Evolution of Donor Morbidity in Living Related Liver Transplantation

Abstract

Objective:

During the last 14 years, living donor liver transplantation (LDLT) has evolved to an indispensable surgical strategy to minimize mortality of adult and pediatric patients awaiting transplantation. The crucial prerequisite to performing this procedure is a minimal morbidity and mortality risk to the healthy living donor. Little is known about the learning curve involved with this type of surgery.

Patients and Methods:

From January 1991 to August 2003, a total of 165 LDLTs were performed in our center. Of these, 135 were donations of the left-lateral lobe (LL, segments II and III), 3 were of the left lobe (L, segments II–IV), 3 were full-left lobes (FL, segments I–IV), and 24 were of the full-right lobe (FR, segments V–VIII). We divided the procedures into 3 periods: period 1 included the years 1991 to 1995 (LL, n = 49; L, n = 2; FR, n = 1), period 2 covered 1996 to 2000 (LL, n = 47), and period 3 covered 2001 to August 2003 (LL, n = 39; FR, n = 23; FL, n = 3; L, n = 1). Perioperative mortality and morbidity were assessed using a standardized classification. Length of stay in intensive care unit, postoperative hospital stay, laboratory results (bilirubin, INR, and LFTs), morbidity, and the different types of grafts in the 3 different periods were compared.

Results:

One early donor death was observed in period 1 (03/07/93, case 30; total mortality, 0.61.%). Since 1991, the perioperative morbidity has continually declined (53.8% vs. 23.4% vs. 9.2%). In period 1, 28 patients had 40 complications. In period 2, 11 patients had 12 complications, and in period 3, 6 patients had 9 complications. Within the first period, 1 donor underwent relaparotomy because of bile leakage. Postoperative hospital stay was 10 days, 7 days, and 6 days, respectively. Donation of the full right lobe, in comparison with that of the left lateral lobe, resulted in a significantly diminished liver function (bilirubin and INR) during the first 5 days after donation but did not increase morbidity. One donor from period 1 experienced late death caused by amyotrophic lateral sclerosis.

Conclusions:

In a single center, morbidity after living liver donation strongly correlates to center experience. Despite the additional risks associated with temporary reduction of liver function, this experience enabled the team to bypass part of the learning curve when starting right lobe donation. Specific training of the surgical team and coaching by an experienced center should be implemented for centers offering this procedure to avoid the learning curve.

From January 1991 to August 2003, 165 living donor liver transplantations were performed in our center. We divided the procedures into 3 periods using a new standardized classification to evaluate postoperative donor morbidity. One early donor death out of 30 cases occurred in period 1 (03/07/93). Since 1991, perioperative morbidity has continually declined (53.8% vs. 23.4% vs. 9.2%). Morbidity after living liver donation strongly correlates with our experience. The new classification system should be used in further studies.

Living donor liver transplantation (LDLT) was introduced in 1989^1,2 to overcome the severe shortage of size-matched cadaveric organs for pediatric recipients and was able to help reduce mortality on the pediatric waiting list.^3–7 LDLT in children is now common practice at numerous transplantation centers worldwide⁸ and achieves results comparable with those of cadaveric organ transplantation.^3–7

The first successful LDLT in an adult recipient was performed in Japan in 1994 because of limited availability of cadaveric grafts in this country.⁹ The increasing shortage of cadaveric grafts in Western countries has forced transplant surgeons to follow this path.^10–12 Early on, it became clear that the left lateral lobe did not provide sufficient hepatocyte mass to allow successful transplantation in adults. Therefore, harvesting of either the full left or full right liver lobe was required, thus significantly enlarging the donor operation.

The beginning of LDLT was followed by an extensive ethical discussion. The balance between the benefit to the recipient and the risk to the donor (equipoise of risk and benefit) plays a central role in the justification of LDLT.^13–15 In worldwide reports of living liver donor operations, donor morbidity ranged between 0% and 67%,^8,16–20 depending on the individual definition and recognition of morbidity. The lack of a standardized assessment of perioperative complications is a serious limitation to the analysis of donor-related morbidity. Data regarding the effect of era and center experience on postoperative donor outcome are, until now, failing.

In this study, we analyzed the first 165 consecutive living donor operations since starting our program in 1991, comparing different periods and using a standardized classification of severity of complications.

PATIENTS AND METHODS

Study Population

The charts and scientific documentation of all 165 living donors operated between January 1991 and August 2003 were analyzed. Mean follow-up was 69 (±47.7) months. The cases were divided into 3 chronological periods. From the beginning of LDLT in Hamburg until 2000, the procedure was restricted to pediatric recipients. This population was subdivided into period 1 (1991–1995; n = 52) and period 2 (1996–2000; n = 47). In 2001, after approval from the local ethics committee, the LDLT procedure was extended to adult recipients. The positive decision was based on the overall good outcome of living related transplantation for pediatric recipients in our center and the worldwide successful development of living related transplantation for adult recipients. The last period was defined as period 3 (2001–2003; n = 66).

