Comparison of the Estimated Glomerular Filtration Rate (eGFR) in Diabetic Patients, Non-Diabetic Patients and Living Kidney Donors.

BACKGROUND/AIMS We have reported that the eGFR overestimates renal function when glycemic control is poor. It has been reported that eGFR calculated by serum creatinine underestimates GFR in living kidney donors. We compared the utility of the eGFR in diabetic patients, non-diabetic patients and living kidney donors. Forty diabetic patients, 40 non-diabetic patients, and 40 living kidney donors were enrolled. METHODS GFR was measured by inulin clearance (C(in)). eGFR was calculated based on serum creatinine (eGFR(cr)) or serum cystatin C (eGFR(cys)). We compared the agreements between each of the eGFR and C(in) in each group. RESULTS There were significant and positive correlations between each eGFR and C(in) in diabetic patients and non-diabetic patients. However, the intraclass correlation coefficients (ICC) between each eGFR and C(in) in diabetic patients (ICC: eGFR(cr) 0.699, eGFR(cys) 0.604) were weaker than those in non-diabetic patients (ICC: eGFR(cr) 0.865, eGFR(cys) 0.803). The correlation coefficients between each eGFR and C(in) (eGFR(cr); r = 0.422, p = 0.0067 and eGFR(cys); r = 0.358, p = 0.0522) in living kidney donors were significantly weaker than those in non-diabetic patients. The ICCs between each eGFR and C(in) (ICC: eGFR(cr) 0.340, eGFR(cys) 0.345) in living kidney donors were significantly weaker than those in non-diabetic patients. CONCLUSIONS Based on C(in), eGFR was accurate in non-diabetic patients. However, eGFR was inaccurate in living kidney donors and relatively inaccurate in diabetic patients.