Impact of hyperglycaemia and cholesterol levels on the outcome of hepatitis C virus (HCV) treatment in HIV/HCV-coinfected patients

Authors


Dr Laura Milazzo, Department of Clinical Sciences L Sacco, Section of Infectious Diseases and Immunopathology, University of Milan, Via GB Grassi 74, 20157 Milan, Italy. Tel: +390 239 043 350; fax: +390 250 319 758; e-mail: laura.milazzo@unimi.it

Abstract

Objectives

High serum total cholesterol and low-density lipoprotein (LDL) levels have been demonstrated to increase the probability of a sustained viral response (SVR) in chronic hepatitis C. Conversely, insulin resistance reduces SVR rates. We investigated the influence of baseline glucose and lipid values on the outcome of hepatitis C virus (HCV) treatment in HIV-1 infected subjects.

Methods

We retrospectively reviewed the charts of HIV/HCV-coinfected patients treated with an interferon-based regimen from 2002 to 2008. Fasting glucose levels and total cholesterol, LDL and triglyceride levels were recorded prior to the initiation of treatment.

Results

Of the 96 patients enrolled in the study, 36 (37.5%) had genotype 1, 48 (50%) genotype 2 or 3 and 12 (12.5%) genotype 4. SVR was obtained in 25% (nine of 36) and 70% (42 of 60) of patients with genotype 1 and other genotypes, respectively. In the multivariate analysis, the independent predictors of SVR were: genotype other than genotype 1 [adjusted odds ratio 9.64, confidence interval (CI) 2.7–34.3; P<0.0001], HCV viraemia [adjusted odds ratio 0.36, CI 0.15–0.9; P=0.028], fasting glucose ≥100 mg/dL [adjusted odds ratio 0.13, CI 0.034–0.51; P=0.003], and cholesterol level ≥190 mg/dL [adjusted odds ratio 5.96, CI 1.6–22.3; P=0.008].

Conclusions

Higher baseline serum glucose and cholesterol levels may be significant prognostic indicators for anti-HCV treatment outcome in HIV/HCV-coinfected patients.

Introduction

The widespread use of highly active antiretroviral therapy (HAART) has drastically changed the natural history and prognosis of HIV infection in developed countries. However, a broad range of metabolic complications, such as the lipodystrophy syndrome, with possible increased cardiovascular risk, have emerged in treated HIV-infected patients as a consequence of a combination of glucose homeostasis dysregulation, dyslipidaemia and abnormalities in body fat mass distribution [1]. Adipose tissue alterations in HIV-infected patients represent a complex and multifactorial spectrum of manifestations, which have been attributed either to HIV itself or to drug toxicity, with effects being mediated by mitochondrial damage or lipid metabolism derangement.

Moreover, hepatitis C virus (HCV) infection appears to be closely related to metabolic changes. HCV-infected patients show an increased prevalence of insulin resistance, hepatic steatosis, and hypocholesterolaemia, mainly associated with genotype 3a, which could contribute to the lack of response to antiviral treatment of HCV infection [2,3]. Recent cross-sectional and prospective studies highlighted an increased prevalence [4] and incidence [5] of type 2 diabetes in the HCV-infected population. Insulin resistance and type 2 diabetes were also associated with a reduced response rate to pegylated interferon plus ribavirin [6,7]. Conversely, a sustained virological response to treatment reduces the incidence of glucose dysmetabolism and improves insulin resistance as defined by the homeostasis model of assessment method [8,9].

It is currently unknown whether the metabolic changes that characterize HCV infection should be considered a mere prognostic marker of antiviral response or whether they have a causal role in treatment outcome through influencing interferon activity.

The aim of our study was to assess the role of metabolic factors and lipodystrophy on the sustained virological response rate in HIV/HCV-coinfected patients treated with combined pegylated interferon plus ribavirin.

