The saccule resembled a vertically placed coffee bean with the convex side lying against the spherical recess and the contralateral side facing the vestibule. The superior wall was flattened, lay against the floor of the utricle, and formed an imprint. This suggests that expansion of the saccule may compress the utricle near its macula and vice versa. Basally, the inferior margin of the macula faced the entrance of the RD, where the distance between the otolith membrane and RD was in the region of 0.2–0.3 mm. Here, dislodged otoconia could easily reach the RD by gravity. The saccule wall consisted of a thick and thin region where the thick part lay near the spherical recess. In lower mammals, amphibians, and fish, the saccule contains only the thin part, portentous that the macula may respond to both acoustic and static pressure (26). The orientation of the human saccular macula, the spherical recess and the re-enforced membranes were thought to shield the macula from acoustic oscillations such as stapedius-induced vibrations at large sound levels, in contrast to lower forms. The thin wall appears more flaccid and is probably more prone to dilate and collapse in MD. Ruptures may occur at the junction between the regions by increased pressure due to the resilience. The thin wall of the SD was difficult to image in 3D as it emerged near the band-like reinforcement ~0.9 mm from the utricle floor.

The utricle is a horizontal cylinder with several extremities. In this study, for the first time, the UEV was viewed three-dimensionally and it appeared as a fantail with a slit-like opening. Its semi-lunar shape differed somewhat from the descriptions by Bast (27) and Anson and Wilson (28). It was located near the floor of the posteromedial wall of the utricle at the internal opening of the VA, as described in humans by Bast (27, 29) and in other mammals by Hoffman and Bast (30). Wilson and Anson (31) described the valve in a 2-year-old child, and they defined it in an adult and named it the “utricular fold.” We used the nomenclature of Perlman and Lindsay (32) for the UD, SD, and ED. They are alleged to form a Y-shaped type I junction in 83% of the cases, with a wide angle between the UD and ED (27). In type II (15%), the UD is short with a broad angle to the SD. In type III (2.8%), there is almost no UD at the right angle to both the SD and ED. We identified a short fan-like UD reaching the UEV with a valve facing the sinus portion of the ED.

There are several theories as to the function of the UEV. Bast (27, 29) suggested that the valve closes the pars superior to regulate endolymph volume. Bast (29) described this as follows (p. 64):

Bast went on to claim that fluid pressure on the valve in the utricle should force it against the opposing duct wall. This was also supported experimentally in the guinea pig, after reducing perilymph and endolymph pressure in the cochlea (33). The valve remained closed, and the utricle and semicircular canals distended even after the collapse of the saccule and cochlear duct. Bast (34) also showed cases of rupture and collapsed saccules, where the utricle was intact. The UD was firmly closed by the UEV. Despite the rupture of the saccule and collapse of the saccule and cochlear duct, the utricle did not collapse. This suggests that endolymph pressure in the utricle can be maintained independently of the pressure in the saccule and cochlear duct. It also proposes that the resolutely closed UEV is responsible for maintaining utricle pressure and therefore acts as a valve. The findings also indicate that fluid is produced in the utricle/semicircular ducts (9). Several regions within the vestibular labyrinth, such as the cupula, have the potential to secrete endolymph similar to the lateral cochlear wall (35, 36). Guild's (5) opinion was that endolymph flows from the utricle and semicircular duct to the ES.

Bast (29) found that in the fetus, the valve is lined by columnar epithelium and at the base with a highly cellular connective tissue that continued into “paralymphatic” tissue around the endolymph system between the utricle and ED. The loose tissue may permit movement of the stiffer valve in case of increased pressure in the utricle. He could not characterize the connective cells but did not exclude the possibility that smooth muscle fibers are present, but there were no indications of nerve fibers. The tissue was independent of the bone surrounding the aperture of the VA. According to Anson and Wilson (28), the fold contains areolar tissue and a fibrous web with a spear-shaped projection of periosteum projecting from the osseous wall but not into the fold proper. This was denied by Bast (27). Anson and Wilson thought that the stiffer valve or lip may not move by pressure changes, while the outer wall could close the orifice and cause movement of fluid in the UD. Schuknecht and Belal (37) studied the valve in humans and found it ideal to protect the pars superior from collapse after dehydration of the rest of the labyrinth. Bachor and Karmody (38) speculated similarly that reduced pressure in the endolymph system, secondary to collapse of the RD, may close the UEV and prevent loss of endolymph from the utricle. Our study showed fibrous connections or pillars between the utricle and the inner surface of the bony labyrinth wall that may stabilize the utricle and also deter it from collapse in cases of reduced endolymph pressure.

Hofman (39) presented 3D imaging of the UEV in laboratory animals using orthogonal-plane fluorescence optical sectioning microscopy and Lim using scanning electron microscopy (40). The valve was flat and funnel-like at the utricle and ran into a narrow and short duct leading to the sinus portion of the ED. Hofman (39) described the valve as fairly rigid. The opposing utricle membrane was thin and appeared to be compliant. They speculated that the outer wall may play a greater role in the opening and closure of the valve.

