Only about four percent of clinical laboratories in the U.S. use the CKD-EPI creatinine equation, developed by the Chronic Kidney Disease Epidemiology Collaboration, to estimate glomerular filtration rate (GFR), a 2011 CAP survey shows.1 But the authors of a new study of the CKD-EPI equation believe that, with their findings, that percentage is about to radically increase.
The study, published May 9 in the Journal of the American Medical Association, compiles and analyzes data from 1.1 million adults, adding to the already substantial evidence demonstrating the value of the CKD-EPI equation over the commonly used MDRD (Modification of Diet in Renal Disease) Study equation in estimating GFR (Comparison of risk prediction using the CKD-EPI equation and the MDRD Study equation for estimated glomerular filtration rate. JAMA. 2012; 307:1941–1951).
Across a broad range of populations, the study authors find, the CKD-EPI equation classified fewer individuals as having CKD and more accurately categorized the risk for mortality and ESRD (end-stage renal disease) than did the MDRD Study equation. “We found that the newer CKD-EPI equation consistently classified future risk better than the older MDRD Study equation,” says lead author Kunihiro Matsushita, MD, PhD, assistant scientist in the Department of Epidemiology at Johns Hopkins University School of Public Health. “This was true for both mortality and need for dialysis, and across a wide range of studies and subgroups.”
“Our paper is significant in adding a huge amount of data showing that, consistently across the world and across different populations, both general populations and those at high risk of CKD, one can improve risk predictions by using the CKD-EPI equation,” says study co-author Josef Coresh, MD, PhD, MHS, professor of epidemiology at Johns Hopkins University School of Public Health and principal investigator of the Chronic Kidney Disease Prognosis Consortium.
With this study, the CKD-EPI equation “should now replace the MDRD equation,” says an editorial in the same issue of JAMA (Kalantar-Zadeh K, Amin AN. 2012; 307:1976–1977). Dr. Coresh is optimistic that clinical laboratories will follow that advice. Among those that have already made the switch are Quest Diagnostics and LabCorp.
Clinical guidelines typically recommend reporting estimated GFR when serum creatinine level is measured, and 84 percent of U.S. laboratories report eGFR. The first strong evidence that the CKD-EPI equation performs better than the MDRD equation appeared in 2009 with a study that described the CKD-EPI equation in 10,000 people with measured GFR and found it improved on the validating of estimating GFR particularly by reducing bias at the high GFR range (Levey AS, et al. Ann Intern Med. 2009;150:604–612). “Since then, there have been a number of papers and the majority have confirmed that conclusion,” says Dr. Coresh.
“I think this study is particularly significant to those laboratories reporting eGFR because, first of all, most labs currently report eGFR from creatinine in an automated manner, and second, I think the field is mature enough now to have all the evidence people need to switch from the MDRD equation to the CKD-EPI equation,” Dr. Coresh says. A Canadian consortium member, he adds, just released a similar study of almost 1 million people that demonstrates the same finding (Matsushita K, et al. Am J Kidney Dis. 2012;60:241–249).
Like the MDRD equation, the CKD-EPI equation estimates GFR using the variables of age, sex, race, and creatinine level, but it applies different coefficients to those variables. The study in the May 9 issue of JAMA, a meta-analysis of data from 45 cohorts—25 general population cohorts, seven high-risk cohorts of vascular disease, and 13 CKD cohorts—was conducted between March 2011 and March 2012 through a collaboration of more than 100 leading researchers. More than 1 million individuals in 40 countries were included, and the participants were followed up for 9.4 million person-years. The primary adverse outcomes analyzed were all-cause mortality (84,482 deaths from 40 cohorts), cardiovascular mortality (22,176 events from 28 cohorts), and ESRD (7,644 events from 21 cohorts).
The researchers stratified patients according to their eGFR into six categories: 90 or greater, 60–89, 45–59, 30–44, 15–29, and less than 15 mL/min/1.73 m2. When the CKD-EPI equation was used instead of the MDRD Study equation, significant percentages of patients were reclassified to a higher eGFR category: 24.4 percent of participants from the general population cohorts, 15.4 percent from high-risk cohorts, and 6.6 percent from CKD cohorts. The CKD-EPI equation found a lower prevalence of CKD in all cohorts except the elderly.
