Department of Biomedical and Nutritional Sciences, Zuckerberg College of Health Sciences, University of Massachusetts Lowell, Lowell, MA, United States.
Friedman School of Nutrition Science and Policy, Tufts University, Boston, MA, United States.
Adv Nutr. 2024 Aug;15(8):100272. doi: 10.1016/j.advnut.2024.100272. Epub 2024 Jul 14.
Magnesium (Mg) plays a key role in neurological functioning and manifestations. However, the evidence from randomized controlled trials (RCTs) and cohorts on Mg and cognitive health among adults has not been systematically reviewed. We aimed to examine the associations of various Mg forms (supplements, dietary intake, and biomarkers) with cognitive outcomes by summarizing evidence from RCTs and cohorts. PubMed, Embase, PsycINFO, and the Cochrane Central Register of Controlled Trials were searched for relevant peer-reviewed articles published up to May 3, 2024. Three random-effects models were performed, when appropriate, to evaluate the relationship between Mg and cognitive outcomes: 1) linear meta-regression, 2) nonlinear (quadratic) meta-regression, and 3) meta-analysis using Mg variables categorized based on pre-existing recommendations. Three RCTs and 12 cohort studies were included in this systematic review. Evidence from the limited number of RCTs was insufficient to draw conclusions on the effects of Mg supplements. Cohort studies showed inconsistent dose-response relationships between dietary Mg and cognitive disorders, with high heterogeneity across populations. However, consistent U-shape associations of serum Mg with all-cause dementia and cognitive impairment were found in cohorts, suggesting an optimal serum Mg concentration of ∼0.85 mmol/L. This nonlinear association was detected in meta-regression (P = 0.003) and in meta-analysis based on the reference interval of serum Mg (0.75-0.95 mmol/L) [<0.75 compared with 0.85 mmol/L: pooled hazard ratio (HR) = 1.43; 95% confidence interval (CI) = 1.05, 1.93; >0.95 compared with 0.85 mmol/L: pooled HR = 1.30; 95% CI = 1.03, 1.64]. More evidence from RCTs and cohorts is warranted. Future cohort studies should evaluate various Mg biomarkers and collect repeated measurements of Mg intake over time, considering different sources (diet or supplements) and factors affecting absorption (for example, calcium-to-Mg intake ratio). This systematic review was preregistered in PROSPERO (CRD42023423663).
镁（Mg）在神经功能和表现中起着关键作用。然而，关于成年人的镁与认知健康的随机对照试验（RCT）和队列研究的证据尚未得到系统审查。我们旨在通过总结 RCT 和队列研究的证据，检查各种镁形式（补充剂、饮食摄入和生物标志物）与认知结果之间的关联。我们在 PubMed、Embase、PsycINFO 和 Cochrane 中央对照试验注册库中搜索了截至 2024 年 5 月 3 日发表的相关同行评议文章。在适当的情况下，我们进行了 3 个随机效应模型，以评估 Mg 与认知结果之间的关系：1）线性元回归，2）非线性（二次）元回归，3）基于预先存在的建议对 Mg 变量进行分类的元分析。这项系统评价纳入了 3 项 RCT 和 12 项队列研究。RCT 的证据不足以得出关于镁补充剂影响的结论。队列研究显示，饮食镁与认知障碍之间的剂量反应关系不一致，不同人群之间存在高度异质性。然而，在队列研究中发现了血清镁与全因痴呆和认知障碍之间一致的 U 形关联，表明最佳血清镁浓度约为 0.85mmol/L。这种非线性关联在元回归（P=0.003）和基于血清镁参考区间（0.75-0.95mmol/L）的元分析中均有检测到（<0.75 与 0.85mmol/L 相比：汇总危险比（HR）=1.43；95%置信区间（CI）=1.05，1.93；>0.95 与 0.85mmol/L 相比：汇总 HR=1.30；95%CI=1.03，1.64）。需要更多来自 RCT 和队列研究的证据。未来的队列研究应评估各种镁生物标志物，并随着时间的推移重复测量镁的摄入量，同时考虑不同的来源（饮食或补充剂）和影响吸收的因素（例如，钙与镁的摄入比）。这项系统评价已在 PROSPERO（CRD42023423663）上预先注册。
