Pressure-induce nystagmus (PIN), also known as Hennebert's sign (HS), represents a peculiar finding indicating a third window mechanism within the inner ear. It can be elicited either by pressure changes exerted on the external auditory canal (EAC) with tragal compressions or a Politzer bulb, or increasing intracranial/middle-ear pressure through Valsalva maneuvers. It can be found in several labyrinthine disorders with normal otoscopic findings including otosyphilis (1), perilymphatic fistula (2, 3), Meniere's disease (4, 5), vestibular atelectasis (6, 7), hypermobile stapes footplate (8), semicircular canal (SC) dehiscence (9, 10), and other dehiscences between the otic capsule and surrounding structures (11, 12). Similarly, chronic inflammatory pathologies involving the middle ear may lead to labyrinthine fistula (LF) due to otic capsule erosions, accounting for pressure transmission from tympanic cavity into the inner ear (13–19). In the latter case, other classical symptoms and signs of third window syndromes including bone-conducted hyperacusis, sound-induced vertigo (Tullio phenomenon), abnormally enhanced amplitudes and reduced threshold for vestibular evoked myogenic potentials (VEMPs) may be partially or totally hidden by underlying middle ear pathologies.

The pathogenetic mechanism underlying HS is still controversial. In particular, it is unclear whether endolymphatic or perilymphatic flows are involved and which subgroup of hair cells among ampullary and/or otolith receptors represents the target sensor (4–7, 20–28).

Unlike caloric irrigations measuring horizontal SC (HSC) activity in low-acceleration ranges, the video head impulse test (vHIT) represents a recently introduced test assessing the vestibulo-ocular reflex (VOR) of each SC for high accelerations (29). Its diagnostic accuracy is based on Ewalds's laws, stating that stimulation of each canal produces eye rotations around an axis parallel to that of the canal, that ampullopetal endolymphatic flows represent excitatory stimuli for HSC while inhibitory for vertical SCs, and that stronger oculomotor responses are derived from excitatory inputs (30).

Here, we describe three patients with LF due to different temporal bone pathologies presenting with HS and positional nystagmus despite global SCs hypofunction on vHIT. We discuss the possible pathomechanisms underlying PIN and positional nystagmus and we also review the relevant literature.

Oval and round windows represent the only sites with reduced impedance in the inner ear, whereas the remaining membranous labyrinth is entirely encased within the otic capsule. Other anatomical openings connecting the inner ear fluid spaces to the surrounding structures, such as the cochlear and vestibular aqueducts, are functionally closed to sound and pressure flows as they usually offer high impedance (11). Additional bony openings such as a labyrinthine dehiscence can result in increased inner ear compliance leading to abnormal pressure transmission into the vestibular system from surrounding compartments (2). LF represents an interruption of the otic capsule connecting perilymphatic spaces with the middle ear. It may occur in <15% of patients affected by COM with cholesteatoma (13–19) or may represent a late complication in subjects being previously submitted to CWD mastoidectomy (32, 33). Affected sites can be identified in <60% of temporal bone CT scans of patients with intraoperative findings of LF (18, 19, 34) while MRI may provide additional information about inner ear involvement (35). Whereas HSC represents the most commonly affected site, both cholesteatoma matrix and inflammatory tissue may sometimes erode other structures as PSC, superior SC (SSC), vestibule, and cochlea (13–18). As LF may also represent a pathway for toxins and pathogens invasion from the middle ear to the membranous labyrinth accounting for serous or suppurative labyrinthitis, symptoms can include variable combinations of vertigo and HL depending on LF location, on the size of bony defects, and on the extent of inner ear damage (13–19, 36). PIN represents a peculiar finding in LF, although it may show a low diagnostic sensitivity. In fact, it might be underestimated due to a possible complete functional loss of inner ear sensors/afferents or the mass-induced canal plug exerted by concurrent middle ear pathologies might prevent pressure change transmissions to the endolymphatic spaces (16, 18, 19, 28, 37). Its pathomechanism is mainly attributed to the stimulation of SC ampulla, according to the site of erosion (14, 20–23, 26–28). On the other hand, HS has been also ascribed to a possible otolith activation in other inner ear disorders, including Meniere's disease, perilymphatic fistula and vestibular atelectasis. In fact, fibrous adhesions between the stapedial footplate and the saccular membranous labyrinth (vestibulofibrosis) either due to a collapse of the membranous labyrinth or due to a saccular distension on a hydropic basis have been hypothesized (4–7, 24, 25). Nevertheless, clinical observations with VEMPs testing in a patient with endolymphatic hydrops (27) and experimental studies on LF in animal models (20) reported the onset of nystagmus after pressure changes despite saccular areflexia and after removal of the otolith membranes, respectively.

