Article Text

Download PDFPDF

Prognosis and management of chronic kidney disease (CKD) at the end of life
  1. Rachel Davison,
  2. Neil S Sheerin
  1. Renal Services, Freeman Hospital, Newcastle upon Tyne, UK
  1. Correspondence to Professor N S Sheerin, Institute of Cellular Medicine, Newcastle University, Newcastle upon Tyne NE2 4HH, UK; neil.sheerin{at}


The prevalence of chronic kidney disease (CKD) increases with age. As people are living longer, nephrologists are responsible for a progressively older cohort of patients with substantial comorbidities. Patients with CKD have a significant symptom burden and can benefit from intervention and symptom control from an early stage in the illness. It is also increasingly recognised that renal replacement therapy may not always offer an improvement in symptoms or a survival advantage to older patients with high levels of comorbidity. For these reasons, non-dialytic (conservative) management and end-of-life care is becoming part of routine nephrology practice. Such patients will also frequently be encountered in other specialities, requiring generalists to have some renal-specific skills and knowledge. Although there have been significant advances in this field in recent years, the optimum model of care and some of the care preferences of patients remain challenges that need to be addressed.

Statistics from


In the last 50 years, the availability of dialysis has expanded rapidly. Originally only a treatment in the acute situation for otherwise relatively young and fit patients, maintenance dialysis has now become commonplace and criteria for acceptance onto such programmes more relaxed. However, dialysis does not prevent the development or progression of other life-limiting illnesses—patients may be living a life on dialysis, but they are also dying on dialysis. For others, dialysis offers no survival benefit. Identifying and managing such patients approaching the end of their lives is increasingly complex and ever more commonly encountered.

The prevalence of chronic kidney disease (CKD) increases with age, being present in only 1% of those under 35 years of age, but in 40% of those over 75 years (figure 1).1 Registry data (figure 2) confirms that the rate of starting dialysis is highest in those over 65 years.2

Figure 1

Age-specific and sex-specific population prevalence of chronic kidney disease (CKD), showing increase with age (data taken from Stevens et al1). OR indicate common comorbidities that are often found in patients with CKD.

Figure 2

Renal Registry data indicating increasing dialysis take-on rates over time by age group.2 Access the article online to view this figure in colour.

Many factors are responsible for the increasing prevalence of older dialysis patients. First, the population is ageing. Figures from the Global Burden of Disease study show that the life expectancy for most countries has increased by around a decade over the past 40 years, leading to an increase in age-related disease, including CKD (figure 3).3 The birth rate also climbed significantly after World War II, with this generation now contributing towards the increased numbers of older people. Second, there is increasing prevalence of other chronic diseases in westernised societies, such as hypertension, diabetes and obesity1 ,4 ,5 which are associated with the development of CKD. Last, although specific data is sparse, there is a general acceptance of more relaxed criteria for considering a patient for dialysis.6 The result of these changes is that nephrologists are now responsible for an older patient cohort for whom end-of-life care is a critical part of their clinical management. This paper reviews the evidence underpinning the current trends in management and provides information for the general physician who will encounter such patients as part of their routine practice. Areas of uncertainty and future development will also be highlighted.

Figure 3

Data showing change in life expectancy at birth by country over time (1970–2010) (Reproduced from Wang H et al3 with permission from Elsevier).

Renal replacement therapy versus conservative care

The 5-year survival of patients over the age of 65 years starting haemodialysis (HD) is only 35% (see figure 4). For those over 75 years of age, survival is even worse, with only a 20% 5-year survival.7 This is significantly worse than most cancers.8

Figure 4

Kaplan–Meier survival of incident patients 1997–2009 cohort (from day 0), without censoring at transplantation (Registry data7). Access the article online to view this figure in colour.

However, age is not the only factor that impacts upon survival on dialysis. There is evidence that poor functional status and increasing number of comorbidities may have a larger effect on survival than age alone.9 These factors are synergistic rather than additive.9–11 One factor that predicts poorer outcome when starting dialysis is the degree of functional capacity (as defined by the Karnofsky scale) in the 3 months prior to the present illness. Late presentation to renal services (within 3 months of needing dialysis) also negatively affects survival.9

In patients over the age of 65 years, being on dialysis leads to lower life expectancy9 and worse quality of life12 than healthy, age-matched controls. In the older population with multiple comorbidities, renal replacement therapy (RRT) may offer neither a survival benefit,9 ,13 nor an improvement in symptom burden.14 ,15 For this reason, in recent years, there has been growing use of conservative care (also called active supportive care or maximal conservative management) as an alternative to dialysis.

