基于PLS模型的血尿酸与代谢综合征各组分的相关性研究

Correlation analysis of plasma uric acid and otherfactors of metabolic syndrome based on PLS Model

目的探讨血尿酸(uric acid,UA)与代谢综合征(metabolic syndrome,MS)各危险因素的相关关系。方法对30例患者按MS危险因素的多少划分为MS组和MS高危患者组,分别测量UA、胆固醇(TC)、三酰甘油(TG)、高密度脂蛋白胆固醇(HDL-C)、低密度脂蛋白胆固醇(LDL-C)、空腹血糖(FPG)、体质指数(BMI)、血压(SBP/DBP)等指标,分析UA和MS危各危险因素的相关性。结果普通线性回归显示变量之间存在多重共线性,故引入偏最小二乘回归方法(partial least-squaresreg ression,PLS)。通过PLS回归方程可以得知,UA与BMI、FPG、TC、TG、LDL-C、BP呈正相关,与HDL-C呈负相关,尤其与血脂中的TG相关性最强。拟合结果表明,UA水平与变量之间存在明显的线性相关关系,说明变量之间存在相关关系。对危险因素重要性的VIP分析,UA与血脂的关系最显著,其次是BMI、BP和FPG。预测建模分析显示,模型的拟合效果是令人满意的,且随着UA水平的升高,MS组分中的危险因素也逐渐增加,并有MS高危患者向MS发展的趋势。结论 UA水平与MS的发生密切相关,高尿酸血症也是MS的一个重要危险因素和预警因子。

Objective To evaluate the relationship between plasma uric acid and other risk factors of metabolic syndrome.Methods Thirty subjects were divided into MS group and MS high-risk group according to risk factors of MS.UA,BMI,FPG,TC,TG,LDL-C,HDL-C,SBP,DBP and other factors of metabolic syndrome were recorded for correlation analysis.Results Partial least-squares regression was introduced for analysis since multicollinearity among variables was revealed by ordinary linear regression.Significant positive correlations between plasma uric acid and BMI,FPG,TCHOL,TG,LDL-C and BP and negative correlation between the plasma uric acid and HDL-C were observed.Significant linear correlations were observed between plasma uric acid and other factors with fitting analysis.BMI,BP and FPG were revealed to be the leading correlated factors of plasma uric acid according to VIP analysis.Successful fitting result was proved with predictive modeling analysis.Increase of uric acid induced increase of risk factors and severity of metabolic syndrome.Conclusion Serum uric acid is closely associated with MS,thus hyperuricemia is proved important risk factor and warning factor of metabolic syndrome.