Liver transplantation has been available in British Columbia since 1989, with a growing number of transplants performed each year. In 2005, 31 liver transplants were performed in BC. Only 10 years later, in 2015, 82 transplants were performed in a single year. All liver transplants in BC are performed at Vancouver General Hospital using a collaborative process for preoperative evaluation, transplantation, and postoperative management. A multimodal specialty group that includes transplant surgeons, hepatologists, psychologists, social workers, and dietitians evaluates and cares for these complex patients.
Clinical and demographic information on all transplant recipients is collected by BC Transplant, which publishes data regarding wait lists and patient outcomes. As well as providing an extremely valuable overview of the status of liver transplantation in BC, the BC Transplant database permits a detailed examination of the specific surgical complications following transplantation.
Liver transplantation requires the division and reconstruction of multiple structures, including the hepatic artery, hepatic vein or inferior vena cava, portal vein, and bile duct. Sometimes bowel reconstruction is required as well. Despite advances in surgical technique, immunosuppressive therapy, and critical care management of transplant recipients, the recovery phase of liver transplant can be very challenging because of unanticipated severe complications such as bile leak, sepsis, or thrombosis. Furthermore, many patients present for transplantation in advanced stages of liver disease and frequently with previous abdominal operations and significant medical comorbidities.
Post-transplant complications can result in severe morbidity, require additional procedures, and lead to prolonged hospital stay, graft loss, or death. Thus, evaluating surgical complications and their management and comparing these rates and outcomes with those of other liver transplant centres is a crucial part of assessing program performance and identifying areas for improvement.
Ethics approval was obtained from BC Transplant as well as from the UBC Clinical Research Ethics Board and the Vancouver Coastal Health Research Institute. All deceased-donor liver transplants performed in BC from 2007 to 2011 were identified in the prospectively and independently managed BC Transplant database. Information available included patient age and gender, severity of liver disease as measured by the model for end-stage liver disease (MELD) and Child-Pugh (CP) scores, and the cause of liver failure.
A retrospective review of data for 201 individual transplants performed in 191 patients was completed using both paper and electronic records, including operative reports, radiology studies, outpatient clinic notes, and all available follow-up information obtained at Vancouver General Hospital until 31 August 2014. The mean duration of patient follow-up was 168 weeks (median 161 weeks) with a range of 2 days (early postoperative mortality) to 7 years.
Patient records were reviewed specifically to evaluate rates of reoperation, biliary complications, vascular thrombotic complications, lengths of ICU and hospital stays, and mortality.
Patients undergoing deceased-donor liver transplantation in BC from 2007 to 2011 ranged in age from 15 to 71 years (mean 51) and included more males (61.0%) than females (39.0%). In the preponderance of cases liver failure was caused by hepatitis B or C or alcoholic cirrhosis (Figure). Most patients waiting for a liver transplant were able to remain at home without hospital care (Table 1).
The mean MELD score for patients studied was 19.2 (range 2.0–41.0). The mean CP score was 9.7 (range 6.0–15.0). Based on the mean MELD score, the estimated 3-month mortality rate for patients without transplantation was 6.0%. The mean length of hospital stay was 30.0 days, and the mean length of ICU stay was 6.3 days. The in-hospital mortality rate was 5.2% and the overall 1-year mortality rate was 8.4%. Surgical complications were defined as reoperative, biliary, or vascular, and occurred in 107 of 201 cases (53.0%) during the study period (Table 2).
A total of 65 reoperations (32.3% of all cases) occurred during the study period. These included both planned and unplanned reoperations. In 10 cases (5.0%), re-transplantation was required because of acute graft failure (Table 3)
Planned reoperation involved intentionally leaving the abdomen open after the initial transplant when required as part of a damage-control approach. In 24 cases (11.9%), this was necessary because of cardiovascular instability, coagulopathy, high risk of abdominal compartment syndrome, or inability to complete bile duct reconstruction.
