| |
Glucose Abnormalities Are an Independent Risk Factor for Nonresponse to Antiviral Treatment in Chronic Hepatitis C
|
| |
| |
The American Journal of Gastroenterology
Volume 102 Issue 10 Page 2189-2195, October 2007
Albert Lecube, M.D.11Diabetes Research Unit, Endocrinology Division,
Cristina Hernandez, M.D.11Diabetes Research Unit, Endocrinology Division,
Rafael Simo, M.D.11Diabetes Research Unit, Endocrinology Division,
Juan Ignacio Esteban, M.D.22Liver Unit and Ciberehd Hospital Universitari Vall d'Hebron, Institut de Recerca Vall d'Hebron, Universitat Autonoma de Barcelona, Barcelona, Spain, and
Joan Genesca, M.D.22Liver Unit and Ciberehd Hospital Universitari Vall d'Hebron, Institut de Recerca Vall d'Hebron, Universitat Autonoma de Barcelona, Barcelona, Spain
1Diabetes Research Unit, Endocrinology Division, and 2Liver Unit and Ciberehd Hospital Universitari Vall d'Hebron, Institut de Recerca Vall d'Hebron, Universitat Autonoma de Barcelona, Barcelona, Spain
"In the present study we provide evidence that the presence of glucose abnormalities is an independent factor for nonresponse to antiviral treatment in chronic hepatitis C. Therefore, glucose abnormalities should be added to the host-related factors (i.e., age, male gender, alcohol use, presence of cirrhosis, Afro-American ethnicity, obesity, hepatic steatosis, and HOMA) that adversely influence the rate of SVR in HCV infected patients (5, 6). Since the prevalence of both IFG and type 2 diabetes reaches 30% among noncirrhotic HCV infected patients (3), it seems likely that our findings will be relevant not only in the evaluation of the response to antiviral therapy, but also in giving support to therapeutic strategies addressed to normalize blood glucose levels before antiviral therapy."
Abstract
OBJECTIVES: The influence of glucose abnormalities on the efficacy of antiviral treatment is unknown. This study investigated whether glucose abnormalities (impaired fasting glucose and type 2 diabetes) influence the response to antiviral therapy with interferon plus ribavirin in patients with chronic hepatitis C.
METHODS: A total of 178 treatment-naive patients with chronic hepatitis C treated with combination therapy were retrospectively studied. SVR was assessed after completing treatment. Fasting plasmatic glucose was measured prior to therapy.
RESULTS: Compared with nonresponders (N = 111), patients with SVR (N = 67) had lower plasma glucose (94.1 ± 12.7 vs 104.4 ± 25.8 mg/dL, P = 0.001) and a lower prevalence of glucose abnormalities (24.24% vs 44.14%, P = 0.012). The SVR rate was 45.13% in patients with normoglycemia (N = 113), 28.26% in patients with impaired fasting glucose (N = 46), and 15.78% in type 2 diabetic patients (N = 19) (P < 0.001). Multivariate logistic regression identified genotype 1 (OR 1.55, 95% CI 1.01-2.41, P = 0.05), ϒ-glutamyltranspeptidase level (OR 6.41, 95% CI 1.86-22.07, P = 0.003), and presence of glucose abnormalities (OR 2.33, 95% CI 1.04-5.20, P = 0.039) as being independently associated with the absence of an SVR. In addition, patients with glucose abnormalities (N = 65) showed a lower virological response rate when compared with a subgroup of normoglycemic patients (N = 65) matched for sex, age, and liver fibrosis (24.6% vs 44.6%, P = 0.001).
CONCLUSIONS: Glucose abnormalities are an independent predictor of poor virological response to combined therapy in hepatitis C virus infected patients.
INTRODUCTION
Hepatitis C virus (HCV) infection and type 2 diabetes mellitus are two common disorders with a high impact on worldwide health. A high prevalence of type 2 diabetes among HCV infected patients has been consistently reported (1-3), and there is growing evidence to support the concept that HCV infection is a risk factor for developing type 2 diabetes (4). The Third National Health and Nutrition Examination Survey (NHANES III) has shown that among persons of 40 yr of age or older, subjects with HCV infection were more than three times more likely to have type 2 diabetes than those without HCV infection (2). The specific mechanisms by which HCV leads to type 2 diabetes remains to be elucidated, but it seems that an increase of insulin resistance associated with overproduction of proinflammatory cytokines, enhanced steatosis, and liver fibrosis plays an essential role (4).
