Fixed-dose combination , , and versus -containing regimens for initial treatment of HIV-1 infection: week 144 results from two randomised, double-blind, multicentre, phase 3, non-inferiority trials

Chloe Orkin, Edwin DeJesus, Paul E Sax, Jose R Arribas, Samir K Gupta, Claudia Martorell, Jeffrey L Stephens, Hans-Jurgen Stellbrink, David Wohl, Franco Maggiolo, Melanie A Thompson, Daniel Podzamczer, Debbie Hagins, Jason A Flamm, Cynthia Brinson, Amanda Clarke, Hailin Huang, Rima Acosta, Diana M Brainard, Sean E Collins, Hal Martin, on behalf of the GS-US-380-1489 and GS-US-380-1490 study investigators

Summary Background In the primary week-48 analyses of two phase 3 studies, coformulated bictegravir, emtricitabine, and Lancet HIV 2020; 7: e389–400 tenofovir alafenamide was non-inferior to a dolutegravir-containing regimen in treatment-naive people with HIV. See Comment page e374 We report week-144 efficacy and safety results from these studies. Queen Mary University of London, London, UK Methods We did two double-blind, active-controlled studies (now in open-label extension phase). Study 1 randomly (Prof C Orkin MD); Barts Health NHS Trust, Royal London assigned (1:1) HLA-B*5701-negative adults without B co-infection to receive coformulated bictegravir Hospital, Ambrose King Centre, 50 mg, emtricitabine 200 mg, and tenofovir alafenamide 25 mg, or coformulated dolutegravir 50 mg, London, UK (Prof C Orkin); 600 mg, and 300 mg once daily. Study 2 randomly assigned (1:1) adults to bictegravir, emtricitabine, Orlando Immunology Center, and tenofovir alafenamide, or dolutegravir 50 mg given with coformulated emtricitabine 200 mg and tenofovir Orlando, FL, USA (E DeJesus MD); Division of Infectious Diseases, alafenamide 25 mg. We previously reported non-inferiority at the primary endpoint. Here, we report the week-144 Brigham and Women’s Hospital secondary outcome of proportion of participants with plasma HIV-1 RNA less than 50 copies per mL at week 144, and Harvard Medical School, by US Food and Drug Administration Snapshot algorithm, analysed in the same manner. These studies were Boston, MA, USA registered with ClinicalTrials.gov, NCT02607930 and NCT02607956. (Prof P E Sax MD); Infectious Diseases Unit, Hospital Universitario La Paz, Findings 629 participants were randomly assigned and treated in study 1 (314 to bictegravir, emtricitabine, and IdiPaz Madrid, Spain tenofovir alafenamide, and 315 to dolutegravir, abacavir, and lamivudine) and 645 in study 2 (327 to bictegravir, (Prof J R Arribas MD); emtricitabine, and tenofovir alafenamide, 325 to dolutegravir, emtricitabine, tenofovir alafenamide). At week 144, Department of Medicine, Division of Infectious Diseases, bictegravir, emtricitabine, and tenofovir alafenamide was non-inferior to both dolutegravir-containing regimens Indiana University School of for efficacy. In study 1, 256 (82%) of 314 participants had plasma HIV-1 RNA less than 50 copies per mL in the Medicine, Indianapolis, IN, USA bictegravir, emtricitabine, and tenofovir alafenamide group and 265 (84%) of 315 in the dolutegravir, abacavir, and (Prof S K Gupta); The Research lamivudine group (difference −2·6%, 95% CI −8·5 to 3·4). In study 2, 262 (82%) of 320 participants had plasma Institute, Springfield, MA, USA (C Martorell MD); Department HIV-1 RNA less than 50 copies per mL in the bictegravir, emtricitabine, and tenofovir alafenamide group and of Internal Medicine, Mercer 273 (84%) of 325 in the dolutegravir, emtricitabine, and tenofovir alafenamide group (difference −1·9%, University School of Medicine, −7·8 to 3·9). In both studies, no participant had treatment-emergent resistance to study drugs up to week 144. All Macon, GA, USA (Prof J L Stephens MD); treatment regimens were well tolerated with additional exposure. Adverse events that led to study drug Department of Internal discontinuation were reported for no participants in the bictegravir, emtricitabine, and tenofovir alafenamide Medicine, Infectious Diseases, group versus five (2%) of 315 in the dolutegravir, abacavir, and lamivudine group (study 1), and six (2%) of 320 in University of Hamburg, the bictegravir, emtricitabine, and tenofovir alafenamide versus six (2%) of 325 in the dolutegravir, emtricitabine, Hamburg, Germany (Prof H-J Stellbrink MD); and tenofovir alafenamide group (study 2). In study 1, statistically significant differences were observed in median Department of Medicine, changes from baseline in fasting total cholesterol (14 mg/dL vs 10 mg/dL; p=0·034), direct LDL (21 mg/dL vs Institute of Global Health and 14 mg/dL; p=0·004), and total cholesterol to HDL ratio (–0·1 vs –0·3; p=0·007) at week 144; no differences were Infectious Diseases, University observed between groups in study 2. Weight gain was seen across all treatment groups in both studies, with no of North Carolina at Chapel Hill, Chapel Hill, NC, USA differences in median changes from baseline in weight at week 144 for either study. (Prof D Wohl MD); Unit of HIV-related Diseases and Interpretation These long-term data support the use of bictegravir, emtricitabine, and tenofovir alafenamide as a safe, Experimental Therapies, well tolerated, and durable treatment for people with HIV, with no emergent resistance. Azienda Ospedaliera Papa Giovanni XXIII, Bergamo, Italy (F Maggiolo MD); AIDS Research Funding . Consortium of Atlanta, Atlanta, GA, USA (M A Thompson MD); Infectious Disease Department, Hospital Universitari de Bellvitge, Barcelona, Spain Research in context (D Podzamczer MD); Georgia Department of Public Health, Evidence before this study other guidelines-recommended treatment regimens Coastal Health District, Data on bictegravir regimens for HIV treatment extend to a (dolutegravir, abacavir, and lamivudine; and dolutegravir plus Chatham Care Center, maximum follow-up time of 96 weeks. We searched PubMed co-formulated emtricitabine and tenofovir alafenamide), Savannah, GA, USA for randomised clinical trials of bictegravir in people with HIV, co-formulated bictegravir, emtricitabine and tenofovir (D Hagins MD); Department of Adult and Family Medicine, with title and abstract search terms of “bictegravir” and alafenamide showed non-inferior efficacy at week 144. No Kaiser Permanente Medical “randomised” or “randomized”. We used internet search participants on any regimen failed with emergent resistance, Group, Sacramento, CA, USA engines to identify governmental, non-governmental, and further showing the high barrier to resistance of the study (J A Flamm MD); Central Texas professional medical society practice guidelines relevant to regimens. We found less and fewer drug-related Clinical Research, Austin, TX, USA (C Brinson MD); bictegravir. Searches were limited to publications in English adverse effects in those treated with co-formulated Elton John Centre, Royal Sussex between Jan 1, 1997, and Nov 1, 2019. Our search yielded bictegravir, emtricitabine, and tenofovir alafenamide County Hospital, 14 publications, of which we excluded nine because they compared with those who received dolutegravir, abacavir, and Brighton & Sussex University lamivudine. Bone mineral density, glomerular filtration rate, Hospitals NHS Trust, reported results from switch studies in virologically Brighton, UK (A Clarke MD); and suppressed adults with HIV, meta-analyses, systemic analyses, and biomarkers of renal tubular function were similar between Department of Biometrics phase 1, or patient-reported outcomes. The remaining five bictegravir, emtricitabine, and tenofovir alafenamide and (H Huang PhD), Department of reports summarised outcomes from weeks 48 or 96 from dolutegravir, abacavir, and lamivudine. These studies provide Virology (R Acosta BS), Department of HIV Clinical three phase 2 and phase 3 studies of bictegravir, emtricitabine, evidence of the durable efficacy, continued tolerability, and Research (S E Collins MD, and tenofovir alafenamide in treatment-naive adults. Across no treatment-emergent resistance for participants taking D M Brainard MD, H Martin MD), these reports, bictegravir, emtricitabine, and tenofovir bictegravir, emtricitabine, and tenofovir alafenamide. Gilead Sciences, alafenamide was non-inferior to standard-of-care regimens, Foster City, CA, USA Implications of all the available evidence including those containing dolutegravir. Bictegravir, Correspondence to: Data from these analyses confirm findings from reports emtricitabine, and tenofovir alafenamide had a similar renal, Dr Sean E Collins, MD, Gilead from weeks 48 and 96 from randomised studies comparing bone, and lipid safety profile to that of the comparators. Sciences, Foster City, CA 94404, bictegravir, emtricitabine, and tenofovir alafenamide with USA Bictegravir, emtricitabine, and tenofovir alafenamide became standard-of-care regimens, including those containing [email protected] a guidelines-recommended regimen for the initial treatment dolutegravir. Co-formulated bictegravir, emtricitabine, and of adults with HIV on the basis of week-48 data; longer-term tenofovir alafenamide can be administered once daily, does data are needed to inform clinical care. not require HLA B*5701 testing, and provides Added value of this study guideline-recommended therapy for people with HIV and with These randomised, blinded, phase 3 studies provide evidence HIV– virus co-infection. Together, these studies of the long-term safety and efficacy of bictegravir, provide long-term efficacy and safety data to guide treatment emtricitabine, and tenofovir alafenamide. Compared with decisions for people with HIV.

