November 2011
Feature Story

Karen Titus

When it comes to reporting acute kidney injury, creatinine is not the swiftest of messengers. For those who want to know if an injury has occurred, waiting on a rise in creatinine levels is just a little too slow, just as the firing on Fort Sumter was hardly the first clue there was serious tension between the North and South.

Nor is creatinine always a reliable marker for AKI. It can remain low when injury has occurred, and it can rise when an injury is nonexistent.

Nonetheless, no other diagnostic marker has gained FDA approval, and the typical definition of AKI is a 50 percent increase in serum creatinine.

But like restless citizens who realize they can no longer endure a tyrant, change appears inevitable. At recent meetings, “I’ve heard it called the bronze standard, or even the lead standard,” says Michael R. Bennett, PhD, assistant professor of pediatrics at Cincinnati Children’s Hospital Medical Center, Division of Nephrology and Hypertension, and director of the biomarker laboratory.

The search is on for markers that will help physicians identify AKI earlier. No one is happy with the status quo.

Still, it may take some time to un-seat the current king, no matter how little it deserves to remain on the throne. Joseph Bonventre, MD, PhD, notes the not-always appreciated distinction between monitoring acute changes and monitoring chronic changes in kidney. The classic biomarkers—creatinine and blood urea nitrogen—are indicators of function, not indicators of injury. “Especially in the acute setting, there’s often a dissociation between injury and changes in those markers,” says Dr. Bonventre, the Samuel A. Levine professor of medicine at Harvard Medical School and chief of the Renal Division and the Biomedical Engineering Division, Brigham and Women’s Hospital, Boston.

The disconnect likely is a consequence of how the kidney functions, Dr. Bonventre says. Because of renal reserve, normal kidney can withstand some injury with little change in its ability to filter creatinine, among other things. Changes in creatinine might be quite small even when there is a significant amount of injury, he says. On the other hand, there could be changes in kidney function that are then reflected in an increased creatinine, even when no injury has occurred—for example, when the glomerular filtration rate is reduced as a physiological response to a change in volume status.

Reliance on this poor marker has slowed matters considerably. “The flawed definition has hampered this field from the get-go,” says Dr. Bennett. “If you’re using a definition that lags behind a couple of days, you’re not going to be able to prevent an injury—you’re just trying to mop up the damage.”

How poor is serum creatinine? Let Dr. Bennett count the ways:

• It’s influenced by non-renal factors, including age, gender, muscle mass, metabolism, nutrition status, and hydration status.
• It’s typically reduced in critically ill patients due to early fluid resuscitation after surgery.
• Because of renal reserve, plasma creatinine levels begin rising only when more than 50 percent of kidney function is lost. “One of my colleagues says, ‘What other organ do you wait until 50 percent loss of function before you intervene?’” says Dr. Bennett.
• It’s not a true marker of decreased glomerular filtration in the acute setting—several hours to days may elapse before equilibrium is reached. And sepsis (a common cause of AKI in hospitalized patients) in general is associated with reduced creatinine generation, leading to falsely low creatinine concentrations.
• It’s an indication of functional change in GFR, which lags behind structural changes that occur in the acute setting.

Ideally, physicians would have a defining biomarker that’s directly linked to the disease. “Everyone says that we’re looking for the troponin of the kidney,” Dr. Bennett says. It’s an apt comparison, but may not go far enough, given the complexities of AKI. Using just one marker could result in the same flaws that affect serum creatinine. “You might have an earlier definition of AKI, but there will be exceptions to the rule with just one marker,” he says. A panel of biomarkers might be useful in identifying the location of injury (such as proximal tubule, distal tubule, or interstitium) as well as distinguishing between AKI and chronic disease. Markers might also be used for risk stratification and in monitoring, and possibly even providing, treatment.

Like rivals in a Shakespeare history play, there are multiple contenders for a new marker; like lovers in a Shakespeare comedy, it’s possible all will wind up happily engaged at some future point. The four most promising are NGAL, KIM-1, L-FABP, and IL-18. All have their strengths and weaknesses.

NGAL (short for neutrophil gelatinase-associated lipocalin) has been the most widely studied. “If you’ve looked at the literature over the last several years you’re going to find more papers on NGAL than any of the other ones,” says Dr. Bennett. That’s not to say it’s the best marker in all situations, he cautions.

NGAL is increased very early after cardiac surgery, which is one of the largest causes of AKI—it’s thought to occur in 30 to 40 percent of patients undergoing cardiopulmonary bypass, Dr. Bennett says. Creatinine takes 24 to 48 hours to rise post-surgery, but NGAL is detected, in urine and in plasma, between two to four hours after injury has occurred.

