Evaluation of creatinine-based estimates of glomerular filtration rate in a large cohort of living kidney donors

• Published in: Transplantation Vol. 86; no. 2; p. 223
• Main Authors: Issa, Naim, Meyer, Kathryn H, Arrigain, Susana, Choure, Gautam, Fatica, Richard A, Nurko, Saul, Stephany, Brian R, Poggio, Emilio D
• Format: Journal Article
• Language: English
• Published: United States 27.07.2008
• Subjects: Adult; Cohort Studies; Creatinine - blood; Creatinine - urine; Female; Glomerular Filtration Rate; Humans; Kidney Function Tests; Kidney Transplantation - methods; Living Donors; Male; Middle Aged; Retrospective Studies; ROC Curve; Sensitivity and Specificity; Treatment Outcome
• ISSN: 0041-1337
• DOI: 10.1097/TP.0b013e318176aefc

Accurate determination of kidney function is critical in the evaluation of living kidney donors and higher donor glomerular filtration rate (GFR) is associated with better allograft outcomes. However, among transplant centers donor kidney function evaluation varies widely. The performance of creatinine clearance (CrCl), Modification of Diet in Renal Disease (MDRD), the re-expressed MDRD equations with standardized creatinine, and the Cockcroft-Gault (CG) formula as compared with (125)I-iothalamate GFR (iGFR) was analyzed in 423 donors. All methods of GFR measurement were then evaluated for their association with graft function at 1 year. The MDRD and re-expressed MDRD equations underestimated iGFR whereas CG showed minimal bias (median difference=-11.0, -16.3, and -0.5 mL/min/1.73 m(2), respectively). CrCl overestimated iGFR (10 mL/min/1.73 m(2)). The MDRD, re-expressed MDRD, and CG formulas were more accurate (88%, 86%, and 88% of estimates within 30% of iGFR, respectively) than CrCl (80% within 30% of iGFR). Interestingly, low bias and high accuracy were achieved by averaging the MDRD estimation with the CrCl result; both methods available to the clinician in most transplant centers. We also showed that predonation GFR as measured by isotopic renal clearance or any of the creatinine-based estimation formulas may be associated with allograft function at 1 year, whereas the widely used CrCl was not. Variable performance was seen among different GFR estimations, with CrCl being the poorest. Recent recommendations to use the MDRD equation with standardized serum creatinine did not improve its performance. However, recognizing the limited availability of GFR laboratories, these methods are still clinically useful if used with caution and understanding their limitations.