Tel.: 886-7-7317123 ext. Fax: +886-7-7322402
Infectious Causes of Cancer
Impact of diabetes mellitus on incidence of hepatocellular carcinoma in chronic hepatitis C patients treated with interferon-based antiviral therapy
Article first published online: 28 JUL 2010
Copyright © 2010 UICC
International Journal of Cancer
Volume 128, Issue 10, pages 2344–2352, 15 May 2011
How to Cite
Hung, C.-H., Lee, C.-M., Wang, J.-H., Hu, T.-H., Chen, C.-H., Lin, C.-Y. and Lu, S.-N. (2011), Impact of diabetes mellitus on incidence of hepatocellular carcinoma in chronic hepatitis C patients treated with interferon-based antiviral therapy. Int. J. Cancer, 128: 2344–2352. doi: 10.1002/ijc.25585
- Issue published online: 25 MAR 2011
- Article first published online: 28 JUL 2010
- Accepted manuscript online: 28 JUL 2010 12:00AM EST
- Manuscript Accepted: 13 JUL 2010
- Manuscript Received: 30 MAR 2010
- Chang Gung Memorial Hospital, Taiwan. Grant Number: CMRPG880241
- hepatitis C virus;
- diabetes mellitus;
- hepatocellular carcinoma;
There is strong evidence linking chronic hepatitis C virus (HCV) infection and Type 2 diabetes mellitus (DM). Recent studies have suggested that DM is associated with increased risk of developing hepatocellular carcinoma (HCC). The aim of our cohort study was to assess whether DM influence the incidence of HCC in chronic hepatitis C patients treated with interferon (IFN)-based antiviral therapy. A total of 1,470 chronic hepatitis C patients treated with IFN or pegylated-IFN plus ribavirin therapy were enrolled. Of them, 253 (17%) patients had DM at entry. Evaluation of HCC incidence was performed by Kaplan–Meier method and Cox proportional hazards analysis. Patients with baseline DM were significantly older and had higher body mass index, serum transaminase levels and fibrosis scores and lower platelet counts compared to non-DM subjects. Sustained virological response (SVR) was achieved in 160 (63%) of DM and 867 (71%) of non-DM patients (p = 0.008). During a median follow-up period of 4.3 years, HCC developed in 21 (8.3%) of DM and 66 (5.4%) of non-DM patients (p = 0.017). However, DM was not an independent covariate by Cox proportional hazards analysis. In a subgroup analysis, DM (hazard ratio, 4.32; 95% confidence interval, 1.23–15.25; p = 0.023) was an independent predictor of HCC in the SVR patients without baseline cirrhosis, despite a low HCC incidence. In conclusion, DM has a selective impact on HCC development among chronic hepatitis C patients after IFN-based therapy. DM may increase the HCC risk in chronic hepatitis C without cirrhosis after eradication of HCV.
Hepatitis C virus (HCV) has been identified as the major causative agent of chronic liver disease and greatly contributes to the etiology of cirrhosis and hepatocellular carcinoma (HCC).1–3 Prevention of complications associated with chronic HCV infection is one of the most important issues in public health worldwide. Interferon (IFN)-based antiviral therapy has been extensively adopted for treating chronic HCV infection.4–10 Over the past years, accumulating evidence has indicated that IFN or IFN plus ribavirin therapy can reduce HCC development and prolong survival in patients with chronic hepatitis C, especially in those who have achieved a sustained virological response (SVR).11–14 Further improvement in prognosis may be expected, because the current standard therapy of combined pegylated-IFN and ribavirin shows higher efficiency than IFN or IFN plus ribavirin.8–10
Clinical and experimental data suggest that chronic HCV infection is associated with increased risk of developing insulin resistance and its major late feature, Type 2 diabetes mellitus (DM).15–19 This evidence is in keeping either with a direct defect of insulin sensitivity and glucose disposal related to cirrhosis15–17 or with HCV-related mechanisms of insulin resistance.18, 19 Recently, DM has been recognized as a cofactor that can modify the course of chronic HCV infection. Previous studies have shown that DM not only accelerates the histological progression of chronic hepatitis C but also impairs the treatment response to IFN-based therapy.17, 20–22 In addition, it has also been suggested that DM is associated with a higher risk of HCC development in patients with chronic hepatitis C compared to hepatitis B virus (HBV)-infected subjects or those without HBV and HCV infections.23–26 However, this finding has not always been supported by other reports.27, 28
To our knowledge, the impact of DM on long-term prognosis of IFN-based therapy for chronic HCV infection has not been well established. We therefore attempted to undertake a large cohort study to assess the impact of DM on the incidence of HCC and overall survival among chronic hepatitis C patients treated with IFN or pegylated-IFN plus ribavirin therapy.
