Feature Story

The latest on IMA levels, high and low

March 2004
Karen Titus

In their efforts to not miss a single myocardial infarction, physicians appear to have taken a cue from President Bush’s No Child Left Behind program. Just substitute MI for "child." But as with the educational plan, their efforts have been costly, with unintended fallout. And like the president’s school scheme, much of the cardiac efforts turn on the use, and usefulness, of test scores.

The usual tests-troponin, CK-MB, and perhaps the occasional myoglobin-have done a reasonably good job of triaging patients who present with chest pain. A relatively new biomarker, ischemia modified albumin, may do the same, only earlier. IMA is now moving from research settings to clinical use, pushed by physicians who fear overlooking an MI.

One of their biggest worries is missing an infarct in a low-risk patient. Some 8 million chest pain visits are made to U.S. emergency departments annually, says W. Frank Peacock, MD, emergency medicine director of clinical operations and event medicine at the Cleveland Clinic, and more than half-4.5 million-neither meet admission criteria for a coronary event intervention nor "have something so obviously noncardiac that they can go home."

This group of patients fall into a sort of medical purgatory, like dazed travelers socked in by a blizzard at Chicago’s O’Hare airport. "You can’t send them home," says Dr. Peacock. These patients are placed either in a chest pain center, as they are at Cleveland Clinic, or otherwise hospitalized. Then the real sorting begins, usually based on serial cardiac markers and then, depending on risk level, some type of stress test.

Yet the vast majority of these workups end up being little more than high-priced busywork: Only about five percent of these chest pain patients will have an acute coronary syndrome requiring intervention. "It’s a huge inconvenience to the patient, and it’s a gigantic cost, measured in the billions of dollars, to the American health care system," Dr. Peacock says.

Nevertheless, this method is entrenched, representing what Dr. Peacock calls the hidden cost of litigation against emergency physicians. "It’s to the point now where emergency physicians will not send anybody home, no matter how unlikely their symptoms are to represent an MI," he says.

Is it possible to avoid this and to send healthy patients home sooner? What about identifying patients who do have a cardiac problem-how early can they be risk stratified, and possibly treated?

Biomarkers such as IMA are the odds-on favorite to answer these questions. While there’s no shortage of efforts to come up with better tools-ECG pads with 40 leads and ultrafast CTs are the latest shiny objects being dangled in front of ER physicians-nothing is at the starting gate. "Markers are the answer," says Dr. Peacock.

Right now IMA is the only one ready for clinical application. Last year the FDA approved the albumin cobalt binding, or ACB, test to detect IMA for use in ruling out acute coronary syndrome when used with troponin and an electrocardiogram, and it is currently considering a rule-in claim. "What we’re basically saying is, the data show that IMA can be used in conjunction with other tools as an aid to identify patients who are at higher risk," says Peter Crosby, chief executive officer of Ischemia Technologies, Denver. The company filed a 510(k) with the FDA last November, suggesting that IMA could be used for 30-day risk stratification of patients with chest pain. And, in a study set to begin later this year, researchers will further assess IMA’s risk stratification potential and whether intervention with a low-molecular-weight heparin is helpful.

The rule-in possibilities have even the most conservative researchers excited, who thrill at the idea of identifying patients who have ischemia before troponin becomes positive. "I can hardly rein myself in," says Rob Christenson, PhD, professor of pathology at the University of Maryland School of Medicine and director of rapid response and point of care laboratories, University of Maryland Medical Center.

He’s quick to add that he’s not advocating use of IMA for its positive predictive value-"not yet," he says. "We’ll have to wait and see what the positive results actually mean. It’s prudent to be cautious about off-label uses at this point." But as part of a research team that looked at IMA’s risk stratification potential, he’s had an early glimpse of an exciting possibility.

The study, which has not yet been submitted for publication (though Ischemia Technologies’ 510(k) submission to the FDA is based largely on this paper, according to one of the co-authors), evaluated 322 clinically judged low- or medium-risk patients presenting to a hospital emergency room with chest pain of suspected cardiac origin. Patients subsequently had a blood draw within three hours of presentation, which was tested for IMA and troponin (as well as a handful of other markers that the company has yet to report on); they also underwent myocardial perfusion imaging, or MPI. Test results, along with the attending physicians’ interpretation of the ECG and MPI, were included in the case report forms, which were then sent to an adjudicated board of five expert cardiologists. "The intention of the board was to get as reasonably close to the gold standard as we could for determining whether the ECG was normal or suggestive of ischemia, and whether MPI was abnormal or suggestive of ischemia," Crosby says. "In other words, we weren’t relying on the attending physicians’ interpretation."

