Are CKD Definitions Relevant in Older Adults?

Hogan, Michelle

doi: 10.1097/
National Kidney Foundation Spring Clinical Meetings

NATIONAL HARBOR, MD—Older patients make up the largest segment of the population defined as having chronic kidney disease (CKD) and, in some ways, the most controversial. While the condition is pervasive in this group, disease progression, at least until recently, has been reported to be less common in older than in younger patients, prompting concerns about potential overdiagnosis of CKD in older individuals.

In a debate here at the National Kidney Foundation (NKF) 2012 Spring Clinical Meetings, Michael G. Shlipak, MD, MPH, Professor of Medicine, University of California, San Francisco, and San Francisco VA Medical Center, and Richard J. Glassock, MD, Professor Emeritus, David Geffen School of Medicine at UCLA, strove to bring clarity to the controversy.

The 2002 NKF Kidney Disease Outcomes Quality Initiative (KDOQI) guidelines defined chronic kidney disease as kidney damage or glomerular filtration rate (GFR) less than 60 mL/min/1.73 m2 for at least three months, with the condition categorized into five GFR-based stages.

Kidney Disease: Improving Global Outcomes (KDIGO) has proposed to keep that definition but update the classification by emphasizing cause, if known; adding albuminuria stages; and splitting CKD Stage 3 in two, with one subclassification including GFR 30–44 mL/min/1.73 m2, and the other 45–59 mL/min/1.73 m2. This consensus came out of a 2009 KDIGO conference in London, during which a meta-analysis of 1.5 million people from 43 cohorts revealed a strong and consistent relationship of risk with estimated GFR and urine albumin-to-creatinine ratio.

At the NKF meeting, Dr. Shlipak took the side that the current and proposed definitions make sense in older adults, while Dr. Glassock argued that they do not.

Back to Top | Article Outline


Three main points underpinned Dr. Shlipak's support for the application of CKD definitions to older adults:

“First, I think we need to separate issues of measurement of GFR from actual biology of CKD,” he said during his presentation at the conference.

“Second, I'm going to argue that CKD is a disease, not just a risk factor, and, third, I'm going to argue that CKD can be a disease in the elderly, even if everybody has it.”

The relationship between estimated GFR and mortality is attenuated in elderly patients, Dr. Shlipak noted, referring to a study led by Ann O'Hare, MD, now of the VA Puget Sound Health Care System and the University of Washington (J Am Soc Nephrol 2006;17:846–853).

“So why does the estimated GFR–mortality association weaken progressively with age?” he said. “I think there are basically two possibilities.

“First, GFR just becomes less important in elderly patients because of all their other comorbidities, and so, proportionally, it's not that important, or, alternatively, our measurements of estimated GFR are progressively imprecise as people get older.

“I'm going to argue that it's the second, and I'm going to argue that GFR is actually at least as important in elderly people as in younger.”

On average, creatinine production is lower in older people, women, individuals with chronic illness, and vegetarians, and it's higher in African-Americans and in people who have more muscle or are more active. Only race, sex, and age are accounted for in GFR-estimating equations.

“So we really need a better kidney function marker to use in elderly persons because the measurement issue is clouding the entire story,” Dr. Shlipak said. “I think cystatin C is a better marker for elderly persons.”

Unlike creatinine, cystatin C is produced in all nucleated cells, is constantly released into the blood, is freely filtered, does not undergo tubular secretion, and is not affected by muscle mass.

“Creatinine is not adequate to categorize mild CKD, 45–60, in older adults,” he said. “I think we should use cystatin C to confirm CKD in older patients with eGFR 45–60, and that will distinguish the high-risk CKD patients that we care about. This suggestion is actually a suggested practice that's being debated as part of the new KDIGO guidelines.”

Back to Top | Article Outline

Disease & Risk Factor

Chronic kidney disease is both a risk factor and a disease, Dr. Shlipak said. Elderly people with CKD have worse health status as defined by frailty, functional impairment, decreased mobility and strength, and higher infection risk.

“Although clearly a risk factor, I think CKD—a GFR less than 60—is also associated with signs and symptoms consistent with disease, and most of the data on CKD's impact on health came from elderly cohorts.”

CKD's designation as a disease is not negated by its prevalence in the elderly, he added.

“Kidney function, we know, declines steadily with age, even in the absence of any risk factors,” Dr. Shlipak said. “So kidney disease is expected, and decline is expected, but is it normal?”

He drew an analogy between the kidney and the brain.

“Underlying kidney and brain pathologic changes may not actually be benign, even when the function appears normal. The function is maybe adequate under stable conditions, but the organs lack reserve.

“So, for example, during hospitalization, the brain will go into delirium; the kidney has AKI, and so I would argue that age-related changes in the brain and kidney are not normal, even though they're quite usual and typical.”

Back to Top | Article Outline


While the current and proposed CKD definitions are not entirely applicable to older patients, they represent a significant advance, Dr. Glassock noted.

“The original CKD KDOQI classification, which emerged in 2002, I will not only admit but enthusiastically endorse as one of the great ideas in contemporary nephrology. There is no doubt that this was a transformative event.

