eGFR indicates estimated glomerular filtration rate. Adjusted for sex, income quintile, residence, First Nations status, proteinuria, diabetes, hypertension, and Charlson comorbidities (cancer, cerebrovascular disease, congestive heart failure, chronic obstructive pulmonary disease, dementia, myocardial infarction, mild liver disease, paraplegia/hemiplegia, peptic ulcer disease, peripheral vascular disease, and rheumatologic disease).
Error bars indicate 95% CIs.
Hemmelgarn BR, James MT, Manns BJ, et al. Rates of treated and untreated kidney failure in older vs younger adults. JAMA. 2012;307(23):2507-2515.
eFigure 1. Adjusted Risks (Hazard Ratios; 95% Confidence Intervals) of Treated Kidney Failure and Untreated Kidney Failure Accounting for Competing Risk of
Death (Common Reference: Age 45-54 Years; eGFR 60-89 mL/min/1.73 m2)
eFigure 2. Adjusted Rates (per 1000 Person-Years) of Treated Kidney Failure and Untreated Kidney Failure, by Age and eGFR, With Untreated Kidney Failure
Defined by an eGFR <10 mL/min/1.73 m2
eFigure 3. Adjusted Rates (per 1000 Person-Years) of Treated Kidney Failure and Untreated Kidney Failure, by Age and eGFR, With Untreated Kidney Failure
Defined by at Least Two Measures of eGFR <15 mL/min/1.73 m2 During a Six
eFigure 4. Adjusted Rates (per 1000 person-years) of Treated Kidney Failure and Untreated Kidney Failure, by Age and eGFR, With Untreated Kidney Failure
Defined by eGFR <15 mL/min/1.73 m2 Irrespective of Renal Replacement Therapy
Customize your JAMA Network experience by selecting one or more topics from the list below.
Hemmelgarn BR, James MT, Manns BJ, et al. Rates of Treated and Untreated Kidney Failure in Older vs Younger Adults. JAMA. 2012;307(23):2507–2515. doi:10.1001/jama.2012.6455
Context Studies of kidney failure in older adults have focused on receipt of dialysis, which may underestimate the burden of disease if older people are less likely to receive treatment.
Objective To determine the extent to which age is associated with the likelihood of treatment of kidney failure.
Design, Setting, and Participants Community-based cohort study of 1 816 824 adults in Alberta, Canada, who had outpatient estimated glomerular filtration rate (eGFR) measured between May 1, 2002, and March 31, 2008, with a baseline eGFR of 15 mL/min/1.73 m2 or higher and who did not require renal replacement therapy at baseline. Age was categorized as 18 to 44, 45 to 54, 55 to 64, 65 to 74, 75 to 84, and 85 or more years and eGFR as 90 or higher, 60 to 89, 45 to 59, 30 to 44, and 15 to 29 mL/min/1.73 m2.
Main Outcome Measures Adjusted rates of treated kidney failure (receipt of dialysis or kidney transplantation), untreated kidney failure (progression to eGFR <15 mL/min/1.73 m2 without renal replacement therapy), and death.
Results During a median follow-up of 4.4 years, 97 451 (5.36%) died, 3295 (0.18%) developed kidney failure that was treated and 3116 (0.17%) developed kidney failure that went untreated. Within each eGFR stratum the rate of treated kidney failure was higher in younger compared with older people. For example, in the lowest eGFR stratum (15-29 mL/min/1.73 m2), adjusted rates of treated kidney failure were more than 10-fold higher among the youngest (18-44 years) compared with the oldest (≥85 years) groups (adjusted rate, 24.00 [95% CI, 14.78-38.97] vs 1.53 [95% CI, 0.59-3.99] per 1000 person-years, respectively; P < .001). Rates of untreated kidney failure were consistently higher at older ages. In the eGFR stratum of 15 to 29 mL/min/1.73 m2, adjusted rates of untreated kidney failure were more than 5-fold higher among the oldest (≥85 years), compared with the youngest (18-44 years) groups (adjusted rate, 19.95 [95% CI, 15.79-25.19] vs 3.53 [95% CI, 1.56-8.01] per 1000 person-years, respectively; P < .001). Rates of kidney failure overall (treated and untreated combined) demonstrated less variation across age groups; eg, the adjusted rate per 1000 person years for those with eGFR of 15-29 mL/min/1.73 m2 was 36.45 (95% CI, 24.46-54.32) among participants aged 18 to 44 years and 20.19 (95% CI, 15.27-26.69) among those aged 85 years or older (P = .01).
