Abstract
Background/Aim: Several studies have reported that DM is closely associated with an increased incidence of hepatocellular carcinoma (HCC). To clarify the effects of diabetes mellitus (DM) and antidiabetic medications on the prognosis of patients with non-B non-C (NBNC) HCC following curative initial hepatectomy. Patients and Methods: HCC patients (n=156) were divided into three groups according to the presence or absence of chronic viral hepatitis: hepatitis B virus (HBV) group, hepatitis C virus (HCV) group, and NBNC group. The clinical characteristics and survival outcomes were compared. In the NBNC group, univariate and multivariate analyses were conducted to determine prognostic factors. Results: The NBNC group had a higher incidence of DM, ethanol intake, and large nodules than the other groups. Disease-free survival (DFS) was significantly worse in the NBNC group than in the HBV group. In the NBNC group, insulin treatment was an independent prognostic factor for DFS and overall survival (OS). Conclusion: Medications for DM that affect insulin resistance might be appropriate prognostic factors for NBNC-HCC.
Hepatocellular carcinoma (HCC) is the fifth most common cancer and the fourth leading cause of death worldwide (1). HCC mainly occurs as a result of viral hepatitis in hepatitis B virus (HBV) or hepatitis C virus (HCV), which accounts for approximately 75% of patients in Japan (2). In recent years, the incidence of virus-related HCC decreased due to some healthcare projects to prevent hepatitis B (HBV) and hepatitis C virus (HCV) infections (3). Antiviral treatment, such as interferons, direct acting antivirals (DAAs), and nucleoside analogues (NAs) have also reduced the risk of HCC (4-6). However, the incidence of non-B non-C (NBNC)-HCC without either hepatitis B surface antigen (HBsAg) or anti-hepatitis C antibody (HCVAb) has been increasing (3).
The etiology of the NBNC-HCC could be: i) alcohol use, ii) nonalcoholic fatty liver disease (NAFLD), iii) congestive disease, iv) metabolic disease (Wilson’s disease, hemochromatosis, etc.), v) autoimmune liver diseases (primary biliary cirrhosis, autoimmune hepatitis, primary sclerosing cirrhosis), or vi) unknown (7). NAFLD is the hepatic phenotype of metabolic syndrome, which is deeply associated with obesity, diabetes mellitus (DM), hyperlipidemia, and hypertension (8). A small number of NAFLD cases cause steatohepatitis and progress to nonalcoholic steatohepatitis (NASH) (9), while metabolic syndrome is considered to be pathophysiologically insulin-resistant (10).
Several studies have reported that DM is closely associated with an increased incidence of cancer; in particular, the relationship between DM and an increased risk of incident HCC is clinically reliable (11). Among diabetic patients in Japan, approximately 8.7% die as result of HCC, which is the first leading cause of death in malignant diseases (12). DM, obesity, and NAFLD are strongly associated with insulin resistance (13). In the presence of insulin resistance, insulin concentrations in blood rise, inducing insulin-like growth factor-1 (IGF-1) production (14). This hormone is known to stimulate cellular proliferation and inhibit apoptosis within the liver. Insulin and IGF-1 stimulate insulin receptor substrate-1 (IRS-1), which is the major isoform that transduces insulin signaling in the liver and is overexpressed in human HCC (15). Overexpression of IRS-1 leads to activation of phosphatidylinositol-3 kinase (PI3K) signal transduction cascade and enhances tumor progression of HCC (16).
These findings prompted us to evaluate the clinical characteristics and prognosis of surgical patients with NBNC-HCC compared to those with HBV-related HCC (HBV-HCC) and HCV-related HCC (HCV-HCC) in order to investigate the clinical significance of diabetic treatment on the prognosis of patients with NBNC-HCC.
Patients and Methods
Patients. From January 2007 to December 2014, 183 HCC patients, who had not received any other preoperative treatment for HCC, such as transcatheter arterial chemoembolization, radiofrequency ablation, and liver transplantation, underwent initial hepatic resection at our institution. Among these patients, those who underwent non-curative resection (n=4), had macroscopic vascular invasion (n=2), or had lymph node metastasis (n=3) were excluded from the study. HBV or HCV infection was diagnosed based on positive results of viral marker screening for HBsAg and HCVAb, respectively. The 175 patients were divided into three groups as follows: The HBV group (n=31) was defined as positive for HBsAg and negative for HCVAb, the HCV group (n=92) was defined as positive for HCVAb and negative for HBsAg, and the NBNC group (n=47) was defined as negative for both HBsAg and HCVAb. Patients with concomitant HBV and HCV infection (n=1), primary biliary cirrhosis (n=2), and fibrolamellar HCC (n=1) were excluded from the analysis. A total of 156 patients were included in the final analysis, 28 patients in the HBV group, 83 patients in the HCV group and 45 patients in the NBNC group. Fifteen patients that could not be followed up for more than 5 years were also excluded. The study protocol was approved by the Medical Ethics Committee of the Yamanashi University (approval code: 1720), and the study was performed in accordance with the provisions of the Declaration of Helsinki. The requirement for written informed consent from the patients was waived because of the retrospective design of the study.
