Slides
Transcript
Thank you so much. Thank you, Dr. Friedman, that was excellent. I’ll tell you a secret – one of the real reasons that I helped organize this was just to get all these people that I look up to so much in the same room so I could learn from them. I feel incredibly lucky that everyone’s here today, so thank you.
Here are my disclosures – none particularly relevant to the content of this talk. I want to start with a case of somebody that we saw recently. This is a 38-year-old woman who, in November of 2023, suddenly developed back pain and a metallic taste in her mouth. Then, a couple of months later, she woke up while lying flat in her bed with a holocephalic headache. She had pressure behind her eyes and in her suboccipital region behind her head, and she noticed that it was worse when she bent over to tie her shoes and if she coughed or sneezed. She had nausea, she felt like she had to pop her ears all the time, and there was ringing in her ears. It was a non-pulsatile, constant, high-pitched tinnitus in both ears, and she had stiffness in her neck.
She went to her primary care physician, who prescribed her antibiotics, they said this must be a sinus infection; it didn’t make her feel better. She went to neurology, and they ordered a brain MRI. Here’s her brain MRI, and we’ll go over some of the features of this study in particular in detail soon, but the most important part, I’ll argue, is that the report said this is a “normal MRI of the brain.” Because it was a normal MRI of the brain, the neurologist said, “Well, maybe this is idiopathic intracranial hypertension.” They performed a lumbar puncture; her opening pressure was within the normal range. They said, “Okay, maybe it’s not IIH,” and they started Topamax and rizatriptan for migraines. The Topamax made her feel horrible – it made her feel depressed and lethargic – didn’t help with her symptoms at all. The rizatriptan one time kind of helped a little bit with her head pain. They switched to Ajovy; this did nothing for her symptoms, and then she was referred to us and we saw her in clinic in July of 2024.
Now, I want to take a second and look at this MRI of the brain very carefully. So, the sort of hallmark imaging feature of a leak is absent here – there’s no dural thickening or pachymeningeal enhancement (the fancy term for this). There’s no venous engorgement per se. Looking at our transverse venous sinus here, this margin is not particularly convex upward. There is some subtle brain sag here, now it’s not as dramatic as some cases I’m going to show you later. And there’s two ways you could look at this. We could apply the Bern criteria and measure the suprasellar, mamillopontine, and prepontine intervals that I’ve shown in red. These are all very mildly narrowed – nothing profound that you would look at and say, “This is dramatic.” There’s also another finding here, which is that the third ventricular floor is very slightly down sloping, okay, whereas normally we see this as upsloping or flat. It’s very slightly down sloping – a little bit unusual. And she indeed had a CSF-venous fistula with this brain MRI.
I want to take a second here to talk about what is a “normal” MRI of the brain in somebody with intracranial hypotension. Is it somebody who has a Bern score less than two, and that makes them have low probability? Is it somebody who has a Bern score of zero? This Bern score we hear about – the probability of localizing a leak on myelography – zero points, that means it’s “normal.” I argue that a “normal” MRI of the brain is one that the radiologist said is “normal,” because once that is in the patient’s chart, it stays with them for the rest of their medical adventures, right? Just like a pathology report, a lab test, this is now ingrained in part of their medical history. And while we have some of the best radiologists in the world here, who may look at this brain MRI and say, “Well that’s not normal, I see some subtle brain sag,” I would argue that the vast majority of radiologists across the – this country and the world would look at this brain MRI and call it “normal.”
This is where I think the Bern score can be useful. You know, for somebody less familiar with intracranial hypotension and the sort of subtle manifestations therein, just they can apply this number, perhaps would result in less delayed treatment for some patients, and it’s a way that we addressed some of our findings in a study that I’m going to talk about here today.
I want to talk about one more case example. This is a 61-year-old woman who was referred to us by a memory and aging clinic because of incidental “findings” on her brain MRI. She came with her brother to clinic. I always encourage patients to bring a family member or loved one with them to clinic for this exact reason. She said, “I don’t know why I’m here – there’s nothing wrong with me. I have no symptoms at all.” Her brother said, “That’s not true. She has had rapid personality change and memory loss recently.” I went through a review of systems with her. She did have mild left ear hearing loss and ringing in her ear, but she said she had absolutely no headache, she had no neck pain, she had no nausea, she had no dizziness or disequilibrium.
Here’s her brain MRI. She has some subtle dural thickening and profound brain sag. Her posterior fossa structures are very much downward displaced, and she also had a CSF-venous fistula. When we think about the clinical presentation and manifestations of intracranial hypotension, we think about an orthostatic headache. We’ve heard about it in the last couple of lectures. What does that mean? I think most of us understand it to mean that people are worse when they’re upright and they’re better when they’re flat, right?
