Answers

Short answers

1 The axial view of the brain (fig 4) shows bilateral subdural haematomas and venous engorgement of the superior sagittal sinus. The sagittal view (fig 5) shows caudal descent of the brainstem with protrusion of the cerebellar tonsils. The ventricles are reduced in size because of a decrease in cerebrospinal fluid (CSF). The net effect is that the brain seems to sag. In addition, the pons is flattened. A sagittal view of the lumbrosacral spine (fig 6) shows a capacious vertebral canal consistent with dural ectasia and a large meningocele at the second sacral vertebra.

View larger version (87K): [in this window] [in a new window] [PowerPoint Slide for Teaching]   Fig 4 T2 weighted magnetic resonance imaging of the brain: axial view showing bilateral subdural haematomas and venous engorgement of the superior sagittal sinus

View larger version (104K): [in this window] [in a new window] [PowerPoint Slide for Teaching]   Fig 5 T2 magnetic resonance imaging of the brain: sagittal view showing caudal descent of the brainstem with protrusion of the cerebellar tonsils. In addition, the ventricles are reduced as a result of a decrease in cerebrospinal fluid. The net effect is that the brain seems to sag. Note also that the pons is flattened

View larger version (71K): [in this window] [in a new window] [PowerPoint Slide for Teaching]   Fig 6 T2 magnetic resonance imaging of the lumbosacral spine: sagittal view showing a capacious vertebral canal consistent with dural ectasia and a large meningocele at the second sacral vertebra

2 The condition is spontaneous intracranial hypotension, which has an incidence of around half that of subarachnoid haemorrhage.¹ The estimated annual incidence of spontaneous intracranial hypotension is 5/50 000.² The classic presentation is an orthostatic headache, which disappears when the patient is recumbent, together with stereotypical findings on magnetic resonance imaging (MRI).²

3 Treatment should be directed towards stopping the CSF leak. Computed tomography myelography is the study of choice to identify spinal leakage.² Treatments include bed rest, "blind" or targeted epidural blood patching after computed tomography myelography, and surgical repair of the CSF leak.² It is thought to recur in 10% of patients, irrespective of the management strategy used.²

Long answers
1 and 2. MRI findings and diagnosis
Spontaneous intracranial hypotension has an incidence of around half that of subarachnoid haemorrhage.¹ The estimated annual incidence of spontaneous intracranial hypotension is 5/50 000 people.²

The classic presentation is an orthostatic headache, which disappears when the patient is recumbent, together with stereotypical findings on MRI.¹ These include subdural collections, meningeal engorgement, engorgement of the venous sinuses, pituitary hyperaemia, and sagging of the brain.² One series showed that 94% of cases were misdiagnosed at presentation, which led in some cases to unnecessary neurosurgical intervention.³ All patients had a positional headache, which was worse when standing and relieved when recumbent. The rate of conversion from the isolated form to spontaneous intracranial hypotension with subdural haematoma as a result of mistreatment is largely unknown because the most common misdiagnosis was migraine, a diagnosis that does not require diagnostic imaging.

Spontaneous intracranial hypotension results from leakage of CSF from the spine. It is linked with connective tissue disorders, which are seen in up to two thirds of patients with spontaneous leakage.² The same molecular defects in collagen that tend to cause connective tissue disorders are thought to compromise the integrity and durability of the dura.² The net effect is a decrease in the volume of intracranial CSF. We know of no evidence of an association between intracranial hypotension and anticoagulation with warfarin.

