Intended for healthcare professionals


Early detection of chronic kidney disease

BMJ 2008; 337 doi: (Published 01 October 2008) Cite this as: BMJ 2008;337:a1618
  1. J Feehally, consultant nephrologist1,
  2. K E Griffith, general practitioner2,
  3. E J Lamb, consultant clinical scientist3,
  4. D J O’Donoghue, national clinical director for kidney care for England45,
  5. C R V Tomson, consultant nephrologist6
  1. 1University Hospitals of Leicester NHS Trust, Leicester LE5 4PW
  2. 2Wenlock Terrace Surgery, York
  3. 3East Kent Hospitals NHS Trust, Canterbury
  4. 4Department of Health, London
  5. 5Salford Royal Hospitals Foundation Trust, Salford
  6. 6North Bristol Hospitals NHS Trust, Bristol
  1. Correspondence to: J Feehally jf27{at}
  • Accepted 6 July 2008

J Feehally and colleagues explain the value of reporting estimated glomerular filtration rate

as part of a comprehensive management

programme for chronic kidney disease

The number of patients receiving long term renal replacement therapy in the United Kingdom is growing inexorably and treatment already consumes more than 2% of the total NHS budget. The cost of managing patients who present late in the course of progressive chronic kidney disease, the evidence that progression can be slowed or halted, and evidence that chronic kidney disease is associated with increased risk of cardiovascular disease have focused attention on how to optimise the management of early disease in primary care. Policies and guidelines have been designed to improve the recognition, management, and referral of patients with chronic kidney disease by non-specialists. However, the routine laboratory estimation of glomerular filtration rate (GFR) using the simplified modification of diet in renal disease (MDRD) equation has proved controversial.1 Here we describe the UK policies on identifying and managing chronic kidney disease and address some of the concerns about estimating GFR.

Strategic purpose of estimating GFR

The NHS leads the world in the development of a coherent strategy to improve the quality of care for people with chronic kidney disease. The national service framework for renal services for England was introduced in response to recognition of the growth (and cost) of the treatment of established renal failure.2 3 Renal replacement therapy costs around £30 000 (€37 000; $53 000) a year for each patient, and late referral may cost a further £15 000. The framework adopted a five stage classification of chronic kidney disease based in part on GFR and recommended systematic monitoring of kidney function in at risk populations—those with hypertension, diabetes, vascular disease, urological abnormalities, a family history of kidney disease, or taking drugs that can affect kidney function.

UK consensus guidelines on the identification, management, and referral of adults with chronic kidney disease were published in 20054 and have been widely adopted. They emphasise continuing primary care for most patients as part of integrated management of vascular risk, and specialist referral only for the few in whom this will provide “added value.”

In 2006 the Department of Health issued guidance to laboratories on use of the MDRD equation to report estimated GFR whenever serum creatinine is measured in adults, including adjustment for the biases of different creatinine assays to improve accuracy. In the same year chronic kidney disease was added to the quality and outcomes framework, rewarding primary care practitioners for developing registers of patients with stage 3-5 disease and for some aspects of treatment.5

The National Institute for Health and Clinical Excellence has also recently produced a draft guideline on chronic kidney disease. Many of the recommendations are similar to those in the UK consensus guidelines: it endorses use of estimated GFR and the five stage classification but also proposes subdivision of stage 3 into stages 3A and 3B and the addition of a p suffix to denote proteinuria (table). The Scottish Intercollegiate Guideline Network guidelines are broadly similar.6 The UK guidelines, in common with US,7 Australian,8 and international9 recommendations, all endorse the use of the simplified MDRD equation for estimation of GFR. Together these developments are designed to improve recognition and management of chronic kidney disease, which is expected to reduce the growth in the number of patients requiring renal replacement therapy and those presenting needlessly late in the course of their disease.

Staging of chronic kidney disease

View this table:

Why use estimated GFR?

Only a small proportion of patients with chronic kidney disease progress to established renal failure.10 11 Those that do so commonly have a progressive fall in GFR, marked proteinuria, and hypertension. Because of the non-linear (inverse reciprocal) relation between serum creatinine concentration and GFR, it is easier to follow changes in kidney function over time by estimated GFR than by using serial serum creatinine measurements. Early recognition, appropriate referral, and treatment of those with potentially progressive disease can prevent established renal failure in some cases and greatly improve the outcome for others.12 13

Patients with chronic kidney disease are more likely to have a cardiovascular event than to require dialysis. Lower glomerular filtration rates are associated with higher relative risk of cardiovascular death, particularly when the rate is below 60 ml/min/1.73 m2.10 This relation is most obvious in younger people. The association between cardiovascular risk and chronic kidney disease is partly due to the operation of traditional cardiovascular risk factors (obesity, hypertension, diabetes, lack of exercise), which are all also risk factors for incidence and progression of kidney disease, but the association remains after correction for multiple recognised risk factors. Proteinuria is also a powerful predictor of cardiovascular disease. As kidney function declines patients may develop renal specific cardiovascular risk factors, in particular anaemia, abnormal calcium metabolism, fluid retention, electrolyte disturbances, and lipid abnormalities.14

Accuracy of measurement

One of the concerns about using estimated GFR relates to accuracy of diagnosis. A range of factors modify the relation between serum creatinine concentration and GFR including age, sex, ethnicity, muscle mass, certain drugs, recent meat intake,15 and analytical interferences. The simplified MDRD equation adjusts for the influence of three of these variables (age, sex, ethnicity) to generate an improved, but imperfect, estimate of GFR from the serum creatinine concentration. The other limitations of creatinine (and hence GFR) in this setting remain, but this is no more an argument for not reporting estimated GFR than it is for stopping use of serum creatinine.

All laboratory measurements are associated with a degree of uncertainty. The introduction of estimated GFR has resulted in a welcome, if belated, recognition of the variation in creatinine assays in use. In North America it was shown that a reported GFR of 60 ml/min/1.73 m2 could in effect range from 42 to 67 ml/min/1.73 m2.16 In the UK, laboratory correction factors have reduced between laboratory variation, and standardisation of the method for estimation of GFR according to the creatinine assay used reduces the risk of overdiagnosis of stage 3 disease considerably.17

Misdiagnosis of stage 3 disease occurs for two reasons: underestimation of GFR (which the steps above have minimised) and use of a single estimated GFR instead of two measurements taken at least three months apart. Misdiagnosis of stage 2 disease based on an estimated GFR of 60-90 ml/min/1.73 m2 is also common. This is not because of the use of estimated GFR but because practitioners are unaware that diagnosis of kidney disease at this range of GFR requires the presence of other evidence of disease—for example, from imaging, biopsy, or urine analysis.

Implications for patients

Concerns have been raised about the psychological harm caused by diagnosing patients with early stages of disease, when the risk of progression is low. Most patients diagnosed with chronic kidney disease already have one or more chronic diseases (such as diabetes, hypertension, coronary heart disease). Although being told that they have kidney disease may cause concern, whether it causes harm probably depends on how people are told, rather than the terms used.

Insurance providers assess risk from population data describing the increased morbidity and mortality associated with chronic kidney disease. Not all insurers will make identical assessment of such risk, but our discussions with the insurance industry do not indicate any systematic misunderstanding of the interpretation and implications of reporting estimated GFR. Patients should therefore not be disadvantaged by measurement.

Value in older people

Low GFR (<60 ml/min/1.73 m2) is extremely common among older people, leading some to conclude that this is normal for age and not a disease state. There is no evidence of poorer performance of the prediction equation in older people.18 Established renal failure and late referral for renal replacement therapy are also more common among older people. Not all older people have a low GFR, and those that do are more likely to have a history of hypertension and of other risk factors for kidney disease. If “disease” refers to an increased risk of symptoms, complications, or of an adverse prognosis,19 then older patients with lower GFR are rightly labelled as having chronic kidney disease. An analogy can be made with bone density, where age related decline is associated with substantial morbidity; like GFR, bone density is expressed in terms of how far it deviates from that seen in young healthy people.

Practical implications for primary care

Identification of patients with stage 3-5 chronic kidney disease is now relatively straightforward in primary care by searching practice computer databases. Evaluation of newly identified patients includes exclusion of serious obstruction of the bladder outlet and ensuring the patient is not taking nephrotoxic drugs (such as non-steroidal anti- inflammatories).4

Up to 90% of patients identified as having chronic kidney disease will already be on one or more disease registers for coronary heart disease, hypertension, or diabetes. Practices can maximise efficiency by establishing one stop, combined vascular clinics to manage these conditions, focusing first on blood pressure control, with preferential use of angiotensin converting enzyme inhibitors or angiotensin receptor blockers in the presence of diabetes, proteinuria, or established vascular disease. Most patients with underlying vascular disease and diabetes are likely to need two or more antihypertensive drugs.

Results from the quality and outcomes framework for 2006-7 indicated that 99% of general practices had a register of chronic kidney disease; 3% of the population had been identified with stage 3-5 disease; 98% of identified patients had had their blood pressure measured in the past 15 months; and 86% of non-excepted patients had a blood pressure of 140/85 mm Hg or less.20 This is an important first step to improving management of chronic kidney disease in primary care. The audit measures fell short of ideal blood pressure control, and did not require testing for proteinuria, a key risk discriminant in assessing both cardiovascular and renal risk. Indicators have been revised for 2008-9 so that treatment with angiotensin converting enzyme inhibitors or angiotensin receptor blockers is now indicated in patients with hypertension and proteinuria.5 Implementation of this will require that all patients who have chronic kidney disease and hypertension are tested for proteinuria.

A study of detection of chronic kidney disease in primary care using systematic estimation of GFR found that in the first year a practice of 10 000 would identify 148 new patients with stages 3-5 disease, predominantly reflecting patients with previously undetected disease.21 It concluded that for the practice to recoup the costs, the strategy would have to delay end stage renal failure (avoiding the cost of renal replacement therapy) in only one patient for one year.

Early achievements

Current UK policies have increased awareness of chronic kidney disease as a common, harmful, and treatable condition in primary and secondary care. The understanding that estimated GFR is a measure of percentage of normal kidney function is helping to demystify kidney disease. Three per cent of the adult population of England has been identified as having kidney disease and placed on primary care registers in the first year of the programme. There will inevitably be patients in whom the classification and guidelines are difficult to interpret or implement, but this should not detract from the substantial improvements in renal and vascular care for the majority of people with chronic kidney disease.


Cite this as: BMJ 2008;337:a1618.


  • Contributors and sources: This article was drawn up in response to articles hostile to the introduction of routine reporting of estimated GFR in the BMJ and is based on evidence used for the national service frameworks and additional literature reviews undertaken during the preparation of the UK guidelines. All authors contributed sections of this article and approved the final draft. JF is guarantor.

  • Competing interests: JF is immediate past president of the Renal Association and a member of the renal advisory group at the Department of Health for England. KEG is a member of the Department of Health’s renal advisory group and secretary of the UK Primary Care Cardiovascular Group. EJL represented the Association for Clinical Biochemistry on the working party that drew up the UK chronic kidney disease guidelines. He was also on the Guideline Development Group for the National Institute for Health and Clinical Excellence guidelines. CRVT is past chair of the joint committee on renal disease of the Royal College of Physicians and the Renal Association and chaired the working party that drew up the UK chronic kidney disease guidelines.

  • Provenance and peer review: Not commissioned; externally peer reviewed.