Validity of estimated glomerular filtration rates for assessment of baseline and serial renal function in patients with atherosclerotic renal artery stenosis: implications for clinical trials of renal revascularization.

BACKGROUND Despite routine use of estimated glomerular filtration rates (GFRs) as major renal end points in clinical trials of renal revascularization, serial GFR estimates have never been validated in patients with renal artery stenosis (RAS). The purpose of this study was to evaluate the validity of GFR estimates in patients with atherosclerotic RAS. METHODS AND RESULTS Serum creatinine (SCr) and (125)I-iothalamate GFR (I-GFR) were measured in patients with RAS. GFR estimates were calculated from Modification of Diet in Renal Disease (MDRD), Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI), and Cockroft-Gault (CG) formulas. Using I-GFR as the reference standard, the sensitivity, specificity, and receiver operating characteristic area under the curve (AUC) were determined for MDRD, CKD-EPI, CG, and reciprocal SCr for identifying I-GFR <60 mL/min per 1.73 m(2) and a 20% change in I-GFR over time. Between 1998 and 2007, 541 I-GFR measurements were performed in 254 consecutive patients with RAS. MDRD, CKD-EPI, and CG GFR estimates demonstrated good sensitivity (86% to 95%), modest specificity (67% to 71%), and good reliability (AUC, 0.86 to 0.94) for identifying I-GFR <60 mL/min per 1.73 m(2). GFR estimates had good specificity (87% to 95%), poor sensitivity (0% to 45%), and poor reliability (AUC, 0.61 to 0.65) for detecting 20% changes in I-GFR over follow-up. CONCLUSIONS In patients with RAS, GFR estimates demonstrate good sensitivity and modest specificity for identifying I-GFR <60 mL/min per 1.73 m(2) but poor sensitivity and reliability for detecting 20% changes in I-GFR. GFR estimates should not be used in clinical trials as major end points to assess serial GFR after renal revascularization.