Search tips
Search criteria 


Logo of clinbiorevLink to Publisher's site
Clin Biochem Rev. 2003 August; 24(3): 95–98.
PMCID: PMC1853341

The National Kidney Foundation Guideline on Estimation of the Glomerular Filtration Rate


The Kidney Disease Outcomes Quality Initiative (K/DOQI) is a set of evidence-based clinical practice guidelines from the National Kidney Foundation in the USA. All K/DOQI guidelines are published in the American Journal of Kidney Diseases and can also be accessed without restriction on the internet at: Our aim in this article is to critically review one section of the recently released guideline "K/DOQI Clinical Practice Guidelines for Chronic Kidney Disease: Evaluation, Classification, and Stratification". Part 5 of this document, "Evaluation of laboratory measurements for clinical assessment of kidney disease" includes Guideline 4: "Estimation of GFR".1 The summary section of the guideline is reproduced in the box and is presented in the original with 16 pages of explanation and supporting material. In this review we will critically assess the various components of this guideline under headings based on the statements in the summary.

“Estimates of the Glomerular Filtration Rate are the best overall indices of the level of kidney function”

Glomerular filtration rate (GFR) is an assessment of the filtering capacity of the functioning nephrons in the kidneys. It is a sensitive marker to detect changes in the overall renal function. It is useful in the early detection of renal impairment and for monitoring renal function and also provides guidance in the dosing of drugs. The rate of change of GFR is a good predictor of the time of onset of the need for renal replacement therapy such as dialysis. The accepted gold standards for measuring GFR are based on determining the clearance of exogenous substances such as inulin or various radio-labelled markers. As these tests are not routinely available, and other candidate markers such as cystatin C are not yet fully evaluated, available methods are based on the measurement of creatinine in serum. This guideline expands on various approaches to routine assessment of the GFR using creatinine measurements.

“Serum creatinine concentration alone should not be used to assess the level of kidney function”

Serum creatinine measurement is the simplest and most widely used estimate of GFR. In an individual the serum creatinine will approximately double for each halving of the GFR. As described in the guidelines document, the use of serum creatinine to estimate the GFR is sub-optimal for a number of reasons. These include the tubular and extra-renal secretion of creatinine, the wide between-individual variation in creatinine production and the differences between various analytical methods for creatinine. Additionally, while increases in serum creatinine are quite specific for renal impairment, the test has low sensitivity, often requiring a 50% fall in GFR to cause an appreciable rise in serum creatinine, the so-called “creatinine – blind” region of renal impairment. While we accept these limitations, we still believe that serum creatinine measurement remains an important test because of the vast clinical experience associated with the test, the simplicity of monitoring changes in a patient over time and the limitations of calculated estimates of GFR, which depend on the serum creatinine measurement in the first place.

“Measurement of creatinine clearance using timed urine collections does not improve the estimate of GFR over that provided by prediction equations”

A traditional method for estimating the GFR is to calculate the creatinine clearance from serum creatinine and a 24-hour urine collection. As indicated in the guidelines document, creatinine clearance is not an identical parameter to the GFR as tubular secretion and extra-renal creatinine loss lead to creatinine clearance exceeding the GFR. Historically the commonly used alkaline picrate assays have suffered interference with so-called “non-creatinine chromogens” in serum which are not found in urine, which has by chance cancelled out this over-estimation compared to the GFR. Paradoxically, more accurate creatinine assays, e.g. enzymatic and modern rate-blanked methods, may have accentuated this difference. Of greater importance is the difficulty in collecting a reliable timed urine collection. Even under study conditions 24-hour urine collections are notoriously unreliable and prediction equations lead to closer approximations of the GFR than measured creatinine clearances. The guidelines however support the use of measured creatinine clearance in certain circumstances where prediction equations may be inadequate; for example, significant abnormalities in muscle mass or protein intake. We support this section of the guidelines.