The preoperative diagnosis of HCC was made by elevated HCC-specific tumor markers, such as alpha fetoprotein (AFP) and des gamma carboxyprothrombin (DCP), as well as early enhanced liver tumor of imaging studies. The imaging modalities used for HCC diagnosis were contrast-enhanced abdominal ultrasonography, dynamic computed tomography (CT), and gadolinium-ethoxybenzyl-diethylenetriamine pentaacetic acid-enhanced magnetic resonance imaging (EOB-MRI).
Surgical indication. Treatment for HCC was generally decided according to the consensus-based treatment algorithm for HCC, proposed by the Japan Society of Hepatology (JSH) (17). Basically, anatomical resection was attempted by considering liver function, the number, size and location of nodules, and the general condition of the patients; limited resection was performed when the liver reserve was poor and the nodules were small.
The nodule size and the width of the surgical margin were recorded before the specimens were fixed. The histologic grade of tumor differentiation, the degree of fibrosis in the background liver, and the presence or absence of vascular invasion were also assessed microscopically based on the classification system proposed by the Liver Study Group of Japan (LCSGJ) (18).
Postoperative follow-up. Following the initial operation, all patients were followed up regularly in outpatient clinics. Recurrence was checked every 1-2 months by monitoring the plasma levels of the HCC-specific tumor markers, and dynamic CT or EOB-MRI performed every 4 months. When intrahepatic recurrence was detected during the follow-up evaluations, for patients with a recurrence-free survival of ≥1 year, the treatment was decided according to the consensus-based treatment algorithm by the JSH (17). Overall survival (OS) was defined as the interval between the date of hepatectomy and the date of death. Disease-free survival (DFS) was defined as the interval between hepatectomy and the diagnosis of the first recurrence.
Statistical analysis. The differences in baseline characteristic were compared by Fisher’s exact test for categorical variables and Mann–Whitney U-test for continuous variables. In the survival analysis, the rates of OS and DFS were calculated by the Kaplan–Meier method and compared using the log-rank test. Multivariate analysis of the prognostic factors related to survival was performed using the Cox proportional hazard model. All statistical analyses were performed by R software (version 3.6.1). Statistical significance tests were two-tailed, and a p-value of <0.05 was considered significant.
The albumin–bilirubin (ALBI) grade was calculated according to the baseline serum albumin and bilirubin levels [ALBI=(log10 bilirubin μmol/l × 0.66) + (albumin g/l × −0.085)] (19). According to previous literature, ALBI grades 1, 2, and 3 were defined as <−2.60, −2.60 to −1.39, or >−1.39, respectively.
Results
Comparison of clinical characteristics between the three groups. The patients in the NBNC group had a significantly higher incidence of DM (HBV group, p=0.006; HCV group, p=0.007) and ethanol intake (>60 g/daily) (HBV group, p<0.001; HCV group, p=0.001), as well as significantly higher glycated hemoglobin (HbA1c) (HBV group, p=0.007; HCV group, p<0.001) and DCP values (HBV group, p=0.001; HCV group, p=0.016), compared to the patients in the HBV and HCV groups. However, differences in the patient backgrounds may also likely play a role in each comparison. In the NBNC group, the patients had significantly lower incidence of liver cirrhosis (HBV group, p=0.013; HCV group, p=0.009) and demonstrated larger tumor sizes (HBV group, p=0.003; HCV group, p<0.001), compared to the other two groups (Table I).
In prognostic analyses, the NBNC group demonstrated intermediate DFS compared to the other groups, while DFS was significantly poorer in the HCV group than in the HBV group (p=0.009). In contrast, the OS of the NBNC group was similar to that of the HCV group, which was significantly poorer than that of the HBV group (p=0.040) (Figure 1).
Prognostic impact of diabetic treatment in NBNC patients with DM. In the NBNC group, 23 patients had DM, of whom, 2 had received alimentary therapy without antidiabetic medications, 6 were treated with insulin, 11 with sulphonylurea (SU), 8 with alpha glucosidase inhibitor (aGI), 8 with dipeptidyl peptidase 4 (DPP4) inhibitors, 6 with metformin, and 3 patients with thiazolidine (Table II). Further prognostic analyses demonstrated that there were no significant differences in DFS between patients treated with insulin and those who were not, as well as between those treated with DPP4 inhibitors and those who were not (Figure 2a, 2c). The OS was significantly worse in patients treated with insulin than those who were not (p<0.001), while OS was significantly better in patients treated with DPP4 inhibitors than those who were not (p=0.048) (Figure 2b, 2d). However, the DFS was significantly worse in patients treated with metformin than those who were not (p=0043) (Figure 2e). There were no significant differences in DFS and OS between the patients treated with SU and aGI and those who were not (data not shown).
