# Estimation of Glomerular Filtration Rate Using Cystatin C Improves Risk Classification Compared to Estimation Using Creatinine

 

 

      EBM Focus - Volume 8, Issue 37 

[Read the Complete EBM Focus and Earn CME](http://archive.constantcontact.com/fs028/1102736301344/archive/1102783085863.html)

Reference: ([N Engl J Med 2013 Sep 5;369(10):932](http://www.ncbi.nlm.nih.gov/pubmed/24004120?dopt=Abstract)), ([level 1 \[mid-level\] evidence](https://dynamed.ebscohost.com/content/LOE))

Glomerular filtration rate (GFR) is a standard measure of kidney function that is used to stage chronic kidney disease (CKD). The KDIGO staging system defines CKD as GFR< 60 mL/minute/1.73 m2. CKD stages correspond to categories of GFR values, each with a range of 15 mL/minute/1.73 m2, with lower GFR values indicating higher stage (more severe) disease ([KDIGO 2013 Jan PDF](http://www.kdigo.org/clinical_practice_guidelines/pdf/CKD/KDIGO_2012_CKD_GL.pdf)). Because direct measurement of GFR is difficult, it is usually estimated (eGFR) by an equation typically using age, sex, and biomarker parameters. Several different estimation methods (CKD-EPI, MDRD, Cockcroft-Gault) have used creatinine as the principle marker of renal function. Though data have been inconsistent across different patient populations, the CKD-EPI creatinine equation has been shown to be comparable to the commonly used MDRD equation ([Scand J Clin Lab Invest 2011 Apr;71(2):129](http://www.ncbi.nlm.nih.gov/pubmed/21208031?dopt=Abstract&holding=caugamlib)). Newer equations have recently been developed using cystatin C, either alone or in combination with creatinine, and they may be more accurate than creatinine-based equations. This is in part because the filtration rate of cystatin C is unaffected by age, race, gender or muscle mass. A recent large validation cohort study compared risk classification based on CKD-EPI equations using cystatin C or creatinine.

Individual data from 90,750 patients from 11 general-population cohorts who had had standardized baseline measurements of both serum creatinine and cystatin C were analyzed. eGFR was calculated for each patient using CKD-EPI equations using cystatin C or creatinine. At baseline, the prevalence of CKD was 13.7% using cystatin C equation and 9.7% using creatinine equation. All-cause mortality was 13.6% overall in mean follow-up of 7.7 years.

CKD staging using the cystatin C equation was compared to staging using the creatinine equation. Using cystatin C, 21.5% of patients were classified to a higher CKD stage, and 19.3% to a lower stage compared to their creatinine-based stage. Classification to a higher stage by the cystatin C equation was associated with significantly increased risk of mortality for each category of GFR values, with hazard ratios for risk of death ranging from 1.36 to 3.04 (p < 0.05 for each GFR category). Conversely, classification to a lower stage was generally associated with reduced risk or mortality, though these reductions were significant only for stages corresponding to GFR categories of 45-59 mL/minute/1.73 m2 and 30-44 mL/minute/1.73 m2.

The improvement in predictive performance using the CKD-EPI cystatin C equation was summarized by the net reclassification index (NRI), which assesses the relative rates of appropriate and inappropriate reclassification (with positive value indicating improvement). For all-cause mortality, the NRI was 23% (95% CI 18%-28%). The cystatin C equation also associated with improved performance for prediction of cardiovascular death (NRI 17%, 95% CI 11%-23%) and end-stage renal disease (NRI 10%, 95% CI 0%-21%). Performance of CKD-EPI equation using both cystatin C and creatinine was similar to that of the cystatin C only equation. Similar results were also obtained in additional analysis of 5 cohorts with 2,960 patients with chronic kidney disease at baseline.

For more information, see the [Chronic kidney disease](http://web.ebscohost.com/dynamed/detail?sid=ea3e08f8-9144-44f4-93df-42f363d1e600%40sessionmgr13&vid=1&hid=18&bdata=JnNpdGU9ZHluYW1lZC1saXZlJnNjb3BlPXNpdGU%3d#db=dme&AN=115336) topic in DynaMed.