The incidence of morbidity and mortality after living donor liver transplantation (LDLT) is not well understood because reporting is not standardized and relies on single-center reports. Aborted hepatectomies (AHs) and potentially life-threatening near-miss events (during which a donor's life may be in danger but after which there are no long-term sequelae) are rarely reported. We conducted a worldwide survey of programs performing LDLT to determine the incidence of these events. A survey instrument was sent to 148 programs performing LDLT. The programs were asked to provide donor demographics, case volumes, and information about graft types, operative morbidity and mortality, near-miss events, and AHs. Seventy-one programs (48%), which performed donor hepatectomy 11,553 times and represented 21 countries, completed the survey. The average donor morbidity rate was 24%, with 5 donors (0.04%) requiring transplantation. The donor mortality rate was 0.2% (23/11,553), with the majority of deaths occurring within 60 days, and all but 4 deaths were related to the donation surgery. The incidences of near-miss events and AH were 1.1% and 1.2%, respectively. Program experience did not affect the incidence of donor morbidity or mortality, but near-miss events and AH were more likely in low-volume programs (≤50 LDLT procedures). In conclusion, it appears that independently of program experience, there is a consistent donor mortality rate of 0.2% associated with LDLT donor procedures, yet increased experience is associated with lower rates of AH and near-miss events. Potentially life-threatening near-miss events and AH are underappreciated complications that must be discussed as part of the informed consent process with any potential living liver donor. Liver Transpl 19:499–506, 2013. © 2012 AASLD.
Living donor liver transplantation (LDLT) is an accepted treatment for individuals with end-stage liver disease. The results are equivalent or superior to those achieved with deceased donor liver transplantation.[1-4] In areas in which deceased donor organs are in short supply or culturally unacceptable, LDLT assumes greater importance as the only effective treatment for liver failure. The obvious benefits for the recipient and the probability of death without a transplant may produce a coercive atmosphere. Donors may face familial or societal pressure to provide a lifesaving organ, and it is against this background that transplant professionals must provide potential donors with the information necessary for them to make an informed decision regarding donation.
Providing potential liver donors accurate and timely information regarding the risks associated with hepatic lobe donation is hampered by the lack of standardized reporting systems and the emphasis on patient/donor death and graft loss as the ultimate benchmarks of successful LDLT. Reports of adverse events in donors vary widely in terms of which events are considered serious enough to be reported and in the incidence of those that are reported.[5-7] Potentially life-threatening events such as donor aborted hepatectomy (AH) or intraoperative hemorrhage are rarely reported if they are successfully managed. These near-miss events constitute a real risk to LDLT donors, but we currently are unable to provide them with a realistic estimate of their occurrence.
We undertook a Web-based, worldwide survey of liver transplant programs with the aim of improving our knowledge of the actual incidence of events that increase risk for hepatic lobe donors. Participants were asked to report the percentage of specific events occurring in their programs and to rate their severity according to the validated Clavien scale. For near-miss and Clavien grade III-V events, details of the incidents were requested. There are obvious shortcomings associated with self-reported data. However, we believe that this report constitutes the most comprehensive assessment of living liver donor risks currently available and will provide potential living liver donors and the transplant teams advising them with more accurate information as they consider the decision to donate.
MATERIALS AND METHODS
Surveys were sent to all liver transplant programs known to have performed LDLT at least once. Program lists were obtained from the published literature, the American Society of Transplant Surgeons, the Japanese Liver Transplant Society, the European Liver Transplant Registry, and the China Liver Transplant Registry. Additional programs were included if they were known by us to be performing LDLT. Respondents were assured that the results would be reported in a blinded fashion. After initial blast e-mail requests for participation, follow-up e-mails (n = 4), phone calls, and personal interactions were used to increase survey completion.
SurveyMonkey, a commercially available Web-based tool, provided the format for our internally created survey. The length of the survey was respondent-driven. At the end of sections describing critical events, participants were asked if they had additional relevant cases. An affirmative answer generated an additional report form, whereas a negative response moved the respondent to the next section.
