Donor safety is the paramount concern of living donor liver transplantation (LDLT). Although LDLT is employed worldwide, there is little data on rates and causes of ‘no go’ hepatectomies—patients brought to the operating room for possible donor hepatectomy whose procedure was aborted. We performed a single-center, retrospective review of all patients brought to the operating room for donor hepatectomy between October 2000 and November 2008. Of 257 right lobe donors, the donor operation was aborted in 12 cases (4.7%). The main reasons for stopping the operation were aberrant ductal or vascular anatomy (seven cases), unsuitable liver quality (three cases) or unexpected intraoperative events (two cases). Over the median period of follow-up of 23 months, there were no long-term complications of patients with aborted donor procedures. This report focuses exclusively on an important issue: the frequency and causes of no go decisions at a single large volume North American LDLT center. The rate of no go donor hepatectomies should be as low as possible without compromising donor safety—however, even with rigorous preoperative evaluation the rate of donor abortions will be significant. The default surgical position should always be to abort the donor operation if there is an unexpected finding that places the donor at increased risk.
Living donor liver transplantation (LDLT) is an acceptable modality to treat endstage liver disease, particularly in regions where wait times for a deceased donor LTx are high. At our center the need for LTx far exceeds local donation rates. As a result there is a roughly 20–25% rate of death while waiting for an LTx. We have shown that LDLT offers a significant survival advantage to the recipient in our program (1). This advantage has now been confirmed across many US centers, even when wait times are short (2).
Regardless of the potential benefit that LDLT offers to the recipient, donor safety must be the prime concern of any LDLT program. There are significant risks to the living donor, including death (0.2%) and substantial morbidity (40%) (3–5). As a result, LDLT has been controversial since its inception. Adult-to-adult LDLT was introduced almost a decade after LDLT for children, grew rapidly, but has declined in North America since 2001—in part due to a well-publicized donor death in New York (6). At present, LDLT comprises less than 5% of adult LTx, and any further growth of LDLT will depend on defining the optimal recipient and donor characteristics for this procedure as well as broader acceptance and experience in the transplant community.
A number of studies have detailed the general safety of the living donor hepatectomy. Common early complications after living donor hepatectomy include bile leaks (3–9%), bacterial infections (10–12%), incisional hernias (6–9%), pleural effusion requiring intervention (4–5%) and intraabdominal bleeding requiring reexploration (1–4%) (3–5). In our own cohort we also a noted a learning curve with significantly fewer complications after the first 50 donor operations (5).
To minimize donor complications, careful donor evaluation is critical. The evaluation process is designed to reveal any condition that may increase the risk of complications for the donor (3,5). The evaluation process typically proceeds in a stepwise fashion so that unsuitable donors can be identified as early as possible.
However, even with the most extensive and careful donor evaluation there will be cases where problems during the donor hepatectomy necessitate the abortion of the donor operation for the sake of donor safety. Although donor abortions have been discussed in the context of donor morbidity as a whole (3), or in the setting of evaluating radiological techniques (7–14), there is no report, to our knowledge, that focuses exclusively on the rate, the reasons and the complications in patients brought to the operating room for a living donor hepatectomy that have their operations aborted by virtue of unexpected intraoperative findings or events.
Data collection and study subjects
Data for this cohort study were derived from the internal web-based Organ Transplant Tracking Registry (OTTR, HKS Medical Information Systems, Omaha, NE) supplemented by relevant paper-based records of individual living donors. The Institutional Ethics Review Boards and Privacy Boards of the Toronto General Hospital, University Health Network, Toronto, ON, Canada, approved the study. The retrospective cohort data included 257 patients who where eligible and proceeded to surgery for right lobe living donor hepatectomy between October 2000 and November 2008.
Medical criteria of donor suitability
As outlined previously (1,5), donor suitability is determined after completing a multidisciplinary evaluation process that includes blood tests, imaging studies and a psychiatric and medical assessment by professionals who are independent from the transplantation team. Absolute exclusion criteria for donation are: ABO incompatibility, underlying liver disease, positive hepatitis serology, steatosis >10%, inadequate graft size, abnormal biliary anatomy or any medical condition that is considered to pose an increased risk for donor complications.
