July 2012
Feature Story

Karen Titus

Laboratorians will have to educate clinicians about the high-sensitivity troponin assays, and it’s not too early to start. “What works is talking to them,” Dr. Fred Apple says. “You can’t send out memos. You can’t put a paper in the mail to them. Just arrange a time to have a conversation.”

For anyone with even a glancing interest in cardiac troponin assays, Paul Collinson, MD, FRCPath, offers a cautionary tale about the ease with which they can be ill-used.

It’s a rather sad story, says Dr. Collinson, a consultant in chemical pathology, St George’s Hospital, London. “A nice gentleman of about 80 years came in. He’d got up to change his television channel—clearly didn’t believe in using the remote control—and had a bit of a faint, fell over, and cut his head.” He was stitched up in the Emergency Department. Then someone in the ED—for reasons best known to himself, says Dr. Collinson—looked at the patient’s troponin level, which was slightly up.

“So they promptly gave him a lot of antithrombotic medication, and of course—surprise, surprise!—he bled into his cut head,” he continues. “He actually ended up needing a transfusion.”

The lessons, in case they’re not obvious: Best not rely on just one troponin value. And never lose sight of a patient’s clinical presentation.

This is not silly advice, especially as the makers of troponin assays get ready to bring more sensitive versions of their tests to the U.S. market. Observers are predicting that at least two companies will submit 510(k)s to the FDA by the end of the year, which might mean the tests could be available by the second quarter of 2013.

Hopes are high for the new tests. (Experts are quibbling over whether to call them high-sensitivity or highly sensitive assays.) “They better make their way into clinical practice,” says Judd Hollander, MD, who envisions using the more sensitive assays to double the discharge rate of patients who present in the ED with chest pain, but who do not have acute coronary syndrome. “That’s the world’s biggest win,” says Dr. Hollander, professor and clinical research director, Department of Emergency Medicine, University of Pennsylvania, Philadelphia.

But as Dr. Collinson’s tale illustrates, troponin assays have retained their power to fuddle physicians. That’s why it’s not too early to start educating laboratory colleagues and clinicians alike about the new troponin assays. In fact, says Fred Apple, PhD, it’s not just a matter of transitioning to the latest assay. Some laboratories still use older-generation assays, and moving from those to current ones is an equally big leap.

On the surface, none of this should be especially tricky. As the analytical assays improve, labs can measure lower values at better precision. Inherently, they’ll detect more cases of increased troponin values. But the cost of increased sensitivity is decreased specificity. While some clinicians may be used to a specificity of, say, 90 percent, the newest assays are closer to 75 percent specificity, says Dr. Apple, medical director, clinical laboratories, Hennepin County Medical Center, Minneapolis, and professor, laboratory medicine and pathology, University of Minnesota School of Medicine.

“If you’re a cardiologist, that raises a lot more concerns,” he says, “because you’re going to get a lot more referrals.”

The flip side is in the Emergency Department, where physicians “don’t want to miss anything. So they want a very high sensitivity,” says Dr. Apple.

That tension may be one of the reasons why, when Allan Jaffe, MD, gave a grand rounds earlier this year on high-sensitivity troponin assays, he titled it, “Get ready, get set—panic.” Add to that lingering confusion over contemporary assays (a term some experts use to describe the most sensitive assays currently approved for use in the United States), and it’s easy to see why this is, ahem, a sensitive topic.

These assays are far better than many physicians realize, says Dr. Jaffe, a cardiologist, professor of medicine, and chair of the core clinical laboratory services, Department of Laboratory Medicine and Pathology, Mayo Clinic, Rochester, Minn. “So one way for clinicians to get used to dealing with the problems that are going to occur with high-sensitivity assays is to use the assays we have now as the guidelines suggest,” Dr. Jaffe says. If they do, they’ll start to experience the same issues that will become even more important, and more complex, with the high-sensitivity assays.

But for these assays to perform at their best, users need to be rather militant about adhering to the universal definition of myocardial infarction, with an assay imprecision at 10 percent in the 99th percentile upper reference limit.

Rather than adopt it, however, many physicians have instead done a clumsy waltz around the definition. Those who prefer a lower sensitivity “have complained and pushed back pretty strongly,” says Dr. Jaffe; consequently, labs have also been reluctant to use the 99th percentile.

