ST2 in the ED: Will it nail cardiac risk?

December 2007
Feature Story

Anne Paxton

In the cardiac marker horserace, ST2 is the greenest of the field, a biomarker still in the experimental stage. But as a prognostic indicator, it already appears to be equaling, if not edging out, many of the markers in widespread use. As more and more studies confirm ST2's promise when compared with natriuretic peptides and other markers, the question already on many researchers' and physicians' minds is: Could ST2 pull far in front of all the rest?

Even though ST2 has not been approved by the Food and Drug Administration, with the recent publication of a key ST2 study in the Journal of the American College of Cardiology, the marker's potential value in risk stratification is turning heads in the medical community.

The study, "Measurement of the Interleukin Family Member ST2 in Patients with Acute Dyspnea" (Januzzi JL, et al. J Am Coll Cardiol. 2007;50[7]:607-613), measured ST2 concentrations of 593 dyspneic patients presenting to an urban emergency department. The authors concluded that among dyspneic patients, ST2 concentrations are strongly predictive of mortality at one year, and might be useful for prognostication when used alone or together with the natriuretic peptide marker NT-proBNP.

"It's very consistent: If you come into the ER with shortness of breath, and you have a value of ST2 greater than .2 ng/mL, then you're a high-risk patient," says Robert H. Christenson, PhD, director of core laboratories at the University of Maryland Medical Center and professor of pathology and of medical research and technology, University of Maryland School of Medicine. Like many others, Dr. Christenson believes ST2 may be an even broader marker than BNP or NT-proBNP.

Since 2006, "there's been literally an avalanche of interest in testing ST2 in a wide range of cardiovascular disease states, notably including heart failure," says the lead author of the ST2 study, James L. Januzzi, MD, associate professor of medicine at Harvard University, and clinical director of the cardiac intensive care unit at Massachusetts General Hospital.

"Researchers around the world have recognized the importance of this marker in each study, and the consistency of the results is really remarkable," Dr. Januzzi says. "They've found, almost irrespective of what the natriuretic peptide value is, if ST2 value is elevated, prognostically speaking these patients do worse."

This finding is potentially of great clinical significance, he says. "While we think natriuretic peptides are very useful, they are not perfect, and to have another cardiac biomarker with profound prognostic value to assist natriuretic peptide testing in evaluating the acutely dyspneic patient—with or without heart failure—would be welcome. We may have found it."

The pressure to find better markers to assess symptomatic patients in the ER is unrelenting, many emergency physicians report. "As emergency physicians, we literally live and die by cardiac markers," says Richard L. Summers, MD, professor of emergency medicine at the University of Mississippi Medical Center, Oxford. "They're one of our few early risk stratification and diagnostic tools we can use in cardiac patients, who represent one of the true emergencies we are charged with identifying early, and also represent a high liability for us."

Not to mention a costly scenario for the health care system. If a cardiac marker like NT-proBNP or potentially ST2, for a price of about $50, can stem $4,500 in more expensive cardiology tests, its savings in dollars and averted stress on patients is clear.

The problem is that the ED population is mostly not cardiac patients, Dr. Summers points out. "Our charge is to try to find that needle in a haystack thing that's really bad, the emergency we want to identify early so we won't send the patient home."

Over the history of cardiac markers, there have already been a few home runs, says Dr. Christenson. "Troponin was one, and more recently the natriuretic peptides BNP and NT-proBNP. And then a lot of markers have come and a lot of markers have gone. Everybody in this field is searching for that Holy Grail, a marker that will not only predict risk but also may be a key to treatment. That's how the interest in ST2 came up."

BNP and NT-proBNP were big leaps forward, he notes. "They signify that hemodynamic stress is going on, for whatever reason, and heart failure could be one of the reasons. "Then ST2 comes along and, consistently in trials for not only heart failure but also with acute coronary syndrome, it's a better predictor of death. The preliminary data indicate that it's better than the best markers out there," Dr. Christenson says.

The general structure of the ST2 gene, Dr. Januzzi explains, puts it in the interleukin receptor family of inflammatory biomarkers, yet concentrations of ST2 rise dramatically after stretch of cardiomyocytes. Furthermore, "knocking out" the ST2 gene in experimental models leads to severe cardiac dysfunction, including heart failure. "It's known that ST2 has a role in the development of immunologic tolerance; it's an anti-inflammatory hormone. In addition to the cardiac role this protein clearly plays, it is also an inflammatory marker, so there is a pluripotent role of this biomarker in the body, which might support its potential application in a wide range of disease states."