The researchers then looked at 30-day adverse cardiac events, including death, MI, need for revascularization, and other diagnostic tests indicative of acute coronary syndrome. Finally, they assessed the ability of all diagnostic tools at presentation to identify or predict which patients would be at higher risk.

Reports Crosby: "We found that ECG and troponin together were able to identify a little bit less than half the patients who were at high risk. When you added IMA to ECG and troponin, you end up with about 90 percent of those patients identified."

The researchers also found that the relative risk for an adverse event within 30 days increased lineally. The 95th percentile upper limit of normal is 85 U/mL; as expected, the relative risk at the upper limit of normal is one. But for IMA values of around 104 U/mL, the relative risk is over three, Crosby reports. "We now think we have data to show that if IMA is significantly elevated in patients, then those patients are at higher risk, and therefore might be considered candidates for accelerated diagnostic testing and perhaps some sort of therapy," Crosby says.

That’s the ideal. But as Dr. Christenson, who is the second author on the paper, notes, this cardiac Camelot is still a ways off. "There does appear to be a magnitude," he says cautiously. "If IMA is quite elevated, say 105, then it does appear likely that the cause of the rise is cardiac in nature. But I may be more conservative than most, and I’m going to kind of wait for evidence of treatment benefits based on an elevated IMA."

The study revealed another, somewhat surprising, finding, according to Crosby. A significant number of patients enrolled in the study were found to have left ventricular hypertrophy, which in itself is not unexpected, given the patient population being studied. But "IMA was elevated in almost all of those patients," he says, "and it may be an indication of ischemia in those patients." The researchers intend to submit these findings as an abstract.

Physicians familiar with the research are hovering above IMA like kestrels waiting to dive, though there are good reasons to be cautious about taking the plunge. Alan Wu, PhD, director of chemistry at Hartford (Conn.) Hospital, points out the most obvious one: Until the FDA weighs in on the use of the test for risk stratification-and indeed, until the company’s latest study is made available through publication-it’s difficult to pinpoint other uses for IMA. The cutoff value for risk stratification is an unknown, and whether there will be utility for values between 85 and 105 U/mL remains to be determined.

Nor is IMA cardiac-specific, though this isn’t keeping IMA researchers up at night. Dr. Wu notes that situations that might cause a positive-noncardiac ischemia, certain cancers, end-stage renal disease, gastrointestinal ischemia-would not necessarily also present with chest pain. He likens the situation to myoglobin, which is falsely increased in renal disease and in skeletal muscle trauma and turnover. "If you can eliminate those as entities that are less likely to be present, then the test becomes more specific," he says.

Others also downplay the significance of IMA’s relative lack of specificity. "It’s the biggest concern among my colleagues when I talk to them about IMA. And though I am the first to acknowledge that is a valid concern, I think it’s overblown," says Charles Pollack, MD, chairman, Department of Emergency Medicine, Pennsylvania Hospital, and associate professor of emergency medicine, University of Pennsylvania, Philadelphia.

Dr. Pollack suggests the conventional, statistical view of sensitivity/ specificity is not an issue here because physicians lack another practical means of identifying myocardial ischemia before muscle death occurs. "In a strict statistical sense, you can’t calculate the specificity without a true gold standard, which we don’t have," Dr. Pollack says. "We’ve made some of those calculations anyway, because we need to put it in a context so people can get their arms around it. But I think if that’s taken too literally, then it diminishes the value of the test unnecessarily." Moreover, he says, the 30-day risk stratification study, which "is the first time we’ve got some hard data" on the apparent link between higher IMA levels and myocardial ischemia and attendant complications, puts a better clinical face on IMA’s relative lack of specificity.

Though Dr. Pollack describes himself as "bullish" on IMA, he doesn’t shy away from the other unknowns, including the need to get a better handle on the meaning of abnormals between roughly 85-100 U/mL. "We may find out that it doesn’t mean anything," he says, "but it’s too early to say one way or another."

