[Validation of a diagnostic algorithm in non severe pulmonary embolism at the Annecy general hospital. D-dimers, venous lower limb ultra-sound and spiral CT scan]

Abstract

Objective and method: We have evaluated the sensitivity of a diagnostic algorithm for all patients suspected of pulmonary embolism using: D-Dimer, lower limb venous ultrasonography and helical computed tomography. To validate this approach, a lung scan is systematically carried out if the pulmonary embolism diagnosis is not withheld as a result of the algorithm. Clinical tests are organised between the 3rd and 6th month.

Results: Two hundred patients were involved between January 1998 and October 1999. One hundred and six pulmonary embolisms were diagnosed. Out of the 200 ultrasonography tests carried out we found: 71 proximal deep-vein thrombosis (popliteal or supra-popliteal), 33 distal thrombosis (infra-popliteal). Ninety-two cases were negative (4 tests non conclusive). We have deduced that a deep-vein thrombosis permits the diagnosis of thrombo-embolic illness without any further diagnostic approach (no computed tomography). Out of the 129 computed tomographies carried out we found: 35 pulmonary embolisms and 23 other diagnoses. Seventy-one lung scans were therefore carried out. We recorded 7 discordances (scans showed high and very high probability for pulmonary embolism whilst computed tomographies did not): pulmonary angiography was negative 4 times and diagnosed pulmonary embolism once and two patients refused to take the test (Table 2). There are two recurrences in the follow-up: proximal thrombosis and a pulmonary embolism. This involved two patients who had refused to undergo pulmonary angiography.

Conclusion: This diagnostic approach therefore seems satisfactory but would require further investigation on a wider scale.