Treating the 33-Year-Old with a Severe Headache. Did This EP Do Enough?


A 33-year-old female was brought to the emergency department by her husband for evaluation of a diffuse headache of 4 hours’ duration that began gradually and increased to an 8 of 10 intensity over 30 minutes. The headache was associated with bilateral facial and hand numbness, slurred speech, muffled hearing, and a brief syncopal event. The patient denied fever, neck pain, vomiting, focal weakness, vision changes, or photophobia. Nurses’ notes documented that although the pain was typical of her prior migraines, the additional symptoms were not. Her medications included oral contraceptives and Topamax for migraine prevention.

Physical exam showed a young woman in mild to moderate distress. She had clear speech but a strange speech pattern that her husband stated was different from her baseline. Her neurologic examination was otherwise normal. Examination of the head and neck was unremarkable, but no hearing assessment was documented. Examination of the abdomen, chest, heart and extremities was also normal.

The patient initially received one liter of IV saline, Phenergan 25 mg IV and morphine 4 mg IV. No diagnostic testing was ordered. Two hours later, the patient was noted to be somnolent. The physician wrote discharge orders, but the patient could not remain awake long enough to dress herself. The physician re-evaluated the patient and noted that the patient was now lethargic. Narcan was ordered, but serial doses of Narcan had little effect on the patient’s level of consciousness, prompting lab testing and a brain CT.


The emergency physician read the CT scan as normal. Lab testing was also normal. With no clinical improvement 3 hours after receiving Phenergan and 1 hour after receiving Narcan, a lumbar puncture was performed. LP results were also normal. The patient’s symptoms worsened and a brain MRI was ordered. The patient deteriorated further and was intubated for airway protection.

The brain MRI revealed an acute thrombosis of the basilar artery affecting the cerebellum, midbrain and thalamus. Seven hours into the patient’s ED course a neurologist was consulted and recommended mechanical thrombectomy, which was performed nearly 9 hours after the patient first arrived in the ED. The following morning, radiologist over-read of the patient’s initial CT described an acute cerebellar stroke with a dense basilar artery sign. The patient’s stroke progressed to involve the pons, causing weakness in all four limbs. After 6 months of rehab she was able to walk using a walker. She filed a lawsuit against the emergency physician and the hospital.

At trial, the defense argued that the patient’s presentation was most consistent with migraine headache and hyperventilation, that the neurologic exam showed no concerning abnormalities, and that none of the patient’s symptoms were typical of a stroke. The defense also argued that the patient’s somnolence was a known side effect of medications she had received.


The plaintiff argued that migraine headaches and oral contraceptives are both risk factors for strokes and that the patient’s symptoms were a typical presentation of patients suffering from a posterior circulation stroke. The plaintiff expert argued that it was a breach in the standard of care to assume that syncope, speech changes, hearing changes and facial numbness were due to hyperventilation without first assessing the patient for a stroke – especially when the patient’s symptoms did not fit her typical migraine pattern. The plaintiff’s expert also argued that it was unreasonable for the emergency physician to interpret CT scans when remote night time radiology groups were widely available.

Was the patient’s care within the scope of reasonable practice?
Click here to take the survey or sound off in the comment thread below. Your response may be included in next month’s analysis.


SENIOR EDITOR DR. SULLIVAN, an emergency physician and clinical assistant professor at Midwestern University in Illinois, is EPM’s resident legal expert. As a health law attorney, Dr. Sullivan represents medical providers and has published many articles on legal issues in medicine. He is a past president of the Illinois College of Emergency Physicians and a past chair and current member of the American College of Emergency Physicians’ Medical Legal Committee. He can be reached at his legal web site

EMERGENCY ULTRASOUND SECTION EDITOR Dr. Pregerson manages a free online EM Ultrasound Image Library. He is the author of the Emergency Medicine 1-Minute Consult Pocketbook and the A to Z Pocket Emergency Pharmacopoeia & Antibiotic Guide (available at and the Tarascon Emergency Department Quick Reference Guide (  


  1. I’d be curious to know if she had a history of seeking care in the emergency department for her headaches. Obviously that can’t be enough to rule out a serious condition, but if she’s had migraines for some time and has never felt the need to go to the ED, I would view that as a red flag.

    I also tend to think LOC and her husband reporting a change in speech pattern warrants earlier imaging.

