The Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) proposed in 2009 a new equation for estimated glomerular filtration rate (eGFR). The new equation promised less bias and improved precision compared to the Modification of Diet in Renal Disease (MDRD) equation most labs use. This issue of Strategies explores recent research from the collaboration that examined how well the new equation predicts risk of end-stage renal disease and death.
Clinicians and laboratorians have understood for some time that the MDRD equation is far from perfect, and especially that it tends to underestimate GFR at higher values. The CKD-EPI equation uses the same four variables—age, race, sex, and serum creatinine—but applies different coefficients. Previous studies that examined the new equation found that, compared to MDRD, the CKD-EPI equation had lower bias at eGFR greater than 60 mL/min. One of the most significant differences in the new equation was the use of a mathematical function called a spline term. It is this adjustment that compensates somewhat for the limitations of serum creatinine and reduces bias in the higher range of eGFR. However, evidence was lacking for how well the new equation would predict clinical risk in diverse populations.
Seeking to extend the result of previous studies and address these concerns, Kunihiro Matsushita, MD, PhD, of Johns Hopkins University in Baltimore, and colleagues of the Chronic Kidney Disease Prognosis Consortium conducted a study comparing how well the CKD-EPI and the MDRD equations predicted risk for adverse outcomes (JAMA 2012;307:1941-1951). Their meta-analysis drew on data from 1.1 million adults in 25 general population cohorts, 7 high-risk cohorts of vascular disease, and 13 (CKD cohorts, and included participants from 40 countries or regions of Asia, Europe, North and South America, the Middle East, and Oceania.
The researchers classified eGFR into six categories for each equation: ≥90, 60-89, 45-59, 30-44, 15-29, and <15 mL/min/1.73 m2. They found that the CKD-EPI equation reclassified nearly one-fourth of participants to a higher eGFR category compared with the MDRD Study equation, thus lowering the prevalence of CKD in all cohorts except for the elderly. In the general population cohorts, CKD-EPI reclassified 24.4%, in the high-risk cohorts 15.4%, and in the CKD cohorts 6.6%. Only a small fraction, 0.6%, of participants were reclassified to a lower category.
The study also found that those participants whom the CKD-EPI equation reclassified upward had lower risk of mortality and end-stage renal disease (ESRD) compared with those not reclassified, even after adjusting for other factors. In comparison to MDRD, the new equation also pushed down the prevalence of CKD stages 3 to 5 in the general population cohorts to 6.3% versus 8.7%, respectively, and in the high-risk cohorts 14.6% versus 17.7%.
According to Joseph Coresh, MD, PhD, an author of the study, such findings should motive labs to adopt the CKD-EPI equation quickly. As of 2011, only 4% of U.S. labs had switched to the new equation; 92% of labs still used the MDRD equation and 4% used other equations. “We hope that by introducing this definitive data on such a large and unprecedented scale, with more than one million participants, labs can start to move uniformly and quickly to the new equation,” he said. “In some ways, just three years since the equation was first published in 2009 is not a long time. But given the fact that the new equation uses the same variables, I think the community should be able to move more quickly now, and my impression is that the incremental cost of switching is very minimal.” Coresh is a professor of epidemiology and the director of the George W. Comstock Center for Public Health Research and Prevention at the Johns Hopkins Bloomberg School of Public Health.
Matsushita agreed, noting that due to the bias of the MDRD equation at higher eGFR values, the new equation could help avoid unnecessary care. “I understand that some labs may be reluctant to make a change like this,” he said. “However, if labs continue using the MDRD equation, our paper suggests that a substantial population of patients will continue to be incorrectly identified for being at higher risk, leading to excessive additional testing and care. With the CKD-EPI equation, fewer low-risk patients will be flagged as abnormal, allowing physicians to focus on a narrower, more appropriate subset.” Matsushita is an assistant scientist in the department of epidemiology at the Johns Hopkins Bloomberg School of Public Health.
The researchers’ findings about reclassification are not surprising, given the differing cohorts used to develop each equation, noted Andrew Rule, MD, associate professor of medicine and consultant in the division of nephrology and hypertension and in the division of epidemiology at the Mayo Clinic in Rochester, Minn. “This finding is to be expected, because the CKD-EPI equation generally gives a higher GFR estimate than the MDRD Study equation around 60, the threshold that is commonly used to identify chronic kidney disease. The reason this occurs is the CKD-EPI equation was developed using both relatively healthy patient groups such as potential kidney donors and chronic kidney disease patients, whereas the MDRD Study equation was developed using only chronic kidney disease patients,” he said. “However, the CKD-EPI equation has not been shown to be more accurate than the MDRD equation for the estimation of GFR in several relevant patient groups such as among established chronic kidney disease patients and transplant recipients.” Rule was not associated with the study.
Rule commented that clinicians would likely be supportive of a lab switching to the CKD-EPI equation, and the change should not be very difficult for labs. “The technical aspects of the switch should be relatively simple since you would only need to replace the programming code for the MDRD study equation with the CKD-EPI equation and both equations use the same variables. The bigger challenge may be to ensure that clinicians are aware of the change and its potential impact on patient care. Based on this study, about one quarter of participants currently meeting the criteria for chronic kidney disease by the MDRD study equation would no longer have the disease by the CKD-EPI equation. However, since there have long been concerns that the MDRD study equation is ‘overcalling’ chronic kidney disease, I would expect many clinicians to be supportive of the change.”
The new equation should make labs more comfortable reporting eGFR values above 60, since less bias in the higher range makes these values more meaningful, according to Coresh. “The changes with the new equation are subtle, but significant,” he said. “Being able to report results above 60 is a real advantage, because if someone is at 63, particularly if the person is young, that would be very useful to know. What this means is that the blind zone above 60 is now un-blinded.”
In an accompanying editorial, coauthors Kamyar Kalantar-Zadeh, MD, PhD, and Alpesh Amin, MD, wrote that even though the CKD-EPI equation “seems valid because it produces more meaningful risk profiles, it is premature to conclude that the ultimate tool for GFR accuracy has been found” (JAMA 2012;307:1976-19770).
Coresh’s response was that there likely will not be an improvement on the CKD-EPI equation for some time. “It’s taken us about ten years to get this work done, and we have included essentially every study in the world that has measured GFR,” he said. “I don’t think there is any more data to make a better equation. Of course, learning should never stop. There will be new data on minorities and populations in other countries. But this new equations is a clear improvement in several ways, and it’s worth adopting.”