I wanted to touch on some cool research that was circulated in the SinaiEM Journal Club WhatsApp this morning inspired by the Journal Feed newsletter. If you have trouble staying up to date on EM research, Journal Feed can help! It sends one paper a day and summarizes it pretty well. To sign up, click here.

Title: Sensitivity of modern multislice CT for subarachnoid haemorrhage at incremental timepoints after headache onset: a 10-year analysis.

Authors: Annabel VIncent, Scott Pearson, John W Pickering, James Weaver, Leanne Toney, Laura Hamill, Michael Hurrell, Martin Than


  • Atraumatic SAH usually occurs from a ruptured aneurysm and account for 10% of sudden onset, severe headaches
  • Mortality approaches 50% at 3 months without definitive early intervention. 30% of survivors will have severe disability
  • CT H has historically had a Sensitivity of 93-95% in the first 24 hours
  • It is generally accepted in the EM community that CT H has a sensitivity of 100% in the first 6 hours from onset of headache, base on Perry et al, 2011. (Paper also attached)
  • CT scanner technology has improved since 2011, prompting further analysis if this window can be expanded.


  • This is a single center RETROSPECTIVE COHORT CHART REVIEW study.
  • Center: Christchurch Hospital, New Zealand, ED
  • Time Period: January 1, 2008 – December 31, 2017 (10 years)
  • Study Population: Charts with ICD-10 codes for SAH
  • Exclusion Criteria
    • 1. Cases determined to have been miscoded as SAH
    • 2. Clearly traumatic SAH.
    • 3. The admission was a repeat SAH admission during the time period. 
    • 4. SAH found on postmortem in whom no MSCT was performed. 
    • 5. The day of onset of headache was not recorded. 
    • 6. Patient transferred to Christchurch Hospital from another hospital because of difficulties accessing radiology reports and clinical notes. 
    • 7. Patients with lost or destroyed records.  
  • Of 728 initial patients with SAH, 347 met inclusion criteria
    • 332 were CT positive for SAH, 15 had SAH not detected by CT H
  • Radiologists were a mix of neuroradiologists and standard radiologists
  • There were 4 cases in which attending radiologists overread initially negative resident radiologist reports
  • The authors also additionally searched the hospital neurosurgery department database for SAH, and Coronial database for death from SAH within 6 months of an ED visit
  • Missing onset time was biased towards increasing the number of early false negatives. Basically, a conservative approach was used when data was missing. (224 patients)
  • Primary Outcome:  proportion of patients with spontaneous aneurysmal SAH that had a positive MSCT.
  • Secondary Outcome:  proportion of patients with any type of spontaneous SAH that had a positive MSCT.   
  • The authors presented the data in time windows from 0-6, 0-12, 0-24, 0-72 and 0-96 hours.


  • Primary Outcome:  The sensitivity for aneurysmal SAH at 6, 12, 24, 48, 72 and 96 hours post headache onset the sensitivity was 100% (98.0 to 100), 100% (98.2 to 100), 100% (98.3 to 100), 99.6% (97.6 to 100), 99.6% (97.6 to 100) and 98.7% (96.4 to 99.7), respectively  
  • Secondary Outcome:  The sensitivity for all SAH at 6, 12, 24, 48, 72 and 96 hours post headache onset was 100% (98.3 to 100), 99.2% (97.2 to 99.9), 99.3% (97.5 to 99.9), 99.0% (97.1 to 99.8), 99.0% (97.2 to 99.8) and 97.8% (95.5 to 99.1) respectively  
  • Of the 15 missed SAH by CT, 2 were within the 24 hour window. Both were found by LP. The 13 other SAH were found by LP or MRI


  • Data suggests 100% sensitivity for modern 3rd generation CT scanners for detecting aneurysmal SAH within 24 hours
  • Authors believe it is unlikely they missed SAH data points given the hospital is the only acute hospital in its region and the only neurosurgical center
  • Data is limited beyond 24 hours (due to lack of numbers

My (personal) Takeaway/Critiques:

  • We should be careful applying retrospective cohort studies to our population. Due to missing data points, which the authors address in this article, retrospective cohort studies tend to overestimate the effectiveness of diagnostic studies. In other words, we cannot accurately assess what was missed by CTH if we didn’t use a gold standard (likely LP) to r/o SAH. Furthermore, we don’t see patients with SAH in the ED. We see patients in whom we suspect SAH; this is different! 
  • The primary outcome totes 100% sensitivity up to 24 hours in Aneurysmal SAH. However 99.2% in all SAH. While rare, other causes of non traumatic SAH include vasculitis, dissection, AVMs, etc. (these were both caught by LP). It’s not a bad sensitivity either way, but I’m not sure I would differentiate my patients into aneurysmal vs other non traumatic SAH. 
  • The sensitivity (100%) of CTH for SAH is well recognized from 0-6 hours. the majority of the cases in this study (224 out of 347) occur within the first 6 hours. Including these 224 cases can consequently skew the data towards higher sensitivity. In the study population of all SAH, when the 224 patients who had a negative CTH in 0-6 hours are removed, the sensitivity is closer to 96.9% when CT occurred from 6-24 hours from headache onset.

Bottom Line: This is a cool study with positive implications for expanding the window of use of CTH in order to r/o SAH, but I’m not convinced it will change my current practice. A prospective multicenter trial would be the next step.


Sensitivity of modern multislice CT for subarachnoid haemorrhage at incremental timepoints after headache onset: a 10-year analysis. Emerg Med J. 2021 Nov 24;emermed-2020-211068. doi: 10.1136/emermed-2020-211068. Online ahead of print.

July 2024