Improved GFR Estimation by Combined Creatinine and Cystatin C Measurements

Y-C Ma; L Zuo; J-H Chen; Q Luo; X-Q Yu; Y Li; J-S Xu; S-M Huang; L-N Wang; W Huang; M Wang; G-B Xu; H-Y Wang


Kidney Int. 2007;72(12):1535-1542. 

In This Article

Abstract and Introduction


Plasma creatinine may not reflect glomerular filtration rate (GFR) especially in the early stages of chronic kidney disease (CKD). Plasma cystatin C (cysC), however, has the potential to more accurately determine early GFR reduction. We sought to improve the creatinine-based GFR estimation by including cysC measurements. We derived a reference GFR from standard dual plasma sampling 99mTc-DTPA clearance in a training cohort of 376 randomly selected adult Chinese patients with CKD. We compared reference values to estimated GFR and applied multiple regression models to one equation based solely on cysC, and to another combining plasma creatinine (Pcr) and cysC measurements of the training cohort. The results were validated by testing an additional 191 patients. The difference, precision, and accuracy of the two estimates were compared with the modified Modification of Diet in Renal Disease (MDRD) equation for Chinese patients, and another estimate combining cysC and modified MDRD calculations. The estimated GFR combining Pcr and cysC measurements more accurately matched the reference GFR at all stages of CKD than the other equations, particularly in patients with near-normal kidney function.


The number of patients with chronic kidney disease (CKD) is growing worldwide;[1,2] CKD has been recognized as one of the independent risk factor of cardiovascular disease (CVD), even in its early stages.[3,4] Detection of decreased kidney function and identification of CKD in its early stage are thus very important, but because overestimation of true glomerular filtration rate (GFR) will lead to insufficient treatment and underestimation of true GFR will lead to an unnecessary intervention and waste of medical resources,[5] a simple and accurate GFR estimating method is of great clinical importance.

The Modification of Diet in Renal Disease (MDRD) equations provided reasonable accuracy in Whites and African Americans in GFR estimation,[6] and were recommended by the Kidney Disease Outcome Quality Initiatives clinical practice guidelines.[7] Recently, Levey et al. re-expressed MDRD equations using the standardized plasma creatinine (Pcr) assay,[8] and claimed that clinical laboratories can report more accurate GFR estimates in patients with true GFR of less than 90mlmin-11.73m-2. Also, modified MDRD equations based on Chinese CKD patients showed significant improvement compared with the original ones.[9]

Both original MDRD equations and the modified MDRD equations underestimated GFR in CKD stage 1 and 2,[9,10] and re-expressed MDRD equations, however, do not perform well in patients with higher GFR levels.[8] This is partly because the main GFR predictor in these equations is the Pcr, and Pcr's relationship with GFR is different among stages of CKD,[9,11] and different levels of Pcr do not necessarily reflect the true variation of GFR.[12] For example, there was no significant elevation of Pcr levels with the decrease of GFR at the very early stages of CKD because of tubular secretion. These phenomena contribute to the inaccuracy of Pcr-based equations in early stages of CKD, and may be an unavoidable pitfall of Pcr-based GFR estimation equations.

Plasma cystatin C level (cysC) increases earlier than Pcr as GFR decreases, and may be a valuable marker in detecting early renal function impairment.[13,14] It fulfills a number of criteria for endogenous marker of GFR: It is freely filtered and catabolized in the proximal tubule, without being secreted.[15] Unlike Pcr, cysC does not depend on sex and muscle mass, and does not change with age between 1 and 50 years.[15] In several studies[14,16,17] and in one meta-analysis,[18] cysC has been reported to be superior to Pcr in GFR estimation, particularly in patients with near-normal kidney function.

Researchers have developed cysC-based GFR-estimating equations, and have compared their performance with original MDRD equations. White et al.[19] demonstrated that cysC-based equations are more accurate in GFR prediction in renal transplant recipients, than abbreviated MDRD equations. Poge et al.[20] found that cysC-based equations might offer advantages compared with abbreviated MDRD equations, in cirrhotic patients. Grubb et al.[21] reached the similar conclusions in patients with various renal diseases. More recently, Rule et al.[22] claimed that estimating the geometric mean of a cysC-based equation and a Pcr-based equation improved GFR estimation.

In this study, an GFR equation was developed by a log transformed regression model, which simultaneously used Pcr and cysC as independent variables, and compared with the modified MDRD equation for Chinese and a composite equation constituted from the modified MDRD equation and a solely cysC equation.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: