Controversy illustrates the problems that researchers face in arriving at “normal” clinical values

 

AGREEMENT on what are normal values in medicine often is a prerequisite to the development of clinical guidelines. But what happens if guidelines are based on arguable definitions of normal? The result could be the misdiagnosis of patients, as contended recently by Richard J. Glassock, MD, emeritus professor at the David Geffen School of Medicine, University of California, Los Angeles, and U.K. nephrologist Christopher Winearls, MD. Drs. Glassock and Winearls challenged guidelines for diagnosing CKD and assert that many individuals are being wrongly diagnosed as having renal disease.

 

First, some background. Last October, the U.S. Renal Data System (USRDS) announced that a 30% increase in CKD over the past decade had prompted this division of the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) to issue its first-ever separate report documenting the magnitude of this condition. According to the USRDS, CKD affects an estimated 27 million Americans. The accuracy of this estimate, however, depends on what glomerular filtration rate (GFR) is considered normal.


Dr. Glassock observed that the CKD definition used by the National Kidney Foundation (NKF) Kidney Disease Outcomes Quality Initiative (KDOQI) guidelines “encompasses a significant number of normal people and erroneously—at least in my opinion—defines them as having disease. When patients are diagnosed with CKD but don't have it, the potential harm ranges from causing them anxiety and worry to undergoing expensive testing to possibly even losing their health insurance.”


The controversy stems from the fact that serum creatinine measurements are done in nearly all inpatients and a high percentage of outpatients, even those simply making routine visits for preventive primary care. “Whenever a serum creatinine test is done, an estimated glomerular filtration rate, or eGFR, is also calculated according to the standard MDRD [Modification of Diet in Renal Disease study] equation,” Dr. Glassock explains.

“The KDOQI group did not adjust the thresholds differentiating normal from abnormal based on age and gender, and that, in my view, was a serious mistake because eGFR declines in all people as they age—everyone,” Dr. Glassock asserts. “If the KDOQI guidelines applied only to white males under the age of 45, the KDOQI criteria are just fine, but 80% of people diagnosed as having CKD using these criteria are over 65. The normal GFR in a 75-year-old woman is close to 40 [mL/min] rather than 60, but KDOQI calls less than 60 chronic kidney disease, so the problem is that elderly women in particular are being overdiagnosed as having CKD by the KDOQI criteria. For a 75-year-old man, normal is about 45 or 50.”

Dr. Glassock says the data so far seem to indicate that the number of patients who were “appropriately referred—that is, who have true kidney disease—is far, far less than [the number] who were referred inappropriately and who do not have kidney disease.”

He calls the 2002 adoption of the guidelines premature. “The KDOQI working group should have waited until better data were available to identify what the proper criteria were, or [they should have] come back and revised them after they found the criteria were incorrect,” Dr. Glassock says. “That hasn't happened.”

Drs. Glassock and Winearls took their concerns to the “Controversies in Nephrology” section of the Clinical Journal of the American Society of Nephrology (2008;3:1563-1568). They objected to the screening of unselected populations not known to be at risk of CKD by means of the eGFR formula, KDOQI staging system, and other tools “of dubious value,” warning of the inherent dangers of such screening.  This was published alongside counterarguments presented by the Thomas Hostetter, MD, director of the Nephrology Division at Albert Einstein College of Medicine, Bronx, N.Y., and fellow Einstein nephrologists Michal L. Melamed, MD, and Carolyn Bauer, MD. The group noted that mass or universal screening was not the purpose of eGFR reporting, agreed it did not seem justified, and even agreed that the KDOQI staging system leads to “disturbingly high estimates” of CKD.

But Dr. Hostetter and his co-authors advocated eGFR as just one tool that can be used to help reduce the “even more disturbing fraction of people with serious and progressive renal disease [who] are not diagnosed, counseled, or treated.”

“I think laboratories ought to report if creatinine leads to an estimated GFR less than 60; I think that's perfectly justifiable,” remarks Dr. Hostetter, a one-time director of the NIH's National Kidney Disease Education Program (NKDEP). “All it does is provide a more meaningful reflection of kidney function for clinicians who are not kidney doctors. I'm not a big proponent of staging, of saying you have stage 3 or stage 4 or stage 5. The laboratories just ought to report the estimated GFR, whatever it is.”

Specifically, the eGFR is beneficial in identifying patients who should avoid certain drugs that can further harm the kidneys, who should be taking ACE inhibitors or angiotensin receptor blockers to slow the progression of kidney disease, and, in extreme cases, who need dialysis or kidney transplantation. “I just think that without knowing what the GFR is, those things are hit or miss,” Dr. Hostetter says. “However, the eGFR is kind of rough at its very, very best, and some nephrologists have mistaken ideas about how accurately it can ever be measured. That creates problems.”

All in all, Dr. Hostetter is comfortable with the science that was used to establish the eGFR thresholds for CKD. “With any kind of laboratory test, even a blood pressure measurement, there has to be some kind of judgment in what to do. None of these is going to be driven algorithmically by a robot. With eGFR, we know that there's some error in it, but there's some error in every measurement we make.”

Like Dr. Hostetter, current NKDEP director Andrew S. Narva, MD, believes that the answer is not to abandon use of the eGFR measurement but to make sure that clinicians understand its limitations. “The eGFR using the MDRD equation is the best estimating tool we have, but it's simply an estimate,” he says.

Actual GFR had been employed mostly as a research tool, but it is being used more and more in certain clinical settings. The validity of any GFR measurement is compromised by the fact that as a product of muscle breakdown, creatinine is affected by age, gender, race, and even diet. “Creatinine is only useful when it's stable, and there have been many misuses of eGFR and other creatinine-based estimates of kidney function because they were applied at a time when the creatinine was actually changing,” Dr. Narva says. “That gives a very misleading impression of what kidney function is.”

