PARIS--When pulmonary embolism (PE) is suspected, it may be possible to avoid pulmonary angiography--without sacrificing accuracy--by using a noninvasive approach that includes spiral computed tomography (CT). This approach, according to the findings of a recent French study, can confirm or exclude PE 99% of the time.[1]

It also appears to be highly safe, added the study authors, who are from the hospital Hôtel-Dieu in Paris.

"Only one complication occurred with spiral CT," they reported. Furthermore, thromboembolic events and death were uncommon during follow-up.

The study included 228 patients with suspected PE. Unless a contraindication to a specific technique was present, patients underwent spiral CT, ventilation/perfusion lung scanning, and D-dimer measurement within 24 hours of hospital admission. Those with normal spiral CT findings also received lower-limb ultrasonography if the lung scan results were inconclusive and/or the D-dimer level was above 500 ng/mL.

A diagnosis of PE was made if spiral CT detected a thrombus or if ultrasonography revealed a thrombus when the spiral CT scan was normal. The authors used clearly defined criteria to determine when a diagnosis of PE could be excluded. For example, when spiral CT showed no thrombus, PE was ruled out if the lung scan was normal, if the D-dimer level was below 500 ng/mL, or if a previous lung scan showed similar nondiagnostic findings.

Of the 228 patients, 201 (88%) underwent spiral CT, 178 (78%) received a lung scan, and 227 (99.5%) were given a D-dimer test. Lower-limb ultrasonography was required in 56 cases (24%). Ultimately, 96 patients (42%) were given a diagnosis of PE.

Spiral CT was the most accurate diagnostic technique, confirming PE in 70 patients (73%). In only four cases (4%) was the diagnosis of PE made solely on the basis of a high-probability lung scan. Ultrasonography was required to confirm the diagnosis of PE in 22 patients (23%). In three cases (3%), patients had high-probability lung scans despite normal spiral CT results; in each of these cases, ultrasonography confirmed the presence of PE and thus the need for pulmonary angiography was avoided.

The noninvasive strategy was equally good at ruling out thromboembolic disease in the other 132 patients. In 34 cases (26%), the spiral CT scan was normal, the lung scan was nondiagnostic, the D-dimer level was above 500 ng/mL, and venous ultrasound was normal. PE was excluded by a normal lung scan result in 18 patients (14%), by a D-dimer level below 500 ng/mL in 41 patients (31%), and by a previous nondiagnostic lung scan in 14 cases (11%). In 23 patients (18%), an obvious alternative diagnosis was seen on the spiral CT scan. Two patients (1%) had high-probability lung scans despite normal spiral CT results; both patients underwent pulmonary angiography, which excluded the diagnosis of PE. The only adverse event associated with spiral CT was one case of renal failure that required hemodialysis for three weeks. This patient had been known to be at risk for renal failure.

At three months, 92% of the patients were available for follow-up. During that time, three of the patients with PE and two of those without PE suffered a subsequent thromboembolic event. Mortality in the two groups was 8% and 5%, respectively. In only two cases, both of which occurred in the patients with PE, did the deaths result from thromboembolic disease.

"Only 1% of our patients underwent angiography because the noninvasive strategy was inconclusive," the authors pointed out. Thus, they concluded that their noninvasive approach diagnoses or excludes PE in 99% of suspected cases. Other prospective studies would be needed, they said, to determine if spiral CT should replace lung scanning as the first step in PE diagnosis.

--Timothy Begany

Reference
1. Lorut C, Ghossains M, Horellou M-H, et al. A noninvasive diagnostic strategy including spiral computed tomography in patients with suspected pulmonary embolism. Am J Respir Crit Care Med. 2000;162:1413-1418.

Return to table of contents