Conclusion In Alberta, Canada, rates of untreated kidney failure are significantly higher in older compared with younger individuals.
Although patients aged 65 years or older represent the fastest-growing segment of the population initiating long-term dialysis,1-3 rates of dialysis initiation peak by age 75 years and decline thereafter.3 Potential explanations for this observation include higher mortality among elderly people with non–dialysis-dependent chronic kidney disease (CKD), markedly reduced progression of CKD in this age group, and increased incidence of untreated kidney failure (that is, the decision not to initiate dialysis) in elderly individuals.
Studies of the association among age, kidney function, and clinical outcomes have reported that elderly patients are less likely to develop end-stage renal disease (ESRD) compared with younger patients and are more likely to die than to progress to kidney failure even at the lowest levels of estimated glomerular filtration rate (eGFR).4-6 Furthermore, although reduced eGFR is common among older adults, the clinical significance of moderate reductions in eGFR and risk of mortality in particular has been questioned.7-9 However, previous studies have defined kidney failure by receipt of long-term dialysis, which reflects both disease progression and a treatment decision. Because it is plausible that the likelihood of initiating long-term dialysis among individuals with kidney failure varies by age, earlier studies may provide an incomplete picture of the burden of advanced kidney disease in older adults, based on the incidence of long-term dialysis alone.
We studied a community-based cohort of more than 1.8 million adults to determine whether age is associated with the likelihood of treated kidney failure (renal replacement therapy: receipt of long-term dialysis or kidney transplantation), untreated kidney failure (eGFR <15 mL/min/1.73m2 without renal replacement therapy), and all-cause mortality. The size of the cohort permitted a refined characterization of the association between age and the risk of these adverse outcomes.
We performed a retrospective cohort study using laboratory data and administrative data from Alberta, Canada. The study population included Alberta residents aged 18 years or older who had at least 1 outpatient serum creatinine measurement between May 15, 2002, and March 31, 2008.10 The index date was defined as the date of the first serum creatinine measurement. There were 1 821 850 participants aged 18 years or older with at least 1 outpatient serum creatinine measurement during the cohort enrollment period. We excluded 231 people who died on their index date, 3168 because they had ESRD at the time of cohort entry, and 1627 because their baseline (index) eGFR was less than 15 mL/min/1.73 m2, leaving a final cohort of 1 816 824 participants. The institutional review boards of the University of Calgary and University of Alberta approved the study and granted waiver of patient consent.
We estimated the index eGFR using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation.11 To reduce interlaboratory variation, creatinine measurements were standardized across provincial laboratories to an isotope dilution mass spectrometry reference standard, and a laboratory-specific correction factor was applied where necessary.10 Index eGFR was categorized as 90 or higher, 60 to 89, 45 to 59, 30 to 44, and 15 to 29 mL/min/1.73 m2 based on outpatient creatinine measurements only.
Proteinuria was measured by urine dipstick based on random outpatient spot urine measurements. We defined proteinuria as normal (urine dipstick negative), mild (urine dipstick trace or 1+), heavy (urine dipstick ≥2+), and not measured (no outpatient urine dipstick measurement).12 To establish baseline proteinuria, all measurements in the 6-month periods before and after the index eGFR were used. The median of all measurements was used for patients with multiple measurements.13
Patients were followed up from the date of cohort entry until the date of the renal outcome of interest, death, outmigration from the province, or study end (March 31, 2009). We considered 3 outcomes: all-cause mortality (obtained from the Alberta Health and Wellness Registry file), treated kidney failure (initiation of long-term dialysis [acute dialysis not included], as determined from provincial dialysis registry and administrative data using a validated algorithm, or receipt of a kidney transplant14), and untreated kidney failure (progression to an eGFR <15 mL/min/1.73 m2 without dialysis treatment or a kidney transplant). The definition of untreated kidney failure was consistent with current guidelines15 and required a sustained eGFR of less than 15 mL/min/1.73 m2 during follow-up (with absence of renal replacement therapy, although other medical treatments were permitted). All outpatient serum creatinine measurements following the index measure were considered during the study period (inpatient serum creatinine measurements were not considered). The renal outcomes were mutually exclusive. Specifically, patients were classified as having treated kidney failure if they received long-term dialysis treatment or a kidney transplant at any time, irrespective of their eGFR.
