Department of Biostatistics, School of Public Health, Fudan University, Shanghai, China.
NHC Key Laboratory of Health Technology Assessment, Fudan University, Shanghai, China.
J Alzheimers Dis. 2022;86(2):919-929. doi: 10.3233/JAD-215517.
The relationship between body mass index (BMI) and dementia is inconclusive. Undesirable loss of fat-free mass is a risk factor for cognitive decline while obesity is also a risk factor for cardio-metabolic disorders among the older adults.
This study aimed to examine the association between BMI and incident all-cause dementia among Chinese older adults using a prospective study.
Participants were 1,627 community residents aged 60 or older without dementia from the Shanghai Aging Study. Cox regression models, incorporated with restricted cubic splines, were used to explore a nonlinear association between baseline BMI and risk of all-cause dementia as measured by hazard ratio (HR) using both frequentist and Bayesian approach.
We diagnosed 136 incident dementia cases during the mean follow-up of 5.3 years. Compared with moderate BMI (18.5-24.0 kg/m2), low BMI (< 18.5 kg/m2) were related to an increased risk of dementia with the HR as 3.38 (95% CI 1.50-7.63), while high BMI (≥24.0 kg/m2) showed a decreased risk of dementia without statistical significance (HR = 0.91, 95% CI 0.60 to 1.39). Sensitivity analysis in participants without central obesity indicated that the association was still significant with even higher HR. Bayesian approach presented the similar results.
Our result indicates that low BMI may contribute to high risk of incident dementia, even in individuals without central obesity.
体重指数（BMI）与痴呆之间的关系尚无定论。非脂肪质量的不良损失是认知能力下降的一个危险因素，而肥胖也是老年人代谢紊乱的一个危险因素。
本研究旨在通过前瞻性研究，检验中国老年人 BMI 与全因痴呆发病的关系。
参与者为来自上海老龄化研究的 1627 名年龄在 60 岁及以上且无痴呆的社区居民。Cox 回归模型，结合限制立方样条，用于探索基线 BMI 与全因痴呆风险之间的非线性关系，风险比（HR）采用频率论和贝叶斯方法进行评估。
我们在平均 5.3 年的随访期间诊断出 136 例新发痴呆病例。与中等 BMI（18.5-24.0 kg/m2）相比，低 BMI（<18.5 kg/m2）与痴呆风险增加相关，HR 为 3.38（95%CI 1.50-7.63），而高 BMI（≥24.0 kg/m2）则与痴呆风险降低无统计学意义（HR=0.91，95%CI 0.60-1.39）。在无中心性肥胖的参与者中进行敏感性分析表明，即使 HR 更高，关联仍然显著。贝叶斯方法也得出了相似的结果。
我们的结果表明，低 BMI 可能导致痴呆发病风险增加，即使在无中心性肥胖的个体中也是如此。
Zotero 插件