Unplanned reoperation occurred when patients were deemed stable enough for abdominal closure after the primary transplant procedure, but subsequently presented with multiple indications for repeat surgery. In 31 cases (15.4%) an unplanned reoperation was required. The most common indication for reoperation was early postoperative hemorrhage (15 cases). Other indications for reoperation were management of a bile leak (3 cases), hepatic artery pseudoaneurysm (2 cases), common bile duct stricture (1 case), hepatic artery thrombosis (1 case), and portal vein thrombosis (1 case).
A bile leak occurred in 19 cases (9.5%) and three patients who developed a bile leak subsequently developed a stricture (Table 4). Biliary strictures were identified by endoscopic retro-grade cholangiopancreatography (ERCP) or magnetic resonance cholangiopancreatography (MRCP). Anastamotic stricture occurred in 34 cases (16.9%). The majority of post-transplant biliary strictures occurred in a delayed fashion, with only 9 of the 34 cases (26.5%) occurring within 30 days. The median number of days to detection and intervention was 87, highlighting the insidious nature of stricture formation. ERCP was used as the initial intervention for anastamotic stricture 100.0% of the time, with only two patients (5.9%) requiring reoperation after failure of endoscopic therapy.
Vascular complications were defined as hepatic artery thrombosis (HAT), complete or partial portal vein thrombosis (PVT), and hepatic vein thrombosis (HVT). HAT occurred in 15 transplants (7.5%), with 12 cases defined as early HAT (detected at or before 30 days post-transplant) and 3 cases defined as late HAT (detected after 30 days). Of the early cases, 75.0% were managed operatively, with six patients undergoing re-transplantation, and three undergoing vascular reconstruction or revision. Of the three late HAT cases, one patient was managed with anticoagulation, one underwent open repair, and one underwent an attempt at revascularization by interventional radiology that was ultimately unsuccessful and was managed subsequently with observation.
PVT occurred in nine transplants (4.5%), three of which required urgent re-transplantation. In two of these cases, the patient had a concurrent hepatic artery thrombus. Anticoagulation and observation were used to manage the remainder of patients.
Surgical complications following liver transplantation can involve a complex set of morbidities that have a significant impact on patient care and quality of life. Severe complications can result in graft failure or mortality, and any complication can affect transplant outcomes. In our study, complications occurred in approximately half of all cases.
Despite limitations inherent when comparing health care systems and patient populations, a comparison of transplant outcomes in BC with those in the United States and Europe spanning similar time periods revealed similar 1-year mortality rates (Table 5). A comparison of outcomes related to biliary and vascular complications also revealed similar rates, with the exception of a higher rate of early HAT in BC.
In our study, bile leaks and anastamotic strictures were the most common biliary complications. Biliary complications are usually detected when clinical suspicion is raised by:
• Presence of bile in surgical drains.
• Persistently elevated or newly elevated liver enzymes or bilirubin.
• Symptoms of biliary sepsis.
• Results from an MRCP or a hepatobiliary iminodiacetic acid scan.
The reported incidence of post-transplant bile leaks ranges from 2.0% to 25.0%. Most leaks occur at the anastamosis, cystic duct remnant, or T-tube site because of either technical issues or anastamotic ischemia and breakdown. Early leaks (less than 4 weeks post-op) usually occur because of ischemia or technical issues, while late leaks usually occur during removal of T-tubes.
Anastamotic strictures form secondary to fibrosis as the duct heals. Strictures tend to have a more insidious course and can be detected any time after transplant. Management options include a combination of ERCP, percutaneous drainage, surgical revision and/or observation. ERCP can be both diagnostic and therapeutic, and successful stenting can obviate the need for surgical revision. However, this can be undertaken only in patients who are clinically stable, and patients with severe sepsis or septic shock from their leak may require primary operative revision.