The current treatment for patients with chronic hepatitis C is the addition of ribavirin to interferon-based therapies for 24 to 48 wk. Unfortunately, a sustained virological response (SVR) is achieved in only 42-52% of treatment-naive patients, and the rest of them either show no response or experience a relapse when therapy is stopped (5). The mechanisms underlying the failure of interferon therapy are not well understood, but evidence indicates that in addition to viral factors, several host factors are also involved (6).
Insulin resistance, calculated by the homeostasis model assessment (HOMA), has been found to impair virological response to combined therapy in chronic hepatitis C patients (7, 8). However, since the HOMA model depends on insulinemia and the clearance of insulin is significantly impaired in advanced liver fibrosis (9), results using HOMA as an insulin resistance measurement in patients with chronic liver disease can be seriously flawed and it might be the reason for conflicting results (10). Therefore, the hepatic insulin extraction rate should be considered when HOMA is used for insulin resistance assessment in patients with chronic liver disease. However, these measurements are expensive and laborious and they are out of the routine management of these patients. Another way to determine whether insulin resistance impairs the response rate to antiviral therapies and further simplify its assessment could be the evaluation of other markers of the same biological phenomenon, as fasting glucose and ϒ-glutamyltranspeptidase (GGT). Because glucose abnormalities are the top of the iceberg of the insulin resistance state they can be used as a practical reflection of insulin resistance in patients with chronic liver disease. In addition, as occurs in other infectious processes (11), hyperglycemia itself could be involved in a poor response to antiviral therapy in chronic hepatitis C. However, there are no studies evaluating the influence of glucose levels in the response to antiviral treatment in chronic hepatitis C.
The aim of the present study was to investigate whether glucose abnormalities (impaired fasting glucose [IFG] and type 2 diabetes) could be helpful in identifying HCV infected patients at risk of being nonresponders to combination antiviral therapy.
METHODS
Study Design and Patient Selection
This was a retrospective study investigating the effect of glucose abnormalities on the efficacy of combination therapy in naive patients with chronic hepatitis C. A total of 326 white patients with chronic hepatitis C treated with combination therapy at the liver unit of our hospital between 1998 and 2002 were screened. Information was available from the database of the pharmacy department in which all patients treated with ribavirin are filed. Inclusion criteria included: age older than 18 yr, anti-HCV and HCV RNA positivity, HCV genotype determined by automated DNA sequencing, a liver biopsy performed within the previous 12 months to starting therapy, and no prior antiviral therapy. Eighty-five patients did not meet the inclusion criteria, the majority (N = 65) due to prior interferon therapy, leaving a total of 241 patients. Among the 241 patients who met the inclusion criteria, 46 were excluded because of the following reasons: alcohol consumption (>40 g/day for men and >20 g/day for women) or features of alcoholic disease in the liver biopsy (N = 7), presence of other underlying causes of chronic liver disease apart from HCV infection (N = 10), type 1 diabetes mellitus (N = 3), and anti-HIV positive (N = 26). Fibrosis and inflammation were assessed using the modified histologic activity index (HAI) scoring system described by Ishak et al. Steatosis was quantified as the percentage of hepatocytes that contained fat droplets: 0% (absent) to 100%. Because few patients had marked steatosis, liver steatosis was classified as present or absent.
Patients received antiviral treatment with thrice-weekly injections of three million units of interferon alfa-2b (Intron A, Schering-Plough, Kenilworth, NJ) plus ribavirin administered in a dose of 800-1,200 mg daily depending on body weight (Rebetrol, Schering-Plough). Duration of therapy was performed according to genotype: 6 months for genotypes 2 and 3, and 12 months for genotypes 1 and 4. All patients were followed every 6 months as outpatients by clinical examination and laboratory testing. Dose reduction was defined as any modification of any of the two treatments during therapy that represented less than 80% of the initial therapy. SVR was defined as an undetectable level of serum HCV RNA 6 or more months after completing therapy. According to the criteria recommended by the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus, diabetes was defined as fasting plasma glucose 126 mg/dL in two separate measurements and IFG was defined as baseline glucose levels between 100-125 mg/dL.
The study was consistent with the principles of the Declaration of Helsinki and it was approved by the hospital's human ethics committee.
