January 2010
Feature Story

Anne Paxton

In emergency departments across the U.S., cardiac troponin is the standard biomarker for the 15 million patients a year who present with chest pain—to the point that clinicians sometimes joke about it. “We have a tendency to measure troponins in almost everybody who presents to the ED,” says Allan S. Jaffe, MD, a practicing cardiologist and chair of the Division of Core Clinical Laboratory Services at the Mayo Clinic. “Everyone knows the story about the guy who comes to the ER with his shoulder hurting, they have him lie down quickly, they get an electrocardiogram and troponin, and they tell him, ‘Good news—you don’t have a heart attack.’ And he says, ‘I know, I hurt my rotator cuff.’”

While troponin’s uses are not likely to expand to orthopedic repair, contemporary high-sensitivity troponin assays are already going beyond the detection of acute myocardial infarction to a new role in diagnosis of acute coronary syndrome. When pending FDA applications are approved, the next generation of assays—what have been known as novel, high-sensitivity troponin assays—promise to open up new vistas in clinical triage, differential diagnosis, and risk stratification.

Prospects for such applications leapt forward last August with the publication of two New England Journal of Medicine articles documenting the importance of high-sensitivity troponin assays (Reichlin T, et al. 2009;361:858–867; Keller T, et al. 2009; 361:868–877). The NEJM imprimatur gave them special prominence. “The data they published was not new,” says Alan H. B. Wu, PhD, chief of clinical chemistry, San Francisco General Hospital, and professor of laboratory medicine at the University of California, San Francisco. “They pretty much duplicated what was known in 2007 with the same Siemens assays. But the initial paper was very small and very preliminary. The new studies involved multicenter trials with large numbers and appeared in a high-profile journal. And that tends to get some notoriety.”

Other multicenter studies are now underway, Dr. Wu adds. “Over the last five years I would say there’s been a five- to 10-fold improvement in the sensitivity of troponin assays on the market. There’s been some improvement in precision—but it’s really sensitivity that’s making all this happen.” In the past year alone, four new troponin assays have been made available by Siemens, Abbott, Roche, and Beckman Coulter. “These are the top four vendors in the U.S., and they have large penetration. A lot of hospital labs use them.”

But the next-generation troponin assays are even more analytically robust. Compared with current commercially available assays, they will reduce the limit of detection by another 10- to 100-fold—and open a whole new chapter in cardiac care. Clinicians need to be aware of the changing environment in which the clinical chemistry community is working now, says Elliott M. Antman, MD, director of the Samuel A. Levine Cardiac Unit at Brigham and Women’s Hospital, Boston. “The playing field is changing under our feet.”

Discussions of troponin are often subject to confusion because of the rapid successive introduction of more sensitive assays over the past several years, some of which are already being called “high sensitivity” assays. Virginia Commonwealth University, for example, started using the Siemens (formerly Bayer) Troponin cTnI Ultra assay, when it phased out the older version of the Bayer cTnI assay, says Michael Kontos, MD, associate professor of cardiology. Despite the name “Ultra,” “this would likely be described as a sensitive troponin assay, with a LLD of 0.006 ng/mL, a 99th percentile of 0.04 ng/mL, and a 10 percent CV at somewhere around 0.03 ng/mL.”

Other currently available sensitive assays include Abbott– Architect Troponin I and Roche Troponin I. But as of the end of 2009, none of the next-generation, so-called novel or prototype high-sensitivity troponin assays had yet been approved by the Food and Drug Administration. “It’s terribly important that people understand which assays are which, because of the miscommunication that occurs in the field,” Dr. Jaffe says. (For purposes of this article, the currently available sensitive assays are referred to as contemporary or sensitive assays, while the assays pending approval are called high-sensitivity.)

Adding to the difficulty of sorting out troponin generations is that some clinical trials have clouded the distinction between sensitive and highly sensitive assays, says Rob Christenson, PhD, professor of pathology and director of clinical chemistry laboratories at the University of Maryland Medical Center.

Dr. Jaffe agrees. “If you look at the New England Journal papers, for example, they did use a couple of novel assays that are not commercially available, but most of those assays were also what I would call ‘contemporary’ assays. Mayo has been part of the evaluation process for one of those assays, the Roche hs cTnT assay. But one of the NEJM papers used the Siemens Ultra cTnI assay, which has been on the market for a couple of years and, despite its name, is not any more sensitive than some other assays out there. It’s a solid assay, what I term a contemporary assay.” The next-generation troponin T, like those developed by Roche, Singulex, Nanosphere, Ortho-Clinical Diagnostics, and Beckman Coulter, are of higher sensitivity by an order of magnitude. “They’re coming down the pike, and they’ll exacerbate all the educational issues already surrounding troponin.”