“The level of GFR should be estimated from prediction equations that take into account the serum creatinine concentration and some or all of the following variables: age, gender, race, and body size. The Modification of Diet in Renal Disease (MDRD) Study and Cockcroft-Gault Equations provide useful estimates”

Before assessing the possible use of prediction equations it is necessary to consider variations in the reporting of creatinine clearance and the GFR. The simplest measure is to express the result in mL/s (SI units) or mL/min (non-SI units). A refinement is to express the results per 1.73 m2 body surface area (BSA) on the basis that it is normal for a larger person, as determined by BSA, to have a higher creatinine clearance or GFR than a smaller person. BSA can be calculated from height and weight from available equations, tables or nomograms.2

The Cockcroft and Gault equation was first published in 1976,3 and has been the subject of multiple validations since that time with the guidelines indicating review in over 50 published articles. The equation estimates creatinine clearance without correction for BSA and is based on predicting the daily urine creatinine excretion given the age, weight and sex of the patient. It has generally been found to be more accurate than measured creatinine clearance and is well accepted in the clinical practice. As noted above, changes in creatinine assays since first publication may lead to the creatinine clearance being a poorer estimate of the GFR than was the case originally. The guidelines also recommend the use of one of the MDRD formulae for estimation of the GFR (expressed per 1.73 m2 BSA).4 There are four versions of the formula which all include the variables of age, sex, serum creatinine and race (African American, non-African American) with the inclusion of serum albumin, serum urea and urinary urea in different versions. As the predictive value of the equations is similar, we will restrict further comment to the “Abbreviated MDRD Study Equation” which is based on only the first four parameters listed above.

The data presented in the guidelines support the MDRD formula as a superior predictor of the GFR when compared to the Cockcroft and Gault formula; however, we do not believe that the MDRD formula has yet been subject to sufficient scrutiny to be able to recommend its routine use. The MDRD formula was developed only using patients with renal impairment, and subsequent to its development has been shown to have a significant bias in people with normal renal function and to perform less well than, or similar to, the Cockcroft and Gault equation in different patient populations.57 Of particular concern is the use of the MDRD formula in elderly patients as this formula differs markedly from the Cockcroft and Gault equation with increasing age. Between the ages of 45 and 70 the Cockcroft and Gault formula predicts a fall of about 27% in creatinine clearance if there is no change in serum creatinine or body weight. In contrast, the MDRD equation predicts a fall of only 8% in most cases over the same period. As occult renal impairment in the elderly is common, further studies are required for this population. The issues we have raised were noted in the original paper outlining the MDRD.4 Our conclusion is different from that of the guidelines, namely that we cannot recommend this formula for routine use at this time.

Due to lack of expertise we will not comment on the proposed prediction equations for use in children.

“Autoanalyser manufacturers and clinical laboratories should calibrate serum creatinine assays using an international standard”

If estimation of GFR is to be made using a serum creatinine measurement, either as a raw value or as part of a prediction equation, it is important that all creatinine assays give equivalent results5. As long as non-specific methods such as the alkaline picrate method are in use, this is not as simple as using the same calibrator. Each method must be shown to give acceptable agreement with a highly specific method, when patient samples are tested in both methods. Currently over 70% of laboratories enrolled in the external quality assurance programme for creatinine administered by the Royal College of Pathologists of Australasia Quality Assurance Programs Pty Ltd use an alkaline picrate method. The between-instrument group variation is approximately 10% (coefficient of variation) at low concentrations (0.05 mmol/L) and about 5% at the high end (0.40 mmol/L). While this may not reflect differences for actual patient samples, it suggests that actual differences at high serum creatinine concentrations may not be that large. We applaud the work of the various diagnostic companies in working towards this standardisation but at this time it is not possible to easily determine from a published description what effect non-creatinine chromogens may have on this measurement. Either highly-specific enzymatic assays that are equally inexpensive are required or a program is needed to ensure equivalence of patient results in different picrate assays.

“Clinical laboratories should report an estimate of GFR using a prediction equation, in addition to reporting the serum creatinine measurement”

The abbreviated version of the MDRD equation uses only serum creatinine, age, sex and race as input parameters. In a population, such as in Australia, which is largely devoid of people of African origin, automatic calculation would be possible using data that are already held in pathology computer systems. This is an advantage over the Cockcroft and Gault equation where the patient’s weight must also be obtained. However, for the reasons stated above, we do not believe that the MDRD equation is yet ready for routine use.