Laboratory Assessments
After overnight fasting, blood samples were drawn for routine analyses including determination of glucose, alanine aminotransferase (ALT), aspartate aminotransferase (AST), and GGT. These parameters were measured by standard laboratory techniques used in clinical chemistry laboratories. All patients had positive anti-HCV measured by a second-generation commercial enzyme immunoassay (Abbott Laboratories, Chicago, IL), and positive HCV RNA in serum (Amplicor, Roche, Montclair, NJ).
HCV genotyping was performed by RT-PCR on a segment from the core region and by hybridization of this fragment with oligonucleotide specific probes according to the manufacturer's instruction (HCV genotyping, DNA Enzyme Immunoassay, Diasorin, Italy). The assay is designed to recognize the 1a, 1b, 2a, 2b, 3, 4, 5, and 6 HCV genotypes.
Statistical Analysis
Normal distribution of the variables was evaluated using the Kolmogorov-Smirnov test. Skewed variables were logarithmically (log) transformed. Comparisons between groups were performed using the Student's t-test for continuous variables and χ2 test for categorical variables. All values are presented as mean ± SD. Baseline clinical variables associated with virological response by univariate analysis (age, body mass index [BMI], triglycerides [log], AST [log], GGT [log], fibrosis score, HCV genotype [1 vs non-1], and glucose abnormalities [IFG plus type 2 diabetes]) were entered as predictors into a stepwise logistic regression analysis to control for confounders and to determine their independent association with SVR, and the odds ratios (ORs) and 95% confidence intervals (CIs) were determined. To further stress the independent relationship between glucose abnormalities and lower SVR, a case-control study was performed. Therefore, for each one of the 65 patients with glucose abnormalities another patient matched by gender, age, BMI, degree of liver fibrosis, and level of transaminases was selected among the 113 patients with normal fasting glucose. All P values were based on a 2-sided test of statistical significance. Significance was accepted at the level of P < 0.05. Statistical analyses were performed with the SSPS statistical package (SPSS, Chicago, IL).
RESULTS
During the study period 195 patients with chronic hepatitis C treated with combination therapy who met all the inclusion criteria and none of the exclusion criteria were selected for the study. Seventeen subjects who withdrew from therapy due to adverse effects (anemia [N = 6], fatigue [N = 4], depression [N = 2], and one case each of leukopenia, pruritus, alopecia, nausea, and insomnia) and did not complete at least 6 months of treatment were also excluded leaving 178 subjects for the final analysis. The most common adverse effects reported in these 178 patients were flu-like symptoms, headache, anorexia, nausea, and fever. Eight (4.49%) subjects discontinued treatment more than 6 months after the beginning because of adverse effects, which included depression (N = 3), anemia (N = 2), and one case each of mild injection site pain, hyperthyroidism, and severe malaise. No case of type 1 diabetes or severe hyperglycemia developed during the treatment period.
Sixty-five (36.51%) of the 178 patients presented glucose abnormalities; 19 had type 2 diabetes. Twelve of the 19 type 2 diabetic patients were under treatment with either oral agents (N = 9) or insulin (N = 3). The remainder of the type 2 diabetic patients, as well as the 46 patients with IFG, were only treated with lifestyle modification and diet. The presence of glucose abnormalities in the 17 patients excluded because of side effects of therapy was similar (N = 5, 29.41%) to the 178 included patients (P = 0.792). The HCV genotype distribution observed in our study was: 143 patients with genotype 1 (80.0%), 8 patients with genotype 2 (4.49%), 19 patients with genotype 3 (10.67%), and 8 patients with genotype 4 (4.49%). Differences in the prevalence of glucose abnormalities among these four genotypes were not observed (genotype 1 34%, genotype 2 37.5%, genotype 3 47.4%, genotype 4 37.5%, NS).
An SVR was obtained in 67 of the 178 (37.28%) patients treated with combined antiviral treatment. Including the 17 patients excluded from the original cohort the intention-to-treat response was 34.35%. Dose reductions and discontinuations were similar in both groups. The relapse virological rate of the 111 nonsustained responders was 35.13%. The main baseline clinical and biochemical features of the 178 patients according to the virological response are shown in Table 1. Distribution of sex was similar between groups. Age, BMI, triglycerides, GGT, and fibrosis score were significantly higher in nonresponders than in patients with SVR. HCV genotype 1 was also more prevalent among patients with no SVR. In addition, patients with no SVR had at baseline higher fasting blood glucose levels (104.4 ± 25.8 mg/dL vs 94.1 ± 12.7, P = 0.001) and a higher prevalence of glucose abnormalities (44.14% vs 24.24%, P = 0.012).
The independent variables related to a lack of SVR by multivariate analysis were the presence of glucose abnormalities (OR 2.72, 95% CI 1.12-6.59, P = 0.026), HCV genotype 1 (OR 2.85, 95% CI 1.07-7.62, P = 0.036), and log GGT (OR 6.62, 95% CI 1.70-25.74, P = 0.006) (Table 2).
In addition, the case-control study comparing the 65 HCV infected patients with glucose abnormalities with a subgroup of 65 patients with normal fasting plasma glucose matched by the main variables influencing glucose tolerance (age, BMI, degree of liver fibrosis, and level of transaminases) showed a significantly lower virological response rate in patients with glucose abnormalities (24.6% vs 44.6%, P < 0.001) (Table 3).
As shown in Figure 1, in the whole group of patients, the SVR rate was clearly related to glucose abnormalities: 15.78% (3/19) in HCV infected patients with type 2 diabetes, 28.26% (13/46) in patients with IFG, and 45.13% (51/113) in normoglycemic HCV infected patients (χ2 test for trend: P = 0.004). Similar results in the SVR rate were observed when only HCV genotype 1 infected patients were analyzed: 12.50% in diabetic patients (2/16), 17.64% in patients with IFG (6/34), and 43.01% in normoglycemic patients (40/93) (χ2 test for trend: P = 0.001).
Finally, a significant positive correlation between baseline glucose and GGT levels were detected (r = 0.425, P < 0,001). In addition, GGT levels from type 2 diabetic patients (95.26 ± 59.26 IU/L) were higher than IFG patients (44.44 ± 26.79, P < 0.001) and normoglycemic HCV infected subjects (47.95 ± 50.01, P < 0.001).
DISCUSSION
In the present study we provide evidence that the presence of glucose abnormalities is an independent factor for nonresponse to antiviral treatment in chronic hepatitis C. Therefore, glucose abnormalities should be added to the host-related factors (i.e., age, male gender, alcohol use, presence of cirrhosis, Afro-American ethnicity, obesity, hepatic steatosis, and HOMA) that adversely influence the rate of SVR in HCV infected patients (5, 6). Since the prevalence of both IFG and type 2 diabetes reaches 30% among noncirrhotic HCV infected patients (3), it seems likely that our findings will be relevant not only in the evaluation of the response to antiviral therapy, but also in giving support to therapeutic strategies addressed to normalize blood glucose levels before antiviral therapy.
In the univariate analysis, apart from a higher prevalence of glucose, the nonsustained responders included in our study were older, had higher BMI, and had more liver fibrosis and higher levels of triglycerides, GGT, and AST at baseline in comparison with nonresponders. HCV genotype 1 was also more prevalent among nonresponders. However, only glucose abnormalities, GGT, and HCV genotype were independently related to SVR. It must be taken into account that glucose abnormalities could be considered as the tip of the iceberg of a cluster of alterations associated with insulin resistance such as higher BMI, higher triglycerides, ageing, and liver steatosis and fibrosis. In support of this, when patients with glucose abnormalities were closely matched with HCV infected patients with normal fasting glucose for these variables, the sustained response rate continued to be significantly lower in the group with glucose abnormalities.
It is consequently understandable that glucose abnormalities, but not the other components associated with insulin resistance, were independently related to SVR. In this regard, although obesity has been proposed as an independent risk factor for nonresponse to antiviral treatment in chronic hepatitis C, no information on glucose metabolism was given in these reports (12-14). Therefore, it is likely that glucose abnormality rather than obesity itself was a more important factor accounting for the lower response to treatment observed in patients with obesity. In the present study, apart from the presence of glucose abnormalities and HCV genotype 1, high levels of GGT were also independently related to no response to antiviral treatment. Regarding GGT, it should be noted that it has been closely associated with insulin resistance in the setting of both obese and nonobese individuals with hepatic steatosis due to different etiologies including chronic HCV infection (15-18). It has been proposed that insulin resistance reduces insulin-dependent suppression of lipolysis (19) and the release of very low-density lipoproteins from the liver (20).
