Comparison of Performance of Improved Serum Estimators of Glomerular Filtration Rate (GFR) to ^99mTc-DTPA GFR Methods in Patients with Hepatic Cirrhosis

Glomerular filtration rate (GFR) measurements are critical in patients with hepatic cirrhosis but potentially erroneous when based on serum creatinine. New equations for estimated GFR (eGFR) have shown variable performance in cirrhotics, possibly because of inaccuracies in reference methods for measured GFR (mGFR). The primary objective was to compare the performance of 4 improved eGFR equations with a 1-compartment, 2-sample plasma slope intercept ^99mTc-DTPA mGFR method to determine whether any of the eGFR calculations could replace plasma ^99mTc-DTPA mGFR in patients with cirrhosis. The secondary objective was to test the hypothesis that mGFR using voluntary voided urine collections introduces error compared with plasma-only methods. Methods: Fifty-four patients with hepatic cirrhosis underwent mGFR determinations from 2 plasma samples at 1 and 3 h after intravenous administration of 185 MBq of ^99mTc-DTPA. GFR was also generated by a UV/P calculation derived from blood and urine samples. These mGFRs were compared with the eGFRs generated by 4 estimating equations: MDRD (Modified Diet in Renal Disease), CKD-EPI (Chronic Kidney Disease-Epidemiology Collaboration) (serum creatinine [SCr]), CKD-EPI (cystatin [CysC]), and CKD-EPI (CysC+SCr). eGFRs were compared with mGFRs by Pearson correlation, precision, bias, percentage bias, and accuracy (eGFRs varying by <10% [p10], <20% [p20] or <30% [p30] from the corresponding mGFR). Results: All eGFRs showed poorer performance when the UV/P ^99mTc-DTPA mGFR was used as the reference than when the plasma ^99mTc-DTPA mGFR was used. When compared with the plasma ^99mTc-DTPA mGFR method, the performance of all eGFR equations was superior to most published reports. There was a moderately good positive correlation between eGFRs and mGFRs. When compared with plasma ^99mTc-DTPA mGFR, precision of eGFRs was in the range of 14-20 mL/min and showed a negligible bias. Compared with the plasma ^99mTc-DTPA mGFR, CKD-EPI (CysC+SCr) showed the best overall performance and accuracy, at 85.19% (p30), 75.93% (p20), and 42.59% (p10). Conclusion: Estimating equations for measuring eGFR performed better than in most published reports, attributable to use of the plasma ^99mTc-DTPA mGFR method as a reference. CKD-EPI (CysC+SCr) eGFR showed the best overall performance. However, more discriminating methods may be required when accurate GFR measurements are necessary. mGFR measurements using urine collections may introduce error compared with plasma-only methods.