Systematic Literature Review and Feasibility Analysis

A systematic literature review was undertaken to identify clinical trials conducted in the HTE population published from January 1, 2003, to February 23, 2021. The literature search was informed by the inclusion criteria defining the HTE population for the BRIGHTE trial and is described fully in Supplemental Appendix Section 1.4. This study complied with all applicable laws regarding participant privacy. Analyses omitted participant identification, and no direct contact or collection of primary data from participants occurred. All trials included in the analysis obtained institutional review board or ethics committee approval before initiation. Because this study only used preexisting, deidentified participant data, institutional review board approval was not required.

Overall, 78 individual articles reporting 52 clinical trials met the inclusion criteria for the review. After excluding BRIGHTE, 51 trials underwent data extraction and were subsequently filtered based on suitability for informing the relevant comparator groups in the MAIC analyses. Factors considered when filtering studies for suitability included the following: (1) availability of key outcome data and alignment of study definitions; (2) expert clinical opinion on the comparability of trial populations with the HTE population recruited into BRIGHTE; (3) how well the ARV regimens in the comparator studies represented OBT comparable to that used in BRIGHTE; and (4) alignment of trial inclusion and exclusion criteria and reported drug resistance data (resistance to ≥3 ARV classes) or susceptibility data (genotypic susceptibility score [GSS], phenotypic susceptibility score [PSS], or overall susceptibility score [OSS]) with those of BRIGHTE. Susceptibility scores have been described in detail previously13Ackerman P Thompson M Molina J-M et al.Long-term efficacy and safety of fostemsavir among subgroups of heavily treatment-experienced adults with HIV-1. and in Supplemental Appendix Section 1.5. Briefly, susceptibility scores are based on a participant's susceptibility ratings for each ARV. A score of 0 indicates resistance to an ARV, 0.5 indicates partial activity, and 1 indicates full activity. These resistance ratings are summed to provide insight into resistance for each participant. In the absence of a universal optimized regimen for comparison (ie, in this case optimized regimen reflects a heterogeneous group of regimens tailored to individual needs), multiple trials were considered for inclusion in the MAIC analyses, each with a unique set of limitations.Three studies and respective comparators were identified for use in the MAIC analyses: TMB-301 (ibalizumab and OBT), BENCHMRK-1/-2 (OBT alone), and VIKING-3 (OBT alone).10Emu B Fessel J Schrader S et al.Phase 3 study of ibalizumab for multidrug-resistant HIV-1.,14Eron JJ Cooper DA Steigbigel RT et al.Efficacy and safety of raltegravir for treatment of HIV for 5 years in the BENCHMRK studies: final results of two randomised, placebo-controlled trials., 15Steigbigel RT Cooper DA Kumar PN et al.Raltegravir with optimized background therapy for resistant HIV-1 infection., 16Steigbigel RT Cooper DA Teppler H et al.Long-term efficacy and safety of raltegravir combined with optimized background therapy in treatment-experienced patients with drug-resistant HIV infection: week 96 results of the BENCHMRK 1 and 2 phase III trials., 17Castagna A Maggiolo F Penco G et al.Dolutegravir in antiretroviral-experienced patients with raltegravir- and/or elvitegravir-resistant HIV-1: 24-week results of the phase III VIKING-3 study. In BRIGHTE, OBT primarily included dolutegravir, darunavir, and tenofovir disoproxil fumarate.5Lataillade M Lalezari JP Kozal M et al.Safety and efficacy of the HIV-1 attachment inhibitor prodrug fostemsavir in heavily treatment-experienced individuals: week 96 results of the phase 3 BRIGHTE study. Complete information on OBT regimens is not available for all comparator studies. The OBT in TMB-301 was likely comparable to that used in BRIGHTE because it was conducted contemporaneously with BRIGHTE and TMB-301 had mostly similar recruitment criteria and participant characteristics.10Emu B Fessel J Schrader S et al.Phase 3 study of ibalizumab for multidrug-resistant HIV-1. In BENCHMRK, 58% of participants had ≥1 protease inhibitor in their antiretroviral therapy regimen, and dolutegravir was not yet available.15Steigbigel RT Cooper DA Kumar PN et al.Raltegravir with optimized background therapy for resistant HIV-1 infection. The OBT used for comparison with VIKING-3 was more similar to that in BRIGHTE and primarily included dolutegravir, darunavir, and tenofovir disoproxil fumarate/emtricitabine.17Castagna A Maggiolo F Penco G et al.Dolutegravir in antiretroviral-experienced patients with raltegravir- and/or elvitegravir-resistant HIV-1: 24-week results of the phase III VIKING-3 study. Notably, unlike typical indirect comparisons in which comparators are prespecified and evidence synthesis directed toward a specific treatment, in this study, the OBT-alone comparator, reflective of optimized regimens in the absence of ibalizumab, could refer to a wide range of treatment regimens. Furthermore, the diversity in the participants’ resistance profiles added complexity to determining the expected activity of these regimens and matching based on these characteristics. Susceptibility scores (GSS, PSS, and OSS) were used as an indication of existing resistance, and consequently the expected activity of OBT and contribution to the overall efficacy of the regimen, but have limitations in terms of how predictive they are of outcomes.18Van Laethem K Vandamme A-M. Interpreting resistance data for HIV-1 therapy management—know the limitations. Therefore, the systematic literature review focused on identifying treatments in patient populations comparable to BRIGHTE (eg, participants’ treatment experience, resistance profile, and available remaining therapies). In the absence of a definitive comparator, 2 comparator trials were selected to inform the OBT-alone scenario (BENCHMRK-1/-2 and VIKING-3). Key study characteristics and inclusion criteria are presented in Table I.