Prognostic impact of DM status in the NBNC group. Because insulin was given concomitantly in half of the patients treated with metformin, metformin was excluded from further prognostic analysis. The results of the univariate analyses of prognostic factors for DFS and OS in the NBNC group are summarized in Table III. Insulin treatment (p=0.004), serum aspartate aminotransferase (AST) >30 IU/l (p=0.044), and multiple nodules (p=0.043) were correlated with significantly worse DFS. In addition, insulin treatment (p<0.001), body mass index (BMI) >25 kg/m2 (p=0.044), serum albumin >4.0 g/dl (p=0.018), and ALBI Grade 2 (p=0.007) were correlated with significantly worse OS. DM was not shown to be a prognostic factor of DFS and OS in univariate analysis.
The results of the multivariate analyses of prognostic factors for DFS and OS in the NBNC group are summarized in Table IV. Insulin treatment was selected as an independent predictor of adverse DFS in multivariate analysis (p=0.017), while insulin treatment (p<0.001), ALBI Grade 2 (p=0.001), and anti-hepatitis B core antibody (HBcAb) positive (p=0.036) were selected as independent predictors of adverse OS in multivariate analysis. Serum albumin was excluded from multivariate analysis since the ALBI score was calculated from serum albumin and serum bilirubin.
Prognostic impact of insulin treatment in the HBV and HCV groups. In the HBV group, there were 5 patients with DM, 4 of whom received insulin treatment before hepatic resection. In prognostic analyses, DFS in patients treated with insulin was significantly poorer than in those that were not treated with insulin (p=0.022). There was no significant difference in OS with or without insulin treatment (Figure 3a, 3b).
In the HCV group, there were 22 patients with DM, 10 of whom received insulin treatment before hepatic resection. In prognostic analyses, there was no significant difference in DFS and OS between patients with or without insulin treatment (Figure 3c, 3d).
Discussion
In our previous study on HCC from 1997 to 2006, the incidence of NBNC-HCC was 12.6% (20); however, in the current study it increased to 26.9%. Recently, a nationwide study in Japan has reported that the number and rate of patients with NBNC-HCC were increasing (3). This might be to be due to the decrease in the number of patients with chronic hepatitis B and C and the increase of metabolic syndrome. In 1998, the World Health Organization proposed diagnostic criteria for metabolic syndrome, several of which are highly associated with obesity, DM, hyperlipidemia, and hypertension (10).
Several studies have investigated the clinicopathological characteristics and prognosis of patients with NBNC-HCC following hepatectomy. The characteristic features of patients with NBNC-HCC are older age (similar to those with HCV-HCC), increased incidence of hypertension, DM, heavy alcohol consumption, and large tumors, and a reduced incidence of liver cirrhosis (21-24). In the current study, similar characteristics were revealed; in particular, the NBNC group patients had significantly lower serum albumin values than the HBV group patients and higher serum albumin values than the HCV group patients, despite a significantly lower incidence of liver cirrhosis in the NBNC group than the other groups. This discrepancy between levels of serum albumin and liver cirrhosis incidence might have occurred due to the fact that 67.9% of HBV patients maintained effective inhibition of HBV replication with NAs, and 21.7% of HCV patients achieved sustained viral response with interferon therapy, at the time of hepatectomy (data not shown). The antiviral treatment with NAs has been previously reported to improve liver function (25), which may have resulted in higher albumin values, compared to the other groups, regardless of the frequency of liver cirrhosis.
The prognosis of patients with NBNC-HCC compared to those with HBV- and HCV-HCC is controversial. Kondo et al. (21) reported that NBNC-HCC patients had better OS and DFS than HBV-HCC and HCV-HCC patients, while other researchers showed no significant difference in DFS and OS among the three groups (22-24). Although the majority of previous studies examined nearly 100 patients with NBNC-HCC, Utsunomiya et al. (26) revealed that the DFS and OS in the NBNC-HCC patients were significantly better than those in the HBV-HCC and HCV-HCC patients, using data from over one million patients in a nationwide survey. Since, in the current study, 67.9% of the HBV-HCC patients received NAs before liver resection, it was considered that the OS in the HBV group was better than that in the NBNC group. NA treatment in patients with HBV infection has been shown to increase the survival rate by contributing to the low incidence of HCC and improvement of liver function (6). In addition, the following should be considered; the etiologies of NBNC-HCC are numerous, there are patients with indistinct or unidentified etiologies (27), and NFALD and alcoholic liver disease have different patient characteristics (28). In this study, since half of the NBNC-HCC patients had habitual ethanol intake >60 g daily, the other half of patients may have had NAFLD, according to the diagnostic criteria for alcoholic liver disease by the Japanese Society for Biomedical Research on Alcohol (29).