Participants were asked to provide program demographics, donor evaluation information, donor morbidity and mortality data, and the incidence of near-miss events and donor AH. Near-miss events were defined in the e-mail containing the link to the survey as follows:
We are particularly interested in the occurrences of “near miss” events. For the purpose of this survey, we are defining “near miss” as an event or events with potentially fatal consequences. Severe bradycardia or displacement of an arterial clip are examples of such events. We would like you to report this type of event even though the situation was successfully managed and the donor suffered no lasting ill effects.
To help to standardize the morbidity data, respondents were asked to use the Clavien system, which was modified for living donors (Table 1) and was provided with the survey, for all events reported. We reviewed the reported events and designated as near-miss events those that were nonfatal but had a significant risk of death.
Table 1. Clavien Surgical Morbidity Scale Modified for Living Donors
|Grade I. Any deviation from the normal postoperative course without the need for pharmacological treatment or surgical, endoscopic, and radiological interventions. The allowed therapeutic regimens are drugs such as antiemetics, antipyretics, analgesics, and diuretics; electrolytes; and physiotherapy. This grade also includes wound infections opened at the bedside.|
|Grade II. Requiring pharmacological treatment with drugs other than those allowed for grade I complications. Blood transfusions and total parenteral nutrition are also included.|
|Grade III. Requiring surgical, endoscopic, or radiological intervention.|
|IIIa. Intervention not under general anesthesia.|
|IIIb. Intervention under general anesthesia.|
|Grade IV. Life-threatening complication (including central nervous system complications such as brain hemorrhaging, ischemic stroke, and subarachnoid bleeding but excluding transient ischemic attacks) requiring intermediate care/intensive care unit management.|
|IVa. Single-organ dysfunction (including dialysis).|
|IVb. Multiorgan dysfunction.|
|Grade V. Death of the patient.|
Statistical analysis was performed with SPSS for Windows (PAWS Statistics 18). The overall variance within groups was assessed with an analysis of variance. Categorical variables were assessed with the chi-square test or Fisher's exact test where applicable. Means between groups were compared with the Student t test. A P value < 0.05 was considered significant.
Appropriate institutional approval was obtained before the start of the study.
Survey Completion and Geographic Distribution
Survey completion was defined by American Survey Research Organization standards, according to which >80% of items must be completed. Surveys were sent to 148 programs and were completed by 71 (48%); 64 currently perform LDLT. The respondents represented 21 countries, and the programs were distributed as follows: 39 in North America, 13 in Asia, 13 in Europe, 4 in South America, 1 in the Middle East, and 1 in New Zealand. Most of the current programs perform both adult-to-adult and adult-to-pediatric LDLT procedures (n = 54), although 5 centers perform only adult procedures, and 5 centers perform only pediatric procedures.
Center Volume and Graft Type
The experience of the centers varied widely and ranged from 2 to 2785 LDLT procedures. For analytical purposes, the programs were categorized as (1) low-volume (≤50 LDLT procedures), (2) moderate-volume (51-200 LDLT procedures), or (3) high-volume (>200 LDLT procedures). Group 3 included 6 centers that performed more than 500 LDLT procedures and 2 centers that performed more than 1000 LDLT procedures.
The centers varied greatly in the length of time that they had performed various types of liver transplantation. The start dates for any type of liver transplantation ranged from 1983 to 2007; the start dates for centers performing pediatric LDLT ranged from 1990 to 2007. The earliest centers performing adult-to-adult LDLT began in 1997, and the most recent began in 2007.
Right hepatic lobes were the most common graft type, with 95.6% reporting their use in some or all procedures. More than 50% of reporting programs had performed >60 right lobe procedures. The use of left hepatic lobe or left lateral segment graft types was reported by 77% of respondents. Thirty-eight adult programs (53.5%) reported the use of left lobes in appropriate recipients. Extended right hepatectomy or right trisegmentectomy grafts with the inclusion of the middle hepatic were relatively infrequent (8 programs or 11.3%).
