Should intensivists routinely perform bedside echos in suspected PE?
In pulmonary embolism (PE), right ventricular (RV) strain on transthoracic echocardiography increases the likelihood of shock and mortality. One study showed among patients with PE and normal blood pressure, 10% of those with RV strain on echocardiogram developed shock, and 5% died in hospital. Those without RV strain maintained their blood pressure and survived (but important to note, so did 90% of those with RV strain).
Right ventricular dysfunction from pulmonary embolism can take a variety of forms, but the simplest and most consistent is dilation of the right ventricle. The RV is usually just over half the volume of the left ventricle on the 4-chamber apical view. RV dilation (or dilatation) at >1:1 the size of the LV (by 'eyeball' or gestalt) by experienced operators is considered right heart dysfunction.
But how experienced do those operators need to be? A new study suggests that with a little practice, critical care fellows can become nearly as good as board-certified cardiologists at identifying RV strain in suspected PE.
In a prospective observational study by Filopei et al, fellows at Mount Sinai Beth Israel in New York performed 154 limited echocardiograms on patients with suspected PE. In 110 of these patients, a cardiologist reviewed a complete echocardiogram within 48 hours.
The area under the curve for size and function were both 0.83 for the fellows. Their attending intensivists were slightly better at 0.87-0.88. The cardiologists' readings were considered the reference standard.
But those cardiologists didn't complete their echo interpretations until almost a day later, in most cases. Authors concluded,
Screening for right ventricular dysfunction using goal-directed echocardiography can and should be performed by pulmonary critical care physicians in patients with acute pulmonary embolism."
Right Heart Strain on Echo in Diagnosis of Pulmonary Embolism
In addition to prognosis, checking right ventricular function could be useful in the triage and differential diagnosis of suspected pulmonary embolism, when numerous etiologies for hypoxemia, dyspnea, or hemodynamic instability are being considered. While not diagnostic of PE, detection of right heart strain could encourage the earlier use of anticoagulants or thrombolytic therapy, while awaiting confirmatory testing.
A prospective observational study by Dresden et al found that among 146 patients with moderate to high pretest probability for PE, 30 of whom were found to have a PE, RV dilation was 98% specific for pulmonary embolism (i.e., very good at ruling in PE), although only 50% sensitive (as you would expect, since many patients with PE don't have RV strain).
When a test as easy, relatively accurate, fast, and cheap as bedside echocardiogram is widely available, there's a good argument to perform it in most patients in whom pulmonary embolism is being considered. This study suggests the learning curve is short, with good results achievable with a modicum of training and practice.
On the other hand, a CT-angiogram remains the gold standard diagnostic test for PE, and is indicated in the large majority of patients with moderate-or-greater suspicion of PE. CTA usually has a rapid turnaround time, and empiric anticoagulation (heparin infusion) can be initiated pending the results. Does early echocardiography improve that process, or just complicate it unnecessarily?
An enlarged right ventricle on bedside echocardiogram could lead to anticoagulation being provided more quickly. Awareness of right ventricular dilation could also add to a physician's diagnostic confidence, and vigilance to deterioration in the minority of patients with RV strain who will develop shock. These patients might receive expedited prescription and delivery of thrombolytics. These are theoretical benefits, unproven in comparative studies.
It's likewise possible that echo-savvy physicians could occasionally miss the diagnosis of PE through overconfidence, by failing to order CT-A in patients with great-looking echos (which do not rule out PE). None of this will get settled soon, or probably ever, since these questions don't lend themselves well to randomized trials.
[polldaddy poll=9878462]