September 2012
Feature Story

Anne Paxton

Like Ruby Keeler in “42nd Street,” the kidney function biomarker cystatin C has taken its turn waiting in the wings. Despite the fact that in studies of kidney disease prognosis, cystatin C has been consistently shown to be a better marker than creatinine, it has not been part of customary equations for estimating GFR (glomerular filtration rate), used to diagnose chronic kidney disease. But a new study suggests that cystatin C’s moment may have arrived—if not for stardom, at least for equal billing with creatinine in estimating GFR and detecting renal disease.

The study, “Estimating glomerular filtration rate from serum creatinine and cystatin C,” published July 5 in the New England Journal of Medicine (2012;367:20–29), was conducted by the Chronic Kidney Disease Epidemiology Collaboration with funding from the National Institute of Diabetes and Digestive and Kidney Diseases. The researchers developed estimating equations based on cystatin C alone and in combination with creatinine in 5,352 participants from 13 studies. They then validated the equations in 1,119 participants from five other studies in which GFR had been measured. While the authors concluded that cystatin C may not replace creatinine in general practice, they found that the creatinine-cystatin C equation performed better than equations that used creatinine or cystatin C alone.

The advantage of the combined equation is that it’s more precise than creatinine, says the lead study author Lesley A. Inker, MD, MS, associate professor of medicine, Tufts University School of Medicine. “I wouldn’t say that the creatinine-cystatin C combination is going to replace creatinine for now, given creatinine’s worldwide use, but it certainly is a step toward more precise measurements. Together, cystatin C and creatinine give you a fuller picture of what is the GFR.”

That’s a finding that is welcome to Michael G. Shlipak, MD, MPH, chief, Division of General Internal Medicine, San Francisco VA Medical Center, who has been studying outcomes of kidney disease for 15 years, and cystatin C specifically since 2003 (he wasn’t involved in the NEJM study). And he believes clinical laboratories should take heed. With the number of chronic kidney disease cases in the U.S. estimated to exceed 26 million, clinical laboratories that do not offer cystatin C would do well to consider adding it to their test menu, Dr. Shlipak says. “This study is the end of the era of excuses for not offering cystatin C.”

Andrew S. Narva, MD, director of the National Kidney Disease Education Program, National Institutes of Health, which has been focused for some time on getting more clinicians to regularly use an eGFR equation as opposed to just serum creatinine, views the new cystatin C findings as a step forward, though he points to the broader context. “For the purposes of estimating the risk of CKD, this study adds to our toolbox. But for educating clinicians, I think it’s important for them to understand the uncertainty associated with all of these GFR-estimating methods. Even though a combined cystatin C and creatinine is the most accurate equation at this point for populations, for the individual you see in front of you, you still only have a 90 percent chance of being within 30 percent of the measured GFR. So it’s not exactly like reporting a calcium.” The lesson for clinicians, Dr. Narva says, is to use multiple indicators when assessing patients for CKD.

So far, there is no specific clinical guideline that combines cystatin C into an eGFR equation, says study co-author John H. Eckfeldt, MD, PhD, vice chair for clinical affairs in the Department of Laboratory Medicine and Pathology, University of Minnesota Medical Center. “There’s a lab working group that is part of the National Kidney Disease Education Program. They met in July to discuss whether cystatin C should be recommended for eGFR, and there was basically no decision either way. They decided to remain silent at this point—except to say that it may be useful in problematic cases where the creatinine-based equations are suspect, or there are clinical characteristics of the patient that make you suspicious the eGFR is not accurate.”

The hesitation at this point is due, in part, to the historic lack of calibration of the different cystatin C measurement procedures that have been available commercially. Until recently, there was no established reference material, and labs could not rely on getting the same answer when measuring the concentration of cystatin C. But in 2010, the Institute for Reference Materials and Measurements (IRMM), in collaboration with the International Federation of Clinical Chemistry (IFCC), released a reference material certified for cystatin C, to assign values to calibrators that are part of IVD medical devices.

