- Email: [email protected]
Postoperative liver failure after major hepatic resection for hepatocellular carcinoma in the modern era with special reference to remnant liver volume
Postoperative Liver Failure after Major Hepatic Resection for Hepatocellular Carcinoma in the Modern Era with Special Reference to Remnant Liver Volume Ken Shirabe, MD, Mitsuo Shimada, MD, Tomonobu Gion, MD, Hirofumi Hasegawa, MD, Kenji Takenaka, MD, Tohru Utsunomiya, MD, and Keizo Sugimachi, MD, FACS
Conclusions: The expected remnant liver volume appears to be a good predictor for liver failure in patients who undergo a right lobectomy of the liver. In patients with diabetes mellitus and an expected remnant liver volume of less than 250 mL/m2, a major hepatectomy should be avoided. Careful patient selection based on volumetric analysis in major hepatectomy cases could help prevent the occurrence of postoperative liver failure. (J Am Coll Surg 1999;188:304–307. © 1999 by the American College of Surgeons)
Background: Postoperative liver failure is a lifethreatening complication after hepatic resection. Because of recent advances in liver surgery technique and a more stringent patient selection, mortality after hepatic resection has steadily decreased, but its incidence still ranges from 10% to 20%. The factors linked to postoperative liver failure in major hepatic resection in the modern era should be reevaluated. Study Design: Of 80 patients with viral markers (hepatitis C viral antibody or hepatitis B surface antigen) who underwent major hepatic resections (no less than bisegmentectomies) for hepatocellular carcinoma between 1990 and 1996, 7 patients (8.8%) died of postoperative liver failure within 6 months after hepatectomy. The cause of liver failure was analyzed based on both the preoperative data and the intraoperative findings. In addition, since all the patients who died of liver failure underwent a right hepatic lobectomy, a further data analysis was also done in 47 patients who underwent a right lobectomy of the liver. A volumetric analysis by CT was then done to evaluate the remnant liver volume.
Hepatic resection is the accepted treatment for various types of liver tumors and is increasingly performed with ever expanding indications and surgical limits. In particular, hepatic resection for hepatocellular carcinoma (HCC) is now being performed on patients who have chronic liver disease.1-3 In patients with chronic liver disease, however, postoperative liver failure is still an important cause of death after hepatic resection. Because of recent advances in liver surgery, such as the introduction of intraoperative ultrasonography and ultrasonic dissectors, and more stringent patient selection, the mortality rate after hepatic resection has decreased,3-5 but the incidence of mortality still ranges from 10% to 20%.4-6 An extremely high incidence of liver failure has been reported for patients who undergo a major hepatectomy for HCC.5 Many factors linked to postoperative liver failure have been reported, including preoperative liver function, remnant liver volume, and amount of blood loss during surgery.6-12 The causes of liver failure are complex because of broad differences in liver function, tumor size and location, operative procedures, and different amounts of blood loss for each patient; as a result, an accurate prediction of liver failure remains difficult. Optimal operative procedures have been established and the amount of blood
Results: Between the patients with liver failure and those without liver failure who underwent a right lobectomy, there were no significant differences in preoperative data or intraoperative findings. Volumetric analysis revealed that the remnant liver volume of patients who died of liver failure was significantly smaller than that of patients who lived (p 5 0.008). The incidence of liver failure in patients with a remnant liver volume of less than 250 mL/m2 was 7 of 20 (38%), while it was 0 of 27 in patients with a liver volume of no less than 250 mL/m2 (p 5 0.0012). The only significant risk factor for liver failure in patients with a remnant liver volume of less than 250 mL/m2 was diabetes mellitus (p 5 0.0072). Received April 13, 1998; Revised November 4, 1998; Accepted November 4, 1998. From the Second Department of Surgery, Kyushu University, Fukuoka, Japan. Correspondence address: Ken Shirabe, MD, Second Department of Surgery, Faculty of Medicine, Kyushu University, 3-1-1 Maidashi, Higashi-ku, Fukuoka, 812, Japan. © 1999 by the American College of Surgeons Published by Elsevier Science Inc.
304
ISSN 1072-7515/99/$19.00 PII S1072-7515(98)00301-9
Vol. 188, No. 3, March 1999