Administrative data from the provincial health ministry were used to define demographic characteristics and comorbidity. Aboriginal race/ethnicity was determined from First Nations status in the registry file; it was not possible to identify other racial/ethnic groups, although more than 85% of the Alberta population is white.16 Diabetes mellitus and hypertension were identified from hospital discharge records and physician claims using validated algorithms.17,18 Other comorbid conditions based on the Deyo classification of Charlson comorbidities were identified using validated International Classification of Diseases (ICD), Ninth Revision and ICD-10 coding algorithms applied to physician claims and hospitalization data.19 The presence of 1 or more diagnostic codes (in any position) up to 3 years before cohort entry was used for identification of these comorbidities. Drug use in Alberta is only available for adults aged 65 years or older and therefore was not included. The 6-digit residential postal code for each participant was linked to the 2001 or 2006 Canadian Census (whichever was closest to the index date) using the postal code conversion file to determine median neighborhood household income quintile and rural/urban location of residence.
We performed analyses stratifying age as 18 to 44, 45 to 54, 55 to 64, 65 to 74, 75 to 84, and 85 or more years. Baseline characteristics of the cohort were compared by age category. For each of the outcomes (treated kidney failure, untreated kidney failure, and all-cause mortality), rates per 1000 person-years were estimated using Poisson regression, by age and eGFR strata, with eGFR categorized as 90 or higher, 60 to 89, 45 to 59, 30 to 44, and 15 to 29 mL/min/1.73 m2. Because the primary objective was to determine whether age modifies the likelihood of treated and untreated kidney failure, we assessed the multiplicative interaction between eGFR and age categories on these outcomes. If the Poisson assumption that the variance equals the mean was not met, a quasi-Poisson model was used.20 Rates were adjusted to the sample proportions of the demographic and clinical characteristics listed in the Table. P values for differences in rates between the age groups were determined with age 18 to 44 years as the reference.
We also modeled the cumulative incidence of each of the renal outcomes (treated kidney failure and untreated kidney failure) with death as a competing risk, using the competing proportional subdistribution hazard approach.21 To assess the combined effects of eGFR and age on each of the renal outcomes, we used the full study cohort and a common reference category of age 45 to 54 years with eGFR of 60 to 89 mL/min/1.73 m2 to calculate the hazard ratios.
We performed several sensitivity analyses to assess the robustness of our findings. We repeated all analyses using a composite outcome of treated and untreated kidney failure, an outcome defined as a sustained eGFR of less than 15 mL/min/1.73 m2 irrespective of renal replacement therapy, and an outcome of untreated kidney failure defined as progression to a sustained eGFR of less than 10 mL/min/1.73 m2. Finally, although our primary definition of untreated kidney failure required at least 2 measurements of eGFR (baseline ≥15 mL/min/1.73 m2 followed by ≥1 eGFR <15 mL/min/1.73 m2), we repeated analyses with untreated kidney failure defined as 2 measurements of eGFR of less than 15 mL/min/1.73 m2 during a 6-month period. Statistical analyses were performed with SAS version 9.2 (SAS Institute Inc) and R version 2.13.1 (R Project) software. P < .05 indicates statistical significance by 2-sided test.
The mean age of the cohort was 48.2 (SD, 17.2) years, with 44.3% male and 2.7% of First Nations status. The majority of study participants (45.1%) were aged 18 to 44 years; the oldest age category (≥85 years) included 36 446 participants (2.0% of the overall cohort). Baseline characteristics of participants are shown by age in the Table. The prevalence of lower eGFR increased with increasing age, as did the prevalence of comorbidities and proteinuria.