NGAL has also been detected in contrast-induced nephropathy and is predictive of delayed graft function in kidney transplant recipients.

This latter use has drawn the at-tention of Maria Hollmen, MD, a nephrologist from Helsinki University Hospital, where she works in the Department of Medicine and Division of Nephrology.

Because of Finland’s geography (large) and population (small), Dr. Hollmen and her colleagues face a few tricky variables apart from the normal challenges posed by delayed graft function. The only transplant center is in Helsinki, which can create taxing logistics for transporting patients and organs.

A faster marker would be ideal, Dr. Hollmen says. “Nowadays we just sort of wait and see. Every day we do the rounds, we see whether the creatinine has gone down, if the patient has urine output, or needs dialysis.” It may be a week before physicians know if the kidney is not functioning properly.

Intrigued by NGAL’s promise, the Helsinki physicians decided to take a closer look. Because of Finland’s size quirks, “We could do a nationwide study and include every transplant patient,” Dr. Hollmen says. “We found it was good in predicting delayed graft function.” She and her colleagues use a serum and a urine assay to measure NGAL. She says urine NGAL is more specific for kidney injury, but not all patients will have urine output. By identifying delayed graft function earlier, Dr. Hollmen and her colleagues can avoid use of nephrotoxic drugs, optimize the patient’s fluid balance, and schedule dialysis sooner.

For all its engaging qualities, however, NGAL lacks specificity. That’s kept the door open for other markers.

Enter KIM-1, or kidney injury molecule-1. This molecule, discovered in Dr. Bonventre’s laboratory, is not detectable in normal kidney tissue, but is upregulated in proximal tubular epithelial cells after injury. Its ectodomain is released into urine, Dr. Bonventre explains, making it a sensitive marker for proximal tubule injury; since the ectodomain is heavily glycosylated, the molecule is stable. It’s very specific to that part of the nephron, which is affected by 90 percent or more of the agents that are toxic to the kidney. “And it’s clearly affected by conditions of ischemia or even obstruction to the kidney,” Dr. Bonventre says.

It’s not yet clear whether changes reflective of early injury will have long-term consequences, Dr. Bonventre concedes. Like Dr. Bennett, he cites troponin—where even relatively small changes appear to reflect an adverse cardiac outcome—as a possible model, but knowing whether the same will be true for AKI will require long-term outcomes studies, he says. And even with troponin, there’s reason for pause. “More recently, the [troponin] assays have become so sensitive that people are questioning whether small changes may be as predictive,” Dr. Bonventre says.

On the other hand, he says, the negative predictive value of the KIM-1 test is likely to be quite high. In animal models, the negative and positive predictive values have been shown to be “very, very good,” he says. Proving its worth in humans has been more complicated, as it invariably is.

KIM-1 was among a number of markers evaluated in preclinical studies by a large consortium of pharmaceutical companies in collaboration with the FDA, the results of which filled much of the May 2010 issue of Nature Biotechnology. Though the focus was on preclinical evaluation, Dr. Bonventre says the research contains clear extrapolations to humans.

For now, KIM-1 remains at the clinical evaluation stage, with large studies underway to evaluate its use in cardiovascular disease, in transplantation, and in nephrotoxicity. In another study, Dr. Bonventre and his colleagues are looking at more than 300 patients who’ve undergone kidney biopsies, monitoring disease progression after biopsy to see if the biomarkers correlate with the pathology in the kidney and with disease progression.

KIM-1 rises a bit later than NGAL, says Dr. Bennett—typically around the 12-hour mark. That’s still well before creatinine; moreover, he says, that later rise might be beneficial.

A recent study at Cincinnati Children’s compared four markers post-cardiopulmonary bypass to assess their timing as well as combined values. NGAL and L-FABP, or liver fatty acid binding protein, rose early, at the two- to four- (and occasionally six-) hour mark. IL-18 and KIM-1 rose slightly later, which can be an advantage if the timing of the AKI is unknown. NGAL and L-FABP drop after peaking, though not necessarily returning to baseline levels.

IL-18, a proinflammatory cytokine that is induced and cleaved in the proximal tubule, appears to be very specific to ischemic AKI, says Dr. Bennett, and it doesn’t seem to be altered by nephrotoxic drugs or chronic kidney disease.

