Recurrence-free survival of 204 patients with cirrhosis who underwent hepatectomy for single hepatocellular carcinoma of 5 cm or smaller. Patients were grouped according to the following viral status: hepatitis B virus (HBV) positive and hepatitis C virus (HCV) negative (group 1); HBV negative and HCV positive (group 2); HBV negative and HCV negative (group 3); and HBV positive and HCV positive (group 4). Group 1 vs group 2, P = .007. Group 1 vs group 3, P = .05. Group 1 vs group 4, P = .80. Group 2 vs group 3, P = .90. Group 2 vs group 4, P = .10. Group 3 vs group 4, P = .10.
Recurrence-free survival according to scores resulting from the presence (1 point) or absence (0 points) of each factor with a significant effect at multivariate analysis (hepatitis B virus infection and Edmonson grade III or IV). A score of 0 vs a score of 1, P = .001. A score of 0 vs a score of 2, P < .001. A score of 1 vs a score of 2, P = .03.
Cescon M, Cucchetti A, Grazi GL, Ferrero A, Viganò L, Ercolani G, Ravaioli M, Zanello M, Andreone P, Capussotti L, Pinna AD. Role of Hepatitis B Virus Infection in the Prognosis After Hepatectomy for Hepatocellular Carcinoma in Patients With CirrhosisA Western Dual-Center Experience. Arch Surg. 2009;144(10):906-913. doi:10.1001/archsurg.2009.99
The role of hepatitis B virus (HBV) infection in determining the prognosis after hepatectomy for hepatocellular carcinoma (HCC) in patients with cirrhosis is controversial.
Retrospective study based on multicenter prospectively updated databases.
Two tertiary referral centers specializing in hepatobiliary surgery.
Two hundred four consecutive patients with cirrhosis undergoing hepatectomy for single nodules of HCC of 5 cm or smaller from January 1, 1997, through September 30, 2006.
Patients were divided into the following groups according to their preoperative viral status: HBV positive and hepatitis C virus (HCV) negative (group 1); HBV negative and HCV positive (group 2); HBV negative and HCV negative (group 3); and HBV positive and HCV positive (group 4).
Main Outcome Measures
A multivariate analysis was performed to determine factors associated with recurrence-free survival (RFS) among demographic, clinical, pathological, and surgical variables.
The 2 centers had comparable RFS and early and late recurrence rates. Five-year RFS was significantly higher in groups 2 and 3 compared with group 1 (38%, 34%, and 9%, respectively; P = .007 and P = .05). Factors independently associated with RFS were HBV infection (P = .009; odds ratio, 1.79; 95% confidence interval, 1.15-2.78) and poor tumor differentiation (P < .001; odds ratio, 2.01; 95% confidence interval, 1.36-2.96). The concomitance of 0, 1, or 2 risk factors led to 5-year RFS rates of 49%, 20%, and 8%, respectively (P < .001).
Infection with HBV is a strong predictive factor for lower RFS after hepatectomy for a single nodule of HCC of 5 cm or smaller in patients with cirrhosis, providing a further basis for adjuvant antiviral treatment. Patients who are seropositive for HBV with poorly differentiated HCC should also be considered to be at a high risk of recurrence and possibly included in a policy of salvage liver transplantation.
Hepatocellular carcinoma (HCC) is 1 of the 5 most common malignant neoplasms, and its frequency is increasing in Western countries.1 For patients with early-stage tumors and preserved hepatic function, liver resection remains the best therapeutic choice. However, the rate of tumor recurrence after hepatectomy remains high, ranging from 60% to 100% in the long-term observation.2- 17
Factors related to tumor recurrence in patients undergoing liver resection for HCC have been studied extensively.2,4- 11 Among them, the preoperative levels of α-fetoprotein and transaminases, degree of liver fibrosis, presence of multiple tumors, tumor size, vascular invasion, satellite nodules, incomplete tumor capsule, and requirement for intraoperative blood transfusions are the most widely reported.4- 11
The presence of hepatitis B virus (HBV) infection has a controversial role in determining the risk of posthepatectomy tumor recurrence.7,12,13 The carcinogenetic potential of the nontumorous liver can be defined by elevated hepatitis activity and viral replication indexes, such as high serum HBV DNA levels and hepatitis B e antigen (HBeAg) positivity.7,13
Although most studies comparing hepatitis C virus (HCV) and HBV infections demonstrated that the former generally carries a higher risk of recurrence after resection,14,15 some recent studies showed that serum HBeAg positivity is a predictor of tumor recurrence.13,16 All of these studies are from Eastern countries and included both cirrhotic and noncirrhotic patients (ie, those with variable degrees of chronic hepatitis), with single or multiple HCCs.
