Notice: Wiley Online Library will be unavailable on Saturday 27th February from 09:00-14:00 GMT / 04:00-09:00 EST / 17:00-22:00 SGT for essential maintenance. Apologies for the inconvenience.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Surgical resection is the preferred treatment for primary hepatic malignancies, colorectal liver metastases, and various other metastatic tumors. While orthotopic liver transplantation (OLT) represents the best therapeutic option for early-stage hepatocellular carcinoma (HCC) and for patients with advanced liver disease,1 use of OLT is limited by the scarcity of donor organs, and liver resection continues as the initial treatment for most patients with resectable disease.
Once associated with high perioperative morbidity and mortality,2 major hepatic resection is now accomplished with mortality rates of less than 5% in major centers.3–5 Despite these improvements and advances in preoperative selection of candidates for hepatectomy, postresection liver failure remains a potentially devastating complication and often proves fatal.3–6 Factors that have been shown to play critical roles in the development of posthepatectomy liver failure include underlying chronic liver disease, portal hypertension, extent of resection, volume of residual functional hepatic reserve, intraoperative liver ischemia, and comorbid diseases.3–5, 7, 8
With increasing incidence of chronic liver diseases and HCC, patients requiring liver resections often have some degree of hepatic dysfunction that can leave insufficient functioning parenchyma following hepatic resection.9–13 Unfortunately, no reliable disease-specific therapy exists for postresection hepatic failure, and mortality for the condition remains at 60 to 80%.7 The role of OLT in posthepatectomy liver failure has not been well defined. The present study was undertaken to review our experience with postresection hepatic failure and the use of OLT as treatment.
We performed a retrospective review of all major hepatic resections performed at UCLA Medical Center from 1990-2004. Data were collected in accordance with requirements of our Institutional Review Board and included patient demographics, indication for hepatic resection, comorbidities, underlying liver diseases, Child's classification and Child-Turcotte-Pugh (CTP) scores, intraoperative factors, perioperative morbidity and mortality, perioperative liver tests, and pathology. Transplant data collected included ischemia times, graft survival, patient survival, type of graft, type of arterial and biliary reconstruction, use of venovenous bypass, and posttransplantation complications.
All patients with posthepatectomy liver failure were considered for OLT. Patients who were judged to be appropriate candidates were listed and categorized based on the United Network for Organ Sharing criteria. One patient transplanted in 2003 was listed and classified according to the Model for End-Stage Liver Disease system.
OLT was performed using standard techniques.14 For immunosuppression, a triple drug regimen of cyclosporine (Sandimmune; Novartis, Basel, Switzerland), azathioprine (Imuran; GlaxoSmithKline, Triangle Park, NC), and prednisone was utilized from 1988 through 1994. Routine use of tacrolimus (Prograf; Astellas Pharma, Tokyo, Japan) was begun in 1995 as part of a dual or triple drug regimen with prednisone and mycophenolate mofetil (CellCept; Hoffman-LaRoche, Nutley, NJ), the latter starting in 1997.
Univariate statistical comparisons were performed for patients who developed posthepatectomy liver failure and those who did not. Chi-squared and Student's t-test analyses were utilized where appropriate using Excel statistical software (Microsoft Corporation, Redmond, WA); significance was assigned at the 0.05 level.
Between 1990 and 2004, 435 patients underwent major hepatic resections; average age was 53 yr (range 1-85 yr), and 247 (57%) were female. Indications for resection included metastatic colorectal carcinoma (n = 93), HCC (n = 91), hemangioma (n = 58), cholangiocarcinoma (n = 35), benign cyst disease (n = 32), and focal nodular hyperplasia (n = 20). Resections included right lobectomy (n = 101), extended right lobectomy (n = 44), left lobectomy (n = 49), left lateral segmentectomy (n = 54), segmental resections (n = 110), and wedge resections (n = 35). Complications occurred in 145 patients (33%), and 30-day in-hospital mortality was 1.7%.
Posthepatectomy liver failure developed in 9 of the 435 patients (2.1%); 2 additional patients were transferred to UCLA with liver failure after resections that were performed elsewhere. Of these 11 patients, 7 were male, and average age was 55 years (range 1-71 yr). Of the patients, 6 had chronic liver disease; mean preresection CTP score was 5.5. Indications for resection included primary hepatic malignancies (n = 8), colorectal metastases (n = 2), and echinococcic cyst disease (n = 1). Nine patients were Child's Class A; the remaining 2 were Class B. Five patients (45.5%) had preexisting cirrhosis, 1 (9%) had bridging fibrosis, and 1 (9%) had steatosis. Six of the resections were extended right lobectomies, 2 were formal lobectomies (1 right, 1 left), and the remaining 3 were multisegmental resections. Postresection failure was manifested by encephalopathy in 8 patients, coagulopathy in 6, and ascites in 5. All patients with postresection hepatic failure received hemodialysis for correction of acidosis and plasma transfusions for correction of coagulopathy. One patient underwent completion hepatectomy with portosystemic shunting for postresection hepatic necrosis before receiving OLT 24 hours later.
