Age and the Prevalence of Chronic Kidney Disease

The prevalence of chronic kidney disease (CKD) increases markedly with age, and available data suggest that the number of older adults with advanced kidney disease will continue to increase over time. Among adults in the general population, CKD prevalence increases from less than 5% for those under the age of 40% to 47% among those age 70 and older. While the high prevalence of CKD in older adults may in part reflect a high prevalence of comorbidities associated with CKD at older ages (e.g., diabetes and hypertension), a strong age-associated increase in the prevalence of CKD is present even among those without these conditions.

The marked increase in prevalence of CKD at older ages largely reflects the estimated glomerular filtration rate (eGFR) threshold of 60 mL/min per 1.73 m ² used to define CKD and the distribution of eGFR within the population. Estimates of GFR in the general population follow a normal distribution with a median value around 80 to 90 mL/min per 1.73 m ² . The midpoint of this distribution decreases with age and moves close to 60 mL/min per 1.73 m ² in some groups. For example, in women without comorbidity participating in a community-based cohort study in the Netherlands, median eGFR ranged from 90 mL/min per 1.73 m ² for those aged 18 to 24 years to 60 mL/min per 1.73 m ² for those aged 85 and older. Less than 5% of women aged 85+ without comorbidity had eGFR close to 90 mL/min per 1.73 m ² , the median value for the youngest age group. In summary, median eGFR decreases with age, but there is substantial heterogeneity in eGFR values among patients of the same age.

The amount of urinary protein excretion also increases with age. However, the distribution of albumin-to-creatinine ratio (ACR) and its relationship to the threshold value of 30 mg/g selected for defining CKD are both quite different than for eGFR. Almost half of adults in the general population have ACR below the level of detection, whereas values of 30 mg/g or higher occur in a small minority. Because the percentage of patients who meet ACR criteria for CKD varies less as a function of age than the percentage who meet eGFR criteria for CKD, the majority of older adults with CKD have a low eGFR without significant albuminuria, whereas the majority of younger adults have albuminuria but normal eGFR ( Fig. 50.1 ).

Proportions of patients with chronic kidney disease identified by albumin-to-creatinine ratio (ACR), estimated glomerular filtration rate (eGFR), or both in the US population.

(From James MT, Hemmelgarn BR, Tonelli M. Early recognition and prevention of chronic kidney disease. Lancet . 2010;375:1296–1309. Adapted from McCullough PA, Li S, Jurkovitz CT, et al. CKD and cardiovascular disease in screened high-risk volunteer and general populations: the Kidney Early Evaluation Program [KEEP] and National Health and Nutrition Examination Survey [NHANES] 1999–2004. Am J Kidney Dis . 2008;51:S38–S45.)

Prevalence estimates for CKD in older adults vary widely depending on the methods used to estimate GFR (see Chapter 3 ). The Cockcroft-Gault and Modification of Diet in Renal Disease (MDRD) equations were not developed in populations that included a representative sample of older adults and also rely on serum creatinine, which is a marker of muscle mass. Small improvements in the accuracy of GFR estimation can substantially affect the estimated prevalence of CKD, because a disproportionately large number of patients with a low eGFR have levels that are only slightly less than 60 mL/min per 1.73 m ² . In recent years, the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation has supplanted the MDRD equation in clinical practice, resulting in slightly lower prevalence estimates of CKD in the general population, particularly for younger populations.

Progression to End-Stage Renal Disease

Patients with CKD are at risk for progressive loss of kidney function and of end-stage renal disease (ESRD), traditionally defined as initiation of chronic dialysis or receipt of a kidney transplant. The burden of kidney failure is particularly high in older adults: in the United States the mean age of new patients undergoing dialysis is 65 years, and the crude incidence of ESRD is highest among adults aged 75 and older. The high incidence of ESRD at older ages parallels the high prevalence of CKD in the elderly. However, among patients with similar levels of eGFR, those who are older are actually less likely to go on to initiate kidney replacement therapy. This phenomenon likely reflects a number of different factors, including a greater competing risk of death, lower uptake of kidney replacement therapy, differences in the accuracy of eGFR estimates, and perhaps slower loss of kidney function among older compared with younger adults.

