Predicting long-term cognitive outcomes and Alzheimer’s disease and related dementias after major noncardiac surgery for older adults

项目摘要/摘要 高达10%的老年人在术后1年或更长时间内经历了持续性的术后认知能力下降 非心脏手术，这与阿尔茨海默病和阿尔茨海默氏症的风险增加高达50% 疾病相关痴呆(AD/ADRD)。然而，大多数老年人在认知上不会受到伤害，也会 享受更长的寿命和/或因为手术而改善的功能。关于预期的信息很少 认知结果可用于指导考虑择期手术的老年人，因为先前的研究已经 没有纳入手术前的纵向轨迹，侧重于临床相关的认知结果，或 旨在促进个性化预测。尽管它被称为“道德上的必然性” 医生将术后不良认知结果作为手术知情同意的一部分进行讨论，但他们缺乏 与患者分享准确、个性化的信息。在这项研究中，我们建议使用大量的人口- 具有代表性的纵向数据源，用于开发和验证长期临床预测模型 老年人选择性大手术后的认知结果，包括AD/ADRD。我们将(目标1)使用 健康和退休研究(HRS)，链接到联邦医疗保险账单数据，以开发临床预测模型 根据术前已知因素预测(A)记忆力、(B)AD/ADRD风险和(C) 独立管理自己的财务和药物，这是日常生活中两项认知密集型的工具性活动 生活(IADL)，术后两年。在方法上，我们将平衡模型的复杂性和临床 实用性和使用方法来解释选择性生存。我们将(目标2)增强临床预测 手术后出现并发症或不良事件的模型，例如精神错乱、再次住院和 新的中重度疼痛。这一目标的发现将增进对“最好的情况，最坏的情况”的理解 包括在术前认知风险中的认知结果(记忆、AD/ADRD和认知功能) 讨论。这些发现也将支持未来关于导致不良反应的关键机制的假说。 结果，指导手术后处理的优先事项，以改善长期认知结果。最后， 我们将(目标3)再次使用国家健康和老龄化趋势研究对HRS模型进行外部验证 链接到Medicare数据，以评估预测记忆力、AD/ADRD和IADL的模型的泛化能力 手术后的独立性。经过验证，这些模型可以用于临床预测两年认知能力。 结果，包括AD/ADRD风险，以便老年人拥有关于可能发生的 接受或不接受手术的认知结果。这些信息将使长期认知能力得以纳入 手术共同决策的结果，对老年手术患者来说是一个变革性的进步。

PROJECT SUMMARY / ABSTRACT Up to 10% of older adults experience durable postoperative cognitive decline at 1 year or beyond after major noncardiac surgery, which is associated with up to 50% increased risk for Alzheimer’s disease and Alzheimer’s disease related dementias (AD/ADRD). However, most older adults will be cognitively unharmed by, and will enjoy longer life and/or improved function because of, surgery. Very little information about anticipated cognitive outcome is available to guide older adults considering elective surgery, because prior research has not incorporated longitudinal pre-surgical trajectories, focused on clinically relevant cognitive outcomes, or been designed to facilitate individualized predictions. Although it has been called “ethically imperative” that physicians discuss adverse postoperative cognitive outcomes as part of surgical informed consent, they lack accurate, individualized information to share with patients. In this study, we propose to use large, population- representative, longitudinal data sources to develop and validate clinical prediction models for long-term cognitive outcomes, including AD/ADRD, after elective major surgery in older adults. We will (AIM 1) use the Health and Retirement Study (HRS), linked to Medicare billing data, to develop clinical prediction models based on preoperatively-known factors to predict (a) memory, (b) risk of AD/ADRD, and (c) ability to independently manage one’s finances and medication, two cognitively-intensive instrumental activities of daily living (IADLs), at two years after surgery. Methodologically, we will balance model complexity with clinical practicality and use methods to account for selective survival. We will (AIM 2) augment the clinical prediction models with complications or adverse events that occur after surgery, e.g., delirium, hospital readmission, and new moderate-severe pain. Findings from this Aim will enhance understanding of “best case, worst case” cognitive outcomes (memory, AD/ADRD, and cognitive-functional) to include in preoperative cognitive risk discussion. These findings will also support future hypotheses about key mechanisms that lead to adverse outcomes, guiding priorities for post-surgical management to improve long-term cognitive outcomes. Finally, we will (AIM 3) externally validate the HRS models using the National Health and Aging Trends study, again linked to Medicare data, to assess generalizability of the models predicting memory, AD/ADRD, and IADL independence after surgery. After validation, these models can be used clinically to predict 2-year cognitive outcomes, including AD/ADRD risk, so that older adults have accurate, personalized information on the likely cognitive outcomes with or without surgery. This information will enable inclusion of long-term cognitive outcomes in surgical shared decision-making, a transformative advance for older surgical patients.