Pitfalls in Management of Intracranial Hypotension

Case Presentation
LK, age mid-40s, had a history of hypothyroidism which was adequately managed with hormone replacement therapy, with thyroid hormone levels normalized. Sixteen years before the current presentation, LK had experienced severe head trauma resulting in right frontotemporal subdural and extradural hematoma, for which decompression craniotomy and cranioplasty were performed. Twelve years after the head trauma, LK developed excessive daytime sleepiness, sleep-disordered breathing, and low mood. Continuous positive airway pressure support was prescribed, which resulted in transient improvement in excessive daytime sleepiness. However, after 2 years, LK experienced gradually progressive difficulty in walking along with imbalance, slowness in activities, difficulty in recall which improved with cuing, dysphagia, nasal regurgitation, and drowsiness. Orthostatic holocranial headache with intermittent diplopia was present. Intracranial hypotension was suspected.

Diagnostic Process
Evaluation revealed a relatively low opening cerebrospinal fluid (CSF) pressure (80 mm of water), lower cranial nerve palsies, and ataxia. Contrast-enhanced MRI of the brain and spine demonstrated brainstem sagging with significantly decreased mamillopontine distance (3 mm), tonsillar herniation, and mild pachymeningeal enhancement of cerebral convexities (Figure 1, A and C), without any spinal longitudinal epidural collection (Figure 1B). To rule out any spinal longitudinal epidural collection and negative spontaneous intracranial hypotension (SIH), a dynamic prone CT myelogram (CTM) was performed but failed to demonstrate a CSF leak or delayed contrast excretion in the renal pelvis. Therefore, dynamic digital subtraction myelography (DSM) in the right followed by left lateral decubitus positions was also performed, but the results were unremarkable (Figure 1, D and E). 

Figure 1. Sagittal 3-dimensional fluid-attenuated inversion recovery MRI scan shows evidence of brainstem sagging with tonsillar herniation (white arrow) and a reduced mammillopontine distance (broken white arrow), consistent with spontaneous intracranial hypotension (A). Spinal MRI screening shows no evidence of spinal longitudinal extradural cerebrospinal fluid (CSF) collection (B). Axial contrast-enhanced T1 MRI scan shows postcranioplasty changes with diffuse dural thickening and enhancement (C). Lateral decubitus digital subtraction myelography of the left (D) and right (E) sides shows no obvious CSF leak or CSF–venous fistula. Axial CT scan of the head shows a cement cranioplasty graft with improper opposition (white asterisks) (F).

After thorough multidisciplinary discussion, an epidural blood patch (EBP) was performed with 16 mL of autologous blood slowly injected into the L3-4 epidural space.¹ After the EBP, there was substantial clinical improvement in previously existing symptoms including difficulty walking, imbalance, dysphagia, and drowsiness, but LK’s status returned to baseline status preceding the EBP procedure within 10 to 15 days. A second DSM performed 3 months later failed to demonstrate a leak. A second EBP was attempted at the L1-2 level with an 18-mL autologous blood patch, and CT imaging showed that the injected blood had spread to 3 spinal levels. After transient improvement which lasted 7 days, relapse occurred. A third EBP was administered over 2 sites: 14 mL in the thoracic (D9-10) space with spread of 2 levels above, and 17 mL in the lumbar (L3-4) space with spread of 3 levels above. However, only mild improvement in headache severity (visual analog scale), drowsiness, and speech was noted, which lasted less than 12 hours. 

Case Resolution
After EBP, refractoriness was demonstrated, all imaging was again reviewed, and the possibility of inadequacy of the cranioplasty flap was considered. It was hypothesized that inadequate covering of calvaria by a defective cranioplasty flap may lead to inward atmospheric force, causing brain sagging syndrome (BSS) and secondary intracranial hypotension.² Contrast CT of the head showed dural thickening and enhancement at the cranioplasty graft site with improper opposition (Figure 1F). On re-exploration, the implant was found to be fractured into 3 pieces (Figure 2, A to C). Revision and reconstruction of the right frontoparietal cranioplasty using bone cement was performed and fixed with 4 plates. 

