Predictors and Biomarkers for VCID versus Cognitive Recovery after Stroke Thrombectomy in Rural Appalachia

血管性认知障碍和痴呆（VCID）存在于25-30%的中风患者中，并且难以治疗。 在中风发生时预测这些患者的情况。解决阿尔茨海默病的国家计划， NHLBI-NINDS VCID研讨会呼吁加强对VCID预测的研究，特别是在服务不足的地区 人口数量。这项拟议中的研究将确定在农村，服务不足的患者认知恢复的预测因素 在紧急大血管闭塞（ELVO）卒中的机械血栓切除术后。这些预测者将针对 这些患者将从强化康复中受益。我们的新型生物样本登记用于识别 血栓切除术期间采集的全身血液样本中的蛋白质作为认知功能的独立预测因子 卒中后出院和90天随访时的表现。假设这些蛋白质将提供 对认知障碍（包括痴呆）的增量预测有效性，超过临床 特征、健康的社会决定因素和传统的ADRD血浆生物标志物。拟定的研究将 前瞻性入组225例患者，在血栓切除术时采集全身动脉血，并分析 蛋白质表达使用针对服务不足的农村参与者的保留策略，神经认知测试数据 和静脉血液样品的纵向蛋白质表达将收集后3个月至2年， 血栓切除术研究地点位于美国中风地带，那里有一些医疗服务最差的农村地区。 国家的县，将有助于国家卫生研究院优先关注健康差距。具体目标1：测试 假设增加的患者家乡乡村预示着卒中后更差的认知结果， 无论是急性的还是慢性的广义估计方程模型将检验家乡县的主效应 农村和其他社会健康决定因素对认知状态的影响（正常，认知障碍无痴呆， 痴呆），时间对各时间点认知表现的主要影响，以及农村和 各时间点的认知状态时间。乡村生活、合并症和血栓切除术时间之间的关系 血蛋白表达将作为辅助分析。具体目标2：检验ADRD 血栓切除术时全身血液中的生物标志物是急性卒中后认知障碍的生物标志物。 两年后出现慢性VCID。将评估血样中已知的蛋白质组学 在UKY生物标志物中心的认知生物标志物。具体目标3：检验新的假设 血栓切除术时全身血液中的生物标志物可预测慢性VCID，包括痴呆， 并高于ADRD标志物的作用。将对血栓切除术时间血液进行蛋白质组学分析 样品以测量炎症相关蛋白的表达。递阶广义估计 方程模型将评估随着时间的推移，认知状态预测的增量改善， 这些附加标记。分析将使用LASSO方法和拆分样本交叉验证来创建 基于ADRD和预测认知结果的新型生物标志物的预测模型。

Vascular cognitive impairment and dementia (VCID) is present in 25-30% of stroke patients, and it is difficult to predict at the time of stroke who those patients will be. The National Plan to Address Alzheimer’s Disease and the NHLBI-NINDS VCID Workshop called for increased research on VCID prediction, especially in underserved populations. The proposed research will identify predictors of cognitive recovery in rural, underserved patients after mechanical thrombectomy for emergent large vessel occlusion (ELVO) stroke. These predictors will target patients that would benefit from intensive rehabilitation. Our novel biologic sample registry was used to identify proteins in systemic blood samples taken during thrombectomy as independent predictors of cognitive performance at post-stroke discharge and 90-day follow-up. The hypothesis is that these proteins will provide incremental predictive validity for cognitive impairment including dementia, over and above clinical characteristics, social determinants of health, and traditional ADRD plasma biomarkers. The proposed study will prospectively enroll 225 patients collecting systemic arterial blood at the time of thrombectomy and analyze protein expression. Using a retention strategy tailored to underserved rural participants, neurocognitive test data and venous blood samples for longitudinal protein expression will be collected from 3 months to 2 years post- thrombectomy. The study site, located in the US Stroke Belt with some of the most medically underserved rural counties in the nation, will contribute to the NIH priority focus on health disparities. Specific Aim 1: Test the hypothesis that increased patient home county rurality predicts worse cognitive outcomes post-stroke, both acutely and chronically. Generalized estimating equation models will test the main effect of home county rurality and other social determinants of health on cognitive status (normal, cognitive impairment no dementia, dementia), the main effect of time on cognitive performance across time points, and the interaction of rurality and time on cognitive status across time points. Associations among rurality, comorbidities, and time-of-thrombectomy blood protein expression will be auxiliary analyses. Specific Aim 2: Test the hypothesis that ADRD biomarkers in systemic blood at time of thrombectomy are biomarkers for acute post-stroke cognitive impairment and chronic VCID two years later. Blood samples will be assessed for known proteomic biomarkers for cognition at the UKY Biomarker Center. Specific Aim 3: Test the hypothesis that novel biomarkers in systemic blood at time of thrombectomy predict chronic VCID including dementia, over and above the effects of ADRD markers. Proteomic analysis will be conducted on time-of-thrombectomy blood samples to measure the expression of inflammatory-related proteins. Hierarchical generalized estimating equation models will assess the incremental improvement in prediction of cognitive status over time added by these addition markers. Analyses will use LASSO methods and split-sample cross-validation to create a predictive model based on both ADRD and novel biomarkers that predict cognitive outcomes.