The systematic staging of chronic kidney disease (CKD) by glomerular filtration measurement and proteinuria has allowed the development of rational and appropriate management plans.
One of the barriers to early detection of CKD is the lack of a precise, reliable and consistent measure of kidney function.
The most common measure of kidney function is currently serum creatinine concentration. It varies with age, sex, muscle mass and diet, and interlaboratory variation between measurements is as high as 20%.
The reference interval for serum creatinine concentration includes up to 25% of people (particularly thin, elderly women) who have an estimated glomerular filtration rate (eGFR) that is significantly reduced (< 60 mL/min/1.73m2).
The recent publication of a validated formula (MDRD) to estimate GFR from age, sex, race and serum creatinine concentration, without any requirement for measures of body mass, allows pathology laboratories to “automatically” generate eGFR from data already acquired.
Automatic laboratory reporting of eGFR calculated from serum creatinine measurements would help to identify asymptomatic kidney dysfunction at an earlier stage.
eGFR correlates well with complications of CKD and an increased risk of adverse outcomes such as cardiovascular morbidity and mortality.
We recommend that pathology laboratories automatically report eGFR each time a serum creatinine test is ordered in adults.
As the accuracy of eGFR is suboptimal in patients with normal or near-normal renal function, we recommend that calculated eGFRs above 60 mL/min/1.73m2 be reported by laboratories as “> 60 mL/min/1.73m2”, rather than as a precise figure.
Chronic kidney disease (CKD) is a morbid condition that is common and may be preventable. In the general Australian community, there is evidence of at least one indicator of CKD (proteinuria or reduced kidney function) in about 16% of adults aged over 25 years.1 CKD progresses to end-stage kidney failure at a rate that requires about 1900 Australians each year to commence renal replacement treatment — either dialysis or transplantation.2 Furthermore, in people with CKD there is actually a 20-fold greater chance of death (mainly from cardiovascular disease) than of starting renal replacement therapy.3 However, CKD is frequently asymptomatic, and, although in some instances it can be detected by the presence of proteinuria, many afflicted people have significant reduction of kidney function without overt urinary abnormalities. Therefore, a reliable means of readily assessing the early stages of reduced kidney function is a priority.
The diagnosis and management of CKD has been facilitated in recent years by the Kidney Disease Outcomes Quality Initiative (K/DOQI) clinical practice guidelines of the US National Kidney Foundation. The K/DOQI guidelines4 advise that CKD can be defined and appropriately managed by a staging approach that relies on estimating the extent of kidney damage based on the degree of proteinuria and impaired kidney function, assessed as a reduction in the glomerular filtration rate (GFR).
The most common measure used to assess overall kidney function is the serum creatinine concentration. Interpretation of this index is complicated, as it is inversely proportional to the GFR and varies between individuals based on differences in age, sex and muscle mass. Using serum creatinine concentrations to determine an absolute level of kidney function, including distinguishing normal from abnormal function in the individual patient, is inherently difficult. The broader use of serum creatinine concentration as a tool to increase the detection of asymptomatic CKD is therefore problematical.
GFR is widely accepted as the best measure of kidney function, yet in clinical practice beyond nephrology it is infrequently utilised. The main impediment to its regular clinical use has been the perception that it was necessary to estimate GFR by performing a creatinine clearance test that is dependent on a timed urine collection (usually 24 hours). More recently, calculating estimated GFR (eGFR) using an empirical mathematical formula has been encouraged through the provision of handheld or desktop semi-automated calculators designed for this purpose. The Cockcroft–Gault equation is the most frequently used eGFR formula in Australia, where a general population study has shown that 11.3% of adults have a Cockcroft–Gault eGFR below 60 mL/min/1.73m2 (the threshold value for CKD).1 There are now at least 46 different equations for estimating GFR, but most (including the Cockcroft–Gault equation) require additional information, such as a measure of body surface area (based on height and/or weight measurements), leading to additional complexities that limit the wider use of this approach. There are recognised difficulties associated with collecting body size measurements (eg, errors in measurement and transcription), and, as pathology laboratories cannot ensure the quality of these variables, they are often hesitant to report eGFR using these formulas.