The reference material issue “has really inhibited wide acceptance of the use of cystatin C in any sort of equation,” Dr. Eckfeldt says. “But I think this will change as manufacturers are now in the process of calibrating the various measurement procedures used in clinical labs to the IRMM reference material. Once the different clinical assays get standardized analytically, cystatin C measurement will be much more useful in the clinical arena.”

Dr. Eckfeldt believes the most useful role for cystatin C may be as a check on creatinine-based eGFR. “It’s a way to confirm that there’s not some problem with the production of creatinine in individual patients that is giving spurious values.” Cystatin C works as a check because the cystatin C-based equations are not strongly affected by age, sex, or race. “And race, in particular, is a problem, because many computer systems, such as ours, don’t include race in patient information. If you use cystatin C, you don’t seem to need to include race in the eGFR calculation.”

In fact, the study authors acknowledge that one of the study’s weaknesses is that it didn’t cover a variety of racial groups. “U.S. blacks and Caucasians are the dominant ones in developing these equations. They haven’t tested cystatin C-based equations in Chinese or Japanese or Southeast Asians,” Dr. Eckfeldt points out. “Creatinine also comes from meat, so it can reflect not only genetic differences in muscle mass and amount of skeletal muscle, but it may be dietarily related too. It’s sort of hoped that cystatin C is better in this regard, but I think the jury may still be out a little bit because there’s not a lot of data looking at different races.”

Dr. Narva’s concern about the time and effort required to educate clinicians on appropriate CKD testing is understandable, Dr. Eckfeldt says. “Clinicians are just starting to understand eGFR based on creatinine, and now you have to educate them on cystatin C, for what some consider a minimally incremental benefit.” But Dr. Eckfeldt himself believes that, eventually, it might make more sense to use a cystatin C-based-only equation rather than the combined equation. “Because if you think creatinine is spuriously high or low, for whatever reason, then why do you want it in the equation at all?”

Says Dr. Shlipak: “We have probably 2 million people who are being labeled CKD with creatinine but don’t have chronic kidney disease, and we have roughly 10 million people who actually have CKD but it’s missed by creatinine, and I think cystatin C can help correct that on both ends. I’d love to be able to say just skip the creatinine and go straight to cystatin C.”

But he thinks such a change won’t happen until far in the future. For the time being, he agrees the next step is to use cystatin C to confirm CKD with creatinine. “Then I think we should be using it to screen people with no CKD indicated on their creatinine test, if they’re really high risk—diabetics, people with cardiovascular disease or heart failure, people with HIV infection, and probably people with liver disease as well.”

Whether any of this happens is contingent on laboratory directors becoming familiar with cystatin C and making it available in their labs. “Currently nephrologists or clinicians say, ‘I can’t order cystatin C because it’s not available in the lab,’ and the lab says it’s not available because no one is asking for it. So it’s a Catch-22 where no progress is getting made,” Dr. Shlipak says. “Labs need to realize this is an easily automated test, it’s quite simple to add to their portfolio, and gradually physicians will understand and start using it and it will start reaching its potential.”

The potential cost, on the other hand, is an often-cited obstacle to any wide-scale embrace of cystatin C. CKD, at epidemic proportions in the U.S., already absorbs close to one-third of Medicare and Medicaid budgets. “If you replace all creatinines with cystatin Cs, that would be a huge expense to the medical system,” Dr. Eckfeldt says. The clinical charges vary quite a bit, but the cost has run $10 to $20, he says, compared with a few dollars for creatinine.

From his point of view, that money might be better spent on measuring albumin in the urine. “It’s pretty clear now that measuring albumin in the urine is probably as predictive or more predictive of risk of developing ESRD as well as cardiovascular disease. And if you’re going to spend money for laboratory testing, maybe it ought to be on measuring albumin in urine rather than measuring a different marker in the blood.”

