Supplementary Material for: Association of depression and cognitive performance in US older adults: a secondary analysis of cross-sectional data using NHANES 2013-2014

Introduction Depression has been associated with cognitive performance, but whether socio-demographic and clinical characteristics might influence this association is not well elaborated. This study aimed to further explore this relationship in older adults.
Methods This cross-sectional study is based on data from National Health and Nutrition Examination Survey (NHANES) 2013–2014. 1433 individuals with complete information on depressive symptoms and cognitive function variables were included in this study. Patient Health Questionnaire 9 (PHQ-9) score ≥ 10 as the cutoff to identify cases of depression in our study. We defined poor cognitive performance as a composite cognitive score < 47. Logistic regression models were used to examine the association of depression with cognitive performance (Model 1). We progressively adjusted the covariates as confounders (Model 2: Model 1+age, and gender; Model 3: Model 2+race, education level, family income, drinking, and smoking; Model 4: Model 3+overweight, arthritis, hyperlipidemia, diabetes, hypertension, heart failure, coronary heart disease, heart attack, stroke, and cancer). We then conducted subgroup, interaction, and restricted cubic spline (RCS) analyses to examine this association.
Results The prevalence of poor cognitive performance was 36.6% (53/145) in the depression group and 14.1% (182/1288) in the non-depression group. In the fully adjusted model, depression was significantly associated with poor cognitive performance (adjusted odds ratio=2.25; 95% CI: 1.31–3.81). The results were robust to sensitivity analyses. Gender and education level may modify the association between depression and poor cognitive performance. RCS analysis revealed that the PHQ-9 score was related to poor cognitive performance in a nonlinear manner (P for nonlinearity < 0.001), and exhibited a J-shaped curve.
Conclusion Depression is associated with poor cognitive performance in US older adults. Early recognition and treatment of depression may be potential intervention strategies to protect cognitive health.

引言 抑郁与认知功能表现存在关联，但社会人口学与临床特征是否会对该关联产生影响，目前尚未得到充分阐释。本研究旨在针对老年人群体，进一步探究二者间的关联。
方法 本项横断面研究基于2013-2014年美国国家健康与营养检查调查（National Health and Nutrition Examination Survey, NHANES）的数据。本研究共纳入1433名拥有完整抑郁症状与认知功能变量信息的受试者。本研究以患者健康问卷9项（Patient Health Questionnaire 9, PHQ-9）得分≥10作为抑郁判定截断值，以认知综合得分<47定义为认知功能不良。采用逻辑回归模型分析抑郁与认知功能表现的关联（模型1）。随后逐步校正混杂协变量：模型2为模型1基础上校正年龄与性别；模型3为模型2基础上进一步校正种族、受教育程度、家庭收入、饮酒与吸烟情况；模型4为模型3基础上校正超重、关节炎、高脂血症、糖尿病、高血压、心力衰竭、冠心病、心肌梗死、脑卒中与肿瘤病史。此外，本研究还开展了亚组分析、交互作用分析与限制性立方样条（restricted cubic spline, RCS）分析，以进一步验证该关联。
结果 抑郁组中认知功能不良的患病率为36.6%（53/145），非抑郁组为14.1%（182/1288）。在完全校正模型中，抑郁与认知功能不良存在显著关联（校正后比值比=2.25；95%置信区间：1.31~3.81）。敏感性分析结果显示该结论稳健可靠。性别与受教育程度可能对抑郁与认知功能不良的关联存在调节作用。限制性立方样条分析显示，PHQ-9得分与认知功能不良呈非线性相关（非线性检验P<0.001），且呈现J型曲线特征。
结论 美国老年人群中，抑郁与认知功能不良存在关联。早期识别并干预抑郁，或可成为保护认知健康的潜在干预策略。