Material and Methods
From January 1999 to December 2007, 1,672 consecutive patients with chronic HCV infection who were treated with IFN or pegylated-IFN plus ribavirin therapy in a single center were enrolled. All patients had positive HCV antibody, detectable HCV RNA in serum and elevated alanine aminotransferase (ALT) levels for more than 6 months before enrollment. The exclusion criteria included decompensated liver disease, hepatitis other than hepatitis C (hepatitis B, autoimmune hepatitis and alcohol abuse) and major contraindications to IFN or ribavirin therapy. Alcohol abuse was defined as a high consumption of alcohol more than 50 g/day.29 Patients in whom HCC developed before treatment (n = 90) or within 1 year after the start of therapy (n = 21) were excluded. Also, patients with less than 12 months of follow-up (n = 61) or undetermined response (n = 30) were excluded and none did not have HCC. Thus, a total of 1,470 patients were further analyzed.
The diagnosis of Type 2 DM was based on the American Diabetes Association revised criteria, using a value of fasting blood glucose ≥126 mg/dl on at least two occasions,30 nonfasting blood glucose ≥200 mg/dl or ongoing treatment with hypoglycemic agents. Pathologic diagnoses were performed by percutaneous liver biopsies (n = 1299) within 6 months before treatment. Liver biopsy specimens were evaluated according to a modified Knodell histology index; fibrosis score 4 was defined as cirrhosis.31 Clinical diagnosis of cirrhosis was based on the repeated ultrasound findings suggestive of cirrhosis at least twice 3 months apart,32 supplemented with clinical criteria, such as thrombocytopenia or other signs of portal hypertension. Our study was approved by the institutional review board of the hospital, and informed consent was obtained from each patient before treatment.
Treatments and follow-up
Patients were treated with combination therapy consisting of either pegylated-IFN alfa-2a (Pegasys, F. Hoffmann-La Roche, Basel, Switzerland; 180 μg/week subcutaneously; n = 391), pegylated-IFN alfa-2b (Peg-Intron, Schering-Plough, Kenilworth, NJ; 1–1.5 μg/kg/week subcutaneously; n = 495) or IFN alfa-2b (Intron-A, Schering-Plough, Kenilworth, NJ; 3 or 5 million units subcutaneously thrice weekly; n = 584) and oral ribavirin (Rebetol, Schering Plough, Kenilworth, NJ) daily for 12–48 weeks. Ribavirin was given at a total daily dose of 1,000 mg for patients who weighed 75 kg or less and 1,200 mg for patients who weighed more than 75 kg. Ribavirin dosing was modified based on the drop of hemoglobin. Serum HCV RNA was assessed at the end of treatment and 24 weeks after completion of therapy. SVR was defined on the basis of the absence of serum HCV RNA at week 24 post-treatment. All other cases were classified as non-SVR.
After week 24 post-treatment, all patients were followed up every 3–6 months or more frequently as required.33 Routine follow-up studies included clinical assessment, conventional biochemical tests and HCC screening using serum α-fetoprotein and ultrasonography. A new space-occupying lesion detected or suspected at the time of ultrasonography was further examined with computed tomography, selective hepatic angiography or fine needle aspiration biopsy. The diagnosis of HCC was compatible with the guidelines of the American Association for the Study of Liver Disease in 2005,34 which was based on either the histopathological findings in tumor tissue, one typical HCC feature on a dynamic image and/or α-fetoprotein >200 ng/ml if the nodule was larger than 2 cm in a cirrhotic liver, or two typical HCC features of dynamic images if the nodule was between 1 and 2 cm in a cirrhotic liver.