Introduction of dolutegravir, abacavir, and lamivudine, but fewer strand-transfer inhibitors anchor HIV treatment-related adverse events, largely because of treatment worldwide. Treatment regimens containing an higher incidence of nausea in the dolutegravir, abacavir, integrase strand-transfer inhibitor and two nucleoside and lamivudine group.5,6 reverse-transcriptase inhibitors comprise most European, We report 144-week outcomes of these two trials to US, and WHO guideline-preferred regimens for the ini­ provide data on the relative efficacy and safety of bictegravir, tial treatment of HIV.1–4 Bictegravir is the most recent emtricitabine, and tenofovir alafenamide versus two addition to the integrase strand-transfer inhibitor class different guidelines-recommended, dolutegravir-contain­ and was compared with dolutegravir for initial treatment ing regimens in treatment-naive individuals initiating of HIV as a part of a complete three-drug regimen in two HIV therapy. different phase 3 randomised, double-blind, active-con­ trolled trials. Bictegravir, coformulated with emtricita­bine Methods and tenofovir alafenamide, was compared with cofor­ Study design and procedures mulated dolutegravir, abacavir, and lamivudine in study 1 GS-US-380-1489 (study 1) and GS-US-380-1490 (study 2) and with dolutegravir plus emtricitabine and tenofovir were randomised, double-blind, multicentre, active- alafenamide in study 2. All treatments had high efficacy; controlled, non-inferiority, phase 3 trials. Study 1 was the bictegravir regimen was non-inferior to either done at 122 outpatient centres in nine countries dolutegravir-containing regimen for up to 96 weeks. (Belgium, France, Germany, Italy, Spain, UK, Dominican In study 1, when the nucleoside reverse-transcriptase Republic, Canada, and USA). Study 2 was done at inhibitors also differed between treatment groups, 126 outpatient centres in ten countries, including the bictegravir, emtricitabine, and tenofovir alafenamide nine countries in study 1 and Australia. Detailed had similar bone and renal safety compared with that methods have been previously published.7,8 Briefly, study 1 investigators enrolled treatment-naive adults integrase genotyping was done by deep sequencing living with HIV, with plasma HIV-1 RNA levels of at after randomisation (Seq-IT; Kaiserslautern, Germany). least 500 copies per mL, who were HLA-B*5701 negative, Resistance testing (Monogram Biosciences; did not have infection, and who had an San Francisco, CA, USA) as defined in the protocol was estimated glomerular filtration rate (eGFR) of at least done for any participant who had HIV-1 RNA of at least 50 mL per min (Cockcroft–Gault equation). Similarly, 50 copies per mL, with a confirmed HIV-1 RNA of at study 2 investigators enrolled treatment-naive adults least 200 copies per mL, or who had an HIV-1 RNA of at living with HIV, with an eGFR of at least 30 mL per min; least 200 copies per mL at weeks 48, 96, 144, or the last participants with chronic hepatitis B infection were visit on treatment after week 8, and who did not permitted to enter. Both studies required virological subsequently re-suppress HIV-1 RNA while on resistance testing showing sensitivity to emtricitabine treatment. and tenofovir. Safety was assessed by physical examinations, These studies were undertaken in accordance with the laboratory tests, 12-lead electrocardiogram, concomitant Declaration of Helsinki and were approved by central or drugs, and recording of adverse events, which were site-specific review boards or ethics committees. All coded using the Medical Dictionary for Regulatory participants gave written informed consent. Activities (MedDRA, version 22.0). Relatedness of adverse events to blinded study drugs was indicated by Randomisation and masking the investigator in a binary manner (yes or no). Participants were randomly assigned (1:1) to bictegravir, In study 1, we did dual-energy x-ray absorptiometry emtricitabine, and tenofovir alafenamide; or either scans for hip and lumbar spine bone mineral density dolutegravir, abacavir, and lamivudine (study 1) or before drug administration at baseline and then at dolutegravir, emtricitabine, and tenofovir alafenamide weeks 24, 48, 96, and 144. All scans were read at a (study 2). Participants also received placebo tablets matching the alternative treatment; thus, investigators, participants, and study staff giving treatment, assessing 739 participants screened outcomes, and collecting data were masked to treatment group. A computer-generated allocation sequence (block 108 not randomly assigned size 4) was created by Bracket (San Francisco, CA, USA). 89 did not meet eligibility criteria 8 lost to follow-up Randomisation in each study was stratified by HIV-1 6 withdrew consent RNA (≤100 000 copies per mL, >100 000 to ≤400 000 copies 2 investigator decision 1 outside visit window per mL, or >400 000 copies per mL), CD4 count (<50 cells 1 adverse event per μL, 50–199 cells per μL, or ≥200 cells per μL), and 1 other region (USA or outside of the USA) at screening.

631 randomly assigned Procedures Participants received a once-daily regimen of coformulated bictegravir 50 mg, emtricitabine 200 mg, and tenofovir alafenamide 25 mg; or, either dolutegravir 316 randomly assigned to bictegravir, 315 randomly assigned to abacavir, 50 mg, abacavir 600 mg, and lamivudine 300 mg; or emtricitabine, and tenofovir alafenimide dolutegravir, and lamivudine dolutegravir 50 mg, emtricitabine 200 mg, and tenofovir alafenamide 25 mg. All regimens were given regardless 2 never treated of food intake. We did post-baseline study visits at weeks 4, 8, 12, and every 12 weeks thereafter, with 314 received bictegravir, emtricitabine, and 315 received abacavir, dolutegravir, and masked treatment visits planned until week 144 as tenofovir alafenamide lamivudine previously reported.7,8 Participants were given the option of participating in an open-label extension period for an additional 96 weeks (5 years of cumulative exposure). 54 discontinued 48 discontinued 2 died 1 died Laboratory tests included haematological analysis, 1 5 adverse events serum chemistry tests, fasting lipids, CD4 cell counts, 6 at investigator’s discretion 2 at investigator’s discretion 3 non-compliance with 6 non-compliance with measures of renal function by eGFR (Covance Labora- study drugs study drugs tories, Indianapolis, IN, USA) and HIV-1 RNA plasma 2 protocol violations 3 protocol violations concentration (Roche TaqMan 2.0; Roche Diagnostics, 16 at participant’s decision 15 at participant’s decision 24 lost to follow-up 16 lost to follow-up Rotkreuz, Switzerland) in both studies. Biomarkers of renal tubular function, including urine albumin to creatinine ratio, retinol binding protein to creatinine 260 continued treatment 267 continued treatment ratio, and β2-microglobulin to creatinine ratio (Covance Laboratories) were done in study 1 only. Baseline HIV-1 Figure 1: Study 1 trial profile proportion of participants with HIV-1 RNA less than 742 participants assessed for eligibility 20 copies per mL at week 144 by the Snapshot algorithm, and change in CD4 cell count from baseline at week 144. In study 1, the percentage changes from baseline in hip 85 not randomly assigned 60 did not meet eligibility criteria and lumbar spine bone mineral density were assessed as 14 withdrew consent a secondary outcome at week 144. Renal safety 3 lost to follow-up 2 investigator decision assessments included the change from baseline in serum 2 outside visit window creatinine and eGFR at week 144 (both studies), 1 adverse event 3 other treatment-emergent proteinuria until week 144 (both studies), and percentage changes from baseline in urine retinol binding protein to creatinine ratio, urine 657 randomly assigned β2-microglobulin to creatinine ratio, and urine albumin to creatinine ratio at week 144 (study 1 only). Adverse event incident rates through week 144 and

