Si’ for CT for PE

Abstract & Commentary

Synopsis: Helical CT scan is a definitive way to rule out significant pulmonary embolism.

Source: Donato AA, et al. Arch Intern Med. 2003;163:2033-2038.

This was a prospective, observational study of outcomes of patients who underwent helical computed tomography (CT) scans when they were suspected of having pulmonary emboli. The aim was to determine the outcome of those patients with negative scans who did not receive anticoagulation. There were 433 sequential patients enrolled. Patients were excluded from follow-up analysis for many reasons, including indeterminate CT (14 patients), positive studies with anticoagulation initiated (119 patients), anticoagulation begun or continued without respect to scan results (57 patients), and high-probability V/Q scan with negative CT (2 patients). Thus, there were 243 patients who had negative helical CT results for pulmonary embolus (PE) who were not anticoagulated. Follow-up at 3 months was accomplished for 239 of them. Their mean age was 59 years, and 98 of them (41%) had one or more risk factors for PE. Their pretest probability1 for PE was low probability for 65%, moderate probability for 27%, and high probability for 8%.

Follow-up was accomplished by phone (63%), direct encounter (17%), or review of subsequent hospital records (19%). A total of 33 patients died in the follow-up period; one of these deaths was felt by reviewers to be suspicious for PE. The causes of death in this group were lung cancer (5 patients), sepsis (5 patients), pneumonia (4 patients), disseminated carcinoma (5 patients), leukemia/lymphoma (4 patients), COPD (2 patients), and a variety of other causes for one case each.

Four patients who had a negative CT and did not receive anticoagulation had venous thromboembolic events (1.7 %). Two had DVT, 2 had PE. One of those with PE died, and it is worth examining this death in some detail. This patient was admitted with Proteus mirabilis sepsis and developed Doppler-documented venous thrombosis associated with an indwelling catheter 5 days after her admission and negative CT. Conservative care was undertaken as requested by her family, and she died on hospital day 11. Sepsis was listed as the cause of death.

Comment by Barbara A. Phillips, MD, MSPH

Everything about venous thromboembolic (VTE) disease and its workup is risky and complicated. The entity itself can kill, the diagnostic modality of choice (pulmonary angiogram) is morbid and fallible, and treatment (anticoagulation) is arduous. PE is estimated to occur somewhere between 50,000 to 175,000 times annually in the United States,2 but the condition is suspected 3 or 4 times more frequently than it is proven to occur. Pulmonary angiography, long considered the diagnostic gold standard, is difficult to interpret and expensive. It is associated with a complication rate of 6.5% and a death rate of 0.5% and often requires getting a radiologist out of bed. Clinicians have long searched for a better diagnostic test, and clinical prediction formulae, radionuclide scans, Doppler’s, and d-dimer blood assays all have some promise and usefulness in the diagnosis of VTE.

Helical CT has reasonable sensitivity (86-96%) for central emboli,3,4 but there has been concern about its ability to detect subsegmental emboli.5 The current paper is important because it indicates that the detection of subsegmental emboli may not matter much in a real-world situation. What is important is the clinical outcome of the patient. Not anticoagulating those patients with negative CTs in this study resulted in a 0.4% death rate, which is actually similar to the death rate associated with pulmonary angiography itself. Several other studies have reported similar findings, including follow-up of 6 or more months.6,7 The current study is notable in that it is the first such study to be conducted at a community hospital without full-time academic thoracic radiologists, and thus, most likely represents a "real life" condition. Donato and associates compare the result of this and other studies of helical CT with outcomes for V/Q scans and note that low-probability V/Q scans are associated with a 2.7-3.7% of patients with subsequently documented VTE if anticoagulation is withheld;6,8 they also note high interobserver variability for low- and intermediate-probability V/Q scans, which is much less troublesome with helical CT.

Move over, angiograms. Helical CT is here! 

Dr. Phillips, Professor of Medicine, University of Kentucky; Director, Sleep Disorders Center, Samaritan Hospital, Lexington, KY, is Associate Editor of Internal Medicine Alert.


1. Wells PS, et al. Ann Intern Med. 1998;129:997-1005.

2. Silverstein MD, et al. Arch Intern Med. 1998;158: 585-593.

3. Remy-Jardin M, et al. Radiology. 1992;185:381-387.

4. Van Rossum AB, et al. Thorax. 1996;51:23-28.

5. Goodman LR, Lipchik RJ. Radiology. 1996;199:25-27.

6. Garg K, et al. AJR Am J Roentgenol. 1999;172: 1627-1631.

7. Lomis NNT, et al. J Vasc Interv Radiol. 1999;10: 707-712.

8. Hull RD, et al. Arch Intern Med. 1994;154:289-297.