The starting date of follow-up of each patient began at the initiation of antiviral treatment. The end of follow-up was the date of diagnosis of HCC, date of death or the closing date of the study, May 31, 2009. Among 98 patients who did not return to outpatient clinics for more than 1 year before the closing date, the patients' national citizen identification numbers were correlated with the national mortality datasets through the Statistics Office, Department of Health, Taiwan. The national mortality datasets were based on data in the Certificate of Death, which contained time, place and cause of death and details of the person issuing the document. The underlying cause of death was classified according to the International Classification of Disease (ICD) as HCC for ICD-9 code 155, as chronic hepatitis and liver cirrhosis for ICD-9 code 571 and as DM Type 2 for ICD-9 code 250.
Antibody to HCV was assessed using third generation enzyme-linked immunosorbent array (Ax SYM HCV 3.0, Abbott Laboratories, Chicago, IL). Qualitative detection of HCV RNA was performed by a standardized qualitative reverse transcription-polymerase chain reaction (RT-PCR) assay (Amplicor™, Roche Diagnostics, Branchburg, NJ), using biotinylated primers for the 5′ noncoding region. The lowest detection limit of this assay was 100 copies/ml (50 IU/ml). Serum HCV RNA levels were determined by a branched-DNA (b-DNA) signal amplification assay (VERSANT HCV RNA 3.0. Assay, Bayer Diagnostics, Emeryville, CA). This assay was a sandwich nucleic acid hybridization procedure with a detectable limit at 3,400 copies/ml. Genotyping of HCV was performed by reverse hybridization assay (Inno-LiPA™ HCV II; Innogenetics N.V., Gent, Belgium) using the HCV-Amplicor products.
Data are presented as median (interquartile range, IQR). Comparisons of differences in categorical data between groups were performed using the chi-square test. Distributions of continuous variables were analyzed by the t-test or Mann–Whitney U test for the two groups, where appropriate. Kaplan–Meier curves were generated for development of HCC and patient survival. The prognostic relevance was evaluated by univariate analysis with the log-rank test and by multivariate Cox proportional hazards regression analysis. All analyses were carried out using SPSS software version 15.0 (SPSS, Chicago, IL). All reported p values are two-sided, and p values less than 0.05 are considered statistically significant.
Missing data is an inevitable problem when conducting such hospital-based analysis based on reviews of patients' charts. Most statistical procedures, such as SPSS, exclude observations with any missing variables from the multivariate analysis, as in Cox proportional hazards model. However, working only with complete cases throws out a great deal of potentially useful data and also ignores possible systematic differences between complete and incomplete cases. Therefore, missing data were considered as one isolated group and further analyzed.
Treatment outcome of IFN-based therapy was modeled as a time-dependent covariate to represent the ability of patients undergoing more than one treatment course to change their status from nonresponders to responders in sequential courses. As the definition of SVR is undetectable serum HCV-RNA at week 24 post-treatment, we used this time point as time 0 for classifying response versus nonresponse.26
Comparison of clinical and virological characteristics between patients with and without baseline DM
Patient characteristics are shown in Table 1. There were 766 men and 704 women, 16–78 years old, with a median age of 53 years at enrollment. Pre-existing cirrhosis was diagnosed by biopsy histology (n = 365) or clinical criteria (n = 44). A total of 1,027 patients (70%) achieved a SVR. They included 385 (66%) of 584 patients treated with IFN plus ribavirin and 642 (72%) of 886 treated with pegylated-IFN plus ribavirin therapy, including 62 (37%) of 166 patients who received a second course of therapy.