Adjusted Rates by eGFR. During a median follow-up of 4.4 years (interquartile range, 2.7-5.6 years), 97 451 (5.36%) of cohort participants died, 3295 (0.18%) developed treated kidney failure, and 3116 (0.17%) developed untreated kidney failure.
Within each age stratum, adjusted mortality rates were higher for those with an eGFR of 90 mL/min/1.73 m2 or higher compared with those in eGFR strata of 60 to 89 mL/min/1.73 m2 (Figure 1), but rates generally increased with lower levels of eGFR. Within each age stratum, adjusted rates of treated and untreated kidney failure during follow-up were higher with lower baseline eGFRs (Figure 2).
Adjusted Rates by Age. Within each eGFR stratum, adjusted rates of death increased with increasing age. For example, in the eGFR stratum of 15 to 29 mL/min/1.73 m2, rates of death were almost 11-fold higher for participants aged 85 years or older compared with those aged 18 to 44 years (adjusted rate, 131.93 [95% CI, 116.62-149.27] vs 12.07 [95% CI, 4.69-31.06] per 1000 person-years, respectively; P < .001) (Figure 1). The increased risk of death associated with lower eGFR was higher in the younger age strata (P <.001 for interaction).
The risk of treated kidney failure increased with declining eGFR but was attenuated among older participants (P <.001 for interaction) (Figure 2A). Within each eGFR stratum, rates of treated kidney failure were consistently higher among the youngest age group. The adjusted rate per 1000 person-years of treated kidney failure among those with a baseline eGFR of 15 to 29 mL/min/1.73 m2 was 24.00 (95% CI, 14.78-38.97) among participants aged 18 to 44 years but was 1.53 (95% CI, 0.59-3.99) among those aged 85 years or older (P < .001).
The opposite results were evident for untreated kidney failure (Figure 2B). The risk of untreated kidney failure increased with lower vs higher eGFR categories, and this association was stronger with increasing age (P <.001 for interaction). For the lowest eGFR stratum (15-29 mL/min/1.73 m2), adjusted rates of untreated kidney failure were more than 5-fold higher among the oldest age stratum (≥85 years) compared with the youngest age stratum (18-44 years) (adjusted rate, 19.95 [95% CI, 15.79-25.19] vs 3.53 [95% CI, 1.56-8.01] per 1000 person-years, respectively; P < .001).
Adjusted Rates of Composite Outcome (Treated and Untreated Kidney Failure). Adjusted rates of the composite outcome (treated and untreated kidney failure) declined with increasing age and lower vs higher eGFR categories (P <.001 for interaction). However, these adjusted rates of the composite outcome of kidney failure demonstrated less variation across age strata (Figure 3) than observed when each outcome was considered separately. For example, the adjusted rate per 1000 person-years of the composite kidney failure outcome for those with an eGFR of 15 to 29 mL/min/1.73 m2 was 36.45 (95% CI, 24.46-54.32) among participants aged 18 to 44 years and 20.19 (95% CI, 15.27-26.69) among those aged 85 years or older (P = .01).
Adjusted Rates of Kidney Failure Accounting for the Competing Risk of Death. In analyses that formally accounted for the competing risk of death, similar patterns in the cumulative incidence of treated and untreated kidney failure by age were
evident (eFigure 1). Using a common reference (age 45-54 years and an eGFR of 60-89 mL/min/1.73 m2), the risk of treated kidney failure increased with declining eGFR but was attenuated among older
participants (eFigure 1, top). The opposite pattern of results was evident for untreated kidney failure (eFigure 1, bottom), with the highest hazard ratios in older age groups, in particular for those aged 65 years or older with an eGFR of less than 30 mL/min/1.73 m2.
We found similar results in sensitivity analyses that defined untreated kidney failure as progression to a sustained eGFR of less than 10 mL/min/1.73 m2 (eFigure 2) or alternatively by at least 2 follow-up measures of eGFR of less than 15 mL/min/1.73 m2 during a 6-month period (eFigure 3), or when the outcome was defined as an eGFR of less than 15 mL/min/1.73 m2 irrespective of renal replacement therapy (eFigure 4), with an attenuation in risk estimates. In all of these analyses, rates of treated kidney failure were lower (and untreated kidney failure higher) in older participants compared with those who were younger. The percentage of participants experiencing untreated kidney failure (and the composite of treated or untreated kidney failure) during the study period progressively increased with age, with the highest proportion observed in those aged 85 years or older (Figure 4).
