Liver stiffness change with HCV cure in HIV-infected patients on non-nucleoside analogues

Abstract

Background

Liver stiffness (LS) at sustained viral response (SVR) is strongly associated with a lower incidence of subsequent hepatic events. HIV NNRTIs may have a beneficial impact on fibrogenesis.

Objectives

Our aim was to analyse the influence of NNRTI-based therapy on the change in LS from starting direct-acting antiviral (DAA) therapy to achieving SVR in HIV/HCV-coinfected patients.

Methods

Three hundred and thirteen HIV/HCV-coinfected patients who fulfilled the following criteria were included: (i) had achieved SVR with an IFN-free, DAA-including regimen; (ii) LS ≥9.5 kPa before therapy; (iii) LS measurement available at SVR; (iv) seronegative for HBsAg; and (v) ART containing 2 NRTIs plus either 1 NNRTI or 1 integrase inhibitor (INI) or 1–2 NRTIs plus 1 PI. LS changes were assessed.

Results

Seventy-four patients received NNRTI-based combinations [53 (71.6%) rilpivirine and 16 (21.6%) efavirenz] and 239 patients received other regimens. At baseline, the median (IQR) LS was 16.7 kPa (11.8–25.6) in the NNRTI group and 17.3 kPa (11.9–27.4) in the non-NNRTI group (P = 0.278). The median (IQR) percentage of LS decrease from baseline to SVR was 35.2% (18.2%–52.3%) for NNRTI-based therapy and 29.5% (10%–45.9%) for PI- or INI-based therapy (P = 0.018). In multivariate analysis, adjusted for sex, age, HCV genotype, NRTI backbone and propensity score for HIV therapy, NNRTI-based regimen use was associated with a higher LS decrease [β = 11.088 (95% CI = 1.67–20.51); P = 0.021].

Conclusions

Treatment with NNRTI plus 2 NRTI combinations is associated with a higher LS decline than other ART combinations in HIV/HCV-coinfected patients receiving DAA-based therapy.

Introduction

Achieving sustained viral response (SVR) with direct-acting antiviral (DAA)-based therapy is associated with a dramatic decrease in the risk of liver complications in HIV-infected patients.^1,2 However, a reduced number of liver complications still may occur, even in the long-term^3,4 In this regard, in HIV/HCV-coinfected patients who achieve SVR after treatment with DAAs, liver stiffness (LS) measured by vibration-controlled transient elastography (VCTE) at SVR is a marker of subsequent hepatic outcome.^5,6 Unfortunately, the factors underlying LS change with DAA-based therapy are incompletely understood.

The role of ART in reducing or contributing to liver fibrosis in persons with HIV/HCV is still unclear. Some ART combinations, especially NRTIs, could cause several detrimental effects.^7–10 Conversely, it has recently been observed that rilpivirine, a second-generation NNRTI, could have a beneficial effect on liver fibrogenesis, thus promoting liver regeneration both in vitro and in vivo in a murine model.¹¹ Hence, NNRTIs may have a significant effect on change in LS from starting DAA-based therapy to SVR achievement. However, there are no data available about this issue.

Therefore, our aim was to analyse the impact of concomitant NNRTI-based ART, compared with other ART, on the change of LS from starting DAA to achieving SVR in HIV/HCV-coinfected patients treated with IFN-free, DAA-based regimens.

Patients and methods

Study population

In this prospective multisite study, HIV/HCV-coinfected individuals were recruited from the GEHEP-011 cohort (clinicaltrials.gov ID = NCT04460157). The inclusion criteria in this cohort are: (i) SVR achieved 12 weeks after a DAA-based regimen, either with or without pegylated-IFN; (ii) LS measurement before starting DAA therapy ≥9.5 kPa; (iii) LS measurement available at the SVR timepoint; and (iv) seronegative for HBsAg. For this study, individuals were included if their ART regimen during the HCV treatment period included 2 NRTIs plus 1 NNRTI, 2 NRTIs plus 1 integrase inhibitor (INI) or 1–2 NRTIs plus 1 PI. Individuals were excluded from the study if pegylated-IFN was given as a part of the DAA-based regimen.

