Clinical Citations: Fine-tuning the evaluation of acute pulmonary embolism

March 1, 2006

One of the keys to successfully managing acute pulmonary embolism (PE) is to quickly and reliably identify which patients require anticoagulant therapy. A diagnostic algorithm that combines clinical probability, D-dimer results, and CT findings appears to be quite effective in achieving this, according to a multicenter study conducted in the Netherlands.

One of the keys to successfully managing acute pulmonary embolism (PE) is to quickly and reliably identify which patients require anticoagulant therapy. A diagnostic algorithm that combines clinical probability, D-dimer results, and CT findings appears to be quite effective in achieving this, according to a multicenter study conducted in the Netherlands.

The study included 3306 patients in whom acute PE was clinically suspected. The Wells clinical decision rule was adapted so that it could be used to categorize the patients' PE as "likely" or "unlikely." Those with unlikely PE underwent D-dimer testing, and PE was ruled out if the results were normal. Patients with likely PE underwent CT scanning. Anticoagulant therapy was withheld from patients in whom PE was ruled out. Patients were monitored for 3 months.

PE was considered unlikely in 2206 patients; D-dimer test results were normal in 1057 of these patients. Nonfatal venous thromboembolism (VTE) subsequently occurred in 0.5% of those who did not receive anticoagulants.

The results of CT ruled out PE in 1505 patients. Among those who did not receive anticoagulants, the incidence of VTE was 1.3%. PE was considered a possible cause of death in 7 patients who had normal CT findings.

In an accompanying editorial, Hull notes that spiral CT now plays a key role in the evaluation of PE and that it is more widely available than ventilation-perfusion lung scanning. However, single-detector spiral CT does not have adequate sensitivity to warrant its use as a sole diagnostic test for PE; rather, it should be used in conjunction with the assessment of clinical probability and D-dimer testing or ultrasonography.

In addition, Hull points out that simplifying the Wells decision rule so that PE would be classified as either likely or unlikely--with no intermediate category--makes the decision rule more practical.

In a separate study, Wells and associates evaluated the accuracy of clinical prediction rules for deep venous thrombosis (DVT) and the use of D-dimer testing in conjunction with the clinical probability estimate. Their analysis included 14 studies involving more than 8000 patients.

They found that diagnostic accuracy improved when clinical probability was assessed before diagnostic tests were done. Wells and associates concluded:

• A validated clinical prediction rule should be used to assess outpatients in whom DVT is suspected. The prevalence of DVT is less than 5% in patients with a low probability of DVT.

• Once the clinical probability has been estimated, the results of D-dimer testing can be used to help decide whether DVT can be ruled out without the use of diagnostic imaging. In low-probability patients who have a negative D-dimer test result, DVT can be ruled out without ultrasonography.

• All high-probability patients should undergo diagnostic imaging, regardless of the D-dimer result.

• The specificity of D-dimer testing decreases as clinical probability increases, and false-positive results are more likely. D-dimer should not be used as a screening test.