All cases herein described presented at our attention with functional loss for all SCs on vHIT besides different degrees of sensorineural HL consistent with global cochleo-vestibular damage. This condition likely represented the result of either previous or current labyrinthitis due to LF, accounting for contralesional nystagmus after mastoid vibrations detected in the second patient and for paretic spontaneous nystagmus enhanced by head shakings in the third case. Conversely, slightly reduced VOR-gains for contralesional PSCs (functionally coupled with affected SSCs) may likely result from the severe functional impairment of injured SSCs, in accordance with studies on contralesional function following vestibular neuritis and vestibular deafferentation, where an involvement of both central compensation processes and peripheral impairment of the “push-pull” mechanism have been hypothesized (38–40). On the other hand, PSC was the only hypoactive canal in the contralateral ear in all cases and the expected corrective saccades after head impulses were lacking in most cases, raising the possibility that our findings might be due to artifacts. Nevertheless, most healthy PSCs exhibited highly reduced VOR-gain values compared to ipsilateral SCs also in larger cohorts with unilateral vestibular loss (41). Moreover, the functional impairment for the PSC of the unaffected ear in the first two cases might reflect the greater effect of aging on PSC VOR-gain compared to the other SCs (42).

Unfortunately, low-acceleration VOR for HSCs could not be assessed with caloric test in the patients of our report due to concurrent external/middle ear pathologies preventing water irrigations. Nevertheless, further HSC VOR responses to different velocity and acceleration ranges could have been provided by rotational testing, but a rotatory chair was not available in our departments. Also, otolith activity could not be measured, as bone-conduction could have been the only possible way to test cervical and ocular VEMPs bypassing middle ear barriers, but it was not available in our institutions. However, given cochlear and SCs impairment, it is reasonable to assume that macular hair cells were also damaged, in accordance with studies in serous labyrinthitis (43–45).

Nevertheless, PIN could still be elicited in all subjects, likewise positional nystagmus in two cases. These apparently incongruent findings might be explained assuming a functional dissociation between ampullary hair cells encoding angular accelerations. In particular, detectable PIN despite SCs impairment on vHIT might imply a spared activity for type II hair cells and regular canal afferents encoding cupular displacements which generates nystagmus (46–48). This hypothesis is supported by studies on animal models of canal dehiscence providing evidence that sound-evoked eye movements (comparable in principle to PIN) do not only arise from sustained sound-evoked activation of phase-locking irregularly-discharging canal afferents, but also to slowly developing but sustained excitation/inhibition of regularly discharging afferents (48–50). Our assumptions are also in line with data from human temporal bone surveys showing that vestibular degeneration following serous labyrinthitis starts from type I hair-cells (44, 45). Similarly, preserved caloric responses were found in clinical studies on patients with COM (43, 51), perilymphatic fistula (52) or in cases exhibiting HS despite otolith functional loss (27). Conversely, investigations with rotatory testing have mainly documented reduced vestibular responses for the affected ear (53–55), strengthening the hypothesis of a greater impaired function of hair cells encoding for higher range frequencies. On the other hand, it may be assumed that there is a different functional outcome for damaged ampullary hair cells/afferents following acute labyrinthitis, where a selective recovery of sensors encoding for low-velocity inputs was matched by deficient high-acceleration VOR responses at long-term evaluation (38, 56). Nevertheless, a possible role of the residual function of phasic afferents in the genesis of PIN could not be ruled out a priori as all patients herein described did not present with complete canal loss on vHIT. In fact, the amount of residual canal function that is actually needed to still generate a response is not yet fully understood. Moreover, it could not be excluded that the different maximal eye velocities observed during each single canal impulse in each subject could have affected evoked nystagmus amplitudes and the interpretation of eye movements during PIN (26).