327 randomly assigned to bictegravir, 330 randomly assigned to dolutegravir plus changes in fasting lipids at week 144 were assessed by emtricitabine, and tenofovir emtricitabine, and tenofovir treatment group. alafenamide alafenamide Statistical analysis 7 never treated 5 never treated For each study, we did the week-144 efficacy analysis (proportion of participants with plasma HIV-1 RNA <50 copies per mL at week 144, days 967–1050 inclusive) 320 received treatment 325 received treatment in the full analysis set (all participants who were randomly assigned and had received at least one dose of the study 59 discontinued 47 discontinued drug) after enrolled participants had completed their 6 adverse events 6 adverse events week 144 study visit or had prematurely discontinued the 3 died 4 died 5 pregnancy 3 pregnancy study drug. One the basis of the normal approximation,­ 6 at investigator’s discretion 2 at investigator’s discretion we assessed non-inferiority with a Mantel-Haenszel 3 protocol violations 4 had non-compliance with 18 at participant’s decision study drugs stratified risk difference and its associated 95% CI in 18 lost to follow-up 1 protocol violation virological response rates of each study individually 13 at participant’s decision (bictegravir, emtricitabine, and tenofovir alafenamide 14 lost to follow-up minus dolute­gravir, abacavir, and lamivudine in study 1; bictegravir, emtricitabine, and tenofovir alafenamide 261 continued treatment 278 continued treatment minus dolutegravir plus emtricitabine, and tenofovir alafenamide in study 2) with a prespecified non-inferiority Figure 2: Study 2 trial profile margin of −12%, based on published US FDA regulatory No participants discontinued treatment due to reasons related to efficacy. guidance.9 Sample size justification was based on the primary outcome.7,8 Statistical analysis followed the centralised centre that was masked to treatment group methodology previously reported for the primary assignment (BioClinica, Newtown, PA, USA). endpoint7,8 and included pre-specified subgroups, per- protocol analysis, the proportion of participants with Outcomes plasma HIV RNA less than 20 copies per mL, and missing We have previously reported the primary and secondary as failure and missing as excluded imputations. Changes efficacy outcomes: the proportion of participants who from baseline in CD4 cell count at week 144 were had plasma HIV-1 RNA less than 50 copies per mL at summarised by treatment group with descriptive statistics weeks 48 and 96,5–8 as defined by the US Food and Drug based on the full analysis set using observed on-treatment Administration (FDA) Snapshot algorithm.9 Another data. secondary efficacy outcome was the proportion of Methods for assessing baseline characteristics and participants who had plasma HIV-1 RNA less than safety outcomes including adverse events, bone mineral 50 copies per mL, by Snapshot algorithm, at week 144. density, renal biomarkers, and fasting lipids were Sensitivity analyses assessed virological efficacy at also previously reported.7,8 Significance testing was done week 144 in pre-specified subgroups of age, sex, race, for adverse events with more than 5% treatment baseline HIV-1 RNA, baseline CD4 cell count, geo­ difference using Fisher’s exact text. A post-hoc analysis graphical region, and study medication adherence. compared the changes from baseline in weight at Other secondary efficacy analyses included the week 144 using a two-sided Wilcoxon rank-sum test. proportion of participants with plasma HIV-1 RNA less We used SAS Software (version 9.4) for all analyses. than 50 copies per mL at week 144 when imputing These studies were done according to protocol with­ missing-as-failure and missing-as-excluded data, the out substantial deviations and are registered with Study 1 Study 2 Bictegravir, emtricitabine, Dolutegravir, abacavir, Bictegravir, emtricitabine, Dolutegravir plus emtricitabine, and tenofovir alafenamide and lamivudine and tenofovir alafenamide and tenofovir alafenamide (n=314) (n=315) (n=320) (n=325) Median (range) age (years) 31 (18–71) 32 (18–68) 33 (18–71) 34 (18–77) Sex Women 29 (9%) 33 (10%) 40 (13%) 37 (11%) Men 285 (91%) 282 (90%) 280 (88%) 288 (89%) Race White 180 (58%) 179 (57%) 183 (57%) 195 (60%) Black 114 (36%) 112 (36%) 97 (30%) 100 (31%) Asian 6 (2%) 10 (3%) 7 (2%) 10 (3%) Ethnicity Hispanic or Latino 72 (23%) 65 (21%) 83 (26%) 81 (25%) HIV disease status Asymptomatic 286 (91%) 286 (91%) 286 (89%) 288 (89%) Symptomatic 16 (5%) 14 (4%) 10 (3%) 11 (3%) AIDS 12 (4%) 15 (5%) 24 (8%) 26 (8%) HIV risk factor* Heterosexual sex 61 (19%) 62 (20%) 81 (25%) 77 (24%) Homosexual sex 251 (80%) 250 (79%) 237 (74%) 250 (77%) Intravenous drug use 5 (2%) 4 (1%) 3 (1%) 6 (2%)

HIV-1 RNA log10 copies per mL 4·42 (4·03–4·87) 4·51 (4·04–4·87) 4·43 (3·95–4·90) 4·45 (4·03–4·84) HIV-1 RNA concentration 53 (17%) 50 (16%) 66 (21%) 54 (17%) >100 000 copies per mL CD4 count (cells per μL) 443 (299–590) 450 (324–608) 440 (289–591) 441 (297–597) Participants with CD4 count (cells per μL) <200 36 (11%) 32 (10%) 44 (14%) 34 (10%) ≥200 to <500 156 (50%) 149 (47%) 158 (49%) 171 (53%) ≥500 122 (39%) 134 (43%) 118 (37%) 120 (37%) Creatinine clearance by 125·9 (107·7–146·3) 123·0 (107·0–144·3) 120·4 (100·8–141·8) 120·6 (102·8–145·1) Cockcroft-Gault formula (mL/min) Body-mass index (kg/m²) 25·1 (22·4–28·7) 24·9 (22·5–29·1) 25·0 (22·2–28·3) 24·6 (22·2–28·0) Primary resistance-associated mutations† INSTI 4 (1%) 4 (1%) 3 (1%) 6 (2%) NRTI 6 (2%) 5 (2%) 10 (3%) 5 (2%) NNRTI 36 (12%) 51 (16·2%) 41 (13%) 41 (13%) PI 12 (4%) 11 (4%) 4 (1%) 10 (3%) HIV–HBV coinfected ·· ·· 8 (3%) 6 (2%) HIV–HCV coinfected ·· ·· 5 (2%) 5 (2%)

Data are median (IQR), or n (%), unless otherwise indicated. HBV=hepatitis B virus HCV= INSTI=integrase strand transfer inhibitor. NNRTI=nonnucleoside inhibitor. NRTI=nucleoside/tide reverse transcriptase inhibitor. PI=protease inhibitor. *A participant may fit more than one HIV risk factor category; therefore, percentages can add up to more than 100%. †Primary INSTI substitutions were Thr66Ala/Ile/Lys, Glu92Gly/Gln, Thr97Ala, Phe121Tyr, Tyr143Cys/His/Arg, Ser147Gly, Gln148His/Lys/Arg, Asn155His/Ser, Arg263Lys in IN. Primary NRTI substitutions were Met41Leu, Lys65Glu/Asn/Arg, Asp67Asn, Thr69 insertions, Lys70Glu/Arg, Leu74Ile/Val, Tyr115Phe, Gln151Met, Met184Val/Ile, Leu210Trp, Thr215Tyr/Phe, Lys219Glu/Asn/Gln/Arg in RT. Primary NNRTI substitutions were Leu100Ile, Lys101Glu/Pro, Lys103Asn/Ser, Val106Ala/Met, Val108Ile, Glu138Ala/Gly/Lys/Gln/Arg, Val179Leu, Tyr181Cys/Ile/Val, Tyr188Cys/Leu/His, Gly190Ala/Glu/Gln/Ser, His221Tyr, Pro225His, Phe227Cys, Met230Ile/ Leu in RT. Primary PI substitutions were Asp30Asn, Val32Ile, Met46Ile/Leu, Ile47Ala/Val, Gly48Val, Ile50Val/Leu, Ile54Leu/Met, Gln58Glu, Thr74Pro, Leu76Val, Val82Ala/Phe/ Leu/Thr/Ser, Asn83Asp, Ile84Val, Asn88Ser, Leu90Met in PR.