As shown in Table 1, 253 (17%) patients were diabetic at entry, but none was considered as having Type 1 DM on the basis of their medical files. At baseline, 199 (79%) patients received oral hypoglycemic agents, whereas 25 (10%) received insulin therapy, and 29 (11%) had no regular control. Patients with DM were significantly older (p < 0.001) and had higher BMI (p < 0.001), serum aspartate aminotransferase (AST; p < 0.001), ALT levels (p = 0.004), α-fetoprotein levels (p < 0.001) and fibrosis scores (p < 0.001) and lower platelet counts (p < 0.001) compared to nondiabetic subjects, but there were no differences in virological features (genotype and viral load) between these two groups. Overall, diabetic patients achieved a lower SVR rate (63%) than that in nondiabetic subjects (71%; p = 0.008). Further multivariate analysis showed that DM [odds ratio, 1.45; 95% confidence interval (CI), 1.04–2.01; p = 0.029] was an independent factor associated with SVR (Supporting Information Table 1).
Impact of baseline DM on the cumulative incidence of HCC development
The median follow-up time was 4.4 years (IQR, 2.6–6.7). Patients with baseline cirrhosis have a higher median number of visits after IFN therapy [9 (IQR, 5–15) vs. 5 (IQR, 3–9), p < 0.001] and a shorter median follow-up time between visits [4.2 months (IQR, 3.5–5.2) versus 6.0 months (IQR, 5.1–7.0), p < 0.001] than those without.
A total of 87 patients developed HCC after a median follow-up of 4.3 years (IQR, 2.5–6.6). As shown in Figure 1a, cumulative incidence of HCC development in patients with baseline DM was 1.8% at year 2, 8.2% at year 4, 11.7% at year 6 and 17.4% at year 8; whereas the incidence in those without DM was 1.2% at year 2, 3.9% at year 4, 7.6% at year 6 and 10.3% at year 8 (p = 0.017). In addition to DM, Kaplan–Meier method showed that old age (≥60 years) at entry (p < 0.001), non-SVR (p < 0.001), baseline cirrhosis (p < 0.001), genotype-1 (p < 0.001), low platelet count (<15 × 104/μl; p < 0.001), high α-fetoprotein (≥20 ng/ml; p < 0.001) and high BMI (≥25; p = 0.028) were associated with HCC development among chronic hepatitis C patients treated with IFN-based antiviral therapy (Table 2).
Based on multivariate Cox proportional hazards model, age (hazard ratio, 1.07; 95% CI, 1.05–1.10; p < 0.001), non-SVR (hazard ratio, 4.15; 95% CI, 2.43–7.09; p < 0.001), baseline cirrhosis (hazard ratio, 3.87; 95% CI, 2.25–6.67; p < 0.001), pretreatment platelet count < 15 × 104/μl (hazard ratio, 2.93; 95% CI, 1.58–5.44; p = 0.001), male gender (hazard ratio, 2.19; 95% CI, 1.40–3.42; p = 0.001) and α-fetoprotein ≥ 20 ng/ml (hazard ratio, 2.14; 95% CI, 1.30–3.50; p = 0.003) were independent risk factors for HCC occurrence (Table 3). In contrast, baseline DM was not an independent risk factor for HCC after adjusting other factors in the multivariate analysis.
Subgroup analysis of HCC development according to treatment response and baseline cirrhosis
Apart from the significant continuous variable (age), treatment response and baseline cirrhosis were the two most important predictors of HCC development. We therefore hypothesized that DM might have a selective impact on HCC when the patients were classified into four groups according to SVR and the presence of baseline cirrhosis. As shown in Figure 1b, the incidence of HCC development was the highest in non-SVR patients with baseline cirrhosis and the lowest in SVR patients without baseline cirrhosis (p < 0.001). In the subgroup analysis, DM was a significant risk factor for the development of HCC only in the patients with SVR and nonbaseline cirrhosis among four groups (Figs. 1c–1f). As shown in Supporting Information Table 2, HCC was developed in 5 (4.9%) of 102 DM and 5 (0.7%) of 687 non-DM patients (p = 0.005) in this subgroup. Seven of the ten HCC patients had still not reached established cirrhosis (histology, n = 2 and clinical criteria, n = 8) at the time of HCC occurrence. As all HCC cases were male in this group, multivariate Cox proportional hazards analysis was performed in male patients, showing that age (hazard ratio, 1.14; 95% CI, 1.04–1.25; p = 0.004), baseline DM (hazard ratio, 4.32; 95% CI, 1.23–15.25; p = 0.023), pretreatment platelet count < 15 × 104/μl (hazard ratio, 5.87; 95% CI, 1.44–23.95; p = 0.013) and high BMI (≥25; hazard ratio, 4.06; 95% CI, 1.02–17.28; p = 0.047) were independent factors.
