Pipeline ART: tenofovir alafenamide (TAF)

CROI logoSimon Collins, HIV i-Base

Of several new pipeline compounds with new data at CROI 2015, the most advanced in development was tenofovir alafenomide (TAF), the new version of tenofovir.

TAF is notable for a low milligram dose (10 mg or 25 mg) that achieves intracellular drug levels that are 4-fold higher and plasma levels that are 90% lower, than with 300 mg of tenofovir disoproxil fumarate (TDF).

In the first TAF oral presentation, David Wohl from University of North Carolina presented results from a prespecified combined analysis of two randomised TAF studies (0104 and 0111). [1]

The development programme for TAF is notable for prioritising coformulations rather than the individual drug and these studies compared two fixed-dose combinations (FDCs): elvitegravir/cobicistat/FTC/TAF (n=866) vs elvitegravir/cobicistat/FTC/TDF (Stribild), (n=867), starting with similar baseline characteristics and finding similar results. Dosing for each FDC is one pill, once-daily.

Both studies were randomised, double-blind, placebo controlled non-inferiority studies (lower margin of 95% CI = 12%), with primary endpoints of viral suppression to <50 copies/mL at week 48.

Baseline characteristics were closely matched between arms. Median age was 44 years; 85% were men, 15% women; 25% had black or African descent and 19% were Hispanic/Latino. The median CD4 and viral load were 405 cells/mm3 and 4.58 log copies/mL, respectively; with 12% <200 cells/mm3 and 23% >100,000 copies/mL.

At week 48, viral load was <50 copies/mL in 92% (TAF) vs 90% (TDF), difference +2.0% (95% CI: 0.7% to +4.7%). Approximately 5% (n=45) and 8% (n=71) discontinued treatment during the study, leaving 821 vs 796 participants in the primary analysis. When stratified by baseline viral load below/above 100,000 copies/mL results were 94% vs 91% (above) and 87% vs 89% (below), in the TAF vs TDF arms respectively. CD4 increases were also similar when stratified by baseline CD4 below (86% vs 89%) and above (86% vs 89%) 200 cells/mm3.

Efficacy results were not affected by age (above/below 50 years) or sex, but were slightly lower in both arms in black vs non-black participants: 88% vs 83% (below 50) and 94% vs 93% (above 50).

Increases in CD4 count were significantly higher with TAF: +211 vs +181 cells/mm3, p=0.024.

Of 4% of participants in each arm with viral failure, 1.8% (n=16) vs 2.2% (n=19) had resistance testing, with similare low rates of resistance to either NRTI (n=7 vs 5) or NNRTIs (n=5 vs 3).

The pattern of similar results continued for side effects: approximately 45% were drug-related but only 1% were grade 3/4, and 0.3% were serious – with discontinuations by only 0.9% (n=8) and 1.5% (n=15), in the TAF vs TDF arms, respectively. Five deaths were attibuted to stroke (1), alcohol intoxication (1), alcohol and drug intoxication (1) and myocardial infarction (2).

Common side effects (18% diarrhoea, 16% nausea, 13% headache, 12% upper respiratory tract infection, etc) were mild, with no cases of proximal renal tubulopathy in either study.

With marginal differences between FDCs for efficacy endpoints, the interest shifted to bone, renal and lipid results presented the following day by Paul Sax from Brigham and Women’s Hospital and Harvard Medical School, Boston.

In this analysis, renal differences included less reduction from baseline in eGFR (Cockroft-Gault) in the combined TAF compared to the TDF groups: median -6.6 vs -11.2 mL/min; p <0.001. The four renal discontinuations were all in the TDF groups: renal failure (2), decreased GFR (1), nephropathy (1). Significant median percentage changes from baseline of four sub-clinical markers of proteinuria all favoured TAF: protein -3% vs +20%, albumin -3% vs +7%, retinol binding protein +9 % vs +51% and beta2 microglobumin -32% vs +24%; all p < 0.001.

Mean percentage changes in bone measured by DEXA also favoured TAF vs TDF: at spine -1.30 % vs -2.86 % and hip -0.66 % vs -2.95 %; both p <0.001.

Finally, median changes from baseline in lipid parameters (mg/dL) all increased more in the TAF vs TDF arms: TC +29 vs +14; LDL +14 vs +5; HDL +7 vs +4; (all p <0.001); and TG +19 vs +8, p=0.027. However, there were no significant differences between arms in change from baseline of the TC:HDL ratio (p=0.84) or in the use of lipid lowering drugs (p=0.42).

