Background: Estimation of glomerular filtration rate (GFR) is usually determined from 24-hour urine collection, but it is time-consuming, and difficult in clinical practice. The authors attempted to select an accurate and safe, but more convenient test to obtain an estimated GFR. Objective: To compare estimation of GFR by Bioelectrical impedance analysis (BIA) with GFR calculated by 24-hour urine averaged creatinine clearance and urea clearance (Ccr-Cu-GFR). Material and Method: The authors examined 79 non-diabetic chronic kidney disease (CKD) patients that had estimated GFR between 15 and 89 ml/min/1.73 m2. Subjects were categorized into three subgroups according to K/DOQI-CKD classification: GFR of 60-89 ml/min/1.73m2 (stage 2, 5 subjects), 30-59 ml/min/ 1.73m2 (stage 3, 31 subjects), and 15-29 ml/min/1.73m2 (stage 4, 43 subjects). Results: The mean value of Ccr-Cu-GFR was 33.79 ± 14.78 ml/min/1.73 m2 and GFR by BIA (BIA-GFR), 34.63 ± 14.86 ml/min/1.73 m2 with no overall statistical differences (p = 0.838). In stage 3 CKD patients, the mean BIA-GFR and Ccr-Cu-GFR were similar (38.84 ± 12.47 vs 41.16 ± 9.17, p = 0.399) while in stage 2 CKD, BIAGFR tended to underestimate (63.50 ± 19.35 vs 70.94 ± 7.82, p = 0.407) and in stage 4 CKD, BIA-GFR significantly overestimated Ccr-Cu-GFR (27.31 ± 9.11 vs 23.76 ± 5.68, p = 0.040). Conclusion: The findings suggest that BIA-GFR in non-diabetic CKD patients closely resembled with Ccr-Cu-GFR especially in stage 3 CKD patients. BIA-GFR may be considered as a more convenient test for an assessment of GFR in non-diabetic CKD patients.