Impact of baseline DM on overall survival
A total of 57 patients died during the follow-up period; 35 of them suffered from liver-related death: 25 died of HCC and 10 of complications associated with cirrhosis. Of the 22 patients with nonliver-related death, seven died of severe infection, two of cardiovascular disease, two of DM complication, six of malignancies other than HCC and five of major trauma. As shown in Figure 2, patients with baseline DM had significantly higher overall mortality and liver or nonliver-related death compared to non-DM subjects (p < 0.001, p = 0.002, p < 0.001, respectively). Kaplan–Meier method also showed that old age (≥60 years) at entry (p < 0.001), baseline cirrhosis (p < 0.001), genotype-1 (p = 0.005), non-SVR (p < 0.001), low platelet count (<15 × 104/μl; p < 0.001) and high α-fetoprotein (≥20 ng/ml; p = 0.001) were associated with overall mortality (Supporting Information Table 3).
By multivariate Cox proportional hazards model, age (hazard ratio, 1.04; 95% CI, 1.01–1.07; p = 0.018), baseline cirrhosis (hazard ratio, 2.92; 95% CI, 1.52–5.63; p = 0.001), baseline DM (hazard ratio, 2.49; 95% CI, 1.42–4.37; p = 0.001) and non-SVR (hazard ratio, 1.99; 95% CI, 1.08–3.68; p = 0.027) were independent risk factors associated with overall mortality. As regards to liver-related death, baseline cirrhosis (hazard ratio, 4.07; 95% CI, 1.67–9.91; p = 0.002) and non-SVR (hazard ratio, 2.40; 95% CI, 1.08–5.38; p = 0.032) were independent factors (Table 4)
In our large cohort study, we assessed the impact of DM on incidence of HCC and overall survival among chronic hepatitis C patients who were followed up for a median period of 4.4 years after IFN or pegylated-IFN plus ribavirin therapy. By the Kaplan–Meier analysis, we demonstrated that patients with baseline DM had a higher risk of HCC development than non-DM patients after IFN-based therapy. However, this significance disappeared after statistical adjustment of other factors in the multivariate analysis. The main possible explanation is that diabetic patients were significantly older in age and had higher BMI and fibrosis scores and lower platelet counts compared to non-DM patients. Also, diabetic patients achieved a lower SVR rate (63%) than that in nondiabetic subjects (71%). Thus, the influence of DM may be masked, and the predictive power is impaired in the presence of other stronger risk factors.
Our result differs from that in a recent study enrolling a Western population, which showed that DM increased the risk of developing HCC in patients with HCV-related advanced cirrhosis treated with IFN-based regimen.26 These conflicting data might be associated with the different cohorts of patient collection and treatment regimens as well as different diagnostic criteria of DM. In contrast to our study, Veldt et al. defined DM as either by elevated fasting serum glucose (6.1 mmol/l) or a positive glucose tolerance test. Further studies are necessary to clarify the relationship between HCC development and impaired glucose intolerance without overt DM. On the other hand, in our cohort, the majority (60%) of patients received combined pegylated-IFN plus ribavirin therapy, which provided a higher SVR rate and might increase the benefit in HCC prevention, thus probably affecting the impact of DM on the prognosis of antiviral therapy. The superior response in our patients is probably due to the lower percentage (49%) of genotype 1 infection and the higher average doses of IFN and ribavirin on the lower mean body weight in Taiwanese patients when compared to Caucasian patients.35 In contrast to this report,26 our data suggest that successful HCV eradication can significantly reduce the risk of HCC occurrence in patients with chronic HCV infection, even for the subjects with underlying cirrhosis.
