Vestibular Nucleus Involvement in Patients With Acute Vertigo Due to Herpes Zoster Oticus or Vestibular Neuritis

Abstract

Objectives

Herpes zoster oticus (HZO) typically provokes vestibular symptoms and is traditionally viewed as a cranial nerve equivalent of shingles, but in contrast to vestibular neuritis (VN), it is unclear whether the pathology of HZO is limited to the vestibular nerve (neuritis) or can also involve the brainstem (nucleitis).

Methods

We retrospectively compared brain MRIs of patients with HZO with those of patients with VN to study radiologic changes in the brainstem.

Results

Five of 10 patients with HZO showed signal abnormalities in the vestibular nuclei, which lie in multiple vascular territories, whereas no patients with VN exhibited such findings.

Discussion

HZO may at least in part reflect vestibular nucleitis, as opposed to a pure neuritis.

Introduction

Herpes zoster oticus (HZO) is characterized by painful auricular and external auditory canal skin rash and audiovestibular dysfunction. When accompanied by facial palsy, this is eponymously designated as Ramsay-Hunt syndrome. HZO is believed to be caused by reactivation of latent varicella zoster virus (VZV) in the geniculate ganglion, which may affect all the nerve trunks within the internal auditory canal.¹ Many neurologic complications of VZV reactivation can occur (including meningoencephalitis, neuromyelitis, cerebellitis, vasculitis, and cranial neuritis) in rare cases even without the skin rash.

The pathogenesis of HZO-related complications likely involves both infectious and inflammatory processes. The histopathologic features of HZO are intense inflammation with hemorrhagic tissue necrosis and neuritis resulting in the characteristic skin rash and acute pain.² Active replication of the virus can progress toward the brainstem inducing brainstem encephalitis.³ However, most of the cases reported are associated with viral-induced vasculitis.⁴

Among the clinical spectrum of HZO, vestibular dysfunction is one of the most disabling. A recent study showed that the vestibular nerve could be widely affected.⁵ Evidence of inflammation (neuritis) within the auditory and vestibular nerves and the labyrinth has been noted on postmortem examinations of patients with HZO⁶ and extensively reported in MRI studies.¹ Using contemporary vestibular testing, Eduardo Martin-Sanz et al.⁷ have reported that HZO is associated with more severe symptomatology and poorer recovery in comparison with the more prevalent vestibular neuritis (VN); they concluded that other mechanisms besides purely peripheral neural inflammation could be involved.

The vestibular pathway is a simple oligoneuronal circuit, in which most neurons are situated in the vestibular nucleus (Scarpa ganglion) and brainstem. Following the mechanism involved in herpes zoster–induced neuromyelitis (centripetal infection with spinal cord MRI changes), the vestibular nucleus may theoretically be also involved in patients with HZO.

With this parallelism in mind, we conducted a retrospective case-control study to investigate any radiologic changes in the vestibular nucleus in the brain MRIs of patients with vertigo due to HZO in the acute stage and compared those with brain MRIs of patients with VN.

Methods

We retrospectively searched for patients with VN and HZO diagnoses in a dizzy clinic database (from 2014 to 2022) of patients who, at their original evaluation, had provided written informed consent for the use of their anonymized data in subsequent studies. The confirmed VN cluster comprised patients reported in a separate publication.⁸

Inclusion Criteria

The diagnostic criteria for VN⁸ include (1) spontaneously occurring acute vestibular syndrome (AVS) lasting more than 24 hours within the past 15 days; (2) physical examination showing nystagmus consistent with peripheral vestibular involvement (horizontal-torsional, unidirectional, diminished, or suppressed with fixation); (3) 6-axis video head impulse testing compatible with unilateral vestibular weakness; (4) MRI performed after 72 hours but before 15 days from symptom onset showed no discernible posterior fossa lesion (1.5 T or higher); (5) audiogram showed no synchronic hearing loss ipsilateral to the side of the vestibular weakness, and the patient had no subjective complaints of hearing loss; and (6) no evidence for a more convincing alternative diagnosis.

The diagnosis of HZO with vertigo was based on (1) herpetic eruption on the auricle and external ear canal ipsilateral to vestibular loss; (2) with or without ipsilateral facial palsy and hearing loss; (3) acute vertigo and nystagmus with the same clinical characteristics than the VN group; and (4) brain MRI obtained between 48 hours and 2 weeks from the acme of symptoms (1.5 T or higher).

All MRI images (digital imaging and communications in medicine) were reviewed by a trained neuroradiologist (M.P.A.) and neurotologist (D.A.Y.). Any imaging consistent with a lesion of the vestibular nuclei complex (pontomedullary region) was secondarily analyzed with a more detailed anatomic description⁹ and brainstem atlas.^10,11

Ethics Issues: Approval and Informed Consent

In accordance with the Institutional Review Board requirements of the institution where the study took place, all patients provided written informed consent at their initial evaluation for use of their anonymized data in subsequent studies.