The possibility that pathology laboratories might routinely report an eGFR derived from the serum creatinine concentration has recently become feasible with the development of a formula whose only variables are age, sex, race and serum creatinine concentration. Most importantly, it does not require body surface-area measurements. This formula, the “abbreviated MDRD equation” (named after the US Modification of Diet in Renal Disease Study5), has been validated in many clinical situations. However, the adjustment for race in the MDRD equation is limited to “African-American”, which may affect the formula’s applicability to the Australasian population. In particular, the MDRD formula has not yet been validated in Aboriginal and Torres Strait Islander populations. Moreover, although there is evidence that automated laboratory reporting leads to greatly enhanced detection of CKD by health professionals,6 there is no high-level clinical evidence that this in turn leads to improved clinical outcomes. A reanalysis of the AusDiab study data recently showed that 7.5% of the Australian adult population had an eGFR (based on the abbreviated MDRD formula) of < 60 mL/min/1.73m2 (Associate Professor Steve Chadban, Nephrologist and Director of Kidney Transplantation, Royal Prince Alfred Hospital and University of Sydney, personal communication). The K/DOQI and UK Renal Association guidelines recommend automatic reporting of eGFR from serum creatinine measurements, and advocate using the MDRD formula for this purpose.4,7,8
Because of these developments, an Australasian Creatinine Consensus Working Group (see end of Position Statement), met in November 2004 to develop recommendations on the desirability of automatic reporting of an eGFR from each serum creatinine measurement performed in pathology laboratories. The opportunity was taken to address issues relating to inconsistencies in the measurement and reporting of serum creatinine concentration that might affect its use for calculating eGFR.
The Working Group meeting was sponsored by the Australasian Association of Clinical Biochemists, the Australian and New Zealand Society of Nephrology, Kidney Health Australia and the Royal College of Pathologists of Australasia, and was attended by 21 representatives of these organisations. The following recommendations emanated from the meeting. All resolutions were endorsed unanimously, with the exception of Recommendation 6, where there was one abstention. The Australian Diabetes Society has also endorsed the recommendations.
The estimation of GFR from serum creatinine levels contains a degree of imprecision. Variations of eGFR from the direct measurement of GFR arise from a number of factors, including variability in serum creatinine measurements and the imperfect nature of the estimation equation. For example, the US National Kidney Disease Education Program (NKDEP) has estimated that variability of ±15% may be attributed to the MDRD equation itself. The NKDEP has set a goal of overall accuracy of ±30% for the estimation of GFR, and thus, by allowing for ±15% variability inherent in the estimation equation, recommends that the total error in serum creatinine measurement should be less than ±15%.
In accepting this quality specification, an important issue is assigning a target against which to compare assay performance. The MDRD formula was derived using a serum creatinine assay from Beckman Coulter run on a CX3 analyser, and assays that produce results within ±15% of this type of assay would thus fulfil the accuracy criterion. The CX3 creatinine assay is known to have a small positive bias compared with the recognised international reference method for creatinine measurement (isotope dilution mass spectrometry [IDMS]). As most current routine serum creatinine assays produce results equal to or higher than the IDMS reference method (for results within the reference interval 40–110 μmol/L), assays producing results within ±15% of methods aligned with IDMS will show a total error compared with the CX3 method of less than ±15%, and therefore also satisfy the criterion. In accepting the accuracy criterion, it is hoped that manufacturers will be encouraged to develop creatinine assays that are more closely aligned with the IDMS reference method to allow improved method standardisation in the future.
A review of the status of serum creatinine measurement in Australia and New Zealand using data from national9 and international10 sources and local sample-sharing studies (Dr Graham Jones, Staff Specialist in Chemical Pathology, St Vincent’s Hospital, NSW, personal communication) was presented to the Working Group. These studies indicate that, at serum creatinine concentrations of about 100 μmol/L, creatinine assay results supplied by the major manufacturers generally meet the total error requirement of ±15% deviation from the Beckman Coulter method. At higher creatinine concentrations, the assays meet this criterion without difficulty, but at lower concentrations, there is some variation in achieving this standard. Professional bodies must develop methods for confirming that assays meet the criterion, and laboratories must ensure that their creatinine assays conform to these requirements.