Dr. Jaffe says it’s hard to discuss the matter publicly “in a way that is palatable because it sounds like clinicians really do not want to know about people who are sick,” he says.

Observers trace the difficulties to the prior iteration of blood tests used to identify possible myocardial infarction. Because of their high specificity, the vast majority of elevations could be attributed to ischemic heart disease. “Elevations used to mean, Go to the cath lab, because they’re having an acute MI,” says Dr. Hollander.

But as troponin assays became available, and as their sensitivity improved, the proportion of elevations due to ischemic heart disease has dropped. Laboratories then became victims of their own success, Dr. Collinson suggests. “We initially gave people this duality of good and evil—MI or non-MI—and now we’re saying there’s sin in between.” Physicians are picking up a panoply of other disease states—some well understood, others less so—that can cause troponin elevations, including myocarditis, heart failure, an underlying inherited disease, sepsis, metabolic disease, toxicities, structural heart disease... the list goes on.

When physicians fail to look beyond MI, they can end up like Macbeth. “They’re stuck in ways they don’t know how to get themselves out of,” says James de Lemos, MD, professor of medicine and holder of the Sweetheart Ball-Kern Wildenthal, MD, PhD Distinguished Chair in Cardiology, UT Southwestern Medical Center, Dallas. “I think people forget about Bayes’ theorem when they think about interpreting these results, when they apply a test to a patient population that has a very low likelihood of heart attack, and then the test comes back abnormal. It happens to all of us. It happens to me.”

He agrees that sticking to the universal definition is helpful, because it reinforces that troponins are specific for cardiac injury. But that still doesn’t end a cardiologist’s (or a patient’s) woes. As Dr. de Lemos notes, clinicians need to decide whether the pattern of troponin release indicates an acute or chronic injury. The latter, they’re learning, is much more common than previously thought. And even when an injury is acute, it doesn’t automatically mean it’s ischemic necrosis.

But not everyone appreciates that fact. There may even be subtle biases in the medical system toward ischemic heart disease—follow the money straight to the cath lab—though don’t expect anyone to say that on record.

Clinicians who’ve trod one sure path (see an elevated troponin, hospitalize the patient, order an angiogram, detect disease) may now find themselves lost in the woods. “Now we’re faced with patients who don’t have coronary disease,” says Dr. Jaffe. While that sounds good, these patients do have something, though it might not be clear what. Says Dr. Jaffe: “Clinicians would argue that we’re actually hurting people by doing a lot of unnecessary testing, to say nothing of overwhelming our systems with people who maybe don’t need to be in the hospital.”

Frustratingly for labs, clinicians may resort to calling the elevated troponins false-positives, “which I think is a terrible term,” says Dr. Hollander.

These are not false-positives; these are real positives. “I think what they’re trying to say is that the patient doesn’t need a stent,” Dr. Hollander says. “But that’s not what gets said in practice. What gets said in practice is, ‘It’s a leak, or it’s a false-positive.’”

Physicians also need to remember that more troponin is in almost all circumstances worse than less troponin—it’s that simple, he says. It indicates there’s heart disease, and a patient with heart disease will do worse than the patient without heart disease. “In fact, I think people with stubbed toes and troponin elevations probably have a worse prognosis than people with stubbed toes and no troponin elevations.”

He has no patience for equivocation. Too often, he says, physicians waste their time explaining away troponin elevations, when they should spend more time trying to dig deeper into what it means.

If, for example, a patient with renal failure has a mildly elevated troponin, physicians commonly dismiss it as inconsequential. “I still hear too many people walking around—even my shop, where we talk about this a lot—saying, ‘Oh that’s just a leak from heart failure.’ Well, it’s not a leak from heart failure,” he says, sounding annoyed. “Some part of their heart is dying.”

Dr. Hollander holds little hope that clinicians will learn to care about the universal definition of MI as much as laboratorians do. Clinicians live in a black and white world, he says: Troponin is elevated or it’s not. “No one knows whether they’re using the 10 percent CV, the 20 percent CV, the 95th percentile, the 99th percentile—the lab guys may know that, but a clinician looks at the upper limit of normal. And it’s either above that or below that. That’s all they do,” says Dr. Hollander.

