Daniel M. Keller, PhD

November 03, 2012

San Diego, California — Low estimated glomerular filtration rates (eGFR) and high levels of albuminuria are independently associated with mortality and end-stage renal disease (ESRD) at all ages across a variety of populations, Josef Coresh, MD, PhD, professor of epidemiology, medicine, and biostatistics at Johns Hopkins University in Baltimore, Maryland, reported here at Kidney Week 2012 for the Chronic Kidney Disease–Prognosis Consortium. He said the consortium found higher absolute but lower relative risk differences for mortality at older ages. The work was published in JAMA to coincide with its presentation here.

An impetus for the study was a controversy about whether chronic kidney disease, which is very common at older age, is related to substantial risk.

"We tried to definitively look at how risk of mortality, end-stage renal disease, and cardiovascular disease varied by markers of kidney function across the entire age spectrum from age 18 to over 100 [years]," Dr. Coresh commented to Medscape Medical News.

To assess the possible interaction by age between eGFR and albuminuria with clinical risk, investigators performed a meta-analysis of individual-level data from 2,051,244 participants from 33 general or cardiovascular high-risk populations and 13 chronic kidney disease (CKD) cohorts from around the world. Data were gathered between 1972 and 2011 with a mean follow-up time of 5.8 years (range, 0 - 31 years).

Information was available on demographic and cardiovascular risk factors for all participants. Age ranges were categorized as 18 to 54 years, 55 to 64 years, 65 to 74 years, and 75 years and older. Primary endpoints were all-cause mortality and ESRD. Interactions were expressed as hazard ratios for eGFR compared with age category 55 to 64 years and for albumin (albumin-to-creatinine ratio, ACR), to a reference point of 10 mg/g (20 mg/g in CKD cohorts). The reference eGFR was 80 mL/minute/1.73 m2. Hazard ratios were adjusted for sex, race (black vs nonblack), history of cardiovascular disease, diabetes, systolic blood pressure, serum total cholesterol, body mass index, smoking, albuminuria, and, for clinical trials, the randomized intervention.

Risks for mortality (n = 112,325 deaths) and ESRD (n = 8411 events) were higher with lower eGFR and higher albuminuria for every age category. For both general and high-risk cohorts, the relative risk for death with decreased eGFR lessened as age increased, but the absolute risk for death increased with increasing age. In other words, reduced eGFR had a diminishing relative effect for older ages, although older people were at an overall higher risk for death.

Table 1. Mortality Effect of eGFR for Excess Deaths per 1000 Patient-Years*

Adjusted Age Categories Hazard Ratio 95% Confidence Interval
18 - 54 years 3.50 2.55 - 4.81
55 - 64 years 2.21 2.02 - 2.41
65 - 74 years 1.59 1.42 - 1.77
75 years or older 1.35 1.23 - 1.48
Adjusted Age Categories Absolute Risk Ratio 95% Confidence Interval
18 - 54 years 9.0 6.0 - 12.8
55 - 64 years 12.2 10.3 - 14.3
65 - 74 years 13.3 9.0 - 18.6
75 years or older 27.2 13.5 - 45.5

*eGFR of 45 mL/minute/1.73 m2 vs 80 mL/minute/1.73 m2; P < .05 for age interaction.

In contrast to the effect of eGFR, the relative effect of increased albuminuria did not diminish as much with age, that is, increased albuminuria still exerted a rather strong effect even in the older age categories.

Differences in absolute risk were higher in the older age categories (excess deaths per 1000 patient-years) at an ACR of 300 mg/g vs 10 mg/g.

Table 2. Absolute Mortality Risk for ACR of 300 mg/g vs 10 mg/g for Excess Deaths per 1000 Patient-Years

Adjusted Age Categories Absolute Risk Ratio 95% Confidence Interval
18 - 54 years 7.5 4.3 - 11.9
55 - 64 years 12.2 7.9 - 17.6
65 - 74 years 22.7 15.3 - 31.6
75 years or older 34.3 19.5 - 52.4

In the CKD cohorts, which investigators analyzed separately from the general ones, adjusted relative hazards of mortality did not decrease with age. "Their findings were quite consistent," Dr. Coresh said. "Risk was still present at all ages, but it was even stronger at older age in the CKD cohorts."

In all cohorts, the relative risks for ESRD and the absolute risk differences for ESRD at lower eGFR or higher ACR were similar across age categories.

In summary, Dr. Coresh said, "We find that on a relative risk basis in the general population, the youngest people have the highest relative risk of mortality, and the oldest people have the lowest relative risk of mortality. But in terms of attributable risk, the oldest people have the highest attributable risk. So they have the lower relative risk but higher excess risk associated with low GFR.

"For proteinuria, the attenuation is much smaller, so older people have just a subtly lower relative risk but a markedly higher excess risk," he said. "And then when we look separately in the kidney disease cohorts, really we get an association of low GFR and albuminuria at all age groups with end-stage renal disease and with mortality and end-stage renal disease in the kidney cohorts."

Dr. Coresh said that using current definitions of kidney disease, approximately half of the population older than 75 years has chronic kidney disease. "I don't think we can ignore it.… Most of those people have a milder form of disease, and probably the best thing we can do is make sure that when we use medications, when we do interventions, we're aware that those guys have issues and could be harmed more than somebody else."

He added that people with more severe kidney disease are at risk for dialysis, and at an older age that may be handled differently than at a younger age. "So there's something to be done at every age.... The plan needs to be tailored for age, but chronic kidney disease is a risk factor at every age," he emphasized.

Ian de Boer, MD, associate professor of medicine in the Division of Nephrology and the Kidney Research Institute and adjunct associate professor of epidemiology at the University of Washington in Seattle, who wrote an editorial to accompany the published study, commented to Medscape Medical News, "This report confirms the strong association of chronic kidney disease with death in older adults, and its novel contribution is its ability to evaluate that association across a very large age range and demonstrate that the association is as strong or stronger in older adults than it is in middle age.... [It] clearly comes down on the side of chronic kidney disease being important in that population."

So a large proportion of older adults are living with chronic kidney disease as evidenced by albuminuria or low eGFR. But "to be honest, we don't know the best way of treating people who are older with chronic kidney disease," Dr. de Boer said. Some proven treatments are lowering blood pressure, using ACE [angiotensin-converting enzyme] inhibitors or angiotensin receptor blockers for people with proteinuric kidney disease, and lowering lipids with statins. But "I think we are in need of new therapies to safely and effectively target this particular population," he said.

For physicians treating individual patients, "I think vigilance is the message," Dr. de Boer advised. "I think from a public health standpoint, this article makes it clear that further refining treatment strategies that exist now and developing new treatment strategies for people who are older with chronic kidney disease is a high priority."

The Chronic Kidney Disease–Prognosis Consortium (CKD-PC) Data Coordinating Center has received a program grant from the US National Kidney Foundation (NKF), which receives funding from Abbott and Amgen. Other CKD-PC funding comes from a variety of sources, including government agencies, foundations, and industry sponsors (listed in eAppendix 3 of the paper). Dr. Coresh reports grant support to his institution from the NKF and Amgen, and he has received consultancy fees from Amgen and Merck. The study includes data from multiple cohorts, including organizations Dr. de Boer works with, but he was not directly involved in the analysis of those data presented in the study. Dr. de Boer has disclosed no other relevant financial relationships.

Kidney Week 2012: Abstract TH-PO329. Presented November 1, 2012.

JAMA article. Editorial.