One-two punch in evaluating
  PE risk

November 2006

Feature Story

Karen Lusky

In evaluating the cause of dyspnea in the emergency department setting, physicians typically order D-dimer to help rule out pulmonary embolism—and the natriuretic peptides to identify heart failure. Yet tantalizing new data suggest that combining D-dimer and natriuretic peptide just might pack a bigger diagnostic punch than either test alone. A growing body of evidence also suggests that natriuretic peptides can help risk stratify patients who have pulmonary embolism.

As for the potential for D-dimer and NT-proBNP as a test duo, a study reported in the September 2006 Archives of Pathology & Laboratory Medicine found that the median levels of both D-dimer and amino-terminal pro-B-type natriuretic peptide (NT-proBNP) were higher in patients with acute pulmonary embolism.

In the study, the tests combined showed 85 percent specificity for PE compared with 72 percent for D-dimer and 58 percent for NT-proBNP, respectively. That means the tests combined were 85 percent effective at identifying people who didn’t have a pulmonary embolism, which is a “big jump” in specificity, says study co-author and pathologist Michael Laposata, MD, PhD, of Massachusetts General Hospital, Boston.

The study showed that the combination of D-dimer and NT-proBNP maintained the high negative predictive value of D-dimer for ruling out PE, which is 99 percent for an immunological-based assay that has the appropriate cutoff for excluding the condition. (NPV tells you the percentage of patients with negative results who don’t have the disease for which you are testing, which in this case is PE.)

Yet despite D-dimer’s high NPV for pulmonary embolism, some clinicians still hesitate to use it in excluding PE without performing additional expensive radiographic studies, given the consequences of missing a case of PE, Dr. Laposata notes. So the researchers wanted to know if a combination of D-dimer and NT-proBNP could improve the ability to sort out dyspneic patients with or without PE. “That was the framework with which we conducted the study,” he says

To answer that question, the researchers retrospectively selected and analyzed a subset of patients with acute PE from a previously performed study, the ProBNP Investigation of Dyspnea in the Emergency Department, or PRIDE, study of 599 ED patients with dyspnea. Computed tomographic angiography exams had confirmed that 19 of those 599 patients had acute PE. For the study reported in Archives, Stacy E.F. Melanson, MD, PhD, a pathologist at Brigham and Women’s Hospital, and research cohorts compared the 19 patients with acute PE to a randomly selected control group of 199 “normal” patients in the PRIDE study who didn’t have acute PE, acute congestive heart failure, or acute coronary syndromes. (The PRIDE study by design excluded people with renal failure.)

Not a “single patient with acute PE had normal values for both NT-proBNP and D-dimer,” according to the Archives study.

“Patients without PE had a median D-dimer level of 0.30 FEU/mL” compared with 2.83 FEU/mL for those with PE, the researchers write. (The study used the Roche Tina-quant D-dimer immunoturbidimetric assay.) “Dyspneic patients without PE had a median NT-proBNP level of 66 pg/mL.” Those with PE had a whopping 1921 pg/mL. The study used the Elecys NT-proBNP test with a cutpoint of 74 pg/mL for excluding PE.

How D-dimer and the natriuretic peptide testing work in tandem isn’t a chicken-and-egg phenomenon. The clot obviously always comes first—hence, a positive D-dimer. If the clot produces a distressing effect on the heart, it releases BNP and NT-proBNP.

“D-dimer is going to be positive in the dyspneic patient who has PE because D-dimer identifies the process of thrombosis and thrombus turnover...,” the study’s principal investigator, James L. Januzzi, MD, a cardiologist at Massachusetts General Hospital, told CAP TODAY. As Dr. Melanson notes, the natriuretic peptides are “presumably released secondary to strain on the heart” when a clot in the lung causes “increased pulmonary vascular resistance.” The Archives article notes that “elevations of natriuretic peptides have been described in patients with PE, particularly those with right ventricular strain.”

D-dimer is used for ruling out PE, but it can’t be relied on to help rule in the condition. As internist Mark Wurster, MD, of Ohio State University Medical Center, Columbus, explains: “A lot of people have elevated D-dimer tests” due to other conditions—for example, “liver disease, active inflammation from rheumatoid arthritis, occult malignancy, or disseminated intravascular coagulation, etc.”

