The accuracy of cystatin C and commonly used creatinine-based methods for detecting moderate and mild chronic kidney disease in diabetes

Background The accuracy of measuring serum cystatin C levels for detecting various stages of chronic kidney disease (CKD) in diabetes is still unclear. Methods In a cross-sectional study of 251 subjects, a reference glomerular filtration rate (GFR) was measured using Tc-99c-DTPA plasma clearance (iGFR). Multivariate analysis was used to identify independent clinical and biochemical associations with serum cystatin C and iGFR levels. The diagnostic accuracy of cystatin C and commonly used creatinine-based methods of measuring renal function (serum creatinine, the MDRD four-variable and Cockcroft-Gault formulae) for detecting mild and moderate CKD was also compared. Results In the entire study population the same five variables, age, urinary albumin excretion rates, haemoglobin, history of macrovascular disease and triglyceride levels were independently associated with both cystatin C and iGFR levels. A serum cystatin C level cut-off > 82.1 nmol/l (1.10 mg/l) had the best test characteristics as a screening tool for detecting moderate CKD (< 60 ml/min per 1.73 m(2)) when compared with creatinine-based methods. At the upper threshold for mild CKD (< 90 ml/min per 1.73 m(2)), cystatin C also had greater diagnostic accuracy than creatinine, but had similar diagnostic accuracy when compared with creatinine-based formulae for predicting renal function. Conclusions This study suggests that the clinical and biochemical parameters associated with serum cystatin C levels are closely linked to those associated with GFR and highlights the potential usefulness of screening for moderate or mild CKD in subjects with diabetes by simply measuring serum cystatin C levels.