Slides
Transcript
Hello and welcome back. We have heard amazing things about spontaneous CSF leaks and seen some amazing images. And we’re moving on now to something that’s sort of near and dear to my heart – and that is the leaks that we cause as doctors. I think there’s a lot of work that needs to be done here. I want to say that one of my mentors and friends, when I was finishing training, told me that the most important year of your training is your first year as an attending. A lot of what I’m going to show you today are things that I didn’t learn when I was in residency and I learned taking care of patients.
But I first want to talk about the scope of the problem when it comes to iatrogenic leaks. What I mean by iatrogenic – or doctor-caused leaks – are, one, you know, both intended lumbar punctures, those that we’re very familiar with where we’re putting a needle through the dura on purpose, and then unintended lumbar punctures. By that, we can break that down further into epidural needles where we’re trying to be in the epidural space, but we’re not, and then spine surgery.
So, let’s get a sense of the scope of this problem and put it in perspective in terms of what we’re dealing with. So, lumbar punctures – estimates in the United States – there’s around 400,000 of these that are done per year, and around 10 to 30% of those result in a post dural puncture headache, meaning that we have around 40,000 to 120,000 new cases of post dural puncture headache per year. We know that this risk is a modifiable one – that if we use a different type of needle, an atraumatic non-cutting needle type, we can reduce that risk by about 60%, which is huge. Unfortunately, the standard trays that most lumbar punctures are done with are done with a cutting needle type, or provided with a cutting needle, which is very unfortunate and something we as a group have to work on advocating for.
What about epidural steroid injections? Over 10 million in the United States are done per year. And recognized dural punctures of around 3% during an ESI, or epidural steroid injection. That means that, at a minimum, in the United States, we’re talking about 270,000 epidural steroid injection leaks per year.
What about labor epidurals? So, around 3 million women in the United States receive a labor epidural a year. There are estimates, in terms of a known dural puncture, that around 1% of them have a dural puncture during that epidural, and the vast majority of those will have a post dural puncture headache. 30% of those women who have an unintended dural puncture after a labor epidural will go on to have a new chronic headache syndrome, meaning that every year from labor epidurals alone we have somewhere in the range of 20,000 new mothers with a post dural puncture headache and around 8,400 new chronic post dural puncture headaches per year.
Now, it’s actually pretty tough to get good data about post-surgical leaks, but there is good data on lumbar spine surgical leaks. And there’s around 500,000 of these performed in the United States per year. 9% of them have a recognized – and I keep saying that – I’m going to show you a case of why that’s important – intraoperative leak, meaning that at a minimum we have around, in the United States, 45,000 surgical leaks per year.
So, let’s put those all together. That means between lumbar punctures and post dural puncture headache, epidurals with unintended dural punctures, and spine surgery with unintended dural violation, we have somewhere in the range of 400,000 new cases of post dural puncture headache or iatrogenic leaks per year.
Now, it’s hard to sort of conceptualize that number or put it in a framework, but compare it to diseases that we’re more familiar with. That is multiples more than Parkinson’s disease or colorectal cancer incidence in the United States, and about half of that of heart attacks per year in the United States. So, this is a humongous problem. Maybe not as, you know, “sexy” or “cool” as these spontaneous leaks that we’re finding – but I would argue as big of a problem, if not a bigger one, that we need to deal with.
I think there are two major barriers to care in post dural puncture headache and iatrogenic leaks. Number one is their underestimated rate of occurrence. As doctors, we don’t like to think that we cause problems and that the things that we do are harming our patients. And so – like Dr. Kranz said – we just don’t see them, right? We don’t see the problems that we’re causing. And then, underrecognized imaging findings. That’s something I want to touch on today, and I think things that neuroradiologists, and radiologists in general, can pay more attention to.
Let’s talk about underestimated. So, this patient had spine surgery. Two days after her spine surgery, she called her nurse and said, “I have a new headache. I have a new postural headache immediately after my spine surgery.” And the nurse said, “Well, there were no dural tears during your surgery, so it’s not due to a CSF leak, and we don’t know what it is. But if it gets worse, go to the ER. Probably, you just need pain control.”
Well, she did go to the ER, and she had subdural hemorrhages on her head CT. And what’s very interesting is that, even though MRIs are a more sensitive test for looking at features of intracranial hypo or hypertension, we could see that not only does she have subdural hemorrhages, but her pituitary gland is more engorged than it was eight years ago – suggesting that this is not a case of elevated intracranial pressure and subdural hemorrhage that’s post-traumatic or otherwise. But, in fact, it is likely due to a leak. We said that in the report. They took her back to surgery, and indeed, yeah – she had multiple dural tears requiring repair.
Now, there’s something else I want you to pay attention to on this image that was obtained when she presented back to the hospital. There is fluid signal along her laminectomy bed, but she does not have a longitudinal epidural fluid collection. She does not have a fluid collection up and down her spine. She just simply has fluid in the laminectomy bed. And I think, as radiologists, most of the time we would interpret that as not a big deal or normal.