High Competing Risk of Death at Older Ages

The incidence of ESRD increases exponentially as eGFR decreases. Mortality rates also increase with falling eGFR, but this increase is more linear and far less dramatic. Most patients with CKD have eGFR between 45 and 59 mL/min per 1.73 m ² , and because death is more common than ESRD at this level of kidney function, most patients with CKD will die before they reach ESRD. At lower levels of eGFR, the risk of ESRD eventually exceeds the risk of death for most patient groups; however, the threshold level of eGFR at which this transition occurs varies depending on age. For example, among a cohort of US veterans, the threshold level of eGFR at which the incidence of ESRD exceeded that of death was as high as 30 to 44 mL/min per 1.73 m ² for some younger age groups but as low as 15 mL/min per 1.73 m ² for some older age groups ( Fig. 50.2 ). Similar patterns are present when ESRD is defined more broadly to include patients with sustained eGFR less than 15 mL/min per 1.73 m ² who do not receive kidney replacement therapy. The relationships among eGFR, ESRD, and the competing risk of death are probably also modified by other factors, such as level of proteinuria, sex, race, and other comorbidity.

Age differences in the threshold level of estimated glomerular filtration rate (eGFR) at which risk of end-stage renal disease (ESRD) exceeds risk of death among a US cohort of veterans.

(From O’Hare AM, Choi AI, Bertenthal D, et al. Age affects outcomes in chronic kidney disease. J Am Soc Nephrol . 2007;18:2758–2765.)

Age Differences in End-Stage Renal Disease Treatment Decisions

Because most ESRD registries include only patients treated with kidney replacement therapy, little is known about those patients with advanced CKD who either prefer not to undergo dialysis or are not offered dialysis. A growing number of single-center studies outside the United States have described relatively high rates of conservative management among older adults with advanced kidney disease. Joly and colleagues reported that, among 146 consecutive patients aged 80 years or older referred to a renal unit in Paris between 1989 and 2000 with an estimated creatinine clearance less than 10 mL/min, conservative nondialytic management was recommended in 37 (25%). Similarly, among 321 patients referred to a renal unit in Britain, palliative nondialytic care was recommended in 20%. Those receiving nondialysis care were older, had lower functional status, and were more likely to have diabetes. An Australian study reported that older patients were more likely to receive information on conservative care and also were more likely to choose not to receive dialysis.

Because many elderly patients with advanced kidney disease are not referred to nephrologists, single-center studies in referred populations likely underestimate the number and percentage of elderly patients with advanced kidney disease who either prefer not to receive or are not offered dialysis. Using laboratory and administrative data from Alberta, Canada, Hemmelgarn et al. found that, while the incidence of ESRD treated with dialysis was lower in older compared with younger patients with similar levels of eGFR, the incidence of sustained eGFR less than 15 mL/min per 1.73 m ² not treated with dialysis was higher in older adults. These results suggest that there are age differences in treatment practices for advanced kidney disease.

Little is currently known about treatment practices for advanced kidney disease in the United States; however, several lines of indirect evidence support the possibility that a substantial number of elderly US patients with advanced CKD do not receive dialysis. First, across hospital referral regions in the United States, there are large differences in the incidence of treated ESRD among older adults that are not accounted for by differences in age, race, and sex. Hospital referral regions with the highest levels of health care spending in general have the highest incidence of treated ESRD. Furthermore, regional differences in the incidence of treated ESRD are most pronounced in the very elderly. Second, despite an increasing prevalence of CKD at older ages, the incidence of ESRD per million of population among US adults peaks in the 75 to 79 age group and decreases thereafter ( Fig. 50.3 ). Similar trends across age groups appear to exist for treatment of acute kidney injury (AKI) among hospitalized patients. Hsu and colleagues described age differences in the management of AKI among members of a large health maintenance organization in northern California from 1996 to 2003. While the incidence of AKI not treated with dialysis increased linearly with age, the incidence of AKI treated with dialysis peaked among those aged 70 to 79 and declined at older ages. More recently, Wong and colleagues described dialysis initiation practices in US veterans with advanced kidney disease and found that, although an implicit decision not to pursue dialysis was relatively rare among members of this cohort (<15% of patients), it was far more common in older than in younger patients. Nevertheless, even among the oldest patients with the highest level of comorbidity, most were either treated with dialysis or were preparing to be treated with dialysis at the most recent follow-up. The percentage of older patients in this study treated with dialysis was much higher than for other developed countries, suggesting that there are marked international variations in kidney failure treatment practices for older patients.

