© 2004 American Thoracic Society
Cost-effectiveness of Computed Tomography in Diagnosis of Pulmonary EmbolismTo the Editor:In their article, Perrier and colleagues (1) found that single-detector helical computer-assisted tomography (CT) is not cost-effective when used alone. However, when the higher sensitivity of multidetector scanners is considered, a diagnostic strategy of D-dimer/lower extremity ultrasound/CT is the most cost-effective strategy for patients with intermediate or high clinical probability. This strategy obviates the need for pulmonary angiography in all clinical probability subgroups. There have been advances in CT imaging that must be considered. CT venography studies the deep venous system from the common iliacs to the popliteal veins, approximately 3 minutes after the initiation of the injection to assess the pulmonary arteries. No additional contrast is required. A growing body of literature has reported that CT venography and lower extremity ultrasound findings are similar. In 116 patients undergoing both CT venography and lower extremity ultrasound, Cham and colleagues (2) found concordant results in 93% of cases. In four cases with discordant results, CT detected deep venous thrombosis not apparent by ultrasound (2). Garg and colleagues (3) found that CT venography had a positive predictive value of 71% and a negative predictive value of 100%. Others have reported similar results (4). The obvious advantage of CT venography is the ability to complete the venous thromboembolism workup in one 20-minute test. An intangible, but important, additional benefit is the efficient use of the time of both the physician and the patient. The disadvantages are added cost and additional radiation. In addition, a number of studies have also documented good clinical outcome when anticoagulation is withheld if no pulmonary embolism is found by helical CT (5, 6). It should be emphasized that for a patient without lung disease and with a normal chest x-ray, a ventilation–perfusion scan is more likely to provide meaningful results. Therefore, we propose the following scheme, which we believe is the most cost-effective and efficient. When there is a low clinical probability of venous thromboembolism and normal D-dimer, no further testing is necessary. With an intermediate or high clinical probability, a normal D-dimer, and a normal chest x-ray, the sequence of ultrasound, ventilation–perfusion scan, CT scan should be used. In the setting of an abnormal D-dimer and abnormal chest x-ray, go straight to CT scan with CT venography. We would be very pleased to know how Perrier and colleagues' findings (1) would change if they include this new technology, now in use in many centers, in their analysis.
Medical College of Wisconsin Milwaukee, Wisconsin REFERENCES
From the Authors:
In their comment on our article (1), Lipchik and Goodman raise two interesting issues: (1) Computed tomography (CT) venography was not included in the strategies evaluated in our decision model, and (2) our strategies were not stratified according to the chest radiograph result. CT venography is indeed an interesting technique. However, data on its diagnostic performances are still scarce and rely on small series (2, 3) that did not adhere to important methodological recommendations on the validation of new diagnostic tests (4). Moreover, to include CT venography in a cost-effectiveness analysis, we would have to know the added costs of the procedure, which are not simply its added charges to the patient or third payer, but all additional related costs (personnel, increased image acquisition time, and so on), which to our knowledge have not yet been evaluated. Be that as it may, the advantage of lower limb compression ultrasonography (performed in our strategies before ventilation–perfusion (
Lipchik and Goodman also propose that
Geneva University Hospital Geneva, Switzerland REFERENCES
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
/
) scan or CT) is that patients with a deep vein thrombosis are not submitted to CT at all. Hence, because CT venography is performed after thoracic CT and CT is more expensive than ultrasound, it is unlikely to be cost-effective, unless the added costs of CT are counterbalanced by a lower false-positive rate of CT venography compared with ultrasound, which is unlikely. Indeed, false-positive results entail hospitalization and anticoagulant treatment and, hence, higher costs. In summary, we lack important information to evaluate the cost-effectiveness of CT venography, but we would be more than willing to redo the analysis when those data are available.