| |
Insulin resistance and response to therapy in patients infected with chronic hepatitis C virus genotypes 2 and 3
|
| |
| |
Jnl of Hepatology Jan 2008
Hossein Poustchi12, Francesco Negro3, Jason Hui4, Ian Homer Y. Cua1, Laura Rubbia Brandt3, James G. Kench5, Jacob George
"....the odds of non-response to treatment was 6.5-fold (P=0.02) higher in individuals with a HOMA-IR>2 compared with those in whom HOMA-IR was <2 after adjusting for the fibrosis stage (Table 6). By comparison, the odds for non-response to treatment for persons with severe fibrosis (F3-4) was 3.6 (P=0.02) when compared to those with mild fibrosis (F0-2), highlighting the greater relative importance of insulin resistance as a predictive variable for treatment response.... This effect appears to be independent of BMI and hepatic steatosis.... insulin resistance within the liver up-regulates the lipogenic transcriptional machinery, and perhaps results in intra-hepatic increases in HCV viral load."
Background/Aims
Obesity is associated with impaired treatment responses in chronic hepatitis C. The aim of this study was to determine the relationship between the insulin resistance frequently seen in obese subjects and sustained virological response to anti-viral therapy (SVR) in patients with genotype 2 or 3 infection.
Methods
Eighty-two patients were studied; 59 received interferon/ribavirin while 23 received peg-interferon/ribavirin.
Results
The overall SVR was (77%). Patients with a SVR had lower mean serum insulin (10.7±0.8μU/ml vs. 22.2±4.9; P=0.03), fibrosis stage (1.9±0.1 vs. 2.7±0.3; P=0.007) and insulin resistance measured by the homeostasis model (HOMA-IR) (2.5±0.2 vs. 6.1±1.5; P=0.03). Age, gender, ethnicity, alcohol consumption, treatment regimen, viral load, portal activity and steatosis did not influence the SVR. By linear regression, body mass index (P<0.001) and fibrosis stage (P<0.001) were independently associated with HOMA-IR. After adjusting for fibrosis stage, patients with HOMA-IR of <2 were 6.5 times more likely to achieve SVR than those with HOMA-IR>2.
Conclusions
Even in treatment-responsive genotypes 2 and 3, high HOMA-IR is associated with a reduced response. Improving insulin sensitivity may be a useful adjunct to anti-viral therapy in these individuals.
Results
The baseline demographic characteristics of the patient cohort are summarised in Table 1. Most of the subjects were male (72%) and Caucasian (87%), with 82% undergoing treatment for genotype 3 CHC infection. The treatment regimen for 59 patients (72%) was standard interferon and ribavirin, while the remaining 23 (28%) received pegylated-interferon in combination with ribavirin. In the overall cohort, a SVR was achieved in 63 (77%) patients, including 43 (73%) treated with standard interferon plus ribavirin and 20 (87%) undergoing treatment with pegylated-interferon and ribavirin. There was no difference in the SVR rates based on treatment regimen (P=0.2), or whether weight-based regimens were used (P=1.0). Hence the group was analysed as a whole.
Table 2 describes the clinical, biochemical, histological and viral characteristics of the patients in relation to their virological response to treatment. Subjects who achieved a SVR had lower indices of insulin resistance, including reductions in the serum insulin level (10.7μU/ml±0.8 vs. 22.2±4.9; P=0.03), the values for HOMA-IR (2.5±0.2 vs. 6.1±1.5; P=0.03), and a trend to lower levels of serum glucose (Table 2). As expected the mean fibrosis stage was lower in subjects who achieved a sustained viral response (1.9±0.1 vs. 2.7±0.3 P=0.007). There was no association between age, gender, ethnicity, alcohol consumption, treatment regimen, viral load, portal/periportal activity or steatosis, in relation to treatment response. There was a trend for greater BMI among the non-responders to anti-viral therapy, but this did not reach statistical significance (P=0.1).