Whereas in the first two subjects the easiest explanation for HS is represented by pressure-induced endolymphatic flows toward the ampulla of the involved HSC activating spared/recovered regular afferents (Figures 4A,B), simultaneous cupular deflection toward the utricle in all dehiscent SCs could likely account for contralesional torsional nystagmus resulting from left tragal compression in the latter case. In particular, while ipsilesional nystagmus resulting from excitatory pressure-induced input within left HSC was mitigated by underlying baseline spontaneous nystagmus beating in the opposite direction, opposed vertical components generated from simultaneous inhibiting flows within vertical SCs canceled each other. Hence, such a PIN vector (right-torsional nystagmus) could be likely derived from the sum of concurrent inhibitions of left-sided vertical SCs afferents (14, 26) (Figure 4C). In accordance with Ewald's laws, pressure removal from the affected ears elicited coplanar weaker opposite nystagmus due to inhibitory ampullofugal flows within HSC in the first two cases. Conversely, the same maneuver generated stronger ipsilesional torsional nystagmus as a result of overlapping excitatory ampullofugal inputs within the left vertical SCs in the latter case (30).

The same reasoning could be applied in the interpretation of nystagmus behavior in the last patient during head movements along the pitch plane. In this case, vertical/torsional positional nystagmus due to simultaneous left-sided PSC BPPV evoked by head bending likely superimposed underlying baseline paretic spontaneous nystagmus (57). In the same patient, paroxysmal nystagmus could be elicited despite PSC VOR-gain loss on vHIT, strengthening the assumption of spared or recovered low-velocity afferents (58). Similarly, in the second case, positional geotropic direction-changing horizontal nystagmus closely matched with the expected oculomotor findings resulting from a buoyancy mechanism likely due to penetration of toxic agents and/or inflammatory mediators into the affected HSC (59–61). On the other hand, positional nystagmus has been reported in patients with labyrinthine-intracranial fistula. In fact, it has been hypothesized that intracranial pressure variations related to sudden changes in head positions could be conveyed into the dehiscent canal and result in excitatory/inhibitory endolymphatic flows accounting for direction-changing nystagmus (62–64). Similarly, head movements have been assumed to evoke subtle mass-induced pressure changes on the membranous labyrinth at the LF area in a patient with cholesteatoma eroding the HSC who presented with geotropic positional horizontal nystagmus (65). Nevertheless, the patient of the second case herein reported should not have developed similar mechanisms, as the LF was in contact neither with intracranial spaces nor with middle ear masses. Furthermore, an outward protrusion of the membranous duct through the right HSC fistula, that should be expected to occur in right-sided head positionings due to the gravity vector, should have resulted in ampullofugal flows, and in turn to apogeotropic nystagmus, which was not the case for our patient.

Although a reverse functional dissociation pattern impairing low-velocity while sparing high-velocity canal VOR has been often observed in several vestibular diseases (66–69), other vestibular pathologies have been related to a loss of sensitivity for high-acceleration head movements while low-acceleration behavior remains intact, likewise LF herein reported (70–72). In particular, SSC dehiscence represents another condition accounting for a third window mechanism that has been demonstrated to result in ocular movements aligning with the plane of the affected canal in response to loud sounds and/or pressure changes despite selectively impaired canal function on vHIT. However, in this condition, either a plug effect exerted by middle fossa structures on membranous labyrinth or dissipation of mechanical energy through the dehiscence have been assumed as underlying factors accounting for reduced high-velocity VOR-gain for the affected canal (73, 74). Whereas this latter mechanism could hypothetically account for the global canal hypofunction on vHIT in the last case with erosion of all the three SCs, it could neither explain the impaired responses in high-velocity domain for vertical canals detected in the first two cases with isolated HSC fistula, nor could it account for the concomitant sensorineural HL.