And I want to use that as a jumping off point to talk about this next case. This is a person who had spine surgery 5 years ago and has had a new orthostatic headache since. Has a brain MRI that does not show features of intracranial hypotension and has this spine MRI. Again, we have our laminectomy down here, and I don’t know about the other radiologists in this room, but when I was in training, I was just told this is normal. You see fluid here in the laminectomy bed – that’s just normal – either a seroma that’s there or granulation tissue, vascularity, what have you. It’s no big deal.
So we looked at this, and in the context of this patient’s new headache after spine surgery, we said, you know, is this maybe a leak? There’s no longitudinal collection, but there’s fluid there. When you scrutinize the sort of dural contour along the posterior thecal sac at the site of decompression, you might wonder – is there a slight sort of discontinuity here or rarefaction of the signal, something going on there? And indeed, there was a leak there – right there. They’re leaking into that granulation and scar tissue, not resulting in a longitudinal epidural fluid collection, but nonetheless, that fluid signal was a clue.
This is not something that I think the majority of radiologists are looking at and calling abnormal, but in the appropriate clinical context of a new headache following a procedure, any fluid signal in the epidural space should be suspicious.
Here’s another really important lesson that I learned – not as a resident, but as an attending. This finding of the dura being distorted after a laminectomy or spinal decompression, I think, is something we see all the time and think is just normal. Well, there’s no bone in the back, and the dura has now moved backwards and is sort of pouching out. This patient also had headaches following her spine decompression, and I looked at this contour and I said, “This just looks weird to me, okay?” The cauda equina nerve roots are maintaining their normal configuration in the sort of native thecal contour. And maybe that’s fine, maybe there’s just sort of arachnoid here, and this is indeed the dura itself. But I didn’t like the sort of acute angles along this margin with the native dural sac that were there, and I just said, “This looks wrong.” And indeed, there was a humongous dural tear here – okay, along the dura. And this was a contained fluid collection, a leak in this patient.
Now, I think probably the radiologists in this room and who are listening online are jogging their memory through every time they’ve seen a decompression where the dura looked something like this. And I think we see it pretty frequently. You have to wonder – how many of those are just the dura that is now distorted, and how many of them are actually a leak?
What if the fluid is there – this fluid that I’ve told you to pay attention to – but it doesn’t opacify on a myelogram? This patient had a spine surgery, and intraoperatively, there was a leak. They knew there was a leak. And they repaired it or attempted to, and there was this large fluid collection in the paraspinal soft tissues on their lumbar spine that I’m pointing to here. Because of this large fluid collection, she had a myelogram done via a cervical puncture. For those of you unfamiliar with that, it means the needle is going in your neck, and the dye is going down towards the bottom of the spine. They didn’t want to traverse the needle through that pseudomeningocele fluid collection in the spine, and that didn’t fill. Okay, there was no dye in that, so we said, “Oh, must not be a leak. It must be a seroma.” But you can see here that the blue arrow is pointing to something else, and that patient has a lumbar drain that was placed to help decompress and heal any potential leak that was associated with that surgery.
And so, we wondered, perhaps there was a pressure gradient differential here that was preventing the dye from getting from the intrathecal space to the pseudomeningocele. And we said, “What if we just inject the pseudomeningocele itself?” which we did here, and you can see that there’s readily apparent communication of the dye into the subarachnoid space intrathecally when we injected this way. And there’s our lumbar drain again. And indeed, even on the axial imaging here on the bottom image, you can almost appreciate the site of the leak – this little irregularity here along the ventral aspect of that pseudomeningocele.
Intraoperatively, you could see their prior repair, and at the edge of that prior repair, the dura was splitting, and there was continuing leaking here. So, moving away from post-surgical leaks to more post dural puncture leaks, you know, we think of these as being imaging-negative. And we know now that the brain MRI is most frequently normal or negative in cases of post dural puncture headache. I really want to focus on the spine. In the acute setting, like with this patient who was imaged immediately following a lumbar puncture, who had a post dural puncture headache. They could be very abnormal – we can have a big epidural fluid collection, and we can even have this so-called dinosaur tail sign, which we’ll talk about here in a little bit. And after patching, this patient did very well, and those signs went away along with their symptoms.
What is this dinosaur tail sign? So, this was first described in a paper in 2017. The authors envisioned that this fluid signal outlining the periphery of the dorsal epidural space sort of mimicked that of a Stegosaurus tail. I don’t love the name, but nonetheless, that’s what it’s called, and so that’s what we say. In their paper, they looked at patients with both proven intracranial hypotension or who had a post-puncture leak on imaging. They had a myelogram, and they saw extravasation of contrast or what have you, and the majority of those patients had this sign. Some patients had the sign and didn’t have a leak, but compared to their control group, only 2 of the 35 in their control group who did not have a spontaneous or post dural puncture leak had the sign, and it was only confined to one inner space. Okay, so never more than one inner space. There wasn’t that undulating contour – it was just one inner space there.
