Comparative performance of glomerular filtrate rate estimated by creatinine, cystatin C or both to predict contrast induced nephropathy in acute coronary syndrome

Introduction: Contrast-induced nephropathy (CIN) is an usual complication of percutaneous coronary interventions (PCI). Here we report the validation of the new cystatin C (CyC)-based glomerular filtrate rate GFR equations (CKD-EPI\_CyC and CKD-EPI\_Cr&CyC) to predict CIN in patients with ACS. We evaluated their performance, as compared with existing creatinine (Cr)-based GFR equations (CG, MDRD-4 and CKD-EPI_Cr), and tested their usefulness in improving the classification of patients in risk of CIN. Methods: 497 subjects admitted consecutively to our hospital (2008-2010) with a diagnosis of ACS and underwent PCI were included. Discriminative power and calibration index of each eGFR equation were calculated and compared, together with the predictive values (sensitivity, specificity). The incremental value that the new CyC-based eGFR equations brings over the sCr-based eGFR equations was assessed by the Net Reclassification Improvement (NRI). Results: 33 (6.6%) patients developed CIN. For GFR, the greatest C statistic was for CKD-EPI_CyC, significantly superior to CG (p<0.001), MDRD-4 (p=0.001) and CKD-EPIsCr (p<0.001). Among CG, MDRD-4 and CKD-EPICr no differences were found in terms of C statistic. However, by reclassification analysis, we found that CG presented a global improvement in the reclassification of 20.9% compared with MDRD-4 (p=0.01), using the cutoff value of 60 mL/min/1.73 m2. There was no difference according to the reclassification between CG and of CKD-EPI\_Cr. CKD-EPI\_Cr&CyC had also higher values of C statistic than CG (p<0.001), MDRD-4 (p=0.003) and CKD-EPI\_Cr (p=0.002). However, in comparison with CKD-EPI\_CyC, the combined equation (CKD-EPI_Cr&CyC) did not improve the discriminative capacity, even it showed a trend to have lower C index (p=0.074). Regarding to the calibration, CyC-based equations presented the highest indices. Based on cut-off of 60 ml/min/1.73m2, the CKD-EPI\_CyC was the equation with the highest percentage of patients correctly classified as at CIN risk (86.7%), similar to the CKD-EPI\_Cr&CyC (83.7%). For sensitivity, the values of the five equations were very similar. For specificity, CKD-EPI\_CyC was the best (87.7%), followed of CKD-EPI\_Cr&CyC (84.3%), whereas it was inferior to 80% for the other 3 equations. Conclusions: The new CyC-based GFR equations represent an advance over currently available Cr-based GFR equations to predict CIN in patients with ACS, improving the discrimination and calibration. Within CyC-based equations, the one that combines Cr and CyC do not improve the performance for CIN risk prediction of the simple equation based only on CyC.