肾小球滤过率评估公式在老年慢性肾脏病患者中的适用性研究

An applicability study of glomerular filtration equation in elderly patients with chronic kidney disease in China

目的了解常用的。肾小球滤过率评估公式在中国老年慢性肾脏病（CKD）患者中的适用性。方法选择在我院就诊、年龄75～96岁的非外籍老年CKD患者324例，分别采用Cockcorft—Gault公式（CG）、校正CG公式、肾脏疾病饮食改良系列公式（MDRD）和慢性肾脏病流行病学合作研究公式（CKD-EPI）估测。肾小球滤过率（GFR），并与体表面积标准化以^99mTc-二乙三胺五乙酸法（^99mTc—DTPA法）测定的GFR（mGFR）结果进行比较；以mGFR≥或〈60ml·min^-1·1．73m^-2为界，将患者分为高mGFR组与低mGFR组，采用偏差、准确率、均方预测误差、线性相关及直线回归分析、Bland—Altman曲线分析、工作特征曲线（ROC）下面积分析等方法评估各公式在总体以及各亚组中的适应性的优劣，并探索使用数据转换的方式进一步对估测公式的应用价值进行评估。结果（1）总体老年CKD患者中，CKDEPI公式较其他公式具有较小的偏差、更优的准确性和更强的相关性，但各公式30％准确率均低于70％，Bland—Altman曲线分析显示校正CG公式一致性较好，其余公式一致性均欠佳；经数据转换，可改善各公式的准确性。（2）以GFR=60ml·min^-1·1．73m^-2为临界值，各公式ROC曲线下面积比较差异均无统计学意义（P〉0．05），诊断价值无差异。（3）在高mGFR水平时，CKD-EPI公式偏差、准确率明显优于其他公式；在低mGFR水平时，简化MDRD公式估测偏差低于其他公式，CG公式15％准确率高于其他公式。结论当血肌酐的测定方法为酶法时，以临床常用的GFR评估公式预测老年CKD患者的GFR，可能产生明显偏差。CKD-EPI公式（尤其在高GFR水平）的适用性优于其他公式，可能更适合我国老年CKD患者。个体化选用公式和适当数据转换可能会提高预测公式的准确率。

Objective To evaluate the applicability of glomerular filtration equation in elderly patients with chronic kidney disease （CKD） in China. Methods Based on measured glomerular filtration rate（mGFR）obtained with 99mTc-DTPA renal dynamic imaging method, 324 elderly Chinese CKD patients were enrolled in this study. The GFR was estimated using the CG equation, calibrated CG equation, simplified MDRD equation （aMDRD）, re-expressed MDRD equation （rMDRD equation）, modified MDRD equation for Chinese CKD patients （cMDRD equation） and CKD epidemiology collaboration （CKD-EPI） equation. The patients were divided into high mGFR group [mGFR〈60ml·min^-1·1.73m^-2] and low mGFR group [mGFR〈60ml·min^-1·1.73m^-2]. Performance in bias, accuracy, mean squared predication residual （MSPR）, linear correlation, linear regression analysis, receiver operating characteristic （ROC） curve and Bland Altman plot analysis of these equations were assessed in the entire population and in subgroups. Data conversion was used to further evaluate the application value of these equations. Results In the entire population, when comparing the bias and accuracy of estimated GFR （eGFR） with mGFR,the eGFR estimated by CKD-EPI equation showed better results. But the agreement limits of the equations, except for the calibrated CG equation, exceeded the prior acceptable tolerances, defined as 60ml·min^-1·1.73m^-2. The 30% accuracy of all the equations was lower than 70% level. Data conversion might improve the performance. The receiver operating characteristic （ROC） curve showed that the differences in maximum diagnostic accuracy among all the equations for the diagnosis of renal failure [GFR〈60ml·min^-1·1.73m^-2 were not statistically significant （P〉0.05）. In the high GFR level subgroup with a mGFR≥60 ml·min^-1·1.73m^-2, CKD-EPI equation showed less bias and higher accuracy than other equations. In the low GFR level subgroup with a mGFR〈60 ml·min^-1·1.73m^-2 , simplified MDRD equation showed less bias than other equations. The 15% accuracy of CG equation was higher than that of other equations. Conclusions None of these equations were suitable for estimating GFR in the elderly Chinese population investigated. At present, based on overall performance, especially in the high GFR level subgroup, the CKD-EPI equation may be the most accurate for estimating GFR in elderly Chinese patients with CKD. Individual choice of equations and proper data conversion should be used if necessary.