loss has also decreased significantly during hepatic resection,3 but a reevaluation of the factors linked to postoperative liver failure is necessary. We analyzed the cause of liver failure occurring in patients who underwent a major hepatectomy in the modern era. METHODS Between January 1990 and February 1996, 284 patients with HCC underwent a hepatic resection at the Second Department of Surgery, Kyushu University. Postoperatively, 32 patients without viral markers such as hepatitis B surface antigen (HBs) and hepatitis C viral (HCV) antibody, or both, did not die of liver failure after hepatectomy. Of the 252 patients with viral markers such as HBs and HCV antibody, or both, 80 patients who underwent a major hepatectomy (no less than bisegmentectomies) were entered into this study. The study group consisted of 70 men and 10 women. The operative procedures performed on these patients were as follows: 10 patients underwent a trisegmentectomy (right trisegmentectomy in 9, left trisegmentectomy in 1), and 70 patients underwent a bisegmentectomy (right lobectomy in 47, left lobectomy in 16, central bisegmentectomy in 7). Postoperative liver failure was defined as appearance of hepatic encephalopathy, progressively increasing hyperbilirubinemia without hemolytic or obstructive mechanism, decreasing activity of blood coagulation assessed by a prothrombin time, and lecitin–cholesterol acyltransferase (LCAT) deficiency, according to the criteria by Takenaka and associates.13 Postoperative liver failure was determined as death from liver failure within 6 months after hepatic resection. Seven of 9 patients (77.8%) who had postoperative liver failure died within 6 months after surgery; these 7 patients were regarded as postoperative liver failure cases in this study. None of the patients died of variceal bleeding or technical failure such as ipsilateral portal vein obstruction. None showed evidence of recurrent cancer at the time of death. Once the patients were diagnosed as having postoperative liver failure, plasma exchange was carried out. Plasma exchange therapy was given by the membrane filtration method using Plasmaflo AP05H (Asahi Medical, Tokyo) using 2000–2500 mL of fresh frozen plasma at 50–80 mL/min blood flow rate and 15–25 mL/min plasma separation rate. Several factors linked to liver failure were analyzed for both liver failure and nonfailure groups: (1) clinical background, including preoperative liver
Shirabe et al
Postoperative Liver Failure
305
function test-age, gender, HBs, HCV antibody, diabetes mellitus, esophageal varices, platelets, serum levels of total protein, total bilirubin, albumin, glutamic oxaloacetic transaminase, glutamic pyruvic transaminase, gamma-globulin, thymol tumidity test, cholesterol, prothrombin time, hepaplastin test, LCAT, indocyanine green retention test at 15 minutes; (2) intraoperative findings including operative time, blood loss, and resected liver weight; (3) a histologic analysis of the liver for cirrhosis; (4) volumetric study (remnant liver volume [mL/m2] was measured after a major hepatectomy). The clinical background, including the remnant liver volume, was compared among trisegmentectomy, right lobectomy, central bisegmentectomy, and left lobectomy groups. Histologic analysis of the tumor-bearing liver was obtained by pathologic evaluation of resected specimens that were classified as having hepatitis without fibrosis, precirrhosis, or cirrhosis. Volumetric study was performed before surgery. All measurements were taken by one of us (KS). In a volumetric study, serial transverse scans at 1-cm intervals were performed from the dome to the most inferior portion of the liver. Each slice of the liver was outlined and the area was calculated using an electroplanimeter. The right and left lobes of the liver were defined by a line passing through the gallbladder and the inferior vena cava, while the left lateral segment was defined by the left sagittal fissure on CT slices.13 The remnant liver volume was expressed as mL/body surface area. All data were expressed as means 6 standard deviation. We analyzed the association of each of the variables studied regarding liver failure, using Student’s t-test with continuous variables and the chisquare test of independence with categorical variables. RESULTS Seven of 80 patients (8.8%) who had major hepatectomies experienced postoperative liver failure. The patients who eventually died of liver failure lived for a range of 17 to 167 days after hepatectomy (mean survival after hepatectomy 85 6 55 days). There was a patient who died within 30 days, while 2 (28.6%) lived more than 3 months. In the major hepatectomy patients, no significant risk factors for liver failure were detected regarding age, gender, or preoperative liver function on intraoperative findings except for remnant liver volume (Table 1). Remnant liver volume in liver failure was 163 6 63 mL/m2 and 335 6 112 mL/m2, re-
306
Shirabe et al
Postoperative Liver Failure
J Am Coll Surg
Table 1. Risk Factors for Postoperative Liver Failure in Patients Undergoing a Major Hepatectomy Risk factor
Liver failure (n 5 7)
No liver failure (n 5 73)
p Value
Age (y) Gender (M/F) Hepatitis C virus antibody (1) Hepatitis B surface antigen (1) Platelets (3 103/mm3) Total protein (g/dL) Albumin (g/dL) Total bilirubin (mg/dL) Glutamic oxaloacetic transaminase (IU/L) Glutamic pyruvic transaminase (IU/L) Alkaline phosphatase (U/L) Cholesterol (mg/dL) g-globulin (%) Thymol turbidity test (U/L) Lecithin-cholesterol acyltransferase (U) Prothrombin time (%) Activated partial thromboplastin time (sec) Indocyanine green retention test (%) Resected liver weight (g) Operation time (min) Blood loss (g) Tumor size (cm) Cirrhosis (non/precirrhosis/cirrhosis) Remnant liver volume (mL/m2)
65 6 7 6:1 6 (86%) 1 (14%) 128 6 22 7.2 6 0.5 3.9 6 0.2 0.7 6 0.4 64 6 18 93 6 39 188 6 35 160 6 34 17.7 6 3.6 4.7 6 2.4 77 6 19 93.7 6 16.6 28.5 6 2.0 14.2 6 5.2 649 6 137 243 6 31 1,350 6 699 3.5 6 1.3 3:1:3 163 6 63
58 6 10 64:9 22 (30%) 50 (68%) 164 6 84 7.3 6 0.6 3.8 6 0.4 0.8 6 0.6 70 6 55 95 6 72 267 6 168 164 6 34 19 6 4.7 4.5 6 3.3 74 6 21 95 6 11 28.7 6 2.6 14.2 6 7.2 628 6 409 297 6 96 2,329 6 2,544 5.8 6 4.4 31:5:37 335 6 112