    • Something like this actually happened to my husband. He was taken to the ED at 0400 in the morning for a head ache that woke him out of sound sleep. He presented with hypertension and the inability to sit in one place for any period of time…pacing. He repeated stated that his head hurt to the staff but no one came to evaluate. Three hours later, after shift change and NEVER being evaluated by an MD, he was given a CT scan which revealed a bleed in all four ventricles of the brain. He was sent home with a follow up with a neurologist in hopes that it would resolve on it’s own. His headache continued and we took him back to the ED where he was intubated and flown to a level I trauma center. There he had a shunt placed. Luckily, after he crashed in the helicopter, he did survive with minimal deficits. He, however, is disabled and can no longer work.
      The problem with that story is that, even though he was ignored for a long period of time, there are no parameters set for treatment of a stroke/bleed. Just suggestions. So malpractice was not a viable option after consulting an attorney.
      I just wish that people would pay attention to the patient at the time they present, not the patient visit history or anything outside of that visit.

  2. I would have scanned her earlier, but in this case it wouldn’t have changed anything, as there was no nighttime radiologist. That’s the biggest issue in my view. I would not be comfortable having to make definitive reads on CTs/MRIs overnight.

  3. Brad Bourkland MD on

    I am biased because I have read through the case and know the final diagnosis. However, the patient’s unusual neurological symptoms accompanying her headache concern me, especially when it is stated that this is not typical of her usual migraines to have these additional symptoms. I think I probably would have done the imaging sooner, followed by the LP. That being said, would that have made a difference? No, especially if I had read the CT as negative and the LP was negative. The real question here is why was a radiologist not reading this CT at the time the patient was in the ED? I think that is pretty common now days to have a Nighthawk or StatRad type night radiology service for advanced imaging interpretations. I enjoy looking at the CT’s I order for my own amusement, curiosity, and education, but I certainly would not rely solely on my own interpretation, especially when missing something could be catastrophic.

  4. Care seemed reasonable except for the EP interpreting a CT Head without a formal radiologist interpretation. Standard of care for CT interpretations should be a residency trained bc/be radiologist interpreting the study (vrad should have been utilized)

  5. Everything seemed to be within the standard of care until the EP read and presumably acted on his/her own CT interpretation.

  6. According to the synopsis, this headache was not typical of “her usual migraines,” as new associated symptoms were present (i.e. numbness, slurred speech and syncope). This new symptom complex would raise suspicion for SAH, stroke, vertebro-basilar insufficiency, vertebral artery dissection and atypical migraine,. In my opinion, the ED physician erred by 1) not ordering a CT scan immediately and 2) not having a radiologist interpret the CT. This could have been accomplished within an hour or so. The CT findings (as interpreted by the radiologist) and associated symptoms, which were apparently new and not present with her usual headache pattern, should have then prompted emergent neurology consultation and additional emergency imaging (MRI/MRA). In all likelihood, this could have been accomplished within the first 2-3 hours after her presentation and anticoagulation and thrombectomy could have been performed 5-6 hours earlier (within 3-4 hours of her presentation).

    Having said that, the effect of earlier thrombectomy on the final outcome would be speculative. She had apparently been symptomatic for four hours prior to ED presentation.

    In my opinion, negligence and possibly causation could probably be established. Damages resulting from the delay in treatment would be more debatable.

  7. For such a atypical presentation, this somewhat delayed, but thorough, evaluation, especially in a busy ER, would be the standard of current care.

  8. I’ve worked at a number of small (and not so small) rural hospitals with xrays being read by me overnight but I’ve never heard of a place where the ER guys needed to read the CTs

  9. EPs should not take upon themselves CT interpretation. Yes, we will, most likely, identify a large bleed, but not a basilary artery hyperdense signs.

  10. Difficult case indeed. I will always struggle with the notion that, common things being common, a pt with a hx of migraines is far more apt to suffer recurrent migraine than a new intracranial process, and that it is not UNcommon for a relapsing condition, ie., migraine, may present in various ways, ie., I’m never sure what to make of the pt who states, “I have “X” and these symptoms don’t feel like “X””, even though- independent of this particular pt- the symptoms may otherwise be consistent with “X” (similarly, the pt who says, “doesn’t feel like my GERD or angina”, or “I always get kidney stones on the other side”).
    I suppose I would have imaged this person earlier, despite the hx of migraines, only if her speech pattern was manifestly odd to me.
    As others have remarked though, the critical issue here is the (apparent) lack of real time radiology interpretation of the head CT. I cannot imagine working at a place where I don’t have a 24/7 radiologist for all advanced imaging studies. If I were the EP here, I would sue the hospital in turn for failing to provide an undeniably critical service, and thereby mandating that I practice entirely outside the scope of my training.