Table 1: Baseline demographic and clinical characteristics

ClinicalTrials.gov, numbers NCT02607930 (study 1) and data in the study and CO and HM had final responsibility NCT02607956 (study 2). for the decision to submit the manuscript for publication.

Role of the funding source Results The funder of the study had the lead role in data collection, Between Nov 13, 2015, and July 14, 2016, 631 participants data analysis, data interpretation, and helped to write the were randomly assigned in study 1 (316 to bictegravir, report. The corresponding author had full access to all the emtricitabine, and tenofovir alafenamide and 315 to Study 1 Study 2 Bictegravir, Dolutegravir, Bictegravir, Dolutegravir plus emtricitabine, and abacavir, and emtricitabine, and emtricitabine, and tenofovir alafenamide lamivudine tenofovir alafenamide tenofovir alafenamide (n=314) (n=315) (n=320) (n=325) HIV-1 RNA <50 copies per mL (full analysis set) 256 (82%) 265 (84%) 262 (81%) 273 (84%) Difference in percentages (95% CI)* –2·6% (–8·5% to 3·4%) ·· –1·9% (–7·8% to 3·9%) ·· HIV-1 RNA ≥50 copies per mL 2 (<1%) 9 (3%) 15 (5%) 10 (3%) HIV1 RNA ≥50 copies per mL at week 144 1 (<1%) 2 (<1%) 1 (<1%) 4 (1%) Discontinued because of lack of efficacy 0 0 0 0 Discontinued for other reasons† and last available 1 (<1%) 7 (0<1%) 14 (4%) 6 (2%) HIV-1 RNA ≥50 copies per mL No virological data 56 (18%) 41 (13%) 43 (13%) 42 (13%) Discontinued because of adverse event or death‡ 2 (<1%) 6 (2%) 8 (3%) 9 (3%) Discontinued for other reasons† and last available 50 (16%) 34 (11%) 35 (11%) 29 (9%) HIV-1 RNA <50 copies per mL Missing data but on treatment 4 (1%) 1 (<1%) 0 4 (1%) HIV-1 RNA <50 copies per mL (per-protocol set) 256/257 (99%) 260/262 (99%) 257/258 (99%) 270/273 (99%) Difference in percentages (95% CI)* 0·4% (–1·6% to 2·4%) ·· 0·7% (–1·4% to 2·7%) ·· HIV-1 RNA <50 copies per mL by missing as failure 259/314 (83%) 267/315 (85%) 268/320 (84%) 276/325 (85%) Difference in percentages (95% CI)§ –2·3% (–8·1% to 3·6%) ·· –0·9% (–6·6% to 4·7%) ·· HIV-1 RNA <50 copies per mL by missing as excluded 259/260 (99%) 267/269 (99%) 268/270 (99%) 276/280 (99%) Difference in percentages (95% CI)§ 0·4% (–1·6% to 2·3%) ·· 0·7% (–1·6% to 2·9%) ·· HIV-1 RNA <20 copies per mL (full analysis set) 245 (78%) 259 (82%) 248 (78%) 257 (79%) Difference in percentages (95% CI)§ –4·2% (–10·5% to 2·1%) ·· –1·1% (–7·4% to 5·3%) ··

Data are n (%). The week-144 window is between days 967 and 1050 (inclusive). *The difference in percentages of participants with HIV-1 RNA less than 50 copies per mL between treatment groups and the 95% CI were calculated on the basis of the Mantel-Haenszel proportions adjusted by baseline HIV-1 RNA stratum and region stratum. †Other reasons include participants who discontinued treatment because of investigator’s discretion, participant decision, lost to follow-up, non-compliance with treatment, protocol violation, pregnancy, and study terminated by sponsor. ‡One death in the bictegravir, emtricitabine, and tenofovir alafenamide occurred after the participant reached the week 96 outcome. §Difference in percentages and 95% CI were based on a dichotomized response: HIV-1 RNA less than 50 copies per mL versus HIV-1 RNA at least 50 copies per mL or missing for missing-as-failure approach, and HIV-1 RNA less than 50 copies per mL versus HIV-1 RNA at least 50 copies per mL for missing-as-excluded approach. Difference in percentages of participants with HIV-1 RNA less than 50 copies per mL between treatment groups and the 95% CI were calculated on the basis of the Mantel-Haenszel proportions adjusted by baseline HIV-1 RNA stratum and region stratum.

Table 2: Virological outcomes at week 144 dolutegravir, abacavir, and lamivudine; figure 1). Of In study 1, 256 (82%) of 314 participants who received these, 314 (50%) received at least one dose of bicteg­ bictegravir, emtricitabine, and tenofovir alafenamide ravir, emtricitabine, and tenofovir alafenamide, and versus 265 (84%) of 315 who received dolutegravir, 315 (50%) dolutegravir, abacavir, and lamivudine. abacavir, and lamivudine had plasma HIV-1 RNA less than Between Nov 11, 2015, and July 15, 2016, 657 participants 50 copies per mL in the week-144 analysis window were randomly assigned in study 2 (327 to bictegravir, (difference −2·6%; 95% CI −8·5 to 3·4; table 2). An HIV-1 emtricitabine, and tenofovir alafenamide, and 330 to RNA of at least 50 copies per mL at week 144 or at the last dolutegravir plus emtricitabine and tenofovir alafe­ test while on treatment was observed in two (<1%) namide; figure 2). 320 (49%) participants received at participants in the bictegravir, emtricitabine, and tenofovir least one dose of bictegravir, emtricitabine, and alafenamide group and nine (3%) in the dolutegravir, tenofovir alafenamide, and 325 (49%) dolutegravir plus abacavir, and lamivudine group, and we found no coformulated emtri­citabine and tenofovir alafenamide. discontinuations due to lack of efficacy. A total of 56 (18%) Demographics and baseline characteristics were participants in the bictegravir, emtricitabine, and tenofovir balanced between the treatment groups in both studies alafenamide group and 41 (13%) in the dolutegravir, (table 1). abacavir, and lamivudine group did not have data at At 144 weeks, the fixed-dose combination of bictegravir, week 144 but had a last on-study HIV-1 RNA test result of emtricitabine, and tenofovir alafenamide was non- less than 50 copies per mL. The most common reasons inferior to both dolutegravir-containing regimens, as for missing data at week 144 were loss to follow-up (24 [8%] defined by the US FDA Snapshot algorithm for the in the bictegravir, emtricitabine, and tenofovir alafenamide secondary efficacy outcome of the proportion of group vs 16 [5%] in the dolutegravir, abacavir, and participants with plasma HIV-1 RNA less than 50 copies lamivudine group) and participant decision (16 [5%] in the per mL (table 2, figure 3). bictegravir, emtricitabine, and tenofovir alafenamide A Virological outcomes by US FDA Snapshot in Study 1 Treatment difference, % (95% CI) 100

n=256 n=265 Bictegravir, emtricitabine, 80 and tenofovir alafenamide (n=314) Dolutegravir, abacavir, 60 and lamivudine (n=315) –2·6 –8·5 3·4 40 Participants (%)

n=56 20 n=41 n=2 n=9 0 HIV-1 RNA HIV-1 RNA No virological –12 –6 0 6 12 <50 copies per mL ≥50 copies per mL data Favours dolutegravir, Favours bictegravir, emtricitabine, abacavir, and lamivudine and tenofovir alafenamide

B Virological outcomes by US FDA Snapshot in Study 2 Treatment difference, % (95% CI) 100 n=273 n=262 Bictegravir, emtricitabine, 80 and tenofovir alafenamide (n=320) Dolutegravir, emtricitabine, 60 and tenofovir alafenamide –1·9 (n=325) –7·8 3·9 40 Participants (%)

20 n=43 n=42 n=15 n=10 0 HIV-1 RNA HIV-1 RNA No virological –12 –6 0 6 12 <50 copies per mL ≥50 copies per mL data Favours dolutegravir, emtricitabine, Favours bictegravir, emtricitabine, and tenofovir alafenamide and tenofovir alafenamide

