Switch from TDF to TAF  GS-US Study

 Design  Endpoints –Primary: proportion of patients maintaining HIV RNA < 50 c/mL at W48 (ITT, snapshot) ; non-inferiority if lower margin of a two-sided 95% CI for the difference = -12%, 99% power –Secondary: percentage change in hip and spine bone mineral density, change in serum creatinine, and change in EFV-related symptom score at W48 Switch to E/C/F/TAF Continue TDF-based regimen Mills A. Lancet Infec Dis 2016;16:43-52 GS-US Randomisation 2 : 1 Open-label HIV+ ≥ 18 years On E/C/F/TDF or EFV/FTC/TDF or boosted ATV + FTC/TDF > 96 weeks HIV RNA < 50 c/mL eGFR (Cockroft-Gault) > 50 mL/min N = 477 N = 959 W48 GS-US Study: Switch TDF to TAF W96 *Randomisation was stratified by previous treatment regimen

E/C/F/TAF N = 959 TDF-based regimens N = 477 Median age, years4140 Female11%10% Race : white / black / asian68% / 18% / 6%66% / 21% / 7% HIV-1 RNA < 50 c/mL98% CD4 cell count (/mm 3 ), median eGFR, mL/min, median Cockroft-Gault CKD-EPI Dipstick proteinuria : grade 1 / grade 29% / < 1% Discontinuation by W48, n (%) For adverse event For lack of efficacy Investigator discretion Consent withdrawal Loss to follow-up Non-compliance Death 32 (3.3%) (8.4%) Baseline characteristics and patient disposition Mills A. Lancet Infec Dis 2016;16:43-52 GS-US GS-US Study: Switch TDF to TAF

Virologic outcome at W48 (ITT, snapshot) E/C/F/TAF (N = 959) TDF-based regimens (N = 477) % HIV RNA < 50 c/mL Difference (95% CI) = 4.1% (1.6 to 6.7) Virologic failure 1* 1.3 N = 10N = No virologic data Superiority of E/C/F/TAF * One patient in the TAF group with virologic failure (W8) had genotypic resistance : M184I/M. The patient resuppressed 4 weeks later without a change of regimen Mills A. Lancet Infec Dis 2016;16:43-52 GS-US GS-US Study: Switch TDF to TAF

HIV-1 RNA < 50 c/mL at W48 by prior treatment regimen, % Mills A. Lancet Infec Dis 2016;16:43-52 GS-US GS-US Study: Switch TDF to TAF E/C/F/TAF TDF-based regimens p < 0.001p = NSp = 0.02 Primary Endpoint 0.5 to to to to % CI = All prior regimensPrior EFV/FTC/TDF Prior boosted ATV + FTC/TDF Prior E/C/F/TDF

Changes in bone mineral density from baseline to W48 Mills A. Lancet Infec Dis 2016;16:43-52 GS-US GS-US Study: Switch TDF to TAF E/C/F/TAF TDF-based regimens Difference in least square means (95% CI) Hip BMD change (N) T-score mean change from baseline (0.10 – 0.15) ; p < Percent change from baseline1.47%- 0.34% 1.81 (1.49 – 2.13) ; p < to 3% increase in BMD > 3% increase 56% 21% 38% 8% p < Spine BMD change (N) T-score mean change from baseline (0.16 – 0.23) ; p < Percent change from baseline1.56%- 0.44% 2.00 (1.55 – 2.45) ; p < to 3% increase in BMD > 3% increase 41% 33% 34% 13% p <

Changes in bone mineral density from baseline to W48 in patients previously on ATV + r + FTC/TDF Mills A. Lancet Infec Dis 2016;16:43-52 GS-US GS-US Study: Switch TDF to TAF E/C/F/TAFATV + r + FTC/TDF Hip BMD mean change (N)4113 During 144 weeks of prior ATV + r + FTC/TDF- 4.58% Percent change from baseline to W %- 0.77% Spine BMD mean change (N)4216 During 144 weeks of prior ATV + r + FTC/TDF- 3.1% Percent change from baseline to W %- 0.74%

Change in Osteopenia/Osteoporosis Diagnosis (defined by T-Score) Spine Normal Osteopenia Osteoporosis Differences between E/C/F/TAF and TDF-based regimens were statistically significant (p < 0.001) BaselineWeek 48 E/C/F/TAF N = 912 BaselineWeek 48 TDF-Based Regimen N = 457 Hip BaselineWeek 48 E/C/F/TAF N = 902 BaselineWeek 48 TDF-Based Regimen N = 452 GS-US GS-US Study: Switch TDF to TAF % % Mills A. IAS 2015, Abs. TUAB0102; Mills A. Lancet Infec Dis 2016;16:

