Discussing:

Prandoni et al. New England Journal of Medicine. Prevalence of pulmonary embolism among patients hospitalized for syncope. NEJM 2016; 375(16): 1524 – 31.

On your next shift, you see a patient with first time syncope who you feel should be hospitalized. The admitting hospitalist reads the above article, and requests that you obtain a CTPE prior to admission because, “one in six patients with syncope have a PE.” What do you do?

Main (counter) Points:

Author’s conclusion, “pulmonary embolism was identified in nearly one in every six patients hospitalized for a first episode of syncope.”

• Significant selection bias! 72% of undifferentiated patients were discharged and not included in analysis.
• High proportion of included patients already had tachycardia, tachypnea, DVT signs/symptoms.
• Of the entire population, 4% of undifferentiated syncopal patients had a PE.

Bottom line: If you suspect syncope secondary to PE, then work the patient up as you already would have based on history, risk factors, vital signs, and physical examination. This trial should not be utilized as evidence to support performing imaging on all undifferentiated syncope patients that present to the ED.

Background:

Syncope is defined as a transient loss of consciousness with spontaneous resolution believed to be due to cerebral hypoperfusion. Despite representing 3-5% of all ED visits, both ED and inpatient workups often are unrevealing for an underlying etiology. The authors posit that the current guidelines of the American Heart Association and the European Society of Cardiology pay little attention to PE as a cause of syncope.

Details:

This was a cross-sectional study of 11 Italian hospitals aimed at determining the prevalence of PE in >18 year old hospitalized patients with first-time syncope. Exclusion criteria were prior episodes of syncope, current anticoagulation, or pregnancy. Diagnosis of PE was ruled out if the patient had low pretest probability according to the Simplified Wells score <4 and had a negative D-dimer. If either were positive, a computed tomography pulmonary angiography or ventilation-perfusion scan was performed.

2584 patients visited the EDs between March 2012 and October 2014. 1867 were discharged from the ED, and another 157 were excluded. 560 patients were included in study, with 330 (58.9%) patients ruling out based on negative Wells score and D-dimer. 97/230 (42.2%) of the remaining population had a PE identified on computed tomography pulmonary angiography or ventilation-perfusion scanning. Of note, in the entire cohort, the prevalence of PE in hospitalized patients with syncope was reported as 17.3% (97/560).

The prevalence of tachycardia (33% vs 16.2%), tachypnea (45.4% vs 7.1%), and signs/symptoms of DVT (40.2% vs 4.5%) were all statistically significant when performing a subgroup analysis and comparing patients with confirmed PE vs without PE. Aside from the large selection bias introduced, this data also suggests that these high-risk patients may have benefited from a more in-depth workup in the ED. Furthermore, the clinical significance of the discovered PEs is also questionable since some were likely incidental findings. 50% (12/24) of positive VQ scans showed diffusion deficit in only 1-25% of both lungs. Likewise, CTPA results showed segmental (19/72) and subsegmental (5/72) PEs that are known to be of questionable clinical relevance.

Level of Evidence:

ACEP Clinical Policy Committee Level III

Reviewers:

Resident Reviewer: Dr. Anatoly Kazakin
Faculty Reviewer: Dr. Otis Warren

Relevant Articles:

Huff JS et al. Clinical policy: critical issues in the evaluation an management of adult patients presenting to the emergency department with syncope. Ann Emerg Med 2007; 49(4): 431 – 44. 

Quinn J, Mcdermott D, Stiell I, Kohn M, Wells G. Prospective validation of the San Franciscso Syncope Rule to predict patients with serious outcomes. Ann Emerg Med. 2006;47(5):448-54.

Source Article:

Prandoni P et al. Prevalence of pulmonary embolism among patients hospitalized for syncope. NEJM 2016; 375(16): 1524 – 31.