Inaccuracy of glomerular filtration rate estimation from height/plasma creatinine ratio

Use of a height/plasma creatinine formula to estimate glomerular filtration rate (GFR) is simpler and less invasive than renal or plasma clearance methods. The aim of this study was to determine whether these formulas enabled accurate prediction of GFR measured from the plasma clearance of 51Cr labelled ethylenediaminetetra-acetic acid (51Cr-EDTA). Thirty nine patients underwent GFR measurement at least six months after potentially nephrotoxic chemotherapy. Altman-Bland analysis was performed on the measured GFR and that estimated simultaneously using the original and a modified Counahan-Barratt formula and the Schwartz formula. The limits of agreement of the estimated GFR with the measured GFR were unacceptably wide in each case, despite highly significant correlation coefficients. The bias was smallest for the modified Counahan-Barratt formula. Use of these formulas to estimate GFR in children is insufficiently accurate for research purposes and has limitations in clinical practice. Furthermore, use of correlation coefficients to evaluate different methods of measuring GFR is inappropriate.