When BC Transplant outcomes for biliary complications were compared with those of a large systematic review of 61 studies by Akamatsu and colleagues, BC rates were found to be similar to those for more than 11 000 patients treated between 1990 and 2009. The BC Transplant bile leak rate of 9.5% was comparable to Akamatsu’s finding of 7.8%, and the BC anastamotic stricture rate of 15.9% was comparable to Akamatsu’s rate of 12.8%. Furthermore, BC was found to utilize a very high rate of primary ERCP. For example, while Akamatsu and colleagues found that 38.0% of bile leak patients were managed with ERCP, 73.7% of BC Transplant patients were managed with ERCP.
The diagnosis of a vascular thrombotic complication is based on clinical suspicion and is part of the differential diagnosis for rising liver enzymes post-transplant. Patients are screened using Doppler ultrasound. Decreased or absent flow is confirmed using CT angiography or conventional angiography to identify the point of obstruction. Hepatic artery thrombosis is a particularly devastating complication of liver transplantation. Since the process of explanting the donor liver involves removing all of its collateral feeding vessels, the arterial supply to the graft is entirely dependent on the hepatic artery. Thus, HAT is a major cause of graft loss in the early post-transplant period. Occlusion of this vessel with thrombus will result in graft ischemia, ultimately leading to necrosis, sepsis, and death. Management of HAT depends mainly on the timing of occurrence, typically divided into early (30 days or less post-transplant) and late (more than 30 days post-transplant).
Early HAT typically requires reconstruction or re-transplantation, depending on the viability of the graft and the stability of the recipient. The onset of late HAT tends be more insidious, as affected patients may have had vascular reconstructions during their initial transplant that allow collateral circulation to support the liver parenchyma without necrosis. A comparison of BC Transplant outcomes for HAT with those found in other studies5-8 indicates a particularly high rate of early HAT in BC (Table 6). Surgical risk factors for early HAT are multiple, and include use of conduits for reconstruction, prolonged cold ischemia time, cytomegalovirus mismatch, re-transplantation, variant arterial anatomy, and prolonged operation time.[5-8]
Evaluation of patients who sustain HAT in BC can elucidate the specific contributory risk factors and identify areas for improvement. Post-transplant portal vein thrombosis occurs in 1.0% to 13.0% of patients.[9-11] PVT preferentially localizes to the anastamotic site, and predisposing risk factors include small portal vein diameter, presence of PVT prior to transplant, splenectomy, and prior surgical shunting. In our study population, two of seven patients who developed a post-transplant PVT also had PVT on preoperative imaging. Occlusion of the portal vein can result in the rapid onset of liver failure, ascites, portal hypertension, and variceal bleeding. The definitive treatment of PVT can include anticoagulation and interventional radiology approaches for thrombolysis. More rarely, surgical thrombectomy is required.
Similarly, hepatic vein thrombosis is a rare complication following liver transplantation. It has been found to have an incidence rate of 1.0% to 7.0%.13 The frequency of hepatic vein thrombosis in our review was 1.0%.
Limitations of study
A limitation of this study was our focus on surgical complications of transplantation, which means we did not take into account medical complications such as sepsis, renal failure, or cardiovascular complications that can also contribute to overall morbidity and mortality. In addition, patient data may have been incomplete in cases where investigations were not performed at Vancouver General Hospital. Also, as the BC Transplant database is meant to track global outcomes, detailed data regarding surgical complications had to be reviewed in a retrospective manner from individual patient charts, which may have captured only short-term and in-hospital complications.
Further investigation needed
Rates for 1-year mortality, biliary complications, and hepatic and portal vein thrombosis after deceased-donor liver transplantation in BC are similar to rates reported in the literature. However, rates for early HAT associated with graft loss requiring re-transplantation are higher than reported in the literature. Early HAT presents a significant morbidity risk to the patient and the specific factors that may be contributing to its occurrence in BC should be investigated.
Dr Segedi has received honoraria and reimbursement for attending meetings from Ipsen, a biopharmaceuticals firm, and Baxter, a medical products and services firm. The other authors of this article have not received fees from these firms and have no additional competing interests to declare.