"The nice thing about ST2," Dr. Januzzi adds, "is it is clearly a cardiac biomarker. Unlike other inflammatory markers like CRP [C-reactive protein], for example, intact ST2 signaling is critical in the heart. So, while it is an inflammatory marker, it's a cardiac inflammatory marker, which gives it a good future for potential cardiac-specific applications."

The JACC study was based on data from a prospective clinical trial, known as the PRIDE study, that was designed to validate the utility of NT-proBNP, Dr. Januzzi says. "It was a 600-patient study of subjects presenting to the ED with shortness of breath, and it allowed us to have a nice platform on which we could examine the relationships between ST2 and the presence of heart failure as well as outcomes."

The assay used for the PRIDE samples, made by MBL (Medical and Biological Laboratories), was hampered by problems that included a prohibitively high coefficient of variation at the lower ranges, Dr. Januzzi says. "It was an ELISA assay that I think everyone pretty much agrees was not acceptable as a clinically useful assay." Critical Diagnostics, the holder of the patent for the ST2 marker, is working on a novel assay for ST2 to resolve the imprecision and sensitivity problems.

Dr. Januzzi says work by Alan Maisel, MD, and Rob Fitzgerald, MD, lends confirmatory and expanded results supportive of ST2 in heart failure. In a study submitted for publication from the University of California, San Diego, ST2 results not only supplemented natriuretic peptide results for prognosis, but similar to the PRIDE experience, ST2 appeared to be a stronger prognosticator than natriuretic peptides. "The study from San Diego examines the relationship between BNP, ST2, and outcomes among patients with acute heart failure, something similar to PRIDE, except in the San Diego study, the investigators did serial measurements as opposed to a single measurement at presentation," Dr. Januzzi says.

Drs. Maisel and Fitzgerald found that patients treated for heart failure whose ST2 levels remained high were destined for a higher likelihood of mortality, irrespective of their BNP levels. In fact, among those patients destined to die, BNP levels often trended downward, while the ST2 concentrations did not.

"So the take-home message—since we're lately focusing on BNP as a tool to identify those patients not responding to medical therapy—is that ST2 may be superior for this indication," Dr. Januzzi says. These data were subsequently replicated in preliminary results of an outpatient heart failure study. Other researchers in the rapidly growing portfolio of ST2 risk stratification studies are in Spain, Austria, and New Zealand. (Related article: Drawing international scrutiny).

Not reported in the JACC study from Massachusetts General Hospital was that, in addition to learning that ST2 was a powerful prognostic tool, the research group observed that ST2 was less affected by body mass index, sex, prior heart failure status, or renal status than were the natriuretic peptides, which might suggest a supplementary diagnostic role in heart failure when BNP or NT-proBNP is less reliable. "While we really think natriuretic peptides are outstanding markers for what we use them for and do not think we will replace them with any other markers for the diagnosis of heart failure, they do have limitations that ST2 may not have, including less accuracy in those who are obese or who have very severe degrees of renal failure," Dr. Januzzi says. "Preliminary data suggest that ST2 may be independent of all those other factors.

"And this was the older, less precise, and less sensitive ST2 assay," Dr. Januzzi adds, saying the results from the newer ST2 assay with better analytical precision and more sensitivity will be of great interest.

Those data are still unpublished, but Dr. Januzzi does not think ST2 should be discounted as a potential diagnostic marker. Though the natriuretic peptides will remain the "foundation" for biomarker diagnosis of heart failure, he says, "it's always nice to know we might have something new coming down the pipeline for diagnosis that might not be as complicated as the natriuretic peptides have turned out to be with their vulnerabilities."

W. Frank Peacock, MD, vice chief of emergency medicine at the Cleveland Clinic and a coauthor of the JACC paper, is director of the "emergency cardiology" fellowship program, which is not one of the traditional specialties. "It's really that area between cardiology and emergency medicine that deserved extra work and expertise because of the development of markers," he explains.

But the sorting process for patients coming into hospitals with chest pain is not in the cardiologist's realm. Unfortunately, Dr. Peacock notes, "the patient doesn't come in with a tattoo across his forehead saying 'I have heart disease.'" Eighty-five percent of people with chest pain will have something like a muscle pull or heartburn or pneumonia, but for the other 15 percent the probability is very high that they have heart disease.

In his view, that's an argument for developing a specialty in emergency cardiology. "Emergency doctors need to take this over. We're already in charge of the sorting process, but we need to develop the tools to become better at it."

