Colorado group as well (24) has recently adopted this approach. Depending on the morphology of the liver, this extended hepatectomy leaves the donor with quite a small remnant requiring either a risky resection, or else the exclusion of a large number of healthy potential donors due to limitations in remnant size. Despite concerns about the safety of the resection, the right lobe graft has become the preferred technique for adult donation and this procedure has been rapidly adopted worldwide (10, 24–26).
To date, five donor deaths have been described at meetings in the scientific community though only one has been published (5, 14–16). There is a clear difference between the two deaths after LL donation (due to pulmonary embolization with normal liver function) and the three deaths after RH donation that were all associated with liver failure (5, 14–16). Based on our documentation of the marked disturbance in hepatic function after right liver lobe donation, it is possible to place the reported deaths in donors in some perspective. Volume measurements in our study would indicate that the right hepatic lobe donor is left with a smaller hepatic remnant than the minimum graft size, which has been deemed safe for the recipient (8, 27, 28). We have previously reported that that delayed graft function occurred if the transplanted liver was less than 50% of expected liver mass (8). However, the outcome after partial liver grafting is due not only to the size of the remnant, but also the condition of the recipient at the time of transplantation. This was best demonstrated in a large series of adults receiving left lobe grafts in which the recipient condition became the dominant predictor of outcome as all the grafts in that series were relatively small for size (18). This must serve to explain why donors, a highly selected, healthy cohort readily tolerated small hepatic remnants. In the current series, the remnant liver volume in the donor after RH ranged from 32% to 54% of the total measured liver volume. Inomata et al. (29) observed a significant difference in the liver function (AST, TBili) of RH donors compared with both LH and LL as long as 7d post-operatively. In our study, all remnants greater than 55% of measured liver volume had prompt return of liver function. In all but one, patients who had significant delayed functional recovery had a remnant liver volume less than 54%. In the setting of delayed recovery of liver function, a second complication such as a bile leak, an infection, or an accident of anticoagulation can easily set up a cycle of fatal complications. Thus, donor deaths (or near misses) probably require a second insult superimposed on the baseline state of transient hepatic insufficiency.
The full left lobe hepatectomy is the most controversial of the procedures for adults LDT. It is appealing because the parenchymal reduction is modest for the donor and seems optimal for donations in which the donor is much larger than the recipient. This hepatectomy has been used exclusively by the Tokyo group (18, 30), but it seems clear that only a limited range of recipients may benefit from these small grafts if the usual criteria for recipient graft size are adhered to (8, 27, 28). Our results indicate an intermediate degree of operative time, blood loss, and difficulty for this operation when compared with right or left lateral resection with only modest impact on post-operative liver function of the donor. More troubling, in our own series, is the high incidence of biliary injuries in these resections, due to the frequency of posterior right duct insertions to the left of the confluence, and, in one of our cases, the posterior duct passed behind the common portal vein and inserted into the left side of the common hepatic duct. The biggest problem with this procedure, however, is the unsatisfactory graft function we observed in many of the recipients (31). Ultimately, the full left resection seems to have limited application for living donation in most circumstances and may be recommended only for larger children or very small adults. It would be wrong to completely eliminate this graft from the spectrum of possibilities, as it may be needed in the extension of the splitting of cadaveric livers for the transplantation of pairs of adult recipients, in which the small left lobe graft can be allocated to very small adults on the waiting list.
We have noted earlier the good results of left lateral donation both in terms of donor outcomes as well as a well-established track record of recipient benefit. Our results pose a marked contrast to a recent report from Belghiti et al. suggesting that left lateral donation is bloody and just as morbid as right hepatectomy (32). It should be noted that the plane of transection in that series was well to the right of the round ligament making the resection more analogous to a full left hepatectomy. We attribute our good results to progressive evolution of our hepatectomy technique to minimize the invasion of the donor (4, 31). We continue to pass quite far to the left in transection of the bile duct because leaving the left duct long seems detrimental to both the donor (proximity to aberrant right duct insertions) and the recipient (ischemic strictures of the duct) (4, 31). Our previous recommendation that the left duct be transected no more than 1cm to the right of the round ligament seems to optimize the balance between protection of the donor common duct and the creation of a suitable duct for re-implantation.
In addition to the issues related to the extent of parenchyma resected, the most challenging technical element in the procedure seems to be associated with the biliary dissection. The current report is the only large series of donor hepatectomies in which there have been no post-operative bile leaks. As alluded to earlier, the combination of a bile leak with post-operative hepatic insufficiency can be the formula for disaster and our technique has emphasized the safety of the donor structures, perhaps at the expense of ease of the biliary reconstruction in the recipient (31). Nonetheless, the left hepatectomy posed some intra-operative difficulties with biliary anatomy requiring immediate reconstructions. In two cases of LH we found right posterior bile ducts, which emerged from behind the right portal vein in one case and the common portal vein in the other to enter into the posterior aspect of the left bile duct in one case and the common hepatic duct in the other. Biliary reconstruction of these ducts was performed in both cases. Early identification of such variations on pre-operative imaging is crucial; however, until recently, MR cholangiography of nondilated biliary ducts was not sensitive enough to detect all of the anatomical variations (33). Intra-operative cholangiography has been used with good success in defining the presence of accessory bile ducts (6, 10, 34) but does not provide the full three-dimensional perspective and correlation with the vascular elements.