Figure 3: Virological outcomes at week 144 Percentage treatment difference was adjusted for both studies. FDA=Food and Drug Administration. group vs 15 [5%] in the dolutegravir, abacavir, and week-144 resistance analysis population; all six (2%) were lamivudine group). The prespecified per-protocol, in the dolutegravir, abacavir, and lamivudine group. Of missing-as-excluded, and missing-as-failure analyses were these, one (<1%) participant had resistance testing consistent with results from the full analysis set, showing between weeks 96 and 144. No emergent resistance high overall efficacy and no differences between the developed to any component of either treatment regimen. treatment groups (table 2). Testing for interactions CD4 cell count increased in each treatment group, with between treatment and subgroup was done using the mean changes from baseline at week 144 of 299 (SD 225) Wald test and identified differences between tr eatment cells per μL for the bictegravir, emtricitabine, and tenofovir groups, favouring dolutegravir, abaca vir, and lamivudine alafenamide group and 317 (220) cells per μL for the in the subgroup with cumulative adherence less than 95% dolutegravir, abacavir, and lamivudine group (p=0·30). (appendix p 14). The difference was driven by subgroup In study 2, 262 (82%) of 320 participants who received participants who did not have data in the analysis window bictegravir, emtricitabine, and tenofovir alafenamide and whose last on treatment assessment of HIV-1 RNA versus 273 (84%) of 325 who received dolutegravir plus was less than 50 copies per mL (31 [30%] in the bictegravir, emtricitabine and tenofovir alafenamide had plasma emtricitabine, and tenofovir alafenamide group vs 15 [13%] HIV-1 RNA less than 50 copies per mL in the week-144 in the dolutegravir, abacavir, and lamivudine group), analysis window (difference −1·9%, 95% CI −7·8 to 3·9; rather than virological failure (appendix p 7). No significant table 2). An HIV-1 RNA of at least 50 copies per mL at treatment differences were noted for other pre-specified week 144 or at the last test while on treatment was subgroups. Results across other efficacy outcomes observed in 15 (5%) participants in the bictegravir, consistently showed the durable efficacy ofot b h regimens emtricitabine, and tenofovir alafenamide group and (table 2). Six (1%) participants with protocol-defined ten (3%) in the dolutegravir plus emtricitabine and teno­ criteria for resistance testing were included in the fovir alafenamide group. Most of these (14 [4%) vs 6 [2%]) observed no differences between treatment groups in the Study 1 Study 2 pre-specified subgroups (appendix p 15). Results across Bictegravir, Dolutegravir, p value* Bictegravir, Dolutegravir plus other efficacy outcomes consistently showed the durable emtricitabine, abacavir, and emtricitabine, emtricitabine, and tenofovir lamivudine and tenofovir and tenofovir efficacy of both regimens (table 2). 15 (2%) participants alafenamide (n=315) alafenamide alafenamide met criteria for viral resistance testing at week 144: (n=314) (n=320) (n=325) eight (3%) in the bictegravir, emtricitabine, and tenofovir Any adverse event 300 (96%) 304 (97%) ·· 291 (91%) 300 (92%) alafenamide group and seven (2%) in the dolutegravir Adverse event ≥10% plus emtricitabine and tenofovir alafenamide group. Nausea 38 (12%) 76 (24%) 0·0001 31 (10%) 42 (13%) One (<1%) participant in the bictegravir, emtricitabine, Diarrhoea 54 (17%) 57 (18%) ·· 66 (21%) 52 (16%) and tenofovir alafenamide group and three (1%) in the Upper respiratory tract 43 (14%) 59 (19%) ·· 43 (13%) 52 (16%) dolutegravir plus emtricitabine and tenofovir alafenamide infection group had resistance testing between week 96 and 144. 44 (14%) 56 (18%) ·· 56 (18%) 57 (18%) No emergent resistance developed to any component of Nasopharyngitis 40 (13%) 52 (17%) ·· 50 (16%) 62 (19%) either treatment regimen. Syphilis 39 (12%) 49 (16%) ·· 33 (10%) 31 (10%) Baseline HIV-1 integrase genotyping was not required Back pain 34 (11%) 38 (12%) ·· 28 (9%) 38 (12%) at study entry and was completed retrospectively for 33 (11%) 38 (12%) ·· 28 (9%) 36 (11%) both studies, showing that primary mutations that Insomnia 25 (8%) 35 (11%) ·· 29 (9%) 24 (7%) are associated with resistance to integrase inhibitors Oropharyngeal pain 21 (7%) 35 (11%) ·· 20 (6%) 18 (6%) were present in eight (1%) participants in study 1 34 (11%) 20 (6%) ·· 25 (8%) 29 (9%) and nine (1%) in study 2, all but one of whom had Grade 3 or 4 adverse event 50 (16%) 50 (16%) ·· 54 (17%) 43 (13%) pre-existing Thr97Ala, which did not affect virological Serious adverse event 41 (13%) 53 (17%) ·· 63 (20%) 40 (12%) outcomes in either study. One participant in Study drug-related adverse 94 (30%) 132 (42%) 0·0021 71 (22%) 95 (29%) study 1 who had Gln148His and Gly140Ser substitutions event with phenotypic resistance to and , Study drug-related adverse event ≥5% partial resistance to dolutegravir, and full sensitivity Nausea 18 (6%) 56 (18%) <0·0001 10 (3%) 17 (5%) to bictegravir was randomly assigned to bictegravir, Diarrhoea 19 (6%) 13 (4%) ·· 10 (3%) 10 (3%) emtricitabine, and tenofovir alafenamide and had HIV-1 Headache 16 (5%) 16 (5%) ·· 14 (4%) 10 (3%) RNA less than 50 copies per mL at week 4 and until Study drug-related serious 2 (<1%) 1 (<1%) ·· 3 (<1%) 3 (<1%) week 144. adverse event In study 2, 14 (2%) participants had HIV-1 and Any adverse event leading to 0 5 (2%) ·· 6 (2%) 6 (2%) hepatitis B co-infection at baseline: eight (3%) in the study drug discontinuation† bictegravir, emtricitabine, and tenofovir alafenamide Death‡ 2 (<1%) 1 (<1%) ·· 4 (1%) 4 (1%) group and six (2%) in the dolutegravir plus emtricitabine Data are n (%). Table shows adverse events reported by 10% or more of participants in any treatment group for each and tenofovir alafenamide group. At week 144, by missing- study. *Significance testing was done for events with more than 5% treatment difference. p-values were calculated as-excluded analysis all 11 (100%) HIV and hepatitis B using Fisher’s exact text. No events in Study 2 met criteria for statistical significance testing. †In Study 1, treatment discontinuations due to adverse events in the dolutegravir, abacavir, and lamivudine group included nausea and coinfected participants had hepatitis B virus DNA less generalised rash (day 4, n=1); thrombocytopenia (day 50, n=1); chronic pancreatitis and steatorrhoea (day 134, n=1); than 29 IU per mL and HIV-1 RNA less than 50 copies per depression (day 248, n=1); and renal failure (day 621, n=1). ‡In Study 1, deaths in the bictegravir, emtricitabine, and mL; five (45%) of 11 from the bictegravir, emtricitabine, tenofovir alafenamide group were recreational drug overdose (day 771, n=1) and suicide (day 656, n=1); neither event was considered to be treatment related. The death in the dolutegravir, abacavir, and lamivudine group was due to and tenofovir alafenamide group and six (55%) from the recreational drug overdose (day 812, n=1); this event was not considered to be treatment related. In Study 2, deaths in dolutegravir, emtricitabine, and tenofovir alafenamide the bictegravir, emtricitabine, and tenofovir alafenamide group were cardiac arrest following appendicitis and septic group. Three (38%) of eight participants in the bictegravir, shock (day 28, n=1), gastric adenocarcinoma (day 376, n=1), hypertensive heart disease and congestive heart failure emtricitabine, and tenofovir alafenamide had hepatitis B (day 412, n=1), and sudden cardiac death (day 1060, n=1). Deaths in the dolutegravir plus emtricitabine and tenofovir alafenamide group were unknown causes (day 174, n=1; day 771, n=1), pulmonary embolism (day 266, n=1), and virus DNA data missing at week 144; two (25%) had no lymphoma (day 422, n=1). None of the deaths in either study were considered to be treatment related. hepatitis B virus DNA detected at the last visit at which Table 3: Adverse events until week 144 hepatitis B virus was tested and one (13%) had no post- baseline visits. Among HIV and hepatitis B virus coinfected participants we found no hepatic adverse were participants who had discontinued treatment events, no grade 3 or 4 adverse events, and no participants because of other reasons not related to efficacy and had a interrupted or discontinued their study regimen because last available HIV-1 RNA concentration of at least of an adverse event. 