Serum albumin and bilirubin, which are affected by liver reserve, serum AST, which expresses the severity of hepatitis, tumor number and AFP, which reflect tumor progression, have been reported as prognostic factors in the NBNC-HCC patients (22, 30, 31). In the current study, insulin treatment, ALBI Grade 2, and HBc Ab positive were shown as significant prognostic factors. The tumor-related factors were not selected in this study. Previous reports have reported that HCC with multiple tumors was independent prognostic factors for DFS (21, 22, 24, 31), however Matsuda et al. have reported a relatively good prognosis of multicentric HCC (32). Our preoperative evaluation including EOB-MRI detected multicentric HCCs, which might affect our results. HbA1c, which is an indicator of blood glucose regulation, was not selected as the prognostic factor; however, insulin treatment probably representing progression of insulin resistance was an independent prognostic factor of worse OS and DFS. Considering that insulin resistance has a key molecular role in hepatocarcinogenesis of NBNC-HCC (33), improvement of insulin resistance may be more important than blood glucose regulation. There are several reports related to the incidence of HCC and diabetic treatments focusing on insulin resistance. Improvement in insulin resistance by treatment with metformin and thiazolidine may be associated with a lower risk of HCC, whereas treatments with SU or insulin, which may promote insulin resistance, seem to be related to a higher risk of HCC (34). In contrast, it is considered that DPP4 inhibitors suppress inflammation and fibrosis independently of glucose metabolism, and may decrease NASH and HCC (35). Herein, we compared the prognosis of diabetic patients treated with different diabetic medications before liver resection. The results suggested that insulin treatment may worsen the OS, while DPP4 inhibitors may improve the OS. The patients treated with metformin had poorer prognosis with regards to DFS then those without, but the difference disappeared when the patients who received insulin and metformin were excluded. There were no significant differences in DFS and OS between the patients treated with SU and aGI and those who were not.
In observational epidemiologic study, Hanley AJ et al. (36) showed that AST and alanine aminotransferase (ALT) were inversely correlated with insulin sensitivity and independently predicted DM. They concluded that baseline elevations of AST and ALT may reflect NAFLD or related pathologies. Clinically, in the present study, insulin resistance had little impact on the prognosis of HBV-HCC and HCV-HCC, but it was found to be an important prognostic factor in the NBNC-HCC group. Therefore, one could say that elimination of insulin resistance might improve the prognosis of NBNC-HCC patients, as inhibition of HBV replication and sustained viral response for HCV with antiviral treatments improve prognosis of the HBV- and HCV-HCC patients, respectively (4-6). Taken together, lifestyle changes leading to improvements in insulin resistance are likely to be the most important factor in preventing the recurrence of HCC, as well as the conservation of liver function.
Antiviral treatment of HBV- and HCV-HCC reduced the incidence of HCC, preserved liver reserve, and improved prognosis. In order to improve the prognosis of NBNC-HCC, it is important to treat diabetic patients by considering the improvement in insulin resistance. Furthermore, since there is a limit to drug therapy alone to improve insulin resistance, the importance of exercise and alimentary therapy should be reinforced to patients.
The present study had several limitations. First, the retrospective design of the study introduces the possibility of selection bias. Second, it is a single-center study’s design with a small cohort, especially for the HBV patients. Finally, the NBNC-HCC has diverse etiologies, including alcohol use, NAFLD, or occult HBV infection. There were patients in the NBNC-HCC with no identified etiology. Therefore, a larger prospective study on the subgroup of patients with NBNC-HCC should be conducted. However, the current study presents interesting results on the effects of insulin treatment on NBNC-HCC prognosis after hepatectomy.
Conclusion
Although there are several etiologies of NBNC-HCC, the proportion of diabetic patients is large, and patients who receive insulin treatment have a significantly worse prognosis. Therefore, it is important to improve lifestyle and drug selection in order to improve insulin resistance.
Footnotes
Authors’ Contributions
Conception and design of study were performed by H. Amemiya and D. Ichikawa. Data acquisition was conducted by H. Amemiya, M. Matsuda, R. Saito and N. Hosomura. Analysis and/or interpretation of data were performed by H. Amemiya, M. Matsuda, and D. Ichikawa. The manuscript was drafted by H. Amemiya and D. Ichikawa. M. Matsuda, R. Saito, and D. Ichikawa critically revised the manuscript for important intellectual content. All Authors approved the final version of the manuscript.
Conflicts of Interest
All Authors disclose no conflicts of interest.
- Received September 14, 2020.
- Revision received December 9, 2020.
- Accepted December 15, 2020.
- Copyright© 2021, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.