Sacral Dural Tears as a Cause of Spontaneous Intracranial Hypotension

Spontaneous intracranial hypotension (SIH) is an acquired disorder usually caused by spinal CSF leaks. The key symptom is orthostatic headache, which may be accompanied by various symptoms with coma as the most serious presentation [1, 2]. Patients who show a spinal longitudinal extradural CSF collection (SLEC) on imaging have an underlying dural tear at the ventral, lateral [3] or dorsolateral [4] aspect of the spinal thecal sac. Patients that are SLEC-negative often exhibit a CSF venous fistula. The exact site of leakage in SLEC-positive patients ranges from the lower cervical to the lumbar spine, with predominant thoracic localization [3, 5].

Schievink et al. [4] also considered complex cysts/dural ectasia as a cause of SIH in their classification system for spinal CSF leaks. While these cysts often occur at the sacral level [4, 6, 7] it is not further specified whether they are associated with an SLEC. There are single case reports in the literature in which a tear of a sacral meningeal cyst was considered the cause of a CSF leak, most of them with a history of a trauma [8‐12]. Some others reported on patients with a CSF leak in the presence of an underlying sacral dural ectasia or anterior sacral meningocele, [13, 14] that can occur, e.g., in Marfan’s syndrome; however, these reports usually did not demonstrate the sacral origin of the CSF leak.

While most dural leaks in SLEC-positive SIH patients can be accurately identified, 10% or more cases remain undetected [3, 4]. Sacral dural tears as a cause of spontaneous intracranial hypotension in SLEC-positive patients may close this diagnostic gap.

The purpose of this study was to evaluate sacral dural tears in SLEC-positive SIH patients and to highlight specific imaging and demographic findings. We also present approaches for the diagnostic work-up and treatment of patients with sacral leaks.

In this study, we demonstrate an underlying sacral dural leak as a cause of SIH in 9 out of 149 patients (6%).

All patients with a sacral leak in this cohort were women, which is notable. In the literature, we found a single case report of a man in whom a rupture of a sacral Tarlov cyst was considered the cause of SIH [12]. Further studies have yet to show whether women with sacral leaks are more than twice as likely as men to be affected, as is usually the case with SIH [18]. The average age of patients in our cohort was 38.5 years and the average BMI was 22.9, which is slightly lower than in the common SIH cohort (38.5 vs. 45.7 years, 22.9 vs. 26) [4, 5].

All patients in our cohort had distinct signs of SIH on head MRI, assessed by the established Bern SIH score [16]. In 8/9 patients the initial score was ≥ 5 points (high probability for SIH) and 1/9 had an initial score of 4 (intermediate probability for SIH). In all patients SLEC was visible on 3D T2-MRM, which has already been shown to reliably demonstrate a SLEC in SIH patients as sensitive as gadolinium-enhanced MR myelography [19] or CT myelography [20]. Once the first patient with a sacral dural leak was identified at our institution, routine MRI examinations of SIH patients were performed with consistent coverage of the entire sacrum. In our experience, the sacrum is often only partially covered on spinal MRI, as it is generally not considered important. Moreover, 2D T2-MRM is also of value, in our study they could show sacral involvement of a SLEC at a glance in 6/8 cases (Fig. 1a).

In the literature, sacral dural tears are generally not assumed as a cause of SIH [21]. In their classification system, Farb et al. [3] did not report on sacral leaks whereas Schievink et al. [4] considered complex cysts/dural ectasia as a cause of SIH (in 4% of all SIH patients). They described them most often to occur at the sacral level, but did not specify whether these have a SLEC. There are a few cases in the literature reporting on intracranial hypotension most likely due to a sacral CSF leak. Most of them reported underlying trauma, except two cases [12, 22]; however, all cases were associated with sacral meningeal cysts, which are a potential expression of dura weakness: Tarlov cysts [8‐11], dural ectasia or anterior sacral meningoceles [13, 14]. Tarlov cysts are CSF-filled sacs within the nerve root sleeves, that usually occur at the sacral level of the spine and only rarely appear with symptoms by compression of spinal nerve roots [23]. Ruptured sacral Tarlov cysts are most likely different from spinal lateral CSF leaks type 1b according to Schievink et al. [7], which is most likely a dural tear in the region of a nerve root sleeve with an arachnoid outpouching, appearing like a meningeal cyst [24]. Dural ectasia is a ballooning of the thecal sac, which most often occurs in the sacrum and can be present in several genetic disorders most commonly in Marfan’s syndrome [6]. Dural ectasia and the severe form of anterior sacral meningocele, which extends through the ventral cortex of the sacrum, may even cause SIH without a dural tear by pooling the CSF [6].

In our cohort, 4/9 patients did not show any sacral cysts, a condition previously not reported with sacral leaks. Consequently, a causative dural weakness in these patients is less likely. Moreover, 7/9 patients in our study had a sacral tear without history of trauma while 2/9 had only minor trauma or yoga lessons and there was no fracture. Therefore, we consider a traumatic origin of sacral leaks as unlikely.

Neither the exact location within the sacrum nor the precise anatomic nature of the leak could be demonstrated in each case in our study. This could be due to the complex anatomy of the sacrum and, as we observed, the often slowly filling meningeal cysts. Microsurgical exploration may help to identify the nature of the leak as well as histological analysis. It is reasonable, however, not to choose a surgical approach in the first line treatment if the site of the leak within the sacrum is not exactly identifiable. Moreover, as reported in surgery of symptomatic Tarlov cysts, complication rate may be high [25]. We decided to perform an EBP at the lower lumbar site with the patient in the reverse Trendelenburg position. On MRI follow-up, 5/9 patients showed response to EBP/CT-guided fibrin patch, whereas 2/9 patients did not respond to EBP so far (Table 1). Clinical follow-up is available in all patients, whereas 6/9 had good clinical outcome. Two patients responded insufficiently and one patient did not respond at all to EBP. In 3/9 patients, clinical follow-up is limited due to a short follow-up period. Our preliminary data suggest that more than one patch is required for treatment in most cases.

This study raises awareness of the presence of sacral dural leaks in SIH patients and may reduce the number of SLEC-positive patients in whom the site of the leak has not been previously identified. Moreover, SLEC might be overlooked in patients with a subtle sacral leak (as in Fig. 3a), while falsely assuming an underlying CSF-venous fistula.

Our study has the limitation of a retrospective evaluation. Second, the number of cases evaluated in this study is low. Third, inhomogeneous diagnostic work-up was performed (not every patient received further diagnostics such as DSM or CTM). Fourth, T2-MRM has been used to confirm the sacral origin of the leak (in cases where the SLEC was localized exclusively at the sacral level) but has the disadvantage of not being dynamic. On the other hand, it is our experience that epidural fluid is always present at the site of dural tear on spinal MRI and never exclusively distant from the rupture site. Fifth, images were reviewed unblinded and by consensus rather than independently. Sixth, imaging and clinical follow-up after EBP or CT-guided fibrin patch remain incomplete to date.

Sacral dural leaks were found to occur in 6% of SIH patients in this study and represent another cause of SIH that has been underreported. In contrast to case reports in the literature, we demonstrated that sacral dural tears can occur without meningeal cysts or trauma. Spinal heavily T2-weighted MR images should be implemented in routine SIH imaging covering the entire sacrum to avoid false negative work-up. Epidural lumbar blood patch with the patient in the reverse Trendelenburg position may be a useful therapeutic approach for these patients.