The survey revealed 23 donor deaths among 11,553 donor hepatectomy procedures (0.2%); 18 were reported directly by centers, and 5 additional reports were received from the European Liver Transplant Registry (Table 2). Single deaths were reported by 8 centers, 2 centers reported 2 deaths each, and 2 centers reported 3 deaths each. These deaths spanned the entire spectrum of donor follow-up and included 4 deaths unlikely to be related to the donation surgery (2 deaths from lung cancer at 22 months and 3.4 years, 1 from asthma at 5 years, and 1 from myocardial infarction at 6 years). With these deaths excluded, the mortality incidence was 0.16%. The relationship of 2 additional donor deaths to donation was questionable (suicides at 4 and 5 years after donation). The majority of the deaths (15/23) occurred within 60 days of donation and ranged from an intraoperative death to suicide at 60 days. Two donors died when liver transplants performed to rescue them from acute hepatic insufficiency failed. There was no association between the type of donor hepatectomy (right versus left versus left lateral segment) and the incidence of death (P = not significant for all groups). Calculations performed without the deaths unlikely to be related to donation (3 right lobes and 1 left lobe) also failed to show significance between groups. Another 11 donor deaths not captured in the survey have been reported in the literature. They are listed in Table 2.
Table 2. Causes of Death
|Intraoperative|| || || || |
|Bleeding/cardiac failure/cardiac arrest||North America||Right||0 days|| |
|Early postoperative (≤60 days)|| || || || |
|Anaphylaxisb||North America||Left lateral segment||1 day||Pomfret et al.|
|Pulmonary embolismc||Europe||Left lateral segment||2 days||Malagó et al.|
|Gastric necrosis (Clostridium perfringens)||North America||Right||3 days||Miller et al.|
|Cardiac arrest||North America||Right||4 days|| |
|Myocardial infarction||Asia||Right||10 days||Polido et al.|
|Sepsis/multiorgan failurec||Europe||Right||11 days||Boillot et al. and Adam et al.|
|Sepsis/multiorgan failurec||Europe||Right||21 days||Adam et al.|
|Fall at home||Asia||Right||28 days|| |
|Cardiac failure/liver transplantationc||Europe||Right||32 days||Adam et al. and Ringe et al.|
|Subarachnoid hemorrhage||Asia||Right||42 days|| |
|Multiorgan failurec||Europe||Right||49 days||Adam et al.|
|Complications of multiple myeloma||Europe||Right||56 days||Melloul et al.|
|Bile peritonitis/sepsis/multiorgan failure||Middle East||Right||60 days||Khalaf et al.|
|Suicide||North America||Left||60 days|| |
|Late postoperative (>60 days)|| || || || |
|Duodenal–inferior vena cava fistula (ulcer)/air embolism||Asia||Right||2.3 months||Chan et al.|
|Nonalcoholic steatohepatitis/liver failure/liver transplantation||Asia||Right||9 months||Akabayashi et al.|
|Lung cancer||Asia||Right||22 months|| |
|Lung cancer||Asia||Right||3.4 years|| |
|Suicide||South America||Right||4 years|| |
|Suicide||South America||Left lateral segment||5 years|| |
|Asthma||Asia||Right||5 years|| |
|Myocardial infarction||Asia||Left||6 years|| |
| ||Previously Reported Deaths Not Captured in Survey|
|Causes Early postoperative (<60 days)|| || || || |
|Cardiac arrest/persistent vegetative state||Asia||Right||2 days||Srinivas|
|Cardiac arrhythmia||South America||Right||2 days||Coelho et al.|
|Massive bleeding||Europe||Right||4 days||Broering et al.|
|Subarachnoid hemorrhage||South America||Right||7 days||Wiederkehr et al.|
|Bile leak/sepsis/multiorgan failure||North America||Right||3 weeks||Zagier et al. and Ochs et al.|
|Berardinelli-Seip/liver transplantation/cardiac failure||Europe||Right||32 days||Malagó et al.|
|Pulmonary embolism||North America||Left||Unknown||Renz and Busuttil|
|Late postoperative (>60 days)|| || || || |
|Suicide||North America||Right||22 months||Trotter et al.|
|Suicide||North America||Right||23 months||Trotter et al.|
|Acute Budd-Chiari syndrome||Europe||Unknown||Unknown||Trotter et al.|
AH was defined as any procedure that was stopped after the donor entered the preoperative holding area. One hundred thirty-six AHs (1.2%) were reported by 44 of the responding programs (62%). The majority occurred after the incision but before the bile duct transection (n = 98 or 72%). Aborted cases were also reported after the hepatic transection (n = 12), in the preoperative holding area (n = 12), and after anesthesia but before the incision (n = 8). After AH, 45% (n = 61) eventually donated, whereas 55% (n = 75) did not. Procedure-related complications were experienced by 13% of the patients after AH, with incisional hernias and wound infections occurring most frequently.