For subsequent analyses, the donor pool was further subdivided according to the graft type. A distinction was made between the donated left-lateral lobe (LL: segments II, III) and the donated hemiliver (left lobe, L: segments II–IV; full-left lobe, FL: segments I–IV; full-right lobe, FR: segments V–VIII).

Demographics and operative, clinical, and pathologic data were collected for all living related liver donors before and after donation. Specific donor personal variables including date of donation, sex, age, body weight, body mass index, serum chemistry, and hematologic parameters at donation were recorded. Outcome analysis included the number of reoperations, length of stay in the intensive care unit (ICU) length of initial hospitalization, morbidity, and death.

Classification of Donor Morbidity After Living Donor Liver Transplantation

Broelsch et al²¹ suggested in 2003 a uniform definition of complications as introduced by Clavien et al²² in 1994 to enable the transplant community to discuss donor outcome critically.²¹ Following the suggestion of Broelsch et al,²¹ we modified the morbidity classification of Clavien et al,²² designed for liver transplantation recipients, to adapt it to the living donor situation (Table 1). Morbidity associated with immunosuppressive therapy in the original classification (rejection, hyperglycemia, hyperlipidemia, or the development of malignancies) was disregarded. Complications which were graded low in the original classification of morbidity after orthotopic liver transplantation because they are associated with conditions of pre-existing disease (higher intraoperative transfusion requirement because of hemorrhagia) were graded higher in our modified LDLT classification because of the absence of pre-existing risk factors in healthy donors.

TABLE 1. Classification of Common Complications of the Donor After LDLT

Selection and Evaluation of Donors

The primary selection criterion for a living liver donor was voluntary and informed consent. Families considering LDLT were informed by a physician of the medical risks of the procedure, including the risk of death. Ethical issues and potential individual psychologic and social consequences were discussed between families and the psychologist. Apart from informed consent, the potential living donors had to fulfill the following acceptance criteria: (1) a relationship with the recipient within the third degree of consanguinity or intensive emotional relationship; (2) blood group compatibility with the recipient, except in recipients younger than 2 years of age without detectable, performed ABO antibodies; (3) no known medical disorder that significantly increases the perioperative risk or contraindicates donation. Donor candidates fulfilling these criteria were evaluated according to the protocol described elsewhere.^23,24 The evaluation protocol for adult to adult liver donation includes an accurate assessment of liver volume and liver quality and an estimation of vascular anatomy. Therefore, a single CT scan volumetric scan measurement is performed to evaluate size of future remnant donor liver. Liver histology is taken to determinate liver quality, especially fatty degree. A conventional hepatic angiography is performed to document intrahepatic arterial anatomy, especially arterial blood supply of segment IV. Exclusion criteria for adult to adult liver donation are a BMI >30, a remnant liver volume <30% of the total liver volume, and a fatty change of the liver >10%.

To reduce the risk, donors were advised to stop smoking.^25,26 Female donors were asked to discontinue oral contraceptives at least 3 months before surgery.²⁷ Accepted donors were advised to donate 2 units of blood and 2 units of FFP 2 weeks before the operation.

Surgical Technique

The detailed surgical technique is described elsewhere.^23,28 Harvesting of the left lateral lobe implements parenchymal transection along the falciforme ligament and sharp transection of the hilar plate at the umbilical fissure behind the left portal vein, thus preserving the segment IV and I bile ducts. Right lobe grafting was performed without taking the middle hepatic vein and without opening of the common bile duct or placement of T-tubes to avoid long-term sequelae in the donor. In general no anatomic variations that would require reconstruction of the donors’ vascular structures or bile ducts were considered as acceptable for donation. From 1991 to 2003, 2 new techniques for the parenchyma transection were introduced. At the start of our LDLT program, from 1991 to 1995 (period 1), electrocautery and clamp fracture only were used for parenchymal transection. In 1997, we started to use the harmonic scalpel (Ultracision) for parenchymal transection. In 1999, the ultrasonic transection technique (CUSA) was introduced. During the period 1996 to 2000 (period 2), the clamp fracture technique, Ultracision, and CUSA were used in 26 (55%), 16 (34%), and 5 (11%) donors, respectively. Since 2001 (period 3), CUSA has been exclusively used (n = 66; Table 2).

TABLE 2. Transection Technique in Living Donor Liver Transplantation 1991 to 2003 at the Hamburg University Hospital Eppendorf

Postoperative Treatment

All donors were admitted to the ICU at least for overnight monitoring. Until 1995, all patients received s.c. low-molecular-weight heparin starting on the day before operation and compressive stockings as antithrombotic prophylaxis. After 2001, additional intermittent mechanical leg compression and ultrasound control of the legs before first mobilization were performed. All further care was identical to that of normal hepatectomy patients.

