Elsevier

The Lancet

Volume 371, Issue 9621, 19–25 April 2008, Pages 1343-1352
The Lancet

Articles
Diagnosis of pulmonary embolism by multidetector CT alone or combined with venous ultrasonography of the leg: a randomised non-inferiority trial

https://doi.org/10.1016/S0140-6736(08)60594-2Get rights and content

Summary

Background

Multislice CT (MSCT) combined with D-dimer measurement can safely exclude pulmonary embolism in patients with a low or intermediate clinical probability of this disease. We compared this combination with a strategy in which both a negative venous ultrasonography of the leg and MSCT were needed to exclude pulmonary embolism.

Methods

We included 1819 consecutive outpatients with clinically suspected pulmonary embolism in a multicentre non-inferiority randomised controlled trial comparing two strategies: clinical probability assessment and either D-dimer measurement and MSCT (DD-CT strategy [n=903]) or D-dimer measurement, venous compression ultrasonography of the leg, and MSCT (DD-US-CT strategy [n=916]). Randomisation was by computer-generated blocks with stratification according to centre. Patients with a high clinical probability according to the revised Geneva score and a negative work-up for pulmonary embolism were further investigated in both groups. The primary outcome was the 3-month thromboembolic risk in patients who were left untreated on the basis of the exclusion of pulmonary embolism by diagnostic strategy. Clinicians assessing outcome were blinded to group assignment. Analysis was per protocol. This study is registered with ClinicalTrials.gov, number NCT00117169.

Findings

The prevalence of pulmonary embolism was 20·6% in both groups (189 cases in DD-US-CT group and 186 in DD-CT group). We analysed 855 patients in the DD-US-CT group and 838 in the DD-CT group per protocol. The 3-month thromboembolic risk was 0·3% (95% CI 0·1–1·1) in the DD-US-CT group and 0·3% (0·1–1·2) in the DD-CT group (difference 0·0% [−0·9 to 0·8]). In the DD-US-CT group, ultrasonography showed a deep-venous thrombosis in 53 (9% [7–12]) of 574 patients, and thus MSCT was not undertaken.

Interpretation

The strategy combining D-dimer and MSCT is as safe as the strategy using D-dimer followed by venous compression ultrasonography of the leg and MSCT for exclusion of pulmonary embolism. An ultrasound could be of use in patients with a contraindication to CT.

Funding

Swiss National Research Foundation, Projets Hospitaliers de Recherche Clinique (France), Pneumologie Développement (France).

Introduction

The contemporary diagnostic approach of pulmonary embolism is based on the combination of clinical probability assessment of disease with sequential diagnostic tests such as plasma D-dimer measurement, venous compression ultrasonography of the leg, and helical CT.1, 2, 3 CT of the chest has emerged as a new way to directly visualise the clot in pulmonary arteries.4 First-generation single-slice spiral CT had a low sensitivity (about 70%) for pulmonary embolism,5, 6 restricting its use as a stand-alone test. Emergence of multislice CT (MSCT) has renewed hope that it could replace pulmonary angiography because of better visualisation of the segmental and subsegmental vessels and thinner collimation. Although the overall sensitivity of MSCT was only 83% in the large Prospective Investigation on Pulmonary Embolism Diagnosis II (PIOPED II) study,3 the negative predictive value of MSCT was 95% in patients with a low clinical probability of pulmonary embolism and 89% in those with an intermediate clinical probability.

To increase the diagnostic yield, the PIOPED II study also investigated the added value of undertaking CT venography of the legs during the same procedure. Although the sensitivity of the combined examination was higher (90%) than it was with chest CT alone, the negative predictive value was only marginally increased (97% vs 95%).3 This finding is compounded by data from two large studies assessing MSCT.1, 2 In the first, which included 756 consecutive patients who were referred to the emergency department for clinically suspected pulmonary embolism,1 the proportion of patients in whom a proximal deep-venous thrombosis was detected by venous compression ultrasonography of the leg despite a negative MSCT was only three of 324 (0·9% [95% CI 0·3–2·7]). In the second study,2 the 3-month thromboembolic risk was low (1·3% [0·7–2·0]) in patients who were left untreated because of a negative chest CT, despite the fact that venous ultrasonography was not undertaken.

Collectively, these results suggest that MSCT might be safe as a stand-alone test and that the added value of venous ultrasonography is questionable. To assess this notion, we compared two strategies: clinical probability assessment and either ELISA D-dimer measurement and MSCT (DD-CT strategy) or ELISA D-dimer measurement, venous compression ultrasonography of the leg, and MSCT (DD-US-CT strategy).

Section snippets

Study setting

The study was designed as a multicentre, randomised, prospective, non-inferiority trial. Data were collected from Jan 10, 2005 to Aug 30, 2006, at six participating medical centres that serve as general and teaching hospitals (Switzerland: Centre Hospitalier Universitaire Vaudois, Lausanne; Geneva University Hospital, Geneva. France: Hôpital Européen Georges-Pompidou, Paris; CHU Angers, Angers; CHU de la Cavale Blanche, Brest. Belgium: Saint Luc University Hospital, Brussels). All patients

Results

Figure 2, Figure 3 show the trial profile in the per-protocol and intention-to-diagnose populations, respectively. During the study period, 1819 patients with clinically suspected pulmonary embolism were randomly assigned: 916 in the DD-US-CT group and 903 in the DD-CT group. Six patients withdrew consent and one died before any test could be undertaken, leaving 1812 patients in the intention-to-diagnose analysis. Table 2 shows the baseline characteristics of the two groups. In the

Discussion

This randomised, multicentre, non-inferiority trial has shown that a strategy combining ELISA D-dimer measurement and MSCT was non-inferior to a similar strategy using D-dimer followed by venous compression ultrasonography of the leg and MSCT for exclusion of pulmonary embolism. The 3-month thromboembolic risk in both groups of the study was similar to that recorded in patients who were left untreated on the basis of a negative pulmonary angiography.19 The study included a large number of

References (27)

  • M Rémy-Jardin et al.

    Central pulmonary tromboembolism: diagnosis with spiral volumetric CT with the single-breath-hols technique. Comparison with pulmonary angiography

    Radiology

    (1992)
  • MD Mullins et al.

    The role of spiral volumetric computed tomography in the diagnosis of pulmonary embolism

    Arch Intern Med

    (2000)
  • SW Rathbun et al.

    Sensitivity and specificity of helical computed tomography in the diagnosis of pulmonary embolism: a systematic review

    Ann Intern Med

    (2000)
  • Cited by (349)

    View all citing articles on Scopus

    These authors contributed equally

    View full text