Purpose: To provide a clinical approach to the diagnosis of pulmonary embolism.
Data identification: An English-language literature search using MEDLINE (1982 to 1990) and bibliographic reviews of textbooks and review articles.
Study selection: In addition to several reviews, studies that evaluated the diagnostic technology of pulmonary embolism were selected. Preference was given to studies with a prospective design, particularly those done within the past decade.
Data extraction: Studies were assessed independently by three unblinded observers. Data were chosen to describe the efficacy of diagnostic technology on the basis of disease prevalence, sensitivity and specificity, and predictive value.
Results of data analysis: A normal lung scan or pulmonary angiogram rules out the diagnosis of clinically important pulmonary embolism with at least 95% certainty. Lung scan interpretations indicating high or low probability have approximately a 15% error in diagnosing or ruling out pulmonary embolism. The accuracy of either scan result improves when the clinical suspicion of pulmonary embolism matches the lung scan result. Serial impedance plethysmography of the lower extremities may exclude thromboembolism with 95% certainty in patients without high-probability lung scan results or cardiopulmonary disease.
Conclusions: The combination of clinical suspicion and the results of the lung scan and impedance plethysmography appear to offer acceptable diagnostic accuracy in evaluating many patients suspected of having pulmonary embolism. The usefulness of this approach for patients with cardiopulmonary disease is still unknown.