It’s true, however, that albumin suffers some of the same issues as cystatin C, in that there is a lot of inter-method variability, he notes. “You have again the problem of different labs producing values 20 percent higher or lower than the true value. So there’s an NKDEP [National Kidney Disease Education Program] lab working subcommittee in conjunction with IFCC trying to develop a reference system and have the different clinical lab methods harmonized for measuring urine albumin.”

“I think we always have to consider cost,” Dr. Inker agrees. But she believes the cost of performing a cystatin C measurement will drop when more people order the tests. Moreover, relative to radiological tests, for example, “as a confirmatory test cystatin C is extremely inexpensive.” She does not believe it’s necessary to measure cystatin C in everyone. “One approach, which I think is a conservative approach, is to say in the creatinine range of around 60 where there’s some uncertainty and we really want to know the answer and avoid false-positives, the addition of cystatin C as a confirmatory test would help us get the right answer.”

Of course, a major virtue of eGFR based on creatinine is that since the assay is part of almost every routine biochemical panel, eGFR is essentially free, Dr. Narva says, and it can alert providers to the presence of CKD in a patient with “normal” creatinine.

“But cystatin C is not that commonly done and has to be ordered specifically.” The additional cost would have to be balanced against the improvement in precision before it could be applied routinely as the screening test for CKD, he says.

After the lack of standardization, the second biggest factor stalling the widespread use of cystatin C has been the need for a special instrument such as a nephelometer, says Dr. Shlipak. “That was something most labs didn’t have space for and didn’t want because it was expensive. But now you can measure cystatin C on basically any chemistry platform, just as you would do a thyroid or anything else. Any clinical laboratory can offer it.” He predicts that the cost will prove manageable, on the order of $2 to $5 per test. While most university-based hospitals and larger reference laboratories are measuring cystatin C, it’s not available in many smaller hospitals and clinical laboratories, Dr. Eckfeldt says. But he hopes that may change with the publication of the New England Journal article.

Cystatin C’s utility in estimating GFR has been known for quite a while, says Dr. Shlipak. “We had a paper in the NEJM in 2005 that showed cystatin C to be a much stronger predictor of mortality and other health outcomes in older people than creatinine. The differences were really striking. In distinguishing people who are really normal from those with mildly decreased function, cystatin C was far superior.” At the time, he adds, there was controversy about how closely cystatin C results were related to kidney function. “And part of that controversy is why, seven years later, most people still haven’t heard of cystatin C and we’re not using it that much.”

A common, though unintentional, bias in most studies, including this new one, might be skewing the results and making creatinine look a little better than it is, he suggests. “Studies that are able to get volunteers to submit to this GFR protocol tend to get very selected populations, mostly younger people with kidney disease. The researchers still find that cystatin C is helpful when added to creatinine, but my opinion is that the studies, though not intentionally, are sort of set up to make creatinine look as good as possible. The difference in result is small relative to what it actually would be in the general population.”

The number of false-positives and false-negatives in the detection of kidney disease is a substantial public health problem, Dr. Shlipak points out. “False-positive eGFRs are an easier problem to correct because they’re a smaller population. Those people can be reassured they can get their full dose of medicine in the case where doctors are under-dosing them, or it might affect their insurance status or general prognosis. But false-negatives are more serious. There are real, tangible problems with not detecting disease when there are potential treatments. With kidney disease, we’re a little short in the treatment area. But we know we can be a little stricter on blood pressure control, try to avoid agents that harm kidneys, and dose medications in a safer way. So in populations at pretty high risk for CKD, I think they should be screened with cystatin C so we know we’re not missing anything,” he says.

Cystatin C needs less manipulation through an equation, and the most important advantage is that it has no race bias, Dr. Shlipak says. “If you’re African American, your creatinine conversion to eGFR gets multiplied by 16 percent because, on average, African Americans have more muscle mass. But not all of them do. If you’re relatively thin, and are developing kidney disease, if you’re African American this equation says a GFR of 55 is 70 because it’s been raised by 16 percent. So you’re not going to be diagnosed with kidney disease for another four years.”