The CKD-EPI equation also reclassified 0.6 percent of participants from the general population cohorts to a lower eGFR category when the CKD-EPI equation was used, as well as 1.2 percent from the high-risk cohorts and 3.2 percent from the CKD cohorts.
The study found that the reclassifications were associated with more accurate risk predictions. Using the outcomes of all-cause mortality and ESRD, the researchers found that, “Participants who were reclassified upward had lower risks of mortality and ESRD compared with those not reclassified, even after adjusting for age, sex, race/ethnicity, and other potential confounders.” Conversely, individuals who were reclassified downward (0.7 percent) had higher risk of mortality and ESRD than those who were not reclassified.
“We focused on the GFR categories that have been established in the clinical practice guidelines for patients’ CKD status,” Dr. Coresh says. “We then watched people’s movement across those categories when they were reclassified from the MDRD equation to the CKD-EPI equation. And the percentage of people who moved, particularly from the CKD range to the non-CKD range, is very large: over 24 percent. So it shows a lower fraction of the entire population have CKD, because CKD is common but not that common. It’s about 11 percent of the population.” The improvement that the CKD-EPI equation brings via reclassification means that “between one in three and one in four individuals will no longer be starred as abnormal with a low GFR.”
Another advantage of the CKD-EPI equation, Dr. Coresh says, is that it allows reporting of eGFR at any point along the continuum of results. “With the MDRD Study equation, because of concerns about bias at GFR levels above 60, many labs reported numbers below 60, and above 60 just reported that the result was ‘above 60.’ So if somebody on MDRD before had a GFR of 58 and now it’s 65, not only would they be moved to a new category so they’re not considered abnormal, but you’d also get the actual value ‘65.’” That added information will be important in certain subpopulations, he notes. “If a person is quite stable and doesn’t have proteinuria, you may say they are above 60 and that’s perfectly fine. But if that person is young and diabetic, or has progressive disease, you may be concerned about that.”
There is a long way to go before a majority of laboratories make the switch, since in the CAP survey last year 92 percent still used the MDRD Study equation. “I think only a few labs are using the CKD-EPI equation right now,” says John Lieske, MD, medical director of the Mayo Clinic Renal Testing Laboratory in the Department of Laboratory Medicine and Pathology. “The older equation has been available for several years and the CKD-EPI equation is somewhat of a moving target that comes out in new versions. In fact there’s a newer version now than what was used in the study. It’s difficult to keep changing those equations in the lab.”
But the recent decisions by Quest Diagnostics and LabCorp to switch to the CKD-EPI equation have already significantly boosted the total number of eGFRs now calculated with the new equation because these national laboratories do such a high volume of tests. Quest Diagnostics announced in a “Laboratory Update” about new tests last year that, effective May 2011, it would be using the CKD-EPI equation to estimate GFR. “We converted to the CKD-EPI equation, exclusively,” Harvey W. Kaufman, MD, senior medical director at Quest Diagnostics, said in an e-mail, adding that Quest is “actively looking into offering the new equation with cystatin C to those who request it.”
It’s probably inevitable that most laboratories will adopt the CKD-EPI equation in the next few years, Dr. Lieske agrees. “And I think that’s a good thing.” But he views the switch as more of a “fine-tuning” than a paradigm shift. “It’s not going to be as big a change as going from not reporting anything to reporting the eGFR with the MDRD equation. But it will be an improvement, especially for those people who are going to be in the range of maybe less than 65 mL/min/1.73 m2. It’s significant in that it has an impact on a certain number of people, though it’s not huge numbers. At this point with the new iterations of the CKD-EPI equation, I don’t know that much more tweaking of this will be all that helpful. They’ve gone about as far as they are going to be able to go.”
Creatinine has many strengths and weaknesses, he points out. “It’s widely available and it’s been in use for a long time. But another iteration of equation is still not going to make it work perfectly for every single patient” in estimating GFR. “So I think there’s not going to be that much more change with creatinine-based equations and there will be a move toward other biomarkers such as cystatin C.”
The federal National Kidney Disease Education Program (NKDEP), headed by Andrew Narva, MD, has adopted a cautious approach in its recommendations on estimating equations for eGFR. The NKDEP Web site says that either the MDRD or CKD-EPI equation can be used. “However, a laboratory that reports eGFR numeric values >60 mL/min/1.73 m2 should consider using the CKD-EPI equation” because it is on average more accurate for values in that range.
