Dr. Marcus Stoodley at the 2023 Cedars-Sinai Intracranial Hypotension Conference
Dr. Marcus Stoodley, Professor and Head of Neurosurgery at Macquarie University Hospital in Australia, presented this talk on “Spinal Manifestations in SIH” at the 2023 Cedars-Sinai Intracranial Hypotension Conference on July 8, 2023. The conference was hosted by Cedars-Sinai with generous support from the Spinal CSF Leak Foundation in Kohala Coast, Hawaii.
Transcript
[00:00:12] I’m going to cover a little bit of spinal CSF physiology, if you can put up with me for that, and then go through what I think might be the physiological impact of that of CSF leak on that physiology and then some of the manifestations of CSF leak in the spine.
[00:00:29] So spinal CSF physiology is really a subset of CSF physiology overall. And of course that involves production, absorption, circulation, and flow in and out of the parenchyma. These are the things to consider. And of course, we all start with, in medical school, thinking about this model of CSF circulation.
[00:00:47] And I won’t go through that because I’m sure everyone knows what that basic model is, but I’ll just draw your attention to the fact that in most diagrams like this, the spinal CSF receives almost no attention. That CSF just goes into the spine and somehow comes back. But we now know that CSF physiology is a lot more complex than that, that CSF, yes, is produced by the choroid plexus, but there’s also a lot of fluid exchange between the parenchyma and the blood vessels all throughout the brain and the spinal cord.
[00:01:21] There’s lymphatic involvement and dural involvement. And I’ve heard this summarized as basically that CSF is produced and absorbed everywhere. And I think that’s a reasonable way to think about it when it comes to some of the more complex aspects of CSF physiology. And this is a clinical example that I think really demonstrates that.
[00:01:41] This is a patient I treated with post traumatic syringomyelia where— so there’s a very large post traumatic syrinx—at operation, there was complete disconnection between the dural sac above and below the level of injury. And the way I treated him was to reconnect those subarachnoid spaces by opening the dural and connecting them. But the CSF in the lumbar subarachnoid space looked completely normal.
[00:02:06] So it is being produced, it had to be being produced and absorbed and circulating in the lumbar spine independently. And we know that CSF, although it’s absorbed everywhere, is probably absorbed in arachnoid villi, both in the intracranial venous sinuses, but also very similar structures in the spinal nerve roots.
[00:02:29] There’s also evidence of lymphatic connection with the CSF. So this is normal physiology. So it’s probably not surprising that we get the development of CSF to venous fistulae through that mechanism, that there is an already established connection and that possibly with patients with intracranial hypertension or connective tissue disorders, that those arachnoid villi open up and you develop direct connections.
[00:02:57] And a similar thing happens less commonly with lymphatic connections. So there are cases of CSF to lymph fistula.
[00:03:08] This is just to make the point that CSF cannot be absorbed or get back into the normal circulation unless it is in the subarachnoid space. So this is a patient with post traumatic syrinx, and she had been treated with a syrinx to spinal shunt, but that treatment failed. And the reason it failed is that the distal catheter is actually in the subdural space.
[00:03:33] So what was happening here is that CSF is going from the post traumatic syrinx and then forming a new collection in the subdural space, and that has no access to the normal absorptive mechanisms of CSF, and so that’s not successful. And so all I did was reposition the catheter into the subarachnoid space, and then it resolved the syrinx.
[00:03:56] So that’s CSF in general, but what about CSF in, in the spine? And as I said, almost every diagram that you see will just end here. So what about CSF circulation in the spine? It’s incredibly complex, and a lot of this work has been done by engineers, including by Bryn Martin, showing that the spinal nerve roots, the dentate ligaments, the arachnoid membranes, have an incredibly complex influences on spinal CSF flow.
[00:04:25] So I don’t want to touch on that in any more detail, but just to point out that it is different to CSF circulation inside the head. And then what about flow in the subarachnoid space and in terms of physiological influences? We’ve known for a long time that arterial pulsations are an important determinant of CSF circulation, both in the head and the spine.
[00:04:47] And we’ve also demonstrated recently that flow in the subarachnoid space increases with an increase in heart rate. But what’s been emerging over the last 10 years or so is that respiratory physiology is actually a much greater, has a much greater influence on CSF flow in the subarachnoid space, both in the spine and intracranially.
[00:05:05] So, changes in respiration are much more important than cardiac changes. And then we come to flow in and out of the parenchyma, and by that I mean that the brain and the spinal cord, and there’s this relatively recent concept of the glymphatics where supposedly CSF enters perivascular spaces from the subarachnoid space, passes from perivascular spaces through aquaporin 4 channels, through the extracellular spaces, washing out all of the metabolites and nasty proteins into spaces around veins, and then back out into the subarachnoid space. Now there’s a lot of controversy about whether that circulation actually exists. But it does exist, at least in part, in the sense that CSF does flow into the parenchyma, into the prevascular spaces. Whether it has this unidirectional pathway is, I think, yet to be determined.