Dr. Shlipak is helping develop CKD guidelines from the international nephrology organization KD-IGO (Kidney Disease—Improving Global Outcomes), an umbrella organization headquartered in Brussels. The guidelines, expected to come out later this year, are not complete. “But I am hoping they will recommend cystatin C be used to confirm CKD. We really need to work on implementing cystatin C’s availability. Then—gradually, disease condition by disease condition—we can establish the appropriate use, particularly to detect CKD in patients at high risk.” There’s good consensus on cystatin C as a confirmatory test, “but to take the next step and start using it as a screening test, we have to get past a lot of misinformation about how expensive it would be. We need studies to find out the prevalence of ‘occult’ CKD which is missed by creatinine.”

Cystatin C is only very infrequently offered in U.S. health care settings, Dr. Shlipak notes. “Companies like Quest and LabCorp offer it as a send-out test, but local hospitals need to have it available because send-out tests are not useful for hospitalized patients.” But cystatin C is a fairly fast test, and “there’s no reason why it couldn’t be part of the same chemistry 10 or chemistry 20 panel.”

The Swedish clinical chemist who first described the amino acid sequence of cystatin C and in 1985 proposed it as a measure of GFR, Anders Grubb, MD, PhD, considers the New England Journal study, with which he was not involved, to be well performed and important. In Sweden, where the health care system and research environment are different, the use of cystatin C for detecting renal disease is far more established. Because variations in muscle mass do not affect cystatin C, adjustments for the patient’s race, an imperfect proxy for muscle mass, do not have to be made, says Dr. Grubb, professor of clinical chemistry at the University of Lund. “With cystatin C you do not really have to add patient factors to the equation,” he says.

He regards the use of these attributes to adjust creatinine results as ill-defined and ethically problematic. In Sweden, “we have been using eGFR equations without sex and age, and we’re not even allowed to use race. And all the comparisons between creatinine and cystatin C have shown that cystatin C is much less affected by these patient factors—if it is affected at all.”

The Swedish government has just completed an inquiry into how to use cystatin C and creatinine and expects to release recommendations shortly. But for now, Dr. Grubb says, “I can tell you that we have read all the papers about creatinine and cystatin C and both of them, and one unanimous conclusion is that using both is best.”

He is a strong advocate of cystatin C. “We believe the clinical importance of knowing the GFR of a patient is too high to refrain from using cystatin C in the initial estimation of the GFR of a patient. The problem with creatinine-based estimations of GFR is that they are often wrong without any suspicion by the physician that this is the case. So the physicians will often not know when they should ‘confirm’ a creatinine-based estimation.

“Knowledge of GFR is so vital for good health care that, at your first meeting with a patient, you should always measure both cystatin C and creatinine to get the best possible assessment of GFR. Then, if the price of the cystatin C test is high in your laboratory, you can use creatinine to follow the GFR of the patient with more confidence that the estimated GFR is not very wrong.”

If he could choose only one equation, he would choose a cystatin C-based equation. “The creatinine level alone is far from ideal as a GFR marker, since it is significantly influenced not only by GFR, but also by muscle mass, amount of ingested meat, and tubular secretion, varying with the GFR-level, with the time of day, and it’s also influenced by many medicines,” Dr. Grubb explains. “Creatinine-based GFR-prediction equations, including, in addition to the creatinine level, anthropometric factors like age, sex, and race, might partially compensate for the influence of muscle mass by considering the mean muscle mass of a person of a specified age, sex, and race. But such equations cannot compensate for differences in creatinine levels due to amount of ingested meat, varying tubular secretion of creatinine, and deviation of the muscle mass of a specific patient—say, from anorexia, paralysis, or bodybuilding—from the mean muscle mass of the population used to generate the GFR-prediction equation.”