50 copies per mL. We found no discontinuations due to The median study drug exposure was 152 weeks (IQR lack of efficacy. Seven (2%) participants from the 147–159 for both treatment groups) in study 1 and bictegravir, emtricitabine, and tenofovir alafenamide 149 weeks (146–153 for the bictegravir, emtricitabine, and group did not have any HIV-1 RNA data after baseline; tenofovir alafenamide group; 146–155 for the dolutegravir, therefore, the only available HIV-1 RNA data, which was emtricitibane, and tenofovir alafenamide group) in used to assess efficacy until week 144, was collected study 2. Most adverse events were reported as mild or before those participants initiated study treatment. We moderate in severity (table 3). In study 1, nausea was reported in significantly fewer participants in the emtricitabine, and tenofovir alafenamide group two deaths bictegravir, emtricitabine, tenofovir alafenamide group occurred after week 96 (one unknown cause and than in the dolutegravir, abacavir, and lamivudine group one lymphoma); the other deaths (one unknown cause and (38 [12%] of 314 vs 76 [24%] of 315; p=0·0001). The one pulmonary embolism) occurred before week 96, as incidence of nausea was highest after starting treatment previously reported.6,8 One death from unknown cause in (16 [5%] in the bictegravir, emtricitabine, and tenofovir study 2 in a participant in the dolutegravir, emtricitabine alafenamide group vs 54 [17%] of 315 in the dolutegravir, and tenofovir alafenamide group was considered possibly abacavir, and lamivudine group at week 4) with a related to study treatment; none of the other deaths on difference between treatments in prevalent nausea until either study were considered related to study treatment. week 144 (appendix p 16). In each study, adverse events Study drugs were interrupted or discontinued by the were similar to those reported at weeks 48 and 96. In investigator when any on-study pregnancy was confirmed. study 1, few participants had adverse events that led to Three (5%) of 62 women in study 1 had confirmed study drug discontinuation (table 3), all of which occurred while on-study, two (7%) of 29 in the before week 96. In study 2, adverse events leading to bictegravir, emtricitabine, and tenofovir alafenamide study drug discontinuation were also uncommon group and one (3%) of 33 in the dolutegravir, abacavir, and (table 3). Only one event (large B-cell lymphoma in the in lamivudine group, all before week 96. The pregnancy the dolutegravir, emtricitabine, and tenofovir alafenamide outcomes were previously reported,5 which were a healthy group) led to discontinuation between weeks 96 and 144. full-term infant and one spontaneous abortion at 2 weeks In study 1, participants in the bictegravir, emtricitabine, gestation in the bictegravir, emtricit­abine, and tenofovir and tenofovir alafenamide group had a lower incidence alafenamide group, and one elective termination in the of drug-related adverse events than did those in the dolutegravir, abacavir, and lamivudine group. Ten (13%) of dolutegravir, abacavir, and lamivudine group (94 [30%] of 77 women in study 2 had thirteen confirmed pregnancies, 314 vs 132 [42%] of 315; p=0·0021; table 3). These events six (15%) of 40 in the bictegravir, emtricit­abine, and were primarily mild or moderate in severity. The tenofovir alafenamide group with nine confirmed difference between groups was driven mainly by the pregnancies and four (11%) of 37 in the dolutegravir, significant differencen i drug-related nausea reported emtricitabine, and tenofovir alafenamide group with four soon after initiation of blinded study drug, which occurred confirmed pregnancies. In the bictegravir, emtricitabine, in 18 (6%) participants in the bictegravir, emtricitabine, and tenofovir alafenamide group, the pregnancies resulted and tenofovir alafenamide group versus 56 (18%) in the in uncomplicated term birth (n=4), spontaneous abortion dolutegravir, abacavir, and lamivudine group (p<0·0001). (n=3), elective termination (n=1), and one pregnancy is The treatment-related adverse events reported in at least ongoing. In the dolutegravir, emtricitabine, and tenofovir 2% of participants in study 1 are shown in the appendix alafenamide group, the pregnancies resulted in (p 8). In study 2, treatment-related adverse events were uncomplicated term birth (n=3) and elective termination reported in 71 (22%) participants in the bictegravir, (n=1). No infant congenital abnormalities were reported in emtricitabine, and tenofovir alafenamide group and either study. 95 (29%) in the dolutegravir, emtricitabine, and tenofovir The overall laboratory safety profiles at week 144 in both alafenamide group. Treatment-related adverse events that studies were similar to those observed at weeks 48 and 96. were reported in at least 2% of participants in study 2 are In study 1, grade 3 or 4 laboratory abnormalities were shown in the appendix (p 9). In study 2, no single adverse reported for 83 (26%) of 314 participants in the bictegravir, event (MedDRA preferred term) met the 5% threshold for emtricitabine, and tenofovir alafenamide group and between group statistical comparison. No drug-related 80 (25%) of 315 in the dolutegravir, abacavir, and adverse events of grade 3 or higher were reported in more lamivudine group (appendix p 10). In study 2, grade 3 or 4 than two participants in either group across both studies. laboratory abnormalities were reported for 77 (25%) of In study 1, two deaths occurred before week 96, as 320 participants in the bictegravir, emtricitabine, and previously reported,5,7 (one drug overdose and one suicide tenofovir alafenamide group and 74 (23%) of 325 in the in the bictegravir, emtricitabine, and tenofovir alafenamide dolutegravir plus coformulated emtricitabine and group). One death (overdose in the dolutegravir, abacavir, tenofovir alafenamide group (appendix p 11). In both and lamivudine group) occurred after week 96. None of the studies, incidence and types of abnormalities were deaths in study 1 were considered to be related to treatment. generally balanced between treatment groups. Most were In study 2, eight (1%) participants died during the study, transient and resolved on therapy. We found no abnormal four (1%) in each group. In the bictegravir, emtricitabine, electrocardiogram findings before week 144 that were and tenofovir alafenamide group sudden cardiac death was associated with either treatment in each study. reported in one participant after week 96; the other deaths No cases of proximal tubulopathy or Fanconi syndrome (one cardiac arrest following appendicitis and septic shock, were reported in either study. No study participant who one gastric adenocarcinoma, and one hypertensive heart received bictegravir, emtricitabine, and tenofovir alafe­ disease and congestive cardiac failure) occurred before namide in either study discontinued because of a renal week 96, as previously reported.6,8 In the dolutegravir, adverse event; one individual in the dolutegravir, abacavir, Spine Hip 2 Bictegravir, emtricitabine, and tenofovir alafenamide Dolutegravir, abacavir, and 1 lamivudine

