Acute pulmonary embolism: external validation of an integrated risk stratification model

Abstract

Background: In hemodynamically stable patients with acute pulmonary embolism, risk stratification is essential to drive clinical management. In these patients, risk stratification for in-hospital adverse outcomes based on markers of right ventricular dysfunction and injury has been proposed.

Methods: The aim of this study was to validate a model based on the incremental prognostic value of right ventricular dysfunction and injury in hemodynamically stable patients with acute pulmonary embolism. Patients from the prospective Italian Pulmonary Embolism Registry were included in the study. Study outcomes were in-hospital death and the composite of in-hospital death or clinical deterioration.

Results: Among 1,515 hemodynamically stable patients, 869 had both echocardiography and troponin assessments. The risk for in-hospital death or clinical deterioration was higher in patients with right ventricular dysfunction and elevated troponin level (8.8%; hazard ratio [HR], 14.2 [95% CI, 1.94-104.16]; P < .01) and with either right ventricular dysfunction or elevated troponin level (4.7%; HR, 7.9 [95% CI, 1.1-59.9]; P < .05) compared with patients without dysfunction and normal troponin levels. The negative predictive value of the model was 100% for in-hospital death and 99% for death or clinical deterioration. C statistics showed an improvement of the discriminatory power for in-hospital death or clinical deterioration by using the overall model (0.66; 95% CI, 0.60-0.73) over either echocardiography (0.59; 95% CI, 0.53-0.67) or troponin level (0.61; 95% CI, 0.53-0.69) alone.

Conclusions: A model that includes both dysfunction and injury of the right ventricle has an incremental prognostic value for risk stratification in hemodynamically stable patients with acute pulmonary embolism. Patients with no dysfunction or injury have a favorable outcome.

Trial registry: ClinicalTrials.gov; No.: NCT01604538; URL: www.clinicaltrials.gov.