Reviews and summaries of the latest EM papers. We try to cover 10 of the best articles each month. Please visit our website for written summaries and to search past editions:

February 21st, 2019    

37. Syncope and near-syncope

Background: Syncope is fairly easily recognised as a brief loss of consciousness. Near-syncope (or presyncope), on the other hand, is a more nebulous phenomenon, both to patients and their doctors. It can be described as an episode of dizziness, feeling hot or cold, blurred vision, nausea, weakness, or a ‘funny turn’. It is generally believed that syncope is a more serious beast than near-syncope, but the current paper puts this idea to the test.


The paper: A prospective observational cohort study across 11 EDs, involving 3,581 patients aged >60 with syncope or near-syncope. The primary outcome was all-cause mortality or serious clinical event (cardiac arrhythmia, heart block, pacemaker insertion, MI, PCI, PE, stroke or subarachnoid haemorrhage) at 30 days. There was no significant difference in this outcome between the groups (syncope 18.2%, near-syncope 18.7%). (Annals of Emergency Medicine, December 2018)


The bottom line: Those patients who present with near-syncope should be investigated with as much care as those with syncope, because they have the same risk of death or serious clinical event.


February 17th, 2019    

36. Is it time to stop checking postural BP?

Background: We often obtain lying and standing blood pressure measurements (OVS – orthostatic vital signs) in our patients, particularly if they have had a syncopal episode. A positive result is usually defined as a drop in systolic BP of >20mmHg, a drop in diastolic BP of >10mmHg, or an increase in heart rate of >30bpm (all measured 3 minutes after standing). We know that OVS are common in asymptomatic nursing home residents, and that they vary according to time of day (Ooi, 1997). We also know that OVS are not helpful in determining the severity of non-blood volume loss, like diarrhoea and vomiting (McGee, 1999). The current paper adds to this picture, considering the utility of OVS in the workup of patients presenting with syncope.


The paper: A systematic review looking to determine the utility of OVS in diagnosing orthostatic syncope, and to determine to what extent OVS can rule out serious life-threatening causes of syncope. Four prospective cohort studies were included, involving over 2,000 patients. They found an average rate of OVS of 25% in undifferentiated ED patients, regardless of their presenting complaint or final diagnosis. Of those ultimately diagnosed with orthostatic hypotension, only 69% had OVS by numbers. Of those ultimately diagnosed with another cause of syncope, 13% had OVS by numbers. (Journal of Emergency Medicine, December 2018)


The bottom line: The finding of a postural blood pressure drop is not very sensitive or specific for a final diagnosis of orthostatic hypotension


Note: If OVS are positive, this is sometimes a false positive (i.e. an incidental, asymptomatic finding, not related to the actual cause of the patient’s syncope). The risk is that you stop looking for another cause. Conversely, if OVS are negative in the ED, this is sometimes a false negative (i.e. there is no BP drop now, but there was when the patient had their syncope). The risk is that you assume it’s fine, and don’t stop some of their twenty anti-hypertensive medications.


Expert commentary:
"A really interesting topic. In light of this evidence clinicians should be mindful when prescribing IV fluids for patients who have postural hypotension: only consider doing so when the patient is symptomatic and intravascularly depleted."

(Dr Dwynwen Roberts, ED consultant)


February 1st, 2019    

35. Myths of A&E: clamping catheters to allow gradual drainage

Background: The other day I overheard one of our ED nurses telling his colleague that he should clamp the catheter to allow gradual drainage for a patient in acute urinary retention (AUR). I must admit I’d never heard of this practice but I certainly wanted to avoid the potential complications of rapid drainage that he mentioned (circulatory collapse and haematuria). Have I been endangering my patients for years by allowing free drainage? What does the literature say?


The paper: A case-control study of 62 male patients with acute urinary retention (AUR). These were randomised to either rapid or gradual drainage of the bladder (gradual meant periodically clamping the catheter to evacuate the entire volume over 2 hours). Although the mean blood pressure did fall post-catheterisation, there was no significant difference between the groups (15mmHg in the rapid group, 10mmHg in the gradual). (Urology Annals, December 2017)


Note: A larger RCT of 294 male patients in AUR also found no significant difference in post-obstruction blood pressure or haematuria, and was the first RCT to consider this question. (Urologia Internationalis, July 2013)


The bottom line: Rapid drainage of the bladder post-AUR does not result in more cases of hypotension or haematuria.