In our study, considering the greater impact of treatment response and cirrhosis on HCC development, we therefore analyzed the effect of DM on the subgroups, which were classified according to SVR and the presence of baseline cirrhosis. Interestingly, DM was an independent predictor of HCC in the SVR patients without cirrhosis, despite the lowest HCC incidence among four groups. In particular, in addition to other predictors of HCC such as age, male gender and lower platelet count, metabolic factors including baseline DM and higher BMI became significant in this specialized subgroup. Although the pathogenic mechanism remains unclear, this can be hypothesized that the effect of DM or obesity may have been more pronounced when the two stronger risk factors are excluded. Our finding was comparable with that in a previous Japanese study, which showed that DM was associated with hepatocarcinogenesis in 279 patients with chronic hepatitis C patients without cirrhosis.36 Recently, Paradis et al. have demonstrated that the development of HCC in patients with features of metabolic syndrome such as DM or obesity as the risk factors has distinct characteristics and occurs, in most cases, in the absence of significant cirrhosis.37 This finding reinforces our observation suggesting that DM and obesity could predispose to HCC formation in chronic hepatitis C without cirrhosis after eradication of HCV by IFN-based therapy.
On the other hand, it has been reported that DM is a negative predictor of survival in patients with chronic liver disease, HCC patients undergoing resection and end-stage cirrhotic patients undergoing liver transplantation.38–41 Recent studies have suggested that the cause of death in cirrhotic patients with DM is mainly related to hepatic failure and not to complications of DM.42, 43 In our study, we identified that DM was an independent risk factor for overall survival among chronic hepatitis C patients after IFN-based therapy. Apart from liver complication, patients with DM had a higher mortality unrelated to HCV than non-DM subjects by the Kaplan–Meier analysis. Moreover, the effect of DM was only statistically significant in the separate analysis of nonliver-related death by the stepwise multivariate Cox proportional hazards analysis, possibly due to the confounding effects in liver-related death. These results concurred with previous observations,39–43 suggesting that optimal glycemic control might be necessary for chronic hepatitis C patients with DM to provide a better long-term outcome.
There are some limitations in our study. First, this is not a truly prospective study. However, we believed that the bias was small because patients were followed by the same physicians throughout the course of disease, with clinical and laboratory assessment and HCC screening using serum α-fetoprotein and ultrasonography every 3–6 months. Second, the cohort of patients was enrolled in an Asian population, with a low prevalence of obesity. Whether DM has a similar impact on the prognosis of IFN-based therapy in areas with high prevalence of DM or obesity remains to be clarified. Third, the prevalence and specific features of the metabolic syndrome such as blood pressure and waist circumference were not obtained. It will be interesting to determine the association of metabolic syndrome with HCC development after IFN-based therapy for chronic HCV infection in further studies.
In conclusion, DM has a selective prognostic impact on HCC development among HCV-infected patients after IFN or pegylated-IFN plus ribavirin therapy. DM may play an important role of HCC development in patients who do not have otherwise stronger risk factors such as non-SVR and underlying cirrhosis, despite a low HCC incidence. In such patients with DM, regular HCC surveillance should be recommended.
- 16Hepatitis C viremia increases the association with type 2 diabetes mellitus in a hepatitis B and C endemic area: an epidemiological link with virological implication. Am J Gastroenterol 2007; 102: 1237–43., , , , , , , , , , , , et al.Direct Link:
- 22Insulin resistance and diabetes increase fibrosis in the liver of patients with genotype 1 HCV infection. Am J Gastroenterol 2008; 103: 1136–44., , , , , , , , , , .Direct Link:
Additional Supporting Information may be found in the online version of this article.
|IJC_25585_sm_suptable.doc||84K||Supporting Table 1|
Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.