Postoperative Liver Function

A simple score, adding points for prothrombine time and total bilirubin level, was used to evaluate liver function at postoperative days 1, 3, and 5. For each donor, 1, 2, or 3 points were awarded for postoperative prothrombin time and total bilirubin values. Distribution criteria for the laboratory levels are shown in Table 3. Mean sums of the points were compared between the 3 periods and within the last period between the FR group and the LL group.

TABLE 3. No Caption Available

Statistical Analysis

Values are shown as means and standard deviations (±SD) or as medians and ranges as appropriate. Analysis of variance (Bonferroni post hoc test) was used for continuous nonparametric data. The χ² test was used to identify significant differences between categorical variables. A P value < 0.05 was considered statistically significant. SPSS version 11.5 (SPSS, Chicago, IL) was used for all statistical analyses.

RESULTS

Study Population and Demographics

The grafts used for LDLT consisted in total of 135 LL, 3 L, 3 FL, and 24 FR. The liver weight of LL ranged from 133 g to 410 g, with a median of 237 g. The mean volume of FR and FL was 733.3 g (±138.12) and 425.3 g (±39.37), respectively. The median volume of the donor remnant liver after harvesting FR was 34.1% and ranged from 30% to 38.8%.

In period 1 (1991–1995), 49 LL, 2 L, and 1 FR were harvested (n = 52). In period 2 (1996–2000), 47 LL resections were performed. During the last period (2001–2003), 39 LL, 1 L, 2 FL, and 24 FR (23 without and 1 with middle hepatic vein) were harvested from the living donors. The recipients consisted of 137 children (<16 years of age) and 27 adults. In 1 case, a living related dual graft transplantation in an adult recipient was performed. The grafts consisted of 1 left lateral and 1 full left lobe (Table 4).

TABLE 4. Cumulative Number of Living Donor Liver Transplantations Performed at the Hamburg University Hospital Eppendorf From 1991 to 2003

From 1991 to 2000 (periods 1 and 2), the donors consisted of 97 parents, 1 aunt, and 1 strongly emotionally related person. Within the last period (2001–2003), the donors consisted of 39 parents, 1 grandmother, 15 children, 3 siblings, 2 aunts, and 2 nephews. Two donors were spouses, and another one was a strongly emotionally related person.

Of the 165 donors reviewed, there were 82 female and 83 male donors (male to female ratio, 1:0.99). In 91% of the cases, the donors were younger than 40 years. Three donors (1.8%) were 50 years or older. The median donor age was 31 (19–60) years. Donor weight ranged from 43 kg to 102 kg, with a mean of 70.9 kg (±11.79). The body mass index ranged from 18 kg/m² to 30 kg/m², with a mean of 23.8 kg/m² (±2.58). There was no significant difference within the periods in sex, age, body weight, or body mass index (Table 5).

TABLE 5. Demographic Data of Living Related Liver Donors at the Hamburg University Hospital Eppendorf

Morbidity Related to Donor Evaluation

Since 2001, every donor evaluation for LDLT has been recorded in our database. Since then, 70% of the potential donors have turned out to be unsuitable for donation.

In 3 cases, donor surgery was aborted intraoperatively. In 1 case, an unexpectedly high degree of fatty degeneration of the liver was the reason, and in the second case, hemiliver donor surgery seemed to be unjustified because of a too small remnant liver volume of the left lobe. In the third case, hypoplasia of the left hepatic artery led to abortion of the intended donor operation. In addition, 2 cases of evaluation morbidity were observed in 2003. After percutaneous liver biopsy, an AV fistula and 1 small hematoma within the right liver lobe were radiologically proven. The patient with the AV fistula was excluded from donor surgery for psychologic reasons, and the small hematoma complicated right lobe grafting in the recipient.

Operative Data

Mean time from skin incision to closure for LL harvesting was significantly reduced from 271.2 minutes (±65.02) in period 1 to 154.8 minutes (±26.9) in period 3 (P = 0.0001). Mean surgical time for harvesting of hemilivers ranged between 230 minutes and 410 minutes, with a mean of 347.6 minutes (±67.6). The operation time of the hemiliver donor surgery differed significantly from LL harvesting within the equal period (P < 0.0001).

Within the first period, 13 patients received autologous and 3 patients heterologous blood transfusions. In 8 and 1 cases, respectively, autologous and heterologous blood transfusions were administered intraoperatively after parenchymal transection. Postoperatively, in 5 and 2 cases, respectively, autologous and heterologous blood transfusions were given. None of the donors who received blood transfusions postoperatively (autologous or heterologous) underwent relaparotomy. Therefore, location of the intra-abdominal bleeding was not reproducible.

