Am Fam Physician. 2009;80(7):742
Background:d-dimer testing is commonly used to help exclude venous thromboembolism (VTE), such as pulmonary embolism and deep venous thrombosis (DVT). Although several guidelines recommend that the d-dimer test be used in addition to the clinical probability of a thrombotic event, physicians may be tempted to use a normal d-dimer test result as stand-alone proof that VTE is unlikely. Gibson and colleagues prospectively evaluated consecutive patients with clinically suspected pulmonary embolism to determine how often d-dimer testing fails when clinical probability is not taken into account.
The Study: Initial risk was assessed using a clinical decision rule (see accompanying table). Patients with low clinical probability received d-dimer testing. Patients with low clinical probability and normal d-dimer levels of 0.5 mcg per L (0.5 mg per L) or less had pulmonary embolism ruled out. All other patients were further evaluated with spiral computed tomography (CT). Reimaging was performed in patients over the next three months if they had further symptoms indicating DVT or pulmonary embolism. Exclusion criteria included pregnancy, low-molecular-weight heparin use for longer than 24 hours at initial evaluation, and hypersensitivity to iodinated contrast fluid.
| Clinical finding | Points |
|---|---|
| Clinical signs and symptoms of DVT (leg swelling and pain with palpation of the deep leg veins) | 3 |
| Alternative diagnosis is less likely than PE | 3 |
| Heart rate greater than 100 beats per minute | 1.5 |
| Surgery or immobilization in past four weeks | 1.5 |
| Previous DVT or PE | 1.5 |
| Hemoptysis | 1 |
| Malignancy (treatment in past six months or receiving palliation) | 1 |
Results: Overall, 1,722 persons were evaluated. The mean age was 54 years, and 78 percent were outpatients. Pulmonary embolism was ultimately identified in 378 patients (22 percent). Forty-five percent of patients with low clinical probability of VTE had a normal d-dimer test compared with 15.2 percent of patients with high clinical probability.
There were 563 patients with a normal d-dimer test; 477 of these patients were calculated to have low clinical probability and 86 to have high clinical probability. Five (1.1 percent) of the patients with a normal d-dimer test and low clinical probability were diagnosed with VTE during the follow-up period. In contrast, eight (9.3 percent) of the patients with a normal d-dimer test and high clinical probability were diagnosed with VTE, of whom all but one were diagnosed at baseline via spiral CT.
Conclusion: The authors recommend that physicians ignore a normal d-dimer test when there is a high clinical probability of VTE. This study reiterates the importance of clinical assessment when evaluating VTE, and that high-risk patients should have further testing, regardless of d-dimer levels.