Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, 1177, Stockholm, Sweden.
Aging Research Network - Jönköping (ARN-J), School of Health and Welfare, Jönköping University, Jönköping, Sweden.
Eur J Epidemiol. 2022 Dec;37(12):1261-1271. doi: 10.1007/s10654-022-00918-w. Epub 2022 Oct 3.
While midlife adiposity is a risk factor for dementia, adiposity in late-life appears to be associated with lower risk. What drives the associations is poorly understood, especially the inverse association in late-life. Using results from genome-wide association studies, we identified inflammation and lipid metabolism as biological pathways involved in both adiposity and dementia. To test if these factors mediate the effect of midlife and/or late-life adiposity on dementia, we then used cohort data from the Swedish Twin Registry, with measures of adiposity and potential mediators taken in midlife (age 40-64, n = 5999) or late-life (age 65-90, n = 7257). Associations between body-mass index (BMI), waist-hip ratio (WHR), C-reactive protein (CRP), lipid levels, and dementia were tested in survival and mediation analyses. Age was used as the underlying time scale, and sex and education included as covariates in all models. Fasting status was included as a covariate in models of lipids. One standard deviation (SD) higher WHR in midlife was associated with 25% (95% CI 2-52%) higher dementia risk, with slight attenuation when adjusting for BMI. No evidence of mediation through CRP or lipid levels was present. After age 65, one SD higher BMI, but not WHR, was associated with 8% (95% CI 1-14%) lower dementia risk. The association was partly mediated by higher CRP, and suppressed when high-density lipoprotein levels were low. In conclusion, the negative effects of midlife adiposity on dementia risk were driven directly by factors associated with body fat distribution, with no evidence of mediation through inflammation or lipid levels. There was an inverse association between late-life adiposity and dementia risk, especially where the body's inflammatory response and lipid homeostasis is intact.
虽然中年肥胖是痴呆的一个风险因素,但老年肥胖似乎与较低的风险相关。导致这些关联的原因尚不清楚,尤其是老年肥胖的反比关联。我们利用全基因组关联研究的结果,确定了炎症和脂质代谢是与肥胖和痴呆都相关的生物学途径。为了测试这些因素是否在中年和/或老年肥胖对痴呆的影响中起中介作用,我们随后使用了来自瑞典双胞胎登记处的队列数据,该数据包含了中年(40-64 岁,n=5999)或老年(65-90 岁,n=7257)时的肥胖测量值和潜在的中介因素。我们在生存和中介分析中测试了体质指数(BMI)、腰臀比(WHR)、C 反应蛋白(CRP)、血脂水平与痴呆之间的关联。在所有模型中,年龄被用作潜在的时间尺度,性别和教育程度被包含为协变量。在血脂模型中,禁食状态被包含为协变量。中年时 WHR 每增加一个标准差(SD),痴呆风险就会增加 25%(95%CI 2-52%),而当调整 BMI 后则略有减弱。没有证据表明 CRP 或血脂水平存在中介作用。65 岁以后,BMI 每增加一个 SD,但 WHR 没有,痴呆风险就会降低 8%(95%CI 1-14%)。这种关联部分是由 CRP 升高介导的,而当高密度脂蛋白水平较低时,这种关联则被抑制。总之,中年肥胖对痴呆风险的负面影响是由与体脂分布相关的因素直接驱动的,没有证据表明炎症或血脂水平在其中起中介作用。老年肥胖与痴呆风险呈反比关系,尤其是在身体的炎症反应和脂质稳态完好的情况下。