[00:06:02] We’ve shown, for example, that in some of our work, particularly on the spine, that we’ve known for a long time that CSF flow into the perivascular spaces does depend on pulsations. And, as has been demonstrated intracranially, that if you change the blood pressure, increase the blood pressure, it actually reduces the flow into the perivascular spaces in the spine.
[00:06:25] And we have demonstrated in animal models that respiration is not only the most important determinant of cSF flow in the subarachnoid space, but it is the most important driver of CSF flow into the spinal cord parenchyma. And then if CSF gets into the parenchyma, it’s got to get out. And in our studies, we’ve demonstrated that cardiac influences are the most important factors here.
[00:06:48] So that heart rate and blood pressure are the most important determinants of CSF getting back out of the spinal cord, and that respiration actually doesn’t have any influence. So respiration causes flow into the spinal cord. There’s some evidence of the glymphatic function in the brain in the sense that of this unidirectional flow that I mentioned.
[00:07:08] But in the spinal cord, I think we’ve demonstrated that it does not happen. Because, this is a experimental study that we’ve done showing that after injection of CSF tracer, which is here in green, the tracer goes out around arteries as much as it does around veins. So there’s not a simple unidirectional flow.
[00:07:29] So, I mean, CSF physiology is a lot more complex than that. I just wanted to touch on a few things to, to paint the picture that we should be thinking about these things in this, in the context of CSF leaks because it’s clearly relevant. CSF is produced and absorbed everywhere. Respiration is the main driver.
[00:07:46] Cardiac factors are important for outflow, but at least in the spinal cord, there is not this specific unidirectional flow. So what does that mean in terms of patients who suffer from spinal CSF leak? And I think particularly of the impaired cognition and even conscious state, then of course, it’s easy to blame the brain sag.
[00:08:04] And it may be that’s the main determinant, but it’s possible also that lowering the CSF pressure is actually changing the glymphatic function in the brain, if that exists. And there is a similar process as has been suggested from normal pressure hydrocephalus and for Alzheimer’s with their perturbations of that clearance of extracellular metabolites and proteins that influence cognitive function.
[00:08:26] So, I just raise the possibility that’s happening. And I tried to find evidence of changes in physiology of brain CSF parenchymal flow with lowering the CSF pressure. And I was, if someone can point me in that direction, I’d be grateful. But the only thing I could find is work that we did nearly 20 years ago, looking at, we were investigating here a spinal cord injury.
[00:08:48] And we showed that. And so here. This is CSF tracer coming into the brain around artery. So this is CSF tracer in perivascular spaces. And on the right panel, these are animals with a shunt going from the spinal subarachnoid space into the peritoneal cavity. So in a sense, creating a spinal CSF leak. And there’s not the same penetration into the brain.
[00:09:12] So it’s not entirely proof, but I think some suggestion that lowering the CSF pressure might be reducing that glymphatic function in the brain and might be a contributing factor to cognitive disturbance. There might also be some perturbations of physiology here that are relevant in terms of rebound hypertension, because if we understood better the CSF productive mechanisms and CSF absorptive mechanisms and the potential alteration of glymphatics from a spinal CSF leak, that might help us to understand why some patients develop rebound hypertension.
[00:09:49] Okay. So thank you for bearing with me on that bit. And I’ll just move now to the actual spinal manifestations. And I’ll go through a few examples. I just make this statement: I think probably that the severity of SIH symptoms depends on the location of the leak and that the more caudal the leak is, the more severe the symptoms are.
[00:10:07] And as we’ve heard that in the most, for most cases, cranial leaks do not lead to spontaneous intracranial hypotension, but I’ll just give a case where it’s not quite cranial. This is a patient who’s had a posterior fossil decompression for Chiari malformation and there’s a pseudomeningocele. And we’ve demonstrated on our dynamic imaging that there’s a leak here with CSF coming into that cavity.
[00:10:30] Now I’ve had the opportunity to operate on quite a few cases with a post-surgical pseudomoningocele for Chiari malformation. Of course, not my own operations, but and in the vast majority of cases, when you get into the pseudomoningocele, there is a seeming membrane. Now, of course, this is not epithelium because that’s not possible. These are modified fibroblasts. But there is a somewhat of a membrane and outside the membrane, the tissue is normal, but when I operated on this young man who had very severe orthostatic headaches, almost bed-bound, the tissue around the pseudomeningocele was all edematous and there was not the same membrane.
