Objectives: To perform a systematic review comparing the diagnostic accuracy of CysC with SCr.
Methods: MEDLINE and EMBASE (January 1984-February 2006) were searched. Studies included i) evaluated CysC against a recognised 'gold standard' method for determining GFR using a receiver operating characteristics (ROC) curve analysis and ii) included data that could be extracted into a 2x2 table.
Results: The search identified 27 population groups in 24 studies (n=2007) that compared the diagnostic accuracy of CysC with SCr. The diagnostic odds ratios (DORs) (95% CI) of predicting renal dysfunction derived from a Moses-Littenberg linear regression model were 3.99 (3.41-4.57) for CysC and 2.79 (2.12-3.46) for SCr.
Conclusion: The diagnostic accuracy for impaired renal function favours CysC. However, the confidence intervals for the pooled DORs for the biomarkers overlap. The ability of CysC (cut-off values between 0.9 and 1.4 mg/L) to rule in renal impairment (as measured by inulin-determined GFR of 60-79 mL/min/1.73 m2) in persons in whom this is suspected is large and conclusive.