And I’m just showing an example that, if you pay attention after that, you know when you’re looking at your cases of leak, you’ll see this actually in your spontaneous leaks too. For example, here in this patient with a large epidural collection tapering off at the edge of it, we have our little dinosaur tail sign – that fluid signal insinuating along the epidural fat. This is important because, in the chronic setting, it may be the only thing we see. This patient was imaged one day postpartum after a labor epidural, has our classic imaging findings of a leak – an epidural fluid collection here with the blue arrows, and then also our dinosaur tail sign here with the orange arrows. After patching, she got substantially better but was not totally back to normal. And then 5 months postpartum, all that’s left is that dinosaur tail sign, right? And this could be, this could be really the only thing that’s there. And this spine MRI was read as normal, and I don’t blame the radiologist for doing so, but pay attention to this in the appropriate clinical context, and it can go away. This patient had a remote history of a lumbar puncture and had this sign on her spine MRI, and after patching, that sign is gone, okay.
So, this, we have to do more work and more studies to understand the utility of this, but I think that there’s something there. Dr. Lennnarson showed this case earlier, but I just want to touch on it here very briefly, because it’s very important to remember that a patient who has a headache after a lumbar puncture, who is not getting better with patching – it may very well be that the needle was not just advanced into the subarachnoid space but was inadvertently advanced too far ventral into the ventral epidural space through the ventral dura. That’s exactly what happened in this patient, and the reason that I wanted to show this case is just because it so beautifully shows the extravasation of contrast into the ventral epidural space at L3 – L4, and it also so beautifully shows the intraoperative photo of how tiny and small that hole is. This patient had a longitudinal epidural fluid collection from a hole that was just a couple of millimeters big.
Finally, sometimes we’ll have that dinosaur tail sign. Sometimes we’ll have an epidural fluid collection, and other times we’ll see this focal irregular contour of the dura or a focal fluid signal. This is something I think we’re going to get into later in today’s talks, but it’s this potential bleb, right? And what we’re talking about in this patient, who had a headache after a lumbar drain, is this finding here – this ovoid fluid signal finding along the dorsal thecal sac. This patient has both of these findings. And what we think that these are – are that they are blebs or pseudomeningoceles. That is a herniation of an arachnoid through a dural defect, kind of pooching out. You can almost imagine in this case, perhaps there’s slow leakage around that bleb, resulting in the adjacent dinosaur tail sign – just a nice, sort of chronological example of how this can develop.
In this patient who had a lumbar drain and has their myelogram here, the lumbar drain is removed, and then on their MRI, we see this very focal outpouching of the dorsal dura here, indicated by the red arrow. Then on the myelogram, it just very nicely shows that this is sort of this exophytic little piece here sticking out. These are often intraoperatively noted to be covered by very vascularized membranes, and you wonder how many of these we’re potentially missing with our, you know, luminal imaging on an MRI. If it’s not sticking out into the epidural space and it’s matted down by one of these vascularized membranes, perhaps it’s there and we just can’t see it – right? For a radiologist in the room, it’s kind of like a CTA versus vessel wall imaging – perhaps we’re missing it because there’s no contour irregularity, and you wonder what exactly is happening here that’s resulting in the clinical syndrome for these patients.
Sometimes it’s noted that there’s leaking from the base of these, where the arachnoid is pooching out and there’s sort of leakage around the neck, but sometimes there’s not as much leakage throughout the neck. What is noticed is that there’s very vascularized membranes covering the bleb. This has led me to wonder if perhaps there’s something that we’re missing here in terms of a CSF venous fistula that we have not yet been able to visualize. Dr. Madhavan from Mayo Clinic actually published this series, noting with traumatic spinal pseudomeningoceles they actually sometimes saw CSF venous fistulas along the course of these, and you wonder with these very vascularized membranes if that could be the case here too.
Finally, I think it’s important to mention recent work that was published by Dr. Schievink’s group, where they looked at patients who were referred to them for a suspected iatrogenic leak. These are patients who had symptom onset following a spine procedure. They were epidural steroid injections, spine surgeries, epidural anesthesia, or a lumbar puncture. These patients had spontaneous leaks, not iatrogenic ones, all removed from the site of their instrumentation. Some of them had ventral tears, some had lateral tears, some had CSF venous fistulas. Why could this be? Well, maybe it was the way they were positioned during surgery. Maybe it was that their pressures were elevated from their general endotracheal anesthesia. Perhaps it’s anchoring bias on the side of the referring clinician. The referring clinician says, my patient just had a spine instrumentation and they have this syndrome. It has to be the spine instrumentation, when in reality they were right but kind of for the wrong reasons – right?
So, in summary – do not under-recognize the rate of puncture. Even if no leak is seen during the procedure – if a patient has new symptoms following a procedure that could result in a leak, consider a leak. Don’t forget that the brain MRI is not going to be positive in the chronic setting in these patients in the majority of cases, and do not use that as a guiding principle for whether or not to pursue treatment. Look very, very carefully at the spine – any contour irregularity of the dura, any fluid signal in the epidural space, including that dinosaur tail sign that we have to come up with a new name for. Finally, consider possible spontaneous leaks too, particularly if your directed treatments at that site are not giving them relief – think outside the box. Perhaps there’s another leak here that is not where the instrumentation was. Thank you very much.