Subarachnoid Hemorrhage, no LP
Al Sacchetti, MD and Chris Carpenter, MD
Print Editor: Whitney Johnson, MD, MS
- If CT using a modern generation scanner is performed within 6 hours of symptom onset, there is strong evidence we do not need to do a lumbar puncture (LP).
- The number needed to LP after negative CT scan to identify one aneurysmal hemorrhage is somewhere between 250 and 15000.
- Use of CTA is controversial and not well-explored in the literature.
- In a previous segment (EM:RAP 2017 June - Subarachnoid, MRI, and the Case of the Vanishing LP) Anand Swaminathan MD and Evie Marcolini MD discussed lumbar puncture in the evaluation of suspected subarachnoid hemorrhage. Sacchetti disagrees.
- Carpenter, CR et al. Spontaneous subarachnoid hemorrhage: a systematic review and meta-analysis describing the diagnostic accuracy of history, physical examination, imaging and lumbar puncture with an exploration of test thresholds. Acad Emerg Med. 2016 Sep;23(9):963-1003. PMID: 27306497
- A patient presents with a severe thunderclap headache associated with vomiting. The CT is negative. What is the next step in the work-up?
- Classically, we were taught that you need to do a lumbar puncture (LP) after the CT scan because the first generation of CT scanners showed we were missing up to 10% of subarachnoid hemorrhages. At Carpenter’s site, they have developed a protocol in conjunction with their neurosurgeons where they do not necessarily do a post-CT LP. They look at how soon the CT was performed after the onset of the headache. If it is within 6 hours, there is strong evidence that we do not need to do an LP and there should be shared decision-making with the patient.
- What goes into the shared decision-making? How many lumbar punctures would you have to perform after a negative CT head with a modern generation CT scanner to find one aneurysmal hemorrhage that you would have missed with the CT? The number needed to LP is anywhere from 250 to 15000.
- Beyond 6 hours, their meta-analysis shows fairly conclusive evidence that a modern generation CT scanner can accurately rule out aneurysmal subarachnoid with a likelihood ratio of 0.7 and tight confidence intervals. Any negative likelihood ratio under 0.1 is extremely powerful for risk-stratifying the patient to a very low likelihood of disease. Their practice is to discuss the limited role of lumbar puncture with the patient if their only concern is subarachnoid.
- They are doing some CTAs after 6 hours from time of onset. However, there isn’t conclusive evidence that CTA is the next best test to do. A few more observational studies are needed to show the number of false positives with CTA. We know that up to 2% of the population is walking around with cerebral aneurysms they did not know about. These will be detected on CTA. We need some randomized controlled trials comparing CTA to follow-up and see if outcomes are improved with CTA.
- Are the patients diagnosed with an aneurysm on CTA going to head to the ER every time they get a headache? Sacchetti doesn’t think so. We diagnose patients with 3 cm abdominal aneurysms all the time and they do not come in every time they have a bellyache. Identifying an illness that does not need immediate intervention is not necessarily a bad thing. However, this could also be viewed similar to a patient with a pulmonary embolism (PE) who ends up getting a CT every time they go to the ER with chest pain and shortness of breath thereafter.
- If you do the lumbar puncture and it looks like a bleed, you will proceed to CTA. Does the order of these studies matter?
- If you are a single-coverage doctor in a busy community emergency department and have to sit down and spend 20 minutes doing an LP while keeping the ER running, there is a lot of appeal to just ordering CTA. This isn’t a great answer but it is the reality of practicing medicine.
- The study also delved into the utility of history and physical examination. They found that history and physical exams were not terribly accurate. The sudden onset of headache is not predictive to rule in or rule out subarachnoid hemorrhage.
Recent Related Material
EMA 2017 Essay: Can We Safely Remove The Lumbar Puncture From The Subarachnoid
Lance M. - May 2, 2020 9:02 AM
Dr. Sachetti references another conversation about the topic. Would be nice to just link or show the issue so we could find it and compare. Thanks.
Chris Navarro: EM:RAP Production Team - May 5, 2020 8:20 AM
I've added the EM:RAP and EMA links to the written summary.
