Evaluation of various formulae for glomerular filtration rate estimation and proposal of new formulae for the Korean population

Background : Some researchers have questioned the necessity of adjusting glomerular filtration rate (GFR) by body surface area (BSA). We compared the relationship between estimated GFR (eGFR) and radionuclide GFR (rGFR) with or without BSA adjustment by comparing the results obtained using various formulae with those obtained using 2 new proposed formulae. Methods : A retrospective study was performed using 204 Korean individuals whose GFR had been estimated by the ^99mTc-diethylenetriaminepentaacetic acid method between March 2004 and July 2008. We used the modification of diet in renal disease (MDRD) II formula, Mayo clinic quadratic (MCQ) formula, Cockcroft-Gault (CG) formula, and lean body mass-adjusted CG formula. Two new formulae, skeletal muscle mass index (SMI)-adjusted CG formula and SMI x 3.4/SCr, were proposed by us. We analyzed each parameter with Pearson's correlation coefficient and also obtained the bias values. Results : BSA did not satisfy the fundamental prerequisites of an adjustment facta for rGFR. MDRD Il and MCQ GFR estimates demonstrated higher Pearson's correlation coefficient with BSA-unadjusted rGFR than they did with BSA-adjusted rGFR. The other GFR formulae estimates showed better correlation with rGFR and more favorable bias (P<0.001) when both GFR estimates and rGFR values were BSA-unadjusted. SMI-adjusted CG and SMI x 3.4/SCr GFR estimates demonstrated correlation with rGFR and bias values similar to those of the MDRD Il and CG GFR estimates. Conclusions : We suggest that absolute, non-corrected GFR and GFR estimate be preferred in daily practice. The absolute, non-corrected GFR and GFR estimate are considered helpful for patients with eGFR ≤. 60 mL/min/1.73 m². We also recommend the clinical use of the new formulae, SMIadjusted CG and SMI × 3.4/SCr (BSA-unadjusted).