Methods

We retrospectively reviewed, in our database, the charts of patients with HIV/HCV coinfection (defined as positive anti-HIV antibody test and detectable serum HCV RNA), followed as out-patients at the Department of Infectious Diseases, University of Milan, who had been treated for chronic active hepatitis C with pegylated interferon and ribavirin between 2002 and 2008. Of the 160 patients reviewed, those with alcohol abuse, hepatitis B virus coinfection or liver diseases other than HCV infection were excluded from the study. Also, subjects with CD4 cell counts <200 cells/μL before treatment initiation and those with no pretreatment glycaemic or lipid profile available, or who had been using hypocholesterolaemic drugs, were excluded from the analysis. Suboptimal therapies or dosages were not considered, and only patients who received a standard antiviral treatment with 180 μg of pegylated interferon α-2a or 1.5 μg/kg of pegylated interferon α-2b subcutaneously once a week, in combination with ribavirin, were included in the analysis. The daily ribavirin dose varied from 800 to 1200 mg depending on HCV genotype and body weight. Virological response was evaluated at 12 and 24 weeks, at the end of treatment (48 weeks), and 6 months after the end of treatment. A sustained virological response was defined as clearance of HCV RNA as measured by qualitative polymerase chain reaction (PCR) assay (COBAS AmpliPrep; Roche Diagnostics, Indianapolis, IN) 6 months after therapy discontinuation. Treatment was withdrawn in patients who did not achieve a virological response, defined as a 2 log10 decrease in HCV RNA (bDNA; Siemens Medical Solutions, Berkeley, CA, USA) at week 12 or undetectable HCV RNA as measured by PCR at week 24.

Demographic [age, sex, height, weight and body mass index (BMI), calculated as weight in kg/height in m2], metabolic [levels of fasting total cholesterol, high-density lipoprotein (HDL), low-density lipoprotein (LDL), triglycerides and fasting glucose, blood pressure and presence of lipodystrophy], viro-immunological (baseline CD4 cell count and HIV viral load) and HCV-related [HCV genotype and baseline viral load, levels of alanine aminotransferase (ALT), gamma-GT and α-fetoprotein, and presence of an ultrasonographic pattern of steatosis] characteristics were recorded at baseline. In particular, hypercholesterolaemia was defined as a baseline cholesterol level higher than the normal limit (190 mg/dL) as assessed by our Institutional Biochemistry Laboratory, based on the recommendations of the European Society of Cardiology [10], hyperglycaemia was defined as baseline fasting plasma glucose ≥100 mg/dL and hypertension as systolic/diastolic blood pressure ≥130/≥85 or by the prescription of antihypertensive therapy. The diagnosis of lipodystrophy, defined as self-reported and physician-confirmed loss of subcutaneous fat, with or without an increase in abdominal girth, was divided according to the Marrakech classification [11] into three types: fat loss, fat accumulation and combined forms (fat loss and fat accumulation in different body regions).

Continuous variables are expressed as median ± interquartile range; categorical variables are expressed as number of cases (percentage). Statistical analyses were performed using spss 15 for Windows XP (SPSS, Chicago, IL, USA). Continuous data were analysed using Student's t-test and Pearson correlation if they appeared normally distributed or the Mann–Whitney test and Spearman correlation otherwise. Categorical data were analysed using the χ2 test. All statistical tests were two-sided and P-values <0.05 were considered statistically significant. Variables associated with a sustained viral response (SVR) with a P-value ≤0.05 in the univariate analysis were included in the multivariate logistic regression analysis where the dependent variable was SVR.

Results

A total of 96 consecutive HIV/HCV-coinfected patients, treated for chronic hepatitis C, were eligible for inclusion in the analysis. The patients' characteristics at baseline are summarized in Table 1.

Table 1.   Baseline characteristics of the patients
Baseline characteristicOverall (n=96)SVR (n=51)No SVR (n=45)UnivariateMultivariate
POR95% CIPAOR95% CI
  1. Values are expressed as median ± interquartile range for continuous variables, and number of patients (%) for categorical variables. Univariate and multivariate models are shown for the associations with sustained virological response (SVR). Significant P-values are shown in bold.

  2. ALT, alanine aminotransferase; AOR, adjusted odds ratio; BMI, body mass index; CI, confidence interval; HAART, highly active antiretroviral therapy; HCV, hepatitis C virus; HDL, high-density lipoprotein; LDL, low-density lipoprotein; OR, odds ratio; Ref., reference group; US, ultrasonography.