Figure 2. Cinematic volume-rendered image of the skull shows a porous cement graft with fracture (white arrow) (A). Intraoperative photographs show the broken previous implant (dotted white lines) (B, C). Sagittal 3-dimensional fluid-attenuated inversion recovery MRI scan before surgery (D) and a comparative MRI scan at 3 months after surgery (E) show substantial resolution of tonsillar herniation and improved mammillopontine distance.

LK started to improve immediately and by 15 days after surgery, was able to perform all routine activities. By 3 months after surgery, LK had resumed professional activities, which continue to be maintained to this date. Follow-up CT showed good approximation of bone flap, and MRI showed improvement in signs of intracranial hypotension (Figure 2, D and E). LK has regained independence with no difficulties in speech, swallowing, or gait.

Discussion
Intracranial hypotension due to low CSF volume or pressure is classified as either primary (ie, spontaneous) or secondary.^3,4 CSF leaks contributing to intracranial hypotension may occur at either the spinal or cranial levels,4 which may vary from simple dural tears to multilevel complex meningeal diverticula, classified as type 1 (dural tear), type 2 (lateral leak), type 3 (CSF–venous fistula), or type 4 (distal nerve root sleeve) leaks.⁵ Secondary intracranial hypotension may be iatrogenic, with possible causes including lumbar puncture, surgery, overshunting, decompressive craniotomy, or cranioplasty. Clinical features of intracranial hypotension include orthostatic headache, loss of consciousness, neck pain, visual symptoms, fluctuating drowsiness, cranial nerve palsies, gait difficulty, and rarely neurocognitive or extrapyramidal changes along with the BSS.²

Once intracranial hypotension is suspected, a sequential diagnostic approach should be used to avoid treatment delays. Documentation of low CSF pressures and systemic causes followed by neuroimaging is indicated. A contrast-enhanced CT of the brain should be performed to evaluate skull bone integrity or a cranioplasty defect. MRI of the brain and spine should be performed to assess characteristic findings of SIH. If SIH is suspected, CTM or DSM should be performed.^4,5 CTM and DSM are the diagnostic procedures of choice to demonstrate the location and extent of a CSF leak. EBP, either targeted or nontargeted, is the treatment of choice pending direct repair of a defect if surgery has been performed in the past.^2,4 The detection rate of a CSF leak ranges from 70% to 90%. When a targeted EBP is performed, the success rate ranges from 50% to 100%. Nontargeted EBP has a success rate of 61% to 76%.^2,6

This case elucidates challenges in the diagnosis and management of intracranial hypotension. Development of insidiously progressive intracranial hypotension 12 years after trauma and cranioplasty is unusual; most cases occur within days to months after injury. Although BSS was demonstrated, the cause was undetectable despite detailed myelographic imaging.⁵ EBP is an accepted treatment modality with best response if targeted at discrete CSF leaks.⁷ Nontargeted EBP was attempted on 3 occasions in LK with substantial near-complete improvement obtained only after the initial EBP. Subsequent EBP failures indicated a need for a more diligent search for the pathogenesis of the intracranial hypotension.⁸

We suspected brain sagging resulting from intracranial hypotension possibly caused by inadequacy of LK’s cranioplasty implant. Demonstration of a broken implant and revision cranioplasty leading to near-complete resolution of symptoms served as a proof of concept of the hypothesis. A recent report by Schievink et al⁹ described similar cases of intracranial hypotension management. Out of 61 participants with BSS, 8 had a history of head trauma and cranioplasty. These participants developed symptoms of frontotemporal dementia due to BSS 1.5 to 13.5 years after their original surgery. When no cause of a CSF leak could be found in the skull or spine, customized polyetheretherketone cranioplasty was performed in 4 participants. Loose craniotomy bone plates or titanium mesh was found in all 4. These findings are remarkably similar to our observations. The late worsening related to lack of complete opposition of implant to skull with subsequent repair further supports the hypothesis for causation of intracranial hypotension and BSS.

In patients with intracranial hypotension who have a history of trauma and cranioplasty, a negative search for the site of CSF leakage should prompt screening for fracture of the cranioplasty implant, which would lead to early detection of a leak and cranioplasty repair, which may prevent complications of BSS.