The last decade has seen a significant improvement in the understanding of spinal cerebrospinal fluid (CSF) leaks and the resulting syndrome of spontaneous intracranial hypotension (SIH). For the reader of Headache and those caring for patients with chronic refractory headaches, three novel concepts are emerging from accumulating data published here and elsewhere: (1) the majority of people with an imaging-proven CSF leak have normal opening pressure; (2) CSF–venous fistulas (CVFs) are a major, rather than rare, cause of SIH; and (3) there is a persistent headache syndrome that occurs at an alarmingly high frequency after unintended dural punctures during epidurals done for analgesia in the obstetric setting. The first article in our virtual issue, “Spontaneous Intracranial Hypotension: 10 Myths and Misperceptions” by Kranz et al. begins with an excellent review of advances up to 2018. The clarity of writing and elucidation of the ways in which SIH can be misunderstood and missed led to it being voted the Members Choice Award winner for 2018. It should not be overlooked by those wanting a primer on the basics of SIH.1 Their myth one: “SIH Is Defined by Low CSF Pressure” traces the history from early reports in which opening pressure was considered a key feature of SIH, to a modern understanding that only a minority of patients with a proven CSF leak have a low opening pressure. When Kranz et al. evaluated their own cohort of 106 symptomatic patients with SIH who underwent thorough imaging regardless of the measured opening pressure, they reported only 34% of patients with symptomatic and imaging-proven CSF leaks had low opening pressure. This has held up since and has been extended to CVFs. One recent cohort of 17 consecutive spinal CSF leaks of which 12 were CVFs reported that only 1 of 17 had low opening pressure.2 In another multi-institution cohort of 48 CVF, only 2 had low opening pressure.3 However, in retrospect, the most impressive aspect of “Spontaneous Intracranial Hypotension: 10 Myths and Misperceptions” may have been its anticipation of the growing importance of CVFs. In fact, their article shows the first image in Headache of a CVF. Discovered accidentally in 2014, the true revolution of CVFs arguably began to unfold only after the publication of “Spontaneous Intracranial Hypotension: 10 Myths and Misperceptions,” and can be traced to the first report in 2019 that the frequency of detectable CVF dramatically increases when patients are imaged in the lateral decubitus position with dense intrathecal contrast layering over nerve root sleeves.4, 5 Two papers in our virtual issue highlight different aspects of the emerging understanding of CVFs. “Headache Response after CT-Guided Fibrin Glue Occlusion of CSF-Venous Fistulas” by Mamlouk et al.6 described CVF as a major, rather than a rare, cause of SIH. Of 78 patients presenting to the authors’ institution with brain or clinical features suggesting SIH, 30 had traditional extradural fluid collections visible on spine magnetic resonance imaging (MRI), while 38 had a CVF that could only be detected with lateral decubitus myelography.6 A similar finding was described in a recent report of sequential patients with SIH in Olmsted County, Minnesota in which 7 of 14 (50%) were found to have CVF as the cause of SIH.7 Mamlouk et al. also demonstrated the utility of computed tomography (CT) imaging for the detection of CVF. Though CVFs were initially discovered using digital subtraction myelography (DSM), imaging following contrast injection—a technique usually reserved for catheter angiography—suites, Mamlouk et al. described that CVFs can be readily detected using decubitus CT imaging, a finding increasingly validated.8 Finally, this work demonstrated that CVFs can be treated effectively in some but not all patients with percutaneous fibrin injection, a finding which has been since reproduced in a multi-site study of more than 100 patients with CVF.9 “Spontaneous Spinal Cerebrospinal Fluid-Venous Fistulas in Patients with Orthostatic Headaches and Normal Conventional Brain and Spine Imaging” by Schievink et al. reports the findings when lateral decubitus DSM was offered to 60 patients with orthostatic headache in whom brain and spine MRI were normal—a group that until 2019 was thought to not have a CSF leak.10 When MRI–negative patients with orthostatic headache were evaluated for CVF, the rate of detectable CVF was 10% overall, but these were non-uniformly distributed. The CVFs were all concentrated in the 31 (52%) patients who had meningeal diverticula, a spinal imaging finding considered normal that has been reported to occur in up to 44% of healthy control patients.11 Thus, the rate of an undetected CVF in patients with a negative MRI of the brain and full spine was 20% among the subset with a meningeal diverticula. In another reported cohort of patients examining the relationship of brain MRI findings and outcome of spinal imaging for CVF, among 24 patients with no pachymeningeal enhancement on brain MRI and no epidural collection on spine MRI, 8 patients (33%) were found to have a CVF.12 These findings broadly validate the findings reported by Schievink et al. in this virtual issue. We anticipate that future work will continue to elucidate the true prevalence of detectable CVFs in patients with normal MRI of the brain and spine. In light of these papers, the pre-2019 literature on SIH needs to be critically