Physical Fitness as an Objective Biomarker for AD/ADRD Risk Modification

大约有600万65岁的美国人患有阿尔茨海默病和相关痴呆症(AD/ADRD)，他们 给患者及其家人带来巨大的情感、身体和经济负担，并使美国 每年经济规模超过2500亿美元。AD/ADRD是一种日益严重的国家公共卫生危机，因为 预计到2050年，美国人的≥65岁将翻一番。目前还没有批准用于治疗的疾病修改药物 AD/ADRD，自2003年以来没有批准过新的症状改善药物。这突显了需要 旨在预防AD/ADRD的新研究方法。体力活动(PA)已被确定为 生活方式干预对预防AD/ADRD的潜力最大。然而，大多数随机化的 PA的受控试验(RCT)，包括目前由NIA赞助的23个正在进行的RCT，规模较小 短期的跟踪调查。需要更大的随机对照试验和更长的随访时间才能提供确切的益处证据。 相关生物标记物的识别和未来风险的量化可以缓解这种情况的长期持续和成本 RCT。大型数据库和长期随访的流行病学研究受到不精确和不可靠的限制 PA的自我报告。我们提出了使用心肺适合度(CRF)作为可靠的 PA降低AD/ADRD风险的生物标志物。与身体成分、血压和心率一样，CRF也是 被认为是PA的生物标志物。同样，良好控制血糖可使HbA1C、PA达到最佳水平 持续时间、频率和强度也会导致最佳的慢性肾功能衰竭。当前的应用程序已设计用于测试 假设较高的CRF独立地与较低的AD/ADRD事件风险相关，并确定 通过标准化运动耐量试验(ETT)评估CRF的最佳水平，将AD/ADRD风险降至最低 在人口和个人层面都是如此。我们的具体目标是(1)确定CRF与 和事件AD/ADRD，考虑到CRF的非线性关系和潜在相互作用 以及(2)定义递增的CRF水平，这些水平与逐渐降低的风险有关 AD/ADRD，总体上，并按年龄、性别和种族分组。我们的目标3是开发和验证一个深层 基于学习的风险预测模型确定最优CRF水平使个体达到最低 AD/ADRD的风险。深度(机器)学习是一项关键的人工智能(AI)技术的最新进展， 允许对更精确的风险预测所需的复杂非线性关系进行建模。这些目标将 通过使用911,698名没有基线AD/ADRD(平均年龄61岁，54,411岁)的退伍军人的ETT数据来实现 女性，20.2万名非洲裔美国人，50107名西班牙裔美国人)，以及长达19年的跟踪数据。 将在VA数据中开发可解释的深度学习风险预测模型，并通过 联邦医疗保险人口将开发一种将CRF建议个性化的“精准医疗”方法 以降低AD/ADRD风险。未来的研究将结合分子生物标记物来验证AD 诊断和确定CRF和CRF修饰对AD风险的影响。

About 6M Americans ≥65 years suffer from Alzheimer's disease and related dementias (AD/ADRD) which poses significant emotional, physical, and financial burden on patients and their families, and costs the US economy over $250 billion annually. AD/ADRD is a growing national public health crisis as the number of Americans ≥65 years is projected to double by 2050. There are no approved disease-modifying drugs for AD/ADRD, and no new symptom-modifying drug has been approved since 2003. This highlights the need for novel research approaches aimed at preventing AD/ADRD. Physical activity (PA) has been identified as the lifestyle intervention with the strongest potential to prevent AD/ADRD. However, most of the randomized controlled trials (RCTs) of PA including the 23 ongoing RCTs currently sponsored by the NIA are small with short follow-up. Larger RCTs with longer follow-up are needed to provide definitive evidence of benefit. Identification of relevant biomarkers and quantification of future risk can mitigate long duration and cost of such RCTs. Epidemiologic studies with large database and long follow-up are limited by imprecise and unreliable self-reports of PA. We propose the novel concept of using cardiorespiratory fitness (CRF) as a reliable biomarker of PA for AD/ADRD risk reduction. Like body composition, blood pressure and heart rate, CRF is considered a biomarker of PA. Just as well-controlled blood glucose leads to optimal HbA1C, PA of adequate duration, frequency and intensity also leads to optimal CRF. The current application has been designed to test the hypotheses that higher CRF is independently linked to a lower risk of incident AD/ADRD and to determine optimal levels of CRF assessed by a standardized exercise tolerance test (ETT), for minimizing AD/ADRD risk at both population and individual levels. Our Specific Aims are (1) to determine the relationship between CRF and incident AD/ADRD, taking into consideration a non-linear relationship and potential interactions of CRF with other risk factors and (2) to define incremental CRF levels that are linked to progressively lower risk of AD/ADRD, overall, and in subgroups by age, sex, and race. Our Aim 3 is to develop and validate a deep learning-based risk prediction model to determine the optimal CRF level for individuals to achieve the lowest risk of AD/ADRD. Recent advances in deep (machine) learning, a key artificial intelligence (AI) technology, allow modeling of complex non-linear relationships necessary for more precise risk prediction. These aims will be achieved by using ETT data on 911,698 Veterans free of baseline AD/ADRD (mean age, 61 years, 54,411 women, 202,000 African American, 50,107 Hispanic Americans) with up to 19 years of follow-up data. Explainable deep learning risk prediction model will be developed in the VA data, and validated with a Medicare population to develop a “precision medicine” approach that will personalize CRF recommendations for individuals to lower their AD/ADRD risk. Future studies will incorporate molecular biomarkers to validate AD diagnosis and ascertain effect of CRF and CRF modification on AD risk.