The Archives study confirmed that a good portion of the ED physicians didn’t trust a negative D-dimer to rule out PE without ordering computed tomography angiography. Twenty-five percent (six out of 24 patients) with negative D-dimer tests in the PRIDE study received unnecessary CTAs that confirmed they did not have PE. The study did not identify those patients’ pretest clinical probabilities of having PE, but Dr. Januzzi says he would assume the patients had “at least a moderate likelihood.” Four of those six patients who had negative D-dimers and CTAs had NT-proBNP readings under the cutoff (74 pg/mL) that the researchers found the data in the study supported for excluding pulmonary embolism.

Dr. Melanson is quick to point out that the study reported in Archives doesn’t prove that clinicians would have skipped the CT angiography because of the additional negative pro-BNP. But a future study and “one valuable to do” could test that hypothesis, she adds. The extra cost of adding BNP to D-dimer is $20, she notes, which is “the incremental cost for the lab,” including labor and reagents.

Dr. Januzzi says the “emphasis” for conducting the study and reporting its findings is that adding NT-proBNP to D-dimer can help clinicians be more “confident in their decisionmaking” when evaluating patients for potential PE. As a result, he says, clinicians might potentially avoid ordering a radiographic study that exposes the patient not only to significant radiation but also “100 cc of IV contrast [material], which comes with the risk of kidney failure.”

And “since physicians normally order NT-proBNP for patients with dyspnea to rule out decompensated congestive heart failure,” Dr. Januzzi continues, “it can make sense for them to use what they already have available in evaluating the potential for PE,” which would be the two tests combined.

“Using similar tools for identifying different disease states is something we don’t do well,” he adds. “Sometimes we miss seeing the forest for the trees. The goal is to avoid potential unnecessary testing for patients with suspected PE.”

Studies indicate that natriuretic peptide testing has prognostic prowess in evaluating patients with PE. A study reported in Circulation in 2003 found that “a proBNP level <500 pg/mL identifies patients who will be potential candidates for an abbreviated hospital length of stay or care on a completely outpatient basis,” according to the study abstract (Circulation. 2003 Apr 1;107[12]:1576–1578).

Patients diagnosed with PE receive anticoagulation therapy on an outpatient or inpatient basis to help prevent clot extension and the development of additional clots. But patients who have a sufficiently large clot in the lungs—or one that is significantly taxing the heart—may require a thrombolytic agent or clot-busting drug. And natriuretic peptide testing may be able to flag candidates for the more aggressive therapy. Thrombolytic drugs “come with the risk of fatal hemorrhage, so it’s not a walk in the park to give them,” Dr. Januzzi says.

Though the study reported in Archives “lacked sufficient numbers to look at the role of NT-proBNP in prognosis,” he says, existing data show a “powerful relationship between the release of NT-proBNP with PE and the presence of right ventricular overload in those patients.” (Similar data exist for BNP, he adds.)

“When the patient has a CT angiogram showing a clot [in the lung] and also has a high NT-pro-BNP,” he or she is likely to have “some right ventricular overload,” Dr. Januzzi continues. “Those are the patients who are going to run into trouble.”

Dr. Januzzi has seen several cases in which patients who show a clot on their CT exam and right ventricular dysfunction on an echocardiogram have received thrombolytic agents. And as the “patient’s PE dissolves and the right heart function improves, the levels of NT-proBNP fall rapidly—within 12 to 24 hours of improvement in heart function.” Thus, “given the relationship between therapy and pro-BNP, we find that serial measurements might provide more information in terms of stratifying the patient’s level of stability and response to therapy,” Dr. Januzzi says.

Cleveland Clinic emergency department physician Frank Peacock, MD, agrees that people with a big PE have elevated natriuretic peptide levels and fare worse. So the idea of using those levels as a risk stratifier is “barking up the right tree.” But in high-risk patients, a negative D-dimer to rule out PE is not clinically helpful, he says. “If someone is high risk, you go ahead with some type of a scan.”

Dr. Peacock agrees that the odds are that a low-risk patient with negative D-dimer results doesn’t have PE—“that’s been pretty well established.” And “it’s probably helpful if BNP is negative, which tells you that if the person has a PE, at least it’s a small one because it hasn’t produced a hemodynamic effect.”

Even so, the ED physician has to consider that a small PE left untreated could grow into a “big clot that becomes a fatal event.” Thus, he says, “it’s not sufficient to only say ‘You don’t have a big PE.’ We have to say, ‘You don’t have any PE.’”

Expert clinical assessment plays a major role in calculating the odds that a person with a negative D-dimer doesn’t have a lurking pulmonary embolism. Cardiologist Paul Stein, MD, notes that clinicians can use various objective criteria, such as the Wells or Geneva scoring systems, to determine the pretest clinical probability that a person has PE.

“If the clinical assessment reveals a low probability of the patient having PE (a 10 percent or less chance) and the D-dimer is negative, then the patient has only a one percent chance of having PE,” says Dr. Stein, director of research education at St. Joseph Mercy-Oakland Hospital, Pontiac, Mich. “If the clinical exam shows a 40 percent probability of the person having PE, and a quantitative rapid ELISA is negative, then the patient has about an eight percent chance of having PE.” With other methods for testing D-dimer, he says, the probability of PE with a negative D-dimer and a 40 percent clinical probability of PE is higher.

Of course, the utility of a clinical risk assessment and combination of D-dimer and natriuretic peptide testing to rule out PE is predicated upon D-dimer having a negative predictive value of 99 percent. And one could easily imagine a scenario in which a patient determined by clinical assessment to be low risk for PE who actually had a small PE could have a false-negative D-dimer and a truly negative NT-proBNP. Says Dr. Januzzi: “With D-dimer, if we honest to goodness are going to trust the technology, then the technology has to perform as promised. When you talk about NPV, you can’t absorb a false-negative more than once.”

Yet some of the so-called false-negatives that clinicians count as an example of being burned by D-dimer testing may not be so negative if clinicians take a closer look. As Dr. Januzzi puts it, “There’s negative and there’s really negative. It’s all about where you put your cut point.”

A similar debate rages about that now for BNP and NT-proBNP, he says. “We should be thinking of these compounds as being released on a continuum. They don’t just turn positive and you have a disease. There are gradients of disease and release of BNP and NT-proBNP. Someone who has 5 picograms under the positive cut point probably has a disease.”

Given the limitations of an assay, one also has to ask, If the patient has a test result “right under the cutoff, might they really be over it?” says pathologist John D. Olson, MD, PhD, of the University of Texas Health Sciences Center at San Antonio. Thus, with “discretion being the better part of valor, if one has a low-risk patient and a value of D-dimer at the threshold, it’s probably wise to pursue that with imaging.”

Clinicians should work with their laboratories to make sure they understand the ability of the D-dimer test to exclude PE, Dr. Olson advises. “Quantitative ELISA assays are the gold standard,” he notes. “Vidas is the one that got the ball rolling using a rapid methodology.” But he “doesn’t think there are good data to indicate it’s much better than a well-controlled microagglutination test,” which is the test used in the study reported in Archives. “Whatever test you use, you have to know what the data show in terms of the appropriate threshold for excluding PE,” Dr. Olson says.

Laura Worfolk, PhD, U.S. scientific and clinical research manager for Diagnostica Stago, Parsippany, NJ, emphasizes that “the manual slide latex agglutination method, which requires subjective interpretation, is less sensitive and should not be used for ruling out deep vein thrombosis or PE.”

While the Archives study supports the hypothesis that NT-proBNP might bring something additional to the diagnostic table for PE, the researchers agree that further studies are needed to validate their conclusions. (Dr. Januzzi says that because NT-proBNP is slightly more sensitive than BNP, one cannot necessarily say the study results are “superimposable” for BNP. But he says “useful data” suggest a similar role for BNP in the context of PE.)

St. Joseph Mercy-Oakland Hospital’s Dr. Stein views the study results reported in Archives as “highly preliminary.” He says “it’s likely that BNP is a good risk stratifier for patients with PE.” And if “both the D-dimer and BNP are negative, one might be somewhat more sure there’s no PE, but the present data indicate the negative predictive values of D-dimer alone and D-dimer plus BNP are quite similar.”

Dr. Wurster says a lot of BNP and D-dimer testing is done in his practice at Ohio State University, though he doesn’t combine the two “for ruling out PE per se.” But, he adds, “It sounds like adding BNP to D-dimer where they are automatically linked for evaluating patients for shortness of breath might be a good combination.”

Biosite Inc., San Diego, has actually done that with its Triage Profiler Shortness of Breath Panel, which includes BNP, D-dimer, and cardiac ischemia markers. Ken Buechler, PhD, chief scientific officer and president of Biosite, says the Archives article “might spark more interest in researching D-dimer and BNP together in the setting of suspected PE.” But he views the study as falling short in recognizing patients presenting with dyspnea who have suffered a silent heart attack as the reason for their shortness of breath.

In Dr. Peacock’s view, the Archives study is “severely limited by not taking all comers.”

“In the ED, we don’t get to exclude patients once it becomes clear they don’t have heart failure or acute coronary syndrome... In real practice you get them all and must take your best shot as they come in the door.” And “that is how a practice-changing study must be designed,” he says. The study suffers, too, from small numbers, he says: “Nineteen pulmonary emboli just isn’t enough, so ultimately its results are hypothesis-generating only.”

Still, the study “does answer the question of ‘Should we study the use of natriuretic peptides in combination with a D-dimer in dyspneic patients further?’ That answer is clearly yes,” Dr. Peacock says.

But the University of Texas’ Dr. Olson emphasizes that the study wasn’t multi-centered. “The researchers are looking specifically at 19 individuals who presented with dyspnea and [were shown through imaging to have PE]...I don’t think you are going to get more than that in many places. I’m sure everyone would love to see 10 times that number, but I don’t think that’s going to occur due to practicalities of performing such a study.”

The logistics and cost of conducting an all-comers study can be an obstacle, Dr. Peacock acknowledges. If he were to design a study of people over the age of 45 with comorbidities presenting in the ED with shortness of breath and chest pain who receive D-dimer and natriuretic peptide testing, “2,000 patients later, we’d have data to divide into subsets and answer all the questions,” he says. The study would cost “a pile of money but the answers could result in clinicians not ordering a lot of CTA exams.” Unfortunately, he adds, one doesn’t get money for not performing health care services.

Dr. Januzzi points out that if a patient with dyspnea had ACS—in which case D-dimer and pro-BNP may be elevated—use of cardiac markers would presumably identify the problem. “In the future,” he predicts, “patients will have a panel run of perhaps 10 valuable markers, with different patterns of marker results meaning different things.”

Biosite is, in fact, conducting clinical trials to develop a multimarker index, or MMX, that will use an algorithm to calculate a single value based on the concentrations in the panel. “Using the MMX, the doctor would not have to look at each individual concentration and use his or her knowledge of D-dimer, BNP, and other markers to make a decision about the diagnosis of the patient,” says Dr. Buechler. “The algorithm would help define that.”

Thus, if the patient had slight elevations of D-dimer and BNP—and a positive troponin, CK-MB, and myoglobin—the algorithm would point to acute coronary syndrome as being the most likely problem. “If D-dimer is up and there’s a slight elevation in BNP but troponin is negative,” Dr. Buechler says, “most likely the patient has heart failure or PE. If BNP is elevated and D-dimer and cardiac markers are not, most likely it’s heart failure.”

Dr. Buechler says that if both D-dimer and BNP are up, in his view, “the patient needs imaging to look for PE, even if the physician believes the patient may have congestive heart failure.”

All agree on the value of doing future studies on lab testing’s role in fast-tracking the accurate diagnosis of PE in the emergency department.

“It may not be that D-dimer plus pro-BNP is the final answer,” Dr. Laposata sums up. But “it might be possible” that doing a combination of “D-dimer plus test B and C” might increase a clinician’s confidence just enough to send patients home without imaging studies.