Diagnosing pulmonary embolism: results vary among the tests

Journal of Family Practice, Nov, 2005

Roy PM, Colombet I, Durieux P, Chatellier G, Sors H, Meyer G. Systematic review and meta-analysis of strategies for the diagnosis of suspected pulmonary embolism. BMJ 2005; 331:259-268.

* Clinical Question In patients with suspected pulmonary embolism, which tests are effective in diagnosing or ruling out the condition?

Study Design Meta-analysis (other)

Setting Various (meta-analysis)

Synopsis

To evaluate the different diagnostic tests available to make the tricky identification of PE, the authors searched Medline, Embase, and Pascal Biomed for English-language studies evaluating such tests. They also searched bibliographies and their personal libraries. Their search strategy did not include attempts to identify unpublished research and excluded articles not written in English, which are 2 minor limitations.

Two reviewers independently selected studies for inclusion and extracted the data, and limited the studies prospective studies that recruited consecutive patients, used pulmonary angiography as their reference standard, and for which the test being evaluated and the reference test were interpreted without knowledge of the results from the other study. The 48 studies selected for inclusion had an average prevalence of embolism of 30%.

The results were reported in terms of likelihood ratio (LR): as a general rule, an test with an LR of less than 0.1 successfully excludes disease and a test with an LR of greater than 10 confidently identifies the disease. For exclusion of PE, negative LRs were: normal lung scan, 0.05 (95% confidence interval [CI], 0.03-0.1); negative spiral CT plus negative ultrasound, 0.04 (95% CI, 0.03-0.06); and a D-dimer concentration of less than 500 [micro]g/L by enzyme-linked immunosorbent assay (ELISA), 0.08 (95% CI, 0.04-0.18). At low probability, these tests lowered the likelihood of PE to less than 5%. D-dimer tests using other methods were not as effective, nor were leg ultrasound, magnetic resonance angiography, and echocardiography.

For diagnosing PE, the LRs were: high probability ventilation perfusion scan, 18.3 (95% CI, 10.3-32.5); spiral CT, 24.1 (95% CI, 12.4-46.7); and leg ultrasound, 16.2 (95% CI, 5.6-46.7). At high probability, these tests increased the likelihood of disease to 85%. Echocardiography is not useful and magnetic resonance angiography did not produce consistent results.

Copyright[c] 1995-2005 InfoPOEM, Inc. All rights reserved. www.infopoems.com.

* Bottom Line

Some tests are better at diagnosing pulmonary embolism (PE) and some are better at excluding it. To exclude PE in patients with a low likelihood of disease, use a lung scan, spiral computed tomography (CT) plus leg ultrasound, or D-dimer by ELISA. To diagnose PE in patients with a high likelihood of disease, use a ventilation perfusion scan, spiral CT, or leg ultrasound. (LOE=1a)