Listen to Drs.Newman & Shreeves muck around in some of the root literature on how we diagnose of Subarachnoid hemorrhage.
With roughly 21-33 000/year in the US, representing 2-5% of new strokes, SAH is a dangerous but uncommon entity.
- SAH can be divided into Aneurysmal ⅘ and non-aneurysmal ⅕, with the former being more worrisome as they tend to do worse.
- 80% patients diagnosed with SAH will have a neurosurgical intervention.
- Prevalence 2% (1/50) of aneurysms in population are healthy normals!
- Risk Factor for aneurysms: PCKD, Ehlers-Danlos, Smoking, HTN
- From historical literature of patients with SAH ⅔ die. ⅓ pts with SAH ( get better with bedrest)
- From root literature diagnosed with Hunt/Hess 1 & 2 we know that 45% are alive at 7 years & 20 years…
- Of those diagnosed with SAH, who undergo surgical treatment, 75% have favorable outcomes, which represents a 30% benefit, an ARR 30% and a NNT 3
The Missed Subarachnoid Hemorrhage
- The Sentinel Bleed only exists with the retrospectoscope
- Spectrum bias must be considered when reading studies pertaining to the sentinel bleed
- Neurosurgery data: Denominator is all those with SAH
- Emergency data: Denominator is all headaches
- Kowalski (PMID:14970066) demonstrated a 20% miss rate (recently seen by provider for HA-related complaint)
- of those missed SAH, approx 20-30% had potentially preventable complications…
- However, the denominator was all SAH (NOT all HA pts)
- In a large retrospective Canadian study (PMID:17322078), evaluating all SAH with clinically important outcomes, 10% had an ED visit within the the previous 2 weeks, however, only 5% (1/20) with HA-related visit
- interestingly when you compare the 30 day mortality:
- Diagnosis was missed : 6%,
- Diagnosis was made: 34%
- which makes sense- 1) sicker patients are easier to diagnose 2) those with milder disease (even if they are missed) still have a good prognosis.
- Patients with thunderclap headache (maximal intensity within seconds to hours of onset) and are associated WITH a focal neuro deficit have 15-25% of having a finding on Head CT.
- Bear in mind, in the ED, we often worry about those with just a headache, not necessarily a neurological deficit (that would increase your pre-test probability).
- In a study by Stiell in 2010 (PMID:21030443) he showed a prevalence 6.5% of SAH
- Of the patients evaluated 78% patients presented with worse HA of life (although more common in pts with SAH however there was poor inter-rater reliability (kappa value) , and is NOT a reliable predictor
- Edlow & Wyer (PMID:11054205), with the assumption of Sensitivity 93% & Specificity 98%, and 15% pre-test probability (high by today’s standard) determined that the likelihood for SAH after negative CT = <2% (post-test prob)
- When considering the kind of patient we are worried about, we don’t worry about the visibley sick patient, we worry about those who present to the ED with a headache, and have a normal neurological exam.
- The Danish study referenced (PMID:20404693) reports Sn 93% and Sp 100% for identifying badness on CT
- therefore, using a pre-test probability of 6.5% (high risk patients), and a negative LR 0.7%, the post-test probability is 0.5% of missing SAH on a HeadCT or 1/200.
- of the 1/200 however, we know that ⅕ are non-aneurysmal (less worrisome due to their “good” prognosis, and not being amenable to intervention, we have 1/250 risk of missing an aneurysmal SAH after a negative Head CT..
- Furthermore of the those that we do identify, approximately (30-)40% will benefit from our interventions (surgical clipping, supportive care)
- so patients with a negative head CT, we can exclude the patients with non-aneurysmal SAH (as source of bad outcomes – which is all we really care about), we can state that performing an LP will ultimately benefit only 1/625.
Please complete the post-test for your asynchronous credits