Both conditions are likely to increase triglyceride contents in the hepatocytes, and consequently, hepatic steatosis. Furthermore, several prospective studies have demonstrated that serum GGT is an independent predictor for developing metabolic syndrome and diabetes (15, 21). In addition, recent data indicate an inverse correlation between serum levels of adiponectin, an adipocyte-derived hormone that protects against insulin resistance and type 2 diabetes, and serum GGT values in patients with HCV-related steatosis (22). Hepatic steatosis may contribute to HCV-associated diabetes by impairing the insulin ability to lower hepatic glucose production and favoring liver fibrosis (4). In our study, where the mean BMI corresponds to nonobese subjects, GGT correlates with fasting glucose when all patients are evaluated, but only type 2 diabetic patients show higher levels of GGT. Moreover, in a recent study, both steatosis and GGT were predictors of insulin resistance in HCV infected patients, but no correlation existed between GGT and the extent of steatosis, suggesting that other factors beyond steatosis per se are responsible for higher GGT levels in these patients (23). Alternatively, insulin resistance can adversely affect the course of chronic hepatitis C and lead to enhanced steatosis, steatohepatitis, and liver fibrosis (4). Therefore, it is not surprising that GGT also appears as an independent risk factor for nonresponse to combined antiviral therapy in our cohort, which has been previously reported by Berg et al. (24).
The relationship between hyperglycemia and the reduced treatment response in HCV infected patients observed in our study may be mediated through several mechanisms. It has been reported that hyperinsulinemia blocks the inhibition of HCV virus replication by interferon, thus reducing the efficacy of antiviral therapy (25). In addition, HCV itself alters intrahepatic insulin signaling through a downregulation of peroxisome proliferator-activated receptor alpha and gamma (26) and the proteosomal degradation of the receptor insulin substrate-1 and -2 (27). Furthermore, HCV can induce insulin resistance directly through a proinflammatory cytokine mediated pathway (28). Supporting this hypothesis, baseline intrahepatic mRNA levels of TNF-alpha have been found to be significantly higher in nonresponders in comparison with sustained responders (29). In this regard, we have recently suggested that insulin resistance mediated by proinflammatory cytokines, but not a deficit of insulin secretion, is the main pathogenic mechanism involved in the pathogenesis of diabetes associated with HCV infection (30). Therefore, inflammatory cytokine imbalance could be the underlying mechanism connecting low response rate to antiviral therapy and glucose abnormalities in patients with chronic hepatitis C.
Apart from insulin resistance, hyperglycemia itself could be involved in the impairment of HCV clearance. Hyperglycemia can impair a wide range of functions in neutrophils and macrophages, including chemotaxis, adherence, phagocytosis, and intracellular killing of microorganisms (11). In this regard, it should be mentioned that the antibody response to hepatitis B vaccine in diabetic patients is poor (less than the titer recommended by WHO to ensure a sufficient level of protection) in almost 50% of cases (31).
Viral response rates to combined therapy in HCV infected patients have been heterogeneous depending on the type of therapy, the proportion of genotypes 2 and 3 included, and the characteristics of the patient population treated. The combination of interferon-alpha with ribavirin achieved SVR rates of 38-46% in treatment-naive patients, while the current therapy with pegylated interferon with ribavirin reaches 52% SVR rates (5). Our results raise the important question of whether or not the normalization of blood glucose levels can significantly improve the response rate. Tarantino et al. (23) have recently reported that patients with metabolic syndrome and chronic hepatitis C (N = 15) presented a response rate of 60% with combined therapy after 3 months of strict diet accompanied by weight and insulin resistance reduction compared to a 17.5% response rate in a similar group of patients (N = 17) on a free diet. However, further studies, including more patients and using not only lifestyle modification but also insulin sensitizers or other oral hypoglycemic drugs, are needed before recommending the achievement of normoglycemia prior to starting antiviral therapy.
In conclusion, we provide evidence that glucose abnormalities, together with more classical risk factors, adversely influence the rate of SVR in HCV infected patients treated with interferon and ribavirin. Some limitations must be taken into account in our study, like the retrospective data collection and the lack of a direct measure of insulin resistance. However, given that insulin resistance measurement by HOMA could be difficult to obtain in the routine management of these patients, our findings have obvious implications for clinical practice. Nevertheless, further prospective studies with a greater number of patients are required to support our results and to advise the fasting glucose threshold that might identify HCV infected patients at risk of being nonresponders to combined therapy. Finally, this study suggests that not only screening for glucose abnormalities should be indicated in HCV infected patients, but also that all measures for achieving normoglycemia may help to improve treatment outcomes in combined antiviral therapy.
|
|
| |
| |
|
|
|