Table ISummary of key study characteristics and inclusion criteria.

ARV = antiretroviral; FDA, US Food and Drug Administration; INSTI = integrase strand transfer inhibitor; OBT = optimized background therapy.

MAIC AnalysisThe BRIGHTE trial consisted of 2 cohorts, randomized and nonrandomized, reflecting underlying patient needs (the nonrandomized cohort had more extensive resistance and no fully active ARV options available).5Lataillade M Lalezari JP Kozal M et al.Safety and efficacy of the HIV-1 attachment inhibitor prodrug fostemsavir in heavily treatment-experienced individuals: week 96 results of the phase 3 BRIGHTE study. To maximize sample size of the analysis population, the 2 cohorts (intention-to-treat–exposed [ITT-E]) were combined to represent the full BRIGHTE population eligible for fostemsavir. Notably, neither TMB-301 nor VIKING-3 included participants comparable to those in the nonrandomized cohort from BRIGHTE, but a similar group of participants was included in BENCHMRK.10Emu B Fessel J Schrader S et al.Phase 3 study of ibalizumab for multidrug-resistant HIV-1.,15Steigbigel RT Cooper DA Kumar PN et al.Raltegravir with optimized background therapy for resistant HIV-1 infection.,17Castagna A Maggiolo F Penco G et al.Dolutegravir in antiretroviral-experienced patients with raltegravir- and/or elvitegravir-resistant HIV-1: 24-week results of the phase III VIKING-3 study.

If key study definitions differed among trials (for inclusion criteria, matching variables, or outcomes), efforts were made to align these between the comparator trial and BRIGHTE, as described below and in detail in Supplemental Appendix Sections 2.2, 2.3, and 2.4. Differences in inclusion criteria, particularly those regarding viral load, were accounted for through exclusion of relevant BRIGHTE participants before the matching process.

The TMB-301 study was a single-group, open-label, Phase III clinical trial used to inform the ibalizumab and OBT comparator.10Emu B Fessel J Schrader S et al.Phase 3 study of ibalizumab for multidrug-resistant HIV-1. This trial enrolled participants contemporaneously with the BRIGHTE trial, and the recruitment criteria and participant characteristics were mostly similar; in the absence of published details of the OBT used in TMB-301, we assume that this was comparable to BRIGHTE. Notably, 17 of the 40 participants (43%) in the TMB-301 trial were receiving fostemsavir in their OBT, and 15 of 371 BRIGHTE participants (4%) were receiving ibalizumab in their OBT.The BENCHMRK-1/-2 studies were selected as 1 of the OBT-alone comparator study candidates because the patient population was closely aligned with the BRIGHTE population. The GSS and PSS indicated that most participants had ≤2 ARV classes remaining and resistance to ≥3 ARV classes,14Eron JJ Cooper DA Steigbigel RT et al.Efficacy and safety of raltegravir for treatment of HIV for 5 years in the BENCHMRK studies: final results of two randomised, placebo-controlled trials.–16Steigbigel RT Cooper DA Teppler H et al.Long-term efficacy and safety of raltegravir combined with optimized background therapy in treatment-experienced patients with drug-resistant HIV infection: week 96 results of the BENCHMRK 1 and 2 phase III trials. and BENCHMRK included an OBT-alone group in an HTE population. However, BENCHMRK, which evaluated raltegravir and OBT, began in 2006, and the ARVs used in the OBT-alone group (eg, darunavir and tipranavir) do not reflect those used in BRIGHTE (notably, BENCHMRK lacked dolutegravir).The VIKING-3 study was selected as the second OBT-alone comparator study. VIKING-3 was conducted in an HTE population comparable to that in BRIGHTE, and the therapies used were generally comparable to BRIGHTE.17Castagna A Maggiolo F Penco G et al.Dolutegravir in antiretroviral-experienced patients with raltegravir- and/or elvitegravir-resistant HIV-1: 24-week results of the phase III VIKING-3 study.,,20Vavro CL Huang J Avatapally C Min S Ait-Khaled M. Durable efficacy and limited integrase resistance evolution in subjects receiving dolutegravir after failing prior integrase inhibitor (INI) regimens: week 48 results from VIKING-3. Similar to the design of BRIGHTE, all participants received the investigational drug dolutegravir as functional monotherapy for a short period (7 days in VIKING-3) before receiving dolutegravir in combination with OBT.

