It remains unclear whether cognitive reserve can attenuate dementia risk among people with different genetic predispositions.

We aimed to examine the association between cognitive reserve and dementia, and further to explore whether and to what extent cognitive reserve may modify the risk effect of genetic factors on dementia.

Within the UK Biobank, 210 631 dementia-free participants aged ≥60 years were followed to detect incident dementia. Dementia was ascertained through medical and death records. A composite cognitive reserve indicator encompassing education, occupation and multiple cognitively loaded activities was created using latent class analysis, categorised as low, moderate and high level. Polygenic risk scores for Alzheimer's disease were constructed to evaluate genetic risk for dementia, categorised by tertiles (high, moderate and low). Data were analysed using Cox models and Laplace regression.

In multi-adjusted Cox models, the hazard ratio (HR) of dementia was 0.66 (95% confidence interval (CI) 0.61–0.70) for high cognitive reserve compared with low cognitive reserve. In Laplace regression, participants with high cognitive reserve developed dementia 1.62 (95% CI 1.35–1.88) years later than those with low cognitive reserve. In stratified analysis by genetic risk, high cognitive reserve was related to more than 30% lower dementia risk compared with low cognitive reserve in each stratum. There was an additive interaction between low cognitive reserve and high genetic risk on dementia (attributable proportion 0.24, 95% CI 0.17–0.31).

High cognitive reserve is associated with reduced risk of dementia and may delay dementia onset. Genetic risk for dementia may be mitigated by high cognitive reserve. Our findings underscore the importance of enhancing cognitive reserve in dementia prevention.

目前尚不清楚认知储备是否可以降低具有不同遗传倾向的人患痴呆症的风险。

我们的目的是研究认知储备与痴呆之间的关联，并进一步探讨认知储备是否以及在多大程度上可以改变遗传因素对痴呆的风险影响。

在英国生物银行内，对 210 631 名年龄 ≥ 60 岁的无痴呆参与者进行了跟踪，以检测痴呆事件。痴呆症是通过医疗和死亡记录确定的。使用潜在类别分析创建了一个涵盖教育、职业和多种认知负荷活动的综合认知储备指标，分为低、中和高水平。构建阿尔茨海默氏病的多基因风险评分是为了评估痴呆症的遗传风险，按三分位数（高、中和低）分类。使用 Cox 模型和拉普拉斯回归分析数据。

在多重调整的 Cox 模型中，与低认知储备相比，高认知储备的痴呆风险比 (HR) 为 0.66（95% 置信区间 (CI) 0.61-0.70）。在拉普拉斯回归中，认知储备高的参与者比认知储备低的参与者晚 1.62 年（95% CI 1.35-1.88）患痴呆症。在按遗传风险进行的分层分析中，与每个阶层的低认知储备相比，高认知储备与痴呆风险降低 30% 以上相关。认知储备低和痴呆遗传风险高之间存在相加交互作用（归因比例 0.24，95% CI 0.17-0.31）。

高认知储备与痴呆风险降低相关，并可能延迟痴呆发作。高认知储备可以降低痴呆症的遗传风险。我们的研究结果强调了增强认知储备在预防痴呆症中的重要性。