Prospective Investigation of Pulmonary Embolism Diagnosis (PIOPED)

BACKGROUND: In 1983, reliable data on the incidence of pulmonary embolism in the adult population and in groups identified at risk were not available because the sensitivity and specificity of the diagnostic procedures had not been determined. Estimates suggested there were about half a million episodes of pulmonary embolism in hospitalized patients each year in the United States. Deaths attributable to pulmonary embolism would be expected in about one third of these patients if left untreated. The clinical diagnosis of pulmonary embolism was subject to a high frequency of false positives and false negatives. The most definitive diagnostic procedure was pulmonary angiography, an invasive, expensive procedure which was not without risk. It required specialized equipment and highly trained personnel both for performance and for interpretation. Therefore, it was used to diagnose pulmonary embolism only in the major medical centers. Another technique utilized as a method to diagnose pulmonary embolism involved a combination of perfusion and ventilation scanning; this method was only minimally invasive. A normal perfusion scan was thought to be of considerable value because it essentially excluded the diagnosis of pulmonary embolism. In selected patient populations, abnormal perfusion scans combined with normal ventilation scans were of substantial help in diagnosis. Although there had been no acceptable validation of the use of perfusion scans in the diagnosis of pulmonary embolism, thousands of patients had been evaluated for pulmonary embolism based on perfusion scanning often using methods of imaging now considered to be inadequate. In the early 1980s, clinical practice interpreted a negative perfusion scan as overwhelming evidence against the presence of pulmonary emboli. This interpretation had not been adequately tested either in a prospective study or by long-term follow-up of patients to determine clinical outcome. With regards to positive perfusion scans, there were data to suggest that as many as two-thirds of positive perfusion scans could not subsequently be confirmed by pulmonary angiography. Prospective studies in which timely angiograms using selective injections and improved imaging techniques were needed to evaluate the usefulness of positive perfusion scans. Phase I was initiated in September 1983. Protocols developed during Phase I underwent independent assessment review in April 1984 and were reviewed and approved by the May 1984 National Heart, Lung, and Blood Advisory Council. Recruitment and intervention started in January 1985 and ended in September 1986. Follow-up was completed on September 30, 1987. DESIGN NARRATIVE: Patients suspected of pulmonary embolism underwent a ventilation-perfusion scan. Patients with an abnormal perfusion scan underwent angiography. All patients were followed for one year.