2241. Utilizing a Clinical Utility Index in a Phase II, Randomized, Bayesian Response-Adaptive Design to Study the Efficacy and Safety of Gepotidacin (GSK2140944) in Patients with Acute Bacterial Skin and Skin Structure Infections
Session: Poster Abstract Session: New Antibiotics in Development
Saturday, October 29, 2016
Room: Poster Hall
Posters
  • T-GSK0660_ID Week Gepotidacin stats poster_for print.pdf (774.4 kB)
  • Background: Gepotidacin (GEP), a first in class novel triazaacenaphthylene bacterial topoisomerase inhibitor, inhibits bacterial replication and has in vitroactivity against key pathogens, including drug-resistant strains, associated with a range of conventional and biothreat infections. Study designs allowing adaptive randomization may limit exposure to suboptimal doses. Methods: This Phase 2 study evaluated efficacy, safety, and pharmacokinetics of 3 IV/oral doses of GEP in subjects with suspected Gram-positive acute bacterial skin and skin structure infections requiring hospitalization. The primary analysis consisted of computing a clinical utility index (CUI) measuring the tradeoff between Day 2-3 clinical outcome and rate of withdrawal due to drug-related adverse events (AE). The dose was successful if the CUI was likely to be greater than a threshold corresponding to approximately 75% efficacy and 2.5% withdrawal rates. An escalation scheme combined with response-adaptive randomization was employed to preferentially randomize subjects to doses with the highest utility. Bayesian decision rules based on utility were used to evaluate futility, allocation to treatment arms, and final dose selection. Randomization frequencies were adjusted after approximately every 18 subjects.   Results: Only 1 subject withdrew from this study due to a drug related AE. Thus decision rules were guided primarily by efficacy results. After the 40 subject run-in with 1:1 randomization to 2 lower doses, subjects were preferentially randomized to optimal doses. The primary endpoint was achieved, with the data suggesting 2 of the 3 doses had clinically significant utility. Bayesian decision rules provided intuitive interpretation of clinical outcome.
    GEP IV/oral
    0.75g/1.5g BID 1g/2g BID 1g/2g TID
    N 58 39 25
    Clinical success 48 (83%) 28 (72%) 23 (92%)
    Drug-related AE withdrawals 1 0 0
    Probability that CUI > 1.1 0.99 0.67 0.99
    Conclusions: Primary endpoints based on safety and efficacy should be given consideration in antibacterial trials given the importance of tolerability and efficacy in this setting. Adaptive designs incorporating Bayesian decision rules allow for dose adjustment and allocation of subjects to optimal treatments, based on emerging results.
    Teri Ashton, PhD1, Ohad Amit, PhD2, Etienne Dumont, MD2, Nicole Scangarella-Oman, MS2, Mohammad Hossain, PhD3, Caroline Perry, PhD2, Mark Fitzgerald, PhD4 and Kert Viele, PhD4, (1)GlaxoSmithKline, Research Triangle Park, NC, (2)GlaxoSmithKline, Collegeville, PA, (3)GlaxoSmithKline, King of Prussia, PA, (4)Berry Consultants LLC, Austin, TX

    Disclosures:

    T. Ashton, GlaxoSmithKline: Employee and Shareholder , Salary

    O. Amit, GlaxoSmithKline: Employee and Shareholder , Salary

    E. Dumont, GlaxoSmithKline: Employee and Shareholder , Salary

    N. Scangarella-Oman, GlaxoSmithKline: Employee and Shareholder , Salary

    M. Hossain, GlaxoSmithKline: Employee and Shareholder , Salary

    C. Perry, GlaxoSmithKline: Employee and Shareholder , Salary

    M. Fitzgerald, Berry Consultants: Consultant , Consulting fee

    K. Viele, None

    Findings in the abstracts are embargoed until 12:01 a.m. CDT, Wednesday Oct. 26th with the exception of research findings presented at the IDWeek press conferences.