Effect of different direct-acting antiviral regimens for treatment of nondiabetic hepatitis C virus–infected Egyptian patients on insulin resistance and sensitivity

The association between hepatitis C virus (HCV), insulin resistance (IR), and metabolic syndrome has been extensively investigated. Direct-acting antivirals (DAAs) have a high sustained virologic response (SVR) rate, reaching > 90%. The effect of SVR after DAA treatment on metabolic parameters and IR in nondiabetic patients could be an important factor in the patient’s long-term outcome. The aim of the study is to evaluate the impact of different DAA regimens on IR and sensitivity in naïve chronic HCV-infected nondiabetic patients (before and after 12 weeks of treatment). This prospective cohort study was conducted on 100 HCV-infected Child A nondiabetic patients eligible for DAA treatment in the Department of Gastroenterology and Hepatology, Ain Shams University, and Kobry El-Kobba Military Hospital among patients attending the outpatient clinic. Patients were categorized into four groups according to the HCV regimens they received for 12 weeks. All patient were subjected to the following tests before and 12 weeks after treatment: HCV quantitative PCR, Fibroscan, fasting insulin level (using insulin quantitative test kit), fasting and postprandial blood glucose (PPG), lipid profile, liver enzymes, BMI, and waist circumference. All patients achieved SVR at 12 weeks. In all treatment groups, lab was assessed before and after treatment, the 2-h PPG, high-density lipoprotein, and low-density lipoprotein levels showed statistically significant increases, whereas triglyceride, fasting glucose, hemoglobin A1C, and fasting plasma insulin levels showed statistically significant decreases. The homeostasis model assessment of insulin resistance (HOMA-IR) exhibited statistically significant decreases, whereas the quantitative insulin sensitivity check index (QUICKI) and Matsuda index showed statistically significant increases, across the four groups. DAA treatment in naïve nondiabetic HCV-infected patients affects metabolic profile and insulin resistance/sensitivity, with similar effect among different DAA regimens.


Background
Hepatitis C virus (HCV) has a known diabetogenic-and insulin-resistance effect on the infected patients, which could be explained by the activation of various inflammatory pathways [1]. HCV affects approximately 200 million

Open Access
The Egyptian Journal of Internal Medicine people worldwide; however, with the introduction of the highly curative direct-acting antivirals (DAAs), several nations have organized eradication plans for the disease. Nevertheless, questions still remain about the effectiveness for the eradication of the associated metabolic effects of the virus [2]. HCV genotype 4 is the most dominant in Egypt, differing from other genotypes in its effect on liver steatosis and response to DAAs [3,4].
Metabolic syndrome comprising steatosis, insulin resistance, and hyperlipidemia is associated with HCV, even in the absence of obesity or diabetes [5,6]. Nonalcoholic steatohepatitis (NASH) was a common finding in liver biopsies collected before the initiation of HCV interferon treatment, and they were incorporated into the scoring system that predicted the antiviral response [7].
The standard method for evaluating insulin resistance is obviously the oral glucose tolerance test (OGTT). Another method widely used in clinics is the hyperinsulinemic-euglycemic clamp [8].
Insulin resistance is associated with a higher risk for cardiovascular disease [9]. Various insulin resistance/ sensitivity scores could be measured using the insulin level. They are surrogate markers for the actual insulin resistance. Multiple validated scores for evaluating insulin resistance/sensitivity are suitable for clinical use, including the quantitative insulin sensitivity check index (QUICKI), homeostasis model assessment of insulin resistance (HOMA-IR), and Matsuda index [10]. The effect of HCV treatment on insulin resistance has been delineated by a recent meta-analysis conducted by Hu et al., which reported that HOMA-IR is decreased after achieving a sustained virologic response (SVR) [11].
Diabetes, which by the time diabetes is diagnosed, it may be too late, is preceded by a prediabetic phase, which could be detected through OGTT and fasting plasma glucose levels [12].
The aim of this study was to evaluate the efficacy of new DAAs for the treatment of chronic HCV infection in treatment-naïve nondiabetic patients on insulin resistance and insulin sensitivity using the following primary parameters for insulin resistance and sensitivity: HOMA-IR, QUICKI, and Matsuda index. Another aim was to investigate the effect of viral eradication on glucose levels and lipid profile, as a component of the complete metabolic profile of the patient.

