How to measure renal function in clinical practiceBMJ 2006; 333 doi: https://doi.org/10.1136/bmj.38975.390370.7C (Published 05 October 2006) Cite this as: BMJ 2006;333:733
- Jamie Traynor, specialist registrar ()1,
- Robert Mactier, consultant nephrologist2,
- Colin C Geddes, consultant nephrologist1,
- Jonathan G Fox, consultant nephrologist2
- Correspondence to: J Traynor
- Accepted 29 August 2006
The reliable measurement of renal excretory function is of great importance in clinical practice and in research. The introduction of routine reporting of estimated glomerular filtration rate and a new definition of chronic kidney disease has renewed interest in methods of measuring renal function. Coupled with this is the fact that several countries are moving towards population screening for renal impairment to try to reduce the associated increased cardiovascular risk. Accurate measurement is methodologically difficult so surrogate measures such as serum creatinine levels and prediction formulas (based on factors such as the patient's age, sex, and serum creatinine level) are more commonly used in routine practice. We describe routine and more specialised methods of assessing renal function and discuss estimated glomerular filtration rate.
The kidney has several interlinked functions (box). These depend on glomerular filtration rate, the unit measure of kidney function. Glomerular filtration rate can be defined as the volume of plasma cleared of an ideal substance per unit of time (usually expressed as ml/min). The ideal substance is one that is freely filtered at the glomerulus and neither secreted nor reabsorbed by the renal tubules.
Creatinine is the closest to an ideal endogenous substance for measuring glomerular filtration rate.w1 Plasma creatinine is almost exclusively a product of the metabolism of creatine and phosphocreatine in skeletal muscle, although ingestion of meat may also contribute slightly.w2 w3 In patients with stable renal function, serum creatinine levels are usually constant, with variability daily of about only 8%.w4 w5 Creatinine is freely filtered at the glomerulus and is not reabsorbed, but up to 15% is actively secreted by the tubules.w6 In advanced renal failure, excretion of creatinine through the gastrointestinal tract increases.w7
Measuring the creatinine clearance using serum creatinine level and a timed urine collection gives an estimate of glomerular filtration rate:
As a result of tubular secretion of creatinine, creatinine clearance tends to overestimate true glomerular filtration rate. This is a systematic error of fairly stable magnitude, however, until advanced renal failure is reached, allowing creatinine clearance to be a reasonable method of following changes of renal function in patients. The main problem with creatinine clearance is the requirement for urine collection over 24 hours; patients find this inconvenient and therefore collections are often inaccurate. Also a 25% daily variation in the values obtained using this method has been reported.w8 Creatinine clearance is therefore no longer much used in clinical practice.
Estimated glomerular filtration rate forms the basis for the classification of chronic kidney disease
An estimated glomerular filtration rate of 60-89 ml/min/1.73 m2 in the absence of other evidence of kidney disease does not signify chronic kidney disease and does not indicate that further testing is required
Patients with chronic kidney disease are at high risk of cardiovascular disease
Estimated glomerular filtration rates, calculated using the “4-v MDRD” based formula, are now routinely reported by biochemistry laboratories alongside serum creatinine results (except in non-validated patient groups)
Serum creatinine levels or estimated glomerular filtration rates may be used to monitor changes in renal function in an individual patient
Formal measurement of glomerular filtration rate is used for accurate assessment of renal function in potential kidney donors and in research studies
Radioisotope or iothalamate methods require multiple blood samples and, if renal function is reduced, the duration of sampling may be up to 24 hours
Sources and selection criteria
We searched Medline from 1966 onwards using the search terms: “measurement of renal function”, “modification of diet in renal disease”, “Cockcroft and Gault”, “radio-isotopes”, “glomerular filtration rate”, “inulin clearance”, and “cystatin C”
We also referred to several textbooks, including the Oxford Textbook of Clinical Nephrology third edition (Oxford University Press) and Comprehensive Clinical Nephrology second edition (Mosby)
Serum urea is a less reliable marker of glomerular filtration rate than creatinine because levels are more vulnerable to change for reasons unconnected to glomerular filtration rate. A high protein diet, tissue breakdown, major gastrointestinal haemorrhage, and corticosteroid therapy can lead to an increase in plasma urea whereas a low protein diet and liver disease can lead to a reduction. Also, 40-50% of filtered urea may be reabsorbed by the tubules, although the proportion is reduced in advanced renal failure.w9
Mean of urea and creatinine clearance
In advanced renal failure the mean of urea and creatinine clearance may give a more accurate estimate of glomerular filtration rate than either clearance alone, as the effects of urea reabsorption and creatinine secretion tend to cancel each other out.w10 It is the recommended method for estimating residual renal functionw11 in patients receiving dialysis.
