Skip to main content

Non-traumatic subarachnoid haemorrhage (SAH) accounts for 5% of acute headache presentations to the emergency department. The urgency to make a diagnosis when patients have a ‘warning leak’ headache, relates to our opportunity to make the diagnosis early, which if missed and the patient has a rebleed, is associated with about a 50% mortality, with half of survivors having significant morbidity.

The approach has generally been to perform a non-contrast CT brain, which is sensitive in the first 6 hours. If the pre-test probability is high, we can then move to a Lumbar puncture (LP) or CT cerebral angiogram (CTCA).

This cohort study looked at the trends in diagnostic testing for the evaluation of headache in the emergency department.

What They Did

This was a retrospective Cohort Study in 21 American Emergency Departments, from 2015 to 2021. There were 198109 encounters during this period.

Primary Outcome: Detection of unruptured intracranial aneurysm within 14 days of visit.

Secondary Outcome: Detection of unruptured aneurysm within 90 days and subarachnoid haemorrhage within 14 and 90 days.

Safety outcomes were missed diagnoses of SAH or bacterial meningitis.

What They Found

Overall there was a significant increase in CTCA, with an associated decrease in LP post CT brain.

  • There was a year-to-year increase in CT use, annualised percentage change of 5.4% (95% CI, 5.1% to 5.8%) (Table 2).
  • There was a year to year increase in CTCA use with an annualised percentage change of 18.8%
  • There was a  decrease in LP use with a annualised percentage of  −11.1%.
  • There was a marked increase in the CTCA:LP ratio with an annualised percentage of 35%.
  • There was a 5% misdiagnosis of subarachnoid haemorrhage (consistent with previous literature)
  • There was an 18% misdiagnosis of bacterial meningitis cases (consistent with previous literature)

Outcomes: There was an increase in the 14 and 90-day unruptured aneurysm to subarachnoid haemorrhage ratio.


This study is of interest as in 2021 ACEP made a recommendation, for CTCA being an acceptable alternative to LP for potential subarachnoid haemorrhage.

However it is also important to note that the CTCA is not sensitive in picking up small aneurysms (<3mm), which count for some 18% of ruptured aneurysms.

We know that a negative CT within 6 hours has a low risk of missing a SAH, comparable to that of CTCA ie., 1 in 700-900, and the risks associated with LP (about 5%), CTCA has increased in popularity.

One of the drawbacks with CTCA is the incidental finding of aneurysms, which may or may not be significant, given that we know that up to 6% of the population have incidental aneurysms found at autopsy.

These findings don’t change my practice.

  • If a patient presents within 6 hours of headache (suspicious for SAH), that patient gets a plain CT and CTCA. If all negative, then we are done, unless my suspicion is very high.
    • If the CT is positive for SAH, but the CTCA is negative, I discuss with neurosurgery as this may be a non-aneurysmal SAH (although a small number of aneurysms in spasm, as well as very small aneurysms, may be missed).
    • If the CT is positive and the CTCA positive: refer to neurosurgery
    • If the CT is negative and the CTCA is positive: This represents the most difficult situation. There has to be some shared decision making here. I will propose an LP in this case, looking for xanthochromia (I delay LP for 12 hours) ie., to determine if the aneurysm has in fact bled , or if it is an incidental finding.
  • If it is greater than 6 hours since headache onset and there is suspicion for a SAH, I discuss the need for LP with the patient.

Listen to the podcast discussion on this topic at EM MASTERY

You might also wish to read: There is no consensus in the way we investigate potential SAH



  1. Dustin M et al. Shifts in Diagnostic Testing for Headache in the Emergency Department 2015 to 2021.JAMA Netw Open. 2024 Apr; 7(4): e247373.

Leave a Reply