Results

Ten MRIs of patients with HZO and 35 patients with VN were retrospectively reviewed. We compared the MRIs of patients in the acute phase of HZO with MRIs of 35 patients with acute VN.

In all patients with HZO a contrast enhancement or thickening of the vestibulocochlear nerve in the cerebellopontine cistern or in IAC fundus was described in the first MRI. Five of 10 patients with HZO additionally displayed pontomedullary lesions. The clinical features are listed in Table. In the 5 patients with HZO, we observed fluid-attenuated inversion recovery (FLAIR) sequence signal abnormalities involving the vestibular nuclei (and in 1 case also the intrapontine segment of the vestibular nerve tract), and in 2 patients, a focal contrast enhancement was observed in the same area. The diffusion-weighted imaging (DWI) was negative in all cases (Figure 1). In one of the cases, a pre-HZO brain MRI was available (for monitoring a known meningioma), and it showed no abnormalities in the region mentioned (Figure 2). Interestingly, in all patients but one, the brainstem lesions were unrecognized or underemphasized in the initial MRI interpretation by a first radiologist. None of the patients with VN exhibited such abnormalities.

Table.

Main Features of Patients With Herpes Zoster Oticus

Figure 1. Brain MRIs in Patients With HZO.

This figure shows imaging of 5 patients, 1 on each row. In the 3-dimensional reconstructions from open-source software (“3D Slicer”), the areas of signal abnormalities are shown in green, and vascular territory of the posterior inferior cerebellar artery is shown in red; the green and red areas are not coterminous because the vestibular structures in question lie in multiple vascular territories. FLAIR = fluid-attenuated inversion recovery; HZO = herpes zoster oticus.

Figure 2. Brainstem MRI Before and After HZO in Patient 5.

The upper row (A–C) includes images from 2.5 months before HZO. The middle row (D–F) includes images taken at 12 days after peak onset of HZO symptoms. The lower row (G–I) includes images taken at 2 months after peak onset of HZO symptoms. The left column (A, D, G) contains FLAIR sequences; the middle column (B, E, H) and right column (C, F, I) contain T1 sequences. Red arrows show (D) FLAIR hyperintensity and (E, F) T1 hyperintensity in the right ventral pontomedullary region, and the white circle shows (F) T1 hyperintensity in the intracanalicular segment of the left vestibulocochlear nerve; those changes were absent at (A–C) 2.5 months before HZO and at (G–I) 2 months after HZO. FLAIR = fluid-attenuated inversion recovery; HZO = herpes zoster oticus.

Discussion

In this study, 5 of 10 patients with HZO showed a focal lesion topographically located on the ipsilesional vestibular nuclei complex, whereas no patients in the VN group had such findings.

The clinical-radiological focality of these findings and the integrity of the surrounding structures (as also reported in the case of Sun et al.⁴) are more plausibly explained as vestibular nucleitis involving retrograde viral infection and inflammation¹² rather than infarction. Because the vestibular nuclei lie in multiple vascular territories,¹³ we are inclined to agree with the conclusions of Mizock et al.³ regarding the pathophysiology, which they suggest to be “adjacent segmental brainstem encephalitis.” The negative results in diffusion (DWI) sequences further support this idea. Other investigators have described similar brainstem radiologic findings in HZO¹⁴ which can occur at some delay after symptom onset.¹⁵ This pattern is distinct from reports of such radiographic abnormalities being restricted to the vestibular/vestibulocochlear nerve.¹

Although this is a small study, the findings have several clinical implications. First, it provides a rationale for a future prospective study of whether brainstem MRI can help distinguish VN and HZO. Second, if brainstem MRI of a patient with acute vestibular syndrome has T2/FLAIR signal abnormality (without DWI evidence of infarction) in the vestibular nuclei or prenuclear fascicle of the vestibular nerve (even without aural vesicular rash), HZO has to be considered, and treatment with corticosteroids is indicated (whereas in VN, corticosteroids have no proven role).

This is a retrospective case-control study whose findings have implications for the pathophysiology of acute vestibular syndrome (HZO may be a vestibular nucleitis; VN is not). This is not intended as a prospective study of the sensitivity/specificity of imaging findings in discriminating HZO from VN.

The limitations of this study include the small number of patients and the absence of serial MRI imaging in most cases to monitor evolution of the lesion or the time window of its occurrence. Although we used MRI of 1.5 T or higher, we cannot exclude possible heterogeneity of resolution between different MRI equipment and/or protocols.

Appendix. Authors

Study Funding

No targeted funding reported.

Disclosure

The authors report no relevant disclosures. Go to Neurology.org/N for full disclosures.