W 42: Adaptive Design in Dose Selection Study of Next-in-Class NNRTI
Poster Presenter
Natalia Vostokova
COO
IPHARMA LLC Russian Federation
Objectives
Adaptive two-stage design with interim analysis for dose selection was used in Phase 2-3 study of new NNRTI VM-1500. The study objective was to select the optimal dose of VM-1500, and evaluate its efficacy and safety vs. Efavirenz (EFV) in combination with 2NRTI in treatment naïve patients with HIV.
Method
ORAL PRESENTATION SCHEDULED: Session 2B at 12:50- 1:00 PM
Randomized clinical trial of VM-1500 was conducted at 13 Russian sites. VM-1500 20 mg, 40 mg, and EFV were studied at Stage 1. Optimal dose of VM-1500 was selected based on viral load at Week 12. Additional patients were enrolled at Stage 2 to demonstrate non-inferiority of VM-1500 vs. EFV.
Results
The primary efficacy endpoints were the decrease of viral load <400 copies/ml by Week 12 (Stage 1) and <50 copies/ml by Week 24 (Stage 2). Safety endpoints included CNS adverse events of special interest (AEOI).
90 patients were randomized on VM-1500 20 mg, 40 mg, or EFV at Stage 1 (1:1:1 ratio). 30 (100%) patients on VM-1500 20 mg, 28 (93.3%) patients on VM-1500 40 mg, and 24 (80%) of patients on EFV competed 12 weeks of treatment. The target viral load of <400 copies/ml at Week 12 was achieved by 28/30 (93.3%) on VM-1500 20 mg, 25/29 (86.2%) on VM-1500 40 mg, and 22/27 (81.5%) patients on EFV. The difference between study groups and EFV was 11.8% for VM-1500 20 mg and 4.7% for VM-1500 40 mg; corresponding 95% CI lower bounds were -2.6 and -11.5% (both to the right of the non-inferiority margin d = -15%). Mean change of viral load (log10) at Week 12 was -2.9 on VM-1500 20 mg, -3.2 on VM-1500 40 mg, and -3.0 on EFV. AEs were reported in 21 (70.0%) patients on VM-1500 20 mg, 25 (86.2%) patients on VM-1500 40 mg, and 24 (85.7%) patients on EFV; AEOIs were reported in 6 (20.0%) patients on VM-1500 20 mg, 11 (37.9%) patients on VM-1500 40 mg, and 15 (53.6%) patients on EFV. Based on the primary efficacy endpoint assessment (<400 copies/ml in 93.3% patients) and the best safety profile with regards to AE and AEOI rate, VM-1500 20 mg was chosen for further investigation at Stage 2.
The additional interim analysis showed that among Stage 1 patients, 30 (100.0%) patients on VM-1500 20 mg and 23 (76.7%) patients on EFV competed 24 weeks of treatment. The target viral load of <50 copies/ml at Week 24 was achieved by 28/30 (93.3%) on VM-1500 20 mg and 19/27 (70.4%) patients on EFV. The difference between study groups was 23.0% with 95% CI lower bound 2.9% (to the right of the non-inferiority margin d = -15%). The preliminary results show potential to demonstrate not only non-inferiority, but also statistically significant difference between groups in the final analysis.
Conclusion
Adaptive clinical trial design refers to studies that allow modifying any design or hypothesis aspect based on the interim data analysis. Any adaptation is appropriate solely in accordance with the predefined plan and at preselected time points. Although adaptations in dose selection studies are still considered less-understood, its methodology can be successfully introduced in clinical programs of next-in-class drugs.
Next-in-class drugs are original patented medications with known targets which are similar to existing drugs in structure and mode of action. This allows for higher predictability of drug effects in humans, and possible achievement of better clinical results owing to the «fine-tuning» of the original molecule. Looking back over the complexity of the clinical paths of original innovative medications, the next-in-class drugs deserve shorter timelines and less expensive development.
Low-risk R&D strategy for the next-in-class drugs is justified by abundant clinical data of medications of the same pharmacological class that support the choice of predictable endpoints, non-inferiority hypothesis, accurate sample size calculation, and well-studied control.
Implementation of adaptive design in Phase 2-3 study of the new NNRTI VM-1500 requires significantly lower sample size and shorter timelines for testing the non-inferiority hypothesis vs. standard of care treatment with EFV in combination with 2NRTI as opposed to similar studies of earlier drugs of the same class. The adaptation includes the choice of optimal dose of the study drug for the second stage of the trial based on the interim analysis of efficacy and safety. Potential biases are controlled by dose blinding design and DMC for assessment of efficacy and safety endpoints. No changes to the initial statistical assumption and methods will be implemented during the trial.
Justified and tailored adaptive design can be recommended for clinical development of next-in-class drugs.