In the U.S., about eight percent to 15 percent of patients presenting to the ER with chest pain will be diagnosed with MI, depending on the assay used, says Fred Apple, PhD, medical director of clinical laboratories at Hennepin County Medical Center, Minneapolis, and professor of laboratory medicine and pathology at the University of Minnesota School of Medicine.

The two August New England Journal articles were about the diagnostic accuracy of troponin for patients presenting with chest pain and suspected of acute myocardial infarction. The first article reported that diagnostic accuracy with four contemporary sensitive assays was higher than with the standard assay and that the sensitive assays can substantially improve early diagnosis of AMI. The second study evaluated the diagnostic accuracy of sensitive troponin I assay as compared with a conventional troponin T assay and traditional myocardial necrosis markers. It found that troponin I improves early diagnosis of acute myocardial infarction and risk stratification, regardless of the time of chest pain onset.

Acute coronary syndrome is a spectrum of conditions: unstable angina, non-ST elevation MI, and ST-elevation MI, explains Dr. Antman, who is senior investigator of the TIMI (Thrombolysis in Myocardial Infarction) Study Group. “What’s happened with cTn assays is that about 30 percent of individuals previously called ‘unstable angina’ because their CK-MB was not considered elevated, we actually now know in retrospect were suffering non-ST elevated MI.”

Many diseases besides acute coronary syndromes have a cardiac component and can elevate troponin. According to one recent study, the mechanisms responsible for the release of very low levels of cardiac troponin T in patients with stable coronary artery disease could include transient, clinically silent ischemic episodes and small-vessel occlusions, inflammatory processes, cardiomyocyte apoptosis, reduced renal clearance, and increased myocardial strain due to pressure or volume overload.

“These patients will present with symptoms suggestive of AMI,” Dr. Wu says. “Their troponin values would be high, but only with the more precise and sensitive assays; with the older generation assays they might be ‘normal.’”

With troponin in the past, Dr. Jaffe says, “we’ve primarily used prognosis as a way of validating diagnosis. To prove that troponin is really measuring something of importance, we’ve looked and said, oh, we see patients with elevated troponin do worse. “

However, he believes that as troponin assays become more and more sensitive, they will reach the point where the level used prognostically may be different from the level used diagnostically. “When you start probing down in patients, particularly those with chronic disease, the cutoff values you have to use are going to fall more into the normal range and overlap with more normal individuals. That means one is going to have to interpret these values as risk values away from the diagnostic situation.”

“We’re going to have increasing numbers of patients where we’ll have to scratch our head because we may not easily be able to determine the etiology of the given elevation. That does lead to some angst on the part of clinicians and patients—and potentially some unnecessary testing,” Dr. Jaffe says.

For example, in a recent study he conducted with colleagues in Heidelberg on the high-sensitivity troponin T assay, it was shown that 50 percent of patients coming in for routine pacemaker implantations, not acutely ill people, had elevations of troponin. “And those levels are going to rise after the pacemaker goes in because you’ve done something to manipulate the heart. A day after insertion, 80 percent of those patients will have elevations. In fact, many pacemaker patients are not necessarily stable; they have underlying disease and complicated presentations. But we can’t think those patients are having heart attacks.”

Dr. Kontos says the high frequency of low-level troponin elevations in patients without clear-cut ischemia has made it necessary for emergency departments to perform serial sampling to determine if the patient requires admission for possible acute coronary syndrome, or for another non-ACS condition, such as congestive heart failure. “The downside of these highly accurate assays is that many patients who come in with heart failure, hypertensive crisis, and other diseases have these low-grade troponin elevations. And it throws the emergency department into confusion. So if we use an even lower cutoff, such as 0.04 ng/mL—which is what most are recommending for the Siemens assays—we have a tremendous number of patients with these elevations. And we could very rapidly be overwhelmed with CCU consults.”

At Virginia Commonwealth University, the policy is to perform serial testing, with the first two samples three hours apart, Dr. Kontos says. “Our residents often won’t evaluate a patient for CCU admission unless two markers come back, particularly if the first troponin is low. In the emergency department it can become very frustrating when you have these patients sitting down there. But if the first value is 0.12 ng/mL and the second value is similar, then you really haven’t seen a change, and maybe you have a low-level necrosis from something else, whereas if you have a rising value you have potentially an acute coronary event. And it should really raise your suspicions.”