"I think we’ve answered to most people’s satisfactions-certainly to mine-what to do with a clearly negative test," he continues. "And these new data make me much more comfortable knowing what to do with a test that’s markedly elevated. What we’re still not sure about is what to do with patients in the middle. It’s analogous to the gray zone in troponin. We don’t know what to do with them, either," he says. "But we still have to work it out with this particular assay."

That won’t be easy, says Dr. Christenson. "My personal feeling is that there is something going on between 85 and 105, but it’s extremely difficult to size a trial to see that," because low-risk patients by their very nature have low numbers of adverse outcomes. "You need a large number to get any outcomes at all, and then to tell a difference between one strategy and another you need still a bigger number to have any confidence in your results. So I think there’s information there, it’s just that we can’t see it."

Even given the unknowns, however, the rule-in use has people talking. "I think the ultimate utility of this test will be in risk stratification," Dr. Wu says. "Because the rule-out doesn’t have the same consequences in terms of health benefits."

Physicians are hardly ignoring the rule-out application, however.

"We’re now transitioning the research into clinical use," says Dr. Christenson. "Clearly there’s compelling evidence that the negative predictive value of IMA is very high. I sort of pinch myself and say, Gee, we finally have a biochemical marker for ischemia. It may not be the perfect marker, but it could be very helpful, frankly."

Dr. Pollack and his colleagues at Pennsylvania Hospital have just started using IMA in real-time to evaluate patients with chest pain syndrome-those with chest pain and its equivalent, such as arm, jaw, or shoulder pain-who clinically appear to be at low risk. The primary goal is to exclude concern about acute coronary syndrome. But given the apparent linearity between IMA elevation and likelihood of adverse events within 30 days, he says, "We’re starting to pay more attention to patients not just with elevated IMA, but markedly elevated IMA," above 100 U/mL.

The response of his emergency department colleagues has been gratifying, he says. They now have a way of objectifying what in the past has been a highly subjective decision-to discharge home a patient with low-risk chest pain sans a lengthy, expensive, and resource-draining evaluation. "It adds some security to a decision that we frequently want to make anyway."

IMA is also making its debut at the Cleveland Clinic, in low-risk chest pain patients as a means of excluding an acute ischemic event. "It’s not appropriate for everyone," Dr. Peacock cautions. "There are people who may have a negative IMA but have too many risk factors to send them home. But in a low-risk patient, with a negative EKG, a negative troponin, and a negative IMA, we think we can send those people home safely without an extensive workup." At some point, they hope to add a specific cutoff number for high risk, he says.

In the peculiar world of emergency medicine, IMA faces a few added difficulties fitting in.

Here’s how Dr. Peacock explains his particular, if theoretical, concerns about possible misuse of the marker. First, because most emergency medicine physicians admit most chest pain patients, a positive IMA almost becomes irrelevant, he says. This may be especially true for younger ER physicians, most of whom "admit everyone, no matter how unlikely the chest pain is, because they live in a world where if you miss the slightest little thing, you get sued." (And, as Dr. Peacock points out, emergency medicine is a young person’s field. "I graduated from medical school in ’85. So in the world of emergency medicine, I’m an old man. It’s rare for emergency docs to last much past 55-it’s so hard on your body, people just find other jobs," he says.)

The upside of such caution is that the miss rate for MIs is less than one percent, versus miss rates of as high as 15 percent 15 years ago. "It does work to find MIs by admitting all these people," Dr. Peacock says. In this context, though, a positive IMA probably will not change ER physicians’ behavior "unless you’re one of the old-school docs who actually sends some chest pain patients home," Dr. Peacock says.

But these physicians may also misuse the test, taking a positive IMA as a reason to admit a patient they would have otherwise sent home. At present this represents an admittedly small group of patients, Dr. Peacock says, citing the early retirement leanings of ER physicians. "But as we get better at this and actually start sending people home, the concern is that tests like IMA will end up driving an investigation that wouldn’t have happened otherwise."

Currently the assay is available on the Beckman Coulter Synchron LX20, the Roche/Hitachi Modular P and 911, and the Roche Cobas Mira Plus. The company is also developing it for other instrument platforms, says Crosby.