In this population-based cohort of more than 1.8 million adults, the rate of progression to treated and untreated kidney failure differed substantially by age. Both the absolute and relative risks of treated kidney failure were highest among younger participants. In contrast, the rate of progression to an eGFR of less than 15 mL/min/1.73 m2 without dialysis or kidney transplantation (untreated kidney failure) was considerably higher among older people. Compared with either treated or untreated kidney failure considered alone, the incidence of the composite outcome of treated and untreated kidney failure was less variable in younger and older participants within each eGFR stratum. Results were robust for varying definitions of untreated kidney failure and when death as a competing risk was taken into account. These findings suggest that kidney disease does progress in older adults and that the true incidence of progressive CKD and kidney failure in older adults is underestimated when defined by receipt of renal replacement therapy (treated kidney failure) alone.
Possible explanations for the lower incidence of ESRD among older CKD patients (compared with otherwise similar younger patients) provided in earlier studies include a greater competing risk of death, slower kidney function loss, and lower uptake of dialysis in older adults.5,6 Our study adds important new information regarding the risk of progressive CKD in elderly persons by taking into account the competing risk of death and by also considering untreated kidney failure—which is particularly relevant given the potential for age to influence the decision to initiate dialysis.22
Most importantly, our results suggest that the incidence of advanced kidney disease in the elderly may be substantially underestimated by rates of treated kidney failure alone and that untreated kidney failure may be more common than initiation of renal replacement at older ages. Untreated kidney failure among adults aged 75 years or older with baseline eGFR of 15 to 29 mL/min/1.73 m2 was approximately 2- to 10-fold more common than kidney failure treated by dialysis. These findings have important implications for clinical practice and decision making; coupled with the finding that many older adults with advanced CKD are not adequately prepared for dialysis,23 these results suggest a need to prioritize the assessment and recognition of CKD progression among older adults. Our findings also imply that clinicians should offer dialysis to older adults who are likely to benefit from it—and should offer a positive alternative to dialysis in the form of conservative management (including end-of-life care when appropriate) for patients who are unlikely to benefit from (or prefer not to receive) long-term dialysis. Given the large number of older adults with severe CKD, these results also highlight the need for more proactive identification of older adults with CKD, assessment of their symptom burden, and development of appropriate management strategies.24,25 Finally, our study demonstrates the need to better understand the clinical significance of untreated kidney failure, the factors that influence dialysis initiation decisions in older adults, and the importance of a shared decision making process for older adults with advanced CKD.26
Rates of incident kidney failure in the United States are 2- to 3-fold higher than those reported in other developed countries27 and are based on receipt of renal replacement therapy. The hypothesis that rates of treated kidney failure are influenced by physician and patient preferences is further supported by the substantial regional variation in rates of treated ESRD in older American adults,23 which suggests that treated kidney failure reflects factors other than disease severity, including situational factors such as country and physician preference. The differences in the likelihood of initiating dialysis by age as illustrated in our study may further explain the between-country differences in ESRD incidence and warrant further exploration.
We defined untreated kidney failure as an eGFR of less than 15 mL/min/1.73 m2 based on current guidelines.15 Given that approximately half of incident North American dialysis patients initiate treatment at an eGFR of 10 mL/min/1.73 m2 or higher,28,29 this definition appears appropriate. We obtained similar results in sensitivity analyses that defined untreated kidney failure by more severe kidney dysfunction (eGFR <10 mL/min/1.73 m2) and by repeated measurements over a period of up to 6 months, further supporting the validity of our primary definition.
Furthermore, our results suggest that the increased mortality associated with CKD is observed at all ages, even among the very old (aged ≥85 years). Similar to a recent meta-analysis of general population studies,30 our results show a fairly consistent relationship between reduced eGFR and increased risk of death in both older and younger adults. The excess mortality associated with an eGFR of 90 mL/min/1.73m2 or higher, consistent for all age groups and previously reported,30,31 is likely due to the relationship of reduced muscle mass (reflecting ill health) to kidney function and mortality.