In contrast to the biliary difficulties, pre-operative MRA and intra-operative ultrasound permit complete definition of the vascular anatomy. We have become extremely interested in the definition of the venous drainage of the right lobe because in some 30% of livers, drainage of the anterior sector (segments 5 and 8), is via large veins, which join the middle hepatic veins (35). This has led to extension of the hepatectomy to the left of the middle vein as proposed by Fan (10). Although we have used that approach in a single case, we prefer to selectively reconstruct hepatic veins based on anatomic assessment of the right graft and keep the plane of transection to the right of the middle vein (31, 36). The need for the middle vein for outflow remains the subject of debate with good recipient outcomes reported by those who drain the middle vein (10, 19) as well as those who do not (25, 31). Marcos et al. insist on identification of all of the accessory hepatic veins, which are of great importance for recipient graft outflow (12, 34). We now routinely reconstruct short hepatic veins entering the inferior vena cava (IVC) or right veins entering the middle vein that approach 10 mm in diameter. Reconstruction of middle hepatic vein tributaries is critical when graft size is limited. Veins are either re-implanted directly into the IVC or reconstructed with an interposition graft using a cadaveric vein graft. In our recent practice this has been performed in 30% of transplants (Kinkhabwala M, New York Presbyterian Hospital, New York, NY, USA, December 2000).
The portal vein is rarely problematic with the left-sided hepatectomies, but a significant number of livers have an early bifurcation of the right portal vein (25). In our early right hepatectomies we attempted to obtain a common orifice for the right lobe in these patients and narrowed the common portal vein in the process. Both cases were recognized intra-operatively and the portal vein was re-closed transversely with a good result as depicted in Figure 4. This was reported in the Lahey series, and required a re-exploration on post-operative day 1 (37). Currently, in cases of portal trifurcation, which are readily diagnosed pre-operatively on MRI we dissect each branch and divide them separately to avoid compromising the donor portal vein.
In Marcos initial series anesthetic times for the donors exceeded 10h in many cases (12). This prolonged period of surgery can only pose hazards, and in our series, the total anesthetic times were nearly 2h longer for the right hepatectomies (8.8h) than the left or left lateral donations. Medical complications have been few in the current series, though a near fatal pulmonary embolus has been reported from Paris, as well as a fatal post-operative hemorrhage from a German center (Broelsch CE, Essen, Germany. Personal communication, December, 2000.). It seems that anticoagulation and the prevention of thrombotic problems is a double-edged sword faced in the management of these patients. Our preference is the avoidance of anticoagulation and the use of mechanical compression devices to prevent lower extremity venous thrombosis. There is some debate about the extent of pre-operative screening for occult disorders of coagulation, which is appropriate as these complications may be devastating. The other important risk, which is faced by the donors, most serious in those right lobe donors with impaired early hepatic function is the occurrence of hypoventilation due to over sedation. For this reason, we maintain the donors in the recovery room overnight so that the airway can be closely monitored. Despite this precaution, one of our donors developed a mucus plug, which fortunately did not cause acute respiratory arrest before it was recognized and treated. Incisional hernias were frequent with the use of absorbable suture closure and have been dramatically reduced to date by the routine use of nonabsorbable materials.
In conclusion, the transition to more extensive hepatectomy for adult to adult live donor transplantation has escalated the risk faced by the living donor. Our technique has evolved with protection of the donor structures as the primary consideration. Nonetheless, the international experience with LDT for adults suggests that the mortality faced by right lobe donors is at least 0.5% and may approach 1%, a risk more than 10-fold that of live renal donation. In contrast, the left lateral donation has resulted in two documented deaths in over 1000 case during the past decade, a rate less than half that of the right lobe donors (5), yet still an order of magnitude greater than renal donors. Analysis of the current experience distinguishes the risk faced by the right and left lobe donors: the right lobe donor faces liver failure as the primary hazard, while the left lateral donor, like the renal donor, faces the vagaries of general anesthesia and the generic risks of major surgery in undergoing donation. Whether this difference is significant is a fit subject for ethical debate, in the meantime the safety of the donor must be protected by meticulous attention to the donor by a highly qualified surgical team and a scrupulous process of informed consent. Finally, resolving the issue of adequacy of training to undertake this procedure is a major challenge for the transplant community, but a requirement that should not be avoided.