Only about 50 percent of laboratories have implemented a true 99th percentile, Dr. Apple estimates, though he concedes there are few reliable statistics that track this. “Even some of the more sophisticated medical centers around the country still give two cutoffs, a 99th percentile value and a higher ROC curve value,” he says. “They call the concentrations in between that the ‘gray zone.’”

That drives Dr. Apple nuts. “I dislike that term.”

Some physicians call it intermediate zone, he says. Whatever words they use, they’re calling it a potential MI. While it is that, it’s just wrong to think this way about a zone, Dr. Apple says. “You have to believe the universal definition of myocardial infarction based on interpreting both clinical and biomarker information. Every major organization in the country has supported the concept that the 99th percentile is the cutoff. Even though it’s an expert opinion, the evidence is very strong. Providers should manage patients with what the definition is. There is no intermediate zone.”

Why the sluggish acceptance? “I can blame everybody,” Dr. Apple jokes.

From the laboratory’s point of view, Dr. Apple says, changing to the 99th percentile requires a huge educational investment. “It’s a lot of work to go to your clinicians and make sure they understand what’s going on.” And, as Dr. Hollander notes, not all clinicians feel inspired to absorb the details.

Many clinicians, including those of the cardiologist persuasion, are resistant to the change, Dr. Apple continues, because they don’t want to see a spike in cardiac troponin positives that are hard to explain. “The reality sinks in, and all of a sudden you’re seeing 35 to 40 percent increase in positive values.” He does see less resistance in physicians who have been in practice fewer years—less time, perhaps, for habits to become fossilized and a keen sense of the changing biomarker landscape. There’s nothing unusual in this resistance, by the way—he recalls similar mulishness when physicians switched from CK-MB to troponin.

Emergency medicine physicians are less likely to tremble at the thought of more sensitive assays. Erring on the side of very high sensitivity is preferable to missing an MI. And, as Dr. Hollander notes, doubling the discharge rate pleases everyone.

But he’s sensitive to the fact that cardiologists won’t want to see more patients unnecessarily. “We’ll need to learn to figure out that not everybody with a low-grade troponin elevation needs a cardiologist,” Dr. Hollander says. When he speaks to ED physicians, he tells them it’s their job to ease cardiologists’ frustrations. “Don’t give them patients they shouldn’t be getting.” It’s ridiculous, for example, to hand off a patient with renal failure (but clearly lacking an ACS) to a cardiologist, based on a slightly elevated troponin.

“If we could get the message out there educationally, we can dampen the degree of problem we’re going to have when these tests hit the market,” he says, echoing Dr. Apple. Not that it will be easy. Like pathologists, ED physicians have to be experts in multiple areas of medicine. “The truth is, people forget what they consider to be little things. I don’t consider troponin to be a little thing, but there are other people who might,” Dr. Hollander says.

Let’s say, for argument’s sake, that users of contemporary troponin assays started using them the way the good Lord intended, or at least with the 99th percentile. Would there even be a need to move high-sensitivity assays when they become available?

Certainly. They’ll let physicians identify additional patients whose acute coronary syndromes are missed by contemporary assays, as well as help clinicians to move patients out of the ED more quickly.

Another benefit will be identifying problems other than MI. Right now, no one knows how to treat every underlying cause.

But that day will come. To take one small example, says Dr. Jaffe, physicians can use even contemporary troponin assays to identify women receiving Adriamycin as breast cancer treatment who are at risk for developing cardiotoxicity to the drug. One strategy uses ACE inhibitors to inhibit cardiotoxicity. The work is early, to be sure, and there’s (so far unproven) concern the therapy to reduce cardiotoxicity might also reduce effectiveness of the chemotherapeutic agents. “But those are the sorts of advances we might make,” says Dr. Jaffe, for other drugs, other circumstances, and other diseases.

Unlike others, Dr. de Lemos predicts the newest assays will be of little value in the ED. Contemporary assays are already sensitive enough for basic rule-in/rule-out applications, he says, especially when used correctly. But the upcoming assays have “unbelievable potential” in chronic disease states and in screening apparently healthy people, he says, as well as in monitoring chronic cardiovascular disease. “It looks like one of the most powerful markers we’ve seen yet.” He and his research colleagues are optimistic that troponin might function as a screening test for subclinical cardiac injury in apparently healthy people, enabling physicians to treat, or even prevent, heart failure.

