Effect of directly acting antivirals for hepatitis C virus infection on proprotein convertase subtilisin/kexin type 9 level

Abstract Background and aims Eradication of the hepatitis C virus (HCV) may affect proprotein convertase subtilisin/kexin type 9 (PCSK9) levels and cardiovascular risk. However, information regarding PCSK9 level after HCV eradication is lacking. Hence, in this case‐control retrospective study, we aimed to evaluate PCSK9 level from pretherapy baseline up to sustained virological response (SVR). Methods Eighty‐four patients treated with directly acting antivirals (DAAs) between July 2015 and May 2018 were enrolled. Differences in baseline PCSK9 level due to absence/presence of recorded baseline characteristics (covariates) were evaluated. Changes in PCSK9 levels from pretherapy to SVR (ΔPCSK9) and their correlations with the covariates were assessed. The repeated measures analysis of variance was used to investigate the differences in PCSK9 level from the baseline to the achievement of SVR due to absence/presence of any covariate. Results The mean age of the patients was 67.6 ± 11 years, and 53.6% were males. Baseline PCSK9 levels were statistically lower in patients using statins than in those not using statins (mean, 70.3 ± 43.1 ng/mL vs 271.8 ± 252.2 ng/mL; P = .017). PCSK9 level decreased significantly from baseline to the time of SVR (255 ± 248 ng/mL vs 169 ± 188 ng/mL; P < .001). PCSK9 levels were statistically higher in the HCV‐infected patients at baseline than in the control group (255 ± 248 vs 166.3 ± 120.2 ng/mL; P = .020); however, this difference was lost after achieving SVR (mean, 169 ± 188 vs 166.3 ± 120.2 ng/mL; P = .464). Changes in PCSK9 level was not statistically related to any of the recorded covariates. The PCSK9 mean level did not differ significantly with absence/presence of any covariate from pretherapy to SVR. Conclusions The reduction in mean PCSK9 level from baseline pretherapy to after HCV eradication was statistically significant. Whether PCSK9 is a new biomarker for cardiovascular risk in these patients remains to be ascertained.


| METHODS
This analysis was a nested study of the South Italian Network for Rational Guidelines and International Epidemiology (SINERGIE), 14 which was approved by the Ethics Committee of the Calabria Region (Project identification code #2012.58.E; June 19, 2013). The study was conducted at "Mater Domini" Teaching Hospital of Catanzaro (Southern Italy), in accordance with the guidelines of the Declaration of Helsinki and the principles of good clinical practice. 15 The patients provided informed consent for participation in the study.
In this case-control retrospective study, all patients who were treated with DAA between July 1, 2015 and May 31, 2018 were included. Patients who had no pre-SVR and post-SVR blood samples available were excluded. A gender-proportional healthy control group was used for comparison of the PCSK9 levels.
The following data were retrieved at baseline and at the time of SVR from the clinical records: demographics (age and gender), medical history (past and ongoing co-morbidities), HCV-RNA viral load, HCV genotype, complete blood count, occurrence of glycaemia, and aspartate aminotransferase (AST), alanine aminotransferase (ALT), gamma-glutamyl transferase (γGT), creatinine, total cholesterol, total bilirubin, albumin, and α-fetoprotein levels. All the parameters were measured following internationally standardized methods. HCV RNA viral load was determined using Cobas AmpliPrep/Cobas TaqMan HCV quantitative test v2.0 (Roche Diagnostics, Milan, Italy). Genotyping was performed using Versant HCV genotype v2.0 assay (LiPA, Siemens, Healthcare Diagnostic Inc., Tarrytown, New York). Data regarding the duration and type of DAA, lipid-lowering agents (ie, statins), and previous treatment with IFN-based regimens were collected.
Liver fibrosis was estimated either via transient elastography (KPa), or by using the fibrosis-4 (FIB-4) score, 16 or AST to platelet ratio index (APRI). 17 According to transient elastography, patients were considered cirrhotic when the estimated liver stiffness was ≥14.5 KPa. 18 Cirrhosis status was excluded or included when APRI index was ≤0.5 or ≥1.5 or when FIB-4 score was ≤1.45 or ≥3.25. Cases were observed at baseline and after achievement of SVR.
High affinity enzyme-linked immunosorbent assay (ELISA) (Elabscience) was performed to measure serum PCSK9 levels in cases and control groups in the laboratories of the University of Perugia. The principle of sandwich ELISA was followed. Serum samples were added to the micro ELISA plate and combined with specific antibodies. Then, a biotinylated detection antibodyspecific for PCSK9 was added and the plate was incubated. The unbound components were removed, and a stop solution was added to halt the enzyme-substrate reactions. Optimal density was measured via spectrophotometry and compared to the standard curve to calculate the PCSK9 levels. The coefficient of variation was <10%. The threshold of detectability of the test was 0.38 ng/mL, and the detection range was 0.63 to 40 ng/mL.
The cases were stratified in subgroups based on the presence or absence of each of the following characteristics (ie, covariates): gender (male vs female), previous use of IFN regimen vs absence of previous IFN treatments, statin co-medication vs absence of statin comedication, hypertension vs absence of hypertension; liver steatosis (ie, bright liver at ultrasound) vs absence of liver steatosis, dyslipidemia (i.e., total cholesterol >200 mg/dL or use of statin comedications) vs absence of dyslipidemia, diabetes vs absence of diabetes, history of cancer vs absence of history of cancer, and liver cirrhosis vs absence of liver cirrhosis.