Dr. Christenson says the University of Maryland Medical Center uses guidelines that are specific about when to sample. “We test when patients come into the hospital, then six to nine hours later. Earlier sampling may well be justified by more sensitive assays. I would say we are almost certainly looking at a different sequence of sampling with the new-generation assays—instead of zero/ six, perhaps zero/ two/four /six. It’s pretty clearly an issue of not just one troponin value. You have to see a rise and fall.”

Unfortunately, troponin has proved to be such a good marker that it has made clinicians in some cases too casual about ordering serial tests, Dr. Wu suggests. “Serial testing has always been part of the guidelines defining MI—not only in 1979 but also 2000, and more recently 2007. But even though the guidelines have always been stern on recommending serial testing, the specificity of troponin is so good that once they had a positive first troponin result that was sufficient to diagnose AMI, people started relaxing on the serial measurement part. So we sort of need to go back to that recommendation because of the high-sensitivity assays. If you have borderline increases that could be caused by other etiologies, serial testing becomes an additional tool to help determine what it is.”

Somewhat surprisingly, it’s still not universally accepted among clinicians that troponin is the gold-standard biomarker and CK-MB and myoglobin assays are no longer needed, Dr. Wu says. At San Francisco General Hospital, which is one of the UCSF medical centers, “We actually never did myoglobin and we ended CK-MB four years ago, even before the advent of the next-generation assay.” But other UCSF hospitals continue to offer CK-MB testing. “It’s accepted in the expert groups of people who do research in this area, but there are multiple reasons why CK-MB is still around, although not so much myoglobin. Some might be clini­cians’ lack of knowledge of what high-sensitivity troponin can produce, or perhaps inappropriate use of cutoff concentrations.”

He’s heard the argument that because troponin is elevated for a week or five to seven days, and CK-MB returns to normal within three days, it’s still very useful for detecting re-infarction that occurs several days after the initial event. “But I don’t subscribe to that,” he says. “I believe troponin can also be elevated a second time, that a second peak does occur with a second infarction.”

Sixteen out of the 17 hospitals in Minneapolis use troponin alone, Dr. Apple reports. “But I’d say across the country, it’s amazing how slowly people are chipping away at CK-MB, because first of all it’s financially very expensive, and there’s really no evidence that it adds anything toward the diagnosis.”

With the new high-sensitivity assays nearing release, could troponin now become valuable in diagnosing cardiac ischemia? It’s possible. “There’s not much out there that has proven to be a successful biomarker for ischemia,” Dr. Kontos says, “and that’s where the next step is going to be.”

“We know that some three to five percent of troponins are found free in the cytoplasm of heart cells,” Dr. Antman says, “and it’s possible under conditions of ischemia that some of that pool is released without actually having any necrosis occur in the cell. In theory, we could be getting to the point where we could see the cytoplasmic pool release—thus ischemia occurring without acute myocardial necrosis.”

Clinical trials are being done now on this hypothesis. “The work that Dr. Mark S. Sabatine [Brigham and Women’s Hospital] and his group have done will show whether we’ll be able to see troponin release in patients who do not have MI but who have ischemia, and that to me is the most exciting possibility,” Dr. Christenson says. “That’s what I think the new-generation assays are going to bring to the party. We’ll be able to see the full spectrum of troponin results.”

Since individuals who have congestive heart failure can release troponin over time, says Dr. Antman, “theoretically we could now categorize patients as to those who are at higher risk of progressively worsening left ventricular function because the extreme wall stress that their left ventricle experiences results in the release of cT. We do know that individuals who have little blips of troponin coming out of their heart—even though they’re feeling fine— do worse.”

It was once assumed that after people reached a certain age their hearts never really regenerated any new heart cells, Dr. Antman says. “We now know that’s not the case, and there is very intense work trying to establish what the rate of turnover of heart cells is. But when an old heart cell dies, it releases its troponin, and in theory if we had a very, very sensitive cTn assay, we could start to detect the background turnover rate of heart cells.” At the very least, in his view, the very high sensitive assays make it necessary to re-evaluate potential sources of variation that have not been meaningful with commercially available assays to date.

