Dilated cervical epidural veins and extra arachnoid fluid collection in orthostatic headaches

Patients and methods.

Cervical MRI were performed in eight patients during the acute phase of severe OH and after headache remission. Severe OH was defined according to the classification criteria of the International Headache Society as a headache intensity >80 on a visual analogue scale (0 = no symptoms, 100 = most severe symptoms) and an onset of pain within 60 seconds after sitting up. The patients’ informed consent was obtained prior to the MRI examination. Imaging was performed on a 1.5 T Magnetom Vision (Siemens, Germany) using a regular neck array coil (T1-weighted spin echo and T2-weighted fast spin echo sequence with sagittal orientation of the cervical spine and the craniocervical junction; T1-weighted spin echo and T2-weighted fast spin echo sequence in axial orientation of the upper cervical spine and the cervical–spinal junction).

Results.

The data of all patients are summarized in the table. All patients had a prominent convex-shaped dilated cervical anterior, internal vertebral venous plexus (CAIVVP) with flow voids, indicating significant flow (figure 1). In the follow-up MRI the no longer prominent CAIVVP had a straight shape in all patients (figure 2). Flow voids were retrograde. We also found an extra arachnoid fluid collection in seven patients during the acute headache phase. A downward shift of the brain with reversible pseudo–Arnold-Chiari malformation was seen in one patient, who was free of pain during the MRI examination.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. T2-weighted MRI (sagittal plane) showing the dilated venous plexus (arrows) of Patient 7 during the acute phase.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. T2-weighted MRI (sagittal plane) of Patient 7 showing normal filling of the venous plexus after headache relief.

Discussion.

All patients with severe OH had significant differences in the filling and flow signals in the CAIVVP, which had a dilated convex shape and flow voids, indicating significant blood flow. Extra arachnoid fluid collection could be demonstrated in seven patients. A pseudo–Arnold-Chiari malformation indicating a downward shift of the brain was observed in only one patient. We did not perform MRI examinations in a control group because follow-up MRI showed resolution of the abnormalities in all patients after headache relief. However, four patients who still had abnormal MRI findings after follow-up did not undergo an additional examination.

The dilatation of the CAIVVP in our patients was more prominent in the cranio-cervical junction than in the lower cervical levels. We gave no contrast medium to optimize imaging of the CAIVVP and performed no additional MRI of the thoracal and lumbar spine. Because similar findings with distended spinal epidural veins and dural thickening and enhancement also have been reported in the mid and lower thoracal and the lumbar levels of single patients with OH, our findings are probably not limited to the cervical spine.2,3⇓

The MRI findings seem to be the spinal correlate of the cranial MRI findings: subdural effusions; intense, diffuse enhancement of the supra- and infratentorial intracranial dura; dilatation of cortical and epidural veins; abnormal dural venous sinus enhancement; and prominence of the pituitary gland.4-6⇓⇓ All MRI findings can be explained best by the Monroe–Kellie hypothesis which states that the sum of the volume of the brain plus CSF plus intracranial blood remains constant. Therefore, each CSF-volume depletion has to be compensated. Because the intracranial cavity and the spinal cord canal are closed systems, the Monroe–Kellie hypothesis must be valid for the spinal chord canal, too. One possible compensatory mechanism for the CSF depletion could be an increase of the venous blood volume with dilatation of the cortical and epidural veins and abnormal dural venous sinus enhancement. Another compensatory step could involve a volume shift into the extravascular compartment.

A shift of fluid into the vascular and the extravascular compartments is per se a physiologic process triggered by a CSF leakage to keep intracranial and spinal volumes constant. In accordance with cranial and spinal MRI findings, it has been shown by the use of color doppler imaging that the mean diameter of the superior ophthalmic vein was substantially larger in patients with intracranial hypotension.7 The meningeal enhancement observed in OH has been interpreted as a marked dilatation of the meningeal veins.6,8⇓ Because we did not explicitly examine the cranium, we cannot comment on whether cranial and cervical MRI findings in OH develop synchronously.

There is increasing agreement that intracranial hypotension is not the key mechanism of OH. Patients with typical OH caused by CSF leakage but with normal CSF pressure have been described,5 and a systematic investigation of the CSF opening pressure data of an initial lumbar puncture and of a second puncture 24 hours later reported no correlation of CSF pressure and OH headaches.9 Therefore modern pathophysiologic concepts suppose that a loss of CSF volume and traction of the pain-sensitive structures are the key features in OH.6

Although all MRI findings in OH might be epiphenomena, one can speculate that—besides the traction of the pain-sensitive structures—dilatation of cerebral and spinal veins might also contribute to the pain in OH. MRI examinations can never reflect the filling of veins in a patient in an erect position. This would explain why even the patients with the most prominent venous dilatation had no or only mild headache in a supine position and why patients with a milder intensity of OH show normal venous filling while in a supine position.

There are arguments to assume that both traction of pain-sensitive structures and venous dilatation work as the pathogenesis of OH. The venous dilatation theory might explain the effectiveness of vasoconstrictive drugs such as caffeine and theophylline in the treatment of OH and the observation that jugular venous compression and Valsalva’s maneuver worsen the pain in OH even when the patient is supine.10 During a lumbar puncture there is an immediate volume loss and a lowering of the intracranial pressure. Therefore the latency to onset of OH cannot be explained by volume loss and traction alone in those individuals who do not develop an OH immediately after a lumbar puncture. Increasing dilatation of veins and a volume shift into the extravascular compartment might better reflect the natural time course of OH in most patients.

The dilatation of the cervical epidural vein plexus might contribute to the neck stiffness (n = 3) and neck/occipital headaches (n = 5) of our patients, because the postcapillary venules and venous sinuses are innervated by trigeminal and upper cervical sensory nerve fibers. However, we did not study whether the clinical cervical findings corresponded with the degree of dilatation of the CAIVVP.