There Are (Almost) No PEs in Syncope, Actually

Last year, we suffered the ignominy of being presented with evidence implying the Emergency Department was systematically failing to adequately evaluate the underlying etiology of patients with syncope. The PESIT trial demonstrated nearly 1 in 6 patients admitted to the hospital had PE, and, worse, most had obvious clinical manifestations of VTE. This is, despite its publication in the New England Journal of Medicine, still nonsense, and flies in the face of every other reasonable estimate of the prevalence of PE.

This study is yet another reasonable refutation of their inflated estimate: a retrospective, secondary re-analysis of a prospectively-collected syncope data set. This analysis reviewed 348 patients previously enrolled in the Emergency Department with a presenting complaint of syncope, about half of whom were observed or admitted to the hospital. Overall, just two of the original 348 were diagnosed with PE in the ED. None of the patients admitted or observed were diagnosed with PE during their hospitalization, but, in their 30-day follow-up period, three total additional PE diagnoses were made.

Without a systematic process for excluding PE, it is reasonable to suggest these numbers are biased towards under-estimating the diagnosis of PE – although the patients in question with 30-day PE each underwent objective testing during their initial presentation with either D-dimer or CTPA. Regardless, the rate of PE in patients hospitalized with syncope is far below the 1 in 6 prominently reported – and we might do well to expunge it from our collective memory.

“Prevalence of pulmonary embolism in patients presenting to the emergency department with syncope”
https://www.ncbi.nlm.nih.gov/pubmed/28811209

6 thoughts on “There Are (Almost) No PEs in Syncope, Actually”

  1. I was disappointed to see this published, as it adds nothing to our understanding while superficially appearing to refute PESIT. I agree that the population studied in PESIT was different than ours, but nevertheless the point was that if you systematically screen for PEs, you find more than you expect. That can only be refuted by a study in which you systematically screen for PEs! That was not done here, so the underreporting bias could easily be very large indeed.

    1. Disappointed might be a little bit strong of a word, but you make a fair point regarding the reliability of a prevalence study where less than a quarter underwent objective testing, and almost a third were lost to telephone follow-up. However, it’s similarly reasonable to suggest those patients in which no work-up was initiated had even lower pre-test likelihood for PE, but stranger things have happened. The level of evidence here is low, but it still functions as a reminder the context of PESIT was different and should have limited impact on ED practice.

  2. It does on the other hand reinforce the idea that syncope in general is a low risk complaint and we are admitting far too many of these people. I think at the very least, PESIT should be repeated in a US ED population. For now I just think about it a little more, just as I do in *admitted* COPD patients.

  3. I don’t understand what the hubbub was about PESIT. It says among high risk patients with syncope requiring admission who don’t have another identifiable cause of syncope that 1/6 had a PE. No one argues that the rate who present to the ED was high. If anything, it suggests that the ED did a good job identifying high risk patients who require further work up as the paper never in any way disputed that the prevalence of PE in all syncope presenting was high or even higher than the ~2-3% usually reported. What PESIT says is if you take out a high risk population (1 screening by the ED) add a negative workup (2nd screening) that a 3rd screening of systematic PE workup THEN AND ONLY THEN becomes high yield. This doesn’t seem controversial to me and seems interesting. All the misinterpretations and misrepresentation of the study are what drove me nuts.

    1. The hubbub was that everyone just took away the “1 in 6” number without placing it into appropriate context – which is, as you say, very limited. It’s probably even more limited than you suggest, because the details of their cohort indicate many of these patients had active signs and symptoms of VTE or PE, and were probably inadequately worked up by their EDs. “1 in 6”, to a hospitalist, sounds like everyone needs objective testing regardless of the ED.

      It reminds me of the PE in COPD discussions, based on inpatient studies showing high prevalence of PE among COPD patients – in which, yes, those patients who never improved needed evaluation for an alternative diagnosis, but the base rate of PE in hospitalized COPD hardly mandates systematic screening.

Comments are closed.