No endogenous ideal substance exists for measuring glomerular filtration rate, so the standard method requires infusion of an exogenous agent, such as inulin. Inulin, a polymer of fructose (5200 daltons), is found in Jerusalem artichokes, dahlias, and chicory and was first used for measuring glomerular filtration rate in 1951.w12 Its use is limited because purified inulin is expensive and difficult to measure and measuring glomerular filtration rate in this way is time consuming for both patients and clinicians. A bolus and infusion of inulin are given to achieve a steady plasma level, followed by collection of regular blood and urine samples over several hours for inulin estimation. This method (nowadays often using polyfructosan (Inutest; Fresenius, Austria) is only used in research studies when very accurate estimation of renal function is necessary.
Functions of kidney related to glomerular filtration rate
Water soluble medicines (for example, digoxin, gentamicin)
Control of blood pressure
Secretion of erythropoietin
Hydroxylation of vitamin D1 (activation)
Gluconeogenesis in the fasting state
Catabolism of peptide hormones (including insulin)
From the late 1960s the use of radionuclides has offered an alternative method of estimating glomerular filtration rate that avoids some of the practical disadvantages of inulin clearance. Estimates using radionuclides correlate closely with inulin clearance.1–4 Radionuclides are usually given as a single dose and the glomerular filtration rate calculated by their rate of disappearance from the plasma, obviating the need for urine tests. When a substance is given under these conditions, two phases of disappearance occur (fig 1):
Computer software is used to calculate the glomerular filtration rate based on data either from both phases (double pool) or from the terminal elimination phase (single pool). Single pool methods have the advantage that fewer plasma samples are required.
Radioisotopic methods have the disadvantage of precautions being required in handling and disposal of radioactive materials. They are also expensive and not suitable for use during pregnancy. Another important consideration is that the terminal elimination phase is significantly prolonged in advanced renal failure. In patients with moderate renal failure (glomerular filtration rate 30-59 ml/min) samples are taken for up to five hours after injection whereas in patients with advanced renal failure samples are required for up to 24 hours after injection.5
Radiocontrast agents were initially available in the 1960s but difficulties in chemical analysis and unacceptable amounts of free iodine in the preparations limited their use in favour of radioisotopic agents.w13 These problems have largely been resolved and radiocontrast agents now offer the advantages of radioisotopes without the concerns of radioactive substances. Agents currently in use are iothalamate (Conray; Mallinckrodt, St Louis, MO), siatrizoate meglumine (Hypaque; Amersham Health, NJ), and iohexol (Omnipaque; Amersham Health, NJ). Iohexol may be the agent of choice as it is relatively quick to use and its results are comparable to inulin clearance.6 7
The past decade has witnessed an upsurge of interest in cystatin C as an endogenous glomerular filtration rate marker. Cystatin C is part of the cystatin “superfamily” of cysteine protease inhibitors. It is freely filtered at the glomerulus. Its use is, however, limited by higher variability of serum levels than creatinine (75% v 7%) between patients.8 Also, serum levels are increased in malignancy,w14 w15 HIV infection,w16 and glucocorticoid therapy.w17 At present cystatin C has no established role, but it may emerge as a useful way of identifying patients with early renal failure as part of screening programmes.w18
To circumvent the practical difficulties of formal measurement of clearance, several prediction formulas have been published. The most commonly used are the Cockcroft and Gault equation and formulas based on the modification of diet in renal disease study (fig 2). Given the limitations of serum creatinine in identifying renal failure, the increased use of prediction formulas, in particular the modification of diet in renal disease formulas, has been advocated.9
Cockcroft and Gault equation
The Cockcroft and Gault equation, which estimates creatinine clearance on the basis of serum creatinine level, age, sex, and weight, was one of the earliest prediction formulas10 and is still widely used. It was based on creatinine excretion in men with normal renal function with a correction for women, based on three other studiesw19-w21: it tends to overestimate renal function at lower levels, particularly when obesity or fluid overload is present, as the resultant increase in weight does not reflect an increase in muscle mass. However, as with creatinine clearance, this is largely a systematic error and the equation remains useful for following changes in renal function in a patient.
Modification of diet in renal disease formula
More recently Levey et al introduced a formula derived from data on patients with advanced renal failure in the modification of diet in renal disease study.11–14 This is referred to as the “6-variable MDRD” or “6-v MDRD” formula.15 This formula gives an estimate of glomerular filtration rate in millilitres per minute adjusted for body surface area of 1.73 m2 and is based on a patient's age, sex, race, and levels of serum urea, serum creatinine, and serum albumin. By avoiding inclusion of weight, the formula is less prone to errors from fluid overload and obesity.
In 2000 a simplification of the modification of diet in renal disease formula using only patient's age, sex, race, and serum creatinine level was derived from the original data.16 This is referred to as the “4-variable MDRD” or “4-v MDRD” formula. With the exception of race, the other variables required are normally provided routinely when a sample is submitted to the laboratory. It is therefore much easier for laboratories to report estimated glomerular filtration rate using this formula.