The studies of dyspnea patients have helped make the marker BNP a staple of the emergency room, he says. "It's only been on the market for the last seven years, but it's clearly the standard of care to measure BNP in someone who's short of breath."

The main reason is that when the heart is under stress from pressure or stretch, it makes BNP. "There are two ways BNP works: It's a vasodilator that decreases blood pressure and takes pressure off your heart, and it causes you to urinate salt water, to decrease the volume of blood and take the stress off the heart. It's a heart defense mechanism."

If it's low, you do not have heart failure, Dr. Peacock says. If it's high, there are other things that could be the cause, such as a big blood clot, which can cause heart stretch and elevated BNP. "So you have to understand the clinical situation."

What ST2 adds is a much sharper risk assessment. "BNP won't identify all the people who do worse. ST2 identifies more, so it's a better prognostic tool. It's a great discriminator." Precisely why is not known, however. "We know there is something in the fibrosis pathway that correlates with very high death rates for patients who come in with elevated ST2. We can speculate it's something about fibrosis that's a final thing before you die, or maybe heart cells are dying and being replaced."

ST2 could also work on acute coronary syndrome as well as heart failure, he says. He is in the middle of a study suggesting it may also signify mortality risk for lung disease. "Because of the way ST2 developed as a heart stretch marker, we focused on heart failure. It could go beyond that. But either way, it's a 'this-dude-is-sick' marker that's extremely valuable in the emergency department.

"I'm not clairvoyant, and a high ST2 would change my level of concern, " Dr. Peacock says.

Many of these disease states overlap in an inflammatory response to the body, Dr. Summers explains, and ST2 is a marker of that inflammatory response, "or something bad going on that the body is responding to.

"So I think any marker that does identify to a high degree of sensitivity an inflammatory response has the potential to help us treat patients who may be having something bad going on that we're not able to discern from the outside—particularly from the ED perspective so we can prevent them from being sent home when the prognosis is bad," Dr. Summers says.

Unlike some biomarkers that add little if at all, "ST2 clearly does add," Dr. Peacock says. "Troponin and BNP are pretty darn good markers, but it looks like ST2 is much better as a prognostic tool, especially in cases of renal failure or obesity—those conditions where BNP is weak."

"It's not out for prime time yet," he adds, "but it deserves to be watched, and it ought to pique a lot of interest in the medical community."

FDA approval, in fact, is the chief milestone on the road to prime time. Dr. Christenson can't speak for the FDA, he notes, but it appears to him that for clearance of ST2, another study similar to the PRIDE study of NT-proBNP may be needed. "The new study was useful," he says, "because PRIDE defined an ST2 cutpoint of 0.2 ng/mL. But the study was run in a single center in Boston, so the cutpoint needs to be validated. The question is, Will 0.2 work in Birmingham, Ala., or in the United Kingdom? It will be important to test that in a separate group of patients to make sure that's the right cutpoint and make sure the marker really behaves the way you predict."

However, ST2 is about to reach a crossroad in the quest for wide acceptance, as Dr. Januzzi prepares to launch a prospective study of ST2 in the cardiac ICU at Massachusetts General Hospital, intending to examine therapeutic interventions and their effects on ST2 concentrations. "The natural question is, Can we change their values? In other words, if they're elevated and bad, can we do something to lower them?"

Targeting the abnormal hemodynamics and filling pressures in those with severe heart failure, the trial will use medications that are beneficial for heart failure, with the expectation that ST2 should respond to favorable interventions, he says. That might make it a useful tool for guiding heart failure management.

The study is sponsored in part by Critical Diagnostics, but it will look at other biomarkers including NT-proBNP and troponins in the same setting, Dr. Januzzi notes. "There is a lot of data about natriuretic peptides in prognosis of acute heart failure. But what everyone wants to know is whether you can and should target your therapies at these biomarkers' concentration. So we'll be looking at that in this study."

A multimarker approach could potentially provide the most accurate prognosis, in Dr. Januzzi's view. "If you look at the predictive value, ST2 appears best, and the natriuretic peptides are exceptionally strong too. Use them together and you have an even more refined ability to predict who's going to run into trouble. Ultimately, we may well use patterns of BNP, NT-proBNP, and ST2 to admit a patient to the ICU, admit to a lesser monitored setting, or even discharge the patient," he says. This makes sense, he adds-you want to use markers with complementary pathophysiology explaining their elevation, as well as complementary prognostic value. "The combination of a natriuretic peptide and a cardio-specific inflammatory marker like ST2 makes perfect sense for this indication."

