Pulmonary embolism (PE) and deep vein thrombosis are two facets of the same disease, that is, venous thromboembolism (VTE). In patients with angiographically proven PE, the prevalence of proximal deep vein thrombosis by venography is around 70%. The sensitivity of compression ultrasonography (US) for the diagnosis of acute VTE in patients with a suspicion of PE is between 40 and 60%, with a high specificity (96 to 100%). Taking into account the 20 to 30% prevalence of PE in a population consulting for suspicion of this disease, the first line use of compression US will allow the diagnosis of acute VTE in half of patients with confirmed PE, that is, in 10 to 15% of patients addressed for suspicion of PE. In outpatients, the first line use of D-dimers which will exclude acute VTE in one-third of the initial population will slightly increase the reliability of compression US as a first imaging test. New tools of looking for deep vein thrombosis, such as computed tomographic venography coupled with computed tomographic pulmonary angiography, could become an interesting approach in the diagnostic strategy of PE, but require adequate evaluation in prospective studies.