  11. I must say I remain a bit confused. This pt. had a normal neurologic exam, with the exception of a “strange speech pattern” (? dysrthria?) and had a thrombosed basilar artery? Really? No cerebellar findings? No Babinski? And she presented with a headache? I think I would have been more concerned with a vertebral artery dissection (if there were no SAH), and I would have obtained a CTA of the head and neck. However, even if diagnosed earlier, I doubt this would have made much difference. Basilar artery strokes are particularly difficult to treat

  12. Mike Duerr, MD, PhD on

    Tough case. Also am a little confused that there are no clinical cerebellar signs. Never considered dysarthria as a tell-tale sign of a posterior CVA. Certainly enough red flags to warrant admission for observation in my mind. Brief LOC and then not regaining a normal level of consciousness, would not send that home. I have worked in places where we read our own CT scans at night, however, that was 20 years ago. Think that an ED physician cannot be expected to detect some of the more subtle signs of an ischemic stroke. Also it appears that the CT interpretation by the radiologist was after the MRI had been obtained. Wonder whether an initial read by radiology without that knowledge would have been the same.
    All in all, IMHO, acceptable care, no breech of SOC. Difficult diagnosis, with overall poor prognosis. Thorough workup, though some time delays, these would appear within reason in an average ED.

  13. It seems an unusual presentation for such an exotic diagnosis. The physician did a neurological examination, a CAT scan ,lab work,and an LP, all of which were normal. At my facility, overreads are done when the radiologist knows the diagnosis. I think the ED physician met the standard of care

  14. Even if the radiologist read the CT at night as acute stroke, would anyone argue that she needs an emergent MRI? She would be outside Tpa window and thrombectomy would not have even been considered as option at most institutions. And there is no role for heparin in acute stroke. I agree that a more expedited ct should have been done and that a rads should read it, I just don’t see this as effecting management. Unfortunate outcome but I don’t think this is malpractice

  15. THis is the fault of the system, not the DEM physician. Biggest flaw– not having neuro-radiology, and perhaps not having access to CTA since it wasn’t mentioned as an option. Post circ syndromes are tricky– requires a high degree of vigilance. Perhaps some anchor bias to the HA?– which doesn’t classically occur with acute ischemic strokes. Would like to have more specific information for the symptoms/exam (ie- could the pt walk initially, dysmetria, dysarthria, diplopia, dysdodokinesis, dizziness/vertigo, etc– other “d”s of the posterior circ). I would agree that the facial/hand symptoms, speech difference, and syncope merit consideration for neuro-imaging, but, obviously that is the call of the MD in the trenches. And that call can be influenced by many factors as cited above (ie- prior presentations, frequent user, h/o migrains, etc). Just my 0.02$ for a challenging presentation……..

  16. “Normal neurologic exam” is a real trap. Most of us do a cursory exam, although specific cranial nerve complaint (paresthesias) should demand a detailed CN exam; cerebellar and posterior column exam should be documented as well. I think we’re all thinking, “what if I had seen that patient?” and feeling a little defensive over the case. The pivotal info here is the reading on the CT, which all agree is not reasonable timely care. What followed the EP reading, tap for SAH, was too delayed, and that had downstream effects on time to further evaluation and diagnosis. This is really a hard case. I might miss this- but I shouldn’t. As noted above, outcomes are poor regardless of timing, but delay in diagnosis doesn’t help.

  17. Yikes- and by “yikes” I am referring to the EM physician’s care of this patient. First there is the 4mg of morphine ordered as a first line medication. While not malpractice in of itself, this is a red flag to me that the physician (presumably EM residency trained and board certified?) is not staying up to date on the latest guidelines on headache treatment and management. I will unequivocally say that opioids are never first line treatment for any headache. Even if you have ICH I think a round of other headache meds (other than NSAIDs) are still indicated as first line therapy. I’ll go one step further to say that opioids have virtually no role the in the ED management of headache. If a patient has a headache so severe and has failed multiple rounds of common HA meds then they should be admitted for status and the inpatient team can order whatever they want. The fact that morphine was used as a first line treatment suggests a need for further education in headache management which then makes me concerned about knowledge deficits in the care of neurological emergencies.

