Comparing measures of cystatin C in human sera by three methods

Abstract

Background: Cystatin C (Cys C) is measured by particle-enhanced nephelometric immunoassay (PENIA), particle-enhanced turbidimetric immunoassay (PETIA) and ELISA.

Aim: To determine differences among these methods.

Method: 80 normal human sera and 20 from patients with renal and/or heart disease were simultaneously assayed. Statistical analyses including receiver operating characteristics (ROC) of the three methods were compared.

Results: There was a highly significant correlation across the assay range between the ELISA and PENIA (r(2) = 0.94) and PETIA methods (r(2) = 0.95). Analysis of variance and bias were poor between the ELISA and the other two methods. Mean difference between ELISA and PETIA was 0.65 +/- 0.63 microg/ml, while it was 0.58 +/- 0.53 microg/ml between ELISA and PENIA. Accuracy (at 30% range) was 17 and 11% between ELISA and PETIA and ELISA and PENIA, respectively. Normalization of the ELISA by a factor of 0.66 improved this relationship. AUC of ROC curves of PENIA, ELISA and normalized ELISA to predict Cys C levels measured from PETIA were all above 0.87 (p = not significant between curves). Criterion values of ELISA*.66 method was close to PETIA measurements.

Conclusion: There is a significant difference in measured human Cys C levels among the three methods, and normalization of ELISA narrows these differences.