Anticoagulant Therapy: Safe to Withhold?

Abstract & Commentary

Synopsis: It is safe to withhold anticoagulant therapy from patients who have a negative helical CT scan for pulmonary embolism.

Source: Goodman LR, et al. Radiology 2000;215:535-542.

Since the advent of helical scanners in the early 1990s, there has been progressive improvement in the ability of CT angiography of the pulmonary arteries to detect emboli directly without the need for the more definitive but invasive pulmonary arteriogram. While many studies published to date have focused on the accuracy of helical CT for central pulmonary emboli, reported as more than 90%, there have been limited data on the safety of withholding anticoagulant therapy in patients whose helical CT pulmonary angiograms are interpreted as normal (i.e., the negative predictive value of CT pulmonary angiography). In fact, it is this limitation of available helical CT data that has led several authors to conclude that the use of helical CT in place of ventilation/perfusion scintigraphy or pulmonary angiography for the diagnosis of pulmonary embolism is not currently warranted. This study seeks to answer the question of the safety of withholding anticoagulant therapy in patients with suspected pulmonary embolism who have a CT pulmonary angiogram interpreted as negative for pulmonary embolism.

At Goodman and colleagues’ institution, patients with suspected pulmonary embolism who had either helical CT angiography or ventilation/perfusion (V/Q) scan performed as the initial diagnostic test were enrolled beginning in 1995. The study group consisted of 548 patients who were followed clinically for three months after one of the following three test results: a normal helical CT angiogram, a normal V/Q scan, or a low probability V/Q scan. The negative predictive values of each of these three test results were calculated based predominantly upon clinical followup of patients to determine whether they developed recurrent pulmonary embolism or thromboembolic disease in the three month period following the index examination.

The incidence of subsequent thromboembolic disease in the patients with negative helical CT exams, normal V/Q scans, and low probability V/Q scans was 1%, 0%, and 3%, respectively, resulting in negative predictive values of 99%, 100%, and 97% for each test, respectively. Goodman et al concluded that the incidence of recurrent thromboembolic disease following a negative CT exam is comparable to that following a normal V/Q scan, and the high negative predictive value of helical CT pulmonary angiography indicated that anticoagulation therapy can be safely withheld following negative examinations.

Comment by Jeffrey S. Klein, MD

While this paper attempts to provide further rationale for using helical CT angiography in the evaluation of suspected pulmonary embolism by showing a high negative predictive value of this examination, there are some significant limitations that preclude drawing widespread conclusions regarding the exact utility of helical CT pulmonary angiography in this setting. The patient populations that comprised the two main study groups with negative test results, those who underwent helical CT and those who had V/Q scans, differed significantly in several respects, most importantly in the source of patient referral and in the performance of Doppler ultrasound examination of the lower extremities—the latter known as a reliable predictor of subsequent thromboembolism and patient outcome. The nature of the follow-up in the patients with negative study results makes it impossible to determine the true incidence of recurrent thromboembolic disease, and a significant number of patients were either lost to follow-up or excluded from consideration because of the administration of anticoagulants. Therefore, more outcome data, preferably with the combined use of CT pulmonary angiography and venography in a randomized, controlled trial with pulmonary angiography as the gold standard and a standardized method of patient follow-up, will be needed before the precise role of helical CT in patients with suspected thromboembolic disease can be determined.


1. Rathbun SW, et al. Ann Intern Med 2000;132:227-232.