Using Standardized Serum Creatinine Values in the Modification of Diet in Renal Disease Study Equation for Estimating Glomerular Filtration Rate

The recent article about estimated glomerular filtration rate (GFR) published by Levey et al is remarkable (1). The authors suggest that the new equation may still have some bias and, especially, less precision in patients with GFR over 60 ml/min/1.73 m². This is even more important if GFR is over 90 ml/min/1.73 m². In the method section, the authors don't give the analytical coefficient of variation (CVa) of their assays (Beckman and enzymatic). This data is of importance especially for low or normal creatinine values. Indeed, the concept of critical difference (CD) is familiar to clinical biologists but should perhaps be reminded to internists. The CD of a biological variable includes the CVa and the intra -individual coefficient of variation (CVi). It is defined as the smallest change in a biological result which is not due to chance (2). The CD is calculated as followed: 1.414 x 1.96 x (CVa² + CVi²)0.5. The CVi of serum creatinine is 4% (3). The CVa of serum creatinine varies belong assays used and laboratories. A creatinine CVa as low as 2 % is rare but conceivable (4). With these CV values, the lowest CD for creatinine is 12%. A value of 0.8 mg/dl is thus not different from 0.704 and 0.896 mg/dl. However, as we have illustrated (5), these differences are not negligible for GFR estimation if creatinine and GFR are normal because small creatinine changes induce large GFR variations in this range. If we take the example of a white, 60 years old man, a creatinine of 0.8 mg/dl gives a result of 98.6 ml/min/1.73 m² with the MDRD equation. If creatinine values of 0.704 and 0.896 mg/dl are introduced, the results of the MDRD equations will be 114.3 and 86.5 ml/min/1.73m², respectively. The low precision of the MDRD equation, when GFR is normal, is thus also linked to the precision of the creatinine assay and to the biological variation of creatinine. This assertion is true for all creatinine-based equations. We think that an improvement of the precision of creatinine- based equation may be illusive in a non renal population. It is important for clinicians to keep this fact in mind when they analyze an estimated GFR and even more when they longitudinally follow a serial of estimated GFR in a patient with GFR over 60 ml/min/1.73m². It is perhaps more cautious to still give MDRD results as "over 60 and 90 ml/min/1.73m²" without giving precise, absolute values of GFR.

References

(1) Levey AS, Coresh J, Greene T, Stevens LA, Zhang YL, Hendriksen S et al. Using standardized serum creatinine values in the modification of diet in renal disease study equation for estimating glomerular filtration rate. Ann Intern Med 2006;145:247-54.

(2) Costongs GM, Janson PC, Bas BM, Hermans J, van Wersch JW, Brombacher PJ. Short-term and long-term intra-individual variations and critical differences of clinical chemical laboratory parameters. J Clin Chem Clin Biochem 1985;23:7-16.

(3) Ricos C, Alvarez V, Cava F, Garcia-Lario JV, Hernandez A, Jimenez CV et al. Current databases on biological variation: pros, cons and progress. Scand J Clin Lab Invest 1999;59:491-500.

(4) Froissart M, Rossert J, Jacquot C, Paillard M, Houillier P. Predictive performance of the modification of diet in renal disease and Cockcroft-Gault equations for estimating renal function. J Am Soc Nephrol 2005;16:763-73.

(5) Delanaye P, Cavalier E, Krzesinski JM, Chapelle JP. Why the MDRD equation should not be used in patients with normal renal function (and normal creatinine values)? Clin Nephrol 2006;66:147-8.

Conflict of Interest:

None declared

To the Editor,

Dr. Delanaye and colleagues question whether the precision of the creatinine assay within the reference range is sufficient for accurate GFR estimation. The Beckman Synchron CX3 assay used during the Modification of Diet in Renal Disease (MDRD) Study had an analytical coefficient of variation (CV) of 4.3% and 1.5% at creatinine values of 1.0 and 5.4 mg/dL, respectively (1). The Roche enzymatic assay used to calibrate the MDRD Study laboratory to standardized creatinine had an analytical CV of 2.0% and 1.8% at creatinine values of 0.89 and 5.86 mg/dL, respectively, in 2004 (n = 194) and 1.6% and 1.1% at creatinine values of 1.00 and 3.84 and gm/dL, respectively, in 2005 (n = 409). Thus, the analytical CV for the Roche enzymatic is as low or lower than that stated by Dr. Delanaye.

We agree with Dr. Delanaye that the effect of imprecision in the serum creatinine assay and biological variation in GFR estimates is greater at lower values for serum creatinine (equivalent to higher values for estimated GFR), and that is one of several important reasons for lesser accuracy of higher GFR estimates (2). For these reasons, current recommendations are to report estimated GFR as a numeric value only when it is less than 60 ml/min/1.73 m2 and to report ">60 ml/min/1.73 m2" for higher values. We believe this is sufficient for most clinical circumstances requiring the clinical assessment of kidney function. New filtration markers, such as cystatin C, and improvement in estimation equations may be required for more accurate GFR estimation at higher values. Until then, if more accurate assessment of kidney function is required in patients with estimated GFR >60 ml/min/1.73 m2, it is necessary to measure the clearance of an exogenous filtration marker or creatinine.

References:

1. Coresh J, Astor BC, McQuillen G, et al. Calibration and random variation of the serum creatinine assay as critical elements of using equations to estimate glomerular filtration rate. Am J Kidney Dis 2002;39:920-9.

2. Stevens LA, Coresh J, Levey AS. Assessing kidney function "“ measured and estimated glomerular filtration rate. N Engl J Med 2006; 354:2473-83.

Conflict of Interest:

None declared