Synthesis of Longer-Term Alzheimer Disease Studies in Order to Model and Improve

描述（由申请人提供）：在过去的几十年中，阿尔茨海默病（AD）和轻度认知障碍（MCI）的2期和3期临床试验通常不成功。疗效主要在6个月或更短的试验中得到证实，只有胆碱酯酶抑制剂治疗轻度AD，而在MCI或更长期的试验中没有。最近基于阿尔茨海默病神经病理学的治疗方法在18个月的试验中没有显示出疗效。虽然这可能反映出缺乏对生物靶点的了解，从而对无效药物进行了测试，但它也反映出试验方法的不足，因此一种适度有效的药物——如果进行了测试——可能不被认为是有效的。特别是，这些试验似乎是根据对预期特征和结果的假设来选择患者样本的，而这些假设可能是无效的，因为它们是基于从以前的个别试验中选择的经验，这进一步提出了有关设计和有效性的问题。本项目广泛而长期的目标是在系统评估以往试验和前瞻性测试设计特点的基础上，为设计试点和二期试验提供合理的基础。一个更广泛的目标是通过为临床试验方法和设计提供模型和模拟来鼓励、补充和加速测试新的和创新的治疗方法的过程。实现目标的研究设计和方法涉及将过去十年来来自NIA ADCS和ADNI的14项临床试验和研究的患者数据整合到一个公共数据库中，其中包括5200多个1至4年的随访。这允许使用基于假设的统计技术进行前瞻性分析，包括模拟，以更好地理解设计参数，结果，并设计更好，更具临床相关性的试验，从而能够检测变化。ADCS研究具有固有的相似性，可以进行组合，包括共同的数据库结构、结果、地点、培训和程序。专家指导委员会制定指导原则和基于假设的研究计划。具体方案解决了诸如选择标准、衰退预测因素和评估新试验设计的模拟等问题。例如，我们将使用回归估计和回归均值调整估计来检查基于伴随用药或严重程度选择试验患者对结果的影响，并随后进行试验模拟以经验检验有效性并计划未来的试验。这项研究的结果将对该治疗领域的试验设计产生广泛而持久的影响，特别是在早期试验阶段，因为设计可以在实施试验之前进行测试和完善。该项目具有高度创新性，因为它为该研究领域提供了新的方法来检查来自大量研究的潜在患者数据，并利用丰富的数据库来实证地探索当前试验所依据的假设。这将导致试验方法的改进，并增加在发展早期发现有效治疗方法的可能性。

DESCRIPTION (provided by applicant): Phase 2 and 3 clinical trials in Alzheimer disease (AD) and mild cognitive impairment (MCI) have been generally unsuccessful over the past decades. Efficacy mainly has been demonstrated in trials of 6 months or less, only with cholinesterase inhibitors for mild AD, and not in MCI or longer-term trials. Recent therapies based on the neuropathology of AD have not demonstrated efficacy in 18 month trials. Although this may reflect lack of knowledge of the biological targets and consequently the testing of ineffective drugs, it also reflects inadequacies of trials methods such that a modestly effective drug - if tested - may likely not be recognized as effective. In particular, the trials appear to choose patient samples based on assumptions about desired characteristics and outcomes that may not be valid because they are based on selected experiences from previous individual trials, further raising questions about design and validity. This project's broad, long-term objectives are to develop rational bases for designing pilot and phase 2 trials based on systematic evaluation of previous trials and testing design features prospectively. A broader goal is to encourage, complement, and accelerate the process of testing new and innovative treatments by providing models and simulations for clinical trials methods and designs. The research design and methods for achieving the objectives involves integrating patient data from 14 clinical trials and studies from the NIA's ADCS and ADNI over the last decade - including over 5200 followed from 1 to 4 years - into a common database. This allows for prospective analyses using hypothesis-based statistical techniques including simulations to better understand design parameters, outcomes, and to design better and more clinically relevant trials that would have the ability to detect change. The ADCS studies have inherent similarities that allow for combining, including common data base structures, outcomes, sites, training, and procedures. An expert steering committee formulates guiding principles and hypothesis-based research plans. Specific protocols address issues such as selection criteria, predictors of decline, and simulations to assess new trials designs. For example, we will examine the effects on outcomes of selecting patients for trials based on concomitant medication or severity using regression estimates and regression to the mean adjusted estimates, and follow this with trials simulations to empirically test validity and to plan future trials. Results of this research will have broad and lasting impact on trials designs in this therapeutic area, and especially on early phase trials in that designs can be tested and refined prior to implementing trials. The project is highly innovative in that it provides new approaches in this research area to examine the underlying patient data from numerous studies, and takes advantage of the rich database to empirically probe the assumptions under which current trials have been instigated. This will lead to improvement in trials methods and increase the likelihood for identifying effective treatments earlier in development. PUBLIC HEALTH RELEVANCE: The results of this research will have a broad and lasting impact on trial designs for Alzheimer disease, mild cognitive impairment, and neurodegenerative disorders, and trials in which cognitive improvement is a targeted outcome. It will address important issues in the assumptions underlying the design decisions. This will lead to better understanding and improvement in the methods for future clinical trials and improve the likelihood for identifying effective treatments, as well as identifying ineffective treatments earlier in development.