SR-PCI showed the semi-lunar lip lying against its inner surface with a somewhat different shape, as shown by Bast (27) (Figure 6). We identified reinforcements around the lip after cropping and adjusting the opacity gradient, suggesting that it is fairly rigid (Figure 7). Consequently, increased pressure within the utricle would seem to open the valve through forces acting on the more flaccid part of the opposite membrane. A lowered pressure within the utricle could perhaps close it. External reduction of pressure in the SD and ED could maintain closure to avoid collapse. Conversely, an increased external pressure could force the valve to open by compressing either the flaccid membrane of the utricle or the lip. The UD was found to be partly situated on the loose basal part of the valve, suggesting a mechanism to externally open the lip (Figure 14). A fibrous wall from the VA was also connected to the UD and the valve to hold it in place.

The RD was first described by Viktor Hensen in 1863 (4). The Swedish anatomist Gustaf Retzius defined the RD in a newborn in 1884 as a fairly wide channel connecting the cecum vestibulare of the cochlea with the saccule (3). It contrasts with the miniscule channel described histologically by several authors. Its patency has been confirmed at serial sectioning, and both physiological and tracer studies suggest an endolymph flow or communication (5, 18, 41, 43, 44).

Earlier micro-CT did not define the human RD after manual segmentation and contrast enhancement (45). The present micro-CT imaging was able to reproduce it when air entered the vestibule. The small dimension may refute the idea of an ongoing flow under normal conditions. However, size variations may exist, and dilation can be seen after obliteration of the ED (46). Bachor and Karmody described the human RD as small, collapsed, or wide in pediatric temporal bones (38). In half of the bones, the RD was collapsed and seemed closed. They speculated that it may open when pressure increases and permanent closure may lead to hydrops. Since the RD epithelium is stratified and rests on connective tissue enhancement, it is reasonable to believe that the RD can dilate under conditions of increased endolymph pressure and generate a flow. The thin connection may protect the cochlea from leakage of electricity to maintain cell currents at high frequencies. The endolymphatic potential has been found to be low in the vestibule except near the receptors (47, 48).

Despite the large number of histopathological studies in MD, only a fraction seem to adhere to the AAO-HNS 1995 criteria. Hallpike and Cairns (49) described two cases where EH was prominent with excessive dilation of the saccule and scala media. There were also changes in the pars superior but no signs of blockage of the communication pathways connecting the utricle, saccule, and ED. The UEV was in its normal position, and the opening of the saccule into the ED contained a dense reddish-staining coagulum in the second case. The absence of loose connective tissue around the ES was a prominent feature. Kyoshiro Yamakawa (50) described similar findings, but there was no mention of the ES. There were edema and calculi in the stria vascularis, the RD seemed open, and the VA was large with colloid-like substance in the duct. He believed that MD symptoms were caused by an augmented pressure in the endolymph which was caused by an increased secretion from the stria vascularis.

EH is a consistent histopathological trait in patients with MD, but its role in the progress of symptoms is unclear. Blockage of the EDS, including the UEV, RD, and SD, is not overrepresented in MD. Dilation of the RD was observed in patients with EH, possibly as a result of increased cochlear hydrostatic pressure (51, 52). Lindsay (53) demonstrated a wide RD in an MD case, but its dimension was not given. According to Shimizu et al. (17) the ED was blocked in 23% in MD, while the UD was blocked in 76% in MD and 52% in normal ears. The SD was blocked in 28% in MD and 76% in normal ears. It suggests that obstruction of the UD and SD may not be the cause of EH in MD. A theory based on CT suggested that displaced saccular otoconia may block the RD, explaining symptoms in MD (54). It is conceivable that a few or even a single displaced otoconia could mechanically obstruct the RD in humans (55). According to Bachor and Karmody (38), there is no correlation between the collapse of RD and cochlear hydrops. Both wide-open and blocked RDs were demonstrated. In patients with MD, the RD was found to have a diameter around 0.1 mm (56). Shimizu (17) showed that the RD in a patient with MD was in the range of 0.05 mm. Kitamura et al. (57) investigated five cases with EH limited to the cochlea with no verified MD. Various obliterations of the saccule and/or RD supported a longitudinal flow of the endolymph. EH depended on the location of the blockage from the RD to the ES. The present study suggests that RD may play a minor role in the exchange of endolymph between the cochlea and vestibule, but under certain conditions, it can dilate and open to maintain cochlear homeostasis.

Schuknecht and Ruther (58) showed that in 42 out of 46 ears, there was either a blockage of longitudinal flow or internal shunts causing fistulae. The UD was blocked in 12, ED in 8, endolymphatic sinus in 9, SD in 7, and RD in 27 bones. They considered that obstructions stopped longitudinal flow from both the pars superior and inferior in 21 cases, only the pars superior in 3 cases, and only the pars inferior in 16 cases. Schuknecht and Belal (37) studied the UEV in 29 human temporal bones with MD and found deposits in the lip and enlargement of the saccule compressing and closing the UD. The 3D reconstructions indicated that the endolymph volume may increase up to three times in MD (59). The relative increase was most prominent in the saccule. A possible explanation is the thin part that may expand even before pressure is built up in the utricle and maintained by the UEV. This could explain MRI findings showing occult or prominent EH preceding symptoms in some patients and in the contralateral asymptomatic ears (60–64).