Age (years)43 (41–46)43 (40–44)44 (42–46)0.0360.8920.801–0.9930.4290.9480.829–1.083
Sex
 Male81 (84.4%)42 (82.4%)39 (86.7%)Ref.Ref.Ref.   
 Female15 (15.6%)9 (17.6%)6 (13.3%)0.5621.3930.454–4.274   
Duration of HIV infection (years)18 (11.3–20)17 (9–20)18 (14–20.5)0.0650.9370.874–1.004   
Patients on HAART87 (90.6%)44 (86.3%)43 (95.6%)0.1380.2920.057–1.487   
Time on HAART (years)10 (5–14.8)9 (3–12)11 (7.5–15)0.0510.9260.858–1.000   
Baseline CD4 count (cells/μL)556 (422–722)552 (415–739)584 (422–720)0.4160.9990.998–1.001   
Baseline CD4%26 (20.8–32)26.7 (20–33.4)25.6 (21.5–31.5)0.8821.0040.957–1.052   
HIV RNA (copies/mL)<50<50<500.0513.0270.993–9.228   
HCV genotype (%)
 136 (37.5%)9 (17.6%)27 (60%)Ref.Ref.Ref.Ref.Ref.Ref.
 Not 160 (62.5%)42 (82.4%)18 (40%)0.0007.0002.748–17.8280.0009.6422.711–34.300
HCV viraemia (log UI/mL)5.79 (5.38–6.23)5.53 (4.94–6.23)5.87 (5.52–6.23)0.0110.4090.205–0.8170.0280.3660.150–0.895
ALT (IU/L)99 (68–135)105 (86–143)81 (59–132)0.0301.0081.001–1.0160.2161.0060.997–1.016
BMI (kg/m2)23.18 (21.61–25.27)23.39 (21.66–25.04)22.89 (21.57–25.46)0.8590.9850.837–1.160   
Glycaemia (mg/dL)92 (84–103)88 (82–96)97 (86–109)0.0280.9720.948–0.997   
Cholesterolaemia (mg/dL)168 (139–193)179 (140–203)158 (137–174)0.1091.0080.998–1.017   
LDL (mg/dL)79 (56–110)100 (56–120)74 (57–100)0.1051.0110.998–1.024   
HDL (mg/dL)41 (32–50)39 (32–48)44 (31–52)0.3870.9870.958–1.017   
Triglycerides (mg/dL)160 (103–219)155 (105–221)161 (100–217)0.9301.0000.998–1.002   
Hyperglycaemia or diabetes28 (29.2%)7 (13.7%)21 (46.7%)0.0010.1820.068–0.4890.0030.1330.034–0.512
Hypercholesterolaemia31 (32.3%)23 (45.1%)8 (17.8%)0.0043.9401.531–10.1380.0085.9661.593–22.349
LDL>100 mg/dL36 (37.5%)25 (49%)11 (24.4%)0.0203.1001.193–8.058   
Hypertriglyceridaemia52 (54.2%)25 (49%)27 (60%)0.3290.6670.295–1.505   
Hypertension45 (46.9%)24 (47%)21 (46.7%)0.8861.0620.469–2.402   
US steatosis31 (32.3%)14 (27.5%)17 (37.8%)0.2820.6230.263–1.474   
Fib-4 index >3.2517 (17.7%)8 (15.7%)9 (20%)0.5640.7110.224–2.262   
Lipohypertrophy7 (7.3%)5 (9.8%)2 (4.4%)0.4242.0450.354–11.820   
Lipoatrophy41 (42.7%)21 (41.2%)20 (44.4%)0.7330.8590.359–2.058   
Combined form8 (8.3%)3 (5.9%)5 (11.1%)0.3720.4910.103–2.339   

Patients were mostly male (81 of 96; 84.4%), with a median age of 43 years (range: 41–46 years). HIV infection had been diagnosed a median of 18 years before the analysis. Eighty-seven of the 96 subjects (90.6%) had been treated with antiretroviral therapy, for a median time of 10 years. The antiretroviral regimen at the time of analysis included a protease inhibitor in 47 patients. The median CD4 cell count at baseline was 556 cells/μL and median HIV viral load was undetectable (<50 HIV-1 RNA copies/mL). Forty-five patients received pegylated interferon α-2a and 51 pegylated interferon α-2b. Demographic characteristics, as well as HIV-related parameters, were similar between patients who achieved a SVR and those who did not (P>0.05 as shown in Table 1). The pretreatment lipid panel [median (interquartile range)] was: total cholesterol, 168 (139–193) mg/dL; LDL, 79 (56–110) mg/dL; HDL, 41 (32–50) mg/dL; and fasting glycaemia, 92 (84–103) mg/dL. Overall, 36 patients harboured HCV genotype 1, six genotype 2, 42 genotype 3 and 12 genotype 4 infection.

