Serum creatinine is an unreliable method of assessing renal function of renal allografts. There are very few and conflicting reports on the reliability of use of various GFR formulae in renal transplant recipients. To assess the performance and accuracy of various GFR formulae in estimating the renal function of renal allograft patients, GFR was measured using an isotope method in 95 stable post renal transplant patients (6 months to 10 years, mean 5.2 years). GFR was estimated in these patients (using calibrated serum creatinine) by the following formulae: MDRD (for whites and blacks), Cockcroft-Gault, Nankivell. In addition GFR was calculated using cystatin C. Comparisons between means of each method to the gold standard were studied by using paired t-test. bias and precision were calculated. The overall agreement was evaluated by median absolute difference, median percent absolute difference and percent of calculated GFR values falling within 20% of measured GFR. GFR Bias( p value) Precision (p value) GFR by isotope 57.4 (27.4) Nankivell 64.5(15.9) +7.1 (0.03) 0.34 (0.01) Cockcroft-Gault 69.5 (23.7) +12.1 (0.007) 0.1 (0.44) Cystatin C 56.0 (23,2) -1.4 (0.634) 0.64 (0.0001) MDRD (whites) 58.9 (15.0) +1.5 (0.45 0.6 (0.0001) MDRD (black) 71.5 (18.1) +14.1 (0.0001) 0.6 (0.0001). The best correlation and least bias (2.4% and 2.6% respectively) was seen when using cystatin C and MDRD (white) formulae. The least correlation and largest bias was seen when using MDRD (black) and Cockcroft-Gault formulae both of which overestimated the GFR by 24.6% and 12.4% respectively.