He likens the relationship between estimated and actual GFR to the estimated date of confinement for a pregnant woman compared with her actual due date. “It's the best predictor of when a woman will deliver, but in fact women will deliver in a bell-shaped curve around that due date; most won't actually deliver on that due date.”

The NKF calls eGFR a useful first step in CKD detection, evaluation, and management but not the last step. “GFR is not the only determinant of risk,” the foundation commented in a statement to Renal & Urology News. “Thus, eGFR should be viewed as a necessary clinical decision tool, but a more complete clinical assessment of patients with CKD is recommended and needed. Studies that use the combination of proteinuria and eGFR to estimate risk are being published and are a first step toward this goal.” (Portions of the full NKF statement, which has been edited here for space, originally appeared in a report by Josef Coresh, MD, PhD; Lesley Ann Stevens, MD; and Andrew S. Levey, MD [Nephrol Dial Transplant. 2008;23:1122-1125]. Additional comments were added by Kerry Willis, PhD, NKF senior vice president for scientific activities.)

The NKF does not agree with the proposal to use age-specific percentiles to adjust the definition of CKD. “If the fifth percentile of creatinine or eGFR for each age, sex, and race group is defined as abnormal, then the prevalence of CKD would be 5% for all groups. In our view, this would lead to far too many young people and far too few older people being considered to have CKD.”

Instead, percentiles “could be based on a healthy elderly group, resulting in a less steep age-related increase in CKD than observed using the current cutoff level for eGFR.” Defining “healthy” in older individuals is problematic, however. “The main rationale appears to be avoiding classification of a large number of elderly people as having CKD with limited treatment options. However, it is not appropriate to define a disease based on the number of people classified as diseased or whether treatment is available.”

The NKF notes that it is “striking” that 38% of individuals older than 70 are classified as having CKD on the basis of decreased eGFR. The organization puts that figure in perspective, however, by noting that diabetes prevalence among people over age 65 is 22% and hypertension and hypercholesterolemia are far more common. “Inadequate treatment for common diseases should be a challenge for future research rather than a reason for changing the definition of what is normal.”

Anemia

Another gray area in the management of kidney disease is the level to which anemia should be corrected. “Almost all people with kidney failure get erythropoietin to keep their hemoglobin up, but we've learned that it shouldn't be kept at the level [considered] normal for people without CKD,” Dr. Hostetter says. “It's still a little uncertain about what the optimal level is, though.”

For people not in kidney failure, the normal range for hemoglobin is 13-15 g/dL of blood. For kidney failure patients, who often feel weak and need transfusions when their hemoglobin falls to 5 or 6 g/dL, 10-11 g/dL is about the right range.

“About 10 years ago, trials started to show us that, surprisingly, bringing dialysis patients any higher would actually make them do worse in terms of feeling weak and needing transfusions,” Dr. Hostetter observes. “Now trials are looking for the sweet spot in between, to refine what the normal level is.”

The NKF has been a leader in formulating opinion-based guidelines to bridge the gap when randomized controlled trials couldn't clarify that normal threshold, but the organization courted controversy a few years ago when its work got support from Amgen, a manufacturer of drugs that counter anemia. “I think in the long run, the NKF has done the right thing and has tried to come up with the best guess. There are some ongoing trials trying to define more scientifically exactly at what level the hemoglobin should be,” Dr. Hostetter concludes. “It's a little different from eGFR, but once again, you try to find the number that suits the epidemiology and the clinical trial data, [then set the point at which] a disease or a condition needs to be treated. Hemoglobin is a little funny because, for reasons I don't think anybody understands, raising the hemoglobin of people with kidney failure to a normal level leads to worse outcomes.”

 

Guideline challenge

Meanwhile, as researchers struggle to arrive at normal values, medical organizations feel compelled to promulgate guidelines as a way to improve patient care and to ensure that medical treatments are appropriate and justified. “Over the last two or three years, there has been a huge push to develop guidelines because there's so much public concern about safety and quality of health care,” says Philipp Dahm, MD, MHSc, clinical research director for the urology department at the University of Florida College of Medicine in Gainesville. “Our country's health-care budget is exploding, so everyone wants to make sure that the things physicians do make sense or are actually effective.” Organizations use various systems to develop guidelines, and Dr. Dahm questions the reliability of these processes. “How do they decide what is good evidence? Just because something is published doesn't mean it's worth the paper it's written on.”

Dr. Dahm believe researchers do a “pretty good” job of reporting how they search for and grade evidence. “But then it all goes into a black box,” he says. “You've got all these people sitting in a room, and then out come the recommendations. The process of what happens in that room and how the people in the room come to decisions about recommendations is completely undefined.”

Dr. Dahm sees a fix: The Grading of Recommendations Assessment, Development and Evaluation (GRADE) Working Group, which was formed in 2000. “Basically the first step is to evaluate the quality of the evidence—how effective is something, how certain are we about its effectiveness, how directly applicable is it to the patients we're talking about--and the importance of the treatment's outcomes,” Dr. Dahm explains.

The framework also addresses how the guideline process should take place. Dr. Dahm says it is “a very transparent process for guideline development.” (Details about the GRADE framework are available at www.gradeworkinggroup.org.)

Regardless of what process is used, guideline developers are limited in large part by the scientific evidence available at a given point in time. As illustrated with PSA testing for prostate cancer and the use of GFR to diagnose renal disease, that evidence can change as researchers gain new knowledge of human biochemistry and physiology.