Serum creatinine and eGFR changes from baseline to W48 Median changes in markers of renal function from baseline to W48 Mills A. Lancet Infec Dis 2016;16:43-52 GS-US GS-US Study: Switch TDF to TAF E/C/F/TAF EFV/FTC/TDF (unboosted regimen) Other TDF-based regimens Mean serum creatinine change,  mol/L Median eGFR (Cockroft-Gault), mL/min E/C/F/TAFTDF-Based regimens p Fractional excretion of phosphate Fractional excretion of uric acid < TmP/GFR *, mg/dL * TmP/GFR : ratio of tubular maximum reabsorption of phosphate (TmP) to GFR

Renal Safety Results (Median % change in proteinuria) UPCR: urine protein:creatinine ratio ; UACR: urine albumin:creatinine ratio ; RBP, retinol-binding protein ; β-2-m: beta-2 microglobulin Mills A. Lancet Infec Dis 2016;16:43-52 GS-US GS-US Study: Switch TDF to TAF E/C/F/TAF TDF-based regimens Each difference between treatment arms was statistically significant (p < 0.001) RBP:Crβ-2-m:CrUPCRUACR Tubular Proteinuria

Adverse events, N (%) * Panic attack; memory loss, speech disturbance, and lack of motivation ; acute renal failure ; Reiter’s syndrome ; nausea, vomiting, and headache ; suicide attempt ; leg swelling and impaired concentration; depression ; and interstitial nephritis ** Abnormal dreams ; depression, insomnia, and irritability ; depression, insomnia, and nightmares ; elevated bilirubin; icterus (N = 2); increased forgetfulness; chronic kidney disease ; elevated serum creatinine ; Fanconi’s syndrome and mild jaundice; increased creatinine; and nephritic colic Mills A. Lancet Infec Dis 2016;16:43-52 GS-US GS-US Study: Switch TDF to TAF E/C/F/TAF N = 959 TDF-based regimens N = 477 Study-drug related adverse event21%16% Grade 3 or 4 adverse event9%11% Serious adverse event65 (7%)35 (7%) Study drug-related serious adverse event02 (< 1%) Premature study drug discontinuation for kidney-related adverse event for jaundice 9 (1%) * (3%) ** 5 (1 Fanconi) 3 (all on boosted ATV)

Most-common drug-related adverse event, % Mills A. Lancet Infec Dis 2016;16:43-52 GS-US GS-US Study: Switch TDF to TAF E/C/F/TAF N = 959 TDF-based regimens N = 477 Upper respiratory tract infection1611 Diarrhea109 Nasopharyngitis98 Headache74 Cough75 Syphilis56 Insomnia56 Arthralgia65 Bronchitis65 Depression46 Osteopenia65 Back pain55 Nausea53 Sinusitis55

Grade 2-4 laboratory abnormalities Mills A. Lancet Infec Dis 2016;16:43-52 GS-US GS-US Study: Switch TDF to TAF E/C/F/TAF N = 959 TDF-based regimens N = 477 Any abnormality25%31% Creatine kinase10% AST5%7% ALT5% Neutropenia4%3% Hypophosphatemia2%3% Hyperuricemia2%1% Alkaline phosphatase< 1% Leukopenia< 1% Platelets< 1% Total bilirubin< 1%24% Hemoglobin0< 1% Creatinine0< 1%

Fasting Lipids (mg/dL, median)  Participants initiating lipid-modifying medications: E/C/F/TAF: 8% ; TDF-based regimens: 6% Mills A. Lancet Infec Dis 2016;16:43-52 GS-US GS-US Study: Switch TDF to TAF Total CholesterolLDLHDLTriglyceridesTC:HDL Ratio p < p < p = p < p = E/C/F/TAF Baseline Week 48 TDF-Based Regimens Baseline Week

 Conclusion –Switching virologically suppressed patients to TAF regimen (EVG 150 mg, cobicistat 150 mg, emtricitabine 200 mg, and TAF 10 mg) was non inferior to the to four TDF-containing regimens at week 48 –It was also superior to continuing with the TDF-containing regimens. –Switching to the TAF group showed several important safety advantages of TAF over any of the TDF-containing regimens : Improvements in hip and spine bone mineral density Improvements in renal function –Decreases in serum creatinine in patients switched from a boosted regimen –Decreases in dipstick proteinuria, in quantitative tests of total urine protein, and total urine albumin, in specific proximal tubular proteins –Improvements in proximal renal tubular function (fractional excretion of uric acid, fractional excretion of phosphate, and renal tubular maximum reabsorption rate of phosphate to the glomerular filtration rate –Slightly higher lipid concentrations with TAF than with TDF-based regimens Mills A. Lancet Infec Dis 2016;16:43-52 GS-US GS-US Study: Switch TDF to TAF