reappraised. For example, reports that MRI of the spine is 90% sensitive for detecting CSF leaks, largely based on its correlation with non-decubitus CT myelography, should now be understood to mislead more than inform.13 Such studies evaluated the performance of MRI for detecting CSF leaks by using conventional CT myelography as a gold standard for establishing the presence or absence of epidural fluid collections, but this is an intellectual framework now clearly undermined by CVF, in which no epidural collection is present. Another important emerging trend in the field of spinal CSF leaks is the growing understanding that after an unintended dural puncture (such as a dural puncture occurring during a labor epidural), a significant number of patients will report new, or worsened, headaches that persist for years. Highlighted in this virtual issue, “Did She Have an Epidural? The Long-Term Consequences of Postdural Puncture Headache and the Role of Unintended dural Puncture” by Barad et al., the authors reviewed several studies which document that up to 30% of women experiencing an unintended dural puncture will go on to have a chronic headache syndrome.14 These data suggest that for some accidentally punctured patients, the large-bore, cutting tip Tuohy needle intended to facilitate threading of an epidural catheter outside the dura may form a uniquely devastating puncture when advanced unintentionally through the dura. While much of these data come from retrospective case–control studies, also highlighted in this virtual issue is 2019's important prospective cohort study of the same phenomenon. In “Incidence of Persistent Headache at 18 Months Following Accidental Dural Puncture in the Obstetric Population: A Prospective Service Evaluation in 45 Patients” by Gauthama et al., the authors report that a third of such patients have new or worsened headache 18 months after an unintended dural puncture.15 This finding closely mirrors the rates of persistent headache suggested by retrospective case control data, and supports their conclusion that “post-dural puncture headache from an accidental dural puncture can no longer be considered a self-limiting condition.” One might expect that the recognition and treatment of patients as having a headache from a CSF leak due to a known dural puncture would be straightforward; however, the data and publications in this virtual issue suggest otherwise. Even when patients report a new and persistent orthostatic headache that dates from the time of a recognized dural puncture, there is evidence that they are not receiving care directed toward a suspected CSF leak. Adding to the clinical confusion is the fact that as time persists, the classic orthostatic nature of the headache may dissipate.16, 17 In “Management and Outcomes of Persistent Headache after Accidental Dural Puncture in the Obstetric Population: A 9-Year Prospective Audit” by Niraj and Critchley, outcomes are described for a cohort of 54 patients with persistent headache dating from the time of a known unintended dural puncture.18 The authors report the longest-term follow-up to date for such patients with an average follow-up of 5.7 years post-puncture. Only half of those with ongoing chronic symptoms years after a known unintended dural puncture were ever treated with an epidural patch in the immediate postpartum period. Subsequently, even when it was clear that chronic symptoms had developed, only 9 of 54 were later offered an epidural blood patch. In the pain clinic where they were assessed and treated, their treatment was focused on palliative medical management and occipital nerve blocks—not escalating treatment for CSF leak. This might be partially accounted for by the fact that only 16 of 54 reported postural features. Furthermore, while 50 of 54 patients had a brain MRI, all were interpreted as normal and thus the potential diagnosis of CSF leak may have been inappropriately dismissed. Unfortunately, most patients with chronic post-puncture headache will have normal brain imaging, even when spine imaging is positive. Schievink et al. reported on 27 patients with chronic symptoms of CSF leak dating from a known dural puncture: only 3 of 27 had either pachymeningeal enhancement or any brain sagging to suggest radiographically that symptoms were related to a CSF leak.19 In contrast, nine had spine findings on imaging: five extradural collections, one had a CVF, and three had small pseudomeningoceles (blebs). Eight of these nine patients had complete resolution of symptoms with fibrin patching or surgery directed at these spinal imaging findings. Of the nine patients with symptoms starting at the time of a known puncture who also had confirmatory spine findings, only three had some brain imaging features associated with CSF leak. The high rate of negative imaging in these patients combined with the progressively less clear orthostatic nature of symptoms may erode confidence in the diagnosis. Indeed, Niraj and Critchley proposed an alternative hypothesis to ongoing CSF leak as the cause of persistent headache after an accidental dural puncture—that there is a period of sensitization that occurs in the peri-puncture period that contributes to long-term symptoms. This hypothesis is challenged by the high rate of clinical response years after a puncture when patching or surgery is used as treatment, as reported by Schievink et al., or Niraj and Critchley's own data that