Analysing the efficacy and tolerability of dolutegravir plus either rilpivirine or lamivudine in a multicentre cohort of virologically suppressed PLWHIV

Abstract

Introduction

Two-drug regimens (2DRs) with dolutegravir, a second-generation integrase inhibitor (INI), are among the suggested switch strategies for virologically suppressed people living with HIV (PLWHIV) in international guidelines^1,2 and have become widely utilized in clinical practice. In particular, the associations of dolutegravir with either lamivudine or rilpivirine have demonstrated high efficacy and optimal tolerability as switch strategies in clinical trials and observational studies, with a favourable effect on metabolic profile and lower toxicity compared with ‘standard’ three-drug regimens with a two-NRTI backbone.^3–7 In our multicentre cohort, we aimed to investigate and compare the efficacy and safety of dolutegravir + lamivudine versus dolutegravir + rilpivirine.

Methods

We retrospectively analysed a multicentre cohort of treatment-experienced PLWHIV switching to either dolutegravir + lamivudine (lamivudine group) or dolutegravir + rilpivirine (rilpivirine group) between January 2015 and December 2021. Inclusion criteria were: virological suppression at baseline (i.e. HIV-RNA <50 copies/mL at the start of the study regimen); age ≥18 years, absence of HBV coinfection; absence of resistance mutations to any of the molecules of the prescribed regimen; and informed consent to data collection. To be sure that no subject had any mutation to study drugs, we decided to exclude data from PLWHIV with no prior resistance mutation test available. Kaplan–Meier survival analysis was used to estimate time to virological failure (VF, defined as two consecutive HIV-1 RNA values ≥50 copies/mL with at least one of them being ≥200 copies/mL or a single HIV-1 RNA ≥1000 copies/mL) and time to treatment discontinuation (TD), defined as the interruption of any of the study drugs or the regimen intensification, and Cox regression analysis to evaluate predictors. Censor was established at the last available visit or death. Changes in immunological and metabolic parameters were assessed by non-parametric tests and their predictors by linear regression analysis. The study was approved by each local Ethics Committee (protocol number of the promoter centre: 5284/15).

Results

Five hundred and ninety-two PLWHIV were enrolled: 306 in the lamivudine group and 286 in the rilpivirine group. In our cohort, the majority of PLWHIV were male (425; 71.8%), with a median age of 51.8 years (IQR 43.5–58.0), a median time from HIV diagnosis of 13.3 years (IQR 7.4–21.7) and a median time of antiretroviral exposure of 10.4 years (IQR 5.9–17.9). Full population characteristics and comparisons between groups are summarized in Table 1.

During 660.3 patient-years of follow-up (PYFU) we observed 8 VFs in the lamivudine group, a rate of 1.2 VF per 100 PYFU. In the rilpivirine group, during 656.3 PYFU we registered 3 VFs, a rate of 0.4 VF per 100 PYFU. PLWHIV who incurred VF did not present evidence of new resistance mutations at a subsequent genotypic analysis. Table 2 summarizes the characteristics of PLWHIV experiencing VF.

The estimated probability of maintaining virological suppression at 144 and 240 weeks were 96.6% and 92.7%, respectively, in the lamivudine group and 98.7% and 98.7%, respectively, in the rilpivirine group (log-rank P = 0.172). No predictors of VF were observed in either group.

In a specific subanalysis analysing peak HIV-RNA, in the lamivudine group the estimated probabilities of remaining free from VF were 97.3% at Week 144 and 94.5% at Week 240 in the group with peak HIV-RNA <500 000 copies/mL versus 87.5% at Week 144 and 70.0% at Week 240 in the group with higher peak viral load (log-rank P = 0.156). In the lamivudine group, PLWHIV with peak HIV-RNA ≥500 000 copies/mL rate of VF was 3.3 per 100 PYFU compared with a rate of 1.0 per 100 PYFU for those with a lower peak HIV-RNA. Similarly, in the rilpivirine group, the probability of remaining free from VF was 98.9% at Week 144 and 98.9% at Week 240 in the group with a lower peak viral load while in PLWHIV with peak HIV-RNA ≥500 000 copies/mL the probability was 96.2% at Week 144 and 96.2% at Week 240 (log-rank P = 0.259). In this group, VF rate for individuals with a peak HIV-RNA <500 000 copies/mL was 0.4 per 100 PYFU compared to 1.3 per 100 PYFU for PLWHIV with a zenith HIV-1 RNA ≥500 000 copies/mL.

Regarding discontinuations, we observed 32 TDs in the lamivudine group during 665.8 PYFU (a rate of 4.8 TD per 100 PYFU), with an estimated probability of remaining in the study regimen of 83.3% at Week 144 and 82.3% at Week 240. Reasons for discontinuation were represented by VF (4 of 306; 1.3%), hypersensitivity reaction (1; 0.3%), CNS toxicity (7; 2.3%), other toxicities (6; 2.0%), switch to a single-tablet regimen (STR) (7; 2.3%), pregnancy (1; 0.3%) and other/unspecified reasons (8; 2.6%). In the rilpivirine group, during 658.7 PYFU, we observed 17 TDs (2.6 per 100 PYFU), due to VF (2 of 286; 0.7%), neuropsychological events (3; 1.0%), other toxicities (3; 1.0%) and other/unknown causes (7; 2.4%). In this group, the estimated probabilities of continuing study treatment were 92.0% and 85.9% at Weeks 144 and 240, respectively (log-rank P value between groups = 0.018). A multivariate model for Cox regression was performed, adjusted for statistically different parameters at baseline between groups as well as time of virological suppression and zenith HIV-1-RNA; no predictors of TD were found.

