We aimed to determine the prevalence of acute pulmonary embolism (APE) in our population and to establish the diagnostic reliability of pulmonary scintigraphy (PS), as well as the influence of the clinical context (CC). During a 2-year prospective study, patients were diagnosed by either pulmonary angiography or normal-perfusion PS. A diagnosis of APE was made in 73 (75%) of a total of 97 patients. In the group of 58 patients diagnosed by ventilation/perfusion (V/Q) PS, 33 were classified as "highly probably" having APE; 32 of these in fact had the syndrome (sensitivity 88%, specificity 94%). Combining the "highly probable" patients with the "highly suspected" patients diagnosed by ventilation/perfusion PS, we obtained a sensitivity of 51% and specificity of 100%. Ventilation scintigrams were unobtainable in 28 older patients with greater dyspnea and tachypnea, who were unable to perform the maneuvers. Of the 21 "highly probable" patients as assessed by PS, 19 were diagnosed as having APE (sensitivity 86%, specificity 85%). Combining the "highly probable" patients with the "highly suspected" group, we obtained a sensitivity of 32% and specificity of 100%. Normal perfusion PS excludes clinically relevant APE. "Intermediate" or "slight" probability PS results, even when combined with CC, have no diagnostic validity.