Capillary Electrophoresis Method For Analysis Of Microalbuminuria In Patients With Type 2 Diabetic Nephropathy

Objective Diabetic nephropathy(DN)is one of the most serious common complications of type 2 diabetes mellitus(T2DM).Microalbuminuria(m ALB)is an important clinical indication for diagnosis of DN.Accurate m ALB detection in early DN is significant for the treatment and prognosis of DN.Immunoturbidimetric assay(ITA)and radioimmunoassay(RIA)are two most commonly used assays for m ALB detection.However,low sensitivity of ITA on m ALB and radioactive contamination of RIA lead to requirement for more accurate and safe clinical methods of m ALB detection.In recent years,capillary electrophoresis(CE)has been used in the detection of trace element in serum and urine.In this research,the effect of CE in the detection of m ALB concentration is studied.We proposed a two factor regression model based on peak height and area of absorbance to estimate m ALB concentration,so as to improve the accuracy of m ALB detection and provide technical support for early diagnosis of DN.Methods 1.Experimental group: we split the diabetes mellitus into 4 groups: 32 cases of healthy control group(control),36 cases of simple diabetes group(DM),19 cases of early diabetic nephropathy group(DN-1),and 22 are clinical diabetic nephropathy group(DN-2).Urine samples(age: between 35 and 75)were collected from a controlled group in the Tianjin Medical University Metabolic Diseases Hospital,and are grouped in accordance wirh immunity tranmission turbidity diagnosis.2.The process of urine samples: After the urine samples were centrifuged.The supernatant obtained was transferred to 5ml tubes,mixed with PMSF saturated solution in a ratio of 1/1000(v/v),and stored at-80 prior to analysis.? 3.Experimental parameters: m ALB(99% purity)were used to configure standard samples,and then used in the separation of CE to systematically investigate the experimental parameters of concentration and p H of the buffer solution,the separation voltage,optimal wavelength,and SDS concentration.4.Experimental data: the different standard concentrations of m ALB were performed by CE under optimum conditions,in order to obtain six peaks height,peak area and m ALB concentration.5.Establishing three regression models: the linear regression model of absorbance peak height and m ALB concentration;the linear regression model of absorbance peak area and m ALB concentration;the binary linear regression model of absorbance peak height,peak height and m ALB concentration.6.Test and evaluation of three kinds of regression models: these three regression models were tested by F-test and multiple correlation coefficient test respectively,the significance and the complex correlation of the model were evaluated,and the optimized regression model was determined.7.Estimation of 4 groups of m ALB concentration with two factor linear regression model : Under the optimum experimental conditions of CE,two factor linear regression model is used to estimate 4 groups of m ALB concentration.8.Statistically analyzed: The variance analysis(ANOVA)method was used to compare the m ALB concentration of 4 groups of urine,and the subjects were re-entered according to the concentration of m ALB in the urine samples,and the differences between ITA and CE were analyzed.Results 1.3 regression models:(1)Absorbance peak height vs m ALB concentration linear regression model: y = 3.34 x1-4.94,F-test,P = 0.001;Correlation coefficient: R = 0.9979,residue is 0.0042.(2)Absorbance area vs m ALB concentration linear regression model: y = 0.12 x2-17.5,F-test,P = 0.001;Correlation coefficient: R = 0.9983,residue is 0.0033.(3)Absorbance peak height and area vs m ALB concentration two factor linear regression model: y=1.05 x1+0.08 x2+6.76,F-test,P = 0.001;Correlation coefficient: R = 0.9992,residue is 0.0017.Correlation coefficient R in two factor linear regression model is significantly higher than that of one factor linear regression model,two factor linear regression model is used to estimate 4 groups of m ALB concentration.2.Estimation of 4 groups of m ALB concentration: Estimation of m ALB in each groups of urine sample is : control group 12.18±0.33 mg/L;DM group 14.23±0.40 mg/L;DN-1 group 107.07±8.30 mg/L;DN-2 group 494.60±50.01 mg/L.DM group has no statistical difference with control group;DN1 group is significantly higher than control and DM group;DN2 is significantly higher than the other three groups.3.Make new groups according to the estimated m ALB concentration: m ALB concentration of 2 patients in DM group is higher than entry criteria of DN-1(20mg/L)and classified to early DN group(DN1);m ALB concentration of 1 patient in DN1 is higher than entry criteria of DN-2(200mg/L)and classified into clinical DN(DN-2).The results imply that CE may detect ITA miss detected DN-1,DN-2 patients.Conclusion 1.Correlation coefficient of two factor linear regression model based on absorbance peak height and area is higher than that of one factor linear regression model.This implied that two factor linear regression model is the optimal model of m ALB concentration estimation.2.Significance of difference is higher with two factor linear estimation model in estimating m ALB concentration in DM group and DN-1 group.This implied that optimized model has higher sensitivity for early DN detection.3.Positive rate DN-1 and DN-2 with CE for 4 groups of urine samples are higher than that of ITA: 2 of 40 samples diagnosed as negative in DM group with ITA are DN-1;1 of 27 patients diagnosed as DN-1 with ITA is DN-2.The results show CE used in m ALB detection is superior to ITA.