Comparison of Patients With and Without Liver Failure (Tables 1–3).
Preoperative assessments demonstrated significantly lower platelet counts in patients who progressed to postresection failure (Table 1). Preoperative Child's classification, CTP scores, and mean age were not significantly different for patients who had postoperative liver failure and those who did not.
Table 1. Preoperative Factors and Postresection Liver Failure
Liver failure (n = 11)
No liver failure (n = 426)
54.7 ± 19
52.9 ± 17
5.5 ± 0.76
5.0 ± 0.23
149.25 ± 104.5
1.15 ± 0.25
1.06 ± 0.18
3.66 ± 0.84
4.15 ± 0.60
0.9 ± 0.17
0.95 ± 0.43
Total bilirubin (mg/dL)
2.08 ± 2.6
1.12 ± 2.8
Aspartate aminotransferase (U/L)
47.89 ± 21.75
51.80 ± 113
Alanine aminotransferase (U/L)
42.86 ± 22.9
52.86 ± 77.10
Alkaline phosphatase (U/L)
177.88 ± 98.6
144.49 ± 119.3
Table 2. Operative Factors and Postresection Liver Failure
Liver failure (n = 11)
No liver failure (n = 426)
Blood loss (mL)
1095 ± 1249
504 ± 741
Extent of resection
Extended right lobectomy
2 segments resected
3 segments resected
4 segments resected
Table 3. Postresection Laboratory Values in Patients With and Without Liver Failure
Regarding operative findings and procedures, biopsy-confirmed cirrhosis and extended right lobectomy were associated with increased risk for postresection failure (Table 2). There were no significant differences in blood loss or use of the Pringle maneuver in the 2 groups. A trend toward increased risk for liver failure was apparent with resection of increased numbers of hepatic segments.
By the second postoperative day, patients who developed irreversible liver failure had significantly elevated total and direct bilirubin levels, significantly lower platelet counts, and prolonged prothrombin time/international normalized ratio (INR) (Table 3). Patients with liver failure also had significant differences in lowest postoperative platelet counts and peak values for total and direct bilirubin, prothrombin time/INR, and creatinine (Table 3).
During hepatic resection, the Pringle maneuver was employed in only 54.6% of the 435 patients. Mean total Pringle time was 18.6 minutes (range 2-60 minutes); 24.8% involved intermittent inflow occlusion. Two patients developing posthepatectomy liver failure had Pringle inflow occlusion for 31 and 15 minutes, respectively.
Of the 11 patients with posthepatectomy liver failure, 7 were listed and received OLT at an average of 25 days (range 2-56 days) after resection. Four patients excluded from consideration included a 60-yr-old male with incompletely resected, advanced stage cholangiocarcinoma (Stage IVa); a 71-yr-old male with significant coronary artery disease and metastatic colorectal cancer with poor prognostic features; a 57-yr-old female with advanced HCC as well as uncontrolled peritonitis and sepsis; and a 69-yr-old female with organ failure requiring pressor support. All of the excluded patients died of multisystem organ failure at a mean of 42 days (range 8-95 days) after hepatic resection.
Whole cadaveric livers were used in all but 1 pediatric patient, who received a left lateral segment split-graft. Veno-venous bypass was used in all but the pediatric recipient, and the mean total time for cold and warm graft ischemia was 6.94 hours (range 2.9-10.1 hours).
Major arterial reconstruction because of complex donor anatomy was needed in 1 patient. Five recipients had biliary reconstruction via primary choledochocholedochostomy, while 2 received Roux-en-Y choledochojejunostomy. One patient had completion hepatectomy combined with portocaval shunt for severe ischemic hepatic necrosis prior to transplantation. Three recipients had episodes of graft rejection requiring treatment, and 1 had primary nonfunction requiring retransplantation twice. One patient developed early biliary obstruction and required conversion to Roux-en-Y bilioenteric anastomosis. Two other patients developed late biliary strictures; one required interventional dilatations, while the other underwent conversion to Roux-en-Y choledochojejunostomy.
Following transplantation, graft survival was 77% at 1 yr, 56% at 3 yr, and 34% at 5 yr. Corresponding patient survival was 88%, 60%, and 40%, respectively.