Follow-up

The baseline timepoint was the date that DAA therapy was started. All participants were evaluated under a common protocol at each participating centre. LS was assessed by VCTE (FibroScan^®, Echosens, Paris, France), according to a standardized procedure,¹² within the 30 days before starting DAA therapy and at the day of SVR. According to prior studies conducted in HIV/HCV-coinfected patients, subjects showing baseline LS >14 kPa were considered as cirrhotics.^6,13

Statistical analyses

The main outcome variable was the percentage of LS change from baseline to SVR. Additionally, achievement of LS values <14 kPa at SVR in those patients with pretreatment cirrhosis was analysed as a secondary outcome variable. Variables associated with LS change from baseline to the date of SVR evaluation in the univariate analysis with a P value ≤0.1, along with age and sex, were entered in multivariate linear regression models. Similarly, variables associated with achieving an LS value <14 kPa at SVR in patients with pretreatment cirrhosis with a P value ≤0.1, as well as age and sex, were entered in a multivariate logistic regression analysis. In order to control potential confounding bias due to differences between the NNRTI-based group and the other ART group at baseline, propensity score (PS) analysis was performed. First, variables that were significantly different in patients on NNRTIs versus other ART at baseline were included as covariables in a logistic regression model. Treatment group was considered as the dependent variable in this model. The PS value thus obtained was also included as a further covariable in the regression models that aimed to identify factors independently associated with the study outcome variables. All data analyses were performed using the SPSS statistical software package release 25.0 (IBM, Armonk, NY, USA).

Ethics

Both the study design and development complied with the Helsinki declaration and were approved by the local Ethics Committee of the Hospital Universitario Virgen de Valme (Seville) (reference = 1612-N-17). All patients gave their written informed consent when entering the cohort.

Results

Patient characteristics

Three hundred and thirteen patients were included in this study. Of them, 74 (23.6%) were on NNRTI regimens and 239 (76.4%) were receiving PI- or INI-based regimens. See Table 1. Patients on NNRTI-based regimens had lower alcohol consumption, higher CD4+ cell counts and showed more commonly undetectable plasma HIV-RNA load compared with those treated with drugs other than NNRTIs. ART regimens are shown in Table S1 (available as Supplementary data at JAC Online). DAA therapy for both groups is show in Table S2.

Table 1.

Baseline characteristics of the patients (N = 313)

Variable	Treated with NNRTI (n = 74)	Treated with PI or INI (n = 239)	P value
Female, n (%)	13 (17.6)	30 (12.6)	0.333
Age (years), median (IQR)	52.2 (47.6–55.5)	51.8 (48.4–54.9)	0.992
People who inject drugs, n (%)	67 (90.5)	192 (80.3)	0.042
Alcohol >50 g/day, n (%)	1 (1.4)	15 (8.3)	0.038
CD4 (cells/mm3), median (IQR)	665 (418–825)	527 (306–776)	0.023
HIV RNA <50 copies/mL, n (%)	71 (95.9)	193 (80.8)	0.003
CD4 nadir (cells/mm3), median (IQR)	193 (70–330)	177 (67–329)	0.668
HCV genotype 3, n (%)	9 (12.2)	45 (18.8)	0.185
Median LS (kPa), median (IQR)	16.7 (11.8–25.6)	17.3 (11.9–27.4)	0.322
LS <14 kPa, n (%)	30 (40.5)	83 (34.7)	0.363
Backbonea, n (%)
tenofovir disoproxil fumarate or tenofovir alafenamide/emtricitabine	54 (73)	113 (47.3)	0.001
other NRTI	20 (27)	126 (52.7)

a

Backbone: NRTI backbone based on tenofovir disoproxil fumarate or tenofovir alafenamide/emtricitabine versus other NRTI backbone.

The comparisons of variables were carried out using the χ² test or Fisher’s exact test when appropriate for categorical variables and Student’s t-test or the Mann–Whitney U-test when necessary for continuous variables.

LS change from baseline to SVR

The percentage of LS decline was 35.2% (18.2%–52.3%) in the NNRTI-based regimen group and 29.5% (0%–45.9%) for patients treated with a PI- or INI-based regimen (P = 0.018) (Figure S1).

The percentage of LS decline among patients with ART including 2 NRTIs plus NNRTI was higher than that observed among those receiving 1–2 NRTIs plus PI or 2 NRTIs plus INI (Figure S2). No significant differences in baseline LS, in LS at SVR or in the percentage of LS decline from baseline to SVR were found according to specific NNRTIs included in ART combinations (Table S3).

NNRTI-based regimen and NRTI backbone were the only variables associated with the percentage of LS change from baseline to SVR in the univariate analyses (Table 2). The PS for NNRTI-based treatment was calculated on the basis of a logistic regression model, where age, sex, drug injection, HCV genotype 3 infection, CD4 count, daily alcohol intake, plasma HIV-RNA <50 copies/mL and NRTI backbone were included as covariables.