The formal application of the International System of Units (SI) recommends using whole numbers rather than numbers frequently less than one. By this standard, the SI units for serum creatinine concentration should be μmol/L. A recent international survey of pathology reporting10 indicated that all countries using SI units (except Australia and New Zealand) reported creatinine levels in μmol/L. In Australia and New Zealand, laboratories are currently divided about equally between using mmol/L or μmol/L as the unit. Conversion of all laboratories to μmol/L as the reporting unit is considered to be a useful step in minimising confusion in clinical interpretation.
The estimation of GFR from serum creatinine level should be performed routinely by laboratories using data rounded to the nearest 1 μmol/L. Measurement of serum creatinine level to the nearest 1 μmol/L allows optimal quality control of such assays, especially for results within the reference interval. Adopting the principle that calculations should be made on raw (unrounded) data where possible, it is recommended that the calculation of eGFR should be performed using serum creatinine measurements reported to this level of precision.
Serum creatinine concentrations associated with a calculated GFR at the important 60 mL/min/1.73m2 decision point are between 80 μmol/L and 120 μmol/L, depending on the age and sex of the patient. To provide information on the assay performance in this range, an internal quality control sample should be run with a creatinine concentration near 100 μmol/L. These data, together with information on biological variation, should be available to clinicians when requested.
The MDRD formula yields an eGFR normalised to 1.73m2 body surface area. Adjusting for body surface area is necessary when comparing a patient’s eGFR with normal values or when determining the stage of CKD. However, an uncorrected eGFR may be preferred for clinical use in some situations, such as drug dosing. To revert to an uncorrected eGFR, the result from the MDRD-derived eGFR should be multiplied by the individual’s body surface area, using the following formula:
BSA = W0.425 × H0.725 × 0.007184/1.73
where BSA = body surface area (m2), W = weight (kg), H = height (cm)
Uncorrected eGFR = GFR estimate (mL/min/1.73 m2) × BSA
In practice, adjusted GFR estimates are adequate except in patients with a body size that is very different from the average.
It should be noted that recommendations for drug-dosing adjustments for patients with reduced kidney function are currently based on the Cockcroft–Gault formula, and this result may differ significantly from the MDRD-derived eGFR.11
In recommending the unit of mL/min, it is recognised that this deviates from the SI unit of time (the second). However, as the overwhelming majority of clinical interpretive information (both Australian and worldwide) uses the unit mL/min, this unit is accepted in order to promote standardisation and remove a source of potential confusion.
The abbreviated MDRD equation13
eGFR = 186 × ([SCR/88.4]–1.154) × (age) –0.203 × (0.742 if female) × (1.210 if African-American)
where eGFR = estimated glomerular filtration rate (mL/min/1.73m2), SCR = serum creatinine concentration (μmol/L), and age is expressed in years.
An automated calculator for MDRD-based eGFR can be found at <http://www.kidney.org.au>.
MDRD = Modification of Diet in Renal Disease.5
The primary reasons for the recommendation are as follows:
Kidney function is poorly deduced from the serum creatinine concentration alone;
Serum creatinine concentration is an imprecise measurement of kidney function, as it varies according to age, sex, muscle mass and diet;
The use of serum creatinine concentration alone to assess kidney function results in undiagnosed cases of CKD, as up to 25% of results may lie within the accepted reference interval yet translate to an abnormal eGFR (< 60 mL/min/1.73m2);12
The abbreviated MDRD formula (Box) was recommended on the basis of the following factors:
Direct comparison of the MDRD equation with other equations such as the Cockcroft–Gault equation and to results from 24-hour urine collections have shown the MDRD equation to be superior for estimating GFR, particularly in the range GFR < 60 mL/min/1.73m2;4,15
There is no requirement for additional information (eg, measurements of body surface area) beyond that already collected by pathology laboratories.
As the MDRD formula has not been validated in children, its use should be restricted to people over 18 years of age.