The test has a very high serum-to-reagent ratio, which means it may not be easily adaptable to every automated piece of equipment, says Dr. Christenson. "Most instruments only take a few microliters of sample, and then they add a lot of ’juice’ to it, a lot of reagents. But this is the opposite, and the different ratio may make it more challenging to adapt the test to some instruments." Laboratories will also need to make sure the test performs as well on automated platforms as it has in research settings.

Dr. Christenson sees an advantage to having the current chemical test evolve into an immunoassay, or even an electrochemical test, based on the fact that it deals with cobalt. There’s also a strong lobby for a point-of-care assay.

Laboratorians will note that even in its current serum-based form, the test can be run in 20 minutes. But that’s lab time only, clinicians respond. While turnaround times of an hour-plus may be acceptable for diagnostic purposes, they’re less helpful if the results are tied to intervention. "The more we have learned about acute coronary syndromes, the more we learn we need to be quicker on them-and not just on the MIs," Dr. Peacock says. In fact, he adds, American College of Cardiology/American Heart Association guidelines suggest biomarker results should be made available between 30 and 60 minutes. "If we’re going to make that return, we’re talking point of care only," he says.

Crosby is less convinced, though the company has started the alpha trial of a POC device and anticipates beta trials will be underway by the third quarter of this year. If all goes well, he says, the company could submit a POC device for regulatory clearance late this year or early next year.

Nonetheless, he steps lightly. "Point-of-care testing has not been particularly successful in the emergency department," he posits, recalling the rollout of B-type natriuretic peptide. Despite being introduced on a point-of-care device, he says, the vast majority of BNP tests were subsequently run in the laboratory. "We suspect we may end up with the same sort of thing with IMA. There will undoubtedly be some hospitals or medical centers that would like a point-of-care device in the emergency department. But we’re not necessarily advocating it," he says.

But echoing Dr. Peacock, he changes his tune when intervention enters the picture. While there’s no current clinical advantage to having rapid IMA results, "the argument gets turned around if you can use a point-of-care device for guiding therapy. But we’re just not there yet," Crosby says.

That’s where another study, scheduled to begin this summer, may come into play. The ULTIMATE trial (Utilization of Lovenox for the treatment of chest pain syndrome patients guided by IMA testing) will enroll approximately 1,000 low-risk chest pain patients with nondiagnostic ECGs and negative initial serum markers. Those whose IMA is normal will simply be followed; those with elevated IMA will be randomized, with half receiving enoxaparin and the other half receiving a placebo. One hypothesis is that patients who have an elevated IMA have a higher prevalence of adverse outcomes than patients with normal levels. The second hypothesis is that patients with elevated IMA will do better clinically if they receive therapy-thereby supplying value to a POC device. "We’re hopeful, but we’re not sure," Crosby says. (As an aside, he reports that IMA is also being looked at for possible use in patients with stroke. Preliminary results from a couple of very small studies have provided suggestive data indicating that IMA may be used to distinguish between hemorrhagic and ischemic stroke. "Again, it’s exciting, but this is only in the early days-too early to be definitive," Crosby says.)

The ULTIMATE researchers may use a higher cutoff-perhaps 99 U/mL-to increase the marker’s specificity, reports Anatoly Langer, MD, chair of the Canadian Heart Research Centre, who will head the study. "More specifically, what we’re hoping to do is enrich the event rate in the sample size. The bane of every clinical trial, including our own, is that the events rates always appear to be lower than you predicted them, and therefore the power of observation becomes lower," says Dr. Langer, who is also a staff cardiologist at St. Michael’s Hospital, Toronto, and associate professor of medicine at the University of Toronto.

The ULTIMATE study may also answer questions about the time frame for measuring IMA. Data from previous studies, which Dr. Langer describes as limited, suggest that IMA should be measured no later than six hours after onset of chest pain. "That’s an important limitation, because obviously many patients come into the emergency room at 12 and 15 and 24 hours."

For any marker attempting to break through to the clinical side, there’s an endless hunger for more information. IMA is no different. "’More data, more data, more data’-that’s what we need. And that’s what we researchers always say," Dr. Christenson jokes. "It’s what keeps us going. But I think we are on the beginning of a very exciting era of biochemical markers for looking at the pre-cell injury phase of acute coronary syndrome. And while I applaud this work and the work going forward, we have a lot of work to show how the markers can add to what we already have."

Karen Titus is CAP TODAY contributing editor and co-managing editor.