During the second and third periods, the transfusion requirement was significantly reduced. Within the second and third periods after LL harvesting, there was no transfusion required. After hemiliver donor surgery within the mentioned period, 3 and 1 donors, respectively, received autologous and heterologous blood transfusions postoperatively (postoperative days 1, 2, and 4). In 3 cases, autologous transfusions were administrated postoperatively because of intraoperative blood loss during parenchymal transection. In 1 case, after right lobe donation, transfusion of heterologous blood was necessary because of bleeding from the cut surface after heparinization in case of pulmonary embolism. LL harvesting and hemiliver donor surgery showed no significant differences in transfusion requirement within the third period (Table 6).

TABLE 6. Operative Data on Living Related Liver Donors at the Hamburg University Hospital Eppendorf

Postoperative Liver Function and Clinical Chemistry Analysis

After LL harvesting, the liver function scores were significantly higher for period 1 than for periods 2 and 3. This was the case for postoperative days 1, 3, and 5 (Fig. 1). Within the last period, the postoperative liver function score after FR lobe donor surgery was higher than after LL lobe harvesting. Statistically significant differences were observed at days 1, 3, and 5 (Fig. 1).

FIGURE 1. Postoperative liver function score at p.o. days 1, 3, and 5 after harvesting of the left lateral lobe, comparing the 3 periods on the left side (period 1 [black column] versus period 2 [white column] versus period 3 [shaped column]) and between different grafts in the period 3 on the right side (LL grafts [shaped column] versus FR grafts [dot column]). *P < 0.05, ANOVA Bonferroni test.

During the first and second periods, aspartate aminotransferase (AST) peaked on postoperative day 1. For LL harvesting, the highest AST levels were observed during the first period. AST level was 158.5 (±159.8) U/L, 42.8 (±27.9) U/L, and 28.8 (±10.6) U/L at days 1, 3, and 5, respectively. Mean AST value of the second period was 70 (±51.6) U/L, 28 (±14.4) U/L, and 25.7 (±17.9) U/L. During the last period, mean AST value of the LL group was 85 (±83.7) U/L, 38.4 (±20.87) U/L, and 32.7 (±35.7) U/L at days 1, 3, and 5, respectively. Compared with the first period, at day 1, AST levels of the second and third period were significantly lower (P < 0.02; Fig. 2). The different types of donor surgery during the last period presented no significant differences of AST values (FR group, mean AST, day 1, 75.1 (±28.6) U/L; day 3, 48.1 (±14.1) U/L; day 5, 41.9 (±35.7) U/L; P > 0.05; Fig. 2).

FIGURE 2. AST level after LL donation in the 3 periods (period 1 [black square] versus period 2 [black triangle] versus period 3 [white circle]). On the right side, postoperative AST level in the LL group (white circle) compared with the FR group (black rhomb). *P < 0.05, ANOVA Bonferroni test.

Perioperative Complications

Of 165 donor surgeries, 1 case of early death was seen in 1993, and a total of 61 complications were observed. Sixty-three percent of the total donor population had no complications. Approximately one third of the donors had minor complications. The total incidence of grade I and grade II complications was 15.7% and 10.9%. Approximately 1/10 of the donors had major complications such as grade III (10.3%) or grade IV (1.2%) complications. A complete list of the observed morbidity in the whole series is shown in Table 7.

TABLE 7. Overall Complications of LDLT, 1991 to 2003

In 6 cases, reoperations were necessary. In 1992, in 1 case, reoperation was performed because of a major bile leak at postoperative day 9. Another reason for reoperation was an accidental injury of the femoral artery during harvesting of the saphenous vein in 1991. In 1993 and 1998, 2 reoperations were necessary because of incisional hernia at postoperative days 218 and 258. In 1993, in 1 case, keloplasty under local anesthesia was performed at postoperative day 202. In 1994, 1 patient showed acute neurologic deterioration hours after the end of the donor procedure. He had developed cerebral hypertension caused by ventricular obstruction by an ependymoma. The tumor was resected successfully, and the patient recovered fully. None of the hemiliver donors underwent reoperation.

The most frequent grade I complications were superficial wound infections (n = 7), followed by nerve compression (peroneal/radialis paresis, n = 2). Minor bile leaks (n = 2) treated conservatively, acute gastritis (n = 2), and urinary retention (n = 2) were observed. Other causes of grade I complications were a pneumatothorax (treated conservatively), a toxic exanthema, a keloid, a subcutaneous wound granuloma, Horner syndrome, postoperative myalgia, and 1 case each of intra-abdominal hematoma after heparinization and effusion of the tympanic cavity.

The most common grade II complications were urinary tract infection (n = 6), infections of unknown origin (n = 2), pneumonia (n = 1), cholecystitis (n = 1), and an infected bilioma, which was treated conservatively using antibiotics (n = 1). Other causes of grade II complications were postoperative bleeding (n = 4) requiring heterologous blood transfusion, locally controlled venous thrombosis without thrombembolic complications (n = 2), peptic ulcer (n = 1), and pericardial effusion (n = 1).