+0·04% 0

−0·37% −1·02% –1

−1·29% Mean change from baseline, % (95% CI) Mean change from baseline, % (95% p=0·26 p=0·39 –2 0 48 96 144 0 48 96 144 Week Week Number at risk Bictegravir, emtricitabine, 304 284 281 264 243 300 278 275 258 236 and tenofovir alafenamide Dolutegravir, abacavir, and 299 287 283 266 244 297 285 279 265 240 lamivudine

Figure 4: Mean percent change from baseline at weeks 24, 48, 96, and 144 in lumbar spine and hip bone mineral density by DXA in Study 1 Bars show 95% CIs. and lamivudine group discontinued treatment in study 1 pp 13, 17). In study 1, the median change in weight from because of renal failure, as previously reported.5,7 Small baseline was +4·1 kg (IQR 0·3–8·7) in the bictegravir, increases from baseline in median serum creatinine and emtricitabine, and tenofovir alafenamide group and decreases in eGFR were seen at week 144 for both groups +3·5 kg (0·0–7·7) in the dolutegravir, abacavir, and in each study (appendix p 12). At 144 weeks in study 1, lamivudine group (p=0·196). In study 2, the median percentage changes in quantitative proteinuria (total change was +4·4 kg (1·0–9·0) in the bictegravir, urinary albumin to urine creatinine ratio) and tubular emtricitabine, and tenofovir alafenamide group and proteinuria (retinol binding protein and β2-microglobulin +5·0 kg (0·5–9·7) in the dolutegravir, emtricitabine, and to urine creatinine ratios) were similar between groups tenofovir alafenamide group (p=0·649). (appendix p 12). In study 1, changes from baseline in fasting HDL and Discussion triglycerides were similar between groups at week 144 Results from these two large, randomised, phase 3 trials (appendix p 13). Significant differences were measured offer 3-year data showing non-inferiority of coformulated in the median changes from baseline in fasting total bictegravir, emtricitabine, and tenofovir alafenamide to cholesterol (14 mg/dL vs 10 mg/dL; p=0·034), direct two dolutegravir-containing regimens for the initial LDL (21 mg/dL vs 14 mg/dL; p=0·004), and total treatment of people with HIV. In 144 weeks of follow-up, cholesterol to HDL ratio (–0·1 vs –0·3; p=0·007) at no participants had treatment-emergent resistance week 144 in the bictegravir, emtricitabine, and tenofovir detected, which underscores the high barrier to resistance alafenamide versus the dolutegravir, abacavir, and for standard-of-care regimens containing either bictegravir lamivudine groups. We found no differences between or dolutegravir­ in combination with two nucleoside groups in initiation of lipid-modifying drugs during the reverse-transcriptase inhibitors. Minimising the risk of study for the bictegravir, emtricitabine, and tenofovir emergent drug resistance is essential to the lifelong alafenamide group compared­ with the dolutegravir, durability of HIV treatment. abacavir, and lamivudine group before week 144 (16 [5%] As expected, given the inhibition of renal creatinine of 314 vs 16 [5%] of 315; p=1·00). No differences in transporters by both bictegravir and dolutegravir, we median changes from baseline in fasting lipid found small increases from baseline in serum creatinine parameters at week 144 were noted for the bictegravir, that occurred early. eGFR remained stable in the emtricitabine, and tenofovir alafenamide group and treatment groups from week 4 until week 144 and did not dolutegravir, emtricitabine and tenofovir alafenamide show evidence of renal toxicity with longer exposure to group in study 2 (appendix p 13). the treatment regimens. Similarly, in study 1, which We found small changes from baseline in hip and included longitudinal monitoring of bone density lumbar spine bone mineral density that were similar using dual-energy x-ray absorptiometry, we found no between both groups in study 1. Mean percentage differences in bone mineral density changes between the changes at week 144 were −1·02% versus −1·29% at the bictegravir, emtricitabine, and tenofovir alafenamide and hip and −0·37% versus 0·04% at lumbar spine (figure 4). dolutegravir, abacavir, and lamivudine groups. Both We found similar changes from baseline in weight at treatments showed similar, well-described declines after week 144 in both groups in each study (appendix antiretroviral treatment was started, reaching a nadir at week 24 for the spine before increasing towards baseline Overall, these results show that all three regimens were and a plateau after week 48 for measurements at the hip. efficacious and safe. In distinguishing among regimens, Lipids changed from baseline in all groups. Participants the higher rates of treatment-related adverse events and in study 1 receiving bictegravir, emtricitabine, and the limitations to the use of dolutegravir, abacavir, and tenofovir alafenamide had modestly larger increases in lamivudine, including among those with hepatitis B total and direct LDL cholesterol, as well as smaller virus co-infection and the requirement for HLA-B*5701 reductions in total cholesterol to HDL ratio, than did screening, support the use of bictegravir, emtricitabine, those receiving dolutegravir, abacavir, and lamivudine. and tenofovir alafenamide, or alternatively dolutegravir, Consistent with data from previous studies in emtricitabine, and tenofovir alafenamide in particular treatment-naive populations,11–16 participants in both populations. studies had weight gain. Interestingly, statistically In summary, as was observed at weeks 48 and 96, significant differences in change in weight from baseline bictegravir, emtricitabine, and tenofovir alafenamide at between the treatment groups were not observed in 144 weeks of therapy continued to be non-inferior to either study at week 144. Weight gain and obesity-related dolutegravir-containing regimens, with no emergent morbidity and mortality among people with HIV is drug resistance or proximal renal tubulopathy detected, increasingly recognised as an emerging concern and has but with a better gastrointestinal tolerability profile. been associated with antiretroviral drugs and patient Further open-label follow-up for a total of 5 years of characteristics.11,17–22 exposure in each study will provide further information Pooled analysis, which included data from Studies 1 about the safety and durable efficacy of bictegravir, and 2, showed that newer antiretroviral drugs, including emtricitabine, and tenofovir alafenamide in people with bictegravir, dolutegravir, and tenofovir alafenamide, were HIV. consistently associated with more weight gain than older Contributors comparators and results suggested that improved CO, ED, PES, JRA, SKG, CM, JLS, H-JS, DW, FM, MAT, DP, DH, JAF, tolerability of treatments might be a contributing factor CB, and AC enrolled participants. HM designed the study. HH and RA 11 did the data analyses, which were reviewed and interpreted by DMB, to differential weight gain over time. The metabolic SEC, and HM. The first draft was written by CO, SEC and HM. All implications of the weight gain and lipid changes authors reviewed and interpreted analyses of data, contributed edits of observed in these studies are unknown and warrant the final report, and approved the draft manuscript. CO and HM made further study. Biological mechanisms that might the decision to submit the manuscript for publication. contribute to differential weight gain on antiretroviral Declaration of interests therapy are yet to be understood. CO reports receiving grants from Gilead Sciences relevant to the submitted work, as well as grants, personal fees, and non-financial from These studies have several limitations. Importantly, Gilead Sciences, ViiV Healthcare, MSD, Janssen, and GlaxoSmithKline the low enrolment of women and people with advanced (GSK) for travel sponsorship and speaker’s bureaus, outside of the HIV disease means that the findings cannot be submitted work. ED has received research support and personal fees assumed to be generalisable to these populations. from Gilead Sciences for participating in advisory boards and speaker’s bureaus and has received personal fees from ViiV Healthcare, Janssen, Given the paucity of data supporting the use of and Theratechnologies for participating in advisory boards. PES has bictegravir, emtricitabine, and tenofovir alafenamide in received a grant from Gilead Sciences for this study, as well as grants pregnant women, we restricted enrolment only to and personal fees from Gilead Sciences, and personal fees from women of childbearing potential who were willing to Janssen, Merck, and ViiV Healthcare for participating in scientific advisory boards. JRA reports receiving payment from Gilead Sciences to conform to the protocol’s stringent his institution for participating in the reported in this Article requirements. An international study23 of 470 women and has received grants from ViiV Healthcare and Gilead Sciences, as with suppressed plasma HIV-1 RNA showed high rates well as personal fees from ViiV Healthcare, Gilead, Janssen, MSD, Teva, of continued suppression after switching to bictegravir, and Aelix, outside of the submitted work. SKG reports research grants received from Gilead Sciences for the recruitment and enrollment of emtricitabine, and tenofovir alafenamide, with no study participants during the conduct of the study, as well as personal discontinuations due to adverse events. Although our fees and non-financial support from Gilead Sciences for participating in results are similar, ours are based on too small a sample advisory boards, including consultation fees travel support to scientific of women to be considered definitively supportive of conferences to present clinical trial data, and grants (unrestricted investigator-initiated grant support) and personal fees (including this finding. Given the known increased risk for weight consultation fees) from GSK and ViiV