Incidence of treated end-stage renal disease by age group over time.

(From Kurella M, Covinsky KE, Collins AJ, Chertow GM. Octogenarians and nonagenarians starting dialysis in the United States. Ann Intern Med . 2007;146:177–183.)

Clinical Significance of Moderately Reduced Estimated Glomerular Filtration Rate

Adults older than 70 years account for approximately half of all US adults with CKD; however, more than half of these older adults with CKD have only moderate reductions in eGFR in the 45 to 59 mL/min per 1.73 m ² range. For many of these older patients, eGFR in this range falls close to the median value for their peers. Because eGFR is thought to decline as part of “normal” aging, some have questioned the clinical significance of such moderate reductions in eGFR in older adults.

Because death is far more common than progression to ESRD among patients with moderately reduced eGFR, the debate about the clinical significance of eGFR in older adults largely revolves around mortality risk. On a population level, the same increase in relative risk of death will be associated with a greater number of deaths in patients with higher background mortality rates (e.g., older patients). However, among individual patients, the same increase in the relative risk of death will translate into a smaller difference in life expectancy among those with more limited life expectancy. For example, a 10% increase in mortality risk translates into 1-year reduction in survival for a patient with a life expectancy of 10 years, as compared with 1-month reduction in survival for a patient with a life expectancy of 10 months. This distinction becomes clinically relevant when considering mortality risk among older patients with very moderate reductions in eGFR, in whom the relative risk of death is only modestly increased compared with the referent of patients with normal kidney function. Such modest increases in relative mortality risk may not translate into a meaningful difference in life expectancy in populations with very high baseline mortality rates.

Consistent with this possibility, several studies have demonstrated that the time to death (or relative hazard of death) associated with a given level of eGFR is attenuated at older ages. In a large national cohort of veterans, O’Hare and colleagues found that at each level of eGFR, the relative hazard of death was attenuated with increasing age. Members of this cohort aged 65 and older with GFR in the 50 to 59 mL/min per 1.73 m ² range (comprising nearly half of the cohort designated as CKD) did not have an increased relative risk of death compared with their age peers with GFR ≥60 mL/min per 1.73 m ² . On the other hand, younger members of this cohort with moderate reductions in eGFR did have a higher risk of death compared with the referent group. Similar results were reported by Raymond et al. for a large UK cohort. More recently, attenuation of the relative hazard of death associated with a given level of eGFR was reported in a large pooled analysis. This study provides the added insight that the threshold level of eGFR, above which mortality risk increases, seems to vary depending on the population studied, the referent category used, the equation used to estimate GFR, and whether analyses include information on level of proteinuria.

Given the very large numbers of older adults with moderate reductions in eGFR and uncertainty about the clinical implications of such modest reductions in eGFR, there is growing interest in efforts to distinguish high- from low-risk members of this group. Several studies suggest that information on other disease markers, such as level of albuminuria, eGFR trajectory, and cystatin C, might be useful in identifying a higher risk subgroup within the large population of older adults with moderate reductions in eGFR. While these measures may be helpful on a population level, it is important to keep in mind that there can be substantial heterogeneity in life expectancy among patients of the same age with similar levels of eGFR and proteinuria. Among a national cohort of older VA patients, the interquartile range in survival time among patients with similar levels of eGFR and proteinuria rivaled differences in median survival across strata. For example, in those aged 80 to 84 years with eGFR of 30 to 44 mL/min per 1.73 m ² and negative to trace proteinuria, median survival time was 5.3 years, but with an interquartile range of more than 6 years (2.6 to 8.7 years).