The most common grade III complications were deep wound infections: liver abscesses/infected bilioma (n = 5), which in all cases required percutaneous drainage using a Sonnenberg catheter. Further important causes of grade III complications were major bile leaks requiring endoscopic therapy (n = 2) or relaparotomy (n = 1) and incisional hernias (n = 2). Complicated venous thrombosis concomitant with pulmonary embolism, complications of vascular surgery (requiring surgical therapy), a pneumatothorax (requiring drainage), and an extended intrapleural effusion (requiring pleurocentesis) were observed in 1 case each. The ependymoma (n = 1) was classified as a type III complication.

Grade IV complications were observed in 2 cases. In 1 case, venous thrombosis concomitant with pulmonary embolism occurred, which led to the death of 1 donor. In another case, a de novo hepatitis B infection of 1 donor was observed. The donor did not receive any blood products, and the cause of infection was not reproducible. Fortunately, there was no chronic course.

Within the first period, 40 complications, consisting of 1 grade IV, 12 grade III, 12 grade II, and 15 grade I complications, were observed after LL harvesting (n = 49). During the second period, a total of 12 complications occurred (n = 47), consisting of 1 grade IV, 2 grade III, 5 grade II, and 4 grade I complications. Within the third period, 4 complications (1 grade II and 3 grade I) occurred after LL harvesting (n = 39). Corresponding to this result for the LL group, there was a significant reduction of all grades of complications over time (Fig. 3).

FIGURE 3. Complications according to our standardized living donor classification of morbidity. Period 1 (black column), 1991 to 1995 (n = 52), versus period 2 (white column), 1996 to 2000 (n = 47), versus period 3 (gray column), 2001 to 2003 (n = 39). *P < 0.05, χ² test.

Particularly within the last period, no severe complication occurred in the LL group. The total number of complications after hemiliver donor surgery was 10 (first period, 5; second period, 5). During the last period, 1 grade III, 2 grade II, and 2 grade I complications occurred after hemiliver donor surgery. Compared with the equal period, there was no significant difference of complication rate between the hemiliver and left lateral groups.

Hospital Stay and Length of Stay in Intensive Care Unit

The donor length of stay in the ICU ranged from 1 to 5 days. The median length of stay in the ICU was 1 day, with no significant differences between the different periods or types of harvested grafts.

The total hospital stay of the living donors ranged between 3 and 35 days. In period 1, the mean hospital stay after LL harvesting was 11.4 (±4.78) days. During periods 2 and 3, the mean hospital stay after the same type of donor surgery was 7.3 (±2.98) and 5.9 (±1.89) days, respectively. The reduction in hospital stay between periods 1 and 3 reached statistical significance (P < 0.0001). The mean hospital stay for hemiliver donor surgery was 8.75 (±3.01) days. Within the same period, the hospital stay for hemiliver donors was significantly longer than for LL donors (P < 0.041).

Mortality and Morbidity

From 1991 to 2003, 1 case of early death was experienced after donor surgery.²⁹ Two days after uneventful LL resection, the donor presented with severe hypoxia and tachycardia after sudden collapse. The donor died of massive pulmonary embolism despite emergency surgical thrombectomy.

We observed a late donor death caused by amyotrophic lateral sclerosis 3 years after uneventful left lateral lobe donation.

Of 165 living related donor operations, the total morbidity rate was 27.3%. During period 1, the total morbidity rate was 53.8%; after LL harvesting, it was 53.1%. During periods 2 and 3, it dropped to 23.4% and 9.2% (Fig. 4). Morbidity after hemiliver donation (9.2%) was not significantly different from that after LL lobe donation in the same period (8.1%).

FIGURE 4. Morbidity after LL harvesting (black column) and hemiliver donor surgery (white column) between 1991 and 2003.

DISCUSSION

The donor's safety is the primary concern in LDLT. Despite careful donor selection, mortality after LDLT has occurred in Europe,¹⁹ the United States,⁸ and Japan.³⁰

In our series, we experienced 1 donor death at the beginning of our program (case 30) caused by fatal pulmonary embolism on day 2 after harvesting of the left lateral lobe in an obese young mother who did not stop smoking and was taking contraceptive medication. Including our case, there have been at least 2 cases of donor death from pulmonary embolism in pediatric LDLT, both in women with a history of smoking.^31–33 The registry of Japan also reported an incidence of pulmonary embolism. Four living liver donors developed pulmonary embolism, although without any mortality.²⁰ Another 3 cases of pulmonary nonfatal embolism in 2 donors for adult recipients and 1 for a pediatric recipient underline the need for extensive screening for procoagulation disorders and other known risks of deep vein thrombosis.³⁴ Therefore, we now accept only donors who stop smoking and taking contraceptives a minimum of 6 weeks before the operation. This criterion is based on reports that oral contraceptives and smoking are associated with an increased risk of venous thrombembolic disease.^25–27 Furthermore, we do not accept living donors with a BMI of more than 30 or with procoagulation disorders. With this strict policy and systematic use of prophylactic measures, like perioperative low-molecular-weight heparin, intraoperative intermittent calf compression, and mobilization of the donor only after exclusion of deep vein thrombosis by Doppler ultrasound on the first postoperative day, we managed to avoid further serious thromboembolic complications. The cumulative world experience with left lateral segmentectomies now exceeds 2000 cases, with a donor mortality in the range of 0.1%.³⁵