Conservative (non-dialytic) care of CKD involves attention to the complications of kidney disease and personalised symptom management to maximise quality of life.16–18 In the subgroup of older patients with significant comorbidities or functional dependency, life expectancy with conservative management is largely equivalent to those on HD.9 ,13 ,19 Some studies have shown a survival benefit with HD, in the region of between 2 and 18 months.20–22 However, extension of life is not without drawbacks. Improved life expectancy may be outweighed by the fact that the majority of that time gained is spent in hospital, either as an inpatient or for dialysis.20 In some studies, when adjusted for age (>75 years) and number of comorbidities, any survival benefit from dialysis as part of a planned care plan was lost.22 ,23 The study by Joly et al21 demonstrating better survival in patients opting for HD may be considered to be biased, in that patients who presented late in the stage of their disease were more likely to be conservatively managed and, therefore, expected to have a poorer survival rate.

In patients with pre-existing impaired functional status, this may be substantially worsened by the initiation of HD, leaving patients more dependent than prior to starting planned dialysis.13

There is conflicting evidence about life expectancy with different dialysis modalities.24–29 Peritoneal dialysis (PD) has been shown to provide an equivalent quality of life to HD in those over the age of 70 years.24 ,25 Recent work has suggested that PD may actually be a better option for the older, frail, patient due to the lack of significant haemodynamic shifts and the ability to undergo treatment at home, minimising illness intrusiveness and improving quality of life.28 ,29 However, some patients may need extra assistance to be able to carry out PD, because of either cognitive, visual or dexterity problems.30 Assisted PD, where carers support the patient by setting up the delivery of PD, is a potential solution to these issues31 and is roughly cost equivalent to in-centre HD.32

Recent work has suggested that when asked to choose between dialysis and conservative management, patients are willing to accept a significantly reduced life expectancy in order to reduce the burden and restrictions placed on them by dialysis, up to 15 months in some situations.33 Decision making is also not always as logical as healthcare professionals may think, and can be influenced by a number of factors, including other patients’ reported experiences.34 ,35

Illness trajectories and predicting death

Dialysis is a life-prolonging treatment, but 5-year survival rates are worse than many cancers8 predominantly due to cardiovascular disease.7 ,8 Despite the high mortality rate for patients with CKD, it is difficult to predict when patients need support towards the end of their life. This is partially due to the unpredictable trajectory of CKD and end-stage renal disease (ESRD)36 ,37 which may be distinct from the traditional models of decline in chronic illness37 ,38 (see figure 5).

Figure 5

Traditional trajectories of chronic illnesses, as provided by Murray et al.38 Access the article online to view this figure in colour.

The use of the ‘surprise question’ (‘would I be surprised if this patient died within the next 12 months?’) was originally developed to aid palliative care decision making and overcome the natural tendencies of physicians to overestimate prognosis.39 It has since been validated in a number of chronic illness cohorts, including HD.40 It has been shown to be a reasonably accurate way of determining which patients may be in the last year of their life.

Patients with CKD and ESRD have a significant symptom burden.41 For patients on a conservative care pathway, this tends to remain fairly static until the last 2 months of life when a sharp increase is described (figures 6 and 7).36 Similarly, functional state and quality of life is relatively preserved until the final 4–6 weeks before death.40

Figure 6

Trajectory of mean Memorial Symptom Assessment Scale–Short Form (MSAS-SF) and Global Distress Index (GDI) scores over the last year of life in patients with chronic kidney disease (CKD) 5 (Reproduced from Murtagh et al36 with permission from the American Society of Nephrology Publishing Group Ltd).

Figure 7

Trajectory of mean total Palliative Care Outcome Scale (POS) scores over the last year of life in patients with chronic kidney disease (CKD) 5 (Reproduced from Murtagh et al36 with permission from the American Society of Nephrology Publishing Group Ltd).

Attempts have been made to create validated scoring systems to predict a need to change management strategies,42 ,43 or 6-month mortality for patients either on,27 ,44 or commencing,45–47 maintenance HD. These confirmed the independent predictive power of the ‘surprise question’, hypoalbuminaemia (serum albumin <25 g/L) and older age, while additionally highlighting the association of low starting body mass index (<18.5 kg/m2),45–47 certain comorbidities (peripheral vascular disease,27 ,44 dementia,27 congestive cardiac failure New York Heart Association (NYHA) classes III or IV,44–47 diabetes,44–47 cerebrovascular disease,44 active malignancy44–47), poor functional status (defined as an inability to transfer without assistance)45–48 or an unplanned start to dialysis.44–48 Taking all these factors into account may be useful in determining those who are in need of more intensive clinical input or more appropriate palliative care and symptom control measures.

Dialysis withdrawal

People who are on dialysis and approaching the end of their life need special attention. In the UK, dying after withdrawal of dialysis is the third most common cause of death (15%)7 in this patient group after cardiovascular disease and infection (20% in the USA49). Patients may choose to stop dialysis for many reasons, but the most frequent are new diagnoses affecting health or the impact of dialysis on quality of life.

The choice to stop dialysis is frequently complicated by the patient lacking capacity at that point,50 or family conflict.51 In one study,52 the majority of patients felt that their nephrologist should be responsible for making the decision to withdraw treatment if they were not able to. However, it has been shown that physicians tend to overestimate prognosis,39 and are poor at predicting patient preferences for life-sustaining treatment.53 A study of nearly 400 Japanese HD patients showed that physicians correctly predicted the patients’ wishes in only 44% with regards to cardiopulmonary resuscitation (CPR), in 47% with regards to continuing HD with severe and progressive cognitive dysfunction, and in 43% to continue HD with a diagnosis of incurable advanced cancer. Family members were no better, scoring 50%, 44% and 47% in the same scenarios.53 However, this may be culturally dependent, and there is little similar work in Europe and North America.

Another consideration is that patients often express high rates of regret at starting dialysis (up to 60% in some studies52), or that they felt pressured by doctors and their family to accept dialysis.51 ,52

The decision to stop dialysis may be a difficult one. Responsible physicians must be certain that there are no treatable, reversible causes of the patient's deterioration, and that depression is not impacting on the patient's capacity before considering dialysis withdrawal.54 Withdrawal from dialysis may be considered unacceptable in some religions and cultures.54 ,55

Survival after withdrawal from dialysis varies from 3 to 17 days,55 with a median of 4 days56 and a mean of 8–10 days.56 This can be considerably longer (up to 47 days in one study56) if a patient has significant residual renal function.

Patients and their families need to be prepared for what to expect if dialysis is withdrawn. Reassurance can be offered that for most patients, progressive impairment of consciousness as a consequence of uraemia results in a peaceful death with little associated suffering.57–60 However, some patients may develop symptoms requiring directed intervention including confusion, agitation or seizures. Sedatives will generally manage these, but may mean that patients are not as responsive or able to interact with their loved ones, something families need to be aware of.56

Pain is commonly reported,54 ,57–63 but with use of appropriate opioid-based analgesia and careful monitoring, is rarely a problem.57–60 ,63 Allowing a free diet is generally advocated, although a high salt diet may increase oedema and breathlessness. Continued fluid restriction may also help.54 Breathlessness is stated to occur in around half of all patients withdrawing from dialysis54 ,61 ,63; generally this can be managed with opiates.54 ,62 The role of intermittent continued ultrafiltration as a symptom relief tool remains controversial, mostly due to confusion from families at the use of a technique that appears similar to dialysis.56 ,61 ,64

In some patients, consideration of withdrawal of dialysis may be prompted by the diagnosis of another, different, life-limiting illness such as cancer. Although in this situation some patients will decide to withdraw from dialysis, others, for personal reasons or to control symptoms, will decide to continue dialysis.64 This can sometimes lead to problems in accessing palliative care and hospice services if there is a perception of continued ‘active treatment’.55 ,62–65

Symptom burden and control

Patients with advanced CKD are known to have a high symptom burden14 ,41 ,61 ,66–70 reported to be comparable to terminal cancer.41 Not all patients with CKD will progress to ESRD, but many will experience distressing symptoms from their renal insufficiency, requiring specialist input. The same is true for patients who choose conservative management or who withdraw from dialysis. There is evidence that healthcare professionals may underestimate the presence and severity of symptoms in patients with CKD.69 ,70 Fatigue, pain and breathlessness tend to be the most commonly reported problems.41 ,66

One of the most common complaints of patients with established CKD is of fatigue,14 ,69 ,71 affecting up to 97% patients in some studies72 and being rated as one of the most troublesome in others.73 For some patients, improvements in energy levels are rated as more important than survival on dialysis.73 The aetiology of fatigue is unclear, but appears to be linked to autonomic dysfunction and sleep disturbance.74 ,75

Pain is another common complaint14 ,61–63 which is often poorly recognised and treated.62 ,63 ,69 ,70 Pain, however, can be successfully managed17 ,18 ,57–60 ,63 ,64 with careful prescribing and avoiding opioid agents, such as morphine, which can accumulate in renal impairment.76–78

Fluid retention and oedema are, to an extent, expected in advanced CKD.14 ,68 Their severity and impact on a patient’s life can be alleviated through the use of diuretic therapy, restriction of salt and fluid intake and critically, management of patients’ expectations.17 ,18 ,61 ,62 ,66–68

Breathlessness is another frequently experienced symptom.14 ,61 ,66–68 It is commonly multifactorial; involving problems with fluid retention, fatigue, anaemia and acidosis. Comorbid conditions, such as heart and lung disease also may contribute.11 ,16 ,26 ,50 Anaemia can be corrected with the use of iron supplementation and erythropoietin therapy and acidosis minimised with bicarbonate supplementation.17 ,18

Anorexia is often experienced14 ,61 ,66–68 and remains difficult to manage alongside the restricted diet that many patients require.19 There is an increasing role of the dietician within the multidisciplinary team (MDT) to provide tailored nutritional information and supplements to those who need it.17 ,18 Nausea and vomiting can be managed through the use of antiemetic agents; the choice depends on the presumed cause of the nausea.17 ,18

Bone pain and fragility fractures may be a problem if calcium and phosphate metabolism are not adequately controlled through the use of phosphate binders, dietary restriction and vitamin D supplements.17 ,18

In common with many chronic illnesses, depression and anxiety often coexist alongside CKD14 ,15 ,36 ,41 ,43 ,52 ,61 ,62 ,66–68 ,73 and can have a significant impact upon health-related quality of life.66 ,67 This needs to be screened for on a regular basis and treated appropriately.17 ,18 Due to the altered handling of some antidepressant drugs, the involvement of the full MDT is required with input from psychiatry or clinical psychology services, as well as the consideration of the use of medications.

Advance care planning

In a survey by Davison,52 most patients reported wanting to discuss end-of-life care and advance care planning (ACP) but only 10% had ever had a discussion with their nephrologist about ACP. The issue that they most wanted to talk about were symptoms and place of care in the last days of life.

ACP is a continuous process that needs to be reflected upon, and the goals of care updated. ACP should cover physical, emotional, psychological and spiritual needs. It is as much about living well as death approaches as it is about dying itself.79 Physicians worry more than patients do about destroying hope by having such conversations.80–82 Most patients and their carers welcome such discussions,81 and there is evidence that timely ACP leads to fewer hospital visits, greater use of hospice facilities, and increasing chances of dying at home.82

Discussing matters relating to the end of life can be hard.43 ,65 The use of open-ended questions, such as ‘under what circumstances would you not want dialysis?’, or ‘what things are most important to you enjoying your life?’, can be helpful to start to broach such issues.43 ,65 ,83 ,84

Unpredictable illness trajectories36 ,37 and the high prevalence of cognitive impairment in dialysis patients85 mean that early ACP is crucial. Nearly three quarters of HD patients have moderate to severe cognitive impairment,85 most commonly due to vascular causes.85 ,86 Worsening cognitive impairment is seen as CKD progresses87 and is associated with an increased risk of death and dialysis withdrawal.88 Dialysis withdrawal in the setting of cognitive impairment is often initiated by the responsible physician, but this can cause conflict with the patient's family.51 The ultimate decision for dialysis withdrawal lies with the medical team, but it is important to stress to families that all decisions are being taken in the best interests of the patient, and wherever possible, to proceed with agreement from the family. ACP can help to alleviate some of the distress in such a situation for the medical team and the families, by documenting the patient's wishes should they lose capacity to express them, as well as specifying conditions which the patient would feel intolerable to continue dialysis or life-prolonging treatment. Each individual circumstance must be carefully considered and the involvement of the MDT is essential, as no two situations are likely to be the same.

Patients on dialysis are often unaware of the life-limiting nature of their disease despite evidence that they wish to know.52 ,89 The decision for CPR is an example of this. CPR rarely prolongs survival in HD patients in the event of a cardiac arrest, unless there is a rapidly reversible cause such as hyperkalaemia.90 ,91 The majority of dialysis patients do not show a good understanding of what CPR entails, or their chance of survival in this situation.92 ,93 However, despite not knowing much about CPR, most HD patients wished to be resuscitated in the event of a cardiac arrest.94 The need for appropriate and timely communication to help patients understand their disease and make truly informed decisions cannot be underestimated.42 ,43 ,65 ,95

Information about the preferred place of death for patients with CKD is limited, with few observational studies suggesting a preference to die at home.48 ,52 HD patients in particular, due to their long history of regular hospital attendance, may choose to receive their care in hospital, or this may be the default position due to unanticipated deterioration during frequent visits.

