The SHED Trial for SAH in ED

Acute headache acounts for up to 2% of Emergency Department presentations, with up to 10% of those patients having a serious cause for their headache.  Subarachnoid Haemorrhage(SAH) is the most frequent serious pathology found. Most patients with suspected SAH will undergo non-contrast head CT. There is some concensus, that a negative CT, performed within 6 hours of headache onset, requires no further investigation. However the sensitivity of CT scan in these situations is a not perfect, ie., 98.7% and so the debate continues. The area that needs more work is what we do with patients presenting greater than 6 hours following headache onset. There were some interesting insights in this study.

The Study
The Trainee Emergency Research Network (TERN). Subarachnoid Haemorrhage in the emergency department (SHED): a prospective, observation, multi centre cohort study. Emerg Med J 2024;42:719-727.

Bottom Line
CT imaging within 6 hours of headache onset had a high sensitivity 97% CI (92.5 - 99.2) and a very low post-­test probability for SAH. The sensitivity of CT reduced over time, interestingly, the post-­test probability for aneurysmal SAH after a negative CT up to 24 hours was 0.1%.

What They Did

The Subarachnoid in the Emergency Department study (SHED) had two aims:

1. To externally validate the use of CT brain imaging within 6 hours to exclude SAH.
2. To look at the use of the CT scan betond 6 hours and externally validate the Ottawa SAH Clinical Decision Rule in a UK population.

This was a multicentre, prospective, observational cohort study in the UK. Clinical data such as headache onset time, time to peak headache and each part of the Ottawa CDR were collected prospectively. Patients were followed up within 28 days.

Inclusion Criteria

• >18 years
• GCS 15
• Presenting with non-­traumatic acute headache reaching maximal intensity within 1 hour.

Exclusion Criteria

• Head trauma in the previous 7 days
• Returning patients with the same headache
• Prior diagnosis of
  • SAH
  • Brain neoplasm
  • Ventricular shunt or hydrocephalus
• Focal neurological deficit
• Headache onset >14 days prior to attendance
• Recurrent headaches (> 3 similar headaches)
• Transfers with confirmed SAH;
• Prisoners and patients currently detained under the Mental Health Act.

Outcome measures

Definition of SAH:

1. Subarachnoid blood reported as present
   1. on unenhanced CT-­ brain
   2. on CT-­angiogram or MR-­angiogram
2. CSF Investigation:
   1. Spectrophotometry cerebrospinal fluid (CSF) 
   2. Visible xanthochromia on LP
   3. Red blood cells > 5×106/L in the final tube of CSF collected and an aneurysm identified on cerebral angiography (digital subtraction, CT or magnetic resonance angiography).

In cases where results were contradictory or inconclusive a panel of clinicians that were independent of the study adjudicated as to whether the result was a positive or negative for  SAH.

N = 3232 patients. 

CT brains were performed in:

• 772 (23.9%) within 6 hours of headache onset
• 708 (21.9%) within 6-12 hours of headache onset
• 323 (10%) within 12-18 hours of headache onset
• 205 (6.3%) within 18-24 hours of headache onset
• 1223 (37.8%) at > 24 hours of headache onset

What They Found

A total of 237 patients were diagnosed with SAH (6.5% prevalence)

• 208 (87.8%) were diagnosed on initial CT
• 29 met other diagnostic criteria
• 23 had a positive LP
  • 3 of these patients had a false negative CT performed within 6 hours.

    • 2 were non-aneurysmal with further investigation.

    • 1 was found to have two aneurysms, that were not coiled.

• 1 had RBC >5×106/L and aneurysm on further imaging,
• 4 had evidence of SAH on further imaging without LP
• 1 patient had a subsequent diagnosis of SAH at 28 day 

Of the 237 patients diagnosed SAH:

• 133 (56.1%) had aneurysmal SAH,
• 67 (28.3%) had non-­aneurysmal SAH, and
• 37 (15.6%) had no aneurysm status determined

Please see table 2 reproducednbelow for full CI. 

External Validation of the Ottawa SAH Clinical Decision Rule

They found a sensitivity of 98.3% (95% CI 96.2% to 99.4%) and a specificity of 8.1% (95% CI 7.3% to 8.9%) for the diagnosis of SAH. The NPV was 98.6% (95% CI 96.7% to 99.5%) and a negative LR of 0.21 (95% CI 0.08 to 0.53).

Discussion and My Take on This

CT imaging within 6 hours of headache onset had a high sensitivity 97% CI (92.5 - 99.2) and a very low post-­test probability for SAH. The sensitivity of CT reduced over time, interestingly, the post-­test probability for aneurysmal SAH after a negative CT up to 24 hours was 0.1% (95% CI 0.0% to 0.4%).

"The Ottawa CDR showed high sensitivity but very low specificity. The rule conveyed minimal impact on post-­test probability, given the low baseline pre-­ test probability."

This was a large observational study that fell short of the initial pre-specified sample size of 9000 patients. There was no intervention of care as it was an observational study ie., usual care was carried out.  This may have resulted more investigations were being performed than usual, especially as the Ottawa SAH decision making rule was used. 

I think that this is an important study as it reinfoces that CT up 6 hours of headache onset, when there is suspicion of a SAH, is very sensitive. However I am one that still has concerns about saying that these patients are clear, purely because we know that we can miss up to 3% of patients in this way. We need to perform about 300 LPs to pick up one of these, however the potential catastophic results of a rebleed, require that we discuss this realistically with our patients. Shared decision making has to be open and honest.

For me the most important part of this study was the very low post test probability of aneurysmal SAH (0.1%) following a negative CT up to 24 hours following headache onset. The negative predictive value at 24 hours was 98.5 as compared to 98.6 at 6 hours.If it needs no intervention, then it may not be significant. I still perform a CTA, when the story is good. What do I need for a good story?.... Sudden severe headache.

Validation on these figures for delayed CTs will help us make further decisions. If I was sure that the NPV at 24 hours was 98.5%, then I could add a CTA and have a maeningful discussion with my patient.

Full table 2 from the study: