Diagnosing pulmonary embolism

D-dimer needs rigorous evaluation

All doctors with patients suspected of having an acute deep venous thrombosis or a pulmonary embolism would love a diagnostic magic bullet. They long for a test that would rule in or rule out the diagnosis of venous thromboembolism with a high degree of certainty. Ideally, such a test would be rapid, non-invasive, and widely available.

The search for a reliable test has continued because the clinical diagnosis of thromboembolism is neither sensitive nor specific. Neither symptoms nor examination of the legs are reliable; nor can clinical signs or symptoms define the presence or absence of pulmonary embolism. Clinical signs and symptoms can only suggest these diagnoses—and asymptomatic venous thrombosis and pulmonary embolism are both fairly common.

Not surprisingly, then, over the years a succession of simple, relatively non-invasive tests has been explored. Blood tests have included measurement of the “diagnostic” triad of lactic dehydrogenase, bilirubin, and glutamic oxaloacetic transaminase; immunoglobulin E1; arterial oxygen and the alveoloarterial oxygen gradient; and fibrinopeptide A2—and others have been proposed.3 4 Lung physiologists have suggested the use of measurements of dead space and tidal volume. Radionuclide venography enjoyed a brief career.

For various reasons each approach followed a parabolic course from the rapid ascendancy of initial excitement to rapid descent as its value failed to be validated. Still in play are techniques that use radiolabelled components of thrombus—such as platelets and monoclonal antibodies to fibrin. Radioactively labelled fibrinogen, which provided invaluable epidemiologic data, is no longer seen as the answer to the diagnosis of venous thrombosis.5 But computed tomography and magnetic resonance imaging are still being investigated.6

As this galaxy of non-invasive tests has moved across the horizon a new and promising entrant has appeared: measurement of D-dimer. D-dimer is a specific degradation product of crosslinked fibrin that is released when the endogenous fibrinolytic system attacks the fibrin matrix of fresh venous thromboemboli. Many recent research papers have made claims for the value of measuring D-dimer concentrations in patients suspected of having venous thromboembolism.7 8

One issue has been clarified: raised concentrations of D-dimer should not be interpreted as supporting the diagnosis of deep vein thrombosis or pulmonary embolism. The test is non-specific; raised values can be found in many other conditions and, indeed, in the postoperative period, a high risk time for thromboembolism. On the other hand, measurements of the concentration of D-dimer may be used to exclude the diagnosis—that is, the absence of a raised concentration implies that there is no fresh thromboembolic material undergoing dissolution in the deep veins or in the pulmonary arterial tree. Such a (nearly) exclusionary test would be valuable, especially when combined with other well studied non- invasive tests for deep vein thrombosis (impedance plethysmography, ultrasonography) and pulmonary embolism (lung scanning).

Will this promise be fulfilled? It is still too early to tell; indeed, we are some distance from knowing. Too many assays have been used in different studies—at least four enzyme linked immunosorbent assays (ELISAs) and four latex assays. Normal values differ for each: as Bounameaux et al recently said, standardisation and calibration are urgently needed.9 Furthermore, any test born in the tightly controlled incubator of research tends to slide a bit in performance when it comes into general use.

Other issues need to be addressed. Normal values tend to vary with age. How long does it take for a deep vein thrombosis or pulmonary embolism to become sufficiently organised that it no longer releases D-dimer? How severely will the value of the test be reduced by all of the competing causes for a raised D-dimer concentration in patients in hospital?

Venous thromboembolism is potentially lethal, so the margin of tolerable error is not wide. D-dimer has not yet been put to the ultimate test—a trial in which treatment is withheld from patients suspected of having deep vein thrombosis or a pulmonary embolism with D-dimer concentrations below a specified cut off value. Such trials of outcome should take account of the existing diagnostic approaches that have already gone through such trials impedance plethysmography and ultrasonography10 for detection of deep vein thrombosis and perfusion ventilation scanning11 12 and pulmonary angiography for exclusion of pulmonary embolism. In these instances, the available data indicate that treatment may be withheld without adverse outcomes.

What is the way ahead? After standardisation and calibration, assays for D-dimer must be subjected to the definitive study: they should be combined with other established tests to support decisions not to treat. Let us hope that D-dimer will not descend the parabola of other diagnostic blood tests, but only time and further investigations will tell.