The recommendation that a value of eGFR calculated to be above 60 mL/min/1.73m2 should not be reported as a precise figure is based on:
the less precise nature of the relationship between MDRD-derived eGFR and direct measurement of GFR (eg, based on inulin or isotopic methods) in the higher ranges;
the fact that the MDRD formula was initially derived from data acquired from people with CKD. Comparisons of this formula with direct measures of GFR (eg, methods using iothalamate) in healthy people indicate a significant underestimation of measured GFR, most marked in the normal GFR range;16
the greater interlaboratory variation in serum creatinine concentrations (related to calibration and choice of laboratory method) in the range of values that translate to an eGFR > 60 mL/min/1.73m2;
The age-related decline in GFR that has been described with inulin-based GFR measurements,21 and more recently with eGFR methods, appears to be about 8 mL/min per decade.12 No Australian data have been published in this area. Automatic reporting of eGFR from serum creatinine concentration will likely reveal that 25% of the Australian population aged over 70 years has an eGFR < 60 mL/min/1.73m2, as previously demonstrated by US data.12 It thus seems prudent to report a reference interval for people over 70 years to guide decision-making for older age groups. The mean eGFR for people aged 70 years and over in the United States has been calculated to be 75 mL/min/1.73m2.22 Further work is in progress to refine the recommendations for Australia and to decide whether a qualifying statement is needed to help interpret eGFRs automatically generated for older age groups.
Comprehensive education initiatives are required to help health practitioners understand the limitations of the eGFR. The information they will need includes:
an appropriate management pathway for those with an eGFR < 60mL/min/1.73m2 (including indications for nephrologist referral);
changes in eGFR with age;
decreased accuracy of eGFR above 60 mL/min/1.73m2;
implications of body surface area for consideration of drug doses;
decreased accuracy of eGFR in acute/unstable conditions;
precision of eGFR result; and
lack of applicability of eGFR to dialysis-dependent patients.
Specific clinical settings in which eGFR is not appropriate for use and GFR should be measured directly include:
populations in which the MDRD equation is not validated (eg, Asian people23) or in which validation studies have not been performed (eg, Aboriginal and Torres Strait Islander populations);
severe malnutrition or obesity;
extremes of body size and age;
exceptional dietary intake (eg, vegetarian diet or creatine supplements);
disease of skeletal muscle, paraplegia, etc; and
rapidly changing kidney function.
A concerted educational campaign is planned to coincide with the implementation of these recommendations. In addition, pre- and post-implementation audits will be undertaken to assess the impact of automatic eGFR reporting on awareness, detection and management of CKD in primary health care, as well as on nephrologist referrals. This information will also assist workforce and health resource planning.
Automatic reporting of eGFR on each occasion a serum creatinine test is requested will significantly increase the likelihood of early detection of CKD and allow appropriate management. In making the conservative recommendations outlined here, we recognise the limitations of existing knowledge, including the current imprecision of creatinine measurement and the imperfections of the MDRD formula, particularly as it applies to GFR in healthy people. The restrictions and qualifications recommended should allow the benefits of this approach to be realised without causing unnecessary concern and unneeded investigations. We recognise that this is an evolving area — creatinine measurements and GFR estimation formulas are likely to improve, and alternative methods for GFR measurement will be developed. However, we believe it is vital to begin a coordinated national effort to improve the identification of patients with renal impairment, providing a firm base on which to build future developments.
- 1. Chadban SJ, Briganti EM, Kerr PG, et al. Prevalence of kidney damage in Australian adults: the AusDiab kidney study. J Am Soc Nephrol 2003; 14 (7 Suppl 2): S131-S138.
- 2. ANZDATA Registry. Twenty-sixth report. Adelaide: Australia and New Zealand Dialysis and Transplant Registry, 2003. Available at: http://www.anzdata.org.au (accessed Jun 2005).
- 3. Keith DS, Nichols GA, Gullion CM, et al. Longitudinal follow-up and outcomes among a population with chronic kidney disease in a large managed care organization. Arch Intern Med 2004; 164: 659-663.
- 4. National Kidney Foundation. K/DOQI clinical practice guidelines for chronic kidney disease: evaluation, classification and stratification. Am J Kidney Dis 2002; 39 (2 Suppl 1): S1-S266. Available at: http://www.kidney.org/professionals/kdoqi/guidelines_ckd/toc.htm (accessed Jun 2005).