Anatomical criteria of donor ‘acceptability’
All potential living donors are subjected to careful assessment of the anatomical details of the potential liver graft. Potential donors undergo a liver ultrasound and CT scan of the abdomen, and if these are deemed acceptable proceed to a magnetic resonance cholangio-pancreaticogram (MRCP) for assessment of the biliary tree. Since 2008, if the biliary anatomy is not clear on the MRCP, a CT-cholangiogram is performed. For right lobe donors, we prefer that the graft-recipient weight ratio be greater than 0.8, and that the residual donor liver volume be greater than or equal to 30%. We have accepted up to three separate right bile ducts, but will not proceed with the donor operation if it is felt that the left biliary system could be compromised. A decision about whether or not to recover the middle hepatic vein is made based on the area of the right liver drained by this vein, and by whether the middle hepatic vein drains a dominant portion of the residual left lobe.
Where appropriate, statistical differences between two groups were compared using a paired Student's t-test using SPSS software.
The characteristics of living donor candidates who underwent surgery are depicted in Table 1. Of 257 right lobe donors for adults, the donor operation was aborted in 12 cases (4.7%). In 11 donor procedures an incision was made, but the donor liver was not divided. In one case, the resection was completed after a small arterio-vascular malformation ruptured intraoperatively. In the nine patients whose case was aborted due to aberrant anatomy or unexpected intraoperative events, a cholecystectomy was performed: after determination of liver quality, donors proceed to an intraoperative cholangiogram through the cystic duct. One patient whose right hepatectomy was aborted (due to an anatomical variation) proceeded to a planned open fundoplication for an axial hiatus hernia. In 10 cases, the operation was stopped due to unexpected findings of anatomical variations or poor liver quality, and in two cases due to intraoperative issues that placed the patient at unacceptable risk.
Table 1. Donor characteristics
Right lobe donors, n = 275
Aborted donor cases, n = 12
37.6 ± 11.6
38.3 ± 12.9
Sex (% male)
Body mass index
26.4 ± 3.8
27.5 ± 4.9
Relatedness to recipient
Year of LD
Of the five cases aborted due to unexpected biliary anatomy, one case, early in the series, was aborted when the intraoperative cholangiogram revealed that the parenchymal transection line would traverse three separate right hepatic ducts rather than the two that were anticipated based on preoperative imaging. The decision was made to abort the procedure due to the complexity of the recipient operation. In another aborted case, the intraoperative cholangiogram showed an extremely narrow crotch between the left and the right bile duct system, with a posterior duct running parallel to the resection line (Figure 1). Although this narrow angle was seen on the preoperative MRCP, the implications were not fully appreciated until the day of surgery. There was no radiological angle that opened the angle between the right and left duct on the intraoperative cholangiogram. Although technically feasible, it was felt that this variation placed the left duct at an unacceptable risk, and the procedure was aborted. In two cases, the intraoperative cholangiogram showed an aberrant segment 4 duct (or possibly a large caudate duct) draining from the right hepatic duct, endangering the residual liver drainage after donation. An example of this is shown in Figure 2. As seen in the figure, the aberrant duct was not seen on the preoperative MRCP, highlighting the limitations of this radiological technique. In the final case aborted due to biliary variations, the size of a segment one duct was best visualized with the intraoperative cholangiogram, and the case aborted due to concerns that the duct would need to be sacrificed in a patient with borderline volumetrics.
Two cases were aborted due to unexpected vascular anatomy. In both cases, the right artery was found in close proximity to the left bile-duct system, and it was felt that dissection of the artery would potentially compromise the blood flow to that system. As discussed later, it can be very difficult to predict subtle variations in portal arterial anatomy, particularly as they relate to the biliary system.
Unexpected liver quality
In 3 instances (25% of the series), the donor hepatectomy was aborted due to an unexpected finding of poor liver quality—there was one case of unexpected steatohepatitis, one case of regenerative nodular hyperplasia and one case of multiple bile duct hamartomas. Intraoperative macroscopic findings were verified by intraoperative biopsies. The biopsy results were discussed with the liver pathologist, hepatologist on service and at least one other transplant surgeon before finally aborting the case. None of these findings were predicted on the preoperative radiological review.
The case of steatohepatitis deserves discussion, as it illustrates the fact that preoperative liver biopsies do not necessarily identify unsuitable donors. In this case, the female donor had a negative history for liver disease, but an ALT of 40 and a BMI of 33 prompted a preoperative biopsy. The biopsy showed a mild patchy steatosis of 5–10% and therefore no clear contraindication for donation. In the operating room the macroscopic appearance of the liver was more worrisome with blunt liver edges, a fatty liver parenchyma and patchy perfusion disturbances. An intraoperative biopsy showed a portal lymphoplasmacytic inflammation and large droplet steatosis of 5–10%, suggesting steatohepatitis. In retrospect, the patient admitted higher use of alcoholic beverages.
