Surrogate markers for adverse outcomes in normotensive patients with pulmonary embolism

Abstract

Background: Although echocardiography has proven utility in risk stratifying normotensive patients with pulmonary embolism, echocardiography is not always available.

Objective: Test if a novel panel consisting of pulse oximetry, 12-lead electrocardiography, and serum troponin T would have prognostic equivalence to echocardiography and to examine the prognostic performance of age, previous cardiopulmonary disease, D-dimer, brain natriuretic peptide, and percentage of pulmonary vascular occlusion on chest computed tomography.

Design: Prospective cohort study.

Patients and setting: Normotensive (systolic blood pressure of >100 mm Hg) emergency department and hospital inpatients with diagnosed pulmonary embolism who underwent cardiologist-interpreted echocardiography and other measurements within 15 hrs of anticoagulation.

Measurements and main results: End points were in-hospital circulatory shock or intubation, or death, recurrent pulmonary embolism, or severe cardiopulmonary disability (defined as echocardiographic evidence of severe right ventricular dysfunction with New York Heart Association class III dyspnea or 6-min walk test of <330 m) at 6-month follow-up. The two-one-sided test tested the hypothesis of equivalence with one-tailed alpha = 0.05 and Delta = 5%. Of 200 patients enrolled, data were complete for 181 (88%); 51 of 181 patients (28%) had an adverse outcome, including in-hospital complication (n = 18), death (n = 11), recurrent pulmonary embolism (n = 2), or cardiopulmonary disability (n = 20). Right ventricular dysfunction on initial echocardiogram was 61% sensitive (95% confidence interval, 46-74%) and 57% specific (48-66%). The panel was 71% sensitive (56-83%) and 62% specific (53-71%). The two-one-sided procedure demonstrated superiority of the panel to echocardiography for both sensitivity and noninferiority for specificity. No other biomarker demonstrated equivalence, noninferiority, or superiority for sensitivity and specificity.

Conclusion: Normotensive patients with pulmonary embolism have a high rate of severe adverse outcomes during 6-month follow-up. A panel of three widely available tests can be used to risk stratify patients with pulmonary embolism when formal echocardiography is not available.