I am currently in the second year of a 3-year Neuroradiology Fellowship, specialising in Diagnostic and Interventional Neuroradiology, the latter involving both complex spinal and vascular intervention. With regards to my spinal work, I perform lumbar punctures, nerve root injections, spinal biopsies/drainages and myelography under both CT and fluoroscopic guidance.
Having observed how debilitating a cerebrospinal fluid (CSF) leak can prove to be, I started learning and reading more about the condition. I am now being trained on how to safely and successfully perform fibrin glue patches, with the aim of treating the CSF leak. The thirst to further my knowledge led me to apply for the CSF leak bursary; it was a privilege to be able to attend the third Intracranial Hypotension Symposium in Los Angeles and learn from pioneers within the field. This article shall provide a comprehensive review of my ‘Symposium Journey’ highlighting the key pillars of the CSF leak pathway, including: the clinical presentation, diagnostic pathway and treatment options.
CSF bathes the brain and spinal cord, and is incidentally 99% water, a fact new to me! A spinal CSF leak is the underlying cause of spontaneous intracranial hypotension (SIH), occurring in 5 per 100,000, typically in middle age (30-50 years) and with a predilection for females (2:1, female:male). Despite being only half as common as the well-known subarachnoid haemorrhage, it is often under- and misdiagnosed, meaning that patients can suffer unnecessarily for many years, with something both treatable and curable in many cases.
The fundamental issue relating to the lack of awareness and misdiagnosis of this debilitating condition is what drives me to further my passion to manage and treat it.
The symposium commenced with a poignant lecture by one of the leaders in this field, Dr. W Shievink. He aptly illustrated his first case, which was on the 24th November 1991: surgical treatment of spontaneous intracranial hypotension associated with a spinal arachnoid diverticulum. It is fair to say that we have come a long way since then, demonstrated by the fact that Cedars Sinai Medical Centre have since evaluated 2270 patients for a CSF leak.
Patients commonly present with an ‘orthostatic headache,’ worse minutes to hours after being upright, and alleviated upon lying flat. There are also several other reported clinical symptoms, including: neck pain, nausea/vomiting, light sensitivity, fatigue, cognitive decline and movement disorders. Interestingly, as Dr. Deborah Friedman (University of Texas Southwestern Medical Centre) discussed in detail, in her talk ‘Every Cranial Nerve,’ the eighth cranial nerve (Vestibulocochlear nerve) is the most commonly affected, resulting in symptoms including vertigo, imbalance and hearing loss. Dr. Jeremy Cutsforth-Gregory (Mayo Clinic) also described an uncommon but recognised manifestation of SIH called Frontotemporal dementia-like behavioural syndrome, characterised by behavioural disinhibition, loss of empathy and hypersomnolence.
I feel that it is imperative for medical professionals, especially in the community and emergency settings, to be aware of the myriad symptoms that can be encountered with this condition, such that it can be appropriately recognised and the ‘diagnostic pathway’ initiated.
The Diagnostic Pathway:
Dr. Richard Farb (Toronto Western Hospital) illustrated the diagnostic pathway using the SIH MRI Protocol, whereby patients initially have an MRI brain and spine examination.The key brain radiological findings include: subdural fluid collections, engorgement of the venous sinuses, pachymeningeal enhancement, pituitary hyperaemia, and ‘sagging’ of the brain. Of note, there is radiological overlap with Chiari malformation, a condition in which ‘sagging’ of the brain is also seen.
As both conditions have a very different management pathway, it is important to make the correct initial diagnosis. Professor Cutsforth- Gregory (Rochester Mayo Clinic) interestingly described findings in an unpublished study, demonstrating skull thickening or cranial hyperostosis on the MRI examination for SIH patients, also known as the ‘layer cake skull.’ Bony growth can occur to compensate for a decreased volume of CSF. The study showed 13% of patients to have this sign, and therefore this is definitely something worth monitoring in the future!
Spinal imaging is performed to evaluate for a spinal longitudinal extradural collection (SLEC), after which the patient is either SLEC positive or negative. At this stage, it is important to identify the specific type of leak, of which there are four:
- Type 1 leaks consist of a dural tear;
- Type 2 leaks represent a meningeal diverticulum;
- Type 3 leaks consist of an abnormal connection between CSF and the venous system, known as a CSF-venous fistula;
- Type 4 leaks are indeterminate.
A Digital Subtraction Myelogram (DSM) is performed for further evaluation, and is a dynamic form of imaging, where an intrathecal contrast injection is performed, and imaging is acquired before and after the injection. Three main types of leak can be localised: 1) rapid leaks visible on MR and CT myelography as a SLEC; 2) ventral leaks and 3) CSF-venous fistulae. Dr. Richard Farb found that in 43 patients following the SIH MRI Protocol and DSM, the leak was located in 86% of cases. If the patient was SLEC positive, he advised that the DSM should performed with the patient on their front; it was most likely to be a Type 1 or 2 leak with a 96% chance of successful identification.
