There is a bit of a buzz on social media at the moment1 about an interesting paper in the NEJM2 on the prevalence of PE in patients with syncope3. We’re all familiar with the controversies about looking too hard for PE in patients and the difficulties in deciding whether a PE is clinically important. It’s really tricky as PE is not without serious complications (aka death), but the treatment is similarly dangerous as anticoagulation has many complications too (aka death). So we really don’t want to treat too many people, and we don’t want to miss people.
Clots are tricky………..
So this week we have a very interesting paper in the NEJM2 which is open access at the moment. The abstract is not on PubMed as yet, but please go read the full paper (or at least the methods and the results sections). Click on the link.
The authors looked at patients admitted to hospital with a first diagnosis of syncope. Syncope patients are high risk patients in the ED as there are a myriad of causes for a transient loss of consciousness4,5. PE is one of the causes of syncope, but it may not be obvious as patients do not always present with chest pain or shortness of breath. This is a concern because a PE large enough to interrupt blood flow enough to cause syncope is probably more than just lung fluff6. The authors argue that we really don’t know what the prevalance of PE is because we’ve never looked for it before.
Tell me about the patients
This is key. These are medical ward patients in 11 Italian hospitals. The patients in this study are patients admitted to hospital with a first episode of syncope. That’s a variable group depending on which health economy you work in. Even within the UK there is a lot of variation about which patients get admitted after syncope. In Virchester the vast majority of patients without abnormal signs, basic investigations, normal ECGs and a normal examination will be sent home for outpatient investigation. That’s not the case elsewhere. In this study it’s difficult to be sure whether the patients are the same as mine here. They did stratify for cardiac risk and looked at patients with comorbidities, but the group of patients with unexplained syncope were also admitted, that’s not the case in our practice. This suggests that their population was not as sick as our admitted patients, but it’s difficult to be sure. They did send home nearly 3/4 of the patients who were seen in the ED which feels about the same as us here in Virchester, but the key is that these are not all syncope patients that we see. It’s a subgroup of the ones that they (you, me) would be worried about. They did exclude trauma patients, patients with multiple episodes, known PE patients and pregnancy (fair enough).
How hard did they look for PE?
The investigation strategy is well described and follows a process very similar to that which we use in Virchester. Low risk patients got a d-dimer and PE was excluded if this was normal. High risk patients got a CTPA (or V/Q if CTPA could not be performed). It’s not a perfect strategy as there are false positives and negatives, but it is a well trodden and widely accepted one that reflects clinical practice. They looked for PE within 48 hours of admission which is important as hospitalisation itself is a risk for the development of PE. They also gave thromboprophylaxis to patients as per normal guidelines.
What are the main findings?
- After exclusions and patients sent home direct, they were left with 560 patients from 2584 screened (this is a subset remember). 330 of the 560 patients (58.9%) had PE excluded on low risk + d-dimer testing. Of the remaining 230 patients, 229 were investigated for PE of whom 42% had a PE. The numbers are bit tricky to follow so I think it goes something like this….
- 2584 patients screened: 560 eligible for the trial
- 560 investigated: 330 ruled out as low risk and normal d-dimer
- 229/230 patients went on to further testing: 92 had a PE (42%)
- So overall the incidence of PE in the entire study cohort was 96/560 which is 17%
That’s a pretty high incidence of PE in a cohort who were not expected to have a PE. Interestingly this was not just lung fluff either. 42% of the PEs were in a major pulmonary artery.
Is this a good study?
Methodologically I would say yes. The authors have described their populations well, have applied a sensible strategy for PE detection and have found results that are really thought provoking. The question is really about what does it mean for us in the ED. Like many PE studies the finding of clots is not necessarilly the same as determining their significance nor the benefit to treatment.
So what’s the clinical significance?
That’s tricky. We know that PE happens a lot. Whether it matters is a much more complex question. This study diagnoses PE using a standard algorithm but does not tell us whether identifying the PEs makes a difference to prognosis. Do these patients die if they are missed? Do they get further PEs? Do they develop cardiac failure or pulmonary hypertension? Would any of these be better or worse if we treat the patients?
PE is a really tricky diagnosis as the consequences of therapy are not inconsiderable7,8. We must therefore be cautious about treatment unless we know what the benefits and risks are.
Is this association or cause?
As we discussed on the syncope podcast we do need to think about the mechanism of syncope to PE. Large PEs can clearly cause mechanical obstruction, thus leading to a transient fall in cardiac output and then syncope. However, how could a very small subsegmental PE cause this to happen? Mechanically you’d expect something that small to pass through the heart with few problems so what happened? Was it really the cause of the syncope, or just associated with it, or resulting from it? Might a smaller PE result in a dysrhythmia as a cause of PE? Was it a big PE at the time but it broke up as it went through the heart leading to only subsegmental PEs on CTPA? Iain has challenged me about this in the past and I just don’t have the answer, do you? If so we’d love to know the answer – how big does a PE need to be to cause mechanical blockage or dysrhythmia?
UPDATE – Worth reading proposed mechanisms on PulmCCM site for an opinion on mechanisms 9 .
But…. I still don't understand. How does PE cause a transient reduction in cerebral blood flow causing global cerebral hypoperfusion????? https://t.co/WO6qn06Y3j
— Iain Beardsell (@docib) October 20, 2016
Is this relevant to emergency medicine?
Although these are admitted patients, these are patients who transit through the ED and in many hospitals we will be clearly involved in determining the diagnostic pathway and identifying the likely causes. This study suggests that PE should remain, and probably move up our list of differentials for our admitted patients. The difficulties remain on deciding what to do if you find a PE in your patient. I think there will be significant pressure to treat if you find a PE in a patient who presents with syncope, but I’d love to see the trial that follows these patients up, or even better determines best treatment practice. Although this trial is really interesting and may well change practice it is still unclear which patients would benefit from anticoagulation. As Rory Spiegal suggests, we may be chasing ghosts.1
The bottom line.
This study found clots in patients hospitalised following syncope. We still don’t know the clinical significance of this, whether to treat them or whether they were the cause of the syncope.
Clots are tricky……
Before you go please don’t forget to…
- Subscribe to the blog (look top right for the link)
- Subscribe to our PODCAST on iTunes
- Follow us on twitter @stemlyns
- See our best pics and photos on Instagram
- PLEASE Like us on Facebook
- Find out more about the St.Emlyn’s team