The role of the ED and ES in MD is still unclear. A diminished absorption in the ED and ES could be due to mechanical obstruction or molecular or fibrotic changes (17) or disturbed vascular drainage (65). Yuen and Schuknecht (66) found no obstruction of the VA in MD and no difference in caliber compared with normal ears. The ED was smaller probably as a result of MD rather than the cause of it. A reduced radiographic visualization of the VA was described by Wilbrand (67) in MD, conceivably explained by its smaller size (68, 69). Results by Monsanto et al. (70) suggested that the UEV and SD may open as a result of retrograde pressure caused by failure of the ES to absorb endolymph. A wide-open valve could disrupt the crista function, resulting in vertigo since it would not protect against pressure fluctuations potentially harmful to the sensory epithelia. An increased pressure in the ED and ES could lead to a reversed flow of endolymph into the utricle and cause selective vestibular disturbance in MD (71).

Experiments have shown that induced alterations in cochlear endolymph volume result in pronounced changes in the ES with bidirectional responses (41, 72). Acute manipulation of the systemic fluid or chronic malfunctions, such as in cochleo-saccular degeneration, modify ES activity (42, 73–76). This suggests that the endolymph moves in either direction between the cochlea and ES and is reabsorbed by the ES under conditions of excess volume and secreted by conditions of volume deficiency. Thus, volume may be regulated in the entire labyrinth by the ES. Potentially, a dysfunction in the ES could lead to a volatile elevation of endolymph pressure and acute cochleo-vestibular symptoms, such as in MD. Nonetheless, there are noticeable differences in the anatomy of laboratory animals. Physiologic adaptation and human upright position may modify fluid exchange and dynamics amid the ear and cranium, with conceivable changes also in the structure of the EDS.

Imaging data suggest that the ED, VA, and UEV form a functional unit for the pars superior. This unit includes a stabilizing fibrous disc projecting from the vertical crest of the internal aperture of the VA to the UD and UEV (Figures 5C, 9C). The UD is directed against and physically associated with the flexible part of the lip of the valve, indicating that the duct may force the valve to open and transmit increased endolymph/pressure into the pars superior (Figure 14). This framework could be particularly prone to mediating sudden escalations of ES pressure into the utricle through the relatively wide UD and UEV. The ES is an expandable bellow-like structure with an extra-osseous part comprising more than two-thirds of the total ES volume. It could mount considerable osmotic pressure gradients against the vestibular labyrinth, despite its 16 times less volume (1.85/30.4 mm³) (69, 77, 78). A secretion–degradation system of osmotically active complexes was revealed in the ES acting within a short time frame. Secretion may be triggered by altered volume/pressure conditions in the labyrinth (73), linked to its ability to monitor endolymph (41, 42). A merocrine secretion was already observed in the human ES in an ultrastructural investigation (79). Super-resolution immunohistochemistry recently confirmed a dual absorptive–secretory capacity of the human ES (80). In a temporal bone study in MD, an unproportioned large volume of the ES contained secreted material in an otherwise smaller sac (69).

Therefore, a diminished resorption or hypersecretion of endolymph in the ES could result in a retrograde flow/pressure that underlies the acute attack of vertigo in MD. We speculate, based on the present morphological findings and earlier experimental studies, that acute vestibular symptoms in MD arise through a sudden fluid pressure increase and dilatation caused by a “backflow” of an overcompensated ES secreting into the utricle and semicircular canals. ES endolymph has a unique ionic composition (81), which is potentially noxious to the vestibular receptors. Dilation could lead eventually to membrane ruptures (82) additionally extending the ES response. Pressure may impact saccular and ultimately cochlear functions via the SD and RD with corresponding symptoms. Our results support Seymour's (2) notions that the ES has a supporting secretory role necessary to compensate if volume is reduced to insure complete filling of the phylogenetically older utricle and semicircular canals. Microinjections suggest that elevated hydrostatic pressure in the pars inferior or directly into the utricle result in utricle/vestibular receptor dysfunction (83). These changes were believed to be similar to sudden fluctuating changes in MD suggesting that EH is the cause of the typical symptoms (84).

Our concept could have several clinical consequences. The dysfunction of the ES may arise through hormonal disturbances, stria alterations, immune factors, etc. (85). One proposed way to relieve this dysfunction is through surgical destruction of the ES (86). Nonetheless, more studies of the UEV, EDS, and ES function in MD are necessary and could lead to novel treatment modalities against this troublesome disease. There is also a need for more studies on how these small compartments control fluid pressure and ion homeostasis.