At this time we also suggest that reporting of an estimate of GFR based on a prediction equation should be restricted to those circumstances where it is specifically requested by the treating physician, as the limitations of any such estimate may not be well understood by all requesters. While we do believe that the use of a prediction equation is likely to be of benefit, a single algorithm needs to be accepted universally or the situation would be similar to having multiple assays with different standardisations, with confusion arising when different results are obtained from different laboratories. It should also be recognised that the limitations of serum creatinine measurements with regard to standardisation and poor sensitivity to early renal disease also apply to prediction equations based on the measurement of serum creatinine.

Finally, if a laboratory, when requested, wishes to provide an estimate of GFR based on a prediction equation, we suggest the following issues be addressed:

  1. The choice of equation (s)
  2. The source of additional information (e.g. weight)
  3. Reference intervals for the equation and consideration of age-related reference intervals
  4. Additional information to include on the report
  5. Supporting educational material


  • ♦ Estimates of GFR are the best overall indices of the level of kidney function.
  • ♦ The level of GFR should be estimated from prediction equations that take into account the serum creatinine concentration and some or all of the following variables: age, gender, race, and body size. The following equations provide useful estimates of GFR:
    • In adults, the MDRD Study and Cockcroft - Gault equations
    • In children, the Schwartz and Counahan - Barratt equations.
  • ♦ The serum creatinine concentration alone should not be used to assess the level of kidney function.
  • ♦ Clinical laboratories should report an estimate of GFR using a prediction equation, in addition to reporting the serum creatinine measurement.
  • ♦ Autoanalyser manufacturers and clinical laboratories should calibrate serum creatinine assays using an international standard.
  • ♦ Measurement of creatinine clearance using timed (for example, 24-hour) urine collections does not improve the estimate of GFR over that provided by prediction equations. A 24-hour urine sample provides useful information for:
    • Estimation of GFR in individuals with exceptional dietary intake (vegetarian diet, creatine supplements) or muscle mass (amputation, malnutrition, muscle wasting)
    • Assessment of diet and nutritional status
    • Need to start dialysis.

Equations referred to in the text expressed in SI units.

Cockcroft-Gault Equation

equation image

Abbreviated MDRD Study equation (demographic variables only):

equation image

Note 1: Serum creatinine (mmol/L) x 11.3 = serum creatinine (mg/dL)

Note 2: Multiply results by 60 to give results in mL/min.


1. National Kidney Foundation. K/DOQI Clinical Practice Guidelines for Chronic Kidney Disease: Part 5. Evaluation of Laboratory Measurements for Clinical Assessment of Kidney Disease. Guideline 4. Estimation of GFR. Am J Kidney Dis. 2002;39(S1):S76–S110.
2. Geigy Scientific Tables. Volume 3, Ed Lentner C. Ciba-Geigy, Switzerland. 1984.
3. Cockcroft DW, Gault MH. Prediction of Creatinine Clearance from Serum Creatinine. Nephron. 1976;16:31–41. [PubMed]
4. Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D. A More Accurate Method to Estimate Glomerular Filtration Rate from Serum Creatinine: A New Prediction Equation. Ann Intern Med. 1999;130:461–470. [PubMed]
5. Bostom AG, Kronenberg F, Ritz E. Predictive Performance of Renal Function Equations for Patients with Chronic Kidney Disease and Normal Serum Creatinine Levels. J Am Soc Nephrol. 2002;13:2140–2144. [PubMed]
6. Wuyts B, Bernard D, Van Den Noortgate N, et al. Reevaluation of Formulas for Predicting Creatinine Clearance in Adults and Children, Using Compensated Creatinine Methods. Clin Chem. 2003;49:1011–1014. [PubMed]
7. Harmoinen A, Lehtimaki T, Korpela M, Turjanmaa V, Saha M. Diagnostic Accuracies of Plasma Creatinine, Cystatin C, and Glomerular Filtration Rate Calculated by the Cockroft-Gault and Levey (MDRD) Formulas. Clin Chem. 2003;49:1223–1225. [PubMed]

Articles from The Clinical Biochemist Reviews are provided here courtesy of The Australian Association of Clinical Biochemists