Since measures of insulin resistance appeared to be associated with the treatment response category by univariate analysis, we next examined the correlates of an elevated HOMA-IR. As shown in Table 3, and as expected, HOMA-IR correlated with BMI and the fibrosis stage (P=0.001). Viral load was not associated with HOMA-IR (Table 3) even after log transformation of the data or using other cut-off values for a high viral load (data not shown). We next performed linear regression analysis to determine the factors independently associated with elevated values of HOMA-IR. Again, both BMI (P<0.001) and fibrosis stage (P<0.001) remained independent variables associated with the extent of insulin resistance as measured by the HOMA score. Indeed, patients with more advanced stages of hepatic fibrosis (F3 and F4) had greater mean HOMA-IR values than those with lesser degrees of fibrosis (F0-2) (5.5±1.1 vs. 2.2±0.2, respectively, P<0.001). When the HOMA scores were categorized into three groups (<2, 2-4, and >4), there remained a highly significant relationship between HOMA-IR and the severity of hepatic fibrosis by Metavir score (Table 4).
Given the close association between insulin resistance and treatment response category (Table 2), and between insulin resistance and fibrosis stage (Table 4), we next sought to determine the independent effects of insulin resistance and of hepatic fibrosis stage to treatment response. As shown in Table 5, there was a highly significant relationship between insulin resistance categories and treatment response (P=<0.001). Thus, SVR rates of 94% were obtained in those with a HOMA-IR of <2, while that in those with HOMA-IR>2 was 65%, akin to that observed in subjects with genotype 1 CHC treated with pegylated interferon and ribavirin [13]. Most importantly, we observed that the adverse effect of insulin resistance to treatment outcome was independent of the fibrosis stage. Thus, the odds of non-response to treatment was 6.5-fold (P=0.02) higher in individuals with a HOMA-IR>2 compared with those in whom HOMA-IR was <2 after adjusting for the fibrosis stage (Table 6). By comparison, the odds for non-response to treatment for persons with severe fibrosis (F3-4) was 3.6 (P=0.02) when compared to those with mild fibrosis (F0-2), highlighting the greater relative importance of insulin resistance as a predictive variable for treatment response. Genotype and viral load status were included as input variables in the multiple logistic regression analysis for the predictors of response to therapy (Table 6). They were not significant predictors in the model. HOMA-IR remained an independent predictor when these variables were included in the model and this indicates that the influence of HOMA-IR on response rate was unrelated to the genotype or viral load.
Since steatosis development in genotype 3 CHC, and the pathogenic mechanisms therein, may be different to that in patients with genotype 2 CHC, separate analyses were performed including only the 67 subjects with genotype 3 infection. The data were essentially unchanged with SVR rates of 93% in those with a HOMA-IR of <2, while that in those with HOMA-IR>2 was 64% (P<0.01). Likewise, steatosis was not a predictor of SVR among patients with genotype 3 CHC.
Discussion
In this study, performed exclusively in patients with the treatment responsive genotype 2 and 3, we sought to clarify the inter-relationships between the BMI, hepatic steatosis, insulin resistance and fibrosis stage, and the response to anti-viral therapy. Our major finding was that as with genotype 1 CHC, there is a close and inverse relationship between insulin resistance and SVR, even in subjects traditionally considered to respond very well to treatment. We identified a threshold discriminant HOMA-IR value of 2 to segregate those with a lower response to anti-viral therapy. Treatment responsiveness declined markedly with pre-treatment HOMA-IR values >2 these individuals were 6.5 times less likely to achieve a SVR. This effect appears to be independent of BMI and hepatic steatosis. The latter adds further credence to the notion that insulin resistance is the major independent determinant of treatment response in HCV positive individuals. While a larger cohort size would have provided greater credence to the data, the fact that we were able to demonstrate a highly significant and inverse relationship between insulin resistance categories and treatment response in this relatively small cohort of subjects suggests that the effect is both real, and likely to be clinically important.
Given the present findings, and the known relationship between insulin resistance and hepatic fibrosis [14], [17], [18], [22], [23] that we have confirmed herein, it was important to distinguish between these two clinically important variables. By logistic regression, our results demonstrated that insulin resistance is an important and independent predictor of SVR even after adjusting for fibrosis stage. It is acknowledged, however, that these results will need to be confirmed by other groups in larger cohorts.