In the present study, we analyzed factors affecting the risk of tumor recurrence in a series of 204 hepatectomies for single nodules of HCC measuring 5 cm or less in diameter and arising in cirrhotic livers, with particular reference to the effect of HBV infection on tumor recurrence.
From January 1, 1997, through September 30, 2006, a total of 142 consecutive curative hepatic resections of single nodules of HCC 5 cm or smaller in diameter in cirrhotic patients were performed at the Department of Surgery and Transplantation, University of Bologna. During the same period, 69 hepatectomies with the same indications were performed at the Unit of Surgical Oncology, Institute for Cancer Research and Treatment. Data were collected from the 2 institutional prospectively updated databases. Of these 211 patients, 7 (3%) died within 30 days after surgery for causes unrelated to HCC recurrence and were excluded from the study. Among the remaining 204 patients who formed the study population, 25 (12%) were HBV positive and HCV negative (group 1); 130 (64%) were HBV negative and HCV positive (group 2); 35 (17%) were HBV negative and HCV negative (group 3); and 14 (7%) were HBV positive and HCV positive (group 4).
Infection with HBV was defined as positivity for hepatitis B surface antigen (HBsAg) or for anti–hepatitis B core antibodies (HBcAb) at the time of surgery. Accordingly, 21 patients in group 1 (84%) were seropositive for both HBsAg and HBcAb, and 4 (16%) were seropositive for HBcAb alone. Only 1 of these patients (4%) was seropositive for HBeAg, and 17 (68%) had detectable serum HBV DNA at surgery. In group 4, 9 patients (64%) were seropositive for HBsAg, 5 (36%) were seropositive for HBcAb alone, 1 (7%) was seropositive for HBeAg, and 4 (29%) had detectable serum HBV DNA.
Infection with HCV was defined as positivity for serum anti-HCV antibodies. Serum HBV and HCV markers were detected using common commercial kits.
Curative resection was defined as the removal of the entire neoplastic mass with a histologically proved tumor-free surgical margin and no tumor thrombus in the major branches of the portal vein, hepatic veins, and bile duct. The nature and size of the tumor and the presence of cirrhosis in the nontumorous hepatic parenchyma were all confirmed by results of the final pathological examination.
The indication for hepatectomy was given when a complete removal of the tumor was considered possible, in accordance with the volume of the remnant liver, the technical feasibility (based on the size and location of the lesion), and the preoperative liver function, which was assessed as previously reported.11,17
The extent of the hepatectomy was based on the International Hepato-Pancreato-Biliary Association classification.18 Major hepatic resection was defined as the removal of more than 2 segments. Anatomical resections included subsegmentectomy, segmentectomy, sectorectomy, left lobectomy, and left or right hepatectomy. Nonanatomical resections consisted of non–segment-oriented (partial or wedge) resections. Surgical techniques have been reported elsewhere.11,17 Intraoperative ultrasonography was always used to detect nonvisible, nonpalpable nodules and to check the resection plane.
After discharge, patients were followed up with measurement of serum α-fetoprotein levels and hepatic ultrasonography every 3 months and with abdominal computed tomography and chest radiographs every 6 months. In the case of suspected tumor recurrence, computed tomography and/or magnetic resonance imaging were performed. In selected cases, bone scan, positron emission tomography, angiography of the celiac trunk with injection of an oily contrast agent (lipiodol), and, if necessary, liver biopsy were performed.
Various treatment modalities were adopted depending on the site and extent of the recurrence and clinical conditions, as previously reported.10,11,17 In particular, repeated liver resections were performed with the same indications as the initial hepatectomy. Liver transplantation was performed in selected cases of posthepatectomy tumor recurrence, if the recurrence fulfilled the Milan criteria.19
The aim of this study was to analyze recurrence-free survival (RFS) according to the patient's viral status (groups 1, 2, 3, and 4) and to investigate demographic, clinical, surgical, and oncological factors potentially affecting RFS and overall survival in the study population.
We evaluated the following variables: sex, age (≤65 or >65 years, corresponding to the median age of the study population), esophageal varices (present or absent), platelet count (≤100 or >100 × 103/μL), serum albumin level (≤3.5 or >3.5 g/dL), aspartate aminotransferase level (≤76 or >76 U/L, corresponding to 2-fold normal values), alanine aminotransferase level (≤82 or >82 U/L, corresponding to 2-fold normal values), total serum bilirubin level (≤1.5 or >1.5 mg/mL), international normalized ratio (INR; ≤1.2 or >1.2), Child-Turcotte-Pugh score (A vs B),20,21 Model for End-Stage Liver Disease (MELD) score (≤9 or >9),22 HBV infection (positive or negative), HCV infection (positive or negative), α-fetoprotein level (≤20 or >20 ng/mL), tumor size (≤3 or >3 cm), tumor capsule (present or incomplete/absent), satellite nodules (present or absent), microvascular invasion (present or absent), tumor differentiation according to the Edmonson classification (grade I or II vs grade III or IV),23 type of hepatectomy (anatomical or nonanatomical), extent of the hepatectomy (minor or major), and intraoperative red blood cell transfusion requirement (yes or no).