Transplantation of the right hepatic lobe from live donors proved technically feasible in the mid-1990s. The operation has saved lives but requires a new level of tolerance for risk to the donor.³⁰ The latest results on mortality from Europe, Asia, and the United States point to a higher risk of mortality after donation of the right liver lobe. Analyzing the reasons of death after adult LDLT, liver insufficiency is the main cause of further complications. In Europe, progressive postoperative liver insufficiency led to the death of the donor in 1 case. In the United States, 1 right lobe donor received liver transplantation at p.o. day 8 caused by Budd-Chiari syndrome, and a second right lobe donor was listed for cadaveric transplantation because of subfulminant liver insufficiency.⁸ Fortunately, the second donor in the United States was removed from the waiting list because of an improvement in liver function. With regard to the results of our postoperative liver function score, it must be emphasized that every patient experiences a phase of liver insufficiency for approximately 2 days. Within this transient phase, the patient is sensitive to developing bleeding complications or significant liver failure. Therefore, preoperative measurement of the future liver remnant combined with preoperative estimation of liver quality, especially the fatty degree of the liver, is a key investigation before right lobe donation.

CT and magnetic resonance imaging (MRI) have been shown to detect significant fatty degeneration of the liver with a high specificity but low sensitivity.^36,37 Minor fatty changes of the liver (<30%) have been shown to be significant for the donor after right lobe grafting,³⁸ especially if the remnant liver volume is between 30% and 35%.

At the moment, there is a lack of knowledge about morbidity from donor evaluation. In our series, we experienced 1 case of intrahepatic AV fistula and 1 case of intrahepatic small hematoma in the right lobe after percutaneous liver biopsy, which complicated right lobe grafting in the recipient. Additionally, we had to abandon 3 donor operations because of unexpected fatty degeneration, a too small remnant liver volume, and, in the third case, a hypoplasia of the left hepatic artery. The underestimation of the fatty change of the liver despite liver biopsy underlines the possibility of a sample error. The potential donor with the too small remnant future liver volume showed a demarcation of segment IV after intraoperative occlusion of the right portal vein and right hepatic artery, which underlines the need for volume measurement based on the portal tributaries and not on liver veins. In case 3, the diameter of the left hepatic artery was overestimated by angiography. We suggest that morbidity, which derive from the donor evaluation itself, should be reported in further registries about living donation.

Morbidity is also an important concern of the potential donor and the physicians involved in the procedure. Potential morbidity is equal to what is known from conventional liver resections (bile leakage, bleeding, pleural effusion, pneumonia, wound infection, urinary tract infection, bowel obstruction, incisional hernia). The main problem is the uniform definition and recognition of these complications. For example, the Registry of the Japanese Liver Transplantation Society defines bile leakage as needing a postoperative hospital stay of 30 days or more and needing any intervention (surgical or radiologic).²⁰ Other surveys do not publish any definition of bile leakage.⁸ Therefore, we modified the standardized classification of postoperative morbidity developed by Clavien et al²² to the need of the living donor.

Using this classification, we experienced a total of 61 complications in 37% of the donor population. At the beginning of our program, the leading morbidity was superficial wound infection and urinary tract infection (grade I and II). These complications were eliminated during the last period by systematic perioperative antibiotic prophylaxis, removal of the urine catheter within the first postoperative day, and shortening of the total operation time.

Three donors (3%) developed an abscess of segment IV and 6 donors a bilary complication on the cut surface after LL donation during the first 2 periods. Bilary complications after LL harvesting did not occur in our center after the operative technique was changed. We moved behind the left portal vein for transection of the left hilar plate, thus preserving segment IV bile duct and artery. After left lobe (segment II–VI) harvesting, a transient small bile leakage occurred in only 1 case, most likely from the stump of the closed segment I bile duct. After right lobe grafting, we did not detect any biliary complication. According to reports from the United States, Japan, and Korea, biliary complications were the most frequent morbidity after right lobe harvesting. Umeshita et al²⁰ reported that biliary complication occurred with an incidence of 4% in Japan. In the United States and Seoul, the incidence of biliary complications was 6% and 7%, respectively.^8,39 According to the ELTR data, biliary complications occurred in 8% of the cases after adult donor surgery in Europe.¹⁹ Surgical tools to reduce the risk of bile leakage are a meticulous transection of the hilar plate without leaving undrained segments, and a careful oversewing of the stump of the entire hilar plate using fine sutures (6/0 or 7/0 monofilament resorbable).

