The workup for a possible SAH is a topic of much interest to ED doctors. The problem is the condition is fairly rare, the potential consequences of a missed SAH are devastating, and the workup remains problematic.

The LP in particular is a cause of much angst. Most patients are apprehensive about the procedure. It is time consuming, and can be technically difficult in some patients (especially the obese). There are frequent false positives – up to 20% (depending on RBC cut-off used). Spectrophotometry is advocated to help sort out these false positives. This however brings its own problems – spectrophotometry is not universally available, and differentiation between fresh blood (traumatic LP) and older blood (SAH) necessitates a 12 hour delay to do the LP.

Given all this, it is fair to say a patient awaiting a CT for a possible SAH is not the most popular handover!

There are 4 conditions we need to consider

  1. An incidental aneurysm. It is estimated that 4-6 % of the population have an intra-cranial aneurysm. Fortunately the vast majority of these will never cause problems and (as with prostate cancer), far more people will die with an aneurysm than will die because of the aneurysm.
  2. A symptomatic enlarging aneurysm which has not ruptured. This can be a cause of a headache, or occasionally a cranial nerve palsy (classically 3rd nerve). We want to find these as we believe a painful (enlarging) aneurysm is at significantly increased short term risk of rupturing – as with a painful enlarging AAA.
  3. An aneurysm which has ruptured and caused a SAH. This is the one we all think about. Some people are lucky and get a small warning bleed prior to a second larger bleed. We definitely want to catch this group. Large bleeds have a high mortality, and there is a high risk of permanent neurological damage if it is not fatal.
  4. A non-aneurysmal SAH. That is, a SAH arising from an angiographically normal artery. These can present like an aneurysmal SAH, with a sudden severe headache. However there is not much we can do about these bleeds – we can’t find the bleeding point, and we can hardly go around clipping off normal Circle of Willis arteries. Fortunately this condition has a better natural history and more benign long term prognosis than aneurysmal haemorrhages.

So we definitely want to diagnose 2 and 3. These are the conditions a neurosurgeon (or interventional radiologist) can intervene to change short term outlook.

It is probably helpful to know about 1 and 4. The patient can be advised they have an aneurysm or a bleed, they can be counselled to stop smoking, and hypertension (and other vascular risk factors) can be closely managed.

But the question we want to ask is “Does the patient have an aneurysm which has either bled, or which might bleed in the near future, and which is amenable to intervention?” This is the real question underlying our workup for ? SAH.

Prior to CT, the only way of diagnosing an aneurysm was by conventional angiography. This procedure has always carried significant morbidity and even mortality, and therefore was not a procedure to be undertaken lightly. LP was therefore used to risk stratify the patient. If there was no bleed, then it was not appropriate to expose the patient to the hazards of angiography. Only if there was a SAH was the risk of angiography justified. So we looked for the bleed to help decide if we wanted to look for the aneurysm.

Early generation CT scanners changed this a bit. They were good enough to find a large bleed, and were sometimes able to see a small bleed. They were not good enough to reliably find the aneurysm – especially if IV contrast was not used. So the rationale for CT was to find a bleed (rather than an aneurysm).  In addition, CT sometimes found another diagnosis (tumor, intra-cerebral bleed), and also helped exclude a space occupying lesion that would preclude LP. Because early CT missed small SAH, LP remained entrenched as part of the ED workup. So the algorithm then became CT, LP if CT negative, and then angiography if LP positive. From an ED perspective, the job was still to find a bleed, and admit the patient under someone who would go looking for an aneurysm. And so the ED literature is focused on the hemorrhage rather than the aneurysm.

Newer generation CT’s are much better at finding small bleeds. One large study has suggested the sensitivity is 100% within 6 hours of the bleed. If a patient presents with a headache within 6 hours and a good quality CT is performed and is read as normal by an experienced radiologist, then you are done. No LP. Some people are less confident that the sensitivity can be 100%. But most people agree it is pretty close and the need for a LP today is much reduced. Note you are still left doing an LP if the CT is performed more than 6 hours after headache onset.

So right now, most of us are doing a CT and doing this ASAP, and sometimes doing a delayed (12 hour) LP to sort out delayed presentations or perhaps a particularly high risk history. I think that is the state of play at present. Again, we are more intent on finding the bleed than the aneurysm.

However I suspect there is another chapter to be written in this story. The aneurysm (including a high risk non-ruptured aneurysm) is more important than the bleed itself. At present, our focus is perhaps too much on the bleed and not enough on the aneurysm. The original intent of the LP was to screen which patients underwent angiography. But the question the neurosurgeons always wanted to answer was about finding an aneurysm they could do something about, not finding the blood per se.

With current CT, we now have the option of going after the aneurysm. CT angiography has similar sensitivity to conventional angiography, is quick, and is non-invasive. So should we just do the CT and CT angiography for all SAH workup? If there is no aneurysm, then maybe you are done. Perhaps there is little point doing a LP looking for a bleed if there is nothing to be done (or that can be done) for it. The neurosurgeon (interventional radiologist) can do nothing for a non-aneurysmal bleed. They can only do something about an aneurysm.

There are pros and cons to such an approach:


  1. No more LPs. No more admitting patients to EMU for someone else to do an LP the next day. Even if the story is good. Even if the headache was 8 hours ago. If there is no aneurysm, then it doesn’t matter.
  2. Quick and easy. The patients will like it. Busy EDs will like it.
  3. If the patient has an aneurysm, it is probably worth knowing about. Stop smoking. Watch your BP. And do come to hospital quickly if you ever get a really bad headache.
  4. It is easier for us to manage the 5% of patients with a real (but incidental) aneurysm than the 20% with a traumatic LP.
  5. May find other conditions. We have had a missed carotid dissection in a 30ish year old female. She did badly. Her CT was normal. However a CT angiography may have found this.
  6. The small population of patients who present repeatedly with high risk sounding headaches don’t require comprehensive evaluation for a SAH each time. If they didn’t have an aneurysm 2 months ago, they don’t have an important SAH (that we can do something about) today.


  1. Need for IV contrast. This is not usually an issue as we are mostly looking at middle aged (rather than elderly) patients. However it will sometimes be a factor (especially with increased rates of diabetes in this population)
  2. Increased radiation. A 2nd scan with contrast after an initial non-contrast scan. Again, this is probably not a huge problem in the middle aged, but will be important in younger patients.
  3. The biggest concern may be labelling 5% or so of patients as having an aneurysm. It is probably some patients will be made anxious about being told they have an aneurysm. This knowledge may have implications for certain occupations (e.g. pilots), and also for life insurance or travel insurance (given the “duty to disclose” obligations of these policies).

On balance, I suspect this is the way we will go. Patients want a quick answer with minimum short term risk and discomfort. Doctors want quick answers, and want to be safe from medico legal risk.

With ongoing advances in CT, with increased availability of MRI and MRA, with EDs struggling with increased patient numbers, and with the increasing need to sort things out as efficiently as possible, I think all the drivers will push this in one direction only.

To people with a different view about this, I offer the following 5 letters. CT KUB.

Dr Steve Walker, Staff Specialist, Nepean Emergency Department.

Leave a Reply

You must be logged in to post a comment.