In this big meta-analysis that frequently gets referenced, over 90% of people they have an orthostatic headache. This is the classic way that this disease presents. So, let’s look at the history of how we’ve come to understand what intracranial hypotension is to better understand why we think of it this way.
In the late 19th century, orthostatic headaches were noticed after lumbar punctures, where doctors were intentionally putting a needle in the spine and withdrawing fluid for diagnostic purposes. People were having headaches when they stood up after that procedure. In the early 20th century, Dr. Schaltenbrand described a syndrome of orthostatic headaches in people who did not have a lumbar puncture. He called it “hypoliquorrhea.”
As time went on, additional symptoms were sort of brought together and ascribed to this condition: ringing in the ears, hearing loss, changes in vision, cognitive symptoms. They were sort of all lumped into this low CSF pressure syndrome. But the first edition of the International Classification of Headache Disorders in 1988 didn’t mention anything about intracranial hypotension. Dr. Schievink really led the way in describing the pathophysiology of spontaneous lateral and ventral dural tears and their associated clinical syndromes, in the later 1990s in particular. And then, in 2004, the ICHD-2 did describe a headache due to low CSF pressure, which is “diffuse and dull, it worsens within 15 minutes upright” – very specific – and “improves within 30 minutes flat.” Wow, I don’t know where they got that from, what data or study, but then, in 2014, Dr. Schievink’s group also described the CSF-venous fistula, this incredible new source of a leak that we previously did not know existed.
Four years later, in the final edition of the ICHD-3, we have the “headache attributed to low CSF pressure,” and the first two words in the description are that it’s an orthostatic headache. And it can be accompanied by neck pain, tinnitus, changes in hearing, photophobia, and nausea. It also talks about the fact that low CSF pressure is one of the major criteria for its diagnosis, even though we know now that most patients with a leak – and particularly those with the CSF-venous fistula – do have a normal opening pressure. And it really calls into question what else they could potentially be getting wrong here or putting too much emphasis on?
I don’t know if you’ve seen this, but there’s this sort of codicil at the end of the diagnostic criteria, which I think has been recently expanded. When I checked on this again, there’s additional language here where they get even more specific about the features of this orthostatic headache. “There is a clear postural component in most cases of intracranial hypotension. It may occur immediately or within seconds of assuming an upright position and it may resolve quickly within one minute.” And you have to elicit the – this nature of the orthostatic headache when you’re talking to a patient and getting their history. They even talk about that if you have a postural headache after sexual intercourse, this is probably a CSF leak. So much focus on this orthostatic headache.
Now, we know that we’ve gained most of our understanding of the clinical features of leaks based on type 1 and 2 leaks. This means tears in the dura – a hole in that bag of fluid that’s resulting in leakage. There’re these ventral defects, this sort of type 1 leaks, a lateral defect, the type 2 leak, and then, of course, the CSF-venous fistula – this one that we’ve only come more recently to understand. Now in our experience, this is the most frequent leak type that we’re finding. It’s actually not rare – it’s over half of our total cases, and I’m sure it’s probably the same for most centers across the world, if not at least a plurality of them.
And so, we decided that we wanted to focus just on these types of leaks and see what is the clinical syndrome that presents in this leak type specifically. We took 48 of our patients who we found definite CSF-venous fistulas on consecutively at our institution. We looked at every single thing they reported to us in clinic – what did they say was going on with them? We also looked at their brain MRIs and analyzed them for each component of the Bern score, not for severity of findings, but just whether or not it met criteria for that particular category of finding.
What did we find? So, 48 patients with a CSF-venous fistula – 37% of them said, “I don’t feel better when I lay down.” Nearly 13% of them said, “I feel worse when I lay down. I can’t lay down flat.” 8% of them said, “I don’t have a headache at all,” and nearly 13% of them had no relief when they were flat and they had at least one migrainous symptom – phonophobia, photophobia, or nausea. You could imagine these patients, in particular, are particularly susceptible to misdiagnosis with migraine.
What did they present with very frequently? Dizziness or a feeling of disequilibrium or being off-balance, which we’ve heard about already today. Worsening with Valsalva or cough, ringing in the ears – almost invariably described as constant high-pitched ringing in both ears, sometimes one more than the other. Neck pain – very common. Nausea – incredibly common as well. And then one that I found particularly interesting which is worsening with position changes, going from lying to standing or sitting to standing, not necessarily being upright.
