Source: Birdwell BG, et al. The clinical validity of normal compression ultrasonography in outpatients suspected of having deep venous thrombosis. Ann Intern Med 1998;128:1-7.
The lead article for the january 1998Annals of Internal Medicineis of import to emergency physicians. Daily, we are confronted with patients in whom we suspect deep vein thrombosis (DVT). There is misconception about how to manage patients whose initial ultrasound study is negative. This study focuses exclusively on outpatients with first-time suspicion of DVT, simplifies their outpatient work-up, and provides comfort for withholding anticoagulation.

At an academic center and VA hospital, Birdwell and colleagues enrolled consecutive outpatients suspected of first-episode DVT. The main objective of the study was to test the safety of withholding anticoagulation if the initial ultrasonography was normal and if the repeated study (day 5 to 7) was also normal. The main outcome measure was pulmonary embolism or symptomatic DVT within three months of initial enrollment. The patients were enrolled based on clinical suspicion of DVT. Simplified real-time B-mode compression ultrasonography was performed, assessing compression of the common femoral and popliteal veins. All patients with a normal study were tested again as outpatients on days 5-7. All patients with an abnormal ultrasound were recommended for confirmatory venography. Any patient with an abnormal ultrasound on follow-up was also recommended for venography. All patients with normal studies, despite symptoms, had anticoagulation withheld.

In all, 405 patients were enrolled. Three hundred forty-two had normal initial studies, and 63 were abnormal. In the normal cohort, seven patients had abnormal ultrasounds on repeat testing, for a total of 70 patients with abnormal studies. Unfortunately, only 37 of 70 had venography performed, with 23 definitive proximal DVTs by venogram and five normal studies. All 65 of the remaining patients with abnormal studies (ultrasound and/or venogram) were started on anticoagulation therapy.

The main question examined by Birdwell et al was whether it is safe to withhold anticoagulation in the face of two serial normal ultrasounds; all patients were followed for symptoms for three months. Two patients in the normal cohort developed symptomatic and confirmed thromboembolic disease (2 of 335, 0.6%), both of whom were cancer patients. Four patients in the abnormal cohort developed recurrent disease (4 of 70, 5.7%). No patient died of pulmonary embolism. Birdwell et al conclude that anticoagulation may be safely withheld in first-time suspects of proximal DVT if they have two normal simple compression ultrasounds 5-7 days apart.


I like this study for several reasons. For one, it is very well written with clearly defined objectives, main outcome measures, and conclusions. Second, the article clarifies a muddy topic area and provides a simple, cost-effective outpatient approach. Third, a growing cadre of emergency physicians are using ultrasound, and the article reviews the newest literature, describes the technique in detail, and makes it "reachable" for the emergency physician (although, the reader should realize that the ultrasounds performed in this study were done by noninvasive vascular lab technicians).

Previous studies have suggested that anticoagulation may be withheld with a normal initial ultrasound and two negative follow-up studies. This study convinces me that one initial normal and one follow-up normal ultrasound provide a high margin of safety for these patients, a group for which I would comfortably withhold anticoagulation. Birdwell et al estimate a $100 million cost savings could be realized by decreasing from three serial studies to two ultrasounds. Depending on your area of practice, the normal ultrasound group could return to a "fast-track" area of the ED for follow-up to receive their second study. In good technical hands, Birdwell et al claim the simple compression technique takes 15 minutes to perform.