0.09 0.99 0.67 0.67 0.26 0.51 0.79 0.53 0.78 0.96 0.22 0.77 0.47 0.87 0.73 0.75 0.86 0.98 0.89 0.14 0.32 0.17 0.72 0.0002
spectively, and the difference was statistically significant (p 5 0.0002). The comparison of the data among patients who underwent a major hepatic resection according to the operative procedures is shown in Table 2. There was no significant difference in preoperative data regarding the preoperative liver function test. The remnant liver volume in triseg-
mentectomy and right lobectomy was significantly smaller than in central bisegmentectomy and left lobectomy. There was no difference between trisegmentectomy and right lobectomy. All 7 patients who died of liver failure after major hepatectomy underwent a right lobectomy of the liver. For the 47 patients who underwent right lobectomy of the liver,
Table 2. The Comparison of Data in Patients Who Underwent a Major Hepatic Resection, According to the Operative Procedure Risk factor
Trisegmentectomy (n 5 10)
Age (y) Gender (M/F) Hepatitis C virus antibody (1) Hepatitis B surface antigen (1) Platelets (3 103/mm3) Albumin (g/dL) Total bilirubin (mg/dL) Glutamic pyruvic transaminase (IUL) Indocyanine green retention test (%) Resected liver weight (g) Operation time (min) Blood loss (g) Tumor size (cm) Cirrhosis (non/precirrhosis/cirrhosis) Remnant liver volume (mL/m2) Liver failure
55 6 13 10:0 4 (40%) 4 (50%) 192 6 79 3.6 6 0.8 1.0 6 0.5 72 6 40 15.4 6 8.4 995 6 601*† 393 6 113‡ 5,140 6 5,689‡ 11.4 6 4.5*‡ 9:0:1 299 6 96*† 0
*Amount is significantly higher than in central bisegmentectomy, † Amount is significantly higher than in left lobectomy group, (p , 0.05). ‡ Amount is significantly higher than in right lobectomy,
Right lobectomy Central bisegmentectomy Left lobectomy (n 5 47) (n 5 7) (n 5 16) 60 6 10 38:9 36 (77%) 28 (17%) 152 6 66 3.6 6 0.5 0.7 6 0.3 95 6 55 13.8 6 6.9 689 6 378*† 267 6 72 1,943 6 1,778† 4.4 6 2.6 17:5:25 266 6 90*† 7(15%)
64 6 6 6:1 7 (100%) 28 (17%) 132 6 55 3.6 6 0.5 1.0 6 0.5 75 6 32 13.0 6 7.0 328 6 209† 322 6 51 3,000 6 857† 7.3 6 6.8 3:0:4 396 6 43 0
58 6 7 16:0 12 (75%) 28 (17%) 147 6 49 3.6 6 0.5 1.1 6 1.0 95 6 73 15.7 6 7.4 407 6 123 291 6 125 1,778 6 909 5.2 6 3.5 7:0:9 464 6 108 0
Vol. 188, No. 3, March 1999
Shirabe et al
307
Postoperative Liver Failure
Table 3. Risk Factors for Postoperative Liver Failure in Patients Undergoing a Right Lobectomy of the Liver Risk factor
Liver failure (n 5 7)
Age (y) Gender (M/F) Hepatitis C virus antibody (1) Hepatitis B surface antigen (1) Platelets (3 103/mm3) Total protein (g/dL) Albumin (g/dL) Total bilirubin (mg/dL) Glutamic oxaloacetic transaminase Glutamic pyruvic transaminase (IU/L) Alkaline phosphatase (U/L) Cholesterol (mg/dL) g-globulin (%) Thymol turbidity test (U/L) Lecithin-cholesterol acyltransferase (U) Prothrombin time (%) Activated partial thromboplastin time (sec) Indocyanine green retention test (%) Resected liver weight (g) Operation time (min) Blood loss (g) Tumor size (cm) Cirrhosis (non/precirrhosis/cirrhosis)
65 6 7 6:1 6 (86%) 1 (14%) 128 6 22 7.2 6 0.5 3.9 6 0.8 0.7 6 0.4 64 6 18 93 6 40 188 6 35 160 6 34 17.7 6 3.6 4.7 6 2.4 76.9 6 18.9 93.7 6 16.6 28.5 6 2.0 14.2 6 5.2 649 6 137 243 6 31 1350 6 699 3.5 6 1.3 3:1:3
we compared the data of those who died of liver failure and those who did not (Table 3). No significant differences regarding age, gender, preoperative liver function, blood loss, resected liver weight, and operation time were observed in the 2 groups that underwent a right lobectomy (Table 2). But the remnant liver volume in the patients who died of liver failure was significantly smaller than that in the patients who did not die of liver failure (p 5 0.0008). Mean remnant liver volume in the patients who died of liver failure was 163 6 63 mL/m2 while that in the patients who did not die of liver failure was 285 6 82 mL/m2. All 7 patients who died of liver failure had a remnant liver volume less than 250 mL/ m2 (Fig. 1). The risk factors for postoperative liver failure for patients with a remnant liver volume of less than 250 mL/m2 are shown in Table 3. Diabetes mellitus was the only significant risk factor (p 5 0.0072). Four of the 7 patients (57%) who died of liver failure had diabetes mellitus, while none of the patients who did not die of liver failure had diabetes mellitus. Of the 4 patients with diabetes mellitus, insulin had been used before surgery. Two patients had diabetic complications, such as retinopathy, nephropathy, and neuropathy. In other opera-
No liver failure (n 5 40)
p Value
59 6 8 32:8 30 (75%) 11 (28%) 157 6 70 7.3 6 0.5 3.9 6 0.4 0.7 6 0.3 63 6 38 95 6 58 228 6 122 165 6 30 19.3 6 5.0 4.2 6 2.9 77.1 6 19.0 98.1 6 5.8 28.7 6 2.0 13.4 6 6.9 706 6 398 272 6 75 1816 6 1148 4.6 6 2.8 13:2:25
0.19 0.99 0.99 0.65 0.27 0.57 0.96 0.81 0.92 0.95 0.40 0.70 0.42 0.64 0.98 0.20 0.87 0.79 0.71 0.32 0.31 0.30 0.40
Figure 1. A comparison of remnant liver volume of patients who died of liver failure and those without liver failure after a right lobectomy of the liver. Mean total functioning liver volume in patients who died of liver failure was 163 6 63 mL/m2 while that in patients who did not die of liver failure was 285 6 82 mL/m2. Remnant liver volume in patients who died of liver failure was significantly smaller than that in patients who did not die of liver failure (p 5 0.0008), and it was never more than 250 mL/m2.
