PREVENTION OF ALZHEIMER DEMENTIA & COGNITIVE DECLINE

将阿尔茨海默病（AD）的发病推迟数年的干预措施将产生巨大的公共卫生效益。一些研究表明，非甾体抗炎药（NSAIDs）可能会造成这种延迟。非甾体抗炎药也可以减轻进行性年龄相关性认知衰退（ARCD），当这种情况代表AD的前驱症状时。这两种预防策略只有在随机试验中才能得到明确评价。这些试验也可以检查长期使用中等剂量的非甾体抗炎药的相关风险，这些剂量似乎对AD和ARCD有保护作用。选择性环氧化酶-2 （COX-2）抑制剂可能提高安全性，但不清楚COX-2抑制剂是否与传统非甾体抗炎药具有明显的保护作用。因此，我们建议对常见的非甾体抗炎药布洛芬和选择性COX-2抑制剂塞来昔布与安慰剂进行平行试验，以预防AD和减轻ARCD。该试验将涉及四个地点，招募2625名年龄在72-88岁之间、一级亲属有阿尔茨海默样痴呆史的无痴呆受试者。因此，定期认知筛查测试结果的显著下降将识别疑似偶发性痴呆的受试者。我们将使用结构化、标准化的评估和诊断方法对这些受试者进行临床评估。所提出的样本假定进行了7年的观察，对死亡率和其他原因造成的损耗以及治疗“中途退出”和“中途退出”进行了现实的估计。它应该提供80%的功率（双尾a=0.05）来检测治疗组中发生率降低30%。在这项申请中，我们建议在最后一名入组受试者完成30个月后进行42-54个月的治疗，并对疗效进行中期分析。在这一点上，研究将有80%的能力检测到任何一种药物的AD发病率降低50%，在这种情况下，试验可以停止。如果没有明显的治疗益处，或者存在安全问题，也应该停止试验。否则，治疗效果的间歇估计将决定竞争更新申请的形式，额外的0.25至4年观察和最终分析。作为次要结果，我们将检查认知评分的轨迹来评估ARCD。

An intervention that delayed onset of Alzheimer's disease (AD) by several years would yield huge public health benefits. Several studies suggest that non-steroidal anti-inflammatory drugs (NSAIDs) may produce such a delay. NSAIDs may also attenuate progressive age-related cognitive decline (ARCD) when this conditions represents a prodrome of AD. Both prevention strategies can be evaluated definitively only in randomized trials. Such trials can also examine attendant risks of long- term NSAID use in the moderate doses that appear to afford protection against AD and ARCD. Improved safety may be available with selective cyclooxygenase-2 (COX-2) inhibitors, but it is not clear that COX-2 inhibition offers the protective effect apparent with conventional NSAIDs. We therefore propose a parallel trial of the common NSAID ibuprofen and the selective COX-2 inhibitor celecoxib vs. placebo for prevention of AD and for attenuation of ARCD. The trial will involve four sites and enroll 2625 dementia-free subjects aged 72-88 with a history of Alzheimer-like dementia for a first degree relative. Therefore, a conspicuous decline in periodic cognitive screening test results will identify subjects with suspected incident dementia. We will evaluate these subjects clinically using structured, standardized methods of assessment and diagnosis. The proposed sample presumes 7 years of observation, with realistic estimates of attrition through mortality and other causes, and of treatment "drop-outs" and "drop-ins." It should provide 80% power (2-tailed a=0.05) to detect a 30% reduction in incidence among the treated groups. In this application we proposed the first 42-54 months of treatment with an interim analysis of efficacy after the last-enrolled subject has completed 30 months. At that point, the study will have 80% power to detect a 50% reduction in AD incidence with either agent, in which case the trial can be stopped. The trial should also be stropped if there is no apparent benefit of treatment, or if safety issues mandate. Otherwise, the intermit estimate of treatment effects will dictate the shape of a competing renewal application an additional 0.25 to 4 years of observation and a final analysis. As a secondary outcome, we will examine the trajectory of cognitive scores to assess ARCD.