INTRODUCTION The glomerular filtration rate (GFR) is an important indicator of renal function, and its precise measurement is essential for guiding clinical management. However, studies evaluating the performance of GFR estimation equations in undialyzed patients with end-stage renal disease (ESRD) remain scarce.OBJECTIVES Our work sought to identify a relatively accurate equation for estimating the residual renal function in undialyzed patients with ESRD.PATIENTS AND METNODS We used the revised Gates method as the gold standard to measure GFR in 101 undialyzed patients with ESRD. We used a total of 23 equations (CKD-EPIScr, MDRDII, FASScr, EKFC, revised Lund-Malmo (LMR), Mayo, XiangYa, XiangYa-s, Vilar, Shafi beta 2M, CKD-EPISCysC, FASSCysC, CAPA, Hoek, Yang, CKD-EPIScr-SCysC, FASScr-SCysC, Adachi, Shafi beta TP, Shafi beta TP-beta 2M, Wong, CKD -EPI_3M, and CKD -EPI_4M) to estimate the patients' GFR.RESULTS The GFR measured by the dual plasma sampling method (dGFR) and the revised Gates method (rGFR) showed high agreement. The median dGFR and rGFR were 13.1 ml/min/1.73 m2 and 10.7 ml/ min/1.73 m2, respectively. In the investigated population, the LMR equation showed a low bias (median, 0.6), high precision (interquartile range [IQR], -3.25 to 1.05), and the highest accuracy (P30, defined as the percentage of eGFR within 30% [70%-130%] of rGFR, 65.3%).CONCLUSIONS Based on comparison of 23 equations, we recommend using the LMR equation, despite its large deviations, to estimate GFR in undialyzed patients with ESRD.