Background: Acute pulmonary embolism (PE) is a life-threatening condition. Making a definitive diagnosis with radiologic studies may delay therapy or be unsafe for the patient. Echocardiography is readily available and can suggest PE by demonstrating right ventricular (RV) dysfunction. McConnell's sign on echocardiogram (ECHO-CG) (RV dysfunction with characteristic sparing of the apex) has been reported to have high sensitivity and specificity for the diagnosis of acute PE. It is hypothesized that McConnell's sign on ECHO-CG in patients hospitalized with suspected acute PE would have a high positive predictive value (PPV).
Methods: Data, from 2005 to 2010, were retrospectively collected on all patients with an ECHO-CG interpreted as revealing McConnell's sign, who had undergone another diagnostic study (computed tomography pulmonary angiography, ventilation-perfusion scan, upper or lower extremity Doppler ultrasound, or autopsy) for venous thromboembolic disease (VTE). The PPV on transthoracic ECHO-CG was calculated for the diagnostic accuracy of McConnell's sign in all patients. To minimize the potential for ECHO-CG reader bias of patients already confirmed to have had a PE by another modality, the PPV was then recalculated only on the patients in whom the ECHO-GM was the first diagnostic study.
Results: Seventy-three patients had findings of McConnell's sign on ECHO-CG. The PPV of McConnell's sign on ECHO-CG was 57% (CI, 45%-67%). Of the 37 patients who underwent an ECHO-CG in the first study for suspected acute PE, 15 patients had VTE confirmed; the PPV in this subset was only 40% (CI, 24%-56%). There were 20 patient deaths overall; of these, only 9 of the patients were confirmed to have VTE.
Conclusion: We concluded that the presence of McConnell's sign has a relatively poor PPV for the diagnosis of acute PE and should not be used in isolation when making a diagnosis of PE in patients.