[D-Dimer determination combined with clinical probability for the diagnosis of leg venous thrombosis]. / Association en pratique clinique du dosage des D-Dimères et d'une probabilité clinique pour le diagnostic de thrombose veineuse des membres inférieurs.

OBJECTIVE: To evaluate the results of combination of D-Dimer test and simple clinical model for the diagnosis of deep vein thrombosis (DVT). MATERIALS AND METHODS: Inclusion: clinical suspicion of DVT. Non inclusion criteria were Clinical model performed by the referring physician included probability varying from high to low. D-Dimer test was performed by five different rapid techniques. Standard of reference was Doppler ultrasonography (DU) performed by a senior radiologist. RESULTS: Eight hundred and fifty-four DU were performed on a 14 months time period, including 206 suspicion of pulmonary embolism, 109 postoperative time period, 120 non-included or excluded patients, 278 incomplete observations, 141 complete observations. DVT was present in 33 cases and absent in the other 108 cases (prevalence 23%). Sensitivity and negative predictive value of the five tests were between 82 and 97% and 90 et 97%. The most sensitive test had a specificity of 36% and a positive predictive value of 32%. Combination of clinical model and D-Dimer test did not improve the diagnostic accuracy. CONCLUSION: None of the test evaluated in the present study, even when combined with the clinical model results, did allow the exclusion of DVT.