Drug plasma trough concentrations during treatment with daclatasvir and sofosbuvir are associated respectively with liver impairment and with renal dysfunction in HIV/HCV co-infected patients

Sofosbuvir (SOF) plus daclatasvir (DCV) achieved high rates of sustained virologic response (SVR) with no difference according to HIV serostatus. Only limited information is available on the pharmacokinetics variability of SOF and DCV in HIV/HCV co-infected patients. Aim was to evaluate the association of plasma drug concentrations (Ctrough) of SOF and of DCV with patient-, treatment-, and disease-related factors in the real-world setting of HIV/HCV co-infected persons.


Conclusion
In HIV/HCV co-infected patients receiving SOF plus DCV, plasma drug concentrations are associated with renal dysfunction for GS-331007 and with liver impairment for DCV.
Though clinical and therapeutically relevance of these findings may apparently be limited, growth of clinicians' knowledge on DAA exposure in difficult-to-treat patients, as cirrhotic and renal impaired subjects, can be relevant in single cases.

Background
The development of peginterferon-free (pegIFN-free) oral regimens of direct-acting antivirals (DAA) greatly improved the efficacy and tolerability of HCV treatment in HIV coinfected population [1][2][3]. Achieving sustained virological response (SVR) after treatment is associated to regression of hepatic disease, including fibrosis reduction, as well as to improvement of several comorbid conditions [4][5].
Focusing on the pan-genotypic DAA combination SOF plus DCV, they are both dosed once daily and have favorable safety profiles [8][9]. In phase III randomized clinical trials, this regimen achieved high SVR rates without difference by HIV sero-status, but with worse outcomes in cirrhotic patients [3,10,11]. However, it is not clear if these SVR results may be replicated in the real-world HIV/HCV co-infection setting, where patients are expected to have more advanced liver disease and higher frequency of comorbidities, as renal disease as compared with clinical trials [12]. Indeed, liver and kidney dysfunction can affect drugs metabolism, including DAA that are cleared by liver, except SOF, which is mainly eliminated by kidney [13]. Only limited information is available on the pharmacokinetics of SOF and DCV in HIV/HCV co-infected patients and particularly on their variability based on liver and renal function.
Thus, aim of the study was to evaluate the association of plasma drug concentrations (Ctrough) of SOF and of DCV with patient-, treatment-, and disease-related factors in the real-world setting of HIV/HCV co-infected persons.

Study design
This study was a monocenter, prospective, observational study evaluating DAA Cthrough in HIV/HCV co-infected persons undergoing anti-HCV treatment with SOF and DCV at the National Institute for Infectious Diseases in Rome, Italy, between February 2015 and March 2016.

Patients and HCV treatment
Patients with HIV/HCV co-infection and treated with daily DCV (30 or 60 or 90 mg, based on drug to drug interactions with antiretrovirals) and SOF 400 mg together with antiretroviral therapy, were enrolled. 6-9 Physicians prescribed HCV treatment according to international treatment guidelines6,7 and criteria of the Italian Medicines Agency available during the study period (stage F3-F4 or any stage of fibrosis with HCV-associated extrahepatic manifestations). Biochemical and viro-immunological parameters were assessed at baseline, at week 4 (W4) during DAA treatment, at end of treatment (EOT), and after-EOT. SVR was defined by HCV RNA <12 copies/mL 12 weeks after-EOT (SVR12).
Adherence to antiretrovirals and to DAA during the previous 4 weeks was self-reported through a visual analogue scale (VAS) ranging from 0 (no drug intake) to 100 (intake of all prescribed medications) and a Likert scale from 1 (no drug intake) up to 4 (intake of all prescribed medications).
Transient elastography before DAA start was used to measure stiffness and to classify fibrosis stage. The Fibrosis 4 score (FIB-4) for non-invasive liver disease assessment and the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation for glomerular filtration rate estimate (eGFR), were calculated using standard formulas at baseline, W4, EOT and after-EOT and used as marker of renal and hepatic function before, during and after DAA treatment.

DAA plasma concentrations
Venous blood samples for Ctrough levels were drawn at the end of dosing interval (20)(21)(22)(23)(24) hours post-dose) and assessed at W4 and W8 during DAA treatment. In order to decrease variability due to single Ctrough value, a mean value between the two time-points (mean-Ctrough) was calculated.
As SOF becomes undetectable in blood 4-5 hours after administration due to intracellular activation to triphosphate GS-461203 (not detected in plasma) and ultimately to GS-331007, this latter is the predominant circulating metabolite and the primary analyte of interest in clinical pharmacology studies [8]. Ctrough were determined for SOF, GS-331007, and DCV using a validated UPLC-MS/MS method. Limit of quantification were 11.71 ng/mL for SOF and DCV, and 19.53 ng/mL for GS-331007 [14].

