A Logistic Regression Model for Microalbuminuria Prediction in Overweight Male Population

BaoDe Chen,WeiYi Xu,Chen Yu,ZhenYun Ni,XueFen Li,DaWei Cui
DOI: https://doi.org/10.1038/npre.2010.5372.1
2010-01-01
Nature Precedings
Abstract:Background: Obesity promotes progression to microalbuminuria and increases the risk of chronic kidney disease. Current protocols of screening microalbuminuria are not recommended for the overweight or obese.Design and Methods: A cross-sectional study was conducted. The relationship between metabolic risk factors and microalbuminuria was investigated. A regression model based on metabolic risk factors was developed and evaluated for predicting microalbuminuria in the overweight or obese. Results: The prevalence of MA reached up to 17.6% in Chinese overweight men. Obesity, hypertension, hyperglycemia and hyperuricemia were the important risk factors for microalbuminuria in the overweight. The area under ROC curves of the regression model based on the risk factors was 0.82 in predicting microalbuminuria, meanwhile, a decision threshold of 0.2 was found for predicting microalbuminuria with a sensitivity of 67.4% and specificity of 79.0%, and a global predictive value of 75.7%. A decision threshold of 0.1 was chosen for screening microalbuminuria with a sensitivity of 90.0% and specificity of 56.5%, and a global predictive value of 61.7%. Conclusions: The prediction model was an effective tool for screening microalbuminuria by using routine data among overweight populations.
What problem does this paper attempt to address?