Patient profiles and outcomes after OLT are shown in Table 4. One patient was transplanted for hepatic failure following extended right lobectomy for colorectal metastases; this patient survived 39 months before succumbing to advanced metastatic colorectal disease. Four patients were transplanted for hepatic failure after resection of HCC; 3 remain alive and free of tumor recurrence at 125 months (Stage II, single 11-cm tumor in right lobe), 39 months (Stage IIIA, 2 tumors in right lobe, largest 4.5 cm), and 26 months (Stage II, single 4.8-cm tumor in left lateral segment) after transplantation. The lone death occurred from overwhelming sepsis and multiple intrahepatic abscesses 17 months after transplantation (Stage IIIA, 2 tumors in Segment 3, largest 4 cm). A patient transplanted for hepatic failure after resection of a large echinococcic cyst remains alive with excellent graft function and no recurrent disease 118 months after OLT. The final patient was transplanted for hepatic failure following extended right lobectomy for hepatoblastoma; this patient remains alive and disease free 74 months after transplantation.
The incidence of posthepatectomy liver failure remains unknown, largely because a universally accepted, standardized definition based on clinical and laboratory criteria does not exist.7 Nevertheless, in most series it is reported to occur after 5 to 20% of partial hepatectomies; some investigators even stress that posthepatectomy liver failure is currently increasing in frequency.7 This has not been the case at UCLA, where Longmire's original series from 1955-1980 demonstrated liver failure after 8% of resections and a perioperative mortality of 11%,6 compared with current values of 2.1 and 1.7%, respectively. These improvements have occurred despite increases in the number of patients with underlying liver disease; from 1955-1980, only 4.3% of patients had cirrhosis and less had steatosis, whereas from 1990-2004, 9% of patients had cirrhosis, 12% fibrosis, and 14% steatosis.
Many factors have been shown to be associated with increased risk of posthepatectomy liver failure, including extensive resection, portal hypertension, hepatic steatosis, diabetes mellitus, preexisting cirrhosis, prior hepatic disease, and prolonged liver ischemia during resection.7, 15–17 Other variables associated with increased risk for PLF include more than 1 comorbid condition, perioperative blood loss over 2L, occurrence of concomitant complications, need for reoperation, advanced age, and preresection chemotherapy.18
Proposed pathogenic mechanisms for postresection failure have included insufficient residual hepatic mass (small-for-size), inadequate hepatic regeneration, high portal flow, and venous congestion.17–20 In the current series, patients with large extended resections most likely experienced the “small for size” phenomenon. However, patients with PLF from segmental resections likely experienced acute decompensation of preexisting hepatic dysfunction.
Methods for predicting necessary hepatic reserve have been studied, and strategies have included clinical scoring systems CTP, metabolic testing (indocyanine green), and volumetric computed tomography prediction of postresection hepatic volume.18 Most studies indicate that patients with normal hepatic function require greater than 30% of preoperative hepatic volume to maintain adequate hepatic function after resection.18–20 In the present series, all patients were evaluated with CTP scoring, but none underwent metabolic testing or volumetric computed tomography analysis to predict postresection functional hepatic reserve. We routinely perform volumetric computed tomography to evaluate hepatic reserve in all donor candidates for living-donor liver transplantation.
In the present series, low preoperative platelet counts predicted postoperative liver failure, but liver function tests and the CTP score did not. Steatosis, a known risk factor for primary nonfunction of liver allografts, is also associated with post-hepatectomy morbidity, liver failure, and mortality.15, 16 Cirrhosis was more frequently identified in our patients with liver failure; 5 of 11 had cirrhosis, while only 1 of 11 had steatosis. Several studies have shown that patients with smaller hepatic remnants are at increased risk for hepatic failure, other complications, and perioperative mortality.17–20 Our findings were similar, as patients undergoing extended resections, especially extended right lobectomy, were more likely to experience posthepatectomy liver failure. Thus, during preoperative evaluation, particular attention should be paid to preoperative platelet results, and extended resections should be avoided or approached with caution in those with low platelets and chronic liver disease.
Early recognition of postresection liver failure is crucial to identify potentially treatable causes such as portal vein thrombosis and initiate appropriate therapy. While postoperative increases in serum bilirubin and prothrombin time have been advocated by some authors as predictive for liver failure,6 these authors detected significant differences by postoperative day 5, a time at which postresection hepatic failure may be advanced and difficult to treat. Comparing patients who developed posthepatectomy liver failure to those who did not, the present study demonstrates significant differences in total and direct bilirubin, prothrombin time/INR, and platelets as early as 2 days after resection. Patients with these laboratory profiles should be treated aggressively for liver failure and considered for early liver transplant evaluation.
