|Home | About | Journals | Submit | Contact Us | Français|
Our goals in this review are to describe what is known about the prevalence and clinical implications of non-dialysis dependent chronic kidney disease (CKD) in the elderly and to discuss some of the most common challenges to managing older patients with CKD.
The prevalence of CKD rises dramatically with age. Based on the results of the National Health and Nutrition Examination Survey 1999-2004 (NHANES), more than one third of those aged 70 or older have moderate or severe CKD defined as an estimated glomerular filtration rate (eGFR) <60 ml/min/1.73 m2 [1, 2]. While all stages of CKD are more common at older ages (including earlier stages of CKD, defined as albuminuria with a preserved eGFR), it is the prevalence of moderate CKD (eGFR 30-59 ml/min/1.73 m2) that increases most dramatically with advancing age (Figure 1).
The high prevalence of CKD in the elderly no doubt reflects the presence of a variety of different risk factors for CKD such as diabetes and hypertension in older individuals. However, high rates of CKD in the elderly may also occur because of an age-associated decline in kidney function that is not explained by other known risk factors. Relatively little is known about how renal function changes during the course of “normal” aging. The Baltimore Longitudinal Study of Aging (BLSA) measured change in creatinine clearance over time among a subset of participants without kidney disease or other known comorbidities . Among these participants, creatinine clearance declined on average by 0.75 ml/min/year. However, among some participants in this study, renal function did not decline at all. Results from BLSA thus suggest that on average kidney function tends to decline with aging even among those without comorbidity, but that this decline does not appear to be inevitable.
Regardless of the reason(s) for the high prevalence of CKD in the elderly, it is clear that as the population ages we can expect to see a large expansion in the number of older individuals who meet current criteria for CKD. For example, in 2000, there were approximately 25 million adults 70 years or older, accounting for 9% of the population. By the year 2050, based on US Census Bureau projections, there will be more than 66 million people over the age of 70 and this group will account for approximately 16% of the overall population  (Figure 2).
Perhaps more relevant than whether renal function declines as part of “normal” aging and whether all older patients who meet criteria for CKD actually have CKD, is the question of whether clinical outcomes associated with a given level of renal function vary with age. Evidence to date suggests that there are substantial differences between older and younger patients in the clinical implications of a given level of estimated glomerular filtration rate (eGFR).
Older patients are less likely than their younger counterparts with similar levels of eGFR to progress to the point of needing renal replacement therapy [5-8]. This phenomenon appears to reflect both a higher competing risk of death among older patients with CKD and slower rates of decline in renal function . Supportive of the latter possibility, many older patients with moderate CKD do not experience progressive loss of renal function [6, 9, 10]. For example, Hemmelgarn et al showed over two years that there was either no or minimal progression of CKD among a large proportion of community dwelling elderly patients . Because initiation of dialysis reflects clinical decision making, age differences in rates of progression to end-stage renal disease (ESRD) may also reflect differences between patients of different ages in how often dialysis is offered or accepted when indications arise.
The effect of age on change in level of renal function appears to be more complicated than the effect on progression to ESRD. Older age is a risk factor for the development of CKD , most likely reflecting both lower mean levels of eGFR and higher rates of renal function loss in older compared with younger patients with an eGFR ≥60 mL/min/1.73 m2. Consistent with this possibility, among participants in the Baltimore Longitudinal Study of Aging without CKD, creatinine clearance declined at a higher rate among older compared with younger participants . Eriksen and colleagues described a similar phenomenon among a cohort with moderate CKD . However, in a national cohort of veterans with CKD, rates of eGFR decline were higher for older compared with younger patients at eGFR levels above 45 ml/min/1.73 m2, but at eGFR levels below this, eGFR declined more slowly for older compared with younger patients. Thus, while eGFR declines more rapidly among older patients with normal or moderately reduced eGFR, among those who survive to reach a lower level of eGFR, the reverse appears to be true. This phenomenon may reflect a survivor effect, a possibility supported by a recent study among participants in the Cardiovascular Health Study. Participants in this study elderly community dwelling Medicare beneficiaries in whom eGFR values at baseline were for the most part either normal or moderately reduced. Participants in this study who experienced the fastest decline in renal function had the highest death rates . This interesting finding raises the possibility that in elderly people with normal or moderate reductions in eGFR, those who experience rapid loss of eGFR may be less likely than others to survive long enough to reach the advanced stages of CKD.
Most patients with CKD have moderate reductions in eGFR. In this group progression to ESRD is a relatively rare event and much less common than other outcomes such as death, hospitalization and cardiovascular events [14, 15]. Rates of ESRD increase exponentially at eGFR levels below 30 ml/min/1.73 m2, but these patients with severe reductions in eGFR account for a relatively small proportion of the overall population with CKD.