308
Shirabe et al
Postoperative Liver Failure
tive procedure groups, all patients who underwent central bisegmentectomy and left lobectomy had a remnant liver volume .250 mL/m2, and three of 10 patients who had trisegmentectomy had a remnant liver volume ,250 mL/m2, but none of them had diabetes mellitus. DISCUSSION The incidence of liver failure after hepatic resection has decreased because of recent advances in liver surgery, such as the introduction of intraoperative ultrasonography, the use of ultrasonic dissectors, and a more stringent patient selection.3,14 The incidence of postoperative liver failure has yet to be qualified, so we reevaluated the risk factors for postoperative liver failure in the 1990s. We defined postoperative liver failure as death within 1 year after hepatectomy. The mean survival of patients who died of liver failure was 95 days, while 2 patients survived more than 3 months. In a previous study,4 postoperative death was defined as death within 3 months after hepatectomy. Two patients who died of liver failure in our study survived more than 3 months after surgery, probably because of improvements in postoperative care and the introduction of plasmapheresis.12 We therefore think that the observation period should be extended to as long as 6 months after hepatectomy. All major hepatectomy patients who died of liver failure underwent a right lobectomy of the liver. The most important risk factor for liver failure after a right lobectomy was remnant liver volume. In the patients who underwent right lobectomy of the liver and thereafter died of liver failure, remnant liver volume was never more than 250 mL/m2. In the central bisegmentectomy and left lobectomy group, no patients died of postoperative liver failure and remnant liver volume was over 250 mL/m2 in all patients. There was no difference between trisegmentectomy and right lobectomy. This seems to be because the tumor size in the trisegmentectomy group was extremely large, portal thrombus of the tumor in the right portal vein was seen, and the remnant liver after trisegmentectomy was hypertrophic. It is well known that liver regeneration is impaired in patients with chronic liver disease.13 In cases with a remnant liver volume of less than 250 mL/m2, the incidence of liver failure was extremely high (35% in our study). As a result, remnant liver volume of 250 mL/m2 may thus be a safe limit for patients with chronic liver disease. This finding has not been reported previously.
J Am Coll Surg
Yamanaka and colleagues1 demonstrated that the parenchymal hepatic resection rate measured by CT was a significant predictor for postoperative liver failure. On the other hand, Tanabe and coworkers4 showed that the same parameter was not a significant predictor. This discrepancy between the two previous papers may be the difference in the operative procedures. The incidence of major hepatectomy was 22 of 36 patients (61.1%) in the study by Yamanaka and associates1 and 26 of 121 (21.5%) in the study by Tanabe and coauthors.4 Remnant liver volume seems to be an important predictor for postoperative liver failure in major hepatic resection. In patients with a remnant liver volume of less than 250 mL/m2 after right lobectomy, the only significant risk factor was diabetes mellitus. In 3 of 10 patients who had trisegmentectomy, the remnant liver volume was less than 250 mL/m2, but none of the patients died of liver failure or had diabetes mellitus. Diabetes mellitus seems to be an important risk factor of liver failure in patients with remnant liver volume less than 250 mL/m2. In studies using diabetic rats that had undergone a partial hepatectomy, it was revealed that diabetes mellitus impaired hepatic regeneration because of the decreased hepatic intracellular energy status.14 This is one of the most important reasons why preexisting diabetes mellitus tends to cause postoperative liver failure in patients with a remnant liver volume of less than 250 mL/m2. We believe that a right lobectomy of the liver for HCC should generally not be performed in patients demonstrating viral markers and diabetes mellitus, if the expected remnant liver volume is less than 250 mL/m2. In conclusion, the expected remnant liver volume appears to be a good predictor for postoperative liver failure in patients who undergo a right lobectomy of the liver for HCC. An expected liver volume of 250 mL/m2 seems to be a safe limit for such liver resections. In patients with diabetes mellitus and expected remnant liver volume of less than 250 mL/ m2, a major hepatectomy should be avoided whenever possible. Careful patient selection based on volumetric analysis in such major hepatectomy cases should help to prevent the occurrence of postoperative liver failure. References 1. Yamanaka N, Okamoto E, Kuwata K, Tanaka N. A multiple regression for prediction of posthepatectomy liver failure. Ann Surg 1984;200:658–663.
Vol. 188, No. 3, March 1999
2. Shimada M, Matsumata T, Akazawa K, et al. Estimation of risk of major complication after hepatic resection. Am J Surg 1994;167: 399–403. 3. Matsumata T, Kanematsu T, Shirabe K, et al. Decreased morbidity and mortality rates in surgical patients with hepatocellular carcinoma. Br J Surg 1990;77:677–680. 4. Tanabe G, Sakamoto M, Akazawa K, et al. Intraoperative risk factors associated with hepatic resection. Br J Surg 1995;82:1262–1265. 5. Tjandra JJ, Fan ST, Wong J. Peri-operative mortality in hepatic resection. Aust NZ J Surg 1991;61:201–206. 6. Noguchi T, Imai T, Mizumoto R. Preoperative estimation of surgical risk of hepatectomy in cirrhotic patients. Hepatogastroenterol 1990;37:165–171. 7. Matusmata T, Kanematsu TY, Yoshida Y, et al. The indocyanine green test enables prediction of postoperative complication. World J Surg 1987;11:678–681. 8. Bozzetti F, Gennari L, Regalia E, et al. Morbidity and mortality after surgical resection of liver tumors. Analysis of 229 cases. Hepatogastroenterol 1992;39:237–241.
Shirabe et al
Postoperative Liver Failure
309
9. Hemming AW, Scudamore CH, Shackleton CR, Pudek M, Erb SR. Indocyanine green clearance as a predictor of successful hepatic resection in cirrhotic patients. Am J Surg 1992;163:515–518. 10. Stimpson REJ, Pellegrini CA, Way LW, et al. Factors affecting the morbidity of elective hepatic resection. Am J Surg 1987;153:189– 196. 11. Didokar MS, Fitzpatrick JL, Elias EG, et al. Risk factors before hepatectomy, hepatic function after hepatectomy in cirrhotic patients. Surg Gynecol Obstet 1989;169:17–26. 12. Takenaka K, Kanematsu T, Fukuzawa K, Sugimachi K. Can hepatic failure after surgery be prevented? World J Surg 1990;14:123–127. 13. Yamanaka N, Okamoto E, Kawamura E, et al. Dynamics of normal and injured human liver regeneration after hepatectomy as assessed on the basis of computed tomography and liver function. Hepatology 1993;18:79–85. 14. Mori K, Ozawa K, Yamamoto Y, et al. Response of hepatic mitochondrial redox state to oral load; redox tolerance test as a new predictor of surgical risk in hepatectomy. Ann Surg 1990;211: 438–446.