L-FABP is an inflammatory molecule; it’s also a proximal tubule molecule that’s upregulated following many acute kidney injuries. And it’s been shown to be higher in patients who have poor outcome. Like NGAL, however, it lacks specificity. It’s increased with chronic kidney disease and in patients with hypertension and diabetes, even when they lack overt kidney damage.

NGAL, in fact, has been shown to be a decent marker of chronic kidney disease, says Dr. Bennett. A recent study showed that baseline NGAL in patients with chronic kidney disease tended to be higher in those whose conditions worsened relatively quickly.

But in acute settings, he says, it’s even higher—which points to another limitation of current AKI biomarker studies. “A lot of them haven’t looked at chronic disease patients, because that’s a possible confounding variable in AKI studies,” he says. But these same patients are also at an increased risk of developing AKI.

Interestingly, says Dr. Bennett, these markers have all been shown to be more effective in the pediatric population than in adults. Jokingly, he notes that children are “more like mice”—that is, they have fewer of the confounding variables/comorbid conditions (obesity, diabetes, smoking, heart disease) that affect adults. “They seem to be excellent markers in children,” he says.

Ashift to other markers might also bring about a shift in attitudes toward AKI. Lack of treatment, lack of definition, and lack of a good marker have created some laissez-faire attitudes in the medical community.

If no treatment for AKI exists—and none does—might identifying it sooner be a nice but ultimately inessential activity, like ballet, or eating macaroons? Those who work closely with AKI patients have strong opinions on the matter, of course, suggesting that creatinine’s limitations might have set a booby trap for AKI treatments. Past trials have generally required creatinine levels to be up by 50 percent, with the result that treatment generally was not given until 24 hours or longer after an injury occurred. Therapies may have been seen as ineffective primarily because they were introduced too late—Dr. Bonventre compares it to treating an acute MI 24 to 36 hours after the infarction, then declaring the treatment ineffective.

Dr. Hollmen notes that there is a “really interesting study that found NGAL to be protective of the kidney when injected into mice, inhibiting AKI.” NGAL might one day prove useful as a protective or therapeutic agent, she hypothesizes.

At Helsinki Hospital, she and her colleagues are pushing to measure NGAL in every emergency department patient who has an elevated creatinine or who is otherwise suspected of having kidney injury. Unfortunately, she says, the nephrologists are often kept out of the loop by ED internists, whose focus may lie elsewhere. “If the patient doesn’t need dialysis, they might not even call us.”

Dr. Bennett would like to see more enlightened attitudes prevail among pediatric cardiac surgeons who dismiss a rise in serum creatinine as unimportant, assuming the kidney subsequently returns to normal. “A lot of times parents of children aren’t even notified, historically, that their children have this possible kidney injury,” Dr. Bennett maintains. These patients might need to be followed over time, he adds, since recent research has suggested there may indeed be long-term effects.

He also suggests new markers (NGAL in particular) could be especially useful in guiding the use of potentially nephrotoxic contrast agents. Five to six percent of patients will have some sort of nephropathy associated with contrast media, and in some cases it can be quite severe. “It’s not as common as other sources of injury, but it’s a serious problem when it does occur.” Using a marker to determine who might be at higher risk would help guide physicians to safer options.

At Cincinnati, researchers are attempting to take the next steps in biomarker discovery to see if their use can improve clinical outcomes. Would earlier intervention based on a two- to four-hour NGAL increase alleviate long-term effects of injuries?

That’s been shown in animal models, Dr. Bennett says, but that hasn’t translated to humans yet “because we have the flawed definition of AKI.”

The time is coming when that could well change, Dr. Bennett says. Studies have started gaining momentum only in the last decade or so, and it can take a while, obviously, for subsequent FDA approval and to change clinical practices. Acceptance of new AKI markers could be protracted, he suggests, since they’d be replacing a marker everybody’s been used to using for their entire clinical careers. Then again, change could come swiftly, since everyone knows that creatinine, and the definition of AKI, are flawed.

Dr. Hollmen thinks markers might be applied to pretransplant situations. All donors have normal creatinine, she says, and so donor kidneys are assumed to be healthy. “But some of them actually aren’t, and we have no way of knowing it beforehand,” she says. If a marker such as NGAL could be used up front to screen out an unusable kidney, the benefits would be tremendous, including reducing ICU stays and enabling the transplant team to forego flights to and from, say, Rovaniemi, in Lapland.

For now, the race may not go to the swift—creatinine’s widespread, ongoing use has assured that. But as research continues, there’s little doubt there will eventually be new markers under the sun.