The MELD score was retrospectively calculated on the basis of biochemistry studies on the day before surgery according to the following equation22:
MELD = [(0.957 × LN [Creatinine in Milligrams per Deciliter]) + (0.378 × LN [Bilirubin in Milligrams per Deciliter]) + (1.12 × LN [INR]) + 0.643] × 10,
in which LN indicates loge.
Results were expressed as mean (SD) or as median (range) of values. Differences between continuous variables were evaluated with the 1-way analysis of variance test with the least significant difference for multiple comparisons. Differences were assessed between categorical variables and with the χ2 test or the Fisher exact test, when appropriate. The Kaplan-Meier method was used for the analysis of prognostic factors for RFS and overall survival, and the differences between groups were compared by the log-rank test. Overall survival was computed from the day of surgery to the day of death or the last follow-up visit. Recurrence-free survival was computed from the day of surgery to the date of detection of tumor recurrence. Patients who died of causes unrelated to tumor recurrence or were lost to follow-up without known tumor recurrence were censored at the date of death or the last visit, respectively. Variables achieving statistical significance at the univariate analysis were put in the multivariate analysis, performed with the Cox proportional hazard model. P < .05 was considered statistically significant. Statistical analysis was performed with a commercially available software package (SPSS, version 13; SPSS Inc, Chicago, Illinois).
To convert albumin to grams per liter, multiply by 10; alanine aminotransferase and aspartate aminotransferase to microkatals per liter, multiply by 0.0167; bilirubin to micromoles per liter, multiply by 17.104; α-fetoprotein to micrograms per liter, multiply by 1; and platelet count to cells ×109 per liter, multiply by 1.
The study included 158 men (77%) and 46 woman (23%). Median age was 65 (range, 36-85) years. Demographic, clinical, pathological, and surgical variables according to the viral status of the study population are reported in Table 1.
Groups 1, 3, and 4 had a higher prevalence of male sex compared with group 2. Mean age was lower in group 1 vs group 2. Preoperative levels of transaminases were significantly lower in group 3 vs the other groups, and the highest levels were observed in group 4. Group 2 had a higher prevalence of patients with alanine aminotransferase levels of greater than 2-fold normal levels compared with group 1. Group 4 showed higher total bilirubin levels compared with the other groups.
Regarding tumor variables, the mean tumor diameter and the prevalence of patients with a tumor diameter of greater than 3 cm were higher in group 3 vs group 2. The absence or incompleteness of a tumor capsule was more frequent in group 1 vs group 4. Group 1 had a lower proportion of tumors with satellite nodules compared with group 3. Microvascular invasion and poorly differentiated tumors were more frequent in group 2 vs group 3. Finally, a higher number of major hepatectomies was performed in group 3 vs groups 1 and 2.
In group 1, 9 patients (36%) received postoperative antiviral treatment with lamivudine. In group 2, 5 patients (4%) received postoperative antiviral treatment with interferon alfa-2b. No antiviral or adjuvant treatments were administered to patients in groups 3 and 4.
The median follow-up was 30 (range, 1-111) months. At the end of the study period, 134 patients (66%) were alive and 70 (34%) had died. Causes of death were related to HCC recurrence in 36 cases (18%), to progression of cirrhosis in 25 (12%), and to other conditions in 9 (4%). Overall 3- and 5-year patient survival rates were 73% and 57%, respectively.
Recurrence of HCC developed in 108 patients (53%). There were 68 early (≤2 years after hepatectomy) and 40 late (>2 years after hepatectomy) recurrences. The 2 centers participating in the study had comparable proportions of early and late recurrences: 50 (67%) and 25 (33%), respectively, in Bologna, and 18 (55%) and 15 (45%), respectively, in Turin (P = .20).
Overall 3- and 5-year RFS rates were 53% and 32%, respectively. The 3- and 5-year RFS rates were comparable between the 2 participating centers (51% and 31%, respectively, for Bologna; 58% and 35%, respectively, for Turin; P = .60).
