Abstract

The syndrome of CSF hypovolaemia has been a recognized cause of orthostatic headache (a headache in the upright position relieved by recumbancy) (1–3). Among other clinical manifestations, a change in hearing has also been described in this syndrome (1, 2, 4, 5). However, the pathogenesis of hearing changes in CSF hypovolaemia is unclear.
We report a patient with CSF hypovolaemia who had orthostatic hypacusis (hearing impairment in the upright position relieved by recumbancy) that was clearly demonstrated by pure tone audiometry. The anatomical and physiological relationship between CSF space and inner ear structure may explain the pathogenesis of orthostatic hypacusis in CSF hypovolaemia.
Case report
A 34-year-old previously healthy man presented with severe headache. The headache consistently occurred within 1 min of assuming the upright posture, and was relieved within 2 min of lying flat. The headache was pressing in nature, and was associated with posterior neck pain. Two days after the onset of his symptoms, he developed a change in hearing that was provoked by standing or sitting, and relieved by lying down. The patient's hearing change was described as ‘muffled’ or ‘distant’. This orthostatic change in hearing had gradually worsened, and was severe enough to impede his daily life.
Neurological examination 5 days after the onset of headache was unremarkable. He continued to develop the orthostatic change in hearing 5–10 min after he assumed an upright position. The orthostatic hypacusis of both ears was clearly demonstrated with pure tone audiometry. When he assumed an upright position for 10 min, the average hearing intensity thresholds at lower frequencies (250, 500 and 1000 Hz) were 26.7 dBHL in the right ear and 23.3 dBHL in the left (Fig. 1A). Upon lying flat, average hearing intensity thresholds normalized: 10.0 dBHL in the right ear and 11.7 dBHL in the left (Fig. 1B). The hearings at higher frequencies were stable. The summating potential to action potential ratio (SP/AP) of electrocochleography (ECoG), performed only in the supine position, was 0.36 in the left ear and 0.25 in the right ear.

The pure tone audiometries of both ears demonstrate hearing loss affecting lower frequencies when the patient assumed the upright position for 10 min (A) and significant improvement of hearing loss when the patient lay flat (B).
On CSF examination, the opening pressure was unmeasurably low and other CSF composition was normal. A brain T1-weighted gadolinium-enhanced MRI showed diffuse pachymeningeal enhancement but no evidence of brain sagging or subdural fluid collection. A spine MRI revealed extradural fluid collection at the level of thoracic spine, and engorgement of the epidural venous plexus at the level of cervical spine. MR myelography showed multiple hyperintense lesions along the bilateral paradural space at the level of upper thoracic spine (Fig. 2A). Radioisotope cisternography demonstrated multiple paradural uptakes of tracer at the level of the thoracic spine. The orthostatic headache and orthostatic hypacusis were simultaneously resolved after treatment of four epidural blood patches. MR myelography, performed 1 month after treatment, showed the resolution of multiple hyperintense lesions along the bilateral paradural space at the level of upper thoracic spine (Fig. 2B). The patient was asymptomatic during the follow-up period of 4 months, and a repeat pure tone audiometry was normal regardless of changes in the patient's body position.

MR myelography shows multiple hyperintense lesions along the bilateral paradural space at the level of upper thoracic spine (A). MR myelography, performed 1 month after four epidural blood patches, shows the multiple hyperintense lesions completely resolved (B).
Discussion
A wide variety of symptoms in addition to headache has been reported in patients with CSF hypovolaemia. These include nausea, vomiting, photophobia, dizziness, posterior neck pain, interscapular pain, radicular upper limb symptoms, visual blurring, diplopia and a change in hearing (2, 4). These associated symptoms are often orthostatic in nature and may be more worrisome than the orthostatic headache. Other, less frequent additional clinical manifestations of CSF hypovolaemia include galactorrhoea, parkinsonism and ataxia, frontotemporal dementia, encephalopathy, stupor and coma (6–11). These unusual clinical presentations may sometimes delay the correct diagnosis and treatment for CSF hypovolaemia.
The patient in this report had orthostatic hypacusis in addition to orthostatic headache. This was clearly demonstrated by pure tone audiometry. The change in hearing of the patient was initially described as ‘muffled’ or ‘distant’ in nature. As the disease progressed, he developed the hearing difficulty whenever he assumed the upright posture. The hearing difficulty of the patient affected the lower frequencies. The fact that endolymphatic hydrops may initially induce dysfunction of the hair cells at the apex that would resonate with low frequencies may explain the characteristics of hearing difficulty of our patient (12). As the range of hearing in normal individuals is usually from 500 to 2000 Hz (13), our patient had considerable hearing difficulty severe enough to impede his daily life. The ECoG refers to the measurement of neuroelectric events generated by cochlear structures and the auditory nerve in response to acoustic stimulation. Our patient revealed elevated SP/AP ratio, which may be related to the hydropic condition.
The unique orthostatic hypacusis of our patient had disappeared after the treatment of CSF hypovolaemia with four epidural blood patches. Multiple levels of abnormal findings on MR myelography and radioisotope cisternography may explain why our patient needed to receive as many as four blood patches because the EBP given in patients with multiple CSF leaks may be less effective than given in those with a single CSF leak (2, 14).
The pathogenesis of the orthostatic hypacusis and audiological findings in our patient are currently unclear, but we attempt to propose three hypotheses. First, they may be caused by stretching of the cochlear nerve as a consequence of descent of the brain. However, our patient showed only hearing change without vertigo or dizziness, despite the anatomical proximity between cochlear and vestibular nerves. Additionally, our patient showed no other clinical manifestations that may be related to the traction of remaining cranial nerves or cervical nerve roots. Therefore, it is less likely that only bilateral cochlear nerves were stretched when he assumed an upright position.
Second, assuming the patient has a patent cochlear aqueduct (15, 16), those findings may be explained by the secondary changes of perilymph caused by a reduction in the CSF volume. In an individual with a patent cochlear aqueduct, among three chambers of the cochlear, the scala tympani is connected with the subarachnoid space of the posterior fossa by the cochlear aqueduct (17). The cochlear aqueduct is filled with loose connective tissue that connects the scala tympani with the subarachnoid space and is believed to be involved in controlling the pressure of the perilymphatic fluid. The pressures in the three chambers of the cochlear have been shown to change parallel to those recorded in the CSF. Under normal conditions, the pressure in the CSF, perilymph and endolymph are equal (18). In CSF hypovolaemia, however, CSF loss may secondarily induce the reduction of perilymph with a compensatory expansion of the endolymphatic hydrops. This disturbs the relationship between the hair cells and the basement membrane, leading to a dampened response to auditory input.
Third, if the patient has an obstructed cochlear aqueduct (15, 16), the endolymphatic sac located in the subarchnoid space may be inflated because of a reduction in the CSF volume. This results in a relatively reduced amount of endolymph in the scala media. Subsequently, the hair cells may be inactivated, and the summating potential of the cochlear may increase, resulting in decreased hearing capacity.