The reasons for most AHs were donor-related (n = 106 or 77.9%), but 30 (22.1%) were related to recipient events (Table 3). The most common donor-related reasons were unexpected vascular or biliary anatomies and unexpected pathologies. The most common recipient-related reasons were hemodynamic instability and recipient malignancies more advanced than expected.
Table 3. Reasons for Donor AH
|Vascular and biliary anatomy||2||2|
|Intraoperative liver injury||2||2|
|Small graft or remnant volume||3||3|
|Recipient Reasons||Centers (n)||Cases (n)|
Survey results indicated that 126 near-miss events occurred in 125 patients, so the incidence was 1.1% (Table 4). Hemorrhaging requiring surgical intervention was cited most frequently, and this was followed by thrombotic events, biliary reconstruction procedures, life-threatening sepsis, and iatrogenic injury to the bowel or vasculature. Six instances of transient hepatic insufficiency were reported.
Table 4. Near-Miss Events
|Reoperation for bleeding||20||39|
|Thrombotic events|| || |
|Portal vein thrombosis||10||10a|
|Inferior vena cava/hepatic vein thrombosis||4||5|
|Reoperation for intra-abdominal sepsis||6||7|
|Transient liver insufficiency||5||6|
|Transient hemodynamic instability||1||1|
|Vascular reconstruction for injury||4||4|
|Reoperation for bowel injury||3||3|
|Transplantation|| || |
|Massive intraoperative bleedingc||5||5|
|Anaphylaxis/systemic inflammatory response syndrome||2||2|
|Respiratory failure requiring mechanical ventilation||3||3|
|Reoperation for diaphragmatic hernia||2||2|
|Parietal transient ischemic attack with motor weakness and foot drop||1||1|
|Reoperation for perforated gastric ulcer||1||1|
The overall reported donor morbidity rate was 23.9% ± 13.9%, with most events (85.8%) occurring within the first 30 postoperative days. The most common complications were bile leaks, wound infections, incisional hernias, and unplanned surgical re-exploration. The majority were mild and self-limited (Clavien grade I or II). However, 5 donors required transplantation after donation: 4 required livers, and 1 required a kidney (0.03% and 0.009%, respectively).
Among the donors requiring liver transplantation, 2 cases were secondary to hepatic failure related to hepatic vein thrombosis. Notably, 1 of these cases was heterozygous for factor V Leiden. The other 2 liver recipients died despite transplantation. The kidney recipient developed contrast nephropathy after angiography.
Influence of Program Volume on Donor Outcomes
Table 5 shows the influence of program volume on donor outcomes. The overall morbidity rate and the incidence of Clavien grade III-V complications did not appear to be related to the center volume. Similarly, the incidence of donor death and the need for transplantation were not related to the center volume. However, the incidence of near-miss events was significantly increased at low- and moderate-volume centers in comparison with high-volume centers. The incidence of AH was related to the center volume and was significantly higher in group 1 versus both groups 2 and 3 and in group 2 versus group 3.
