D/C/F/TAF: phase 3 naive results and splitting PI-based FDC tablets

Simon Collins, HIV i-Base

Two oral presentations at EACS 2017 included new data on the fixed dose combination (FDC) of D/C/F/TAF: darunavir (800 mg), cobicistat (150 mg), emtricitabine (200 mg) and tenofovir alafenamide (10 mg).

Phase 3 AMBER study: 48-week results

48-week results from the phase 3 AMBER study in treatment-naive were presented by Chloe Orkin from Barts NHS Trust, London. [1]

This was an international, double-blind, active-controlled study that randomised 725 treatment-naive patients to either the single tablet D/C/F/TAF or a control group taking darunavir/cobicistat plus tenofovir DF/FTC.

This was a non-inferiority study (defined using 10% lower margin for the 95%CI) with a primary endpoint of viral suppression (<50 copies/mL) at week 48, and follow-up to week 96.

This was a generally young cohort, in early infection, reflecting earlier diagnosis in many countries. Baseline characteristics included median age 34 years (IQR: 27 to 42), 88% male, 82% white. Median CD4 count was 453 cells/mm3 (IQR: 333 to 601) with 7% <200 cells/mm3 and viral load was 4.5 log copies/mL (IQR: 4.1 to 4.9) with 17% >100,000 copies/mL. Baseline renal function was also high with median eGFR of 119 mL/min (IQR: 104 to 136).

At week 48, viral suppression <50 copies/ml was reported in 91% vs 88% of the FDC vs control group respectively (difference: 2.6%; 95%CI: –1.6 to +7.1), meeting criteria for non-inferiority. Viral load was >50 copies/mL in 4% vs 3% and discontinuations/missing data accounted for 4% vs 8%, all FDC vs control. Discontinuations due to side effects were lower for the FDC (2.2% vs 4.4%).

No significant differences in viral outcome were reported by sub group analysis (by age, gender, race, and baseline CD4 and viral load). Although the percentage differences generally favoured D/C/F/TAF, the low numbers for some groups generated wide confidence intervals. For example, for the key viral load stratification, response rates were higher with D/C/F/TAF (90% vs 80%; 95%CI: –3 to + 22) in participants with viral load above 100,000 copies/mL. For the subgroup with CD4 counts <200 cells/mm3 however, results favoured the control arm (72% vs 86%; 95%C: –37 to +8).

Side effects were generally mild and similar between groups and no grade 3/4 events occurred in more than 5% of participants. eGFR based on serum creatinine dropped during the first two weeks, significantly more in the control arm – by approximately 5.0 vs 8.0 mL/min/1.73m2 (p<0.0001) and then remained stable to week-48 in both groups. When eGFR was estimated by serum cystatin C, both arms increased over 48 weeks. Reductions in bone mineral density (BMD) at the spine occurred in both groups but were lower in the control arm only, as expected with tenofovir-based ART. BMD at the hip only reduced in the control arm, by –1.7% at week-24 and ­–2.7% at week 48.

Fasting lipids (TC, LDL, HDL and TG) increased between baseline and week 48 in both arms but were significantly greater in the FDC arm, with small absolute differences that were unlikely to have clinically relevant for most people. Although a marker of patients management rather than drug effect, and not protected by randomisation or study protocol, lipid lowering drugs were started by 2 (0.6%) vs 6 (1.7%) of the FDC vs control group (p=0.18 NS).

PK for whole, split and crushed tablet

The second presentation, by Kimberley Brown from Janssen, reported on bioavailability results from splitting and crushing the D/C/F/TAF tablet, with broadly comparable drug exposures for the split tablet but not the crushed version. [2]

This was a randomised, open-label, cross-over study in 30 HIV negative volunteers, using a single dose, taken 30 minutes after a standard breakfast, as (i) a complete tablet, (ii) two halves of a pill=cutter split tablet, or (iii) as a crushed tablet with the powder mixed in apple sauce.

Overall, plasma concentrations over 24 hours for each method of administration were closely similar at all timepoints for darunavir, cobicistat and FTC, with 90% confidence intervals (CI) for the mean difference falling within the 80% to 125% routinely used to show bioequivalence.

Although the mean difference in AUC for TAF was bioequivalent for the whole vs split tablet (mean ratio 97%; 90%CI 90 to 105) the crushed tablet resulted in approximately 20% reduced AUC (mean ratio 81%; 90%CI 47 to 88).


D/C/F/TAF has already been approved in the EU (in September 2017) and is currently filled with the US FDA.

In the questions after the presentations, the generally healthy cohort in AMBER was referred to several times. For example, in producing the highest viral response rates for darunavir than seen in earlier randomised phase 3 studies. This linked to questions about tolerability results in participants who had poorer baseline markers of higher risks for cardiovascular and renal disease.

Nevertheless, the results of this new formulation broadly reflected the individual component drugs.

The pharmacokinetic study results from splitting the single tablet in half are encouraging for people who need to take smaller pills. The lower AUC with the crushed tablet, however, means this should not be recommended until validated with further studies, even though TAF generally reaches concentrations that are well above the IC95.


  1. Orkin C et al. Week 48 results of AMBER: A phase 3, randomised, double-blind trial in antiretroviral treatment (ART)-naïve HIV-1-infected adults to evaluate the efficacy and safety of the once-daily, single-tablet regimen (STR) of darunavir/ cobicistat/ emtricitabine/ tenofovir alafenamide (D/C/F/TAF) versus darunavir/cobicistat (DRV/c) plus emtricitabine/ tenofovir disoproxil fumarate (FTC/TDF). 16th European AIDS Conference (EACS 2017), 25–27 October 2017, Milan. Oral abstract PS8/2.
  2. Brown K et al. Relative bioavailability of darunavir/cobicistat/emtricitabine/tenofovir alafenamide (D/C/F/TAF) single-tablet regimen when administered as a whole, split, or crushed tablet. 16th European AIDS Conference (EACS 2017), 25–27 October 2017, Milan. Oral abstract PS8/3.

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