[00:11:06] And I wonder whether that might answer one of the questions you had this morning about why some patients with this iatrogenic kind of pseudomeningocele are symptomatic. And it probably depends on that, whether there is a pseudo membrane that develops in a robust fashion.
[00:11:23] So disc protrusion, spinal stenosis, and epidural venous malformations are some underlying conditions. I just thought this was an interesting case with MR imaging demonstrating that is an intradural disc protrusion. It’s barely seen here and barely seen here. And at operation, there’s an intradural, this protrusion through a dural defect.
[00:11:51] So what about spinal canal stenosis? And I was taken by this case report where the patient presented with myelopathy and and orthostatic headache. And they reported that this was spinal canal stenosis causing a CSF leak with extradural CSF at C 1-2, and supposedly confirmed with this radionuclide study, and that after decompression of the spinal canal stenosis that the CSF leak resolved. They didn’t actually operate at the C 1-2 level, so they didn’t see the leak.
[00:12:23] And of course, Dr. Schievink has given a nice series of cases showing that sign of C 1-2 CSF is often if not always a false localizing sign. In my own experience, canal stenosis in association with spinal CSF leak is not common. In this case, the defect was ventral and it was, as opposed to the normal case where there’s an osteophyte poking through the dura, it was a more broadly thinned dura.
[00:12:51] And I wonder whether that was caused by the canal stenosis and compression. And there was a defect. We’ve had a couple of cases of venous malformation, spinal venous malformation, where there’s, this is all epidural veins extending into the mediastinum. And the signal characteristics on MR can be quite deceiving.
[00:13:12] So the signal here could be interpreted as being, epidural hematoma, whereas it’s in fact all just venous malformation. And we’ve had a couple of these. And in both the cases, we’ve demonstrated abnormalities, but not any direct fistula with our CT myelography, but presumably there is a fistula.
[00:13:33] And this is the largest series that I was able to find of nine patients with presenting with intracranial hypotension. They, in this series, they treated all of their patients with either blood patch or surgery. They had similar kind of myelographic findings to our cases. We’ve found that sclerosants are effective.
[00:13:53] So on the one case I showed you this is him being embolized with, rather than with Onyx, but with Fibrovein and Bleomycin as sclerosants, and that’s been very effective. I know that other people have been using Onyx in these cases.
[00:14:08] Spinal cord hernia. This is a case we had. A little complicated, but imaging evidence of spinal cord hernia. She was originally diagnosed with Chiari malformation, so presumably had intracranial hypotension. And when we reduced the the hernia, big ventral defect, and I put a GORE-TEX patch graft around in front of the cord because it’s not really possible to directly repair that large defect.
[00:14:36] Spinal cord hernia has not been well published, and so in a in a review a few years ago, there’d only been around 250 reported cases, and I’ve only been able to find two cases where the presentation was with intracranial hypotension. The vast majority present with myelopathy and Brown-Séquard syndrome. I know that we’ve seen some other cases today with where patients have presented with intracranial hypotension. I agree with Jürgen [Beck] that it’s probable that they develop a CSF leak, the cord herniates out through the defect and prevents any further leak. In the literature, surgery has been been used for most cases and with a dural patch graft.
[00:15:14] But interestingly, widening the defect is often used as an alternative technique to stop that kind of knuckle of cord being caught by a tight dural defect, but presumably in a patient presenting with intracranial hypotension, that would not be the preferred treatment. The dural patch graft material in most cases has been DuraGen or similar collagen-based substitute. In the case that I shared with the GORE-TEX graft, she actually developed a recurrent herniation. I wonder whether the GORE-TEX is not really strong enough to be a good repair.
[00:15:46] Superficial siderosis, I’ll just mention very quickly because the following speaker is going to talk on that, can be, these may not demonstrate very well, but it can be a very minor imaging evidence of superficial sclerosis.
[00:15:58] This is a man who’d had a spinal cord ependymoma removed 20 years ago with this iatrogenic pseudomeningocele that had been dismissed as being not relevant by his previous neurosurgeons, but presented with a complete deafness, spasticity, and an ataxia with this very severe siderosis. I’ll skip over this because I’m sure that the next speaker will talk about this, but just to point out that the different types of leak have different rates of superficial siderosis, so that the type 1a is the most likely to be associated with superficial siderosis. Whereas the fistulae interestingly have a relatively low incidence of of siderosis.
[00:16:42] Brachial amyotrophy we’ve heard earlier today is always associated with the large CSF collection. This is the only case I’ve seen. And unlike the majority of cases that seem to be reported, this was unilateral. So right upper limb proximal myopathy. But with this very extensive ventral epidural collection. First reported in 2009, presents with painless weakness and atrophy. It’s presumed to be related to stretching of nerve roots. It’s probably a bit of a stretch to say it’s related to alteration of lymphatics. I think probably stretching the nerve roots is probably the more likely explanation. And surgical repair of these is the preferred treatment. It’s not common. So here’s the long term risk as reported by Dr. Schievink.