“The original KDOQI schema was based not on data but some assumptions, and it defines CKD on the basis of a departure from normality, using eGFR and kidney damage, and it staged CKD based on the degree of this eparture from normality, emphasizing eGFR.

“Now, 10 years later, we're going to, in a sense, redefine CKD based on the risk of an association with a variety of adverse events; of course, for nephrology, all-cause mortality and ESRD are dominant issues. And we're going to stage CKD based on the magnitude of this risk using two different nonoverlapping complementary biomarkers, eGFR and albuminuria.”

Despite their advantages, the CKD frameworks have shortcomings when applied to elderly individuals, Dr. Glassock said.

“The use of an absolute threshold of eGFR, 60 mL/min or less, even in the absence of any signs of kidney damage—this is true in both schema—leads to overdiagnosis of CKD in the older and elderly adult, and I will try to indicate to you that there is a reasonable argument that GFR stratification by age should have been incorporated into the design, and perhaps an eGFR of 45 mL/min or less might have been a better absolute threshold for determining the presence of CKD in the elderly.”

The choice of the 60-mL/min/ 1.73 m2 threshold in the original guideline was based on measured glomerular filtration rate in the younger population, he noted.

“The KDOQI 2002 classification criteria arbitrarily selected this threshold of 60 mL/min based on the presumption that the average young adult had a measured inulin clearance GFR of 120 mL/min and that a 50 percent loss of this measured GFR would equal the GFR of 60.

“Well, that's all well and good, but eGFR by the MDRD [Modification of Diet in Renal Disease] formula, which was available in 2002, was known to be inaccurate at a measured GFR of 60 and underestimated measured GFR by as much as 8–10 mL/min/1.73 m2, and if you accept this notion, if you're going to use an estimate of GFR to classify patients with CKD—you're not going to use the measured GFR—then the threshold might have been set more appropriately at 55 rather than 60.”

MDRD-estimated glomerular filtration rate declines with older age, Dr. Glassock said.

“Above 60 mL/min/1.73 m2, these estimating formulas lose considerable accuracy. In fact, the MDRD formula is so imprecise that it cannot reliably distinguish CKD Stage 1 from 2 CKD, and if one is in a practical sense going to use these formulas to stage CKD, which is what one does in practice, you cannot reliably distinguish Stage 1 and Stage 2 in either formula.”

Compared with the MDRD Study equation, the newer CKD-Epidemiology Collaboration (CKD-EPI) formula decreases the prevalence of CKD in males and females under 70 but increases prevalence in the elderly.

“Simply moving from one formula to another, one significantly alters the frequency of CKD in the population, but the changes in that frequency differ by age.”

Back to Top | Article Outline

Microalbuminuria, CVD

In the absence of any defined decrease in eGFR or any other sign of kidney damage, microalbuminuria leads to overdiagnosis of CKD in the older and elderly adult, Dr. Glassock said, making his second major point.

There is no doubt that rising levels of albumin in the urine enhance the risk of all-cause and cardiovascular mortality at all levels of glomerular filtration rate, he said. However, it's also important to note that both urinary albumin excretion and creatinine excretion contribute to risk, but in opposite directions.

“The question I'd like to pose is, does an increase in albumin excretion rate that is determined by the urinary albumin-to-creatinine ratio in an isolated fashion, without any other manifestation of kidney damage and with a normal eGFR, reflect intrinsic kidney disease, or is it simply a regional manifestation of a more generalized disturbance in endothelial function—that is, it's a marker, not a maker of kidney disease. So should we use it as a tool to define a disease—kidney disease?”

Dr. Glassock concluded his presentation by talking about the relationship of eGFR with cardiovascular disease and mortality.

“Until the eGFR gets less than 45, there is no, no evidence of an increased hazard ratio for all-cause mortality using the best available creatinine formula, CKD-EPI,” he said, referring to a study of 8,705 community-dwelling adults age 65 or older that was led by Benedicte Stengel, INSERM (French National Institute for Health and Medical Research; Nephrol Dial Transplantation 2011;26:3286–3295).

In addition, it has been demonstrated that the predictive power of the Framingham Risk Score for cardiovascular disease is not improved by the addition of eGFR or albuminuria to the mix.

“Most of the excess cardiovascular risk seen in older adults can be captured by the assessment of standard risk factors, such as the Framingham Risk Score, and non-GFR determinants contained within many of the eGFR formulas contribute significantly to their ability for prediction of cardiovascular events,” Dr. Glassock said.

Back to Top | Article Outline

Common Ground

Both Dr. Shlipak and Dr. Glassock agreed that estimating glomerular filtration rate (GFR) by creatinine is too imprecise for the definition of chronic kidney disease in the elderly, particularly in the GFR range of 45–60 mL/min/1.73 m2. Their proposed solutions differed, however, with Dr. Shlipak supporting the use of cystatin C as a confirmatory test, and Dr. Glassock advocating that the GFR threshold be lowered or tailored to age.

© 2012 Lippincott Williams & Wilkins, Inc.