Multiple Biomarkers for Predicting Contrast Nephropathy after PCI in Patients with Acute Coronary Syndrome Based on LASSO Regression
Objective:To explore the value of multiple biomarkers for predicting contrast nephropathy after percutaneous coronary intervention(PCI)in patients with acute coronary syndrome based on least absolute contraction and selection algorithm(LASSO)regression.Methods:One hundred and nineteen patients with acute coronary syndrome were selected as the study subjects.The general conditions,blood biochemical indexes,inflammatory factors,biomarkers,and the occurrence of contrast-induced nephropathy for 3 days after PCI were recorded.The 10 fold cross validation LASSO regression was used to screen the characteristic variables.The independent predictors of contrast nephropathy after PCI in patients with acute coronary syndrome were obtained by substituting characteristic variables as independent variables of the multivariate Logistic regression model.The correlation between biomarkers and contrast nephropathy after PCI was emphatically analyzed.The joint prediction model and nomogram model were constructed based on the independent predictors,and the calibration curve was drawn to verify the prediction efficiency of the nomogram model.Results:Ten fold cross validation LASSO regression screened out four characteristic variables with the most generalization ability:diabetes,urinary kidney injury molecule-1(KIM-1),urinary neutrophil gelatinase related apolipoprotein(NGAL),and urinary cystatin C(CysC),the corresponding LASSO regression coefficients were 0.436,0.624,0.916,and 2.745,respectively.After adjusting and correcting for confounding factors,diabetes,urinary KIM-1,urinary NGAL,and urinary CysC were independent predictors of contrast nephropathy after PCI in patients with acute coronary syndrome(P<0.05).The nomogram model was constructed based on the independent predictors of contrast nephropathy after PCI in patients with acute coronary syndrome:diabetes,urinary KIM-1,urinary NGAL,and urinary CysC,P=1/(1+e-X),X=-2.345+0.824×urine CysC+0.565×diabetes+0.685×urinary NGAL+0.634×KIM-1.The calibration curve of urinary and nomogram model showed that the predicted value of the risk of contrast nephropathy after PCI in patients with acute coronary syndrome was better agreement with the actual observation value.Conclusion:Diabetes,urinary KIM-1,urinary NGAL,and urinary CysC are independent predictors of contrast nephropathy after PCI in patients with acute coronary syndrome.The nomogram model based on independent predictors shows high predictive value.
acute coronary syndromecontrast nephropathyleast absolute contraction and selection algorithm regressionbiomarkerspredictive value
NETL
NSTL
万方数据