This really made me question myself when I talk to patients. And you know, is it really appropriate for us to say, “Do you feel better when you lay down?” and if the answer is no, well then, okay good, we’ve crossed off a CSF leak. Remember, 17% – nearly a fifth – of our patients with a CSF-venous fistula felt worse with a position transition, but not necessarily when after they were upright from the transition. And 17% of them described worsening with head movement. A common thing patients tell me is, “I’m in the car, and I’m scared to turn my head over to look over my shoulder to change lanes,” or “Just looking back and forth – this is very bothersome for me and worsens my symptoms.”
So, I think we need to move away from this simple question of “Are you better when you’re flat?” to a targeted review of systems, getting at the specific symptom profile that we’re coming to understand in our patients. We did what’s called a cluster analysis – which is a fancy term that a statistician helped do (shoutout Dr. Han at Stanford) – where, basically, we’re looking at which symptoms tend to present together frequently compared to how frequently they tend to not present together, right.
And what we found is that there seemed to be a classic symptom cluster. In just under half of our patients, they were experiencing what I think most people would ascribe to a CSF leak headache – they had a pressure or throbbing-type headache located in the occipital region, in the back of their head. They felt better when they were flat and they had frequent comorbid neck pain. There was also sort of this atypical symptom cluster – people who had more of a sharp or stabbing headache quality. Their headache was usually located more frontally than the other cluster. They felt less relief when they were flat. They had less neck pain as well.
And then we had some patients who had no headache at all. What did they present with? Half of them presented with memory issues, and then about 13% each presented with fainting or feeling like they were going to faint, difficulty concentrating (many patients refer to as brain fog), hearing loss, pain that shot down their arms or legs (we call radicular pain), or numbness or tingling (a paresthesia), tinnitus (ringing in their ears), or neck pain alone with no radiation into the head, no headache head pain at all.
What did we find when we looked at those imaging findings? So, that classic mnemonic “SEEPS” that kind of gets at the imaging findings in intracranial hypotension – our dural enhancement, our brain sag, subdural collections, venous engorgement, pituitary engorgement. What we found is that these findings didn’t tend to all present in an all-or-none type fashion, but rather they also tended to cluster together as well. In particular, the dural enhancement, when present, tended to co-occur with venous engorgement. And the features that are encoded in the Bern score of brain sag tended to also co-occur, but they didn’t necessarily happen together all the time. Sometimes one group would be present, sometimes the other group would be present, and occasionally they would present together – and again, this is any degree of severity.
I’m showing two examples of patients here with extreme examples. Our first patient here has severe brain sag – downward displacement of their posterior fossa structures – but they have no dural enhancement at all or venous engorgement. Our second patient has no brain sag. Their posterior fossa structures are normally positioned, but they have diffuse dural enhancement and venous engorgement. And both of these patients had CSF-venous fistulas.
Our cluster analysis showed this in a similar fashion: some patients had more brain sag but less venous and dural engorgement; others had more venous and dural engorgement but less brain sag; and some had both sets of findings together. When we looked at these together, the clinical and imaging findings in concert, we found something sort of interesting was that – the patients who had that classic symptom cluster (the pressure-type headache in the back of their head that felt better when they were flat, that neck pain associated with it) tended to occur more with brain sag compared with the atypical symptoms, which tended to more occur with our venous and dural engorgement.
So, we understand that there are these two pathophysiologic mechanisms that occur with CSF loss that result in changes in the brain – that loss of fluid that leads to that loss of buoyancy and brain sag, and then that venous engorgement, both of the dura and of the dural venous sinuses. From a Monro-Kellie doctrine, as that CSF leaves, the blood-filled structures engorge to take up that space. And that perhaps these are occurring in parallel but not necessarily together, and that perhaps the brain sag is somewhat linked to this classic symptom profile which we’ve come to understand.
So, in conclusion, relief when flat – this thing we’ve all come to understand as a classic hallmark symptom of a leak – may be less common in a CSF-venous fistula, which, at least in our practice, is the majority of leak types we’re finding. Whereas dizziness, tinnitus, nausea, cough, headache, neck pain – these are very, very common. The absence of a headache, the absence of a positional component, or the absence of a migrainous symptom should not preclude a leak workup. And the brain MRI findings of a leak may be incredibly subtle. When you have a radiology report that says “normal,” it deserves a second look by somebody who’s looking at these specific findings. A patient can lack dural thickening, they can have very, very subtle brain sag, and have a leak present. And these imaging and clinical findings may cluster together, and it’s our job to figure out what this means for our patients moving forward.
Thank you very much.