one third of these patients continue to report that their headache is postural. The readers are advised to attend to how these controversies unfold in the next several years, but the implications are enormous. Epidurals are ubiquitous, and accidental dural punctures are a regularly recurring experience for even the best practitioners. Why would these post-puncture leaks become persistent? One reason is highlighted by a case report in this virtual issue. “A Causative Role for Remote Dural Puncture and Resultant Arachnoid Bleb in New Daily Persistent Headache: A Case Report.”20 The images are informing and highlight the small pseudomeningoceles previously described by Schievink et al. above.19 Following a known dural puncture, the patient acquired a persistent structural alteration at the site of the puncture, a small pseudomeningocele or “bleb” predisposing to ongoing CSF leak. The subtle nature of the finding on the spine MRI is highlighted, and one look at the surgical images challenges our conception that a punctured dura will reliably heal in a way that restores its pre-puncture architecture. In cases such as the one presented in this virtual issue, it does not. These blebs, it seems, are also most visible on a thin cut (1 mm) axial or sagittal 3D acquired T2 fat-suppressed sequence and can be easily missed on the standard 2D T2 fat-suppressed sequences (e.g., STIR), which are most often acquired 3–5 mm apart. Both better imaging and perhaps a better appreciation of our existing imaging are needed. Picking up this theme, in “Dinosaur Tail Sign: A Useful Spinal MRI Finding Indicative of Cerebrospinal Fluid Leakage” by Sakurai et al., the authors describe a spinal MRI sign few headache practitioners or neuroradiologists appreciate as an imaging feature suggesting a higher likelihood of an existing CSF leak: linear fluid intensity signal outlining the sagittal perimeter of the dorsal epidural space extending one or more segments on lumbosacral sagittal fat-suppressed T2-weighted MRI.21 This undulating T2 bright perimeter of the dorsal epidural space was imagined by the authors to resemble a dinosaur tail, hence the “dinosaur tail sign.” Seen in both iatrogenic and spontaneous CSF leaks, the fluid signal may represent epidural venous distension, free uncontained fluid, or both. In their study, among patients with CSF leak, there were an average of 3.8 spinal segments, while among the controls these were both uncommon and when present, exclusively limited to a single spinal segment. The reader should appreciate that this is a very different imaging finding than the characteristic one that neuroradiology trainees are educated on: a discrete epidural fluid collection. We suspect it is neither being noted nor interpreted as pathologic when seen as would be suggested by this single study. Hence, these findings require validation. The theme here is that whether looking for a bleb or a dinosaur tail sign, the critical spinal MRI findings are more subtle than an epidural collection and require a careful and up-to-date interpreter. These issues highlight the importance of prevention. Two articles in our virtual issue point to the role of the neurologist and headache practitioner in that prevention: “The Effect of Needle Size on Cerebrospinal Fluid Collection Time and Post-Dural Puncture Headache: A Retrospective Cohort Study” and “Lower Incidence of Postdural Puncture Headache Using Whitacre Spinal Needles after Spinal Anesthesia: A Meta-Analysis.”22, 23 Both studies add to the compelling data that use of pencil-point (atraumatic) needles reduces the rate of early post-dural puncture headache (PDPH) without increasing patient discomfort. A large subsequent meta-analysis in the Lancet reported the same core findings: atraumatic needle use reduces PDPH by just over 50% without increasing the rate of failed procedures.24 Given that at least one study has found that an early PDPH is a risk factor for persistent post-puncture headaches, it stands to reason that needle choice may not only affect the rate of acute PDPH but persistent headache syndromes as well.25 Finally, the last articles highlight the still-surprising clinical heterogeneity of spinal CSF leaks. “Paradoxical Headache in a Case of Chronic Spontaneous Intracranial Hypotension and Multiple Perineural Cysts” highlights the previously reported and incompletely understood phenomena in which some patients with CSF leak report a reverse postural headache improved when upright and worse when flat.26 Chow et al. described a case in which hemifacial spasm was the primary manifestation of SIH, while the aforementioned case of the post-puncture bleb presented initially as a new daily persistent headache a time long remote from the puncture.20, 27 In each case, the ability of SIH to appear as something other than the stereotypical orthostatic headache is highlighted. In summary, SIH remains a condition that eludes detection, humbles the clinician, and remains the great deceiver, but the headache community is increasingly catching on to its tricks. The incredible advances made in the knowledge of CSF leaks over the past decade are inspiring, and we look forward to the future, where undoubtedly new discoveries await. Ian Carroll has no conflicts to declare. Andrew L. Callen is a consultant for Eli Lilly.
No takes yet. Share an insight, caveat, or question.
Carroll et al. (2024) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: