Intended for healthcare professionals


Chronic kidney disease in elderly people: disease or disease label?

BMJ 2016; 352 doi: (Published 18 January 2016) Cite this as: BMJ 2016;352:h6559
  1. Timothy Ellam, research fellow12,
  2. Helen Twohig, academic training fellow3,
  3. Arif Khwaja, consultant nephrologist1
  1. 1Sheffield Kidney Institute, Northern General Hospital, Sheffield Teaching Hospitals NHS Foundation Trust, Sheffield S5 7AU, UK
  2. 2Department of Cardiovascular Science, University of Sheffield, Sheffield, UK
  3. 3Academic Unit of Primary Medical Care, Northern General Hospital, Sheffield, UK
  1. Correspondence to: A Khwaja arif.khwaja{at}
  • Accepted 19 November 2015

Timothy Ellam and colleagues argue for a focus on what diagnosis means for individual patients rather than population risks

Around half of people aged over 75 meet current diagnostic criteria for chronic kidney disease (CKD). However, labelling them all as diseased is controversial and may cause unnecessary anxiety.1 2 The classification system defining this epidemic of CKD in elderly people is validated primarily as an epidemiological risk stratification tool rather than a clinical aid to patient management. We highlight the need to focus the debate on improving patient centred outcomes rather than just identifying everyone at above average risk.

Controversies in ageing kidneys

Older patients make up a large proportion of those reaching end stage renal disease, with 25% of people starting dialysis in the UK being over 75.3 However, the relative risk of end stage disease is much lower for an elderly patient with CKD than for a younger patient with the same level of kidney function. Thus in a large Scottish CKD cohort followed for more than five years, the individual risk of requiring renal replacement therapy (dialysis or transplant) fell from 14.33 per 100 patient years in 15-25 year olds to just 0.65 per 100 patient years in those aged 75-85.4 An estimated 1.8 million people over 75 in the UK have CKD stages 3-5,5 but in 2013 only 1700 patients in that age group started dialysis, fewer than one in 1000.

Clearly most older people currently labelled as having CKD die with abnormal kidney function rather than because of advanced kidney failure.6 One of the reasons for this is the increased risk of cardiovascular disease that accompanies advanced kidney disease. Indeed, managing cardiovascular risk is often cited as an important reason for early identification of CKD. However, as with the risk of starting dialysis, the relative effect of CKD on cardiovascular risk is much reduced in older people.7

As currently defined, CKD prevalence increases dramatically with age, from <1% with CKD stages 3-5 at age 16-24, to 41% at over 75.5 Most elderly people with CKD have CKD stage 3a (estimated glomerular filtration rate (eGFR) 45-59 mL/min/1.73 m2), which in the absence of heavy albuminuria is associated with a low risk of end stage disease and does not generally require referral to nephrology.4 An eGFR<60 was originally used to define CKD because it reflects a loss of 50% of the kidney function of an average healthy young adult.8 However, given that GFR declines with age, even in healthy cohorts such as kidney transplant donors (figure),9 there has been a concern that use of a fixed, age independent threshold risks labelling a natural age related decline as disease.2 Since the kidney is a highly vascular organ age related decline probably partly reflects increasing occult vascular disease.


Fig 1 Fall in estimated glomerular filtration rate with age in men and women from the Nijmegen community cohort (5th, 50th, and 95th centiles) and among healthy kidney donors (mean and 90% prediction interval based on regression analysis of measured GFR). Reproduced from Poggio et al9 with permission

A related controversy is the use of a urine albumin:creatinine ratio >3 mg/mmol (microalbuminuria) to define CKD in elderly people. This degree of albumin leak may also reflect generalised vascular disease and endothelial dysfunction rather than intrinsic renal disease10 and therefore is likely to overestimate the burden of disease.

Population risk stratification or clinical decision aid?

Proponents of the current staging system argue that disease definitions should be based on thresholds where risk increases rather than debating what natural renal ageing is.7 Indeed in terms of population risk stratification the current classification system “works,” with an eGFR <60 mL/min/1.73 m2 identifying an increased risk of cardiovascular mortality across all age groups compared with an eGFR of 80 mL/min/1.73m2.7 However, the increased risk of mortality from an eGFR of 45-60 compared with an eGFR of 80 in elderly people is extremely small—in fact, the risk attributable to CKD3a at age 75 is the same as that seen with albuminuria increments within the normal (submicroalbuminuric) range. A similar small increase in relative risk is seen in younger age groups at an eGFR of 70-75,7 where a diagnosis of CKD is not applied.

Does communication of small statistically significant risk associations defined at the population level translate into meaningful information for individuals? Despite concerns about causing unnecessary anxiety,2 the prevailing view seems to be that it is important to increase awareness of CKD in all groups.11 12 13 Indeed, the UK National Institute for Health and Care Excellence (NICE) guidelines stipulate that regardless of age the 18 combined categories of eGFR and albumin:creatinine ratio should be used to determine an individual’s risk of adverse outcomes and that this should be discussed with them. However, no guidance is provided on how the conversation would differ for a 30 year old and a 75 year old. For an elderly person with CKD3a the discussion should entail an explanation that end stage disease is extremely unlikely but that their cardiovascular mortality risk is very slightly higher than someone of the same age who has the same risk profile but an eGFR of 80. What this means for individual patients is far from clear. There is a negligible difference in life expectancy between elderly people with an eGFR 45-60 and those with an eGFR >60 mL/min/1.73 m2.14

Traditional cardiovascular risk factors (blood pressure, lipids, smoking status, etc) have been incorporated with age into equations for counselling patients about their cardiovascular risk.15 Although albuminuria and declining eGFR are accompanied by exponential increases in cardiovascular disease risk, these equations at best use a binary yes/no input for CKD15 and clinically useful models for predicting personal risk in patients with kidney disease are lacking.16

A continuing emphasis on population risk rather than patient centred prognosis is also evident in the recommendation for cystatin C based eGFR. This marker has been shown to improve the prediction of cardiovascular mortality in large cohorts.17 NICE guidelines recommend that measurement of cystatin C is considered in those with eGFR 45-60 and no other evidence of kidney disease.18 However, there is no evidence this adds meaningful information for counselling patients or triggers any specific intervention to improve outcomes.

Patient benefit

If the principal benefit of diagnosing CKD3a or microalbuminuria in elderly people is to identify those at higher cardiovascular risk, we should ask how this diagnosis changes patient management or improves outcomes. Interventions to reduce cardiovascular risk in the general population comprise blood pressure control, dietary and lifestyle advice, and statin therapy where appropriate. A diagnosis of CKD clearly adds little to this clinical management plan. Although recent guidelines advocate offering statins to all adults with CKD,19 the average 70 year old has a 10 year risk of cardiovascular disease greater than 10% even if they have no risk factors15 and so will be eligible for a statin anyway.19

A diagnosis of CKD3a without albuminuria also does not lead to any specific interventions to protect kidney function. Blood pressure targets are not lower and preferential use of antagonists of the renin-angiotensin system is not indicated. In fact the risk of end stage renal disease in elderly people with kidney disease is so low that few would benefit from current renoprotective interventions. For example, a simulation study in a cohort of around 370 000 veterans aged over 70 examined the effect of an intervention that reduced the risk of end stage disease by 30% over three years; over 2500 patients with CKD3a would need to be treated to prevent one case of end stage disease.20

Limiting the use of non-steroidal anti-inflammatories, radiocontrast, and other nephrotoxic agents may be the most important response to CKD3a in elderly people. However, this does not require a disease label so much as a general awareness of the need to check eGFR when considering these agents in older patients. Inclusion on a kidney disease register does not guarantee optimum blood pressure control or nephrotoxin avoidance,12 and overemphasising CKD prevalence in elderly people may contribute to a perception that it is of little consequence.21

Rational approach

Age clearly modifies the implications of lower eGFR for an individual and thus should be taken into account when counselling and managing patients. We need a better understanding of how elderly people perceive and respond to a diagnosis of CKD before promoting “disease” awareness. A reasonable approach would be simply to lower the eGFR threshold for defining disease in elderly people. Thus elderly people with an eGFR of 45-60 mL/min/1.73 m2 and no other evidence of kidney disease might not be labelled as having CKD. Given that there is no evidence that indefinite eGFR monitoring improves outcomes for patients without risk factors for progression or a history of rapid decline, it may be unnecessary to institute long term monitoring purely on the basis of an eGFR of 45-60 mL/min/1.73 m2.

Considering the complexity of the current risk stratification system, a single age based criterion is a relatively simple step towards more patient centred care. Some may consider that withholding a diagnosis of CKD from elderly patients with an eGFR of 45-60 mL/min/1.73 m2 is paternalistic.13 However, this raises the question whether it is also paternalistic to withhold risk information from those who have the same increased risk attributable to albuminuria increments below the microalbuminuria threshold or from younger people with an eGFR of 70 mL/min/1.73 m2. A line must be drawn somewhere otherwise all measurable epidemiological risk predictors will be “diseases” requiring promotion of patient awareness.

With the availability of large CKD cohort datasets it is perhaps unsurprising that the debate regarding CKD in elderly people has shifted from defining “normal” to emphasising population level attributable risk. However, the use of epidemiological and statistical concepts of risk to define disease in elderly people leads to utilitarian disease labelling that has little effect on life expectancy, quality of life, or clinical decision making. The challenge now is to shift the focus on to considering the benefits of diagnosis for individuals. This should involve clarifying patient perceptions of CKD, integrating associated risks into meaningful patient centred counselling models including age and other risk factors, and ensuring that regardless of the disease label, elderly people receive optimum care.

Primary care perspective

Although CKD3 is primarily diagnosed and managed in primary care, not all general practitioners have accepted the classification of eGFR 45-60 as a disease, especially for older patients. McIntyre and colleagues found that only 41% of a cohort of patients with CKD3 who were on disease registers were aware of having kidney problems and that those aged over 75 were more likely to be unaware.12

Qualitative work with GPs has suggested that they are wary of causing anxiety by giving a disease label to patients, particularly when clinical benefit is uncertain.21 Chronic kidney disease stage 3 sounds serious to many patients22; chronic is often taken to mean severe and stage 3 suggests it is already someway advanced. It is therefore understandable that GPs often explain the problem and its consequences without heavy emphasis on the specific terminology. However, in an age where patients can readily access their electronic clinical records and copying clinic letters to patients is commonplace there is increasingly a need to explain specific terms to patients.

One factor affecting discussions about kidney function during consultations with older patients is that it is unlikely that they will attend their GP or practice nurse for a review specifically triggered by a label of CKD. Most will be attending for monitoring of other conditions (diabetes, hypertension, etc) and kidney function is monitored as part of this. Results are therefore discussed holistically alongside those relating to other conditions. For patients with eGFRs of 45-60 mL/min/1.73 m2, relevant information could still be discussed without a label of CKD.

Patient perspective

A 71 year old woman, already taking simvastatin and antihypertensive treatment, was diagnosed with CKD3a after three successive eGFR results in the range 50-60 mL/min/1.73 m2 over several months. Urine analysis showed no abnormality and her eGFR was stable. This new diagnosis was discussed with her and a few weeks later she was asked how she felt about the discussion and what the diagnosis meant to her.

“To be honest, it did scare me. You hear the words ‘kidney disease’ and immediately think of dialysis. I did feel better when I was told it was more about my blood pressure, and probably partly just getting older, but I went on the internet when I got home and everything you read there scares you. A few days later I felt better though. I thought my GP didn’t seem worried about it and didn’t give me any medication for it so it can’t be that bad. I can understand that it’s just part of getting old. Everybody’s body deteriorates with age and everything wears down eventually. I think it’s the word ‘chronic’ that’s most worrying. It makes you feel like you’re just going to fade away and I did think, is this going to be the start of me being really ill?”

Key messages

  • Half of the elderly population meet current criteria for chronic kidney disease

  • This is based on an age neutral staging system that identifies small increases in cardiovascular risk in population studies

  • Diagnosis of early disease (stage 3a) in elderly people does not by itself lead to any specific treatment

  • Elderly patients may be better served by a system that focuses on optimising patient centred care rather than using epidemiological risk stratification to define disease


Cite this as: BMJ 2015;351:h6559


  • Competing interests: We have read and understood BMJ policy on declaration of interests and have no relevant interests to declare.

  • Contributors and sources: TE is a clinical research fellow with a research interest in the links between CKD and CVD. AK is a consultant nephrologist. HT is an academic GP. All authors helped draft the article and have approved the final version. AK is the guarantor.

  • Patient involvement: The patient quoted has approved the words in the patient perspective box and was given the opportunity to read the whole article.

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


View Abstract