Carp27 - May 4, 2020 3:42 PM
The segment Al refers to was "Don't Feat the Stink" in June 2017 episode of EM:RAP with Evie Marcolini and Swami. There is also an EM Abstracts essay on the topic of post-CT LP for SAH in March 2017 episode of EMA that goes into greater detail on the ED-specific diagnostic evidence.
Alfred S. - May 4, 2020 5:37 PM
Thankfully Chris has a much better memory than me.
Brian L. - May 6, 2020 9:37 AM
A mistake in the text.
LR 0.7, should read 0.07
Alfred S. - May 6, 2020 11:22 AM
Charles M. - May 25, 2020 9:57 AM
Beyond 6 hours, their meta-analysis shows fairly conclusive evidence that a modern generation CT scanner can accurately rule out aneurysmal subarachnoid with a likelihood ratio of 0.7 and tight confidence intervals.
The 0.7 should be .07 (as noted above) and the phrase needs to say it's the <<negative>> LR (word 'negative left out).
Yvonne K. - May 11, 2020 8:31 AM
Question: in the discussion of CTAs for subarachnoid hemmorage beyond 6 hours of sympton onset. if you do a CTA right after the CT head, why and who would follow up with an LP (ie the order may matter because we could spare an LP)
Alfred S. - May 11, 2020 10:44 AM
I guess you can sum this up this way.
1. CT within 6 hours with late generation CT negative for SAH then done.
2. CT at any time positive for SAH, then CTA.
3. CT > 6hrs after onset of symptoms is negative then offer patient option of CTA or LP.
4. If LP positive then need CTA
5. If CTA positive, with or without proceeding LP, then discuss with Neurosurgery / Interventional Radiology to see if meets high risk criteria based on age, location and size.
If it is me in your department, just do the CTA.
Michael L. - May 13, 2020 7:28 AM
I grew up in the era of doing an LP after every negative CT for SAH. The plus side of that practice is that you get really good at doing LPs. I've seen younger docs (can't believe I'm using the term "younger docs") who are less comfortable/facile with LPs given the change in practice with a lower threshold to send the patient to fluoroscopy. Neither here nor there, I suppose, but an interesting unintended consequence.
Alfred S. - May 13, 2020 7:35 AM
Yes, you are old. But there are many procedures that have changed with new technology. Landmark central lines, myleograms for epidural abscesses, and open abdominal surgeries are all some of the procedures that have disappeared.
Rick M. - May 20, 2020 11:01 AM
I have been dong CT/CTA for suspected SAH for what seems like decades now. My experience has been as follows: positive tap and negative CTA, the Neurosurgeon says, "what do you want me to do...an exploratory brain surgery?" If the CTA is negative...what is the neurosurgeon or interventional radiologist going to do? If the LP was negative, but the CTA was positive...I am still going to call the neurosurgeon and/or the IR because the symptoms were suspicious, and the CTA shows an aneurysm. If the CTA is negative...the results of the LP (which may be traumatic) does not and never has changed the end result...a discussion with a neurosurgeon who says, "well there is no aneurysm for me to go fix." So, I get the CT/CTA.
I would like to know how many times a positive LP with a negative CTA has led to any further action, other than "follow-up?"
Alfred S. - May 20, 2020 12:48 PM
Your argument makes sense.
Steven P. - June 26, 2020 11:09 AM
Just to clarify: If patient presents greater than 6 hours since symptom onset and CT brain is negative and a CTA results as negative with no LP performed, am I done?
If so, are you aware of any evidence that supports this approach?
Steve P. MD
Dean B. - August 18, 2020 3:41 AM
In Al’s segment, May 2020 EMRAP, who is reporting the CT scan & does this matter?
Academic/university centres are served by neuroradiologists, and these CTs are checked & reported by neuroradiologists. Smaller centres have in-house reporting by general radiologists till 7.00pm or so, and then the reporting is out-sourced, to a radiologist who is sometimes across the other side of the world. Does this discrepancy between reporting abilities in different centres change the recommendations made in this segment?
Alfred S. - August 18, 2020 7:57 AM
I am not aware of any data that suggest a neuroradiologist is required to read a plain CT looking for subarachnoid blood. I think this is a reading any radiologist should be capable of identifying.
Regina R. - September 8, 2022 5:12 AM
what is defined as a modern CT scanner?