Diet is a modifiable lifestyle factor with a proven role in cardiovascular disease risk reduction that might also play an important part in cognitive health. Evidence from observational studies has linked certain healthy dietary patterns to cognitive benefits. However, clinical trials of diet interventions have demonstrated either null or, at best, small effects on cognitive outcomes. In this Review, we summarize the currently available evidence from observational epidemiology and clinical trials regarding the potential role of diet in the prevention of cognitive decline and dementia. We further discuss possible methodological limitations that might have hindered the ability of previous diet intervention trials to capture potential neuroprotective effects. Considering the overwhelming and continuously expanding societal, economic and health-care burden of Alzheimer disease and other dementias, future nutritional research must address past methodological challenges to accurately and reliably inform clinical practice guidelines and public health policies. Within this scope, we provide a roadmap for future diet intervention trials for dementia prevention. We discuss study designs involving both intensive personalized interventions — to evaluate pharmacokinetic and pharmacodynamic properties, establish neuroprotective thresholds, and test hypothesized biological mechanisms and effects on brain health and cognition through sensitive and precise biomarker measures — and large-scale, pragmatic public health interventions to study population-level benefits.

饮食是一种可改变的生活方式因素，已被证明在降低心血管疾病风险方面的作用，也可能在认知健康中发挥重要作用。来自观察性研究的证据将某些健康的饮食模式与认知益处联系起来。然而，饮食干预的临床试验表明，对认知结果的影响要么为零，要么充其量只是很小。在本综述中，我们总结了目前来自观察性流行病学和临床试验的可用证据，这些证据关于饮食在预防认知能力下降和痴呆方面的潜在作用。我们进一步讨论了可能的方法学局限性，这些局限性可能阻碍了先前饮食干预试验捕捉潜在神经保护作用的能力。考虑到阿尔茨海默病和其他痴呆症不堪重负且不断扩大的社会、经济和医疗保健负担，未来的营养研究必须解决过去的方法学挑战，以准确可靠地为临床实践指南和公共卫生政策提供信息。在此范围内，我们为未来预防痴呆症的饮食干预试验提供了路线图。我们讨论了涉及密集个性化干预的研究设计——评估药代动力学和药效学特性，建立神经保护阈值，并通过敏感和精确的生物标志物测量来测试假设的生物学机制和对大脑健康和认知的影响——以及大规模、务实的公共卫生干预来研究人群水平的益处。