Some of the federal agencies’ concerns may be reflected in the minutes of the NKDEP Laboratory Working Group meeting in July 2011, when there was debate among various researchers over a potential shift to CKD-EPI. At the meeting, Dr. Narva explained the factors dictating the agency’s approach. He noted that the NKDEP has put strong emphasis on both eGFR and urine albumin in the assessment of CKD, and, “We want to address broader challenges as we move forward and develop criteria by which newly developed equations can be considered acceptable.” Rather than endorse a single equation, he said, “NKDEP will list those equations that can be used and under what circumstances a specific equation is useful.” That is now the approach taken on the NKDEP Web site.
Dr. Narva also stressed the importance of coordinating the approach to assessing CKD across the key federal agencies—his own (National Institutes of Health), the Centers for Disease Control and Prevention, and the Centers for Medicare and Medicaid Services.
CDC representative Desmond Williams, MD, PhD, leader of the Chronic Kidney Disease Initiative, expressed reservations about switching equations. Significant resources have gone into the MDRD equation, many other equations are under development, and a good reason is needed to change equations, Dr. Williams said at the meeting.
The qualifications he listed: 1) a statistically significant improvement in the delineation of disease status and risk compared with the gold standard (measured GFR), 2) significant improvement in the prognostic value for progression to ESRD and to development of complications and premature death using outcomes related to kidney disease, 3) reliability and reproducibility of the new standard, and 4) other factors such as simplification, lower cost, and the capability of addressing disadvantages of the current standard. On these criteria, he said the CKD-EPI equation did not present an advantage.
However, at the meeting there was evident controversy over this conclusion, as John Eckfeldt, MD, PhD, medical director of clinical pathology at the University of Minnesota, stressed that the CKD-EPI equation does seem to statistically improve the prediction of ESRD and death. He also made the point that the cost of converting to a new equation is in programming the changes in the laboratory information system, and it makes sense to make the change if it is easy to do in the LIS.
Dr. Coresh believes that many of these earlier reservations have likely been addressed in the research published in the past two years. He also feels the costs of switching will be minimal. “Serum creatinine assay is the same assay and for both equations GFR should be calibrated to a reference standard. The covariates are the same—age, sex, and race—so it’s just a matter of changing the equation in the computer. When we’ve spoken to physicians, it seems like they are basically ready to accept the switch. There doesn’t need to be a massive re-education.” However, he emphasizes, when labs make the switch it is worth making sure eGFR results are reported with a footnote so that people know what equation is used. “If a patient is followed up on, you would want the physician to know there was a switch” in the eGFR equation, Dr. Coresh says.
He does not believe that the collection of racial data that the equation requires will pose a problem. “Both the MDRD and CKD-EPI equations require this term. I think we’re becoming a more multiracial and diverse nation, so needing race isn’t ideal. It is a bit complicated because, biologically, creatinine seems to relate to muscle mass and probably because of that and diet, there seems to be a difference between blacks and whites, so performance of the eGFR equation is improved if one knows the patient’s race.”
While the JAMA editorial published with the study maintains there is a “disconnect” between eGFR and patients’ clinical course, Dr. Coresh contends that, on the contrary, the connection is a strong one. “There’s no substitute for having physicians know their patients,” he points out. “However, we’ve found in a number of studies that the two key markers of chronic kidney disease, which are eGFR and a measure of protein in the urine, typically albuminuria, are very strong risk factors for nearly all outcomes of kidney disease.”
It’s also true that the cause of kidney disease can provide additional information, he says. “And there will be modifying factors other than that, such as heart failure which can aggravate fluid overload, and infection which can lead to sepsis, which is a strong risk for acute kidney injury which can make the course much worse, plus specific immunologic conditions that lead to rapidly progressive disease.”
But as for the best equation for estimating GFR from creatinine, he says, this most recent study confirms that the jury is no longer out. “With the CKD-EPI creatinine eGFR equation, we’re reducing the number of people labeled as having kidney disease and the people we do label are at higher risk than they were previously.” It’s time for laboratories to make the switch from MDRD to the CKD-EPI creatinine equation, he believes. “We think that using the CKD-EPI creatinine equation for eGFR will both improve clinical care of CKD patients and reduce the costs of that care.”
1. Current status of reporting estimated glomerular filtration rate (eGFR) for adults.
Anne Paxton is a writer in Seattle.