Table 2.

Variables associated with the percentage of LS change

Variable	n (%)	Median percentage of LS change (IQR)	Bivariate P value	β (95% CI)	Multivariate P value
Female
yes	43 (13.7)	33.5 (13.1–50.1)	0.503	0.994 (−10.57 to 12.56)	0.866
no	270 (86.3)	30.1 (12.4–46.1)
Age >52 years
yes	150 (47.9)	30.31 (10.3–48.9)	0.862	−0.522 (−1.22 to 0.18)	0.142
no	163 (52.1)	36.6 (14.2–51)
People who inject drugs
yes	259 (82.7)	30.1 (12–45)	0.139
no	54 (17.3)	36.6 (14.2–51)
Alcohol >50 mg/day
yes	16 (5.1)	29.1 (7.3–41.3)	0.486
no	236 (75.4)	32.4 (14–48.8)
CD4 >563 cells/mm3
yes	139 (50.2)	29.9 (10.3–43.4)	0.353
no	138 (49.8)	31.2 (13.9–48.3)
HIV RNA <50 copies/mL
yes	264 (84.3)	30.3 (11–47)	0.546
no	24 (7.7)	33.7 (18.4–49.3)
CD4 nadir >181 cells/mm3
yes	122 (39)	32.1 (14.1–49.2)	0.362
no	122 (39)	30.5 (10.8–44.1)
HCV genotype 3
yes	54 (17.3)	37.3 (18.2–54.3)	0.053	12.311 (1.18 to 23.44)	0.030
no	259 (82.7)	30 (10.7–45.9)
NNRTI
yes	74 (23.6)	35.2 (18.2–52.3)	0.018	11.088 (1.67 to 20.51)	0.021
no	239 (76.4)	29.5 (10–45.9)
Backbonea
tenofovir disoproxil fumarate or tenofovir alafenamide/emtricitabine	167 (53.4)	35.3 (14.4–50.4)	0.011	6.852 (−2.76 to 16.47)	0.162
other NRTI	146 (46.6)	27.1 (5.9–43.4)

a

Backbone: NRTI backbone based on tenofovir disoproxil fumarate or tenofovir alafenamide/emtricitabine versus other NRTI backbone.

Continuous variables were categorized by the median value or clinically relevant cut-off points. Age was entered as a continuous variable into the linear regression model. The comparisons of variables were carried out using the Mann–Whitney U-test for continuous variables.

In a multiple linear regression analysis, including sex, age, HCV genotype 3 infection, NRTI backbone and PS, an independent relationship between NNRTI-based regimen and the percentage of LS change was observed (Table 2).

In an additional subanalysis including only patients with cirrhosis (LS ≥14 kPa) at baseline (n = 200), a multiple logistic regression analysis, adjusted for sex, age and PS, was performed; NNRTI-based regimens were not associated with achievement of LS values <14 kPa at SVR [OR = 1.731 (95% CI = 0.76–3.94); P = 0.191]. See Table S4.

Discussion

This study shows that treatment with NNRTI-based regimens is associated with a higher LS change than other combinations in HIV/HCV-coinfected patients receiving DAA-based therapy. Given that LS after HCV treatment with DAA-based therapy strongly predicts subsequent clinical outcome in HIV/HCV-coinfected patients,⁶ NNRTI-based ART might lead to a lower incidence of ensuing hepatic complications in subjects achieving SVR.

LS, an indirect marker of liver fibrosis, inflammation and portal pressure, declines with HCV treatment.^14,15 However, in subjects with LS >14 kPa at SVR, a non-negligible incidence of liver complications occurs, even in the long-term.¹⁶ Therefore, therapeutic strategies associated with greater LS decline may help to prevent the development of subsequent hepatic complications.¹⁷ A possible explanation of the mechanisms behind the association between NNRTI-based therapy and greater reduction of LS is a selective STAT1-dependent induction of apoptosis of hepatic stellate cells exerted by rilpivirine.¹¹ This may have paracrinal effects in the hepatocytes and could promote liver regeneration. In this regard, most patients receiving an NNRTI in our study were on a rilpivirine-based regimen.

In the subanalysis focusing on patients with pretreatment cirrhosis, subjects taking NNRTI-based combinations tended to reach more commonly LS <14 kPa at SVR, a proven predictor of better liver outcome, but statistically significant differences were not reached, probably due to insufficient statistical power.

