Background: To test the hypothesis that troponin I and echocardiography have an incremental prognostic value in patients with pulmonary embolism (PE).
Methods and results: In 91 patients with acute PE, echocardiography was performed within 4h of admission. Troponin I levels were obtained on admission and 12h thereafter. The 0.06 microg/l troponin I cut-off level was identified as the most useful, high-sensitivity cut-off level for the prediction of adverse outcome by receiver operating characteristic analysis with a sensitivity and specificity of 86%, respectively. Twenty-eight (31%) patients had elevated troponin I levels (4.9+/-3.8 microg/l). Twenty-one (23%) patients had adverse clinical outcomes including in-hospital death in five, cardiopulmonary resuscitation in four, mechanical ventilation in six, pressors in 14, thrombolysis in 14, catheter fragmentation in three, and surgical embolectomy in three. The area under the receiver operating characteristic curve from multivariate regression models for predicting adverse outcome without troponin I and echocardiography (0.765), with troponin I (0.890) or echocardiography alone (0.858), and the combination of both tests (0.900) was incremental. Three-month survival rate was highest in patients with both a normal troponin I level and a normal echocardiogram (98%). Positive predictive value for adverse clinical outcomes of the combination of echocardiography and troponin I was higher (75% (95%CI 55-88%)) compared with each test alone (echocardiography: 41%, 95% CI 28-56%; troponin I: 64%, 95% CI 46-79%).
Conclusions: While troponin I measurements added most of the prognostic information for identifying high-risk patients, a normal echocardiogram combined with a negative troponin I level was most useful to identify patients at lowest risk for early death.