Intraoperative or postoperative bleeding is a frequent morbidity in liver surgery, and homologous blood transfusion should be considered as donor morbidity. When starting our LDLT program, we performed parenchymal transection in the clamp fracture and finger fracture techniques. With the introduction of devices for parenchymal transection (Ultracision and CUSA) in periods 2 and 3, we significantly reduced the need for intraoperative and postoperative blood transfusion. Malago et al⁴⁰ also reported a significant reduction of blood loss since refining the parenchymal transection technique. In contrast, Takayama et al⁴¹ reported no significant reduction in blood loss during liver resection by using the CUSA compared with the clamp transection technique.

Harvesting of hemilivers is associated with a significantly larger resection plane, thus increasing the amount of blood loss. The regular appearance of liver insufficiency after right lobe grafting in addition to significant blood loss during donor surgery leads to a dangerous hypocoagulation stage. Therefore, parenchymal transection with minimal blood loss is especially imperative in right lobe grafting.

In LDLT, minor morbidity should not be neglected, because minor complications are able to disturb the healthy donor significantly. In our series, we observed 24 minor morbidities (grade I) throughout the entire study population, with a significant reduction in the second period as a result of a standardization of postoperative care. Additionally, we reduced the incision of the abdominal wall, moving from a routine Mercedes incision to a simple median upper laparotomy for harvesting of the LL and an inverted L-shaped incision for harvesting of the right lobe. Further reduction of the abdominal wall incision would be possible only with the help of minimally invasive techniques.⁴²

Postoperative morbidity did not differ between the LL group and the hemiliver group in the last period (8.1% vs. 9.2%). In Europe, 34% of the donors after right lobe grafting and 10% after left lateral donation developed complications.¹⁹ Donor morbidity was 14.5% in the United States after harvesting of the right graft.⁸ The registry from Japan has published morbidity after right lobe, left lobe, and left lateral grafting of 19%, 12%, and 8.2%, respectively. The morbidity after right lobe harvesting was significantly higher compared with the other (P < 0.0001).²⁰ Our adult living donor transplantation program has benefitted from our experience of pediatric living donor transplantation and our vast experience in segmental liver transplantation and from our high volume of daily liver resections.

From the beginning of our LDLT program to date, the morbidity rate and length of hospital stay have decreased significantly. The Chicago group also experienced a significant reduction of complications while performing 100 pediatric living donor transplantations.¹⁷ The center and surgeon volume effects, which are described for other surgical procedures,^43,44 may contribute to this development. We agree that the effect of learning is an important aspect that must be recognized. Appropriate training should allow new centers to start with a minimal learning curve. Additionally, sufficient center volume will allow for a general tendency of fewer complications.

This study, like most others, has reviewed only short-term morbidity after LDLT. In the future, detailed studies of the late follow-up of living donor patients are very important to complete the picture and collect evidence-based information for future potential donors.

CONCLUSIONS

Reliable and accurate assessment of morbidity is an important task of a transplantation center, because safety of the donor is the key issue in LDLT. The published morbidity after partial liver donation is hardly comparable because of the variation in definition. To overcome this problem, we adapted the well-standardized classification of morbidity from Clavien et al²² to the need of the living donor.

As a result of an ongoing improvement in surgical technique, selection of the donor, and postoperative care, we were able to reduce perioperative morbidity significantly from 53.8% at the beginning of the program to 9.2% in the last period, despite introducing right lobe grafting. In a single center, morbidity after living liver donation strongly correlates with center experience. Despite the additional risks associated with temporary reduction of liver function, this experience allowed us to bypass part of the learning curve when starting right lobe donation. Specific training of the surgical team and coaching by an experienced center should be implemented for centers starting this procedure to avoid the learning curve. The proposed standardized morbidity classification could help in comparison of outcome after LDLT among different centers.

Discussions

Dr. Slooff: It is a privilege to comment on this paper from the distinguished Hamburg center, and I thank Dr. Broering for sending me the manuscript beforehand. The conclusions of the paper have a solid basis. You have convincingly shown that experience counts. In itself this is not new in surgery but it is still a neglected point in discussions with policy makers, insurance companies, etc.

A strong point in your paper is that you use a morbidity classification system. I would strongly support its use in such studies. Noteworthy is that the biliary complication rate after right lobe donation is much lower than reported in the literature.

There is place for some remarks. I think you need to put the results you have presented into perspective. Eighty percent of the resections were left lateral lobe resections. So it may be questionable if the results pertain also to full right lobe donation.

Secondly you should have stressed the environment were the results were obtained. You are a center with a reputation in partial and split liver grafting and also with a huge experience in Hepato-Biliary surgery. This matters and certainly has been a determining factors in your results.