One important issue concerning the use of prediction formulas is that different laboratories use different methods for creatinine estimation. Some assays are more sensitive than others to non-creatinine chromogens, which falsely increase creatinine values. This error is magnified in prediction formulas, particularly in patients with higher levels of renal function.17 18 To replicate as closely as possible the results obtained in the modification of diet in renal disease study, all serum creatinine values should ideally be determined using the methods of the Cleveland Clinic laboratory in the original modification of diet in renal disease study.17 18 This raises problems for different providers of analytical systems and would not be straightforward to achieve. However, some of the difference in assay methods can be corrected for and there are plans to adopt correction factors throughout the United Kingdom. The United Kingdom National External Quality Assessment Service has played a key part in this change. When using a correction factor, the formula used is a slightly modified form of the 4-v MDRD formula.19
The modification of diet in renal disease formulas have been validated in an ever increasing number of patient groups, including elderly patients and recipients of renal transplants,20–22 although concern has been expressed over reliability in different ethnic groups such as Chinese and Indian patients.23 24 Also, as these formulas were based on data from patients with advanced renal failure, their validity has been questioned in patients with normal or near normal glomerular filtration rates. It is therefore recommended that they are not used routinely at levels greater than 60 ml/min/1.73 m2. It should be stressed that these formulas are not valid in certain clinical settings such as acute renal failure, pregnancy, severe malnutrition, diseases of skeletal muscle, paraplegia, and quadriplegia, in children, or when renal function is changing rapidly.
Plans are in progress to report estimated glomerular filtration rates whenever measurement of serum creatinine is requested throughout the United Kingdom. Some centres have already started providing estimated glomerular filtration rates routinely and most centres should be doing so by the end of 2006. This is likely to identify large numbers of patients with reduced glomerular filtration rates who may have been overlooked when their renal function was assessed by serum creatinine levels alone: data from the United States suggest that around 5% of adults have chronic kidney disease stages 3 (glomerular filtration rate 30-59 ml/min), 4 (15-29 ml/min), or 5 (< 15 ml/min). These people have a markedly increased risk of cardiovascular morbidity and mortality and identification of them should enable earlier initiation of measures to reduce cardiovascular risk and also the rate of progression of renal failure. A similar prevalence figure has been reported in the United Kingdom.25 The table lists the five stages of chronic kidney disease.
The UK Renal Association among associations in other countries, including Canada and Australia, has introduced guidelines for targeted screening to detect reduced renal function in primary care using estimated glomerular filtration rate. These guidelines describe how to manage most patients in primary care and advise which patients should be referred to nephrologists. In the case of Australia and New Zealand, specific attention is recommended to high risk subgroups such as those of aboriginal descent. The potential impact of identifying patients with reduced renal function has been recognised by the United Kingdom's health service and is reflected in the General Medical Services contract for general practitioners, which now provides remuneration for the identification and monitoring of chronic kidney disease.
Additional educational resources
Renal Association (www.renal.org/eGFR/)—this website has clear information on definition of chronic kidney disease and provides guidelines on management. A section is included for patients
NHS Employers (www.nhsemployers.org/primary)—this website provides further details of the general practitioner contract and remuneration of general practitioners within the United Kingdom
National Kidney Foundation (www.kidney.org/kidneyDisease/)—this website provides useful information for patients as well as another good resource for health professionals
Tips for non-specialists
Use estimated glomerular filtration rate as a guide to renal function in conjunction with advice on renal disease from the Renal Association's website (www.renal.org/eGFR/)
If the estimated glomerular filtration rate is < 60 ml/min then:
Review previous results or repeat the measurement to assess if renal function is stable or declining
Measure blood pressure and test urine for protein and blood
Review drugs for potentially nephrotoxic agents, such as angiotensin converting enzyme inhibitors or angiotensin receptor blockers, diuretics, non-steroidal anti-inflammatory drugs, and antibiotics
Check for urinary symptoms, signs of fluid retention or hypovolaemia, and palpable bladder
Enter into a chronic disease management programme and decide whether referral to a renal clinic is appropriate, using local guidelines or those from the Renal Association
Prediction formulas using serum creatinine levels are by far the most widely used methods of measuring renal excretory function in routine clinical practice. One of these, the modified 4-v MDRD estimated glomerular filtration rate formula19 is now being used for direct reporting of estimated glomerular filtration rates by laboratories and has become the standard method used to identify and monitor patients with reduced renal function in the United Kingdom and elsewhere. It is anticipated that recognition and appropriate management of patients with chronic kidney disease will reduce cardiovascular events and slow further deterioration in renal function in these patients.
References w1-w21 are on bmj.com
Contributors JT wrote the main draft of the manuscript and oversaw submission. He is guarantor. RM, CCG, and JGF assisted with writing and amending the manuscript.
Competing interests None declared.