It is also problematic that the creatinine-based GFR prediction equations require specification of the race/ethnicity of the patient, Dr. Grubb points out, since this parameter is biologically ill-defined and ethically questionable. “Cystatin C-based GFR-prediction equations, with cystatin C level as the only term, have virtually the same diagnostic performance as the more complex creatinine-based prediction equations and fewer non-GFR factors influence the cystatin C level.”

As chair of the IFCC working group on standardization of cystatin C, which successfully produced an international calibrator after two years of work, Dr. Grubb has been deeply involved in the adoption of international standards, which have greatly improved the commercial assays and which were used in the New England Journal study. “The fact that the different cystatin C assays have not agreed has been the main problem with its use in the eGFR equations,” Dr. Grubb says. “But the companies have modified their methods to agree almost completely. So I hope that problem is solved now.”

In Europe, the price of a cystatin C test is not much higher than that of a creatinine test using an enzymatic method. “For example, at my own laboratory, we charge $1.30 for a creatinine test and $3.40 for a cystatin C test. Cystatin C tests in the U.S. seem to be much more expensive than in Europe and I believe this is due to the complex and expensive procedure to get the permission of the FDA to use a diagnostic test. The many competing cystatin C assays in Europe keep the price low there, and the few assays available in the U.S. can, due to lack of competition, keep the price high. If we do not consider market mechanisms, it is even more clear that the price of a cystatin C test should not be much higher than the price of an enzyme-based creatinine test,” Dr. Grubb says.

Various research efforts continue to refine eGFR equations and how they fit into diagnosis of CKD. In an editorial accompanying the New England Journal article, Matthew Weir, MD, of the Division of Nephrology at the University of Maryland School of Medicine, urges that, in addition to creatinine and cystatin C, albuminuria be part of a CKD-defining algorithm. Dr. Narva agrees with this recommendation: “We try to make physicians aware that the two simple tests for assessing CKD—eGFR and urine albumin—are equally important.” In fact, Dr. Narva adds, NKDEP does not include a staging algorithm in its materials because any staging based on a single factor like eGFR has to be used cautiously to limit false-positives. “For example, elderly people might be unnecessarily told they have kidney disease and that can be very frightening.” He thinks a multi-factor algorithm will develop over time, something equivalent to the Framingham algorithm for cardiovascular disease. And that eventual CKD algorithm “may include eGFR, albuminuria, diabetes status, and biomarkers we don’t routinely include now.” Research now underway, he adds, will point the way on this issue.

In Sweden, Dr. Grubb says, it can be difficult to obtain funding for clinical research because basic research is favored. However, “We have thousands of CKD patients every week, and if we don’t use anything except sex and age, they are called ‘anonymous’ and we can use their data without ethical permission. So it’s easy for us to get several thousand patients with known GFRs.” He thinks the priority in CKD research should be projects that make it possible for the NKDEP and the IFCC working group for standardization of cystatin C to agree on the best way to use cystatin C and creatinine in clinical practice.

One of the limitations of the New England Journal study, Dr. Inker says, was that the findings may reflect the specific characteristics of the studies included in the database. It is possible that greater precision might be observed within more homogeneous subgroups, and she is working on studies of such subgroups now. “The populations where people would hypothesize that creatinine is very bad and cystatin C alone may be good are frail populations,” Dr. Inker says. “The elderly can be extremely heterogeneous, but in the frail at all ages—people with reduced muscle mass—very well-done studies are needed to compare different methods of measuring GFR.”

Dr. Shlipak, too, is engaged in a study of a subgroup. He is looking at a younger population of Caucasians and African Americans, and he expects results will be available in a few years. But in the meantime, with the new study adding to the weight of the evidence, he feels the jury verdict on cystatin C is in, and that testing of cystatin C should be considered an essential part of capturing kidney disease risk and kidney function. “Basically, I think we have what we need.”