Table 5. Influence of Program Volume on Donor Outcomes
|Near-miss events||2.9 ± 15.7||1.8 ± 6.5||0.5 ± 0.3||0.11||<0.001||<0.001|
|Overall morbidity||23.2 ± 15.4||24.1 ± 12.2||25.0 ± 15.3||0.81||0.74||0.85|
|Clavien grades III-V||8.1 ± 11.8||8.3 ± 9.4||10.7 ± 12.1||0.95||0.54||0.51|
|AH||3.8 ± 23.1||1.5 ± 0.9||0.7 ± 0.8||<0.001||<0.001||<0.001|
|Transplantation (liver)||0.15 ± 0.2||0.03 ± 0.03||0.02 ± 0.1||0.59||0.73||0.61|
|Death||0||0.2 ± 0.03||0.1 ± 0.9||0.52||0.78||0.40|
Although LDLT is a lifesaving procedure for the recipient, it carries a significant risk of morbidity and mortality for the otherwise healthy donor. We conducted the survey to obtain comprehensive data concerning the incidence and outcomes of adverse events following hepatic lobe donation. We particularly wished to evaluate the prevalence of near-miss events and AHs, which are events rarely captured in standard reports. This information may then be used to provide potential donors with realistic estimates of the risks associated with being a living liver donor. It may also be used to identify areas of practice in need of refinement.
In order to encourage complete disclosure, the respondents were assured that the results would be de-identified and reported only in aggregate and that no individual or center would be identified. Self-reported data have obvious drawbacks such as selective reporting of poor outcomes and open interpretation of what constitutes a reportable event. Nevertheless, we believe that this report contains a representative snapshot of LDLT and its attendant complications as it currently exists. The results represent 64 currently active LDLT centers on 5 continents. Approximately 65% have performed more than 50 procedures, and 12 have performed more than 200 LDLT procedures; the latter group includes 1 center that reported its experience with 2785 completed donor hepatectomies.
The overall response rate was relatively low at 48%, but the nature of the nonrespondents is important. The rate was calculated with the initial mailing list, which included all centers known to ever have performed LDLT rather than those currently performing the procedure. For example, the United Network for Organ Sharing Web site indicates that 23 of the nonresponding US programs performed <5 LDLT procedures in the previous 5 years. Similar data are not available for non-US programs, but it is likely that inactive programs in other locations did not respond. The active centers that did respond represent significant experience in the field, as evidenced by published literature and activity in professional societies. We, therefore, believe that our sample represents currently active and reasonably experienced LDLT centers.
The mortality rate observed in this survey was consistent with the rates previously reported.[11, 17, 20, 22, 24, 31, 32] This survey documented 23 deaths in 11,553 procedures for an incidence of 0.2%. This calculation was based on all reported deaths, including some temporally remote from the donor surgery and probably unrelated. However, because this cannot be established with certainty, our data do not support altering current descriptions of the risk of donor death.
We did not find a correlation between center experience and the incidence of donor death. This is similar to reports from the Adult-to-Adult Living Donor Liver Transplantation Cohort Study consortium but is at odds with reports based on general surgical data suggesting a relationship between hospital/surgeon volume and patient outcomes.[33, 34] However, the living donor population cannot be directly compared to general surgical patients. They are healthy and are undergoing an elective procedure rather than one for cause. There also exist national mandates that require only experienced hepatobiliary surgeons to perform living donor hepatectomy. The reported deaths occurred in all geographic areas and in both moderate- and high-volume programs. Low-volume programs did not report any donor deaths. One possible explanation is that most deaths in this relatively ideal patient population are truly random and that a critical number of cases must be performed before the random event is encountered.
In contrast, the incidence of donor AH did show an association with center experience (Table 4). Prospective donors faced a 1.16% overall risk of experiencing AH. However, donors at high-volume centers experienced significantly fewer AHs (0.7%) than donors at low-volume (3.84%) and moderate-volume centers (1.54%). An analysis of the reasons stated for discontinuing the procedure reveals that many were potentially preventable with more rigorous donor and recipient evaluation protocols. For example, 44 cases were discontinued when the vascular or biliary anatomy was deemed unsafe after the procedure was under way; another 34 cases were stopped when an unexpected hepatic pathology was encountered intraoperatively.
