A study at the San Francisco VA Medical Center (SFVAMC) has found that a test of kidney function that measures blood levels of cystatin C—a protein produced by most cells in the body—is a far more accurate predictor of mortality risk in elderly people than the current standard kidney function test, which measures levels of the protein creatinine.
In the study, led by Michael Shlipak, MD, chief of the Department of General Internal Medicine at SFVAMC, the researchers determined that the level of serum cystatin was much more accurate than serum creatinine in predicting risk of death from all causes, and from cardiovascular disease in particular. The results of the study will be published in the May 19 issue of the New England Journal of Medicine.
The researchers measured levels of cystatin and creatinine in blood samples taken from 4,637 participants in the Cardiovascular Health Study, a national, long-term, longitudinal study of elderly people sponsored by the National Heart, Blood and Lung Institute. They found that, for all participants, the higher the level of cystatin, the higher the risk of mortality.
In comparison, the test for creatinine revealed the mortality risk only for those participants whose creatinine levels were in the top 10 percent of those measured. “We were astonished by the difference in the results,” says Shlipak, who is also an associate professor of medicine, epidemiology, and biostatistics at the University of California, San Francisco (UCSF).
Cystatin proved even more accurate at measuring risk of death from cardiovascular causes: Participants with the highest levels of cystatin were 700 percent more likely to die of cardiovascular causes than those with the lowest levels, compared with a 500 percent risk differential for death from all causes. “These are just incredible gradients of risk,” notes Shlipak. “There are very few risk factors this predictive that we see in clinical research.”
In comparison, only the subgroup with the very highest levels of creatinine was identified as being at greater cardiovascular risk.
The researchers believe the reason for the difference in accuracy is in the nature of the two proteins. Both the cystatin test and the creatinine test measure how efficiently the kidneys filter those respective proteins from the blood; the higher the protein level, the less efficient the kidney function. However, Shlipak observes, “creatinine is produced by muscle, so the levels in the body are largely determined by how much muscle mass you have.” Elderly people have less muscle mass than younger people; women tend to have less muscle mass than men; and African Americans tend to have more muscle mass than Caucasians, he says. “So to take creatinine and attempt to figure out renal function, you have to integrate all these different parameters.”
In elderly people, he points out, low creatinine levels may simply reflect lower muscle mass, regardless of kidney health.
Cystatin C, by contrast, seems to be independent of muscle mass, age, gender, and race, so its blood levels “appear almost entirely driven by kidney function,” says Shlipak. The result is a much more precise measure of kidney dysfunction—which, as the study points out, is associated in elderly persons with an increased risk of death.
The researchers found that, measured by cystatin levels, participants fell into three broad groups. Twenty percent were in a high-risk group that had an overall mortality risk of about 9 percent per year; 40 percent fell into a medium-risk group with a mortality risk of about 4 percent per year; and another 40 percent had a lower-than-average mortality risk of about 2 percent per year.
Shlipak points out that only the low-risk group could be considered to have normal kidney function: “The medium-risk group has kidney disease. It’s just never been detected before. It wouldn’t have been detected by creatinine.” This means, he says, that cystatin has real potential as a tool for revealing kidney dysfunction before it becomes symptomatic, particularly among the elderly.
In clinical settings, Shlipak believes measuring cystatin would be most useful in assessing preoperative risk and the risk of side effects from medication and from dyes used in radiology procedures. Currently, creatinine is used as a measure in those situations, “and clearly, that’s just not a good enough test in an older population.”
Shlipak cautions that it has not yet been proven beyond doubt that cystatin C itself does not contribute to risk of mortality in some unknown way; however, “all of our research suggests that it’s simply a measure of kidney function, and that’s why it has such a strong role in prognosis.”
To obtain their results, the researchers analyzed blood samples taken, and subsequently frozen, in 1992 and 1993 from the 4,637 participating men and women who were 65 or older at the study’s inception. Follow-up visits with the participants were conducted over the telephone every six months and in person annually until June 30, 2001, and cause of death was determined for all who died. Creatinine was measured around the time the samples were drawn; the researchers measured cystatin in 2003 from the frozen samples. To ensure the accuracy of the values obtained from the frozen samples, the researchers confirmed that a test sample showed no change in values measured over a course of five cycles of freezing and thawing.
In future research, Shlipak would like to conduct a clinical trial to determine if cystatin C is truly effective as a diagnostic tool. “We need to prove that before we can recommend that people use it.”
Additional authors of the study are Mark J. Sarnak, MD, Tufts-New England Medical Center, Boston; Ronit Katz, PhD, Collaborative Health Studies Coordinating Center, Seattle; Linda F. Fried, MD, PhD, Veterans Affairs Pittsburgh Healthcare System, Pittsburgh; Stephen L. Seliger, MD, University of Maryland School of Medicine, Baltimore; Anne B. Newman, MD, MPH, University of Pittsburgh; David S. Siscovick, MD, MPH, University of Washington, Seattle; and Catherine Stehman-Breen, MD, of Amgen, Thousand Oaks, California.
The research was funded by a grant from the National Heart, Lung, and Blood Institute that was administered by the Northern California Institute for Research and Education (NCIRE); by an award from the American Federation for Aging Research and National Institute on Aging Paul Beeson Scholars Program, also administered by NCIRE; by the Robert Wood Johnson Foundation Generalist Faculty Scholars Program; and by awards from the National Institute of Diabetes and Digestive and Kidney Diseases and the Department of Veterans Affairs Office of Research and Development.