Withdrawing inactive NRTIs in HIV-1 subjects with suppressed viraemia: a randomized trial

Abstract

Introduction

Extensively pretreated HIV-infected subjects with virological failure (VF) and limited treatment options may receive salvage regimens containing NRTIs with only residual or no activity according to genotypic resistance tests.¹ This strategy has been based mainly on the hypothesis that the impaired replication capacity and the bidirectional phenotypic antagonism associated with some reverse transcriptase mutations could contribute to sustained antiviral effects in such cases.^2–4

The potential contribution of these recycled or inactive NRTIs to salvage regimens is complex and incompletely understood. There is evidence that when three active drugs can be combined in a salvage regimen, NRTIs could be safely omitted without compromising regimen failure.^5–7 On the other hand, NRTIs with moderate or high levels of resistance have shown non-inferior efficacy to fully active raltegravir when combined with lopinavir/ritonavir in subjects with VF with first-line ART of an NNRTI plus two NRTIs in a randomized trial in resource-limited settings.⁸ Inactive NRTIs, when given with a boosted PI in second-line salvage therapy, retained substantial efficacy as well, equivalent to adding a new drug class in a trial in sub-Saharan Africa.⁹

Once virological suppression is achieved, the contribution of these NRTIs to the maintenance of the ART activity remains elusive, and they could add unnecessary potential toxicity, regimen complexity and costs.

We hypothesized that withdrawing NRTIs with intermediate to complete genotypic resistance when viral load is suppressed would maintain virological efficacy.

Methods

Study design and patients

A Phase IV, multicentre, randomized, prospective, active control arm, open study was performed at 12 sites in Spain. Eligible patients (age ≥18 years) had HIV-1 infection with at least one prior VF, plasma HIV-1 RNA <50 copies/mL for ≥6 months, any CD4+ cell count, receiving a regimen with at least two active drugs (one of them an active boosted PI) and at least one NRTI without complete activity (assessed using the Spanish RIS Resistance Score) in a cumulative genotype.¹⁰ Exclusions included chronic hepatitis B, active cancer or AIDS-defining illnesses, pregnancy or breastfeeding, and foreseeable non-compliance according to the investigator's discretion. No drug changes were allowed for the last 6 months or at screening.

Ethics Committee approval was obtained in accordance with all participating centres and each patient gave written informed consent.

Randomization

Subjects were randomly allocated 1:1 to stop NRTIs without complete activity, either one or both, (experimental arm) or to maintain the treatment unchanged (control arm). Central randomization was done by telephone at the coordinating site, including stratification by receiving one or two NRTIs with incomplete activity.

Procedures

The primary endpoint was the proportion with HIV-1 RNA <50 copies/mL at week 48 (US FDA snapshot algorithm). Secondary endpoints included the rate of VF (confirmed HIV-1 RNA >200 copies/mL), treatment-emergent resistance, and incidence and severity of adverse events.

Genotypes of reverse transcriptase, protease and integrase genes (in those receiving integrase inhibitors) were analysed in all VFs in the second confirmatory sample.

Patients attended study visits at screening, baseline and then weeks 4, 16, 32 and 48.

Statistical analysis

The primary efficacy analysis used the ITT-exposed population (patients who had received at least one dose of study medication). The trial was designed to show non-inferior efficacy of the experimental arm at week 48 (non-inferiority margin: −12%). Sample size calculations assumed 80% power, a one-sided significance level of 0.025, 90% overall response rate and 5% of patients lost to follow-up. Continuous variables were compared among groups using parametric or non-parametric tests (Student's t-test, Mann–Whitney test and Wilcoxon test). For discrete variables, the χ² test or Fisher's exact test was used as appropriate. SPSS 15.0 and R software version 3.1.0 were used (http://cran.r-project.org/).

This study is registered with EudraCT: 2012-000198-21.

Results

Baseline characteristics

Of 92 subjects screened, 90 were randomized and treated (experimental, n = 45; and control, n = 45). There were no significant differences in baseline characteristics between groups (Table 1). The mean age was 50 years, 80% were male, 40% had chronic hepatitis C and the mean CD4+ cell count was 542 cells/mm³. Subjects had been extensively pretreated, having received a mean of 10 separate antiretroviral regimens and a mean of 4 PIs, and the median number of prior VFs was 3. The baseline regimen for 46 (51%) subjects contained 4 drugs and the baseline regimen for 37 (41%) subjects contained 3 drugs. Seventy-four subjects (82%) harboured the mutation M184V/I and the median number of thymidine-associated mutations was 3 (IQR: 0–4), T215X and D67X being the most frequent. In the experimental arm, thirty-two (71%) subjects removed 1 NRTI and 13 (29%) removed 2 NRTIs. Overall, 29 of 45 (64%) discontinued emtricitabine or lamivudine, and 22 of 45 (49%) discontinued tenofovir disoproxil fumarate (tenofovir DF) (Table 2).