| Statistical analysis
Differences between cases and control group by gender were assessed using the chi-square test. The Student's t-test for unpaired data was performed to assess the differences in mean age and mean PCSK9 levels between that at the baseline and at the achievement of SVR in both the cases and control group. The Student's t-test for unpaired data was also performed to assess the differences between the mean PCSK9 levels in the cases (at baseline and at the achievement of SVR) and the control group, and between the PCSK9 mean levels in the case subgroups mentioned above. The repeated measures analysis of variance (ANOVA) was used to investigate the potential differences in PCSK9 expression pattern in cases from the baseline to the achievement of SVR based on absence/presence of covariates, which are expressed as statistically significant interactions called "covariate by time." 3 | RESULTS

| Characteristics of the control group and comparison with cases
Thirty-eight healthy users were part of the control group. The gender distribution in the control group did not differ significantly from that of the cases: 23/38 (60.5%) individuals were males in the control group vs 45/84 (53.6%) in the cases group (P = .513). The mean age of the control group was lower than that in the cases (46.8 ± 13 years vs 67.6 ± 11 years; P < .001).

| Changes in PCSK9 level and possible predictors
As shown in Figure 1, PCSK9 level decreased significantly from baseline (mean 255 ± 248 ng/mL) to the time of SVR (mean 169 ± 188 ng/mL; P < .001), accounting for a mean ΔPCSK9 of 86 ± 112 ng/mL. PCSK9 levels were statistically higher in the HCVinfected patients at baseline than in the control group (mean 255 ± 248 vs 166.3 ± 120.2 ng/mL; P = .020); the difference was not statistically significant when controls were compared with patients after DAA therapy (mean 169 ± 188 vs 166.3 ± 120.2 ng/mL; P = .464).  significant reduction in PCSK9 level concomitant with HCV eradication using IFN-free regimens. Interestingly, although the mean age of the cases was higher than that of the control, the PCSK9 level did not differ significantly from that of the control group after achievement of SVR.
Our observations are comparable with those of Hyrina et al. 12 However, contrary to our results, they observed a significant increase in PCSK9 level in 27 patients who achieved HCV eradication. 12 The only other study on PCSK9 level in HCV-infected individuals 13 was conducted on patients treated with daclatasvir/asunaprevir, a combination of nonstructural protein inhibitors of NS5A and NS3, which are not available in Europe. Furthermore, in this study, 13 different laboratory techniques were used to quantify the levels of active and inactive PCSK9. This apparent inconsistency between our results and those of previous studies can be explained as follows. 12 Hyrina et al's study 12 was based on only 27 patients. Furthermore, the patients were treated with regimens including PEG-IFN, which significantly affects biomarkers of inflammation, coagulation, and oxidative stress. 19 In this context, it is noteworthy that Butt et al 20  SVR, corresponding to HCV eradication, is obtained in more than 95% of the treated patients for most DAA-containing regimens in clinical trials, 23 which has also been confirmed in real-life studies including our cohort from which patients were selected for the present analysis based on availability of serum samples taken before and after DAA therapy. 24 However, some major clinical adverse events, such as arrhythmia, impairment of cardiac function, or increased prevalence of dyslipidemia have been reported. [25][26][27] In contrast, several studies have shown the clinical benefit of SVR induced by DAA treatment with respect to "hard" clinical endpoints such as CVD-related mortality. 20,28,29 Hence, although DAA regimens are associated with worsening of serum lipid profile 27