Regarding troponin’s potential use in primary care, the jury still seems to be out. Dr. Christenson is skeptical. “There is a trial going on, but there’s no evidence it will ever really work out there for prevention. I think it’s doubtful.” However, Dr. Apple believes that five years down the line, a baseline troponin might be a number clinicians would want. “In the future, once we get a series of high-sensitivity assays that are consistent, troponin may be a very powerful tool—not for screening but as a risk tool for following people from a young age.”

In the meantime, Dr. Antman says, clinical application of cardiac troponin assays continues to evolve substantively, but clinicians sometimes become frustrated because they have to grapple with test results of ambiguous significance. And there can definitely be a burden on coronary units.

The internationally accepted definition of MI, which was expanded in 2007, goes beyond what was traditionally called a heart attack, which involved much more significant heart damage, Dr. Antman says. “We’re now able to detect smaller amounts of heart damage that may not necessarily require a patient to be brought into the cardiac care unit. It took awhile to convince people you don’t have to admit every patient with an elevated troponin to the CCU. But if we change the ability to diagnose an MI, we have to consider the impact of such algorithms.”

It’s a challenge for emergency departments and clinicians, Dr. Wu says. “EDs are resistant to lower cutoff concentrations because they see so many more positive results and they don’t know what to do with them. They have to be worked up, somebody has to spend more time with the patient to make the right decision, and time is not something we have a lot of in the ED.”

The hardest lesson to learn about high-sensitivity troponin, he says, is that it’s a test for injury for which there can be many reasons, and those reasons have to be worked out. “If the cutoff is just set high enough where only AMI can reach that limit, then the decisionmaking is very simple: Positive equals AMI or negative equals no AMI. But ignoring their potential cardiac injury by dumbing down the assay with a huge cutoff, and ignoring all the things that can otherwise cause a high troponin, would be putting our heads in the sand. I would want to know if there is a cardiac injury because any cardiac injury is bad news—just like any cerebral injury. It’s never good.”

Dr. Wu also expresses reservations about the need for point-of-care troponin. “A lot of places are doing it and lots of others are thinking about it. But my professional opinion is the analytical sensitivity of the devices is suboptimal compared to the central lab.” If a clinician is just looking for AMI, “then POC devices are up to the task, and if you have a positive result you may get the patient to a cath lab half an hour earlier, or take whatever therapeutic measures you’re going to take.”

The negative result is the problem. “If it’s negative on a POC device, do you know it’s truly negative, or negative because the assay is not sensitive enough? That’s the dilemma, and so a lot of places actually repeat the sample in the central lab. But if the central lab is positive and the POC device was negative and you had relied on the POC device to send the patient home, think about the consequences. Then it becomes somewhat of a medicolegal issue.”

“From a clinical practice standpoint, the real question is this: If you have symptoms of cardiac chest pain and suspicion of cardiac disease, are you interested in a quick answer so you can go home? Or are you interested in the best answer possible so you can be managed the best? If you put it in that context, it’s a no-brainer.” Increasing the sensitivity of POC troponin is a key priority of manufacturers, however, and Dr. Wu predicts one or more companies will come up with a highly sensitive and precise POC assay in a few years.

New research is already moving clinicians’ understanding of troponin into unpredicted directions. A third New England Journal article that was released Dec. 24, 2009 reported on a sub-study of the PEACE Trial (Prevention of Events with Angiotensin Converting Enzyme Inhibition) (Omland T, et al. N Engl J Med. 2009;361:2538–2547). That study found that troponin T concentrations as measured with a highly sensitive precommercial assay were significantly associated with incidence of cardiovascular death and heart failure—but not with myocardial infarction in patients with stable coronary artery disease. Up to now, with conventional assays, these patients had troponin levels below the limit of detection. “This finding stands in stark contrast to observations in patients with acute coronary syndromes, in whom troponins are considered biomarkers of acute cardiovascular injury…and more accurately predict recurrent myocardial infarction than they do death or heart-failure events,” the authors conclude.