In our series the average time between resection and OLT was 25 days. While we do not propose that the decision for transplantation can be made within 48 hours after resection, our data suggest that patients with significant abnormalities in platelet count (less than 100K), INR (above 2.0), and/or total bilirubin (>6.6 mg/dL) on postoperative day 2 are at significantly increased risk for developing irreversible PLF. Moreover, patients not developing PLF had total bilirubin less than 2 mg/dL in 98% of cases, INR below 2 in 97% of cases, and prothrombin time below 19 seconds in all cases. As our data support the concept that patients with INR >2, total bilirubin ≥6.6 or platelets <100K are more likely to experience relentless progression of postresection liver failure, we advocate urgent OLT evaluation for these patients. Outcomes for OLT clearly demonstrate inferior results for sicker patients with more severe acute hepatic decompensation,14 and delays in initiation of evaluation might be detrimental to the patient's chances for successful liver replacement.
The present series includes four patients with HCC, 3 of whom had preresection disease burden within the Milan criteria.21 Three of the 4 patients developed hepatic failure after limited segmental resection (maximum 2 segments), suggesting the presence of significant underlying hepatic dysfunction. OLT for these patients has been successful in this small series, as 3 of the 4 patients survived disease-free for at least 2 yr after OLT. The lone death occurred secondary to fatal sepsis in a patient without HCC recurrence. While success has been demonstrated with OLT for patients with disease exceeding the Milan criteria,22–24 and while the present series includes a 72-yr-old patient with an 11-cm hepatoma who has survived recurrence-free 125 months after rescue transplantation for PLF, in most instances candidacy for OLT after PLF should be limited to patients with preresection tumor burden within the Milan criteria. Patients with extensive intrahepatic disease, advanced cardiopulmonary comorbidity, poor social support, poor functional status, or limited life expectancy due to other medical diseases should not receive OLT.
Liver transplantation as primary treatment for metastatic colorectal disease has been discouraged since reports in 1994 demonstrated high tumor recurrence rates.25 One young patient in our series developed postresection hepatic failure after curative resection of metastatic colon cancer by an extended right lobectomy; she underwent OLT and survived for 39 months before she died of recurrent and advanced metastatic disease. OLT for this patient was justified by the young age of the patient, the curative nature of the resection, the absence of extrahepatic disease, and the favorable prognostic nature of the original colon cancer. Patients of advanced age, with unfavorable primary colorectal tumors (e.g., advanced T and N stage, synchronous presentation of primary tumor and metastases, poor response to chemotherapy), or with advanced comorbidities should not receive OLT for PLF. An algorithm describing our approach to the evaluation and management of patients with PLF is shown in Figure 1.
Patients with HCC and background cirrhosis are at greatest risk for posthepatectomy liver failure.17–20 In the present series, 6 of 91 patients (6.7%) undergoing resection for HCC experienced liver failure, and 4 of 7 who underwent OLT had HCC. Liver transplantation is the preferred treatment for limited HCC (solitary tumor <5 cm, 2 or 3 tumors <3 cm in diameter).21 However, due to the increasing number of patients with HCC and the relative shortage of liver allografts, patients with HCC, despite the increase in Model for End-Stage Liver Disease points, may spend considerable time on the organ waiting list and may potentially suffer from disease progression prior to transplantation. Hepatic resection has been shown to be an effective therapy for select patients with HCC;26, 27 however, controversy exists with regards to long-term outcomes.28–30 Some maintain that resection of HCC prior to OLT results in higher peritransplantation mortality, increased risk of HCC recurrence, and poorer patient and graft survival.28 In the present series, all patients with HCC undergoing OLT for postresection hepatic failure had significant perioperative complications; 3 had biliary strictures, while 1 had PNF in 2 grafts requiring 2 retransplantations. Only 1 had acute rejection requiring treatment. Following transplantation, all 4 patients with HCC lived 1 yr, and 1 lived for 5 yr. This survival trend reflects the general outcomes for OLT for malignancy, which clearly has the poorest long-term survival for all indications for OLT.14
Overall patient and graft survival for the 7 OLT recipients underscored the heterogeneous nature of those with posthepatectomy liver failure: an admixture of patients with benign and malignant disease, often with underlying chronic liver disease, and experiencing subacute to rapid hepatic decompensation. Results for 1-, 3-, and 5-yr patient (87.5%, 60%, and 40%) and graft (77%, 55.6%, and 33.3%) survival are lower than those in recent reports, and most similar to outcomes for OLT for malignancy.14 Posttransplantation morbidity was considerable, similar to findings in other series evaluating OLT after resection.28 However, OLT in the setting of posthepatectomy liver failure produces survival rates unparalleled by any other available treatment. Therefore, patients with postresection hepatic failure should be considered early as candidates for liver transplantation for definitive treatment of a frequently fatal condition. Until strict criteria can eliminate the problem of postresection hepatic failure, we propose that major hepatic resection should be performed in centers where liver transplantation is an option in the rare instances where it will be needed.