An alternative interpretation of the present data is that insulin resistance is merely a surrogate marker for other factors such as obesity and steatosis which reduce treatment responsiveness. As shown in Table 3, HOMA-IR is associated with elevated BMI and obese subjects may achieve lower and sub-therapeutic concentrations of interferon [24]. However, pegylated interferon-2a has a large 40kDa polyethylene glycol moiety resulting in a reduced volume of distribution, and weight and body mass index (BMI) should not influence its serum concentration [25]. Despite this, weight and BMI have been shown to remain negative predictors of treatment responsiveness for pegylated interferon-2a (40kDa)-based anti-viral regimens [2], [15], [26]. Pegylated interferon-2b is currently administered on a weight-based regimen. However, even with the use of this agent, BMI remains an independent negative predictor of treatment response [1], [2], [7], [15], [26]. These data strongly imply that host factors such as insulin resistance, rather than BMI per se, are responsible for the lower SVR. This is consistent with the multivariate analysis (Table 6) which showed that HOMA-IR, rather than BMI, was an independent predictor of treatment response.
Insulin resistance can lead to hepatic steatosis which is associated with an impaired response to anti-viral therapy [27], [28], [29]. It has therefore been suggested that hepatic steatosis decreases the contact area between drug and hepatocyte membranes, leading to a reduction in anti-viral drug efficacy [30], [31] and therefore, in SVR rates. However, patients with genotype 3 have an excellent response to treatment, despite their greater extent of hepatic steatosis, rendering this an unlikely explanation [32]. Poynard et al. [16] showed that steatosis in genotype 3 patients was not associated with a lower sustained response. Likewise in the present report, steatosis did not influence the outcome of therapy. Hence, it is unlikely that the extent of hepatic steatosis per se is a sufficient explanation for the observed differences in treatment responsiveness related to insulin resistance.
If insulin resistance has a direct effect in reducing treatment response, there are several plausible mechanisms. In both in vivo and in vitro models, HCV infection is specifically associated with insulin resistance and up-regulation of hepatic lipid synthesis [14], [33], [34], [35]. Inhibition of lipid synthesis in vitro is associated with a 90% reduction in viral replication [33]. Insulin resistance per se, unrelated to HCV infection, is also associated with increased intra-hepatic lipid synthesis [36]. In turn, hepatic steatosis exacerbates hepatic insulin resistance. Thus, obesity, as well as direct viral inhibition of insulin signalling pathways [37], [38], can both cause and worsen hepatic steatosis and insulin resistance. In turn, insulin resistance within the liver up-regulates the lipogenic transcriptional machinery, and perhaps results in intra-hepatic increases in HCV viral load. The latter may reduce interferon-associated treatment responses, and be one explanation for the reduced response seen in genotype 1 and 3 patients with a high pre-treatment viral load [2], [5]. This effect might be saturable, accounting for the viral load cut-offs observed in clinical practice. Clearly, other molecular mechanisms to explain the negative association between insulin resistance and treatment response may exist, including insulin-mediated direct or indirect effects within the hepatocyte to either increase interferon responsiveness or to reduce interferon degradation.
Our data indicate that SVR rates of >90% are achievable in persons with low HOMA-IR values. Conversely, SVR rates drop to the levels seen with genotype 1 infection (`60%) at values above 2. An immediate implication is that irrespective of biopsy findings (which is not mandatory in many countries), genotype 2 and 3 infection with HOMA-IR values <2 can be confidently prescribed the currently available anti-viral therapy. On the other hand, since responses in persons with HOMA-IR values >2 are similar to those achieved with genotype 1 infection, consideration should be given to enrolling these patients into clinical trials that seek to improve virological response rates.
In conclusion, the present data suggest that insulin resistance is a powerful predictor of sustained virological response rates to currently available combination therapies for genotype 2 and 3 chronic hepatitis C infection. This effect appears to be independent of fibrosis stage. Those with a low level of insulin resistance as measured by the HOMA score can confidently be initiated on currently available therapies with a high likelihood of viral eradication. On the other hand, treatment decisions in patients with significant insulin resistance should be more circumspect.
Background and aims
Chronic hepatitis C (CHC) is a potentially curable disease. Combination therapy with pegylated interferon alpha and ribavirin results in overall sustained virological response (SVR) rates of 54-66% [1], [2], [3], [4]. Evidence accumulated from clinical trials over the last decade suggests that CHC resulting from genotype 3, and particularly genotype 2, has a high likelihood of a cure. SVR rates in this group of subjects is typically >75%, [1], [2], [3], [5] with rates as high as 93% reported in some studies [6]. The high response rates to treatment of genotypes 2 and 3 CHC have been further consolidated by studies which indicate that weight-based ribavirin dosing is not required [3], and that there is no difference in SVR rates between traditional interferon and pegylated formulations [1]. Indeed, an inadvertent consequence has been the failure to consider non-responders with these genotypes as a significant clinical problem. Instead, research has focused on designing trials to determine if shorter treatment durations are sufficient [5], [7].