The observational nature of this study represents a limitation, as this may lead to bias in the distribution of treatment combinations. The use of PS for received NNRTI combinations counterbalances this potential bias. In this regard, a higher proportion of patients in the NRTI group received tenofovir disoproxil fumarate or tenofovir alafenamide/emtricitabine as the NRTI backbone. This was expected, as efavirenz or rilpivirine are usually used in single-tablet regimens coformulated with tenofovir disoproxil fumarate or tenofovir alafenamide/emtricitabine. Tenofovir disoproxil fumarate or tenofovir alafenamide/emtricitabine use was associated with a higher percentage of LS change at SVR. Consequently, NNRTI use and LS change could be confounded by the tenofovir disoproxil fumarate or tenofovir alafenamide/emtricitabine NRTI backbone. However, after adjustment for treatment differences by PS, tenofovir disoproxil fumarate or tenofovir alafenamide/emtricitabine as the NRTI backbone was not independently associated with LS reduction. Taking into account PS and the tenofovir disoproxil fumarate or tenofovir alafenamide/emtricitabine backbone, the association between NNRTI use and LS change at SVR was confirmed. A second limitation could be the length of follow-up, which was not long enough to analyse the impact on the emergence of liver complications occurring later on. However, the predictive value of LS at the date of SVR evaluation has been proven.⁶ Large randomized controlled trials could provide stronger evidence on this issue, but such studies are unlikely to be undertaken, given the decreasing number of HIV/HCV-coinfected patients who are receiving DAA-based therapy nowadays and the long follow-up the detection of post-SVR liver events would require.

In conclusion, our results show that the use of NNRTI-based regimens is associated with a greater decline of LS with DAAs in HIV/HCV-coinfected patients. Because of this, the use of NNRTI-based therapy in these patients could ultimately contribute to reducing the incidence of liver events in this setting, as well as to reducing the need for surveillance programmes for specific liver complications. Further studies on this issue to assess the long-term impact of differences in LS decline associated with NNRTI-based therapy are needed.

Acknowledgements

This study would not have been possible without the collaboration of all patients, medical and nursing staff and data managers who took part in the study. We want to thank other members of the RISHEP13 and GEHEP-011 study groups who collaborated in this study.

Funding

This work was supported by the Ministry of Science, Innovation and Universities of Spain for A.G.-S. and A.R.-J. who are recipients of Miguel Servet Research Contracts (CP18/00146 and CP18/00111), co-financed by the FSE ‘El Fondo Social Europeo invierte en tu futuro’. A.C.-G. has received a Río Hortega grant from the Instituto de Salud Carlos III (CM19/00251). M.F. is the recipient of a postdoctoral perfection grant from the Spanish Ministry of Science, Innovation and Universities of Spain (CD18/00091). J.M. is the recipient of a grant from the Servicio Andaluz de Salud de la Junta de Andalucia (B-0037). J.A.P. has received a research extension grant from the Programa de Intensificación de la Actividad de Investigación del Servicio Nacional de Salud Carlos III (3SNS).

Transparency declarations

A.C.-G. has received lecture fees from Gilead. F.J.V.-M. has received lecture fees and grants from Gilead, MSD and Janssen-Cilag. D.M. has received lecture fees from ViiV Healthcare, MSD, Gilead and Janssen, has received consulting fees from ViiV Healthcare and has been an investigator in clinical trials supported by GlaxoSmithKline. A.I. has received financial compensation for lectures, consultancy work and educational activities and funds for research from Gilead Sciences, Janssen-Cilag, Merck Sharp & Dohme and ViiV Healthcare. J.M. has been an investigator in clinical trials supported by Bristol-Myers Squibb, Gilead and Merck Sharp & Dome and has received lecture fees from Gilead, Bristol-Myers Squibb and Merck Sharp & Dome and consulting fees from Bristol Myers-Squibb, Gilead and Merck Sharp & Dome. J.A.P. has received consulting fees from Bristol-Myers Squibb, AbbVie, ViiV Healthcare, Gilead, MSD and Janssen Cilag, has received research support from Bristol-Myers Squibb, AbbVie, ViiV Healthcare, Janssen Cilag, MSD and Gilead and has received lecture fees from AbbVie, Bristol-Myers Squibb, Janssen Cilag, ViiV Healthcare, MSD and Gilead. All other authors: none to declare.

Supplementary data

Tables S1 to S4 and Figures S1 and S2 are available as Supplementary data at JAC Online.