MDRD equation estimates of glomerular filtration rate in potential living kidney donors and renal transplant recipients with impaired graft function

Sir, 99mTc DTPA‐GFR was measured in 33 renal‐transplant recipients with chronic allograft nephropathy (group A, 10 female, age range 20–67 years, body mass index 18–32 kg/m2) and 34 potential living kidney donors with ‘normal’ renal function (group B, 18 female, age range 29–66 years, body mass index 19–32 kg/m2). Results were adjusted for body surface area (using the DuBois equation) and for initial disequilibrium of the isotope within its volume of distribution [2,3]. MDRD‐GFR was calculated for each individual using data collected at the time of the isotope study, and compared with the corresponding 99mTc DTPA‐GFR measurement. In group A, additional GFR estimates were obtained using the Nankivell ‘B’ [4] and Walser [5] formulae (derived from renal transplant recipients and ‘low clearance’ chronic renal failure patients, respectively). Bias plots were constructed using Microsoft Excel ‘Analyse‐it’ software [6]. The median values of the corrected 99mTc DTPA‐GFR measurements in A and B were 28 (range 15–53) ml/min/1.73 m2 and 96 (range 65–126) ml/min/1.73 m2. In A, the Levey formula gave the lowest bias (mean difference between 99mTc DTPA‐GFR and MDRD‐GFR) and scatter (median absolute difference between 99mTc DTPA‐GFR and MDRD‐GFR) of the three prediction equations (Figure 1 and Table 1). One patient (AG) was clearly an outlier, with an MDRD‐GFR of only 17 ml/min/1.73 m2, compared to 35 ml/min/1.73 m2 measured by 99mTc DTPA. The mean of his measured urea and creatinine clearances (from a 24‐h urine collection) was 34 ml/min/1.73 m2, suggesting that the MDRD‐GFR estimate was inaccurate. The patient was a body‐builder with an unusually large muscle mass. Cockroft and Gault (C&G) estimates showed a large positive bias, although direct comparison of values is spurious. The C&G formula calculates creatinine clearance rather than GFR. Secondly, the units of C&G estimates are ml/min rather than ml/min/1.73 m2, and the most appropriate method of correcting for body surface area is unclear as a weight term is already included in the basic formula. In group B, MDRD‐GFR estimates tended to be lower than 99mTc DTPA‐GFR (Figure 1), with a bias of 16.6 ml/min/1.73 m2 (15.7% of the 99mTc DTPA‐GFR measurement, 95% confidence interval 11.8–20.8 ml/min/1.73 m2) and a scatter of 17.9 ml/min/1.73 m2 (17.1% of the 99mTc DTPA‐GFR measurement). A difference between GFR values of more than 15 ml/min/1.73 m2 was seen only in patients with a 99mTc DTPA‐GFR measurement that was above 90 ml/min/1.73 m2. The Levey formula has the advantages of relative simplicity, ease of reporting (anthropometric data are not required), and low cost. Our study suggests that its use in renal transplant recipients with chronic allograft dysfunction is valid unless there is an unusual body composition. However, it tends to underestimate GFR in patients with normal or mildly impaired renal function as measured by a reference isotopic method (99mTc DTPA‐GFR). Similar findings have recently been reported by Bertolatus and Goddard [7]. Refinements of the Levey formula to improve its reliability in the higher GFR range may be determined by further studies involving larger numbers of patients. Bias plot for MDRD‐GFR vs99mTc DTPA‐GFR (both patient groups). Bias and scatter values for various prediction equations (group A) A, age; G, gender; R, race; Cr, serum creatinine; U, serum urea; Alb, serum albumin; Wt, weight; Ht, height. Bias and scatter values for various prediction equations (group A) A, age; G, gender; R, race; Cr, serum creatinine; U, serum urea; Alb, serum albumin; Wt, weight; Ht, height. JS is supported by the Yorkshire Kidney Research Fund.