As liver histology was available only in 27% of the study patients, we applied to our cohort the Fib-4 index [age (years) × aspartate aminotransferase (AST; IU/L)]/[platelet count (108/L) × ALT (IU/L)1/2] [12], a recently developed model for the noninvasive assessment of liver fibrosis in HIV/HCV-coinfected patients. In our population, 37.5% of patients had Fib-4 index values <1.45, predictive of mild or absent fibrosis (corresponding to Ishak stages 0–3); 17.7% had Fib-4 values >3.25, predictive of advanced fibrosis (corresponding to Ishak stages 4–6), and 44.8% had intermediate values between the two cut-offs, indicating an indeterminate index.

Sixty patients out of the 96 (62.5%) achieved an early viral response (EVR) and 51 patients (53%) an SVR. According to the HCV genotype, the SVR rate was 25% (nine of 36) in patients with genotype 1 and 70% (42 of 60) in patients with genotypes other than genotype 1. HCV infection with genotypes other than genotype 1 was significantly associated (P=0.0001), in the univariate analysis, with a higher probability of achieving a SVR [odds ratio (OR) 7; 95% confidence interval (CI) 2.75–17.82]. SVR was inversely correlated with basal HCV viral load (OR 0.4; 95% CI 0.2–0.81; P=0.01) and levels of aminotransferases.

Continuous values of metabolic parameters such as BMI, total, LDL or HDL cholesterolaemia and triglyceridaemia did not differ between the two groups of patients. Hypercholesterolaemia and elevated serum LDL cholesterol (>100 mg/dL) were associated with a significantly higher probability (OR 3.94; 95% CI 1.53–10.13; P=0.004 and OR 3.1; 95% CI 1.19–8.0; P=0.02; respectively) of achieving a SVR. The analysis of different ranges showed the highest probability of response to be associated with total cholesterol levels ≥220 mg/dL and LDL ≥120 mg/dL (OR 4.75; 95% CI 0.94–23.98; P=0.05 and OR 5.4; 95% CI 1.32–22.3; P=0.019, respectively). An increased probability of SVR was also seen for a cholesterol range of 191–219 mg/dL (OR 2.4; 95% CI 0.73–7.66; P=0.12) and an LDL range of 100–119 mg/dL (OR 2.1; 95% CI 0.67–6.56; P=0.2), but this was not statistically significant. No correlation was found between anti-HCV treatment outcome and hypertension or the ultrasound detection of hepatic steatosis. Greater age increased the risk of nonresponse (P=0.036). Liver fibrosis, defined as a Fib-4 index >3.25, did not significantly affect the treatment outcome in our analysis. Moreover, no correlation was found between HCV genotype and baseline values of total cholesterol, LDL cholesterol and triglycerides when adjusted for age, gender and time on HAART.

Hyperglycaemia or diabetes was present at baseline in 21 of 45 nonresponders (46.7%) but only in seven of 51 responders (13.7%) and the univariate analysis identified impaired glucose metabolism as the major factor influencing SVR after genotype (P=0.001; OR 0.18; 95% CI 0.068–0.489).

After adjusting for age, genotype and HCV viral load, multivariate analysis showed a reduced probability of achieving EVR and SVR for hyperglycaemic (≥100 mg/dL) or diabetic patients (P=0.02; OR 0.26; 95% CI 0.08–0.82 and P=0.003; OR 0.13; 95% CI 0.034–0.51, for EVR and SVR, respectively), whereas a higher probability of SVR, but not of EVR, was associated with high levels of cholesterol (≥190 mg/dL) (P 0.008; OR 5.96; 95% CI 1.59–22.35) (Table 1).

Finally, we assessed the possible influence of HIV-related lipodystrophy on treatment outcome. Fifty-six of the 96 patients (58.3%) in our cohort were defined as being lipodystrophic, with most of the patients (41 of 96; 42.7%) showing peripheral lipoatrophy without a significant correlation with anti-HCV therapy response in the univariate analysis. Only seven and eight patients showed fat accumulation and combined forms of fat distribution, respectively.

Discussion

Recent studies have identified glucose and lipid metabolic parameters as possible adjunctive predictors of SVR in HCV- and HCV/HIV-infected patients, in addition to the well-known virological (HCV genotype and HCV viral load) and host (age, liver disease stage and CD4 cell count) factors. Particularly, insulin resistance and type 2 diabetes were found to be associated with a reduced sensitivity to anti-HCV treatment in HIV/HCV-coinfected persons [13,14]. Moreover, serum LDL and total cholesterol levels have been identified as prognostic indicators of SVR in patients with HCV infection, particularly infection with genotypes 1 and 2 [14,15].