Cardiologists might be forgiven for thinking these will be slowly accruing dividends. Potential doesn’t help “clinicians who are struggling every day with, ‘What do these elevations mean? I have my patient in front of me right now, and you’re driving me nuts,’” Dr. Jaffe concedes. Sorting it out will tax even the best medical minds for some time.

Complicating matters, there’s some degree of heterogeneity in release of troponin from the heart. Troponin release is flow related, Dr. Jaffe explains. Chest discomfort due to a blood vessel that was intermittently open and intermittently closed would give a fairly quick signal. For someone with a prior MI, whose blood flow from the affected area was poor, it might take longer to see a change in pattern. In addition, says Dr. Jaffe, even chronic elevations may change due to biologic variability.

Such factors will need to be taken into account as the newer assays come into practice. That’s why the time to talk—with laboratory colleagues, cardiologists, emergency medicine physicians, hospitalists—is now.

It might be useful to start by discussing what to call these as-says.

Clinical Chemistry has led the charge to use “high sensitivity” as the term of choice, says Dr. Jaffe.

But Dr. de Lemos takes exception to this definitive, saying it implies the assays have reached the end of the line, “as if we’re done.” In reality, assays become progressively more sensitive. He votes for “highly sensitive.” And, he says, assays that claim to be “high sensitivity” are only marginally more sensitive than the assays currently available in the U.S. market.

Even those who vote for “high sensitivity” recognize the term is flawed. “We don’t have a strong definition of what high-sensitivity means,” Dr. Apple says.

He’s trying to come to the rescue. If troponin had a life coach, it would be Dr. Apple, who tirelessly urges all sorts of improvements for making this assay shine. (“I look forward to the day if and when I have to go into the hospital for a troponin,” he says. It’s not that he welcomes the thought of having an MI. But he anticipates the day when physicians will recognize all the meanings of an elevated troponin, and know how to intervene.) His idea is that high-sensitivity should be defined by the number of individuals in a given reference population detected by any given assay.

All well and good. But “normal” isn’t necessarily any easier to define than “high sensitivity,” as it turns out. Dr. Jaffe reports that a high-sensitivity troponin T assay, in its initial validation study, detected 80 percent of normals. In a subsequent multicenter trial, it detected 33 percent of normals. “There’s some ambiguity,” he says, laughing.

Should “normal” be defined by 20-year-olds, asks Dr. Apple, or by people aged 20 to 75 who appear to be normal? Men, women, or both?

Dr. Apple recalls pushing the NIH to fund a study that looked at this question. The agency told him the assay manufacturers should be the ones to support it—an idea that went exactly nowhere.

He has since sent normal specimens to assay makers to help them develop normal ranges; while the debate could extend to whether his samples are indeed normal, Dr. Jaffe says, “At least you’re talking about the same set of samples.” Some experts in the field are also advocating for a sample bank that the companies could use as well, though Dr. Jaffe says others see this as gilding the lily.

He doesn’t. He would want such an initiative to include history, physical, NT-proBNP, creatinine, and some sort of imaging study as well. He’s helped develop several data sets that include imaging and says the impact on normal values is substantial. “A lot of disease hides beneath the radar,” he says.

Beyond the definitional detour, laboratorians will likely hear the same two aforementioned concerns from clinicians: so-called false-positives, and, as Dr. Jaffe puts it, “What are we going to do with these large numbers of consults?”

This spring Dr. Apple spoke to cardiologists and ED physicians at another health system, which transitioned from a less sensitive assay to a contemporary one. The big concern, not surprisingly, was the jump in positives.

His suggestion is to calculate a delta, using 500 or 1,000 patients from the institution’s database and based on the ordering pattern. He recommends zero-, three-, and six-hour measurements. (Hennepin switched to that pattern from a zero-, four-, eight-hour pattern two years ago. He expects expert guidelines to be published before the end of the year suggesting 0-3-6 is better, based on data showing even three hours can deliver a very high negative predictive value.) “Pretty much everyone who’s done this demonstrates a change of their specificity from zero hours at 70 percent, and with the three-hour delta, an improvement to greater than 90 percent.”

“If you have a lab information system, you should be able to go mine your own data,” he says.

Despite the benefit of calculating a delta, most labs aren’t doing it, Dr. Apple says. (He guesses five percent.) Using a delta will go a long way in helping to educate clinicians, but he says the burden is huge. Few clinicians understand that troponin assays are all different, he says, and that one delta will not suffice. “They can’t treat it like an INR.”

