Estimating glomerular filtration rate by serum creatinine or/and cystatin C equations: An analysis of multi-centre Chinese subjects

    loading  Checking for direct PDF access through Ovid



Various equations based on serum creatinine or/and cystatin C, required further validation in a Chinese population. We compared the performance of six Chinese equations (Mascr, Peiscr, Macys, Fengcys, Mascr-cys and Fengscr-cys) with the CKD-EPI equations in multi-centre Chinese subjects and evaluated their applicability in clinical practice.


A total of 1522 adult patients from four different hospitals of China were enrolled in the study. 99mTc-DTPA renal dynamic imaging was used as the reference GFR (rGFR), and serum creatinine and cystatin C were measured by standardized assays. An optimal score system was implemented in the study.


The average rGFR of recruited subjects was 67.30±28.89 mL/min per 1.73m2. All estimated GFR (eGFR) correlated well with rGFR. In accordance with Bland–Altman analysis, the Fengscr-cys equations achieved optimal overall performance (score 14 vs 0–6), with least bias (median difference, −0.57 mL/min per 1.73m2; median absolute difference, 8.83 mL/min per 1.73m2), best precision (17.99 mL/min per 1.73m2), highest accuracy (percentage of eGFR within 15%, 30% and 50% of the rGFR (P15, P30 and P50; 49.7%, 78.7% and 91.8%, respectively); root-mean-square-error (RMSE, 16.28)). The Fengcys equation, a typical cystatin C based equation, was another well-behaved formula with an impressive performance. The Ma equations performed much poorer than the CKD-EPI equations. Consistent results can be observed in the GFR- /age- and sex-specific subgroups, while none equation yielded ideal accuracy in GFR<60 mL/min per 1.73 m2 subgroup.


The Fengscr-cys equation appeared to achieve the best performance for GFR estimation in overall Chinese adult patients. However, further research is warranted to improve the accuracy of available equations in GFR less than 60 mL/min per 1.73 m2 individuals.


This paper examines the predictive value of eGFR equations in a hospitalised population in China. While these equations were not derived for use in this population, data such as this examining the accuracy of these equations in this setting is useful.

Related Topics

    loading  Loading Related Articles