Indeed, estimated glomerular filtration rate is an excellent predictor of progression to ESRD among patients of all ages and rates of ESRD rise exponentially with falling eGFR among older as among younger patients . However, at any given level of eGFR, rates of death are higher and rates of ESRD are lower among older compared with younger patients (Figure 3). Thus the threshold level of eGFR below which risk of ESRD exceeds that of death varies according to age. Among younger patients, risk of ESRD may exceed that of death even at relatively high levels of eGFR (e.g. <45 ml/min/1.73 m2). On the other hand, in the elderly death is a more common outcome than progression to ESRD even among those with severe reductions in eGFR (e.g. <30 ml/min/1.73 m2).
While older patients experience higher death rates than their younger counterparts with similar levels of eGFR, the relationship between death rates among those with and without CKD varies with age. Older patients with CKD are at higher risk of death than their younger counterparts, and severe reductions in eGFR are associated with very high death rates in the elderly [16, 17]. However, the relative risk of death associated with a given level of eGFR is attenuated at older ages [5, 16, 17]. Thus, many older patients with moderate reductions in eGFR are at no higher risk of death than their age peers with levels of eGFR above 60 mL/min/1.73m 2. For example, among a national cohort of veterans, only among those under the age of 65 were eGFR reductions in the range of 50-59 mL/min/1.73m2 associated with an increased risk of death compared with the referent group with an eGFR ≥60 mL/min/1.73 m2 . Similarly, among a British community cohort, Raymond et al demonstrated that among patients aged 75 and older with an eGFR 45-59 mL/min/1.73m2, mortality risk was no greater than for the referent group with an eGFR ≥60 mL/min/1.73 m2 . This phenomenon probably reflects both the lower mean eGFR and higher absolute mortality risk among members of the referent group of older people with an eGFR ≥60 mL/min/1.73m2. These observations are of more than academic interest because the overwhelming majority of older patients currently classified as having CKD have mild to moderate reductions in eGFR and are no more likely to die than their age peers with “normal” renal function . Thus, regardless of whether older patients with mild to moderate reductions in eGFR do or do not have CKD, their mortality risk may be no higher than for their age peers who do not meet eGFR criteria for CKD.
Because there are large age differences in both the prevalence and clinical implications of CKD, clinicians caring for older patients with CKD face some unique challenges. Below we discuss some of the most common dilemmas faced by clinicians caring for older patients with CKD.
Current guidelines prioritize interventions to reduce cardiovascular risk in patients with moderate CKD. However, such an approach may create inconsistencies in care if older patients with mild to moderate reductions in eGFR are at no higher risk of death and cardiovascular events than their age peers with higher levels of eGFR. Ultimately measures beyond eGFR are probably needed to identify the subset of older individuals with moderate reductions in eGFR at greatest risk for adverse outcomes. Evaluating change in eGFR and stability of eGFR measurements over time represent one possible approach [16, 18]. Level of urinary albumin also appears to be a valuable mortality risk stratification tool, at least for patients with moderate CKD. Among participants in the Framingham Offspring Study, the presence of microalbuminuria conferred a higher mortality risk among those both with and without a low eGFR . Among participants in the second Nord-Trondelag Health Study, level of urinary albumin distinguished between patients with CKD at high and low risk of death . In this study, the relationship between mortality and level of urinary albumin in patients with CKD was similar for older and younger participants.
Although older patients with CKD are less likely than their younger counterparts with similar levels of eGFR to progress to ESRD, because the prevalence of CKD increases markedly with age, the overall incidence of ESRD increases with age and the majority of patients who reach ESRD are 65 years or older.
Furthermore, the elderly represent the fastest growing group with end-stage renal disease (ESRD) in many countries. For example, in Canada, the numbers of incident ESRD patients over the age of 75 more than doubled between 1996 and 2005, while the number aged 20-64 years decreased slightly . While these trends largely reflect population aging, they may also reflect more liberal practices with respect to dialysis initiation in the elderly .
Thus, a major challenge facing clinicians caring for older patients who meet criteria for CKD is to identify the small percentage but large absolute number who will experience progression of their underlying CKD to ESRD. This is a group most likely to benefit from proactive efforts to slow progression and, when such efforts fail, from discussions about and planning for the management of advanced CKD. While level of eGFR is somewhat helpful in identifying those most likely to progress, the majority of older patients with severe reductions in eGFR never progress to the point of needing dialysis . At present few data are available to help clinicians identify elderly members of this group who are most likely to experience progressive loss of renal function and ultimately reach ESRD.