February 1st, 2019    

34. Syncope vs Seizure

Background: Many cases of seizure are obvious, and many cases of syncope are obvious, but some fall into the grey zone in between. Witnesses often think a patient has had a seizure because there were jerking movements during the episode, and it can be tricky for us to figure out what really happened after the fact. That is, it was tricky until this lovely little paper came along…


The paper: This was a comparison of the videotape recordings of 65 cases of syncope and 50 cases of seizure. The first group were undergoing tilt table testing for suspected vasovagal syncope, while the second were being evaluated for epilepsy. The authors focused primarily on motor phenomena, looking for patterns to differentiate syncope from seizure. The results are shown in the table below. (Neurology, April 2018)

     Syncope       Seizure   
Myoclonic jerks?           51% 100%
Bilateral? 78% 100%
Synchronous? 48% 90%
Median number 2 48
Duration of jerking 3.6s 29s
Tonic posture 65% 100%
Loss of tone 100% 0%


The bottom line: The following phenomena favour syncope rather than seizure as the diagnosis: no myoclonic jerks, or a few myoclonic jerks that are unilateral or asynchronous, and loss of tone. The authors coin the 10/20 rule – 10 jerks or less is syncope; 20 jerks or more is seizure.


February 1st, 2019    

33. Do patients need to fast before procedural sedation?

Background: “Hi there, Johnny. I see your arm is bent at an impossible angle and you’re in extreme pain. Unfortunately you had some crisps 30 minutes ago so we’ll have to let you sit here for another 5 hours till we can fix it.” Admittedly, this is a rare scenario these days, but what about the older lady who presents hypotensive with fast AF? Do you wait, bolusing IV fluids and nervously watching the monitor until she has been NBM (NPO for our American friends) for enough time to cardiovert? What is the evidence for fasting before procedural sedation? This is the largest prospective ED study of this topic to date.


The paper: A planned secondary analysis of a multi-centre cohort study, involving 6,183 children who were undergoing parenteral conscious sedation for painful procedures (mostly fracture reduction, but also foreign body removal, abscess I&D, laceration repair and lumbar puncture). The commonest agent used was ketamine, but propofol and midazolam were used in many. The ASA fasting guidelines for solid food (6 hours) were not met in 48% of cases (5% did not meet the 2-hour requirement for liquids). There was no difference in any of the outcome measures across these groups (aspiration, vomiting or other adverse event). There were in fact no cases of aspiration at all. The commonest adverse events were oxygen desaturation (5.5%) and vomiting (5.1%), but most vomits occurred during recovery. Of the 6 children who vomited during the sedation itself, 3 had been fasting for at least 10 hours before the procedure. (JAMA Pediatrics, July 2018)


The bottom line: This study does not support delaying sedation to meet established fasting guidelines in children


Note: This study is by no means a lone outlier. Each of the following papers had a similar conclusion, some extending the evidence to an adult population: Agrawal 2003Roback 2004Treston 2004Bell 2007McKee 2008Thorpe 2010Taylor 2011Beach 2016. In all, there are around 20,000 ED cases of procedural sedation reported in the literature, and only 2 reported clinically important aspiration events (one patient was NBM for 6hr, the other for 24hr). Aspiration is a risk that is trivially small and is not reduced by fasting. I think we can let our patients eat.


February 1st, 2019    

32. Does this patient with chest pain need a chest X-ray?

Background: We often obtain chest X-rays (CXR) as part of a workup for patients presenting with chest pain, even though several studies have shown relatively low yield for this investigation. In 2002, Rothrock et aldeveloped a clinical decision rule, using retrospective data to recommend the 10 highest-yield criteria for finding important pathology on the CXR. Their conclusion was that if none of these criteria were present, a CXR could safely be omitted. Rothrock’s criteria were:
– Age > 60
– Haemoptysis
– Prior or current alcohol abuse
– Prior TB
– Prior VTE
– Oxygen saturation < 90%
– Respiratory rate > 24
– Temperature > 38
– Rales on auscultation
– Diminished breath sounds on auscultation


The paper: A multi-centre prospective observational study, including 1,089 patients with non-traumatic chest pain, enrolled over a two-year period. Each of the Rothrock criteria were recorded, with the addition of history of heart failure (CHF) and smoking (significant variables suggested by Hess et al in 2010). Each had CXR, which were reviewed by radiologists. Only 70 patients (6.4%) had clinically significant abnormalities on the CXR (the commonest was pleural effusion, followed by consolidation, heart failure, a new mass and pneumothorax). The only variables with a statistically significant association with CXR abnormalities were age, history of CHF, low sats and abnormality on auscultation. The Rothrock criteria+2 had a sensitivity of 92.9%, which equates to an NPV of 98.4%. (Academic Emergency Medicine, June 2018)


The bottom line: The diagnostic yield of CXR in chest pain is low (6%). The above decision rule could reduce the amount of X-rays considerably, with little increased risk to patients.