In your results you refer to some problems with liver function, especially in right lobe grafting. You stated that a liver biopsy might discriminate between grafts performing good or poorly. I doubt if that is a valid statement. You did not mention the results of donor liver biopsies but I suspect that most of the liver biopsies in the donors were normal, otherwise you would not have used them. Any use of a biopsy in discriminating good or bad grafts is therefore not possible. My question is, if we should not use Indocyanine Green Testing or MEGX testing in combination with volumetry to provide adequate information about functional capacity of the graft and the remnant?

If I interpreted your numbers well there was still a relatively high thrombosis rate after right lobe retrieval. I wonder if that has to do with your operative technique keeping the central venous pressure low in order to reduce blood loss. This may cause underfilling in patients and induce thrombosis.

I have another question partially related to your paper. When your experience grows does that have consequences also for your donor selection. Has it become less strict and easier for your potential donors?

Dr. Broering: Thank you very much for your comments and questions. We perform liver biopsy routinely only prior harvesting of either the full right or full left lobe and we consider a steatosis of more than 10% as a contraindication for donation of the right liver lobe. We do not perform MEGX test or indocyanin green clearance, since these tests can not predict functional reserve of the remnant liver.

We experienced 2 deep vein thrombosis after right lobe donation in the beginning of our adult living donor program and one was complicated by pulmonary embolism. Since than we extended our prophylactic concept to intraoperative intermitted calf compression in addition to s.c. low molecular weight heparin. Since than we did not observe any further deep vein thrombosis on routinely performed Doppler ultrasound examination on the first postoperative day. We need to further evaluate if the low CVP during parenchymal transection is a risk factor for deep vein thrombosis.

At this moment we strictly follow our donor criteria. But maybe in the future by having more experience we will do so.

Dr. Clavien: Dr. Broering, I would like to congratulate you to provide such detailed outcome data in this field where we are under tight scrutiny from the outside. I have 3 questions. My first question relates to the classification of complications. In your presentation you mentioned a number of grade 3 complications. Originally grade 3 complications were designed to cover negative events associated with lasting disability or the development of malignant disease, while grade 2 complications addressed potentially life-threatening complications such as pneumonia, bile leak, abdominal abscess, etc. It seems to me that most of your complications were grade 2 rather than grade 3. Recently, we revisited our complication system (in press, Ann Surg) with an attempt to minimize confusion among the various types of complications. My second question deals with the dramatic improvement in recovery of liver function after surgery while gaining experience. How do you explain this observation? Did you use different techniques to divide the liver parenchyma? Thirdly, you concluded that there is a need for coaching small centers initiating such program. How do you foresee this in practice?

Dr. Broering: One of the most important goals of our study was to classify postoperative complications after living donation of a part of the liver to facilitate comparing of donor outcome. Therefore we decided to adopt your morbidity classification, which was originally designed for morbidity after liver transplantation to the special situation of a healthy living donor. In general significant morbidity was graded higher in our morbidity classification. For example complications with lasting disability were classified in our classification as grad 4 complications. In fact postoperative liver function improved over the years. The reasons are in our experience the significant reduction of blood loss during parenchymal transsection by using the CUSA, respecting the segment 4 artery in left lateral lobe harvesting and shortening of the operation time. Our series clearly demonstrates a learning curve with living donation. To prevent this learning curve in centers initiating such programs we suggest external personal coaching of these teams during the first cases.

Dr. Senninger: Congratulations and 2 short questions. The first question is relating to the late deaths: this diagnosis that you mentioned is that really procedure related? I would like your definition of late deaths as suitable to your procedure: if you wait long enough the death rate will be 100%.

The second question is how did you make your decision to switch, did you use evidence based methods for that? Was it personal impression or expert opinion? I would like to know that, because it has impact on what new centers will choose.

Dr. Broering: At this moment late death was defined as any death of a donor after the procedure. Prior starting hemi-liver donation, we performed a study on different devices for parenchymal transsection. In this internal study, the CUSA turned out to facilitate parenchymal transsection with the lowest blood loss.

Dr. Margreiter: In periods 1 and 2 you had more female donors than male donors. However, in period 3 you had significantly more male donors. Was this simply due to the fact that you started doing right lobes or were you able to change the attitude of Germans toward organ donation in general.

Dr. Broering: In fact the introduction of adult living donor liver transplantation contributed to the higher number of male donors in the third period.

Dr. Bismuth: Last year your group presented a paper concluding that there is no more use of living donor transplantation in children. I agreed with this conclusion. However, today you conclude that harvesting left lobe in living donor is a good training for harvesting the right liver for adult.

My second question is on the morbidity of the hemi-liver grafts. You put together in the same group morbidity in relation with left lobe and right hemi-livers. Is there any difference between left lobe and right hemiliver morbidity?

Dr. Broering: The short and long term results after split liver transplantation and transplantation of grafts from living donors are still the same. The reason for the ongoing need for living donor liver transplantation for pediatric recipients in Germany is that the authorities failed to implement mandatory splitting of good quality livers. In our small series morbidity after harvesting of the left hemi-liver was not significantly different from that after harvesting of the right hemi-liver.