Another seemingly counterintuitive outcome of increasing the sensitivity of troponin is that it can lower its specificity for AMI. “Let’s use troponin T as an example,” Dr. Apple begins. “Beckman’s assay now in the marketplace has a normal reference range of .04. Let’s say they measure 30 percent of normals with that assay. A high-sensitivity assay measures 95 percent of normals, so they’re able to measure analytically much lower values. If you come into the ER with the current assay, you might be showing a normal value but it’s undetectable, versus a normal value with a measurable number. “

“Because the high-sensitivity assay is so sensitive, you might have an underlying secondary etiology for heart injury, such as really bad flu, and we’ll see an increase in troponin that is not an acute rising pattern due to MI. When we transition to the more highly sensitive assays, we’re measuring even lower concentrations—and with very good precision—and the variability of measurement down at concentrations that are lower than current assays can even detect is less than 10 percent.” A new article in the American Heart Journal looks at the Nanosphere high-sensitivity assay, he says (Wilson SR, et al. 2009;158: 386–391). “They actually showed by two hours they had detected everyone with ischemic disease, while with the contemporary assay they compared against, they detected only 15 to 20 percent.”

“The positive thing about these new high-sensitivity assays is you’re lowering the bar with confidence about what you’re measuring, you’re detecting the acute injury earlier, and you’re detecting your ability to distinguish acute from chronic injury earlier by looking at that delta between the first test and a second test as early as two hours later.”

The six-hour window, as confirmed in a 2009 study (Apple FS, et al. Clin Chem. 2009;55:930–937), is the line that the international cardiology Global Task Force has said is the one recognized to have the highest negative predictive value, Dr. Apple notes. “If troponin is negative six hours after presentation, it’s highly unlikely you have MI.” However, he argues that the goal likely will be to shorten that window to two hours with future high-sensitivity troponin assays.

Dr. Apple has concerns about the FDA’s regulation of troponin assays. First, he’d like to see more consistency of requirements for troponin assay packet inserts. “There’s no uniformity on what the FDA requires the manufacturers to place in their 510(k) clearance package inserts. They need to make sure everyone has the same pieces of information based on the analytical and clinical characteristics that are year 2010-pertinent: What’s the 99th percentile, what’s the imprecision at the 99th percentile, what’s the clinical diagnostic information at the 99th percentile, and so on.”

The second thing that needs to be addressed is that consistent guidelines for what constitutes a “normal” reference population have not been established. “We need to do a better job of defining what’s a normal,” Dr. Apple says. “There’s no clear idea of what the FDA or anybody really considers a normal population and what number of individuals is the minimum that should be analyzed to develop a normal reference range so users can determine a 99th percentile that will be reliable.”

“Is a ‘normal’ a 20 to 30 year old without any disease and not on any medications? The FDA appears to prefer to get an age-matched population of normals that correspond to the age group of patients that show up for heart attack. Rarely is one ‘normal’ once they’ve reached 40 or 50.”

Dr. Apple’s third concern has to do with imprecision and the fact that it’s not as bad as people might expect. “With these new assays, it’s recommended that we have imprecision at the 99th percentile be 10 percent. But the evidence conclusively shows that assays that have up to as high as 20 percent do not give any misleading clinical information for diagnosis or risk stratification. Imprecision up to 20 percent at the 99th percentile should not invalidate an assay and should be acceptable to the FDA.”

While it may seem hard to argue with increasing perfection, some are asking where the improved sensitivity of the troponin assay should stop. “This has been going on since the 1990s,” says Dr. Wu, “and we’re on the fourth or fifth generation of troponin T and the second or third of troponin I. The next-generation prototype troponin assay can now detect 90+ percent of healthy subjects, up from about 50 to 70 percent with the current commercial assay. Where does it all end?” He agrees with Dr. Apple, who opined in a recent Clinical Chemistry article that we do not need to go any further. “That should be the last generation in terms of analytic sensitivity that we will need, because we’ll not need to measure lower than the lowest troponin of a healthy subject,” Dr. Wu says.

Among cardiologists, Dr. Christenson has seen the gamut of opinion about increasingly sensitive troponin assays. “Some are very excited about it—mostly folks that know a lot about cardiac markers. Some are waiting to see, and others think it’s all a pain in the neck and why are you doing this anyway. They say, ‘Troponin used to be a good test until you guys screwed it up.’ There are always nihilists and enthusiasts. I guess I would call myself enthusiastic.”

Nevertheless, the clinical community needs to be sure it is ready for the workings of a more perfect assay, says Dr. Jaffe, who urges caution. “When we go to very high sensitivity assays, there’s a real risk some of the minor analytic problems that we handle very easily for standard assays may become a real problem—because tiny amounts of elevation could become diagnostically terribly important.”

“We don’t know how to deal with the clinical ramifications of a lot of these elevations,” he says. “I think we have to go slowly in making sure we are analytically in the right place.”

Related Links

• Redefining myocardial infarction