After renal transplantation monitoring and detection of slight-to-moderate changes in GFR is a prerequisite for an optimal patient management. Recently, several equations to estimate GFR were developed and verified in the MDRD study cohort. However, little is known about the application of the MDRD formulas in the setting of renal transplantation. We prospectively conducted a study of the GFR estimates of the Cockcroft and Gault (C&G), MDRD6-, MDRD7 and the abbreviated MDRD (aMDRD) with the true GFR as measured by (99m)Tc-DTPA clearance in 95 consecutive patients 6.5, 5.3-7.7 years (mean, 95% CI) after renal transplantation. On average the DTPA clearance was 37.4, 34.4-40.4 mL/min/1.73 m(2), which differed significantly from estimates of GFR by C&G (52.6, 48.3-56.9 mL/min/1.73 m(2)), MDRD7 (44.8, 40.7-49.0 mL/min/1.73 m(2)), MDRD6 (43.8, 39.9-47.7 mL/min/1.73 m(2)) and aMDRD (46.6, 42.4-50.9 mL/min/1.73 m(2)). Bias was lowest for MDRD6 (6.4 mL/min/1.73 m(2)) and highest for C&G (15.2 mL/min/1.73 m(2)). Precision was similar for MDRD7 and aMDRD (10.6 and 11.1 mL/min/1.73 m(2)) but significantly better for MDRD6 (8.6 mL/min/1.73 m(2); p < 0.035). Accuracy within 50% of real GFR was 55.8% for C&G, 83.2% for aMDRD, 87.4% for MDRD7 and 90.5% for MDRD6. MDRD equations perform significantly better than the commonly used C&G formula. Moreover, the MDRD6 equation provides the best diagnostic performance, and should therefore be preferred in renal transplant recipients.