Note: A decision rule with 12 points is rather cumbersome in practice. However, these elements can easily be abbreviated into a tool that can serve as a useful aide memoire as follows (just my suggestion):


Obtain a CXR in patients with chest pain if…
– Age > 60
– History of CHF or smoking
– Abnormal vital signs
– Abnormal findings on auscultation


February 1st, 2019    

31. Which syncope patients need admission?

Background: Syncope is a common presenting complaint and it can be difficult to decide which patients need admission, particularly if no cause is found. Clinical decision rules (San Francisco syncope score, OESIL score, EGSYS score, ROSE score, Canadian Syncope Risk Score) are not used widely due to poor sensitivity and lack of external validation. San Francisco is probably the best-known of these, but this external validation study in 2009 only found a sensitivity of 74%, much lower than in the original research. Thankfully, the ever-busy European Society of Cardiology recently updated their 2009 guidance for managing syncope in the ED. It’s a chunky document, but here we’ll just focus on the recommendations for which patients need admission. (European Heart Journal, June 2018)


Low-risk features:
– Typical prodrome (light-headedness, warmth, sweating, nausea)
– Provocation (unpleasant sensation, hot environment, postprandial, or triggered by cough, defaecation or micturition)
– Position (prolonged standing or on standing)
– Long history of recurrent syncope with low-risk features
– Normal examination
– Normal ECG

These patients are very likely to have reflex, situational or orthostatic syncope and can be discharged directly from the ED


High-risk features:
– Chest pain, breathlessness, abdominal pain or headache
– Syncope preceded by palpitations
– Syncope during exertion or when supine
– Structural heart disease or severe heart failure
– Unexplained hypotension or GI bleed
– Undiagnosed systolic murmur
– Abnormal ECG (acute ischaemia, AV blocks or sinus pauses, tachyarrhythmias, long QT, HOCM, Brugada, WPW )

These patients should be admitted for further investigation


Note: patients with neither high nor low risk features should be observed for a period in a ‘Syncope Unit’ or receive close follow up in an outpatient Syncope Clinic


January 1st, 2019    

30. Management outcomes of children with isolated skull fractures

Head injuries are common in children. Most do not require a CT, and even if they do, the commonest finding is an isolated skull fracture (i.e. just a fracture, with no intracranial haemorrhage). How are these best managed? Previous studies of children with isolated skull fractures suffered from low numbers and poor follow up. The current paper is the largest study to date on this topic. The authors try to determine the risk of deterioration in this group of children.

Annals of Emergency Medicine, June 2018


January 1st, 2019    

29. Effect of steroid on migraine recurrence

Several meta-analyses over the last 10 years have found a benefit to using dexamethasone as an adjunct for reduced recurrence of migraine (for example, BMJ in 2008 and European Journal of Neurology in 2013). Pooled data estimate a number needed to treat (NNT) of 9. The suspected mechanism is that it works by dampening the state of cerebral neurogenic inflammation. The current study looked to see whether a longer-acting steroid would achieve a better response over the 7 days post-ED presentation.

Annals of Emergency Medicine, November 2018


January 1st, 2019    

28. Are biphasic allergic reactions real?

You've given your patient IM adrenaline, they're breathing normally and feeling well again. You'd like to discharge them with an Epipen but there's a little voice at the back of your head: "What if they have a biphasic reaction at home and they can't take the pen in time?" Prior studies have found biphasic reaction rates as high as 20%, and there are reports of reactions happening as long as 4 days after the initial anaphylaxis. Studies so far have tended to be underpowered, retrospective and with widely differing definitions of 'allergy' and 'biphasic reaction' (we're not really concerned about a slight rash appearing on day 2, after all).

Anesthesiology, November 2018

Annals of Emergency Medicine, June 2014


- Older Posts »