Archer et al. reported a cross-ancestry genome-wide association study (GWAS) on memory performance (n = 27,633) and cognitive decline (n = 22,365), using data from four aging cohorts.¹ Associations were adjusted for age, sex, and the first five principal components. The study of this nature is rare and important for understanding cognitive performance and its relation to neurodegenerative diseases. However, by design, Archer et al.’s¹ study had some limitations that are important to address during the analysis phase. Alzheimer's disease (AD) has a protracted pre-clinical period,² which makes studying memory decline challenging, especially when the study sample is predominantly composed of individuals at older ages (considering age-related cognitive decline), and those at risk of AD are overrepresented in the study sample compared to the general population prevalence. In the study, the percentage of apolipoprotein E (APOE) ε4 carriers in the four participating cohorts ranged from 26.16% to 46.41%, with participants having a baseline mean age > 72. Study samples also included individuals with mild cognitive impairment and AD.¹ Consequently, the emergence of the APOE locus as a significant determinant for baseline memory performance and memory decline was not surprising and might be predominantly driven by the inclusion of a higher proportion of APOE ε4 carriers who are at high risk of dementing illness.³ This might also explain why the authors observed a stronger genetic correlation with AD than with cognitive performance and educational attainment, suggesting that the phenotypes primarily capture aspects of AD rather than memory performance or dementia-related memory decline that precedes the diagnosis of AD. Additionally, sensitivity analysis was not conducted after adjusting for APOE allele carriers or removing ε4 carriers. Considering these limitations and absence of proper replication, caution should be exercised when interpreting the purported roles of prioritized genes (e.g., SLC25A44, BSX, and DPP8) in age- and dementia-related memory decline.

阿切尔等人。报告了一项关于记忆表现（ n  = 27,633）和认知能力下降（n = 22,365）的跨祖先全基因组关联研究（GWAS） ，使用来自四个衰老队列的数据。¹关联根据年龄、性别和前五个主要成分进行了调整。这种性质的研究很少见，但对于理解认知表现及其与神经退行性疾病的关系很重要。然而，根据设计，Archer 等人的¹研究存在一些局限性，在分析阶段需要解决这些局限性。阿尔茨海默病 (AD) 具有较长的临床前阶段² ，这使得研究记忆力下降具有挑战性，特别是当研究样本主要由年龄较大的个体组成时（考虑到与年龄相关的认知能力下降），并且那些有 AD 风险的人是与一般人群的患病率相比，在研究样本中所占比例过高。在该研究中，四个参与队列中载脂蛋白E（ APOE ）ε4携带者的百分比范围为26.16%至46.41%，参与者的基线平均年龄> 72岁。研究样本还包括患有轻度认知障碍和AD的个体。¹因此， APOE基因座作为基线记忆表现和记忆衰退的重要决定因素的出现并不令人意外，这可能主要是由于包含较高比例的患有痴呆症高风险的APOE ε4 携带者所致。 ³这也可以解释为什么作者观察到与 AD 的遗传相关性比与认知表现和教育程度的遗传相关性更强，这表明表型主要捕获 AD 的各个方面，而不是 AD 诊断之前的记忆表现或痴呆相关记忆衰退。此外，在调整APOE等位基因携带者或去除 ε4 携带者后，未进行敏感性分析。考虑到这些限制和缺乏适当的复制，在解释优先基因（例如SLC25A44、BSX和DPP8）在年龄和痴呆相关记忆衰退中的所谓作用时应谨慎行事。