Relatively few studies have specifically examined cohorts of genotype 2 and 3 subjects in order to identify predictors of a lower SVR. Male gender, higher body mass index (BMI), older age and most importantly fibrosis stage were not identified as adverse predictors of treatment response, though elevations in GGT and ALT were associated with a lower SVR in two reports [5], [7]. The presence of steatosis was identified as a negative predictor of SVR in two other studies [8], [9].
Recently, obesity has been identified as a modifiable host factor associated with a lower SVR [2], [10]. An elevated BMI is associated with reduced insulin sensitivity [11], [12] and hepatitis C virus (HCV) infected subjects with greater insulin resistance (IR) were found to have a lower response to anti-viral therapy [13]. Likewise, we have shown that patients who previously failed anti-viral therapy have higher insulin resistance values than those who achieved a SVR [14]. In these and other reports, up to 70% of the treated cohort were infected with the relatively treatment unresponsive genotype 1 HCV [11], [13], [15]. Therefore, the general relevance of insulin resistance to treatment response in CHC, particularly for genotype 2 and 3 infection, has not been determined.
The aim of the present study was (a) to determine the relationship between IR measured by the homeostasis model (HOMA-IR) and SVR rates in patients with genotype 2 and 3 CHC undergoing combination anti-viral therapy and (b) to determine whether a threshold of HOMA-IR exists that is associated with a reduced SVR. It was considered that the identification of a clinically useful predictor for treatment responsiveness will allow for tailored anti-viral regimens.
Methods
Patient selection
The study population comprised 82 subjects who underwent combination therapy for CHC. Fifteen (18.3%) were infected with genotype 2 and 67 (81.7%) with genotype 3. The consecutive patients were enrolled and the data were collected prospectively from 1999 at two university teaching hospitals (59 from Westmead Hospital, University of Sydney, Australia, and 23 from University Hospital, Geneva, Switzerland). These patients were part of a larger cohort that has been the subject of previous reports [14], [16], [17], [18]. They were selected from the database with the following inclusion criteria: (a) combination interferon and ribavirin-based anti-viral therapy, (b) a liver biopsy prior to the commencement of treatment and (c) evidence that the subject had received at least 80% of the intended dose of the two therapeutic agents for at least 80% of the intended duration of treatment confirmed by history and chart review. The latter criterion was included in order to ensure that patients could be segregated into true sustained responders and non-responders as previously published [19].
The following patients were excluded from the study by appropriate serological and biochemical tests and by a clinical history: diabetic patients and patients on anti-diabetic therapy, concomitant hepatitis B infection (defined as HBsAg positive), human immunodeficiency virus infection, autoimmune hepatitis, primary biliary cirrhosis, sclerosing cholangitis, hemochromatosis, 1-anti-trypsin deficiency, Wilson disease and drug-induced liver disease. During and for the 6 months prior to the commencement of anti-viral therapy, subjects had consumed <70g of alcohol per week.
Liver histopathology
An ultrasound guided percutaneous liver biopsy was performed in all subjects. The degree of necroinflammatory activity and of fibrosis was scored by an expert hepatopathologist based on the Metavir system [21] (F0, no fibrosis; F1, portal fibrosis without septa; F2, portal fibrosis with rare septa; F3, numerous septa without cirrhosis; F4, cirrhosis). Portal/periportal activity was scored from 0 to 3. Hepatic steatosis was scored as the percentage of hepatocytes containing macrovesicular fat droplets and was graded from 0 to 3 (grade 0, <5% steatosis; grade 1, 5-33% hepatocytes affected; grade 2, 34-66% of hepatocytes affected; grade 3, 67% or more of hepatocytes affected).
Treatment outcomes
Patients were initiated on treatment with either a combination of standard interferon (3 million units three times weekly) and ribavirin or pegylated interferon and ribavirin (which became available during the latter part of this study), for an intended duration of 6 months. A sustained viral response was defined as a negative HCV PCR (Roche Amplicor HCV Monitor ver.2 Roche Diagnostics) test performed at least 6 months after the completion of the therapy.
|
|
| |
| |
|
|
|