State-of-the-art imaging techniques should be used to visualize vascular and biliary structures before surgery, and a close working relationship with radiologists skilled in these techniques is an important part of ensuring overall donor safety.[28, 35] In our opinion, the use of these techniques has the potential to greatly reduce the occurrence of significant abnormalities discovered at the time of surgery and, therefore, to reduce the risks associated with unnecessary surgery and anesthesia. An unexpected hepatic pathology is rarely found in programs that routinely include liver biopsy as part of the predonation evaluation, but this may not be necessary in all donors. Algorithms for selecting evaluation liver biopsy for patients deemed to be at increased risk exist,[36, 37] and they have reduced or eliminated pathology-related AHs in programs in which they are used routinely.
The aforementioned incidents are most likely distinct from AHs that occur because donors, despite passing rigorous cardiac evaluations, develop HD instability after the induction of anesthesia or donors with no known drug allergies develop anaphylactic reactions after antibiotic administration. The former are more likely associated with programmatic learning curves, whereas the latter are probably random events that cannot be ascribed to a particular cause.
The other category of AH reported was the result of issues arising with the recipient (n = 30) after the initiation of donor surgery. Twelve of the reported cases aborted for recipient reasons were associated with recipient malignancies (n = 10) or infections (n = 2). Some centers routinely explore high-risk recipients before the initiation of donor surgery to ensure the absence of extrahepatic malignancies. This seems to be a reasonable strategy for avoiding the subjection of a healthy donor to unnecessary anesthesia and surgery.
We were especially interested in obtaining data concerning the occurrence of near-miss events, that is, those with the potential for catastrophic outcomes. This type of event is not necessarily captured in complication data but is frequently discussed informally at transplant meetings. For example, a vascular clip displaced during surgery, if it is seen and quickly replaced, is unlikely to be counted as a donor complication. This is in contrast to the same type of event when it is unnoticed and results in a blood loss requiring transfusions. However, the events and the risks associated with them are potentially equivalent.
In this study, 126 near-miss events were reported (approximately 1 in every 92 procedures). There were no differences associated with geographic regions. This rate is likely an underestimation representing those most memorable to the reporting individuals, but this report does represent the first comprehensive report of actual risks faced by donors across various health care systems and practice models. The actual reported events are those commonly reported after liver resection, such as bleeding, biliary injury, and thrombotic events. High-volume centers reported larger numbers of near-miss events, but when they were indexed to the number of LDLT procedures, the rates at high-volume centers were significantly lower than those at either low- or moderate-volume centers. This suggests that a prolonged learning curve, significantly greater than the previously reported 20 LDLT cases, is needed to maximize donor safety.
The LDLT volume did not affect the reported incidence of Clavien grade III-V complications. This is consistent with some published reports,[3, 38, 39] although the complication rate reported here was slightly lower. Respondents were asked for estimates of the incidence of Clavien grade III-V complications instead of the detailed descriptions for deaths or AHs. This type of data is retrospective and subject to variations imposed by each center's definition and classification of surgical complications and explains in part the lower reported rates observed in this study versus contemporaneous reports.[3, 40]
This survey has generated reliable data on near-miss events, AHs, complications, and deaths in living liver donors that can be used to better inform prospective donors. It appears that program experience can significantly reduce the incidence of donor AH and near-miss events, but donor deaths occur randomly with a consistent rate of 0.2%. The living liver donor community must develop a universally accessible reporting system that will allow a critical analysis of all events so that donor safety can be maximized. For example, risks such as the increased incidence of thrombotic events in donors with factor V Leiden have been identified by several groups and are reported in this survey, but testing for the defect is not a routine part of all donor evaluations. It is our opinion that this testing increases the margin of donor safety, but LDLT would benefit from studies designed to validate the clinical utility and cost-effectiveness of donor evaluation protocols. Lessons learned from near-miss events and complications should be consistently classified and systematically shared.