Dr. Jaffe is less confident than Dr. Apple, though he makes it clear he supports the use of deltas. “Labs should do it—don’t get me wrong.”

The analytic component is easy to set up. But the biologic component is not trivial. Labs must work with clinicians to find the right balance between sensitivity and specificity. “Otherwise, the lab is going to say, ‘We did it perfectly,’ but the patient died,” Dr. Jaffe says. All physicians think they know more than they do, he says, and each group tends to look at its own circumstances as the index one. That’s why no decision should be made in a vacuum.

Dr. Apple concurs. “Clinicians in general—and this is bad that I’m going to say this—but while they know their clinical practice well, they do not pay close enough attention to what the cardiac troponin biomarker flag is telling them. Everyone thinks they know how to interpret the biomarker. But they don’t. I would say a large majority of my clinical colleagues still need to be educated.

“I’m amazed—and I’m fortunate to talk to a lot of clinicians—that they don’t understand some of these things, like looking at timed specimens,” he continues. “They still think every time a troponin is increased it should be an MI. They say, ‘Gee, I don’t like this, let’s go back to CK-MB.’”

He doesn’t blame clinicians, however. “I put the responsibility on my laboratory colleagues, whether you’re a PhD or MD, to get the key papers for the assay you use. Make a strong effort to get them to your clinical colleagues, so they understand the troponin assay.”

That sets up what could be the start of a joke: How do you get a clinician to listen to a laboratorian?

“What works is talking to them,” says Dr. Apple. “You can’t send out memos. You can’t put a paper in the mail to them. Just arrange a time to have a conversation.”

He estimates that troponin is one of about half a dozen tests where labs have to put in the extra educational effort. TSH and PSA have been others. His approach: He meets quarterly with relevant groups of physicians. “You need to be in the same room with them. You show five slides, you hand them the key papers. Ideally you’ll find one clinical colleague who’s willing to champion it within each department.”

The best way for laboratories to educate ED physicians, Dr. Hollander suggests, is to include an explanation of the result on the lab report. “That reaches everyone who reads the lab report.” Relying on Web-based CME programs, or manufacturers’ reps, he says, won’t work.

Feeling discouraged? Dr. Collinson offers some reassuring words from London.

When he and his colleagues switched to a high-sensitivity troponin assay several years ago, two reasons loomed large: to be able to measure troponin in so-called normal patients, and to achieve earlier diagnoses. Both goals, he says, have been met.

He discussed the switch with the cardiologists, explaining the reasons for the change. “They were perfectly happy,” Dr. Collinson says cheerfully. “They more or less said to me, ‘Well, you just tell us what to do. You’re the expert.’” Jokes Dr. Collinson, “The nice thing about our cardiologists is we practice democracy: one man, one vote—my vote.”

But agreement implies discourse. “You need to have the discussion,” says Dr. Collinson.

Familiarity has bred comfort with the assay. The only downside has been the rise in non-ACS cases in patients with slightly elevated troponin. “We have had one or two instances of cardiologists who have been a bit peeved because of the increased referrals, some of which they think are inappropriate,” Dr. Collinson says. And in a couple of cases, patients were put into ACS pathways when they shouldn’t have been. “The physicians were seduced by the troponin.”

The emergency physicians are also comfortable with the assay, he reports, mostly because it enables them to move patients out of the ED even more quickly than before. His colleagues are now quite comfortable sending patients home if their troponin hasn’t gone up over the course of three hours from hospital admission.

However, “We have had one or two grumbles from our acute medical physicians [who receive patients from the ED].” He attributes the complaints to a couple of physicians who, he says, tend to be overly dependent on the troponin to guide their decisions.

For all the hand-wringing, Dr. Collinson says that if physicians keep their clinical wits about them, the rise in troponin in cases not involving ACS “isn’t such a big deal.” It all goes back to the fact that physicians need to remember the diagnosis of MI isn’t a laboratory diagnosis; it’s a clinical one. (He’s allowed to say this. Like Dr. Jaffe, he’s got a foot in both the lab and clinical camps, as a practitioner in the emerging specialty of preventive cardiology.)

For Dr. Collinson, using a high-sensitivity assay has been worth the effort. “It’s a great improvement over where we were before. We would never go back.”