Delivery of care at the end of life, and current controversies

Although in recent years, there has been an increasing focus on end-of-life care for patients with CKD,42 ,43 ,65 many areas still remain uncertain and at times, controversial.

The question of whether survival in ESRD is improved by following a conservative care or dialysis-based treatment modality is unlikely to be conclusively answered. Ideally, a randomised controlled trial is required, but this would be very difficult ethically. It is probable that we will have to continue to rely on observational evidence as outlined above.

We still need to define what active conservative management should be. The use of erythropoietin, vitamin D analogues and phosphate binders all carry a significant financial and treatment burden without proven benefit in terms of survival.

Funding is a particularly difficult issue. Currently, provision of RRT with HD in the United Kingdom attracts an National Health Services (NHS) tariff of around £20 000/year.96 There is no equivalent funding for conservative care provision. This will prevent the development of high-quality end-of-life services for patients with CKD.

Access to palliative care services at the end of life is fragmented for patients with CKD, in particular, dialysis patients. In the USA, only 13.5% of the 115 239 dialysis patients who died in a 2-year period used a hospice.64 ,97 In the UK, there remains some confusion about when dialysis can be considered an active treatment and when it can be considered ‘palliative’.55 ,62–65 ,98 Increasing involvement of primary and palliative care services is helping to address this.42 ,43 ,65 ,98 Various models of care have been developed for this population, but it remains unclear which is the most useful and who should take the lead role in the process—nephrologists, palliative care physicians, primary care providers, or if it should be a shared responsibility.66 ,99 ,100


Patients with CKD and ESRD have complex needs which require careful thought and integration of services and the MDT.95 Although there have been significant advances in recent years, there remains a host of unanswered questions if we are to provide the best care possible.

Main messages

  • People are surviving longer and living with multiple comorbidities, including chronic kidney disease (CKD).

  • Patients with CKD have a number of symptoms, most of which can be managed successfully without the need for renal replacement therapy (RRT).

  • RRT may not always offer a survival advantage or an improvement in quality of life.

  • Timely advance care planning and symptom control measures, ideally earlier in the illness trajectory, are essential to enable provision of good quality care towards the end of life.

Current research questions

  • In a matched older population with significant comorbidity, which provides the best survival advantage— haemodialysis, peritoneal dialysis or conservative care?

  • What is the preferred place of death for patients with CKD/end-stage renal disease (ESRD), and how well do we achieve these goals?

  • What is the most appropriate and cost-effective model of care for end-of-life care in this population, and which professionals should lead it?

  • How can we improve access to hospice and other residential forms of care for patients with CKD/ESRD at the end of life?

Key references

  • Chandna SM, Schulz J, Lawrence C, et al. Is there a rationale for rationing chronic dialysis? A hospital based cohort study of factors affecting survival and morbidity. BMJ 1999;318:217–23.

  • Murtagh FEM, Addington-Hall J, Higginson IJ. The prevalence of symptoms in end-stage renal disease: a systematic review. Adv Chr Kidney Dis 2007;14:82–99.

  • Chandna SM, Da Silva-Gane M, Marshall C, et al. Survival of elderly patients with stage 5 CKD: comparison of conservative management and renal replacement therapy. Nephrol Dial Transplant 2011;26:1608–14.

  • Murtagh FEM, Sheerin NS, Addington-Hall J, et al. Trajectories of illness in stage 5 chronic kidney disease: a longitudinal study of patient symptoms and concerns in the last year of life. Clin J Am Soc Nephrol 2011;6:1580–90.

  • Davison SN. End-of-life care preferences and needs: perceptions of patients with chronic kidney disease. Clin J Am Soc Nephrol 2010;5:195–204.

Self assessment questions

Answer true (T) or false (F) for the below:

  1. The demographics of the renal community are changing to an older, more comorbid population.

  2. There is an age limit applied to whether you can expect dialysis or not.

  3. For some patients, survival may be roughly equivalent whether they choose to dialyse or not.

  4. The ‘surprise question’ may be useful in identifying patients on dialysis or conservatively managed, who are near the end of their lives.

  5. Dying of kidney failure causes many distressing symptoms which are difficult to manage.


  1. True

  2. False

  3. True

  4. True

  5. False



  • Contributors RD undertook the literature search and created the initial draft. All authors contributed to writing and redrafting the report. Guarantor: NSS.

  • Competing interests None.

  • Provenance and peer review Commissioned; internally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.