Conclusions: The expected remnant liver volume appears to be a good predictor for liver failure in patients who undergo a right lobectomy of the liver. In patients with diabetes mellitus and an expected remnant liver volume of less than 250 mL/m2, a major hepatectomy should be avoided. Careful patient selection based on volumetric analysis in major hepatectomy cases could help prevent the occurrence of postoperative liver failure. (J Am Coll Surg 1999;188:304–307. © 1999 by the American College of Surgeons)
Background: Postoperative liver failure is a lifethreatening complication after hepatic resection. Because of recent advances in liver surgery technique and a more stringent patient selection, mortality after hepatic resection has steadily decreased, but its incidence still ranges from 10% to 20%. The factors linked to postoperative liver failure in major hepatic resection in the modern era should be reevaluated. Study Design: Of 80 patients with viral markers (hepatitis C viral antibody or hepatitis B surface antigen) who underwent major hepatic resections (no less than bisegmentectomies) for hepatocellular carcinoma between 1990 and 1996, 7 patients (8.8%) died of postoperative liver failure within 6 months after hepatectomy. The cause of liver failure was analyzed based on both the preoperative data and the intraoperative findings. In addition, since all the patients who died of liver failure underwent a right hepatic lobectomy, a further data analysis was also done in 47 patients who underwent a right lobectomy of the liver. A volumetric analysis by CT was then done to evaluate the remnant liver volume.
Hepatic resection is the accepted treatment for various types of liver tumors and is increasingly performed with ever expanding indications and surgical limits. In particular, hepatic resection for hepatocellular carcinoma (HCC) is now being performed on patients who have chronic liver disease.1-3 In patients with chronic liver disease, however, postoperative liver failure is still an important cause of death after hepatic resection. Because of recent advances in liver surgery, such as the introduction of intraoperative ultrasonography and ultrasonic dissectors, and more stringent patient selection, the mortality rate after hepatic resection has decreased,3-5 but the incidence of mortality still ranges from 10% to 20%.4-6 An extremely high incidence of liver failure has been reported for patients who undergo a major hepatectomy for HCC.5 Many factors linked to postoperative liver failure have been reported, including preoperative liver function, remnant liver volume, and amount of blood loss during surgery.6-12 The causes of liver failure are complex because of broad differences in liver function, tumor size and location, operative procedures, and different amounts of blood loss for each patient; as a result, an accurate prediction of liver failure remains difficult. Optimal operative procedures have been established and the amount of blood
Results: Between the patients with liver failure and those without liver failure who underwent a right lobectomy, there were no significant differences in preoperative data or intraoperative findings. Volumetric analysis revealed that the remnant liver volume of patients who died of liver failure was significantly smaller than that of patients who lived (p 5 0.008). The incidence of liver failure in patients with a remnant liver volume of less than 250 mL/m2 was 7 of 20 (38%), while it was 0 of 27 in patients with a liver volume of no less than 250 mL/m2 (p 5 0.0012). The only significant risk factor for liver failure in patients with a remnant liver volume of less than 250 mL/m2 was diabetes mellitus (p 5 0.0072). Received April 13, 1998; Revised November 4, 1998; Accepted November 4, 1998. From the Second Department of Surgery, Kyushu University, Fukuoka, Japan. Correspondence address: Ken Shirabe, MD, Second Department of Surgery, Faculty of Medicine, Kyushu University, 3-1-1 Maidashi, Higashi-ku, Fukuoka, 812, Japan. © 1999 by the American College of Surgeons Published by Elsevier Science Inc.
304
ISSN 1072-7515/99/$19.00 PII S1072-7515(98)00301-9
Vol. 188, No. 3, March 1999
loss has also decreased significantly during hepatic resection,3 but a reevaluation of the factors linked to postoperative liver failure is necessary. We analyzed the cause of liver failure occurring in patients who underwent a major hepatectomy in the modern era. METHODS Between January 1990 and February 1996, 284 patients with HCC underwent a hepatic resection at the Second Department of Surgery, Kyushu University. Postoperatively, 32 patients without viral markers such as hepatitis B surface antigen (HBs) and hepatitis C viral (HCV) antibody, or both, did not die of liver failure after hepatectomy. Of the 252 patients with viral markers such as HBs and HCV antibody, or both, 80 patients who underwent a major hepatectomy (no less than bisegmentectomies) were entered into this study. The study group consisted of 70 men and 10 women. The operative procedures performed on these patients were as follows: 10 patients underwent a trisegmentectomy (right trisegmentectomy in 9, left trisegmentectomy in 1), and 70 patients underwent a bisegmentectomy (right lobectomy in 