Methods
This study was conducted at the Gastroenterology and Hepatology Department, Ain Shams University, and Kobry El-Kobba Military Hospital on 100 patients with chronic HCV infection who attended the outpatient clinic between September 2018 and June 2019.
The patients were scored as Child A according to the Child-Pugh score diagnosed based on HCV-Ab by enzyme-linked immunosorbent assay (ELISA) and HCV ribonucleic acid (HCV RNA) by PCR and received treatment with DAAs.
The patients were divided into the following four groups according to the drug received: i. Fasting serum insulin was measured using the insulin quantitative test kit, which is based on a solid-phase ELISA. The assay system uses first an anti-insulin antibody for solid-phase (microtiter wells) immobilization and then an anti-insulin antibody in the antibody-enzyme (horseradish peroxidase) conjugate solution. The standards and test specimen (serum) were added to the insulin antibody-coated microtiter wells. Then, the anti-insulin antibody labeled with horseradish peroxidase (conjugated) was added. The presence of human insulin in the test tube causes the formation of a molecular aggregate by the combination of the antibody on the well and the enzyme conjugate, where the insulin molecules are sandwiched between the solid-phase and enzymelinked antibodies. The wells of the labeled but unbound antibodies were washed after 1 h of incubation at room temperature. A blue color developed due to the addition of TMB solution and incubation for 20 min. The color development stopped after the addition of the stop solution, and then the color changed to yellow, at which time point it was measured spectrophotometrically at 450 nm. The color intensity in the test sample is directly proportional to the insulin concentration. j. Radiological investigations, including pelvicabdominal ultrasound examination to assess the homogeneity, echogenicity, and brightness of the liver, as well as the liver and spleen dimensions, portal vein diameter and splenic vein diameter, and the presence of thrombosis or focal lesions k. Electrocardiogram (ECG) and echocardiography for patients aged > 70 years l. Assessment of insulin resistance: HOMA-IR was evaluated before and after treatment. HOMA-IR = Fasting serum insulin × Fasting serum glucose (mg/dl)/405 [14]. m. Assessment of insulin sensitivity: QUICKI and Matsuda index were measured before and after SVR. 2. Matsuda index: This is used to assess the level of insulin sensitivity, which is calculated using plasma glucose and insulin levels [15]. A level of> 4.3 predicts insulin resistance [10]. After obtaining approval from the local institutional ethics committee, and an individual agreement of each participant in the study, an informed and written consent was obtained from each participant (Ethical Committee Approval number FMASU-MD-321/2017).

Statistical methods
1. Data analysis was conducted using the SPSS program version 23. 2. Quantitative data were expressed as mean and SD. 3. Qualitative data were presented as count and percentage. 4. One-way ANOVA was used to compare quantitative data between different groups followed by post hoc test when significant. 5. Fisher's exact test was used to compare qualitative data between groups. 6. Paired samples t test was used to compare data before and after treatment. 7. Repeated measures ANOVA was used to compare changes in quantitative data before and after treatment between different groups. 8. P < 0.05 was considered to be statistically significant.

Results
SVR was achieved at 12 weeks in all the 100 patients (100%), irrespective of the type of medications they received (groups A, B, C, and D). Tables 1 and 2 show the comparison between the four groups according to the demographic data (age and gender). Group A consisted of 23 men (92%) and 2 women (8%) aged (mean ± standard deviation) 57 ± 9 years. Group B consisted of 25 men (100%) aged 58 ± 7 years.
In group C, there were 23 men (92%) and 2 women (8%) aged 59 ± 9 years. Group D comprised 24 men (96%) and 1 woman (4%) aged 58 ± 7 years showing no statistical significance across groups. Among the study patients, there were 95 men and 5 women with age ranging from 34 to 72 years (mean age = 58.16 years). Patients with chronic HBV coinfection, HIV coinfection, schistosomiasis, and autoimmune hepatitis were excluded. Figure 1 shows the demographic data (height, weight, and BMI) of the four groups. Table 3 shows the comparison between the four groups according to laboratory data. At baseline, the TG level was highly significant (P < 0.001). Platelet count, total bilirubin, serum albumin, and INR showed positive statistical significance (P < 0.05). The remaining laboratory data (hemoglobin level, total leukocyte count, neutrophil count, lymphocyte count, ALT, AST, PT, serum creatinine, urea, total cholesterol, HDL, LDL, and AFP levels and PCR findings) showed no statistical significance.
The "a" and "b" post hoc test revealed statistically significant differences between groups A and C and between groups A and D in terms of TG level. The Kruskal-Wallis test revealed a statistically significant difference between the groups; however, it did not identify the groups that were different.
The liver dimensions showed no changes in groups A and B, but a statistically significant difference was detected in groups C and D between the dimensions at baseline and those after achieving SVR, as shown in Table 4.
Lipid profiles showed statistically significant increases in total cholesterol (mg/dl) from baseline to SVR at 12 weeks as follows: group A (from 172. 6     Regarding blood glucose (mg/dl), the fasting levels were significantly decreased from baseline to SVR at 12 weeks in the four groups as follows: in group A, from 92.64 ± 8.84 to 91.12 ± 8.55; in group B, from 87.28 ± 8.29 to 86.64 ± 8.55; in group C, from 88.68 ± 9.48 to 87.16 ± 10.08; and in group D, from 91.4 ± 8.15 to 90.16 ± 8.51 (Figs. 2B, 3B, 4B, and 5B, respectively).
However, the 2-h postprandial blood glucose levels showed statistically significant increases from baseline to SVR at 12 weeks in the four groups as follows: from 151.72 ± 16.34 to 181.16 ± 10.55, from 160.8 ± 11.47 to 184.48 ± 9.84, from 153.08 ± 31.69 to 184.52 ± 10.91, and from 163.24 ± 13.37 to 183.84 ± 11.25, respectively. A comparison of results could not reveal the effect of DAAs on 2-h postprandial blood glucose levels. These findings are illustrated in Figs. 2B, 3B, 4B, and 5B, respectively.
The HOMA-IR values also showed statistically significant decreases from baseline to SVR at 12 weeks in group A (from 3.19 ± 1.35 to 2.17 ± 0.96), group B (from 3.33 ± 1.31 to 2.38 ± 0.9), group C (from 2.9 ± 1.22 to 2.04 ± 0.76), and group D (from 3.55 ± 1.48 to 2.38 ± 1.09), as shown in Fig. 6A. The overall data and statistical differences are shown in Table 4.
The QUICKI values in our study exhibited significant increases from baseline to SVR at 12 weeks in group A (from 0.33 ± 0.02 to 0.35 ± 0.03), group B (from 0.32 ± 0.02 to 0.34 ± 0.02), group C (from 0.33 ± 0.02 to 0.35 ± 0.02), and group D (from 0.32 ± 0.03 to 0.34 ± 0.03), as shown in Fig. 6B. The overall data and statistical differences are shown in Table 4.
Regarding the Matsuda index, the values showed significant increases from baseline to SVR at 12 weeks in group A (from 5.57 ± 2.03 to 6.57 ± 2.31), group B (from 5.53 ± 2.4 to 6.34 ± 2.16), group C (from 6.02 ± 1.92 to 6.91 ± 2.34), and group D (from 5.17 ± 2.22 to 6.43 ± 2.91), as shown in Fig. 6C. The overall data and statistical differences are shown in Table 4.
Regarding the ultrasound findings, the values showed significant decrease in the liver dimensions from baseline to SVR at week 12 in group B and D only while the other groups showed no change (Tables 5 and 6). All groups did not show any change in spleen dimensions.