If managing patients based on specific biomarker concentrations turns out to be associated with better outcomes, "that would be a revolution," Dr. Januzzi says. "The problems we have these days in modern cardiology with respect to financial outlay as well as outcomes—so much is due to heart failure. Anything that will improve our ability to better treat patients would be welcome financially and would likely avert a lot of pain and suffering for our patients with heart failure."

But Dr. Summers has "mixed feelings" about the multimarker approach. He says: "It depends on how well physicians understand the limitations and powers of the multimarker strategy being used. We don't want to get into a situation where we have so many markers that they would cause treatment or more testing that's unnecessary or even dangerous for patients if you don't have the specificity required."

Researchers in the acute cardiac care program at Virginia Commonwealth University Health System have spent several years studying cardiac markers, says Robert Jesse, MD, PhD, the program's director and a coauthor of the JACC study. "We've been involved through much of the last decade in development of troponin assays, and we've done a lot of studies on IMA [ischemia-modified albumin], one of the ischemic markers we were hoping would pan out better than it did."

ST2 is still very much at the experimental stage, he emphasizes. "Almost all the work with ST2 to date has been on retrospective samples. They're just now moving into prospective studies. BNP and NT-proBNP are relatively good at helping discriminate between people showing with shortness of breath due to cardiac causes rather than those with primarily a pulmonary cause. The way it's looking, ST2 is better at prognosis."

Hospital executives and physicians wonder if cardiac markers could be used as screening tests to save the expense of conducting echocardiograms, he says. "Right now hospitals are overcrowded, and as a result, ERs are overcrowded. And often the difficult distinction an ED physician needs to make is, Do I take this patient and just admit them to initiate their therapy, or can I just send the patient home and see them in a clinic tomorrow, managing this in the outpatient environment."

"These are often very difficult decisions, and sometimes our thinking gets biased more by what's practical, so a good prognostic test would help. The downside of a very prognostic but not diagnostic test like ST2 is that people don't exactly know what to do with the patient or, probably worse, they do something unnecessary."

"You don't want a test that will cause the cardiologist to automatically assume another test needs to be done," Dr. Jesse says, noting that this is one of the problems now with troponins. "The tests are getting so sensitive and reliable, and the analytical noise is almost nonexistent, so a positive troponin makes everybody think acute coronary syndrome. Troponins do define high-risk patients, but does everybody need a catheterization if the troponin is positive? No." He sees a potential for the same thing to happen with ST2 if physicians aren't careful in using and interpreting these tests in appropriate clinical situations.

Aiding therapeutic monitoring is another tantalizing role that ST2 might play.

Looking at patients with chronic stable heart failure, Dr. Januzzi says, "In therapy, not infrequently, they'll be doing fine, but all of a sudden the wheels come off, they start spiraling into repeated hospitalization and eventually death. If we were able to watch them closely on the outpatient side using measures of biomarkers like ST2 or NT-proBNP, that would be a huge advance forward."

In heart failure, Dr. Peacock says, "we do know there are a number of life-prolonging drugs such as enalapril that if our patients are on, their mortality drops, although not all the eligible heart failure patients are on those drugs." It could be that elevated ST2 is associated with elevated therapy such as upping the dose or treating the patient more aggressively.

But more study is necessary to determine that, he says. "The PRIDE database does include that kind of data, but when you subset the data you get very small numbers. So until we have a prospective randomized study of therapies, it is premature to anticipate ST2 will play a role in therapeutic monitoring." He stresses, however, that there are plenty of data to justify such a trial of ST2.

Could there even be a wider purpose for ST2, perhaps as a screening tool like PSA?

Dr. Jesse is agnostic on that question. "In general, I'm skeptical about screening these days because we can create as many problems as we'll solve, unless we just look at very specific populations." Though the average person would not be likely to have elevated ST2, "we haven't yet looked at what an elevated ST2 means in broad populations. I don't think ST2 is a generic marker for badness, so it's not something you'd want to use on everyone who walks in the door."

He cites troponin as an example. "If you only look at troponin in people coming to the hospital with chest pain, you have a population that's very enriched and is going to be positive for what you're looking for—AMI. If you test troponin on everyone who comes to an ICU, you'll start to say, 'A lot of patients with positive troponins don't have apparent coronary disease.' Troponin is elevated because of myocardial necrosis, which can happen in the absence of coronary disease such as in trauma, sepsis, or viral infections."

Ideally, he says, "When we develop new markers, we do it in highly selected populations because we're trying to answer relatively specific questions. And when we think we understand how they perform in those scenarios, from there we can go on and expand the landscape a little bit."