47, left lobectomy in 16, central bisegmentectomy in 7). Postoperative liver failure was defined as appearance of hepatic encephalopathy, progressively increasing hyperbilirubinemia without hemolytic or obstructive mechanism, decreasing activity of blood coagulation assessed by a prothrombin time, and lecitin–cholesterol acyltransferase (LCAT) deficiency, according to the criteria by Takenaka and associates.13 Postoperative liver failure was determined as death from liver failure within 6 months after hepatic resection. Seven of 9 patients (77.8%) who had postoperative liver failure died within 6 months after surgery; these 7 patients were regarded as postoperative liver failure cases in this study. None of the patients died of variceal bleeding or technical failure such as ipsilateral portal vein obstruction. None showed evidence of recurrent cancer at the time of death. Once the patients were diagnosed as having postoperative liver failure, plasma exchange was carried out. Plasma exchange therapy was given by the membrane filtration method using Plasmaflo AP05H (Asahi Medical, Tokyo) using 2000–2500 mL of fresh frozen plasma at 50–80 mL/min blood flow rate and 15–25 mL/min plasma separation rate. Several factors linked to liver failure were analyzed for both liver failure and nonfailure groups: (1) clinical background, including preoperative liver
Shirabe et al
Postoperative Liver Failure
305
function test-age, gender, HBs, HCV antibody, diabetes mellitus, esophageal varices, platelets, serum levels of total protein, total bilirubin, albumin, glutamic oxaloacetic transaminase, glutamic pyruvic transaminase, gamma-globulin, thymol tumidity test, cholesterol, prothrombin time, hepaplastin test, LCAT, indocyanine green retention test at 15 minutes; (2) intraoperative findings including operative time, blood loss, and resected liver weight; (3) a histologic analysis of the liver for cirrhosis; (4) volumetric study (remnant liver volume [mL/m2] was measured after a major hepatectomy). The clinical background, including the remnant liver volume, was compared among trisegmentectomy, right lobectomy, central bisegmentectomy, and left lobectomy groups. Histologic analysis of the tumor-bearing liver was obtained by pathologic evaluation of resected specimens that were classified as having hepatitis without fibrosis, precirrhosis, or cirrhosis. Volumetric study was performed before surgery. All measurements were taken by one of us (KS). In a volumetric study, serial transverse scans at 1-cm intervals were performed from the dome to the most inferior portion of the liver. Each slice of the liver was outlined and the area was calculated using an electroplanimeter. The right and left lobes of the liver were defined by a line passing through the gallbladder and the inferior vena cava, while the left lateral segment was defined by the left sagittal fissure on CT slices.13 The remnant liver volume was expressed as mL/body surface area. All data were expressed as means 6 standard deviation. We analyzed the association of each of the variables studied regarding liver failure, using Student’s t-test with continuous variables and the chisquare test of independence with categorical variables. RESULTS Seven of 80 patients (8.8%) who had major hepatectomies experienced postoperative liver failure. The patients who eventually died of liver failure lived for a range of 17 to 167 days after hepatectomy (mean survival after hepatectomy 85 6 55 days). There was a patient who died within 30 days, while 2 (28.6%) lived more than 3 months. In the major hepatectomy patients, no significant risk factors for liver failure were detected regarding age, gender, or preoperative liver function on intraoperative findings except for remnant liver volume (Table 1). Remnant liver volume in liver failure was 163 6 63 mL/m2 and 335 6 112 mL/m2, re-
306
Shirabe et al
Postoperative Liver Failure
J Am Coll Surg
Table 1. Risk Factors for Postoperative Liver Failure in Patients Undergoing a Major Hepatectomy Risk factor
Liver failure (n 5 7)
No liver failure (n 5 73)
p Value
Age (y) Gender (M/F) Hepatitis C virus antibody (1) Hepatitis B surface antigen (1) Platelets (3 103/mm3) Total protein (g/dL) Albumin (g/dL) Total bilirubin (mg/dL) Glutamic oxaloacetic transaminase (IU/L) Glutamic pyruvic transaminase (IU/L) Alkaline phosphatase (U/L) Cholesterol (mg/dL) g-globulin (%) Thymol turbidity test (U/L) Lecithin-cholesterol acyltransferase (U) Prothrombin time (%) Activated partial thromboplastin time (sec) Indocyanine green retention test (%) Resected liver weight (g) Operation time (min) Blood loss (g) Tumor size (cm) Cirrhosis (non/precirrhosis/cirrhosis) Remnant liver volume (mL/m2)
65 6 7 6:1 6 (86%) 1 (14%) 128 6 22 7.2 6 0.5 3.9 6 0.2 0.7 6 0.4 64 6 18 93 6 39 188 6 35 160 6 34 17.7 6 3.6 4.7 6 2.4 77 6 19 93.7 6 16.6 28.5 6 2.0 14.2 6 5.2 649 6 137 243 6 31 1,350 6 699 3.5 6 1.3 3:1:3 163 6 63
58 6 10 64:9 22 (30%) 50 (68%) 164 6 84 7.3 6 0.6 3.8 6 0.4 0.8 6 0.6 70 6 55 95 6 72 267 6 168 164 6 34 19 6 4.7 4.5 6 3.3 74 6 21 95 6 11 28.7 6 2.6 14.2 6 7.2 628 6 409 297 6 96 2,329 6 2,544 5.8 6 4.4 31:5:37 335 6 112
0.09 0.99 0.67 0.67 0.26 0.51 0.79 0.53 0.78 0.96 0.22 0.77 0.47 0.87 0.73 0.75 0.86 0.98 0.89 0.14 0.32 0.17 0.72 0.0002