    This concern was then brought forth when the patient complained of slurred speech. While we have all seen hyperventilating patients with bilateral arm numbness, slurred speech is not typical of an anxiety reaction. While this could be attributed to a complicated migraine, this needs to be run to ground with a CT and admission for a stroke workup. The CT should have been ordered as soon as the patient was evaluated with an interpretation by a board certified radiologist. Meds (other than morphine) could have been started while waiting for the CT but the delay in CT was a bad move. While I will wet read my own CTs to expedite care in sick or concerning patients, I would never make a disposition decision on what I believe is a negative head CT without a radiologist weighing in. While most of us can be counted on to catch a large bleed, a dense basilar artery sign is way outside of my area of expertise. While nighttime radiologists can be a pain and difficult to get a hold of to expedite reads and clarify CT reads, they need to be utilized for the interpretation of CTs.

    However- all that being said- it’s really hard to say that the delay in eventual diagnosis led to the patient’s poor outcome. We are essentially talking a delay of 5-7 hours. While the neurologists will repeat the mantra “time is brain” until they have a stroke themselves, it’s really hard to say that this delay contributed to a poor outcome as these strokes are known to have poor outcomes despite maximal and timely treatment. My biggest concern here is the lack of knowledge on current headache treatment in the ED and the early diagnostic closure that ignored concerning neurological symptoms.

    • After reading the comments I am also perplexed that a patient with a posterior circulation stroke had slurred speech as this makes no anatomic sense. Assuming the ED physician didn’t skimp on the neuro exam (not clear based on this article) then a cerebellar stroke without cerebellar signs makes no sense. However, I have come to realize that I don’t think we know a lot about how the brain works so I don’t spend any more time trying to rationalize symptoms with a certain lesion and let that sway the need for a workup. If someone complains of a concerning neurological finding they get the indicated workup whether or not I think it makes sense anatomically.

  18. Thanks to everyone for responding!
    I enjoy reading the diversity of opinions and the good points raised by those commenting.
    This was a challenging case in diagnosing an uncommon problem with a high mortality.
    Steve and I had a Twitter discussion about the case and I suggested that he come here to comment so that we could present our views in more than 160 character salvos.
    One of Steve’s big issues was that morphine was used as a first line agent for headaches. During our Twitter discussion he suggested that morphine use in headaches initially showed that the doctor wasn’t on his “A game” and then opined that doing so would be “unreasonable” but not negligent.
    Steve’s comment elaborates on his concerns.
    First, regarding the Twitter comments, I have to raise a few points. The idea of being “unreasonable” but not negligent can’t happen. Negligence is, by definition, unreasonable behavior. The terms are interchangeable. Semantics are precisely why we are still pushing to change the nomenclature in medical malpractice cases to “reasonable practice.” Then we don’t have to worry about misinterpretation of terms such as “negligent” or “standard of care.”
    In addition, a physician doesn’t have to be on their “A game” in order to provide reasonable care. We can’t establish perfect or near-perfect care as the yardstick by which all medical practice is measured. Instead, we need to judge physicians by what another “average” physician would do.
    With regard to Steve’s comments above, I also have to disagree with a couple of points. Steve is strongly opposed to using opioids as first-line treatment for any headache. I’ll agree that we should try to avoid opioids as a first-line treatment, but also note that there are many instances in which we may not have other options available or in which other agents might not be the best option. For example, the patient with multiple medication “allergies” may not have another readily available treatment for headache. A patient with a post-traumatic headache who is at risk for bleeding probably won’t respond to other traditional medications and IV NSAIDs may pose a risk for bleeding – especially those patients on anticoagulants. Opioids are effective and easily reversible in such cases. External factors may also influence this issue as well. If satisfaction scores demand perfection at the risk of job loss and a patient with a headache is requesting opioids, the patient may very well get opioids as first-line treatment. That is the system some of us have let be constructed around us.
    Regardless of the intervening issues, I’m not sure that Steve’s recommendation to deliberately withholding opioids so that a patient must be admitted to get whatever medications the inpatient services deem appropriate for pain relief is a reasonable alternative. If other medications have failed and a patient’s pain would be relieved with opioids, why force an admission?
    I agree that the patient’s symptoms didn’t fit the diagnosis of an anxiety reaction. I also agree that the physicians shouldn’t be responsible for reading CT scans and that the delay probably didn’t make much of a difference in the patient’s outcome. I just think we need to avoid painting headache treatment with such broad brush strokes.

Leave A Reply