Rilpivirine (TMC-278) vs efavirenz in treatment-naive patients: phase 3 results
Simon Collins, HIV i-Base
Results from two large international randomised phase 3 studies (ECHO and THRIVE) comparing rilpivirine to efavirenz were combined in one late-breaker presentation. Rilpivirine was developed with a 25mg dose due to phase 2 studies showing similar efficacy at 25mg, 50mg and 75mg and a caution over cardiovascular toxicity (QTc interval) at higher doses.
The two studies differed only in the use of nucleosides with ECHO using tenofovir/FTC in all patients and THRIVE allowing investigator choice. Each study randomised just under 700 treatment-naive patients with no NNRTI resistance and sensitivity to RTIs. The primary endpoint was viral load suppression <50 copies/mL at week 48 (ITT-TLOVR analysis) to demonstrate non-inferiority to efavirenz (lower margin 12%), with follow-up continuing to week 96.
Baseline characteristics of the 1368 patients included approximate median CD4 count 250 cells/mm3 (range 11,140), median viral load 5 log copies/mL (range 27), with just over 25% having a previous AIDS diagnosis, Gender ratio was 75% male: 25% female and mean age 36 years. Racial demographics were roughly 60% Caucasian, 24% Black and 12% Asian. Between 79% patients were coinfected with hepatitis B or C. Nucleoside choice in THRIVE was 60% tenofovir/FTC, 30% AZT/3TC and 10% abacavir/3TC.
At week 48, suppression to <50 copies/mL was achieved in 84% vs 82% patients in the rilpivirine vs efavirenz groups (pooled results difference +1.6; 95%CI 1.7 to +8.8, p<0.0001). This lower bound for the confidence interval was significantly above the 12% lower limit pre-specified for non-inferiority studies. CD4 increases were +192 vs + 176 cells/mm3 respectively.
Differences in the rilpivirine vs efavirenz arms were more apparent when looking at reasons for treatment failure, with 9% vs 5% reporting virological failure and approximately 2% vs 7% discontinuing due to side effects, respectively. Around 5% patients discontinued from each arm for other reasons.
In the rilpivirine vs efavirenz groups, 5.5% vs 2.6% of people who never suppressed <50 copies/mL and 3.5% vs 2.2% patients suppressed and then
No differences in virological response were reported by gender, race or geographical region, or by nucleoside backbone. However, by baseline viral load the pooled response rates were 90% vs 84% (difference +6.6: 95%CI +1.6, +11.5) in favour of rilpivirine in the <100,000 group and 77% vs 81% (difference 3.6: 95%CI 9.8, +2.5) in favour of efavirenz in the >100,000 group.
People whose treatment failed on rilpivirine developed higher rates of both NNRTI- (63% vs 54%) and NRTI-associated (68% vs 32%) mutations. Rilpivirine was associated with E138K, with 90% of these patients showing phenotypic cross-resistance to etravirine, essentially loosing the NNRTI class. People experiencing virological failure on efavirenz commonly developed K103N, which should retain sensitivity to etravirine.
Tolerability results favoured rilpivirine with comparisons below for rilpivirine vs efavirenz. While >90% of patients in each arm reported at least one side effect, grade 24 events related to study drug occurred in 16% vs 31%, p<0.0001) and discontinuations due to toxicity occurred in 3% vs 8%, p=0.0005. Neurological side effects occurred in 17% vs 38% (p<0.0001), psychiatric side effects in 15% vs 23% (p=0.0002), abnormal dreams in 8% vs 13% (p=0.0061) and rash in 3% vs 14% (p<0.0001).
Grade 3/4 laboratory abnormalities occurred in 11% vs 18% patients (p<0.001), with higher rates of ALT (1.5% vs 3.4%, p<0.05) and increases in LDL (0.7% vs 4.1%, p<0.0001), triglycerides 0.3% vs 2.2%, p<0.001) and total cholesterol (0.1 vs 2.5%, p<0.0001), all favouring rilpivirine.
Minimal change in mean serum creatinine in both groups with no grade 3/4 creatinine increases and no discontinuations due to renal side effects or cases of acute renal failure. No difference was seen in changes in QTc interval between TMC278 and efavirenz groups.
A one-pill once-daily fixed dose combination of rilpivirine plus tenofovir/FTC is already in development and bioequivalence to the separately dosed compounds were presented as a late breaker. 
This study from Gilead was an eight-day, randomised, single-dose, open-label, phase 1 study in 36 HIV-negative adults in fed conditions. Formulation bioequivalence was met based on 90% confidence intervals (CI) for the ratio of geometric least square means (GMR) for Cmax and AUC. All treatments were generally well tolerated with most adverse events mild in severity. Two participants did not complete the study.
Results from a granule formulation for paediatric dosing were also presented. 
Rilpivirine data were submitted to the FDA in July and if approved then both the single and 3-in-1 formulations could be available early in 2011.
While the low dose (25mg) makes it easier to develop as a fixed dose combination the lower rates of virological suppression seen when baseline viral load is >100,000 copies/mL may be related to drug exposure levels in some patients. While no clear relationship to dose and response were seen in the smaller phase 2 studies for rilpivirine, the low dose must increase the risk of suboptimal dosing in at least a small percentage of patients with either poorer absorption or higher clearance. The option of dose escalation might be useful to study in this specific patient group.
The option of an alternative to Atripla that has fewer CNS toxicities will clearly be welcomed, though potentially as a switch drug, given the cross-resistance to etravirine and higher risk of virological failure at higher CD4 counts.
- Cohen C et al. Pooled week 48 efficacy and safety results from ECHO and THRIVE, two double-blind, randomised phase III trials comparing TMC278 versus efavirenz in treatment-naive, HIV-1-infected patients. 18th IAS Conference, 1823 July 2010, Vienna. Oral abstract late breaker THLBB206. http://pag.aids2010.org/Abstracts.aspx?SID=1990&AID=17525
- Mathias A et al. Bioequivalence of the co-formulation of emtricitabine/rilpivirine/tenofovir DF. 18th IAS Conference, 1823 July 2010, Vienna. Oral abstract late breaker LBPE17. http://pag.aids2010.org/Abstracts.aspx?AID=17780
- Crauwels H et al. Relative bioavailability of a concept paediatric formulation of TMC278, an investigational non-nucleoside reverse transcriptase inhibitor (NNRTI). 18th IAS Conference, 2010, Vienna. Poster abstract THPE0158. http://pag.aids2010.org/Abstracts.aspx?AID=12590