Continuing risk of thromboemboli among patients with normal pulmonary angiograms.

PURPOSE The purpose of this investigation was to determine the frequency of future pulmonary embolism (PE) among patients in whom PE was suspected and excluded by pulmonary angiography. METHODS Data are from the Prospective Investigation of Pulmonary Embolism Diagnosis (PIOPED). The present investigation includes 380 patients in whom PE was excluded by pulmonary angiography and who received no anticoagulant therapy while in the hospital or after discharge. Patients were followed 1 year after their normal pulmonary angiograms. RESULTS Pulmonary embolism occurred within 1 year in 6 of 380 (1.6%)(95% confidence interval [CI] 0.3 to 2.9%) patients with suspected PE and normal pulmonary angiograms. Four of six (67%) PE were nonfatal. Three of six (50%) PE occurred in the first 8 days and four of six (67%) PE occurred within the first month. Among patients who subsequently showed PE, a history of thrombophlebitis or an objective test suggestive of deep venous thrombosis was present in five of six (83%). Symptoms, signs, and radiographic abnormalities were similar among patients who showed PE on follow-up and those who did not, although the pulmonary artery mean pressure was higher among those with subsequent PE (34 +/- 25 mm Hg vs 22 +/- 10 mm Hg) (p < 0.01). CONCLUSION The frequency of PE on follow-up among patients with normal pulmonary angiograms is small. There is, however, a real and measureable rate of clinically important PE over the subsequent 12 months that is higher than reported in the general population of hospitalized patients. Patients with suspected PE, therefore, even if their angiogram is normal, might benefit from definitive studies of the lower extremities to exclude deep venous thrombosis. The cost benefit ratio of this, however, has not been evaluated.