Nevertheless, in clinical practice primary care providers and nephrologists are frequently called on make a judgment about which of their older patients with CKD are likely to progress. For example, in deciding whether to refer a patient for hemodialysis access, the nephrologist must, at least implicitly, estimate that patient’s risk of progressing to ESRD and must also decide on an acceptable threshold level of risk at which to make this referral. Clinical practice guidelines in this area are based primarily on eGFR and/or assume that clinicians somehow know ahead of time which patients will progress to the point of needing dialysis within a pre-specified time period. Decision making in this area is particularly challenging in the elderly because even at very low levels of eGFR within the range at which access placement is recommended, the majority of elderly people do not go on to require dialysis over the subsequent 1-2 year period .
There is similar uncertainty about which elderly patients with CKD should be referred to a nephrologist. Several studies have found that older age is a predictor of late referral to nephrology care among patients beginning dialysis [24-27]. Among patients who initiate dialysis, timely referral to a nephrologist before the onset of ESRD survive longer after dialysis initiation . However, these observations were made retrospectively among the select group of patients who began chronic dialysis. With the benefit of hindsight, it is clear that patients who ultimately go on to start dialysis should have been referred beforehand to a nephrologist. However, that older patients who initiate dialysis are less likely to have been referred to a nephrologist beforehand probably speaks to the substantial complexity of prospectively identifying which elderly patients with non-dialysis dependent CKD will in fact progress to this point . While one solution to this problem would be to increase rates of referral to nephrology among older patients with earlier stages of CKD, because many of these patients will not progress and may not benefit from specialist nephrology care, a more efficient solution would be to develop ways to identify those older patients with CKD at highest risk for needing dialysis. Research on how physicians make referral decisions, with a specific focus on the geriatric population may provide greater insight in to the multiple different aspects of referral decision-making in older patients with CKD .
A third major challenge in caring for older patients expected to progress to ESRD lies in evaluating the overall benefit to these patients of recommended therapies. Outcomes both of interventions to prepare for ESRD and of ESRD itself vary by age. For example, Lok et al demonstrated that in CKD patients with functioning arteriovenous fistulae, survival and use of revision interventions were similar between patients of different ages . However, in older patients, fistulae were less likely to mature. Similarly, although survival may have improved over time for older patients initiating dialysis , survival rates are lower than for younger patients and extremely low for the very elderly . Indeed, among elderly patients with a high burden of comorbidity, it is not clear that initiation of dialysis results in improved survival compared with symptom management alone . The presence of specific age-related comorbidities may also be important. For example, based on USRDS data, dialysis patients with a diagnosis of dementia carried an adjusted hazard ratio for death of 1.87 (95% confidence interval 1.77 to 1.98) and the two-year survival rate for dialysis patients with and without dementia were 24% and 66%, respectively. Patients with dementia were also significantly more likely to discontinue dialysis than those without dementia . Withdrawal from dialysis is a common cause of death among elderly patients with end-stage renal disease and voluntary withdrawal increases with age .
Kidney transplantation and dialysis are used interchangeably to define ESRD and choice of modality is one of the decisions patients and providers should discuss before onset of ESRD . However, in reality renal transplantation is performed only rarely in patients over the age of 70. For example, of all US transplant recipients alive in 2005, only 2.4% were older than 75 . Because donor kidneys are a rare commodity, a patient-centric approach to transplant in the elderly may come into conflict with organ allocation priorities, particularly if these explicitly favor younger recipients and do not take into account physiologic (as opposed to chronologic) age. Certainly, level of comorbidity is higher and survival after transplant is lower in older compared with younger transplant recipients, and older patients are far more likely to die with a functioning graft compared with their younger counterparts . However, even among the elderly, those who receive a renal transplant appear to survive longer than those who remain on dialysis . Furthermore, there can be considerable heterogeneity in life expectancy among older patients with ESRD and in older, as in younger patients, graft and patient survival vary substantially as a function of level of comorbidity independent of age .
In conclusion, a large and growing number of elderly people meet criteria for CKD. However, many of these patients are less likely than their younger counterparts with similar levels of renal function to progress to ESRD. Indeed, most have moderate reductions in eGFR and are at no higher risk for adverse outcomes than their age peers with eGFR levels in the “normal” range. Nevertheless, a subset of older patients with CKD will progress to ESRD and these individuals account for a growing proportion of the ESRD population. Thus, a major challenge for primary care providers and nephrologists is to somehow identify the small proportion but large and growing number of older patients with CKD who are most likely to experience progressive loss of renal function and benefit from aggressive efforts to slow progression and to prepare for ESRD. While clinicians are routinely forced to make this assessment, few formal tools are currently available to identify those older patients with CKD at greatest risk for progression to ESRD. Research is thus needed to develop prognostic tools to identify those older patients with CKD who are most likely to progress to ESRD.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.