The other two cases aborted for reasons of unexpected liver parenchymal pathology demonstrate the limits of even a very extensive radiological and biochemical workup. One patient had no history of liver disease and pristine lab work. On opening the abdominal cavity, there were innumerable 2–3 mm soft gray spots throughout the liver. Intraoperative biopsy revealed diffuse biliary harmatomas involving at least two-third of the portal triads. Although the significance of this finding was unclear, the possibility of a ductal plate malformation or an association with polycystic liver disease was raised, and the case was aborted. In retrospect, a preoperative MR of the liver remarked multiple tiny T2 high-signal intensity foci, which at the time were called liver cysts. The other patient had no history and no lab work to suggest liver pathology. The intraoperative macroscopic liver examination showed a liver of large nodularity. Intraoperative biopsy revealed nodular regenerative hyperplasia with fibrosis. Preoperative imaging, including ultrasound, CT and MRI did not reveal any abnormalities.
Unexpected intraoperative event
In the two cases that were aborted due to intraoperative issues that placed the donor at unacceptably high risk, one was due to a sudden intra-parenchymal bleed, and another to unexplained CVP liability. In the first case, the liver parenchymal transection had begun and was halfway completed when the right lobe of the liver became grossly swollen and then cracked over the dome, with some bleeding that was readily controlled by pressure. The donor remained hemodynamically stable. After consultation with the recipient surgeon, the decision was made that completing the right hepatectomy in a formal fashion was the safest course of action for the patient, although the transplant had to be cancelled. The preoperative imaging had demonstrated a small, less than 1-cm vascular lesion in the right lobe of the liver in the general area of the parenchymal bleeding, attributed to an arterio-venous malformation. It is possible that this area was the origin of the intra-parenchymal hemorrhage that led to the rupture of the liver.
In the second case, while mobilizing the right lobe of the liver it was noted that the right lobe had become quite congested, and that the patient's CVP had become quite elevated (over 25 mm Hg). The patient was young and healthy, and there was no history of cardiac disease. He had previously experience stress-related atypical chest pain and had early repolarization changes on his preoperative electrocardiogram. For these reasons he underwent a dobutamine stress echocardiogram and was seen in preoperative consultation by cardiology. Neither the cardiology consult nor the stress echocardiogram were remarkable. At the time of surgery an intraoperative trans-esophageal echocardiogram was performed, and did not reveal any cardiac dysfunction except a small persistent foramen ovale. The latter was judged to be too small to account for the elevated CVP. After consultation with other members of the surgical group, the case was aborted for reasons of unexplained CVP lability.
The living donor team reviewed all 12 cases as this series was being put together. We performed this review with a view to identifying patterns that would predict donor abortion. Apart from the narrow angle between the right and left bile ducts, there was no preponderant pattern that could be used to predict donor abortion. The other donor abortions speak to the limitations of the radiological investigations used in the donor evaluation, discussed later, and to the fact that the unpredictable is always a possibility.
All patients had an uneventful postoperative course and left the hospital on average 4.8 ± 0.6 day (vs. 6.9 ± 1.8 day for all donors). Table 2 summarizes the details for the abortion of the donor procedure and actions taken to manage the complication. There have been no long-term complications with a median follow up of 23 months. In two patients, further investigation after the frustrated donor procedure revealed an undiscovered precondition. The patient with steatohepatitis was advised to abstain from alcohol and to follow dietary restrictions. Under these conservative measurements, the steatohepatitis clinically resolved. Our hepatology department follows the patient with nodular regenerative hyperplasia on an annual basis. Three years later, the patient is still unremarkable in terms of his liver function tests and Doppler ultrasound imaging.
