Evaluation of serum creatinine- and cystatin C-based equations for the estimation of glomerular filtration rate in a Chinese population.

Xiao-Hua Pei,Juan He,Qiao Liu,Bei Zhu,Li-Hua Bao,Chen-Jing Yan,Jian-Qing Wu,Wei-Hong Zhao
DOI: https://doi.org/10.3109/00365599.2012.660985
2012-01-01
Scandinavian Journal of Urology and Nephrology
Abstract:Objective. This study aimed to evaluate the applicability of a selection of glomerular filtration rate (GFR) estimating equations based on serum creatinine (SCr) and serum cystatin C in a Chinese population. Material and methods. Estimated GFR values from 10 equations were compared with reference GFR (rGFR) from the Tc-99m-DTPA renal dynamic imaging method. The study enrolled 569 Chinese participants (41.5% women, 53.5 +/- 16.9 years, range 19-92 years), with mean rGFR 74.80 +/- 26.10 (range 9.8-146.8 ml/min/1.73 m(2)). Results. Bland-Altman analysis illustrated that the 95% agreement limits of all the equations surpassed the acceptable tolerance (<60 ml/min/1.73 m(2)), of which the MacIsaac equation was the closest one, reaching 71.7 ml/min/1.73 m(2). Linear regression analysis also demonstrated a consistent result. When assessed in all participants, the accuracy of the six equations reached and exceeded the acceptable level (>= 70%), of which the Shanghai and MacIsaac equations gained more accuracy than others. When compared in subgroups, the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI), MacIsaac and Cockcroft-Gault (CG) equations were optimal for rGFR stages >= 90 ml/min/1.73 m(2), 30-89 ml/min/1.73 m(2) and < 30 ml/min/1.73 m(2), respectively. Conclusion. The results demonstrated that further improvement is needed for the selected 10 equations. Not all the cystatin C equations were superior to SCr equations. They have their own applicability at various GFR levels. At present, the CKD-EPI, MacIsaac and CG equations may be applied to evaluate GFR in normal, mild to moderate and severe kidney function, respectively.
What problem does this paper attempt to address?