Comparison of NRTI combinations ARV-trial.com Comparison of NRTI combinations ZDV/3TC vs TDF + FTC Study 934 ABC/3TC vs TDF/FTC HEAT Study ACTG A5202 Study ASSERT Study FTC/TDF vs FTC/TAF Studies GS-US-292-0104 and GS-US-292-0111 1

ASSERT Study: ABC/3TC + EFV vs TDF/FTC + EFV ARV-trial.com ASSERT Study: ABC/3TC + EFV vs TDF/FTC + EFV Design Randomisation* 1 : 1 Open-label W48 W96 N = 195 > 18 years ARV-naïve HIV RNA > 1,000 c/mL Any CD4 cell count HLA-B*5701 allele negative ABC/3TC + EFV N = 197 TDF/FTC + EFV * Randomisation was stratified by screening eGFR (MDRD), race (black or non black), and BMI Primary analysis Intention to treat, exposed analysis : 2-sided significance level of 5%, with 90% power to detect a difference of 10 mL/min in eGFR (MDRD) change from baseline between arms ASSERT Post FA. JAIDS 2010;55:149-57 2

ASSERT Study: ABC/3TC + EFV vs TDF/FTC + EFV ARV-trial.com ASSERT Study: ABC/3TC + EFV vs TDF/FTC + EFV Baseline characteristics and patient disposition ABC/3TC + EFV N = 192 TDF/FTC + EFV N = 197 Median age, years 38 36 Female 17% 20% Black 14% 16% BMI (kg/m2) < 25 / > 25 66% / 33% 67% / 33% eGFR (mL/min/1.73m2) < 90 / > 90 32% / 68% 33% / 67% HIV RNA (log10 c/mL), median 5.01 5.12 CD4 cell count (/mm3), median 240 230 10-year coronary heart disease risk (Framingham score), median 2.97 2.90 Discontinuation by W48 63 (33%) 44 (23%) For lack of efficacy N =11 N = 2 For adverse event N = 25 N = 20 Lost to follow-up N = 5 Protocol deviation / Withdrew consent N = 5 / N = 6 N = 1 / N = 4 Other reasons N = 11 N = 12 ASSERT Post FA. JAIDS 2010;55:149-57 3

ASSERT Study: ABC/3TC + EFV vs TDF/FTC + EFV Adjusted mean change from baseline in eGFR by MDRD (ml/min/1.73 m2), ITT-e population Repeated measures mixed model analysis P = 0.435 -4 8 4 ABC/3TC + EFV TDF/FTC + EFV Change from BL in GFR by MDRD (ml/min/1.73 m²) -8 12 24 36 48 Treatment Difference at week 48 Week ABC/3TC EFV QD Better TDF/FTC EFV QD Better Treatment difference (ml/min/1.73 m²) 192 193 172 181 158 176 155 173 144 167 135 159 Subjects at visit Mean change in eGFR at week 96 : + 1.48 (ABC/3TC) - 1.15 (TDF/FTC) ASSERT Post FA. JAIDS 2010;55:149-57 ; Moyle GJ, Antiviral Therapy 2013;18:905-13

ASSERT Study: ABC/3TC + EFV vs TDF/FTC + EFV Secondary renal endpoints No differences between arms in proportion with significant declines in eGFR or with renal failure Percentage change from baseline at week 48 Urinary retinol-binding protein/creatinine ratio : TDF/FTC = + 50% vs ABC/3TC = 0% (p < 0.0001) Urinary b2-microglobulin/creatinine ratio : TDF/FTC = + 24% vs ABC/3TC = - 47% (p < 0.0001) Changes in bone density (DXA) at week 48 ABC/3TC + EFV TDF/FTC + EFV p Mean change in total hip loss - 1.9% - 3.6% < 0.001 Mean change in lumbar spine - 1.6% - 2.4% 0.036 BMD loss > 6% in the hip / in the spine 3% / 5% 13% / 10% Bone turnover markers : increased in both groups over the first 24 weeks (significantly greater with TDF/FTC), stabilizing or decreasing thereafter ASSERT Post FA. JAIDS 2010;55:149-57 ; Moyle GJ, Antiviral Therapy 2013;18:905-13 ; Stellbrink HJ. CID 2010;51:963-72

ASSERT Study: ABC/3TC + EFV vs TDF/FTC + EFV ARV-trial.com ASSERT Study: ABC/3TC + EFV vs TDF/FTC + EFV Efficacy at week 48 25 50 100 75 59.4 71.0 Adjusted difference 95% CI = 11.6% (2.2 ; 21.1) ITT, TLOVR ABC/3TC + EFV TDF/FTC + EFV HIV RNA < 50 c/mL % Protocol-defined virologic failure (< 1 log10 c/mL reduction in HIV RNA by W4, confirmed rebound ≥ 400 c/mL after confirmed HIV RNA < 400 c/mL by W24 or confirmed HIV RNA > 400 c/mL after W24) 6 on ABC/3TC + EFV Emergence of resistance : 3/6 2 patients with NNRTI mutations 1 patient with NNRTI mutation and K65R + D67N 2 on TDF/FTC + EFV = no resistance emergence Median CD4/mm3 increase at W48 : + 150 in both groups ASSERT Post FA. JAIDS 2010;55:149-57 6

ASSERT Study: ABC/3TC + EFV vs TDF/FTC + EFV ARV-trial.com ASSERT Study: ABC/3TC + EFV vs TDF/FTC + EFV Safety at W48 ABC/3TC + EFV N = 192 TDF/FTC + EFV N = 197 Drug-related adverse events 51% 47% Grade 2-4 29% 20% Most common grade 2-4 AEs : dizziness, abnormal dreams, drug hypersensitivity Drug hypersensitivity N = 12 (6%) N = 1 (< 1%) Clinically suspected ABC hypersensitivity N = 6 - Median changes in fasting lipids from baseline at W48 Total cholesterol, mg/dL 1.36 0.66 Triglycerides, mg/dL 0.23 0.05 LDL-cholesterol, mg/dL 0.81 0.39 HDL-cholesterol, mg/dL 0.38 0.28 Total cholesterol/HDL-cholesterol ratio - 0.559 -0.934 ASSERT Post FA. JAIDS 2010; 55:149-57 7

ASSERT Study: ABC/3TC + EFV vs TDF/FTC + EFV ARV-trial.com ASSERT Study: ABC/3TC + EFV vs TDF/FTC + EFV Conclusion No difference in estimated glomerular filtration rate between ABC/3TC + EFV and TDF/FTC + EFV However, increases in markers of tubular dysfunction in the TDF/FTC arm No differences in markers of glomerular dysfunction over 96 weeks Significant higher virologic efficacy for TDF/FTC + EFV Emergence of resistance to NNRTI and NRTI on virological failure with ABC/3TC + EFV but not with TDF/FTC + EFV Incidence of grade 4 adverse events and serious adverse events was higher in the ABC/3TC arm, mainly because of drug hypersensitivity or hypersensitivity Smaller increases in serum lipids with TDF/FTC Greater increases in bone turnover and decreases in hip and vertebral bone mineral density with TDF/FTC + EFV compared to ABC/3TC + EFV ASSERT Post FA. JAIDS 2010; 55:149-57 ; Moyle GJ, Antiviral Therapy 2013;18:905-13 8