However, he advised an initial lateral decubitus position with a SLEC negative leak, which was likely to be a Type 3 or 4 leak, with an identification chance of 67%. Accurate leak detection is probably the most important part of the CSF journey because if the leak site is not found the treatment pathway is much more complex.
Obtaining images of suitable diagnostic quality is essential to be able to answer the different clinical questions posed by this condition. As a Neuroradiologist, the talk delivered by Dr. Peter Kranz (Duke University Medical Centre) titled ‘Imaging Tips and Tricks’ was extremely beneficial.
He opened with the crucial and often understated fact: ‘some imaging is not the same as good imaging.’ It is very well-publicised that imaging of poor diagnostic quality often raises more questions than it is able to answer.
Epidural leaks are easy to see, but hard to localise. Localising a leak enables more targeted treatment, and thus hopefully better outcomes. The ‘shoulder problem’ is one interventionalists are unfortunately accustomed with as they obscure the area of concern, which makes leak detection extremely difficult. Therefore, for suspected ventral leaks above the level of T3 (thoracic spine), an ultrafast CT is advised to avoid the shoulders obscuring the lower cervical spine on the lateral projection. For ventral leaks below T3 and lateral leaks, a DSM is suitable.
For successful CSF-venous fistulae detection, experience is key, and one must know what to look for. One should have technically strong imaging with thin images (0.6mm slices), concentrated contrast, rapid image acquisition and the lateral decubitus position. As Dr. Richard Farb from Toronto also demonstrated, the prone position identified 19% of CSF-venous fistulae, whereas the lateral decubitus position identified 74%. This is further exemplified by the fact that prior to April 2018, 50% of patients with SIH had negative spinal imaging, but post-April 2019 just 25% of patients had negative imaging.
Finally, the Valsalva technique is extremely useful, performed by forcefully blowing out, with the nose pinched and mouth closed. This manoeuvre effectively impedes venous return due to a very positive intra-thoracic pressure. This allows greater visualisation of a CSF-venous fistula, one that may have been hidden during normal respiration.
This was one of the most useful sections of the symposium for me because common practical issues that I had seen or encountered were tackled with sensible tricks, tips and solutions! This will help me to perform these procedures more effectively, with greater success and thus improved patient care.
One of the most fascinating aspects of the conference was learning about advances in treatment, because this is fundamentally what I thrive upon.
The initial treatment is conservative, consisting of bedrest, analgesics, increased caffeine and hydration. Failing this, the mainstay of treatment is the injection of autologous blood into the epidural space, so called epidural blood patch (EBP). 10-20 ml of blood is used and relief of symptoms can be instantaneous in a third of patients, which means that it also serves a diagnostic purpose. If the initial blood patch is unsuccessful, a further treatment can be performed with a larger blood volume (20-100ml). Dr. Joseph Gemmette (University of Michigan Hospital) delivered a brilliant lecture about a novel technique, which I was not familiar with, in which a catheter can be used to access the dorsal epidural space, allowing an epidural blood patch or fibrin glue to be delivered at multiple levels. This mitigates patients having to have multiple EBPs, a procedure that is known to be painful.
If there is no symptomatic relief after the blood patch, the next option is a fibrin glue patch, which requires the exact site of the CSF leak to be known. Dr. Schievink reports that a third of patients for whom an EBP was unsuccessful will have relief following the fibrin glue patch. Surgery is the final option and is reserved for patients who have had unsuccessful treatments. Providing a structural abnormality or specific CSF leak location is identified, this method is both safe and successful.
I currently use catheters for vascular intervention. It would be useful for me to formally observe and learn how to perform multi-level epidural blood patches, with the view of implementing this at my centre in the future.
The symposium provided me with an incredible opportunity to meet and learn from leaders and pioneers within this field, for which I am extremely grateful to the CSF Leak Association. The overwhelming feeling during the symposium was that of a sincere passion and desire to enhance both knowledge and expertise in this field, for the betterment of patients suffering from this condition. The main challenge lies in treating patients who have recurrent leaks (approximately 10%), normal MRI findings, diffuse multi-level leaks, and persistent symptoms despite resolution of the leak.
The future is exciting and developments include: better contrast agents to detect slower or intermittent leaks; better gluing agents that are able to plug leaks without compromising nerve roots or the spinal cord; and better percutaneous devices, such as the curved needle, to more easily access and treat ventral leaks.
Amar A Chotai