spectively, and the difference was statistically significant (p 5 0.0002). The comparison of the data among patients who underwent a major hepatic resection according to the operative procedures is shown in Table 2. There was no significant difference in preoperative data regarding the preoperative liver function test. The remnant liver volume in triseg-
mentectomy and right lobectomy was significantly smaller than in central bisegmentectomy and left lobectomy. There was no difference between trisegmentectomy and right lobectomy. All 7 patients who died of liver failure after major hepatectomy underwent a right lobectomy of the liver. For the 47 patients who underwent right lobectomy of the liver,
Table 2. The Comparison of Data in Patients Who Underwent a Major Hepatic Resection, According to the Operative Procedure Risk factor
Trisegmentectomy (n 5 10)
Age (y) Gender (M/F) Hepatitis C virus antibody (1) Hepatitis B surface antigen (1) Platelets (3 103/mm3) Albumin (g/dL) Total bilirubin (mg/dL) Glutamic pyruvic transaminase (IUL) Indocyanine green retention test (%) Resected liver weight (g) Operation time (min) Blood loss (g) Tumor size (cm) Cirrhosis (non/precirrhosis/cirrhosis) Remnant liver volume (mL/m2) Liver failure
55 6 13 10:0 4 (40%) 4 (50%) 192 6 79 3.6 6 0.8 1.0 6 0.5 72 6 40 15.4 6 8.4 995 6 601*† 393 6 113‡ 5,140 6 5,689‡ 11.4 6 4.5*‡ 9:0:1 299 6 96*† 0
*Amount is significantly higher than in central bisegmentectomy, † Amount is significantly higher than in left lobectomy group, (p , 0.05). ‡ Amount is significantly higher than in right lobectomy,
Right lobectomy Central bisegmentectomy Left lobectomy (n 5 47) (n 5 7) (n 5 16) 60 6 10 38:9 36 (77%) 28 (17%) 152 6 66 3.6 6 0.5 0.7 6 0.3 95 6 55 13.8 6 6.9 689 6 378*† 267 6 72 1,943 6 1,778† 4.4 6 2.6 17:5:25 266 6 90*† 7(15%)
64 6 6 6:1 7 (100%) 28 (17%) 132 6 55 3.6 6 0.5 1.0 6 0.5 75 6 32 13.0 6 7.0 328 6 209† 322 6 51 3,000 6 857† 7.3 6 6.8 3:0:4 396 6 43 0
58 6 7 16:0 12 (75%) 28 (17%) 147 6 49 3.6 6 0.5 1.1 6 1.0 95 6 73 15.7 6 7.4 407 6 123 291 6 125 1,778 6 909 5.2 6 3.5 7:0:9 464 6 108 0
Vol. 188, No. 3, March 1999
Shirabe et al
307
Postoperative Liver Failure
Table 3. Risk Factors for Postoperative Liver Failure in Patients Undergoing a Right Lobectomy of the Liver Risk factor
Liver failure (n 5 7)
Age (y) Gender (M/F) Hepatitis C virus antibody (1) Hepatitis B surface antigen (1) Platelets (3 103/mm3) Total protein (g/dL) Albumin (g/dL) Total bilirubin (mg/dL) Glutamic oxaloacetic transaminase Glutamic pyruvic transaminase (IU/L) Alkaline phosphatase (U/L) Cholesterol (mg/dL) g-globulin (%) Thymol turbidity test (U/L) Lecithin-cholesterol acyltransferase (U) Prothrombin time (%) Activated partial thromboplastin time (sec) Indocyanine green retention test (%) Resected liver weight (g) Operation time (min) Blood loss (g) Tumor size (cm) Cirrhosis (non/precirrhosis/cirrhosis)
65 6 7 6:1 6 (86%) 1 (14%) 128 6 22 7.2 6 0.5 3.9 6 0.8 0.7 6 0.4 64 6 18 93 6 40 188 6 35 160 6 34 17.7 6 3.6 4.7 6 2.4 76.9 6 18.9 93.7 6 16.6 28.5 6 2.0 14.2 6 5.2 649 6 137 243 6 31 1350 6 699 3.5 6 1.3 3:1:3
we compared the data of those who died of liver failure and those who did not (Table 3). No significant differences regarding age, gender, preoperative liver function, blood loss, resected liver weight, and operation time were observed in the 2 groups that underwent a right lobectomy (Table 2). But the remnant liver volume in the patients who died of liver failure was significantly smaller than that in the patients who did not die of liver failure (p 5 0.0008). Mean remnant liver volume in the patients who died of liver failure was 163 6 63 mL/m2 while that in the patients who did not die of liver failure was 285 6 82 mL/m2. All 7 patients who died of liver failure had a remnant liver volume less than 250 mL/ m2 (Fig. 1). The risk factors for postoperative liver failure for patients with a remnant liver volume of less than 250 mL/m2 are shown in Table 3. Diabetes mellitus was the only significant risk factor (p 5 0.0072). Four of the 7 patients (57%) who died of liver failure had diabetes mellitus, while none of the patients who did not die of liver failure had diabetes mellitus. Of the 4 patients with diabetes mellitus, insulin had been used before surgery. Two patients had diabetic complications, such as retinopathy, nephropathy, and neuropathy. In other opera-
No liver failure (n 5 40)
p Value
59 6 8 32:8 30 (75%) 11 (28%) 157 6 70 7.3 6 0.5 3.9 6 0.4 0.7 6 0.3 63 6 38 95 6 58 228 6 122 165 6 30 19.3 6 5.0 4.2 6 2.9 77.1 6 19.0 98.1 6 5.8 28.7 6 2.0 13.4 6 6.9 706 6 398 272 6 75 1816 6 1148 4.6 6 2.8 13:2:25
0.19 0.99 0.99 0.65 0.27 0.57 0.96 0.81 0.92 0.95 0.40 0.70 0.42 0.64 0.98 0.20 0.87 0.79 0.71 0.32 0.31 0.30 0.40
Figure 1. A comparison of remnant liver volume of patients who died of liver failure and those without liver failure after a right lobectomy of the liver. Mean total functioning liver volume in patients who died of liver failure was 163 6 63 mL/m2 while that in patients who did not die of liver failure was 285 6 82 mL/m2. Remnant liver volume in patients who died of liver failure was significantly smaller than that in patients who did not die of liver failure (p 5 0.0008), and it was never more than 250 mL/m2.