A fifth of patients with spontaneous intracranial hypotension display some of the manifestations of Marfan’s syndrome—such as arachnodactyly, high arch palate, and tall stature—without harbouring the fibrillin 1 (FBN1) mutation, which is pathognomonic of Marfan’s syndrome. They do, however, have defects in microfibrils and the extracellular matrix.⁴ An association exists between spontaneous intracranial hypotension and Marfan’s syndrome, although the rate of co-incidence is not known.⁵

After leakage of CSF, the brain is unsupported and sits low in the cranium. In this position the bridging veins are stretched and they can tear, which results in subdural haematomas. This is seen in 20% of cases.⁶ The Monroe-Kellie hypothesis states that the sum of the volume of intracranial CSF, cerebral tissue, and blood is constant. Hence, when the CSF is depleted the venous system becomes engorged to compensate. This causes the typical findings seen on cerebral MRI, including subdural collections, meningeal engorgement, engorgement of the venous sinuses, pituitary hyperaemia, and sagging of the brain.² Given the association of spontaneous intracranial hypotension with the formation of subdural haematomas there is an argument for withholding or reversing anticoagulant treatment in patients on warfarin who presents with spontaneous intracranial hypotension but no subdural haematoma. This argument is merely speculative at the moment because the risk, if any, of conversion to spontaneous intracranial hypotension with subdural haematoma in these patients is unknown.

Spontaneous intracranial hypotension has also been reported in rare spinal abnormalities, such as bone spurs⁷ and disc herniation,⁸ which cause spontaneous leakage of CSF. It is also possible that spontaneous intracranial hypotension could be caused by a decrease in the production of or an increase in the rate of absorption of CSF. This is merely a hypothesis and has not been seen clinically.⁹

Dural ectasia is a major diagnostic criterion for Marfan’s syndrome, but it also occurs in Ehlers-Danlos syndrome and neurofibromatosis. It was thought to be innocuous.¹⁰ However, it can cause or aggravate spontaneous intracranial hypotension. The spinal meningeal defects in patients with Marfan’s and Ehlers-Danlos syndrome predisposes them to CSF leakage via local rupture after minor trauma, which often goes unrecognised.

3. Treatment
Treatment aims to stop the leakage of CSF. Computed tomography myelography is the study of choice to identify spinal leakage.² Treatments include bed rest, "blind" or targeted epidural blood patching after computed tomography myelography, and surgical repair of CSF leakage. Controlled randomised trials on treatments and outcomes are lacking,² and no comprehensive guidelines are available.

Although cerebral herniation could theoretically occur after lumbar puncture for spontaneous intracranial hypotension, this has never been demonstrated clinically. The puncture site is very small and the pressure of the CSF is very low, so any further CSF leakage is minimal. Although 5% of patients with spontaneous intracranial hypotension report an exacerbation of symptoms after lumbar puncture, these are generally mild and an empirical blood patch is not needed.²

Conservative management with bed rest and rehydration has been found to be effective.¹¹ An immediate resolution of symptoms has been seen after the administration of an epidural blood patch—10-20 ml of autologous blood into the spinal epidural space. The site of "blind" epidural blood patch administration is at the level of the "safe zone" for lumbar puncture; namely, between vertebrae L3 and L4 or L4 and L5, which is inferior to the level of termination of the spinal cord. This relieves symptoms in around a third of patients.¹² Higher success rates are achieved with targeted blood patch application, but the level of the CSF leak must be known and the blood patch applied specifically at this level. A fibrin agent that acts as a plug can be used as an alternative. About a third of patients whose symptoms are not relieved by a blind epidural patch find relief with the targeted blood patch method.¹³ Surgical repair of the leakage site is also a possibility but is usually reserved for patients whose symptoms are refractory to epidural blood patch techniques.¹⁴

The dramatic features of spontaneous intracranial hypotension seen on MRI, such as subdural haematomas, do not require surgical intervention. Indeed, surgical evacuation of intracranial haematomas is associated with worse outcomes than treatments aimed at stopping the leakage of CSF.⁶ Patients with spontaneous intracranial hypotension and subdural haematomas who are taking anticoagulants pose a particular therapeutic challenge. The decision to continue, discontinue, or reverse anticoagulation can be difficult. Current evidence supports the transient cessation of treatment until the haematoma resolves, even in high risk patients, such as those with metallic heart valves, as in the present case.¹⁵ The international normalised ratio should be less than 1.5 before lumbar puncture is attempted to avoid the risks of epidural or subdural spinal haematoma.¹⁶

Recurrence is thought to occur in 10% of patients, irrespective of the management strategy used.²

Cite this as: BMJ 2009;338:b911