Table 2. Aborted donor procedures
Reason for abortion
Intraoperative complication management
Days in hospital
25 year old brother
Aberant segment 4 duct originating from the right hepatic duct
Cholecystectomy, surgical consult
46 year old brother-in-law
Aberant segment 4 duct originating from the right hepatic duct
Cholecystectomy, surgical consult
33 year old friend
Right hepatic artery adherent to the left duct system
Cholecystectomy, continued with planned fundoplicatio
20 year old friend
Right hepatic artery adherent to the left duct system
Cholecystectomy, surgical consult
49 year old father
Intraoperative cholangiogram revealed three ducts in unfavorable location
Cholecystectomy, surgical consult
27 year old son
Narrow crotch left and right bilary system, risk of denuding posterior duct
Cholecystectomy, surgical consult
43 year old brother
Disection would have left segment 1 undraigned, borderline volumetrics
Unexplained intraoperative CVP rises up to 25 mm Hg with intermittent liver congestion
Cholecystectomy, Intraoperative TEE, cardiology consult
54 year old anonymous
Intraparenchymal hemorrhage right lobe
Right hepatectomy without donation
Consequences of live donor abortion for the potential recipient
We routinely stagger the donor and recipient operations, such that the recipient is only called to the operating room once the donor team has made the determination that the donor liver anatomy and quality is suitable. As a result we have almost always avoided unnecessary interventions in the recipient. In only one of the 12 donor no go operations described in this study was the recipient called to the room, and in that case (donor right lobe hemorrhage) the recipient case was cancelled prior to an incision being made.
Six of the potential liver recipients who did not receive a liver graft due to abortion of the original living donor procedure went on to receive a deceased donor graft. For four patients, an alternative living donor was found. One recipient died on the waiting list and one patient was taken off the list because their clinical status improved.
Although LDLT is performed worldwide, there is very little information on the incidence and sequela of aborted living donor hepatectomies. In this study, we report exclusively on those patients who go to the operating room but do not donate—a situation where surgeons have to act on the principle ‘primum non nocere’—to avoid harm to the living donor.
This report reflects our experience and the decision making of a single high volume North American center. The results of the A2ALL multicenter group place our study in context (3). Of the 13 donors aborted in the A2ALL study, one was aborted due to the donor declining the operation in the operating room. Of the remaining 12, five were aborted due to recipient issues (one death, one instability, three findings that precluded transplantation). Four of the remaining seven were aborted due to issues of liver quality (steatosis, granulomas or ‘unspecified inadequacies’) and three due to dense adhesions or small, aberrant bile ducts. Thus, 4/405 were aborted due to unexpected liver quality (just under 1% of the total case load), and up to 2/405 (roughly 0.5%) for unexpected aberrant ducts. In our series, 5 of 12 aborted donor hepatectomies (of a total of 257 donor operations) were for reasons of unexpected biliary anatomy (1.9%) and in three cases (1.2%) for unexpected liver quality. Although these rates are slightly higher than those of the A2ALL group, none of the differences are statistically significant (total rate of donor abortions, biliary anatomy, liver quality). Without directly comparing the data sets, it is impossible to be sure what underlies the differences in rates of abortion. While one can speculate on factors that could underlie the different findings, based either on differences in multicenter versus single-center approaches or on individual center expertise, it is probably more appropriate to note that the overall rate of donor abortions is quite close (3.2% vs. 4.7%). Given the overall small number of donor abortions reported (excluding the radiological literature, there are 13 + 12 = 25), a more detailed analysis awaits higher numbers.
The heart of any evaluation of the no go donor hepatectomy rate is the question of how rigorous the evaluation process was that led to a donor being deemed acceptable. For the purposes of anatomical evaluation any donor that has been deemed ‘acceptable’ has undergone a dedicated liver ultrasound, CT angiogram of the liver, MRCP and possibly a CT-cholangiogram if the donor biliary anatomy is not well defined on the MRCP. All four donor surgeons in our program review the CT and MRI images, and a donor only proceeds if a unanimous decision is made regarding anatomical suitability.
However, there are clear limitations to the radiological assessment of potential donors. We have used CT scans for depiction of vascular anatomy, in particular hepatic arterial anatomy. Our CT protocol is a single examination, standard protocol used at most centers; it combines a precontrast image through the liver for assessment of fat, a CT angiography phase and a venous phase to depict the hepatic venous and parenchymal anatomy (15,16). CTA is preferred to MRA technique due to its superior spatial resolution; in fact, in the radiology literature it is accepted as the gold-standard in studies evaluating advanced MRA techniques (10).