Statistical analysis
Median and interquartile range (IQR) were used to summarize continuous variables and absolute frequencies with percentages for categorical variables. Statistical comparisons of SOF and DCV mean-Ctrough, between different categories of factors, were made using Mann-Whitney test. To evaluate correlation between DAA mean-Ctrough and continuous variables, Spearman rank correlation was calculated.
For the purpose of the study, we considered CKD-EPI at baseline as a marker of the pretreatment renal function, and CKD-EPI at EOT and post-EOT to investigate renal function during and after DAA. FIB-4 at baseline and at W4 were used to evaluate hepatic function during the first month of DAA treatment [15]. Evaluated cut-offs were: 12 KPa for stiffness level, 3.25 at FIB-4 for severity of liver impairment, and 60 mL/min at CKD-EPI for impairment level of renal function. STATA 10.1 software was used for statistical analysis.

Patient population
Thirty-five HIV/HCV co-infected patients were included and general characteristics were shown in Table 1 (first column). HCV-RNA at baseline was a median of 5.7 (IQR 5.1-6.1) log10 copies/mL, and decreased by -4.6 [IQR -5.0-(-4.1)] copies/mL at W4 of DAA treatment. HCV-RNA at EOT was available in 32/35 patients because one patient died for non-liver-related cause, one was lost to follow-up, and blood test was not available for the third patient. Overall, SVR12 was achieved in 30/32 cases: 2 patients (6.2%), both cirrhotic with hepatocellular carcinoma, experienced virological failure. As shown in Table 1, no statistically significant associations were found between GS-331007 or DCV mean-Ctrough and most analyzed variables. Only renal (CKD-EPI) and liver function (FIB-4) significantly correlated with, respectively, GS-331007 and DCV Ctrough values.
A statistically significant correlation was observed between DCV plasmatic exposure and

Discussion
The results of our study showed that, in HIV/HCV co-infected patients, plasma concentrations of SOF and of DCV are respectively associated to renal and hepatic function.
On one hand, plasma exposure to the primary metabolite of SOF, GS-331007, that is primarily excreted renally, correlated to kidney function, in terms of higher concentrations with decreasing eGFR. As this association was not present at baseline and disappeared after DAA treatment, a detrimental effect of SOF treatment on kidney function can be hypothesized.
On the other hand, decreasing DCV plasma concentrations were associated to increasing hepatic impairment as assessed by FIB-4, and this association was found throughout the entire observation period. Besides these observations, no further associations between DAA plasma concentrations and other patient-, treatment-, or disease-related factors were observed.
In our study, lower eGFR during DAA treatment (W4) and at EOT significantly correlated with higher GS-331007 plasma exposure. This finding is in accordance with previous observations that found increase of GS-331007 plasma concentration up to 451% in patients with severe renal damage, leading to possible further nephrotoxicity [8,13]. In fact, real-life studies suggested higher risk of eGFR deterioration with SOF treatment in patients with preexisting renal impairment, particularly chronic kidney disease, kidney transplant or HIV co-infection [16]. Our study, conducted in patients with overall normal kidney function at baseline, showed mild (within normal range) decline of median eGFR during SOF and DCV treatment, but highlights the association of minor and transitory renal function modifications with SOF metabolite plasma concentrations. The clinical effect of this observation still needs to be assessed, and may be of particular importance in HIV/HCV co-infected patients. In the meanwhile, strict monitoring of renal function during treatment with SOF-based regimens is warranted.
DCV is hepatically metabolized and biliary excretion is the major route of elimination, as it is highly bound to plasma proteins (99%). In cirrhotic patients, decreased hepatic function impairing metabolic pathways, may lead to lower DCV plasma concentrations [13]. As previously demonstrated and reported in the product information sheet by the European Medicines Agency, among patients with hepatic impairment (Child-Pugh A/B/C) plasma concentrations were lower if compared with unimpaired subjects. Notably, the unbound fraction of DCV remained unchanged in patients with cirrhosis, so that no differences in the active concentration of DCV were found and no need for dose modifications is advised [9][10][11]13]. The results of our study were consistent with these observations: decreasing DCV Ctrough values were found with increasing severity of hepatic disease.
Even though an association between lower DCV plasma concentrations and slower HCV viral kinetics in HCV mono-infected and HIV/HCV co-infected persons has been reported [17,18], data suggested that the presence of liver impairment has only a limited relevance on DCV pharmacokinetics and, consequently, on its clinical effect [18].
In our study, no apparent effect on the treatment's virological efficacy was observed but the limited number of subjects failing HCV treatment (only 2 patients) could represent a limitation to this regard. Furthermore, it is crucial to underscore that achieving sub-optimal DCV drug levels could be an issue for the selection of resistance variants and the choice of future regimen options [19].  Table   Table 1. Characteristics of the HIV/HCV-coinfected patients and statistical analysis.
The last two columns contain p-value from testing the non-parametric correlation with sofosbuvir metabolite (GS-331007) and daclatasvir mean-C trough in correspondence to continuous parameters or contain median sofosbuvir and daclatasvir mean-C trough and pvalue from the comparison among categories in correspondence to categorical variables.

Supplementary Files
This is a list of supplementary files associated with the primary manuscript. Click to download. STROBE_checklist_cohort_MASTROROSA.doc