With ST2 there are very real possibilities, he believes. "It's very interesting that no matter what population we've looked at so far—although there's a lot we need to learn—it's elevated in the groups that have the worst outcomes."

So from the perspective of the ED, "if I've got somebody who's not so sick he needs to be intubated, and I'm sitting on the fence about sending him home or admitting him, there's a lot of scrutiny regarding admissions and whether patients meet specific criteria. So my sense is this is probably where ST2 will be used: for decisionmaking in patients presenting with heart failure and some of the ischemic syndromes as well. Where we know it works best now is in the setting of an acute presentation."

Dyspneic patients without heart failure is another group being explored now, Dr. Jesse says, with researchers asking whether ST2 elevation means the same thing in people with other forms of pulmonary-based blood diseases.

What the positive reports about ST2 mean for the diagnostics industry and laboratories is open to speculation. In general, Dr. Jesse says, one can take several pathways when developing a new marker. One is to develop a specific device on which to test the assay, and the argument would be that it makes sense to know every ST2 test is the same. "The alternative, if the assay is really driven by antibody-based technology, is to sell or license that antibody to anyone who could then run it on their mainframe instrument."

"So now you're talking troponin T versus troponin I," he says, referring to one test that could be performed only on a Roche Diagnostics device and another test "licensed out to just about anybody."

The problem with troponin I, he says, was that "Bayer and Beckman and Abbott and everyone else with a troponin I assay had different endpoints and different cutoffs. You literally could get someone transferred in from outside the hospital with a normal troponin that was four times higher than your normal troponin. So if you weren't cautious about making sure where the assay was done, you might think that change was significant."

However, having an assay able to run on only one platform has its downside, too. "A lot of laboratories are happy with their big robotic automated mainframes, and are willing to add a new assay to the mix, but don't want a new stand-alone instrument because now the tests all have to be done manually."

There are so many variables, he suggests, that once a small company gets an assay to the point of being comfortable with its specifications, sensitivity, specificity, and so on, "how to market it is a really complex business decision." Biosite Inc., which makes the assay for BNP, was successful with a relatively small handheld or desktop instrument and had control over the assay for about three years. "Those are difficult challenges for the presidents and the boards of small diagnostics companies."

The ST2 clinical trials being run here and in Europe will be critical to ST2's fate. "We do know the FDA will no longer accept at face value a number unless it actually means something," Dr. Jesse says. "They're increasingly demanding that we show incremental value in a new assay over what's out there now, and that's costly to do."

Nevertheless, Dr. Jesse envisions a bright future for ST2. "We've spent many years appreciating the nuances of which populations troponin was important in. Now we're looking at populations outside ACS and seeing some interesting things that we currently have no idea what to do with, but we know in certain conditions being troponin-positive is equally bad. This is kind of where we are with ST2 now, so the first step is not to let something bad happen that could be prevented; this is the role of the prognostic signal. While it will take us several years to sort out exactly what the role of ST2 is in this bridge between the inflammatory process and cardiovascular disease states that we're seeing in both ACS and heart failure, the real promise is that ST2 will reveal new insights into the pathophysiology of these diseases that will translate into therapeutic opportunities to improve outcomes."

Dr. Summers hopes there will be more work on the specific physiological mechanism by which ST2 is elevated and what its role is in the patient's physiology. "We know it's a prognostic marker, but why? We have a glimpse of that but not a basic understanding. It's the 'why' that will tell us what we need to do to fix it."

In fact, Dr. Summers says, medicine in general will be taking a more detailed look at the microphysiology of what's going on with patients. "The last 10 to 20 years of medicine have emphasized evidence-based practice. But implicit in that is the idea of treating populations as a whole—for example, everybody with chest pain gets this treatment and this treatment. But while we know a certain treatment for everybody over 55 may improve the majority, for the individual in front of you, that may not necessarily be true."

The next 20 years in medicine will be the era of goal-directed therapy, where clinicians target the patients' physiology to optimize their outcomes, he predicts. "And to do that requires a better understanding of the individual, not just everybody with chest pain."

This is a great time to be talking about ST2, Dr. Christenson says. "The most important thing to do with any new biomarker is to validate it. We've seen too many come—and I'd even put hs-CRP in that category—where there is lots of hype about it, it turns out to be useful, but it also becomes a target that everybody compares their marker with to show theirs is better. So now it's being validated, and folks in the clinical realm, like myself, will be ready when ST2 comes on the market." His forecast: "This is a marker that people will actually see become useful in laboratories in the next few years."

Related Links

• A cardiac marker is born
• Drawing international scrutiny