The literature on Alzheimer’s disease (AD) provides little data about long-term cognitive course trajectories. We identify global cognitive outcome trajectories and associated predictor variables that may inform clinical research and care.

Data derived from the National Alzheimer’s Coordinating Center (NACC) Uniform Data Set were used to examine the cognitive course of persons with possible or probable AD, a Mini-Mental State Examination (MMSE) of ≥10, and complete annual assessments for 5 years.

Thirty-six Alzheimer’s Disease Research Centers.

Four hundred and fourteen persons.

We used a hybrid approach comprising qualitative analysis of MMSE trajectory graphs that were operationalized empirically and binary logistic regression analyses to assess 19 variables’ associations with each trajectory. MMSE scores of ±3 points or greater were considered clinically meaningful.

Five distinct cognitive trajectories were identified: fast decliners (32.6%), slow decliners (30.7%), zigzag stable (15.9%), stable (15.9%), and improvers (4.8%). The decliner groups had three subtypes: curvilinear, zigzag, and late decline. The fast decliners were associated with female gender, lower baseline MMSE scores, a shorter illness duration, or receiving a cognitive enhancer. An early MMSE decline of ≥3 points predicted a worse outcome. A higher rate of traumatic brain injury, the absence of an ApoE ϵ4 allele, and male gender were the strongest predictors of favorable outcomes.

Our hybrid approach revealed five distinct cognitive trajectories and a variegated pattern within the decliners and stable/improvers that was more consistent with real-world clinical experience than prior statistically modeled studies. Future investigations need to determine the consistency of the distribution of these categories across settings.

有关阿尔茨海默病 (AD) 的文献提供的有关长期认知过程轨迹的数据很少。我们确定了可能为临床研究和护理提供信息的全局认知结果轨迹和相关预测变量。

来自国家阿尔茨海默病协调中心 (NACC) 统一数据集的数据用于检查可能患有 AD 的人的认知过程、简易精神状态检查 (MMSE) ≥10，并完成 5 年的年度评估。

三十六个阿尔茨海默病研究中心。

四百一十四人。

我们使用了一种混合方法，包括根据经验进行操作的 MMSE 轨迹图的定性分析和二元逻辑回归分析，以评估 19 个变量与每个轨迹的关联。 ±3分或更高的MMSE分数被认为具有临床意义。

确定了五种不同的认知轨迹：快速下降者（32.6%）、缓慢下降者（30.7%）、之字形稳定者（15.9%）、稳定者（15.9%）和改善者（4.8%）。下降组有三种亚型：曲线下降、锯齿状下降和晚期下降。快速下降与女性性别、较低的基线 MMSE 评分、较短的病程或接受认知增强剂有关。早期 MMSE 下降 ≥3 点预示结果会更差。较高的创伤性脑损伤发生率、ApoE ϵ4 等位基因的缺失以及男性是有利结果的最强预测因素。

我们的混合方法揭示了下降者和稳定/改善者的五种不同的认知轨迹和多样化模式，这比之前的统计模型研究更符合现实世界的临床经验。未来的调查需要确定这些类别在不同环境中分布的一致性。