Latent variable parameter estimates.

Several cardiovascular disease (CVD) risk factors (e.g., hypertension, poor glycemic control) can affect and be affected by autonomic nervous system (ANS) activity. Since excess adiposity can influence CVD development through its effect on hypertension and diabetes mellitus, it is important to determine how adiposity and altered ANS activity are related. The present study employed structural equation modeling to investigate the relation between adiposity and ANS activity both directly and indirectly through biological variables typically associated with glycemic impairment and cardiac stress in older adults. Utilizing the Atherosclerosis Risk in Communities (ARIC) dataset, 1,145 non-smoking adults (74±4.8 yrs, 62.8% female) free from known CVD, hypertension, and diabetes and not currently taking beta-blockers were evaluated for fasting blood glucose (FBG), insulin, and HbA1c concentrations, waist circumference (WC), blood pressure (BP), and markers of ANS activity. WC was recorded just above the iliac crest and was used to reflect central adiposity. Resting 2-minute electrocardiograph recordings, pulse wave velocity, and ankle-brachial index data were used to assess the root mean square of successive differences in RR intervals (RMSSD) and the pre-ejection period (PEP), markers of parasympathetic and sympathetic activity, respectively. FBG, insulin, and HbA1c inferred a latent variable termed glycemic impairment (GI), whereas heart rate and diastolic BP inferred a latent variable termed cardiac stress (CS). The structural equation model fit was acceptable [root mean square error of approximation = 0.050 (90% CI = .036, .066), comparative fit index = .970, Tucker Lewis Index = 0.929], with adiposity having both significant direct (β = 0.208, p = 0.018) and indirect (β = -.217, p = .041) effects on PEP through GI. Adiposity displayed no significant direct effect on RMSSD. CS displayed a significant pathway (β = -0.524, p = 0.035) on RMSSD, but the indirect effect of WC on RMSSD through CS did not reach statistical significance (β = -0.094, p = 0.137). These results suggest that adiposity’s relation to ANS activity is multifaceted, as increased central adiposity had opposing direct and indirect effects on markers of sympathetic activity in this population of older adults.

多种心血管疾病（cardiovascular disease, CVD）危险因素（如高血压、血糖控制不佳）可对自主神经系统（autonomic nervous system, ANS）活动产生影响，同时也会受其调控。鉴于过量脂肪堆积可通过影响高血压与糖尿病的发生发展来作用于心血管疾病，明确体脂与自主神经系统活动异常之间的关联具有重要意义。本研究采用结构方程模型（structural equation modeling），旨在探究老年人群中体脂与自主神经系统活动之间的直接关联，以及通过通常与血糖损伤、心脏应激相关的生物学变量所介导的间接关联。本研究依托社区动脉粥样硬化风险（Atherosclerosis Risk in Communities, ARIC）数据集，对1145名无吸烟史的成年人（年龄74±4.8岁，女性占比62.8%）进行评估，这些受试者均无明确心血管疾病、高血压及糖尿病病史，且未服用β受体阻滞剂。检测指标包括空腹血糖（fasting blood glucose, FBG）、胰岛素、糖化血红蛋白（HbA1c）浓度、腰围（waist circumference, WC）、血压（blood pressure, BP）以及自主神经系统活动标志物。腰围测量位置为髂嵴上方，用于反映中心性体脂含量。静息状态下2分钟心电图记录、脉搏波传导速度及踝肱指数数据被用于评估RR间期连续差值的均方根（root mean square of successive differences in RR intervals, RMSSD）与射血前期（pre-ejection period, PEP），二者分别为副交感神经与交感神经活动的标志物。空腹血糖、胰岛素及糖化血红蛋白共同构成潜在变量血糖损伤（glycemic impairment, GI），而心率与舒张压则共同构成潜在变量心脏应激（cardiac stress, CS）。结构方程模型拟合效果良好[近似误差均方根（root mean square error of approximation）=0.050，90%置信区间（confidence interval, CI）为0.036~0.066，比较拟合指数（comparative fit index, CFI）=0.970，塔克-刘易斯指数（Tucker Lewis Index, TLI）=0.929]，体脂对射血前期存在显著的直接效应（β=0.208，p=0.018）与通过血糖损伤介导的间接效应（β=-0.217，p=0.041）。体脂对RR间期连续差值均方根无显著直接效应。心脏应激对RR间期连续差值均方根存在显著调控通路（β=-0.524，p=0.035），但腰围通过心脏应激介导的对RR间期连续差值均方根的间接效应未达到统计学显著性（β=-0.094，p=0.137）。本研究结果表明，体脂与自主神经系统活动的关联具有多面性：在该老年人群中，中心性体脂升高对交感神经活动标志物存在方向相反的直接与间接效应。