Efficacy

In the primary study endpoint (Table 3), 41 of 45 (91.1%, experimental arm) and 44 of 45 (97.8%, control arm) had an HIV-1 RNA <50 copies/mL at 48 weeks (adjusted treatment difference: −6.7%; 95% CI: −17.4, 4.1); therefore, not meeting non-inferiority evidence. Two subjects (experimental arm) had transient detectable low-level viraemia (<200 copies/mL) and their treating physicians reintroduced the inactive NRTIs (one tenofovir DF/emtricitabine and one emtricitabine). Both subjects had HIV-1 RNA <50 copies/mL at the 48 week window. In an exploratory post-hoc analysis allowing NRTI reintroduction, rates of HIV-1 RNA <50 copies/mL were 95.6% and 97.8%, respectively (treatment difference: −2.2%; 95% CI: −7.2, 2.7). The median CD4+ T cell count remained stable in both arms. One subject was lost to follow-up in the control arm.

Drug resistance

There were no VFs in the control arm. One subject in the experimental arm had VF at week 48. Genotypic resistance tests showed mutations in the reverse transcriptase (M41L, L210W, T251Y, Y188L and G190A) and protease (L33F, M46L, I84V and L90M), all present before randomization. The integrase genotyping showed N155H and T97A. The patient failed with a regimen that included raltegravir in the study. Previously he had received regimens including raltegravir for 5 years, but all available HIV-1 RNA samples during those years were <50 copies/mL. Therefore, it cannot be excluded that integrase resistance was selected during the study failure.

Safety

Rates of discontinuation were low in both arms (2% and 2%, respectively). There were 10 grade 3/4 adverse events, 5 in each arm (none related to study drugs). In the experimental arm: 1 cirrhotic ascites, 1 hepatic encephalopathy, 1 epilepsy, 1 lung carcinoma and 1 bacterial pneumonia. In the control arm: 1 syphilis, 2 uterine cervical carcinomas, 1 hyperbilirubinaemia and 1 Giardia lamblia enterocolitis. There was one death (lung cancer, experimental arm). There were 63 grade 1/2 adverse events in the experimental arm, 3 of them potentially related to ART, and 48 in the control arm, 1 of them potentially treatment related.

Discussion

In extensively pretreated HIV-infected subjects with multidrug resistance and HIV-1 RNA below 50 copies/mL on a salvage suppressive regimen, the withdrawal of NRTIs with only residual or no activity was associated with high rates of continued virological suppression at 48 weeks. All patients were receiving a regimen with at least two active drugs (one of them an active boosted PI). Seventy-one percent of the individuals removed one NRTI and 29% two. The results not only apply for the withdrawal of emtricitabine/lamivudine, but are also relevant for the removal of NRTIs associated with probable toxicities. Forty-nine percent of the subjects discontinued tenofovir DF, which has been associated with potential long-term kidney and bone toxicity, particularly when combined with boosted PI,¹¹ or discontinued abacavir, with mixed data about its association with increased cardiovascular risk.^11,12 These potential toxicities could be prevented with this strategy.

While this antiviral NRTI activity not captured in genotypic resistance tests could have clinical interest in the initial viral decay in subjects with VF or in reducing the viral load setpoint in subjects with continued untreatable VF,^13–16 their contribution to the maintenance of HIV-1 suppression, once achieved, had not previously been explored. This is the first randomized trial to explore this strategy.

The rates of VF were low and drug resistance was uncommon. Only 1 in 90 subjects had emergent resistance in VF. The subject selected integrase resistance while receiving a regimen including raltegravir. Despite documentation of previous raltegravir exposure, we cannot exclude that raltegravir resistance was selected during the current failure.

Not unexpectedly, there were no drug-related grade 3 or 4 adverse events and there were no discontinuations due to drug toxicity.

Our results are in agreement with a preliminary single arm uncontrolled prospective series and add further knowledge to the field.¹⁷ The VERITAS study recruited only subjects with a CD4+ cell count of ≥250 cells/mm³ and receiving a regimen with at least four antiretrovirals, and all 30 patients maintained an undetectable viral load at 48 weeks. Only one NRTI was removed at baseline (lamivudine/emtricitabine in nearly all). Our series seems to suggest that the removal of two NRTIs could also be safe, and that the need of establishing a CD4 threshold limit might be unwarranted. A second single arm pilot prospective study included 31 subjects, with HIV-1 RNA suppressed for more than 12 months.¹⁸ Ninety percent of the subjects remained with suppressed viraemia at 48 weeks. NRTIs were reintroduced in two subjects with a rebound in plasma viraemia (one of them with VF) and their HIV-1 RNA was re-suppressed at 48 weeks. Our series suggests that plasma HIV-1 RNA suppression of 6 months prior to NRTI withdrawal could be enough.

