| Background and Aims: Diabetic nephropathy is one of the most common complications of diabetes, and it is the most common cause of.end-stage renal disease.Therefore it is important to evaluate these patients’ renal function accurately.Glomerular filtration rate(GFR) is an important indicator of renal function, and99mTc-DTPA(99m Tc-diethylenetriamine pentaacetic acid) renal dynamic imaging is widely used for the determination of GFR. However, it is expensive and complex,which is an important factor to restrict its use in wide screening and basic hospitals.Today, estimated glomerular filtration rate(e GFR) equations based on general population are widely used, however, it has become obvious that these equations fail to provide the same performance in certain population because of different features.The aim of this study is to compare the performance of different e GFR equations and explore the most suitable one in Chinese patients with type 2 diabetes.Methods: A total of 326 in-patients with type 2 diabetes were enrolled from 2010 to2014. Patients’ age, sex, height, weight, serum creatinine and cystatin C(Cys C) were recorded. GFR was measured using 99 m Tc-DTPA-measured GFR(m GFR) as the reference standard. e GFR was calculated with Cockcroft-Gauh(C-G) equation(e GFR1 equation), modification of diet in renal disease-4(MDRD-4) equation(e GFR2 equation), Chinese equation(e GFR3 equation), Chronic Kidney Disease Epidemiology Collaboration(CKD-EPI) creatinine equation(e GFR4 equation),CKD-EPI Cys C equation(e GFR5 equation), CKD-EPI Cys C and creatinine equation(e GFR6 equation), Feng Cys C equation(e GFR7 equation) and Feng Cys C and creatinine equation(e GFR8 equation). e GFR was compared with m GFR using the paired t test and Pearson correlation analysis. The performance of e GFR equation was analyzed by Bland-Altman analysis, bias, precision, accuracy and receiver operating characteristic(ROC) curve analysis. The accepted tolerance of Bland-Altman analysis was defined as 60 m L/(min·1.73 m2). Bias is estimated by the mean difference between e GFR and m GFR; precision is estimated by the standard deviation of bias;accuracy is estimated within 10%, 30% and 50% of m GFR. All of the data were performed with SPSS, version 19.0 and Med Calc 12.7.Results: The m GFR of all patients was 60.46±27.20 m L/(min·1.73 m2). All e GFR equations overestimated m GFR except e GFR1 equation [61.29±30.39 m L/(min·1.73m2), P=0.450]. The consistency limits of these equations all exceeded the accepted tolerances according to Bland-Altman analysis, and e GFR6 equation [66.1(-41.1~25.0) m L/(min·1.73 m2)] was better than other equations. e GFR5 equation[-0.83 m L/(min·1.73 m2)] was the least biased among eight equations. e GFR4 equation [18.35 m L/(min·1.73 m2)], e GFR5 equation [19.75 m L/(min·1.73 m2)],e GFR6 equation [16.84 m L/(min·1.73 m2)] and e GFR7 equation [21.67 m L/(min·1.73m2)] had the highest precision among eight equations. e GFR5 equation and e GFR6 equation had the highest accuracy within 10%(24.54% and 26.99%, respectively),30%(68.71% and 65.64%, respectively), and 50%(87.42% and 84.36%, respectively)of m GFR. The area under ROC curves were highest with e GFR6 equation(0.905) and e GFR4 equation(0.891) among eight equations.Conclusions: e GFR5 equation and e GFR6 equation were superior to other six equations in Chinese patients with type 2 diabetes. The consistency limits of these equations all exceeded the accepted tolerances, therefore the two equations should be further improved before they can be applied to clinical use as a routine tool. |
PDF Full Text Request