Estimating glomerular filtration rate

Glomerular filtration rate (GFR) is a parameter of prime clinical significance because it defines kidney function. All those with reduced kidney function, no matter what the cause, have reduced GFR, and GFR correlates well with disease severity. The most accurate assessment of GFR is thus valuable not only for early diagnosis of chronic kidney disease but also for monitoring its progress. 

Furthermore, since safe and effective use of many medicinal drugs depends on knowledge of patient kidney function, assessment of GFR has an important role for delivery of drug therapies.  

Unfortunately it is simply not possible to determine GFR directly. Indirect methods of measurement, based on renal clearance from blood of exogenous (administered) substances come closest to providing a ‘true’ value for GFR, but these are too cumbersome and expensive for routine use.  Estimation (calculation) of GFR from plasma creatinine concentration is the pragmatic solution. 

This is cheap, technically simple and therefore easily applied for routine assessment of kidney function.  Considerable clinical research over the past 10-15 years has been directed at devising and validating equations for most accurate estimation of GFR from plasma creatinine concentration, and more recently, plasma cystatin concentration.  

A summary of these studies is contained in a recently published review article authored by those that have conducted much of this research. The article provides the scientific rationale that underpins current recommendation for routine estimation of GFR. 

By way of introduction the authors first briefly describe the underlying physiology that allows an understanding of what is meant by glomerular filtration rate (GFR).  There follows more detailed discussion of the ‘gold standard’ methods used to indirectly measure GFR. 

The greater part of the article however is devoted to comparison of the different methods (equations) used to estimate GFR from plasma creatinine concentration that have been developed and validated by study over the past decade or so. 

The two principle equations discussed are: the MDRD (Modification of Diet in Renal Disease) study equation developed in 1999 and later modified (simplified) in 2006, and the currently recommended CKD-EPI (Chronic Kidney Disease – Epidemiological Collaboration) equation developed in 2009. 

Discussion however is not limited to these two equations. There are other more recently developed equations that remain to be fully validated.  As the authors make clear this is an evolving area of research that may yield further improvements in estimating GFR.  An important recent development is the use of plasma cystatin concentration, either alone or in conjunction with plasma creatinine concentration, to estimate GFR. 

This article is a useful, highly authoritative overview of GFR, its indirect ‘gold standard’ measurement, and its estimation from plasma concentration of endogenous substances (creatinine and/or cystatin).   

A related clinical study is also recently published. According to the investigators responsible for this study the now recommended CKD-EPI equation for estimating GFR has not been fully validated for use among patients with heart failure.  They wanted first to confirm that the CKD-EPI equation is a more accurate estimation GFR than the previously recommended MDRD equation for this patient group. 

A secondary aim was to test the accuracy of other equations developed since the introduction of CKD-EPI in 2009.  The study population comprised 120 patients with moderately severe heart failure (76% were NYHA class II or III).  All study subjects were submitted for ‘gold standard’ measurement of GFR by a standard research technique that involves continuous infusion of a radiolabelled tracer  [¹²⁵I] iothalamate.   

Blood was sampled for measurement of plasma creatinine concentration and plasma cystatin concentration to allow estimation of GFR by calculation.  Bias of estimated GFR (e-GFR) compared with reference measured GFR (m-GFR) for each patient allowed comparison of different equations to accurately assess GFR. 

In summary the results demonstrated a mean negative bias for eGFR for all equations. However the negative bias was greater for the MDRD equation than for the CKD-EPI equation allowing the authors to conclude that the CKD-EPI equation more accurately estimates measured GFR than does the MDRD equation. 

They suggest, in line with current recommendations, that for heart failure patients the CKD-EPI equation should be used to estimate GFR from plasma creatinine concentration.  The secondary results of this study suggest that it is preferable in terms of accuracy and precision to use plasma cystatin concentration rather than plasma creatinine concentration for estimation of GFR.