A Comparison of Prediction Equations for Estimating Glomerular Filtration Rate in Adults without Kidney Disease

Abstract

Abstract. The ability of the Modification of Renal Disease (MDRD) equation to predict GFR when compared with multiple other prediction equations in healthy subjects without known kidney disease was analyzed. Between May 1995 and December 2001, a total of 117 healthy individuals underwent 125 I-iothalamate or 99m Tc-diethylenetriamine-pentaacetic acid (DTPA) renal studies as part of a routine kidney donor evaluation at either Brigham and Women's Hospital or Boston Children's Hospital. On chart review, 100 individuals had sufficient data for analysis. The MDRD 1, MDRD 2 (simplified MDRD equation), Cockcroft-Gault (CG), Cockcroft-Gault corrected for GFR (CG-GFR), and other equations were tested. The median absolute difference in ml/min per 1.73 m 2 between calculated and measured GFR was 28.7 for MDRD 1, 18.5 for MDRD 2, 33.1 for CG, and 28.6 for CG-GFR in the 125 Iiothalamate group and was 31.1 for MDRD 1, 38.2 for MDRD 2, 22.0 for CG, and 31.1 for CG-GFR in the 99m Tc-DTPA group. Bias was Ϫ0.5, Ϫ3.3, 25.6, and 5.0 for MDRD 1, MDRD 2, CG, and CG-GFR, respectively, in subjects who received 125 I-iothalamate and Ϫ33.2, Ϫ36.5, 6.0, and Ϫ15.0 for MDRD 1, MDRD 2, CG, and CG-GFR, respectively, in those who received 99m Tc-DTPA studies. Precision testing, as measured by linear regression, yielded R 2 values of 0.04 for CG, 0.05 for CG-GFR, 0.15 for MDRD 1, and 0.14 for MDRD in those who underwent 125 I-iothalamate studies and 0.18 for CG, 0.21 for CG-GFR, 0.40 for MDRD 1, and 0.38 for MDRD 2 for those who underwent 99m Tc-DTPA studies. The MDRD equations were more accurate within 30 and 50% of the measured GFR compared with the CG and CG-GFR equations. When compared with the CG equation, the MDRD equations are more precise and more accurate for predicting GFR in healthy adults. The MDRD equations, however, consistently underestimate GFR, whereas the CG equations consistently overestimate measured GFR in people with normal renal function. In potential kidney donors, prediction equations may not be sufficient for estimating GFR; radioisotope studies may be needed for a better assessment of GFR. Further studies are needed to derive and assess GFR prediction equations in people with normal or mildly impaired renal function.

Lower haemoglobin level and subsequent decline in kidney function in type 2 diabetic adults without clinical albuminuria

“…The generalisability of these findings will need to be confirmed in future studies. The original and refitted MDRD equations may have limitations for estimating GFR in individuals with normal or near-normal kidney function [36,37]. Our preliminary observations suggest that estimated GFR using the refitted MDRD equation in Japanese diabetic patients is well correlated with values obtained using a quadratic GFR equation, which is reported to perform better than the MDRD equation in healthy individuals [21].…”

Monitoring renal function: measured and estimated glomerular filtration rates - a review

“…Both MDRD and C-G equations have been reported to be less accurate in populations without chronic kidney disease, such as young patients with type 1 diabetes without microalbuminuria and potential kidney donors (27,28). Some studies report the higher accuracy of the MDRD equation (26) while others report similar performance of MDRD and C-G equations (29).…”

Detection of Chronic Kidney Disease in Patients With or at Increased Risk of Cardiovascular Disease

“…17,18 It is important to note that GFR values derived from either form of the MDRD study equation can be automatically generated and placed in laboratory reports of any patient, because the age and gender information needed to make the calculation is generally included in the database on each patient. Because race is often not included in such databases, automatic reporting will include a comment asking the practitioner to multiply the calculated GFR by 1.21 for black patients.…”

“…19 Although both forms of the MDRD study equation have been validated by other research groups for patients with significant reduction in GFR, their accuracy is reduced in persons with normal or only slightly diminished renal function. As screening tools, they can overestimate the number of patients with CKD, 17,18,20 and their accuracy in patients with cardiovascular disease has not been well substantiated. However, a study of patients with left ventricular dysfunction showed that a GFR Ͻ60 mL/min per 1.73 m 2 , calculated by the MDRD study equation, remained an independent risk factor for death, 21 which suggests that the predictive value of the test was high.…”