Our study has limitations that should be considered when interpreting its results. First, we could not assess the reasons why members of our cohort did not initiate dialysis despite having a low eGFR. It is possible that an eGFR of less than 15 mL/min/1.73 m2 is less likely to result in uremic symptoms in older adults (compared with younger patients), which may contribute to higher rates of untreated kidney failure in the former. However, we obtained similar results when kidney failure was defined as an eGFR of less than 10 mL/min/1.73 m2 (a level at which the majority of patients experience uremic symptoms32); thus, it is possible that other factors (including access to therapy, patient or physician preferences, or absence of symptoms) may explain the observed differences in the incidence of untreated kidney failure by age. Second, our cohort was limited to adults who had at least 1 serum creatinine measurement as part of their medical care in Alberta (with baseline kidney function defined by their first measurement) and, as such, may include a greater proportion of adults with underlying comorbidity. However, given the large study size and the publicly funded health care system in Canada, our results are likely generalizable to other settings. Third, we cannot comment on the long-term prognosis for patients with treated and untreated renal failure due to insufficient follow-up. While country-specific treatment practices for ESRD in older adults may exist, similar to regional variation within other countries such as the United States,23 such differences would not affect the conclusions drawn from our study. Finally, since we excluded people who died or initiated dialysis on the same day as their initial serum creatinine measurement, it is possible that results would have differed if kidney function was classified using values obtained by a population-based survey. However the relatively small number of such participants suggests that this decision is unlikely to have influenced our conclusions. Our study also has several strengths, including its community-based setting in a population with universal access to health care, including long-term dialysis. Furthermore, the large study size permitted a refined characterization of the adjusted association among age, eGFR, and risk of kidney failure, including a large number of individuals aged 85 years or older.
In conclusion, we found that the incidence of advanced kidney disease in elderly persons may be more substantial than previously thought—as reflected by markedly increased rates of untreated kidney failure among the oldest patients. Future studies are needed to better understand dialysis initiation practices in older adults with advanced kidney disease and to determine how to optimize shared decision making in this setting.
Corresponding Author: Brenda R. Hemmelgarn, MD, PhD, Division of Nephrology, Foothills Medical Centre, 1403 29th St NW, Calgary, AB T2N 2T9, Canada (firstname.lastname@example.org).
Author Contributions: Dr Hemmelgarn had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Hemmelgarn, James, Manns, Quinn, Tonelli.
Acquisition of data: Hemmelgarn, James, Tonelli.
Analysis and interpretation of data: Hemmelgarn, James, O’Hare, Muntner, Ravani, Quinn, Turin, Tan, Tonelli.
Drafting of the manuscript: Hemmelgarn.
Critical revision of the manuscript for important intellectual content: Hemmelgarn, James, Manns, O’Hare, Muntner, Ravani, Quinn, Turin, Tan, Tonelli.
Statistical analysis: Hemmelgarn, James, Muntner, Ravani, Turin, Tan.
Obtained funding: Hemmelgarn, James, Manns, Tonelli.
Administrative, technical, or material support: Hemmelgarn, James, Muntner, Tonelli.
Study supervision: Hemmelgarn, Tonelli.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr O’Hare reports receiving royalties from UpToDate. Dr Muntner reports consulting for and receiving salary support through a grant to his institution by Amgen. No other disclosures were reported.
Funding/Support: This work was supported by the Canadian Institute of Health Research (CIHR) and by an interdisciplinary team grant from Alberta Innovates–Health Solutions (AI-HS). Drs Hemmelgarn, Manns, and Tonelli were supported by career salary awards from AI-HS. Dr Hemmelgarn was supported by the Roy and Vi Baay Chair in Kidney Research and Dr Tonelli was supported by a Government of Canada Research Chair. Dr Turin is supported by Fellowship Awards from CIHR and the Canadian Diabetes Association.
Role of the Sponsor: The funding organizations played no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; or preparation, review, or approval of the manuscript.
A complete list of the Alberta Kidney Disease Network has been previously published (BMC Nephrol. 2009;10:30).