We have used MR primarily for the depiction of the biliary anatomy, but also for its superior depiction of fat within the liver as compared to CT. Routine MR cholangiography sequences which are still widely used today are not consistent in the depiction of smaller second-order biliary branches. The lack of consistency led an advanced MR imaging center to conclude that ‘for implementation of MR imaging as the sole preoperative imaging modality for living liver donor evaluation, improvements in MR cholangiography are clearly needed’ (17). In the past few years, the use of parallel imaging in the abdomen, hepatobiliary contrast agents used for excretory MR cholangiography and higher field strength magnets have all been tried to address this deficiency. Nevertheless, the accuracy of MR cholangiography, including excretory cholangiography, in the recent literature has been reported to be between 84.4% and 92.5% (7–9,11–13), when compared to conventional cholangiography. The expectation that MR cholangiography should be able to detect all significant biliary anomalies preoperatively is therefore not attainable, despite all of the latest advancements in MR imaging.
In our series of 257 patients, five (1.9%) had biliary anatomy that appeared as an unexpected significant variant leading to surgical abortion. Given the high incidence of biliary branching variations and the fact that most of these patients were evaluated prior to the latest technological advancements in MR magnets, the performance of our preoperative MR cholangiography performance in the prediction of biliary anatomy appears adequate. Two additional patients had the right hepatic artery in contiguity with the left bile duct, a condition that could not be diagnosed or suspected preoperatively, especially because vascular and biliary anatomy were imaged by two different modalities.
We have changed our imaging protocol recently to improve the preoperative biliary depiction. We are now exclusively using 3.0 Tesla magnets for our potential liver donor population. Furthermore, we have utilized CT cholangiography in cases where the MRCP is of insufficient quality (14). These changes, coupled with our improved appreciation of which cases to avoid (e.g. the ‘narrow angle’ between left and right bile ducts outlined in Figure 1), should decrease the rate of unexpected anatomical variations in the future.
Another central issue is that of preventability—whether there were anatomical factors present on the preoperative workup that, in retrospect, might have precluded a donor being deemed acceptable. In preparing this paper, we performed a review of the seven donors whose cases had been aborted for reasons of anatomy that had not been appreciated on their preoperative assessment. Apart from the narrow angle between the right and left duct that led to one donor procedure abortion, there was no clear pattern of anatomical variations that predicted subsequent donor abortions. Four of the five cases aborted for biliary anatomy had aberrant ducts or ductal relationships that were not appreciated on the preoperative MRCP and speak to the limitations of the radiological assessment. However, the decision to abort was generally not a simple one and involved a decision about the additional risk that would have been required to perform the surgery, considerations of the remnant liver lobe and the degree to which the unexpected finding would affect the quality and complexity of the recipient operation. With our current experience of over 300 performed living donor procedures, we would in retrospect have clearly turned down 1 of the 12 no go patients on the basis of preoperative imaging—the patient with the narrow angle between the right and left duct. We would have deemed the other six as ‘acceptable’, although we might now—with the benefit of experience in both the donor and recipient operation—have proceeded with the donor hepatectomy in at least two (one with three separate right ducts, one with large caudate duct).
Three of the 12 no go donor hepatectomies were the result of unexpected liver quality. At our center we do not routinely perform a preoperative liver biopsy. We reserve liver biopsies for the assessment of fatty liver disease in patients who have elevated liver enzymes and evidence of fat on imaging (5). All patients are thoroughly assessed for a history of or risk factors for liver disease, and undergo serological testing for hepatitis B and hepatitis C viral infection. In the present series, in two cases, the donor procedure was aborted because of a preoperatively undetected liver disease that might conceivably have been diagnosed on preoperative liver biopsy—one case of steatohepatitis, and one of liver fibrosis (nodular regenerative hyperplasia). However, as demonstrated by the first case—in which a preoperative liver biopsy failed to demonstrate the steatohepatitis that ultimately led to abortion of the procedure—preoperative liver biopsies apart from the risk of the procedure have a significant false negative rate and as a result we do not advocate routine liver biopsies in patients who do not have risk factors for liver disease (18).
With continued improvements in the preoperative imaging available for donor assessment the rate of no go donor hepatectomies should decrease. However, it will never be zero—the unpredictable is always possible. Nor should it be zero: the donor surgeon should always be prepared to abort the procedure whenever donor safety is in question, and the threshold for aborting the procedure should be relatively low. There have been no untoward complications in the 12 patients who aborted the donor procedure, suggesting that the potential donors are not at any greater-than-expected risk for complications. At our single, large center and in the multicenter A2ALL study, the rate of no go donor hepatectomy is less than 5%. All efforts should be made to decrease it still further while still respecting the paramount importance of maximizing donor safety.