A poster at CROI provided preliminary renal, bone and lipid data on use of TAF in patients with moderate-mild (stage 3-2) renal impairment, defined as eGFR 30-69 mL/min. [3]

This was 48-week data from a 96 week Phase 3 study in 242 stable patients on treatment (with or without TDF) with an undetectable viral load (<50 copies/mL) for at least six months, who were switched to the single tablet combination of elvitegravir/cobicistat/FTC/TAF (E/C/T/TAF). The primary endpoint was the change in eGFR from baseline to week 24. Hepatitis B/C were exclusion criteria. Patients were stratified by eGFR < 50 mL/min (n=59) and >50 mL/min (n=162).

Baseline characteristics included median age 58 (IQR: 52, 65) years, 21% women, 18% African descent, median CD4 632 cells/mm3. Differences in the <50 vs >50 groups included re-switch TDF use in 58% vs 69%, hypertension in 50% vs 34%, median eGFR 43 vs 60 mL/min/1.73 m2 (57 vs 77 when adjusted for age and sex), clinically significant proteinuria 58% vs 35% and clinically significant albuminuria in 64% vs 32%, respectively. Boosted PI was used by 44% of the overall group (breakdown by TDF use not provided).

At week 48, 222/242 (92%) remained virologically suppressed (<50 copies/mL), with data not available for 22 people (7%). There were seven discontinuations due to side effects: two for renal failure (eGFR decline), diarrhoea, choking, fatigue, joint swelling, sleep disorder and bladder cancer.

At week 24, median (IQR) changes in eGRF (Cockroft-Gault) was -0.4 (-4.8, +4.5) and adjusted eGFR (CKD-EPI cystatin C) was +3.8 (-4.8, +11.2) mL/min/1.73 m2. Changes by baseline stratification appeared similar. Actual eGFR measured by iohexol clearance in a sub group of 32 patients remained similar at baseline and at weeks 2, 4, 8, and 24. Changes in proteinuria and albuminuria remained stable or improved for nearly all patients, with only 5%, 2% and 0% worsening in participants with baseline grade 0, 1 and 2, respectively. There were no cases of proximal renal tubulopathy or Fanconi syndrome.

Median (IQR) changes in bone mineral density (BMD) at week 48 increased by 1.9% (-0.3, +4.3) n the spine and 0.9% (-0.3, +2.7) in the hip (both p<0.001), although BMD was reduced in more than 25% of patients at both regions.

Changes in lipids (mg/dL) from baseline to week 48 showed differences when stratified by prior TDF use: TC +19 vs -11; LDL +7 vs – 4; HDL +1 vs -4; and TG +12 vs -1; with changes in TC:HDL ratio of +0.3 vs +0.2, all in the prior-TDF vs no-TDF groups respectively.


TAF is probably a better drug than TDF, and the sub-clinical markers may have advantages. However, these were not sufficiently important for patient care for Gilead to prioritise it’s development.

Gilead had in vitro data on the potential benefits of TAF (formerly GS-7340) in 2001 and held back on development for over a decade before presenting Phase 1 data at CROI in 2011. [4, 5]


Unless stated otherwise, all references are to the Programme and Abstracts of the 2015 Conference on Retroviruses and Opportunistic Infections (CROI 2015), 23-26 February 2015, Seattle.

  1. Wohl D et al. Tenofovir alafenamide (TAF) in a single-tablet regimen in initial HIV-1 therapy. 2015 Conference on Retroviruses and Opportunistic Infections (CROI 2015), 23-26 February 2015, Seattle. Oral late breaker abstract 113LB.
  2. Sax P et al. Renal and bone safety of tenofovir alafenamide vs tenofovir disoproxil fumarate. 2015 Conference on Retroviruses and Opportunistic Infections (CROI 2015), 23-26 February 2015, Seattle. Oral late breaker abstract 143LB.

  3. Pozniak A et al. Safety of tenofovir alafenamide in renal impairment. 2015 Conference on Retroviruses and Opportunistic Infections (CROI 2015), 23-26 February 2015, Seattle. Poster abstract 795.
  4. Eisenberg EJ, He GX, Lee WA. Metabolism of GS-7340, a novel phenyl monophosphoramidate intracellular prodrug of PMPA, in blood. Nucleosides Nucleotides Nucleic Acids. 2001 Apr-Jul;20(4-7):1091-8. doi: 10.1081/NCN-100002496.
  5. Markowitz M, Zolopa A, Ruane P, et al. GS-7340 demonstrates greater declines in HIV-1 RNA than TDF during 14 days of monotherapy in HIV-1-infected subjects. 18th CROI; 27 February – 4 March 2011; Boston, MA. Late breaker oral abstract 152LB.

Links to other websites are current at date of posting but not maintained.