In this retrospective study, we evaluated, in HIV/HCV-coinfected patients, the influence of serum fasting glucose levels and total cholesterol, LDL and triglyceride levels on HCV treatment outcome. In our study population the SVR rate was 25 and 70% in genotype 1 and other genotypes, respectively, in accordance with response rates reported in other casistics [16]. We analysed hyperglycaemia and type 2 diabetes rather than insulin resistance, as defined by the homeostasis model of assessment – insulin resistance, as baseline insulinaemia measurements were available only in a few patients. In the univariate analysis, age, genotype, HCV viraemia, and ALT, glucose, total cholesterol and LDL levels were found to affect SVR. Multivariate analysis showed a higher probability of achieving SVR with higher cholesterol levels (≥190 mg/dL), whereas glucose values ≥100 mg/dL were negatively associated with sustained response rate.

In individuals with chronic hepatitis C, insulin resistance and type 2 diabetes are more often seen than in healthy controls. In a large cohort of 9841 patients, type 2 diabetes was evident in 8.4% and HCV infection increased the risk of diabetes 3.77 times in patients older than 40 years [4].

HCV infection promotes insulin resistance via direct interference with the phosphorylation of insulin receptor substrate (IRS)1 and IRS2 or via the tumour necrosis factor-α- and suppressor of cytokine signalling (SOCS)-3-dependent proteosomic degradation of IRS1 and IRS2 [17,18]. Moreover, HCV impairs β-oxidation through direct damage of mitochondrial membranes or through the inhibition of peroxisome-proliferator-activated receptors α and γ [19].

Recent studies suggest that HCV infection is not associated with the ‘classic’ metabolic syndrome as a whole but rather with markers of insulin resistance and inflammation. This is the reason why we chose not to assess the possible relationship between the anti-HCV response and the presence of the ‘classic metabolic syndrome’ often defined using the National Cholesterol Education Program Adult Treatment Panel III criteria. We believe that it is important to distinguish HCV-related metabolic consequences from the classic dysmetabolic phenotype. In fact, patients with HCV infection can show a specific hepatitis C-associated dysmetabolic syndrome in which insulin resistance plays a major role, but which is distinguished by the presence of steatosis and hypocholesterolaemia, and the absence of arterial hypertension and obesity [20,21]. In HIV/HCV-coinfected patients, the situation is even more complex. In this population, HIV infection itself and antiretroviral therapy profoundly affect metabolic parameters, sometimes amplifying the HCV effect, as observed for glucose metabolism [22,23], or counteracting it, as described for cholesterol [24]. This is particularly true for coinfected patients harbouring HCV genotype 3, which was recently correlated with lower total cholesterol and triglyceride levels independently of HAART [25,26]. We failed to confirm this association in our cohort, probably because the number of patients was too small for this purpose.

Finally, changes in body fat distribution associated with lipodystrophy impair the correct evaluation of abdominal obesity for the assessment of metabolic syndrome. In agreement with the reported increased risk of developing lipoatrophy for HCV-coinfected patients [27], we found a high prevalence of lipoatrophy in our study population, although this morphological alteration did not modify the probability of SVR in the logistic regression analysis.

A major limitation of this study is the absence in the analysis of liver histology, because of the limited availability of these data in our cohort. The Fib-4, a simple index based on routine laboratory tests, has been recently assessed and has been demonstrated to accurately differentiate mild/moderate fibrosis from bridging fibrosis/cirrhosis in HIV/HCV-coinfected patients [12]. It is noteworthy that 44.8% of the cohort had an indeterminate Fib-4 index (between 1.45 and 3.25) which did not allow prediction of the severity of liver fibrosis. Although severe fibrosis is known to be an independent predictor of SVR for patients with chronic hepatitis C and is characterized by a progressive decrease of serum total cholesterol and LDL levels [28], the low rate of Fib-4 index >3.25 in our cohort could explain the lack of correlation between fibrosis and nonresponse to treatment.

In conclusion, our study identified, in an HIV/HCV-coinfected population, hyperglycaemia, serum total cholesterol and, to a lesser extent, LDL as predictors of response to pegylated interferon plus ribavirin therapy, as previously reported in HCV-infected subjects. A possible explanation of the influence of total cholesterol and LDL levels on virological response to treatment is that the entry of HCV into the cell is mediated by an LDL receptor that could be down-regulated by high levels of total and LDL cholesterol, resulting in reduced HCV infectivity [29,30].

As HIV/HCV-coinfected patients display a complex pattern of metabolic alterations in which viral features, host characteristics and drug toxicity interact and influence each other, prospective studies will better address this important issue and the underlying mechanisms.

Ancillary