308
Shirabe et al
Postoperative Liver Failure
tive procedure groups, all patients who underwent central bisegmentectomy and left lobectomy had a remnant liver volume .250 mL/m2, and three of 10 patients who had trisegmentectomy had a remnant liver volume ,250 mL/m2, but none of them had diabetes mellitus. DISCUSSION The incidence of liver failure after hepatic resection has decreased because of recent advances in liver surgery, such as the introduction of intraoperative ultrasonography, the use of ultrasonic dissectors, and a more stringent patient selection.3,14 The incidence of postoperative liver failure has yet to be qualified, so we reevaluated the risk factors for postoperative liver failure in the 1990s. We defined postoperative liver failure as death within 1 year after hepatectomy. The mean survival of patients who died of liver failure was 95 days, while 2 patients survived more than 3 months. In a previous study,4 postoperative death was defined as death within 3 months after hepatectomy. Two patients who died of liver failure in our study survived more than 3 months after surgery, probably because of improvements in postoperative care and the introduction of plasmapheresis.12 We therefore think that the observation period should be extended to as long as 6 months after hepatectomy. All major hepatectomy patients who died of liver failure underwent a right lobectomy of the liver. The most important risk factor for liver failure after a right lobectomy was remnant liver volume. In the patients who underwent right lobectomy of the liver and thereafter died of liver failure, remnant liver volume was never more than 250 mL/m2. In the central bisegmentectomy and left lobectomy group, no patients died of postoperative liver failure and remnant liver volume was over 250 mL/m2 in all patients. There was no difference between trisegmentectomy and right lobectomy. This seems to be because the tumor size in the trisegmentectomy group was extremely large, portal thrombus of the tumor in the right portal vein was seen, and the remnant liver after trisegmentectomy was hypertrophic. It is well known that liver regeneration is impaired in patients with chronic liver disease.13 In cases with a remnant liver volume of less than 250 mL/m2, the incidence of liver failure was extremely high (35% in our study). As a result, remnant liver volume of 250 mL/m2 may thus be a safe limit for patients with chronic liver disease. This finding has not been reported previously.
J Am Coll Surg
Yamanaka and colleagues1 demonstrated that the parenchymal hepatic resection rate measured by CT was a significant predictor for postoperative liver failure. On the other hand, Tanabe and coworkers4 showed that the same parameter was not a significant predictor. This discrepancy between the two previous papers may be the difference in the operative procedures. The incidence of major hepatectomy was 22 of 36 patients (61.1%) in the study by Yamanaka and associates1 and 26 of 121 (21.5%) in the study by Tanabe and coauthors.4 Remnant liver volume seems to be an important predictor for postoperative liver failure in major hepatic resection. In patients with a remnant liver volume of less than 250 mL/m2 after right lobectomy, the only significant risk factor was diabetes mellitus. In 3 of 10 patients who had trisegmentectomy, the remnant liver volume was less than 250 mL/m2, but none of the patients died of liver failure or had diabetes mellitus. Diabetes mellitus seems to be an important risk factor of liver failure in patients with remnant liver volume less than 250 mL/m2. In studies using diabetic rats that had undergone a partial hepatectomy, it was revealed that diabetes mellitus impaired hepatic regeneration because of the decreased hepatic intracellular energy status.14 This is one of the most important reasons why preexisting diabetes mellitus tends to cause postoperative liver failure in patients with a remnant liver volume of less than 250 mL/m2. We believe that a right lobectomy of the liver for HCC should generally not be performed in patients demonstrating viral markers and diabetes mellitus, if the expected remnant liver volume is less than 250 mL/m2. In conclusion, the expected remnant liver volume appears to be a good predictor for postoperative liver failure in patients who undergo a right lobectomy of the liver for HCC. An expected liver volume of 250 mL/m2 seems to be a safe limit for such liver resections. In patients with diabetes mellitus and expected remnant liver volume of less than 250 mL/ m2, a major hepatectomy should be avoided whenever possible. Careful patient selection based on volumetric analysis in such major hepatectomy cases should help to prevent the occurrence of postoperative liver failure. References 1. Yamanaka N, Okamoto E, Kuwata K, Tanaka N. A multiple regression for prediction of posthepatectomy liver failure. Ann Surg 1984;200:658–663.
Vol. 188, No. 3, March 1999
2. Shimada M, Matsumata T, Akazawa K, et al. Estimation of risk of major complication after hepatic resection. Am J Surg 1994;167: 399–403. 3. Matsumata T, Kanematsu T, Shirabe K, et al. Decreased morbidity and mortality rates in surgical patients with hepatocellular carcinoma. Br J Surg 1990;77:677–680. 4. Tanabe G, Sakamoto M, Akazawa K, et al. Intraoperative risk factors associated with hepatic resection. Br J Surg 1995;82:1262–1265. 5. Tjandra JJ, Fan ST, Wong J. Peri-operative mortality in hepatic resection. Aust NZ J Surg 1991;61:201–206. 6. Noguchi T, Imai T, Mizumoto R. Preoperative estimation of surgical risk of hepatectomy in cirrhotic patients. Hepatogastroenterol 1990;37:165–171. 7. Matusmata T, Kanematsu TY, Yoshida Y, et al. The indocyanine green test enables prediction of postoperative complication. World J Surg 1987;11:678–681. 8. Bozzetti F, Gennari L, Regalia E, et al. Morbidity and mortality after surgical resection of liver tumors. Analysis of 229 cases. Hepatogastroenterol 1992;39:237–241.
Shirabe et al
Postoperative Liver Failure
309
9. Hemming AW, Scudamore CH, Shackleton CR, Pudek M, Erb SR. Indocyanine green clearance as a predictor of successful hepatic resection in cirrhotic patients. Am J Surg 1992;163:515–518. 10. Stimpson REJ, Pellegrini CA, Way LW, et al. Factors affecting the morbidity of elective hepatic resection. Am J Surg 1987;153:189– 196. 11. Didokar MS, Fitzpatrick JL, Elias EG, et al. Risk factors before hepatectomy, hepatic function after hepatectomy in cirrhotic patients. Surg Gynecol Obstet 1989;169:17–26. 12. Takenaka K, Kanematsu T, Fukuzawa K, Sugimachi K. Can hepatic failure after surgery be prevented? World J Surg 1990;14:123–127. 13. Yamanaka N, Okamoto E, Kawamura E, et al. Dynamics of normal and injured human liver regeneration after hepatectomy as assessed on the basis of computed tomography and liver function. Hepatology 1993;18:79–85. 14. Mori K, Ozawa K, Yamamoto Y, et al. Response of hepatic mitochondrial redox state to oral load; redox tolerance test as a new predictor of surgical risk in hepatectomy. Ann Surg 1990;211: 438–446.