 |
| April 7, 2010 Part 1 In this three-part issue of Invenio, we will provide a top-line summary of a case study on how the use of adaptive designs can increase efficiency and conserve resources during Central Nervous System (CNS) drug development. A full version of this paper has been submitted for review. Adaptive clinical trial designs permit modifications of specific aspects of an ongoing study based on accumulating data in real time to optimize the information that the study will provide for subsequent decisions concerning development plans for the compound being studied. To maintain the validity and integrity of the trial, the adaptation options should be specified prior to the start of the trial. These adaptations can include: (a) stopping early either for futility or success, (b) expanding the sample size due to greater than expected data variability, or (c) allocating patients preferentially to doses or treatment options demonstrating the optimal balance of efficacy versus safety and tolerability, (i.e., therapeutic index). The goals are to optimally benefit patients within the trial and increase the efficiency of the drug development process. How well these goals will be met is dependent on time to information (the earlier, the better) and the recruitment speed relative to the readout of observations required to adapt. In this issue of Invenio, we focus on the application of adaptive design to dose-finding studies. Traditional designs usually involve 2-3 doses only and do not routinely use modeling techniques. Adaptive designs by contrast use more doses (5-15) and apply modeling approaches to estimate the dose-response, using all available data accrued in real time to identify an optimal target dose, e.g. the minimal dose yielding maximal efficacy. In light of the estimated dose-response and the emerging data, two adaptations, (dose finding and early study termination for either success or futility), are being considered on an ongoing basis. This issue in the series presents a virtual case study which is based upon a compilation of several real life studies. The goal of the series is to illustrate several features and advantages of adaptive trials designs relevant to CNS drug development. Virtual Study Summary - Two hundred and fifty patients with acute exacerbation of the target illness were randomized and approximately two-thirds of the patients completed the 30-day double-blind treatment period. The investigational study drug was well tolerated and did not cause undesirable side effects. However, the study was stopped early when it became apparent that none of the treatment arms including the active comparator differentiated from placebo. Conclusions - The adaptive design enabled the early stopping of this study. The high variability on the primary endpoint observed particularly in the placebo group might have been due to a suboptimal selection of the patient population and/or the extensive use of rescue medication. The findings of this failed study are not unique to a CNS indication, but it emphasizes the need to optimize all aspects of clinical trial design including patient population, endpoint selection and possible confounding factors such as concomitant medications. Nevertheless, the use of the adaptive design in this virtual case study saved considerable research resources and permitted a decision about what to do with the compound (i.e., stop its development). In next week’s installment we’ll offer an in-depth discussion of how the adaptive trial design was deployed in this virtual case study and offer some suggestions as to why the compound failed to separate from placebo. |
 |