- 5. Levey AS, Bosch JP, Lewis JB, et al. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Modification of Diet in Renal Disease Study Group. Ann Intern Med 1999; 130: 461-470.
- 6. Akbari A. Detection of chronic kidney disease with laboratory reporting of estimated glomerular filtration rate and an educational program. Arch Intern Med 2004; 164: 1788-1792.
- 7. National Kidney Disease Education Program. Information for health professionals. GFR calculator. Available at: http://www.nkdep.nih.gov/professionals/gfr_calculators/index.htm (accessed Jun 2005).
- 8. Joint Specialty Committee on Renal Disease. Chronic kidney disease in adults: UK guidelines for identification, management and referral. June 2005. Available at: http://www.renal.org/CKDguide/full/UKCKDfull.pdf (accessed Jun 2005).
- 9. Chemical Pathology QAP Program. General serum chemistry and therapeutic drugs program. Cycle 66. End-of-cycle report. Sydney: Royal College of Pathologists of Australasia Quality Assurance Programs Pty Ltd, 2004.
- 10. European Commission Joint Research Centre. The International Measurement Evaluation Programme. Trace and minor constituents in human serum. Geel: Institute for Reference Materials and Measurements, 2003.
- 11. Tharp R. Renally excreted drug dosing, 2005. Available at: http://www.rxkinetics.com/renal.html (accessed Jun 2005).
- 12. Coresh J, Astor BC, Greene T, et al. Prevalence of chronic kidney disease and decreased kidney function in the adult US population: Third National Health and Nutrition Examination Survey. Am J Kidney Dis 2003; 41: 1-12.
- 13. Levey AS, Greene T, Kusek J, et al. Simplified equation to predict glomerular filtration rate from serum creatinine [abstract]. J Am Soc Nephrol 2000; 11: A828.
- 14. Hallan S, Asberg A, Lindberg M, Johnsen H. Validation of the Modification of Diet in Renal Disease Formula for estimating GFR with special emphasis on calibration of the serum creatinine assay. Am J Kidney Dis 2004; 44: 84-93.
- 15. Levey AS, Coresh J, Balk E, et al. National Kidney Foundation practice guidelines for chronic kidney disease: evaluation, classification, and stratification. Ann Intern Med 2003; 139: 137-147.
- 16. Rule AD, Larson TS, Bergstralh EJ, et al. Using serum creatinine to estimate glomerular filtration rate: accuracy in good health and in chronic kidney disease. Ann Intern Med 2004; 141: 929-937.
- 17. Cumming RG, Mitchell P, Craig JC, Knight JF. Renal impairment and anaemia in a population-based study of older people. Intern Med J 2004; 34: 20-23.
- 18. Go AS, Chertow GM, Fan D, et al. Chronic kidney disease and the risks of death, cardiovascular events, and hospitalisation. N Engl J Med 2004; 351: 1296-1305.
- 19. Weiner DE, Tighiouart H, Amin M, et al. Chronic kidney disease as a risk factor for cardiovascular disease and all-cause mortality: a pooled analysis of community-based studies. J Am Soc Nephrol 2004; 15: 1307-1315.
- 20. Anavekar NS, McMurray JJV, Velazquez EJ, et al. Relation between renal dysfunction and cardiovascular outcomes after myocardial infarction. N Engl J Med 2004; 351: 1285-1295.
- 21. Davies DF, Shock NW. Age changes in glomerular filtration rate, effective renal plasma flow, and tubular excretory capacity in adult males. J Clin Invest 1950; 29: 496-507.
- 22. National Kidney Disease Education Program. Information for health professionals. Frequently asked questions about estimated GFR values. Available at: http://www.nkdep.nih.gov/professionals/gfr_calculators/gfr_faq.htm (accessed Jun 2005).
- 23. Zuo L, Ma Y-C, Zhou Y-H, et al. Application of GFR-estimating equations in Chinese patients with chronic kidney disease. Am J Kidney Dis 2005; 45: 463-472.
Publication of your online response is subject to the Medical Journal of Australia's editorial discretion. You will be notified by email within five working days should your response be accepted.