Pulmonary embolism. The implications of prospective investigation of pulmonary embolism diagnosis

Abstract

PIOPED represents a milestone in the study of pulmonary embolism diagnosis because of its well-designed protocol, proper execution, and the large number of patients enrolled. The most important conclusions of the study are 1. Interobserver agreement is good for classifying ventilation-perfusion scans either as normal or as high probability for pulmonary embolism, but interobserver agreement is lower for classifying scans as intermediate or low probability. 2. About 40% of patients with pulmonary embolism have high probability ventilation-perfusion scans, 40% have intermediate probability scans, and 20% have low probability scans. Few (less than 1%) patients with normal perfusion scans have pulmonary embolism. 3. Eighty-seven percent of patients with high probability scans have pulmonary embolism, and 30% of patients with intermediate probability scans have embolism. Unfortunately, 14% of patients with low probability scans have pulmonary embolism. 4. Clinical suspicion can be combined with the ventilation-perfusion scan results to improve the accuracy of diagnosis of pulmonary embolism. About 90% of patients with high probability scans and high or intermediate clinical suspicion for pulmonary embolism indeed have embolism. At the other extreme, only 4% of patients with both low probability scans and low clinical suspicion have embolism. In the remaining combinations of categories 6% to 66% of patients have embolism. 5. Suggested modifications of the original PIOPED criteria for classifying ventilation-perfusion scans make the analysis simpler and more useful. New studies have examined subgroups from PIOPED to refine guidelines for clinical practice further and to incorporate the results of tests for deep venous thrombosis into the diagnostic evaluation.(ABSTRACT TRUNCATED AT 250 WORDS)