Healthcare for participating in gain in women and people with advanced HIV disease advisory boards. CM has received grants and personal fees from Gilead who start antiretroviral therapy, a greater degree of Sciences and ViiV Healthcare, and grants from Janssen. JLS reports weight gain might be expected in a cohort with higher grants from Gilead Sciences relevant to the submitted work, as well as other compensation from Gilead Sciences for participating in speaker’s percentages of women and people with advanced HIV. bureaus. H-JS reports personal fees and travel reimbursement from In a pooled analysis24 of week-144 results for both Gilead Sciences as honoraria for scientific advice and presentation as studies, the median changes in weight for women were well as trial documentation fees; personal fees from Janssen, , and 5·0 kg in the bictegravir, emtricitabine, and tenofovir MSD as honoraria for scientific advice and presentation as well as trial documentation fees; and trial documentation fees from ViiV Healthcare. alafenamide group, 7·9 kg in the dolutegravir, abacavir, DW has received grants and personal fees from Gilead Sciences, ViiV and lamivudine group, and 4·9 kg in the dolutegravir, Healthcare, and Merck, and personal fees from Janssen. FM reports emtricitabine tenofovir alafenamide group. grants and personal fees from ViiV Healthcare, Janssen, Gilead Sciences, and MSD, and grants from Abbott. MAT has received research support 6 Stellbrink H-J, Arribas JR, Stephens JL, et al. Co-formulated from Gilead Sciences for this study to AIDS Research Consortium of bictegravir, emtricitabine, and tenofovir alafenamide versus Atlanta relevant to the submitted work, as well as clinical trial research dolutegravir with emtricitabine and tenofovir alafenamide for initial funding from Bristol-Myers Squibb (BMS), Cepheid, Cytodyn, Gilead treatment of HIV-1 infection: week 96 results from a randomised, Sciences, GSK, MSD, and ViiV Healthcare to AIDS Research double-blind, multicentre, phase 3, non-inferiority trial. Lancet HIV 2019; 6: e364–72. Consortium of Atlanta, outside of the submitted work. DP reports receiving grants from Gilead Sciences relevant to the submitted work, 7 Gallant J, Lazzarin A, Mills A, et al. Bictegravir, emtricitabine, and tenofovir alafenamide versus dolutegravir, abacavir, and lamivudine and grants and personal fees from ViiV Healthcare and Gilead Sciences, for initial treatment of HIV-1 infection (GS-US-380-1489): a double- and personal fees from Janssen, outside of the submitted work. blind, multicentre, phase 3, randomised controlled non-inferiority DH reports that her institution receives revenue for the conduct of trial. Lancet 2017; 390: 2063–72. clinical trials from Gilead Sciences, Merck, ViiV Healthcare, and 8 Sax PE, Pozniak A, Montes ML, et al. Coformulated bictegravir, Janssen, and reports receiving personal fees from Gilead Sciences, Merk emtricitabine, and tenofovir alafenamide versus dolutegravir with and ViiV Healthcare for participating in speaker’s bureaus, and personal emtricitabine and tenofovir alafenamide, for initial treatment of fees from Gilead Sciences and ViiV Healthcare for participating in HIV-1 infection (GS-US-380-1490): a randomised, double-blind, advisory boards. JAF reports receiving grants from Gilead Sciences as multicentre, phase 3, non-inferiority trial. Lancet 2017; 390: 2073–82. research support paid to Kaiser Foundation Research Institute. CB has 9 Smith F, Hammerstorm T, Soon G, et al. A meta-analysis to assess received personal fees aand travel reimbursement from Gilead Sciences the FDA DAVP’s TLOVR algorithm in HIV submissions. Drug Inf J for participating in advisory boards and speaker’s bureaus, relevant to 2011; 45: 291–300. this study; personal fees and other potential conflicts of interests from 10 Taiwo BO, Zheng L, Stefanescu A, et al. ACTG A5353: a pilot study Gilead Sciences for participating in advisory boards and speaker’s of dolutegravir plus lamivudine for initial treatment of human bureaus; personal fees and travel reimbursement from ViiV Healthcare immunodeficiency virus-1 (HIV-1)–infected participants with HIV-1 for participating in Speaker’s Bureaus; and research grants from GSK RNA <500000 copies/mL. Clin Infect Dis 2018; 66: 1689–97. Novo Nordisk, BMS, Janssen, Roche, Colucid, Elcelyx, and Sangamo for 11 Sax PE, Erlandson KM, Lake JE, et al. Weight gain following serving as a principal investigator; and from Daiichi Sankyo for serving initiation of antiretroviral therapy: risk factors in randomized as a sub-principal investigator, outside of the submitted work. AC has comparative clinical trials. Clin Infect Dis 2019; published online Oct 14. DOI:10.1093/cid/ciz999. received non-financial support from Gilead Sciences for conference attendance, non-financial support from GSK from conference travel, 12 Taramasso L, Ricci E, Menzaghi B, et al. Weight gain: a possible side effect of all antiretrovirals. Open Forum Infect Dis 2017; and personal fees from ViiV Healthcare, GSK, and Gilead Sciences for 4: ofx239. advisory boards and meetings. HH, RA, DMB, SEC, and HM are 13 Tate T, Willig AL, Willig JH, et al. HIV infection and obesity: employees of Gilead Sciences and shareholders of Gilead Sciences stock. where did all the wasting go? Antivir Ther 2012; 17: 1281–89. Data sharing 14 Lakey W, Yang L-Y, Yancy W, et al. Short communication: from Gilead Sciences shares anonymised individual patient data upon request wasting to obesity: initial antiretroviral therapy and weight gain in or as required by law or regulation with external researchers. Approval HIV-infected persons. AIDS Res Hum Retroviruses 2013; 29: 435–40. of such requests is at Gilead Sciences’s discretion and is dependent on the 15 Yuh B, Tate J, Butt AA, et al. Weight change after antiretroviral nature of the request, the merit of the research proposed, the availability therapy and mortality. Clin Infect Dis 2015; 60: 1852–59. of the data, and the intended use of the data. Data requests should be sent 16 Sharma A, Hoover DR, Shi Q, et al. Relationship between body to [email protected]. mass index and mortality in HIV-infected HAART users in the women’s interagency HIV Study. PLoS One 2015; 10: e0143740. Acknowledgments 17 Norwood J, Turner M, Bofill C, et al. Brief report: weight gain in This study was sponsored by Gilead Sciences. We thank the persons with HIV switched from -based to integrase individuals who participated in this trial and their families, the strand transfer inhibitor-based regimens. principal investigators (appendix pp 1–6) and their staff, the Gilead J Acquir Immune Defic Syndr 2017; 76: 527–31. Sciences study staff, and Anna Kido (Gilead Sciences employee) for 18 Koethe JR, Jenkins CA, Lau B2, et al. Rising obesity prevalence and providing editorial assistance. Parts of this study have been presented weight gain among adults starting antiretroviral therapy in the at the 17th European AIDS Conference 2019, Nov 6–9, Basel, United States and Canada. AIDS Res Hum Retroviruses 2016; Switzerland. 32: 50–58. 19 Finucane MM, Stevens GA, Cowan M, et al. National, regional, and References global trends in body mass index since 1980: Systematic analysis of 1 European AIDS Clinical Society. Guidelines, version 9.1, health examination surveys and epidemiological studies with October 2018. http://www.eacsociety.org/files/2018_ 960 country-years and 9·1 million participants. Lancet 2011; guidelines-9.1-english.pdf (accessed Nov 4, 2019). 377: 557–67. 2 Panel on Antiretroviral Guidelines for Adults and Adolescents. 20 Hill JO, Wyatt HR, Reed GW, Peters JC. et al. Obesity and the Guidelines for the use of antiretroviral agents in adults and environment: where do we go from here? Science 2003; 299: 853–55. adolescents living with HIV. Department of Health and Human Services. (http://www.aidsinfo.nih.gov/ContentFiles/ 21 Bourgi K, Jenkins C, Rebeiro PF, et al. Greater weight gain among AdultandAdolescentGL.pdf (accessed Nov 4, 2019). treatment-naïve persons starting integrase inhibitors. Conference on Retroviruses and Opportunistic Infections; Seattle, WA, USA; 3 Saag MS, Benson CA, Gandhi RT, et al. Antiretroviral drugs for March 4–7, 2019 (abstr 670). treatment and prevention of HIV infection in adults: 2018 recommendations of the International Antiviral Society-USA Panel. 22 Bhagwat P, Ofotokun I, McComsey GA, et al. Changes in waist JAMA 2018; 320: 379–96. circumference in HIV-Infected individuals initiating a raltegravir or protease inhibitor regimen: effects of sex and race. 4 WHO. Updated recommendations on first-line and second-line Open Forum Infect Dis 2018; 5: ofy201. antiretroviral regimens and post-exposure prophylaxis and recommendations on early infant diagnosis of HIV: interim 23 Kityo C, Hagins D, Koenig E, et al. Switching to fixed-dose guidance. Geneva: World Health Organization, 2018. bictegravir, emtricitabine, and tenofovir alafenamide (B/F/TAF) in https://apps.who.int/iris/bitstream/handle/10665/277395/ virologically suppressed -1 infected women: a randomized, open- WHO-CDS-HIV-18.51-eng.pdf?ua=1 (accessed Nov 4, 2019). label, multicenter, active-controlled, phase 3, noninferiority trial. J Acquir Immune Defic Syndr 2019; 82: 321–28. 5 Wohl DA Yazdanpanah Y Baumgarten A, et al. Bictegravir combined with emtricitabine and tenofovir alafenamide versus 24 Orkin C, Ajana F, Kityo C, et al. Efficacy and safety of bictegravir/ dolutegravir, abacavir, and lamivudine for initial treatment of emtricitabine/tenofovir alafenamide vs comparators in cis-women HIV-1 infection: week 96 results from a randomised, double-blind, and girls (living with HIV): an analysis of 5 clinical trials. multicentre, phase 3, non-inferiority trial. Lancet HIV 2019; 17th European AIDS Conference; Basel, Switzerland; Nov 6–9, 2019 6: e355–63. (abstr PS7/6).