Discussion
The effect of DAAs on insulin resistance and sensitivity is an important subject in the post-HCV eradication era, where insulin resistance and hepatic steatosis remain new unresolved dilemmas. Although various studies have been conducted in this regard, each study concentrated on only one facet of the problem, i.e., either HOMA-IR, HOMA-β, or glucose levels and HbA1C or lipid profile and liver steatosis. In the present study, we evaluated the effect of treatment on all these parameters to assess the metabolic problem of the patients in general. Furthermore, we excluded all patients with abnormalities in their glucose tolerance test to ensure that there is no previous metabolic imbalance or an apparent prediabetic risk. Surprisingly, we found that even this strict category of patients (with normal OGTT) benefited from the eradication of HCV infection and showed a marked decrease in their lipid, glucose, and insulin levels. Liver dimensions were decreased after treatment with DAAs in our study, which could be due to the resolution of chronic hepatitis and the decline in the associated inflammatory response [21].
Regarding the HOMA-IR values, we detected a significant "within-subjects" effect, indicating a significant change in score over time. The "between-subjects" effect was not statistically significant, indicating that the mean score did not differ significantly between the study groups. There was no interaction between the variables time and drug type.
Hashim et al. [22] demonstrated that of the 75 patients treated for HCV infection by DAAs, the HOMA-IR value improved in 72% and 73.3% of patients after 12 weeks and 6 months, respectively. Compared with the interferon treatment era, the eradication of HCV infection showed a decline in HOMA-IR, fasting insulin, and glucose levels in the patients [8]. Adinolfi et al. found that HOMA-IR levels were decreased in nondiabetic patients treated for HCV genotype 1, with the viral-clearing effect being irrespective of BMI or baseline liver fibrosis [23].
In another Egyptian study conducted in the National Liver Institute, Menoufia University, the researchers found that different DAA regimens were associated with an increase in insulin secretion as evaluated by HOMA-β, an increase in insulin sensitivity as assessed by HOMA-S, and a decrease in insulin resistance as determined by HOMA-IR when SVR was achieved at 12 weeks [24]. Russo et al. examined the effect of DAAs on nondiabetic patients with HCV infection and found that insulin levels were decreased even in cirrhotic patients, or in those with high BMI, but the levels of BMI, fasting glucose, and HOMA-IR showed no significant changes after 12 weeks of treatment. This could be attributed to the baseline heterogeneity of the examined population, i.e., although they were nondiabetic, OGTT was not performed before treatment to analyze their glucose tolerance. However, their study showed similar effects on lipid profile as those observed in our study. Hu et al. [11] conducted a meta-analysis to investigate the effect of HCV treatment on insulin resistance in interferon-containing regimens and found that HOMA-IR and HOMA-β levels were significantly decreased in patients who achieved SVR compared with those who did not achieve SVR.
We could not find a comparative study regarding the QUCKI and Matsuda indices in patients before receiving DAAs and after achieving SVR at 12 weeks.
Our study showed that for HOMA-IR, QUICKI, and Matsuda index scores, there were significant "withinsubjects" effects, indicating significant changes in score over time. The "between-subjects" effects were not statistically significant, indicating that the mean scores did not differ significantly between the study groups. There