[00:17:29] Radiculopathy, I have to say, we don’t really see, it’s not something I’ve seen, is radiculopathy being a major presenting feature for patients with spinal CSF leaks. But I’m, from the literature, it is associated with large extradural collections, not positional, as opposed to the headaches, and responds to surgical treatment.
[00:17:51] So, myelopathy I presented those couple of cases of spinal canal stenosis and they present with myelopathy, but I think we can understand why that would be.
[00:18:01] It doesn’t seem to be that myelopathy as a result of extradural CSF collections causing compression of the cord is something that is seen at all. Maybe seen very rarely, but if if at all. This case report by Dr. Schievink, I think probably the myelopathy is related to the tonsillar descent and the associated cord edema pre-syrinx state.
[00:18:24] I’ll move now to Chiari malformation, which is, I think herniation of the cerebellar tonsils is relatively common in association with spinal CSF leak and usually resolves with treatment. Occasionally, though and I think, Wouter and I both agreed on this, that occasionally one is forced to do a decompression even if you’ve not been able to find the site of CSF leak or the CSF leak treatment hasn’t been successful, because sometimes these patients are so symptomatic from symptoms that are very suggestive of Chiari malformation.
[00:18:56] It can be complicated though, because sometimes there’s imaging evidence suggestive of intracranial hypertension, such as CSF around the optic nerves, tortuous optic nerves, empty sella sign, that make it difficult to know whether the patient’s maybe got an underlying problem with intracranial hypertension that has provoked a CSF leak, or whether the tonsillar descent is related to the intracranial hypotension.
[00:19:19] So, so, chicken and egg here is definitely a consideration and can make the assessment difficult. This is just to point out that this is actually interesting case where this patient presented with clinical features very suggestive of Chiari malformation, but the imaging really was that there was an arachnoid membrane here and associated with this large syrinx.
[00:19:44] So we operated on that, resected the membrane, uncomplicated, the syrinx went away, but postoperatively she’s got clinical and radiological features of intracranial hypotension. So, she’s gone somewhere else or gone away. So I can’t give you any spinal imaging, but presumably there is an associated spinal CSF leak in that case.
[00:20:07] The last thing I’ll talk about is syringomyelia. So in my experience, syringomyelia and association with spinal CSF leak is not common, but it certainly occurs. And so here’s a couple of cases. But what has always been the observation in our cases is that the syrinx is a particular type.
[00:20:25] It’s very rostral and there’s imaging evidence. On MRI, but better seen with intraoperative ultrasound, if you ever get the opportunity, but these are communicating syrinxes. So that means that the CSF enters the syrinx cavity here from the fourth ventricle. And so presumably the pathophysiology here is that the cerebellar tonsillar descent has obstructed the outflow of CSF from the fourth ventricle.
[00:20:49] So this is really quite different to the pathophysiology of Chiari associated syringomyelia, where, although we see that sometimes, and in Chiari patients, we see that when there is a membrane obstructing the outlets of the fourth ventricle; but in these cases, I think the pathophysiology is that the cerebellar tonsillar descent is obstructing the fourth ventricle outlets, and it’s a particular communicating type of syrinx, and in the reported series that I could find, they’re very similar appearances.
[00:21:18] So these are, there’s evidence of a very rostral communicating syrinx cavity in each of these cases, including in in Wouter’s published case.
[00:21:31] The largest series I could find was of six patients all presented with cough-associated headache rather than with the syrinx as the presenting clinical problem. And the syrinx was reversible with treatment of the CSF leak.
[00:21:47] I just put this slide in to point out that sometimes the CSF leak associated syrinx is iatrogenic. This is a patient I’ve really struggled with intracranial hypertension. I know many of you never struggle with these patients, but I struggle with these patients. So, I’ve put in a cisternal shunt.
[00:22:03] So this is a shunt connected to a T connector and then going into the peritoneal cavity. And she’s developed intracranial hypotension because I’m now draining too much CSF and she’s got an associated syrinx. So, sometimes that is iatrogenic and the reverse actually happens too, that sometimes treatment of the, of a syrinx—and in this reported case a syrinx to pleural shunt has resulted in intracranial hypotension from overdrainage of CSF. So personally I don’t ever use shunts outside the spine and I’d encourage you to consider that when you’re putting in syrinx shunts because I think syrinx to subarachnoid, rather than subdural, syrinx to subarachnoid shunts work much better than trying different body cavities.
[00:22:47] Thanks very much for your attention.