Genotypically, inactive NRTIs have been commonly included in salvage regimens in subjects with VF with multidrug resistance and limited treatment options.¹⁹ Some basic science findings have prompted this prescription in the past. Reverse transcriptase mutations (mainly L74V and M184V/I), selected by nucleoside analogue chain terminators, have been shown to diminish RNA primer usage and impair the replication capacity of HIV-1.^2,20 In addition, they can delay the development of resistance to PIs and NNRTIs.²¹ Furthermore, a bidirectional phenotypic antagonism has been shown between K65R and thymidine-associated mutations, which could contribute to sustained antiviral effects in such cases when both mutation patterns coexist.^3,4

A recently reported ACTG study demonstrated in salvage therapy that omitting the NRTIs had non-inferior efficacy to their inclusion at failure.⁶ However, NRTIs were omitted either being active or inactive, and all subjects received a regimen with at least three phenotypically active drugs. Therefore, the study actually concluded that adding a fourth drug (an NRTI, either active or inactive) was not needed in these salvage regimens.

Finally, NRTIs with moderate or high levels of resistance have shown non-inferior efficacy to fully active raltegravir when combined with a boosted PI in subjects with VF with first-line ART of an NNRTI plus two NRTIs in two randomized trials in resource-limited settings.^8,9 Many subjects received a combination of zidovudine and tenofovir DF in these regimens and the bidirectional phenotypic antagonism between these drugs could have a potential role in their residual activity.

Our study has limitations. We were not able to recruit the predefined number of participants and therefore the statistical power to conclude non-inferiority was limited. Even though non-inferiority was met in a post-hoc analysis allowing NRTI reintroduction in subjects with low-level detectable viraemia and without VF, it was not a predefined subanalysis and could be prone to bias.

In summary, the removal of NRTIs with only residual or no activity has been associated with high rates of continued virological suppression at 48 weeks and appears to be a safe strategy in heavily pretreated HIV-infected subjects with multidrug resistance receiving a salvage suppressive regimen. These results contribute to our understanding of NRTI use in this difficult-to-treat population and warrant a fully powered study to confirm the non-inferiority of this strategy, with specific analyses on potential economic savings and reduction in NRTI-associated toxicity.

Funding

This work was supported by the Spanish Ministerio de Sanidad through an Independent Clinical Investigation Promotion Grant (C11-330 2012), by grant MTM2012-38067-C02-01 from the Ministerio de Economía y Competitividad from the Spanish Government, and 2014 SGR 464 from the Departament d'Economia i Coneixement of the Generalitat de Catalunya.

Transparency declarations

J. M. L. has served as an advisor or speaker for or has been awarded with grants for clinical research from Gilead Sciences, Merck Sharp & Dohme, ViiV Healthcare, Bristol-Myers Squibb and Janssen-Cilag. H. A. has received support for travel to meetings from Janssen-Cilag, Gilead Sciences, ViiV Healthcare, Abbvie and Merck Sharp & Dohme. M. J. P.-E. has received honoraria for lectures or for participation in advisory boards from Abbott, Bristol-Myers Squibb, Boehringer Ingelheim, Gilead Sciences, ViiV, MSD and Janssen-Cilag, has received payment for educational presentations for Bristol-Myers Squibb, ViiV and Abbott Laboratories, and has received unrestricted grants from Abbott, ViiV, Gilead and Janssen-Cilag. A. I. has received honoraria for lectures, consultancies or funds for research from Abbvie, Boehringer Ingelheim, Bristol-Myers Squibb, Gilead Sciences, Janssen-Cilag, Merck Sharp & Dome and ViiV Healthcare, and honoraria for educational activities from Gilead Sciences, Janssen-Cilag and Merck Sharp & Dome. J. B. has served as an advisor or speaker for Merck Sharp & Dohme, ViiV Healthcare, Bristol-Myers Squibb and Janssen-Cilag. All other authors: none to declare.

Author contributions

Study concept and design: J. M. L. Acquisition of data: J. T. Analysis and interpretation of data: J. M. L., J. T. and N. P.-A. Drafting of the manuscript: J. M. L. Critical revision of the manuscript for important intellectual content: H. A., A. A., A. P., M. J. P.-E., A. I., M. M., J. B. and B. C. Statistical expertise: N. P.-A. All authors approved the final version of the manuscript. J. M. L. and J. T. had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Acknowledgements

This work was presented in part at the Twenty-second Conference on Retroviruses and Opportunistic Infections, Seattle, WA, USA, 2015 (Abstract 553).

We wish to thank the study patients and their families for their participation in the study.

Members of the Nuke-Out Study

Josep M. Llibre, Isabel Bravo, Jessica Toro, Aintzane Ayestaran, Núria Pérez-Alvarez, Bonaventura Clotet, Daniel Podzamczer, Arkaitz Imaz, Pere Domingo, Irene Fernández, Manel Crespo, Joaquin Burgos, Silvia Valero, Mar Masià, Maria Jesús Pérez-Elías, Alberto Díaz, Miguel Yzusqui, Ana Isabel Mariño, Hortensia Alvarez, Antonio Antela, José Hernández-Quero, Antoni Payeras and Juan González Moreno.

References

Author notes