Treatment of recurrence consisted of various combinations of percutaneous ablations (ethanol injection and/or radiofrequency) and transarterial chemoembolization (60 cases [56%]), repeated resections (11 cases [10%]), liver transplantation (5 cases [5%]), and conservative treatments (32 cases [30%]).
Tumors recurred in 19 patients (65%) in group 1, 65 (50%) in group 2, 16 (46%) in group 3, and 8 (57%) in group 4 (group 1 vs groups 2 and 3, P = .01; group 1 vs group 4, P = .20; group 2 vs group 3, P = .60; group 2 vs group 4, P = .70; and group 3 vs group 4, P = .40).
Recurrence-free survival according to patient viral status is reported in Figure 1. Three- and 5-year survival rates were 34% and 9%, respectively, in group 1; 57% and 38%, respectively, in group 2; 63% and 34%, respectively, in group 3; and 33% and not computable, respectively, in group 4. The RFS rate was significantly lower in group 1 vs groups 2 and 3 (P = .007 and P = .05, respectively) and was not significantly different between groups 1 and 4 (P = .80), groups 2 and 3 (P = .90), groups 2 and 4 (P = .10), and groups 3 and 4 (P = .10).
By the univariate analysis, male sex (P = .01), HBV infection (P = .002), incomplete or absent tumor capsule (P = .02), the presence of satellite nodules (P = .02), microvascular invasion (P = .001), and Edmonson grade III or IV (P < .001) negatively affected RFS (Table 2).
A platelet count of 100 × 103/μL or less (P = .009), serum albumin level of 3.5 g/dL or less (P = .02), total bilirubin level of greater than 1.5 mg/dL (P = .002), Child-Turcotte-Pugh class B (P = .01), MELD score of greater than 9 (P = .02), tumor diameter of greater than 3 cm (P = .01), the presence of satellite nodules (P = .01), microvascular invasion (P = .01), Edmonson grade III or IV (P = .05), and the requirement for intraoperative red blood cell transfusions (P = .01) were correlated with worse overall survival (Table 3).
By multivariate analysis, HBV infection (P = .009) and Edmonson grade III or IV (P < .001) were independent predictors of lower RFS, whereas a low platelet count (P = .01), an elevated total bilirubin level (P = .002), the presence of satellite nodules (P = .01), and microvascular invasion (P = .003) were correlated with worse overall survival (Table 4).
A score for each factor with a significant effect by multivariate analysis for RFS was assigned, with 1 point if the factor was present and 0 points if not. Accordingly, there were 85 patients (42%) with a score of 0, 97 patients (48%) with a score of 1, and 22 patients (11%) with a score of 2. Three- and 5-year RFS rates were 73% and 49%, respectively, for patients with a score of 0, 42% and 20%, respectively, for patients with a score of 1, and 16% and 8%, respectively, for patients with a score of 2. Recurrence-free survival was significantly lower for patients with a score of 2 compared with patients with scores of 0 and 1 (P < .001 and P = .03, respectively). Recurrence-free survival was also lower for patients with a score of 0 compared with patients with a score of 1 (P = .001) (Figure 2).
We examined factors affecting RFS in a series of consecutive hepatectomies for HCC performed in cirrhotic patients during a 9-year period at 2 tertiary Italian centers. Patient evaluation, postoperative follow-up, surgical techniques, and early results were comparable between centers.11,17 From our analysis, the 2 centers also had similar RFS and a comparable proportion of early and late recurrence rates. Thus, the clinical setting was homogeneous.
We selected cirrhotic patients, eliminating the bias of different tendencies to tumor recurrence due to variable degrees of liver fibrosis,24 and considered single tumors 5 cm or smaller, targeting a population that fulfilled the criteria for liver transplantation.19
Lower RFS rates in HBV-positive patients compared with HCV-positive patients and in HBV-negative and HCV-negative patients were observed. Patients who were both HBV-positive and HCV positive also showed a tendency for a low RFS, but this result could be affected by the low number of patients in group 4.
Patients who were HBsAg negative and HBcAb positive were considered to be HBV positive because HBV infection, although occult, was the only recognized cause of cirrhosis in group 1. On the other hand, taking into consideration group 4, there is growing evidence that occult HBV infection can sustain tumorigenesis and is associated with the development of HCC in patients with chronic HCV infection in the absence of apparent viral replication.25- 27
Infection with HBV and poor tumor differentiation were independently correlated with lower RFS, whereas the factors of liver function, portal hypertension, and tumor behavior predicted overall survival. However, the adoption of multimodal treatment of recurrences, including additional resection and transplantation, can strongly affect patient survival, which is therefore a less reliable outcome variable.