Regarding immunological parameters, we observed a significant improvement in CD4+ cell count at Week 240 in the lamivudine group (median increase +65 cell/mm³; P = 0.012); the only predictor confirmed at a multivariate analysis was baseline CD4+ cell count (per 10 cells/mm³ more, B −3.6, 95% CI −6.5 to −0.7; P = 0.016). In the rilpivirine group we registered a non-significant increase in total CD4+ cell count (median +73; P = 0.056) and a significant increase in CD4/CD8 ratio (median +0.2; P = 0.014).

Finally, regarding renal function, individuals in the lamivudine group presented a significant increase in serum creatinine value at Week 240 (median +0.12 mg/mL; P < 0.001), while PLWHIV in the rilpivirine group did not show significant changes. Baseline creatinine value was the sole predictor of changes in the lamivudine group (per 1 mg/mL more, B −0.24, 95%CI −0.35 to −0.12; P < 0.001). At baseline, median serum creatinine was significantly higher in the rilpivirine group compared with the lamivudine group (0.93 mg/dL versus 0.90 mg/dL; P = 0.023).

Discussion

In our multicentre cohort, we found a similar virological efficacy of dolutegravir + lamivudine and dolutegravir + rilpivirine, with a low rate of VF and no mutations emerging in failing PLWHIV, further confirming the safety of both switch strategies.

In previous work from our cohort^7,8 we observed a trend towards an increased risk of VF for PLWHIV with an HIV-RNA peak ≥500 000 copies/mL; we speculated that, since zenith HIV-RNA is correlated to HIV-DNA,⁹ a marker for HIV latent reservoir predictive of treatment response, PLWHIV with a higher zenith may have a significantly higher HIV-DNA load, that could, in turn, justify a higher risk of viral rebound during treatment with a 2DR. In this work, however, we failed to observe the same association between peak HIV-RNA and VF in this analysis for either analysed regimen. Notably, in accordance with the most recent US Department of Health and Human Services (DHHS) guidelines,² in this work we applied a less strict definition of VF compared with previous works from our cohort,^7,8 thus explaining the different results observed.

Although a higher risk of TD was reported for individuals in the lamivudine group, it is to be noted that seven TDs in this group (7/32; 21.8%) were due to a switch to an STR; in fact, our cohort has collected data on dolutegravir-containing regimens since 2016,¹⁰ and it is to be noted that an STR with dolutegravir/rilpivirine was introduced long before the STR with dolutegravir/lamivudine was introduced, since the disparity in terms of TD due to switch to an STR. Interestingly, the rate of TD due to neuropsychological events was below 2.5% in both groups. This threshold is lower than those observed from other studies analysing discontinuations of dolutegravir-based regimens due to neuropsychiatric events.^11,12

Our secondary analysis on immunological outcome at 240 weeks shows that both regimens are associated with an improvement in the long term. These data are reassuring and contribute to the existing literature since long-term, real-life data on immunological improvement with dolutegravir-based 2DRs are still scarce.

Regarding renal function, it is well documented that dolutegravir is associated with an apparent increase in serum creatinine, attributed to the fact that it inhibits the organic cation transporter 2 (OCT2), reducing the excretion at the tubular level of creatinine.¹³ In our study, however, we observed a disparity between the two groups, with individuals in the lamivudine group showing an increasing trend in serum creatinine while PLWHIV in the rilpivirine group did not present significant changes at Week 240. This disparity may be explained by a significant difference of baseline creatinine between the two groups; in fact, median serum creatinine was higher in the rilpivirine group, hence the drug-related increase could have been less pronounced compared with the lamivudine group.

Our study’s main limitations are the retrospective design and the possibility of study bias due to differences in patients’ baseline characteristics between groups. However, our study also presents major strengths such as the long follow-up time, the sample size and the availability of genotypic tests for all included PLWHIV, which allowed us to include the subjects as close as possible to the labelled indications for both regimens.

In conclusion, our study shows that both dolutegravir + lamivudine and dolutegravir + rilpivirine are effective and safe as switch strategies in clinical practice, with a low cumulative incidence of VF and a favourable immunological recovery, even in the long term.

Funding

This study was supported by ViiV Healthcare Europe, which provided an unrestricted grant to the ODOACRE cohort. No additional external funding was received for this study. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Transparency declarations

A. Ciccullo received travel grants from ViiV Healthcare and at the time of writing is a shareholder in Bavarian Nordic. A. Capetti has received a personal grant from AB, Gilead and ViiV. G. Sterrantino has received funds for speaking by Gilead, Merk, Janssen, AbbVie and ViiV. C. Mussini has participated in advisory boards, received study grants and/or speaker honoraria from: AbbVie, Gilead, ViiV, Janssen, Angelini, BMS and MSD. S. Di Giambenedetto was a paid consultant or member of advisory boards for Gilead, ViiV Healthcare, Janssen-Cilag, Merck Sharp & Dohme and Bristol-Myers Squibb. All other authors: none to declare.

References