While both cases 2 and 3 presented with a widened low-frequency air-bone gap, as expected from a third mobile window pathology (11), several symptoms and signs pertaining to the third window spectrum could not be detected in our patients. Whereas the lack of pulsatile tinnitus, own body sounds hyperacusis and vertigo induced by loud sounds could be explained by the masking effect of the underlying middle ear disorders, nystagmus induced by Valsalva maneuvers could have been missing due to the different location of LF created by middle ear disease (as in our patients) compared to SSC dehiscence at the arcuate eminence. In fact, the lack of PIN in nasal Valsalva could likely be due to the fact that pressure transmission from the nasal cavity to the middle ear through the Eustachian tube was prevented by the coexistent middle ear pathologies. On the other hand, whereas glottic Valsalva maneuver should generate nystagmus through increased intracranial pressure conveyance to the labyrinthine spaces via SSC dehiscence (9), in no case herein described did LF expose the membranous labyrinth to intracranial cavity. Nevertheless, the lack of video-oculographic recording, providing an accurate detection of subtle eye movements to pressure/sounds and slow phase velocity measures, could have prevented detecting these signs in our study.

In general, the results of our study and literature data suggest that conclusions about SCs activity in the case of vestibular impairment only based on vHIT data could be misleading, as these measurements do not reflect the whole VOR response spectrum and dissociation among afferents encoding high and low-acceleration responses could be possible. In fact, as observed in our patients, HS could be elicited despite global canal hypofunction on vHIT, allowing clinicians to combine the analysis of PIN behavior and imaging to identify the location of LF prior to middle ear surgery (14). According to the same reasoning, vestibular hypofunction on vHIT should never authorize clinicians to neglect evaluating for provoked nystagmus in patients with vestibular symptoms, as residual/spared canal activity could account for ampullary activation to endolymphatic flows despite vHIT data.

Nevertheless, as already evidenced, the main limitation of this study, preventing any definite conclusion on the genesis of PIN in our patients, is the lack of objective measurements of both low-acceleration SCs VOR and otolith function. Despite the unlikelihood of a macular contribution to vestibular responses to pressure changes in the patients herein described, we could not exclude a possible activation of otolith receptors, as it has been described how these structures could modify ongoing ocular movements or generate spontaneous horizontal nystagmus (75, 76), and how they could be functionally spared in case of labyrinthitis with concurrent third window pathologies (77). Additionally, even though the presenting instrumental picture and PIN behavior in the patient with EAC tumor overlapped vestibular findings in the other two individuals with COM, we could not exclude that the two pathologies may result in different labyrinthine lesion patterns. Even though it has been reported how squamous cell carcinoma of the EAC could result in inner ear invasion through HSC erosions (78) and how middle ear tumors could lead to both labyrinthitis and labyrinthine ischemia (79, 80), inner ear histopathology in the case of temporal bone malignancy is still mostly unknown.

Although pathomechanisms underlying PIN in LF are still unclear, our case series showing that HS could be detectable even with hypoactive SCs on vHIT might offer additional insights to this aspect. Although vHIT measurements would suggest an impairment of canal function, SCs seem to represent the target sensor of HS in LF as nystagmus axis matched that of the affected canals and its characteristics strictly followed Ewald's laws. In our opinion, this apparently paradoxical finding might be possible through a functional dissociation between low- (active) and high- (impaired) velocity canal afferents. An asymmetrical damage or recovery among different subgroups of hair cells following labyrinthitis might represent the underlying process accounting for this functional behavior. Nevertheless, we strongly believe that further studies on PIN are required to substantiate the assumption that HS in LF results from the stimulation of type II hair-cells and regular afferents of dehiscent SCs.

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.

The studies involving human participants were reviewed and approved by Area Vasta Nord Emilia Romagna. The patients/participants provided their written informed consent to participate in this study. Written informed consent was obtained from the individual(s) for the publication of any potentially identifiable images or data included in this article.

AC, PM, SM, and EA: conceptualization and data interpretation. AC, CB, PM, SM, FC, MF, and FL: investigation and original draft preparation. AC, CB, MB, IJF, PM, LR, and GB: data acquisition. AC: images and artwork. GB, EA, and AG: supervision and manuscript review. All authors approved the final version of the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.