Although the prognostic factors emerging from our analysis have repeatedly been reported as affecting tumor recurrence and survival in previous studies,4- 11 to our knowledge this is the first report showing a strong negative effect on RFS of HBV infection in a Western countries series. Most studies found a higher tendency for earlier recurrence in HCV-infected vs HBV-infected patients.14,15 Furthermore, the prognosis of HBV-positive patients with HCC smaller than 5 cm was comparable if arising in cirrhotic or noncirrhotic patients.7 Together, these studies suggest that liver resection for HCC associated with HBV infection is generally associated with more favorable outcomes compared with other causes of HCC.
Some recent reports from Eastern areas have shown an adverse effect of HBeAg-positive status compared with HBeAg-negative status in terms of HCC recurrence and survival after hepatectomy, probably because of a greater predisposition to multiple carcinogenesis owing to a more active viral replication in the former group.13,16
Our study differed from the study by Sun et al13 because we excluded noncirrhotic patients and those with multiple nodules, thus selecting a more homogeneous population. In addition, although 53% of HBV-infected patients were positive for HBV DNA, only 5% were positive for HBeAg, which put HBV-positive patients at a theoretically lower risk of recurrence compared with the HBeAg-positive populations analyzed in the previous studies.13,16
The preoperative mean levels of transaminases (another index of viral replication) in group 1 were similar to those in group 2, meaning that the hepatitis activity was probably comparable. Even more, the proportion of patients with alanine aminotransferase levels of more than 2-fold normal levels was significantly higher in group 2 vs group 1. Group 4 (coinfected patients) had the worst indexes of hepatic function and the highest levels of transaminases, which might account for the low survival rates recorded in this group. However, transaminase levels were not determinant factors for HCC recurrence in our study population.
When we excluded a lower proportion of patients with complete tumor capsule in group 1 vs group 4, no significant differences in tumor-related factors emerged between patients with different types of viral infection.
Based on almost equivalent tumor aggressiveness and nontumorous microenvironment in subjects with isolate HBV or HCV, it is conceivable that HBV infection may predict a higher tendency for tumor recurrence independent of the necroinflammatory hepatitis activity, even in patients with occult infection. In this light, it has been suggested that the mechanisms by which overt HBV might induce tumor formation are probably conserved in cases of occult infection.25- 27
Our results seem to reinforce the need for antiviral treatment with nucleotide or nucleoside analogues or with interferon in HBV-infected patients with cirrhosis.28,29 However, several points should be further discussed. First, the use of lamivudine in one-fourth of our patients did not seem to prevent or delay tumor recurrence. Second, the development of viral resistance to antiviral drugs may represent a frequent event that should be avoided, particularly in subjects with potentially cured HCC. Third, the use of nucleotide or nucleoside analogues does not necessarily mean complete viral eradication, and the carcinogenetic potential of HBV DNA integration may persist.25 Finally, although promising,30 the treatment with interferon alfa may not be tolerated in cirrhotic patients.31
In the present series, the concomitance of 1 or 2 risk factors led to very low 5-year RFS rates. In particular, one may question whether hepatectomy is the best therapeutic choice for HBV-positive patients with poorly differentiated tumors. Considering that liver transplantation (vs resection) still offers better results for aggressive HCCs,32 from our analysis, this concept is even stronger for HBV-infected patients. We believe that hepatectomy should be indicated in HBV-infected cirrhotic patients with invasive HCC and that early initiation of adjuvant antiviral treatment and possibly a policy of bridging to liver transplantation should be considered.
Correspondence: Matteo Cescon, MD, Unità Operativa Chirurgia Trapianti di Fegato e Multiorgano, Padiglione 25, Policlinico Sant’Orsola-Malpighi, Via Massarenti, 9, 40138 Bologna, Italy (firstname.lastname@example.org).
Accepted for Publication: July 22, 2008.
Author Contributions: Dr Cescon had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Cescon, Grazi, and Andreone. Acquisition of data: Cescon, Cucchetti, Ferrero, Viganò, and Zanello. Analysis and interpretation of data: Cescon, Cucchetti, Grazi, Ferrero, Ercolani, Ravaioli, Zanello, Andreone, Capussotti, and Pinna. Drafting of the manuscript: Cescon. Critical revision of the manuscript for important intellectual content: Cucchetti, Grazi, Ferrero, Viganò, Ercolani, Ravaioli, Zanello, Andreone, Capussotti, and Pinna. Statistical analysis: Cescon and Cucchetti. Study supervision: Cucchetti, Grazi, Ferrero, Viganò, Ercolani, Ravaioli, Zanello, Andreone, Capussotti, and Pinna.
Financial Disclosure: None reported.