Screening viral load inclusion criteria differed among the studies. Therefore, a screening HIV-1 RNA load >1000 copies/mL was selected for BRIGHTE participants (combined randomized and nonrandomized ITT-E study population, N = 371; 24 participants with screening HIV-1 RNA load ≤1000 copies/mL removed, N = 347) for comparison with TMB-301 and BENCHMRK. For comparison with VIKING-3, screening HIV-1 RNA load ≥500 copies/mL was selected for BRIGHTE participants (ITT-E study population, N = 371; 13 participants with screening HIV-1 RNA load <500 copies/mL removed, N = 358). Because the definition of virologic suppression differed among the studies, a threshold of HIV-1 RNA load <50 copies/mL was used.

Overall susceptibility rating scoring differed among the studies. Overall susceptibility ratings of 0.5 (partial susceptibility) from BRIGHTE were reassigned to 0 (no susceptibility) to match the TMB-301 scoring approach. To match resistance data between BRIGHTE and VIKING-3, reassignments were performed for OSS-new (susceptibility score for ARVs not previously used by the participant) reported for the OBT from the VIKING-3 trial to account for the activity of the investigational agent (dolutegravir) in addition to OBT. Dolutegravir was presumed to be fully active in VIKING-3 and assigned an overall susceptibility rating of 1 (full susceptibility) for all participants. Furthermore, partial overall susceptibility ratings from BRIGHTE were reassigned from 0.5 to 0 to match the VIKING-3 scoring approach. Further details of how susceptibility scores were harmonized between BRIGHTE and VIKING-3 are provided in Supplemental Appendix Section 2.4.2.

TMB-301 reported data through 24 weeks, and results were compared with 24-week data from BRIGHTE. In the comparison of BRIGHTE with BENCHMRK, end points were compared at week 96, the latest comparable time point for each study available at the time of this analysis. In VIKING-3, participants could withdraw once commercial dolutegravir was available after week 48, and these withdrawals were recorded as failures. Therefore, BRIGHTE was compared with VIKING-3 efficacy end points at week 48. Safety data available from VIKING-3 reported all events observed through the week 48 data cutoff (median [range] exposure, 507 [14–757] days). Thus, safety data were compared with the BRIGHTE week 48 data cutoff (median [range] exposure, 621 [1–1039] days).

We used MAIC analyses to reweight IPD from the BRIGHTE trial as the index cohort, such that the adjusted (weighted) baseline characteristic summary statistics matched the summary statistics reported for each comparator cohort, using previously described methods2Signorovitch JE Wu EQ Yu AP et al.Comparative effectiveness without head-to-head trials: a method for matching-adjusted indirect comparisons applied to psoriasis treatment with adalimumab or etanercept. and summarized below. Participant-level outcome data from the index trial were then weighted by these values, allowing for comparison of treatment outcomes across balanced populations.For each MAIC analysis, a logistic propensity score model was fitted using the method of moments to derive weights for the index trial to balance the summary statistics of the baseline characteristics between BRIGHTE and the comparator cohort. Summary statistics and effective sample size were derived as described in Supplemental Appendix Section 2.1. Treatment effects were estimated using weighted regression with sandwich SEs. For all continuous outcome variables, mean differences between the index and comparator groups were estimated. For all binary outcome variables, odds ratios (ORs) were estimated to measure relative efficacy or tolerability, with a weighted logistic regression. Variables considered for the purpose of matching were determined using literature on factors prognostic of outcomes in HIV21Antiretroviral Therapy Cohort Collaboration.
  Survival of HIV-positive patients starting antiretroviral therapy between 1996 and 2013: a collaborative analysis of cohort studies., 22May M Sterne JAC Sabin C et al.Prognosis of HIV-1-infected patients up to 5 years after initiation of HAART: collaborative analysis of prospective studies., 23Prognosis of HIV patients receiving antiretroviral therapy according to CD4 counts: a long-term follow-up study in Yunnan, China. and in consultation with clinical experts. Age, sex, number of years treated, disease duration, AIDS history, baseline viral load, baseline CD4+ cell count, and number of active ARVs in initial OBT (based on susceptibility scores) were chosen. For each MAIC analysis, the final set of matching variables was determined by data availability, and conformation of the data with the statistical assumptions was required.

Where possible, analyses quantified efficacy in terms of differences in mean change from baseline in CD4+ cell count and relative difference assessed by OR in proportions of participants with virologic suppression, protocol-defined virologic failure (PDVF), and treatment discontinuation. Secondary analyses assessed the safety profile of fostemsavir, considering proportions of participants with drug-related adverse events (AEs), serious AEs (SAEs), discontinuation because of AEs, and death, where appropriate.