To allow controlled balloon inflation, a bespoke high-pressure precision syringe driver was constructed to be capable of generating pressures up to 10 bar, driven by a stepper motor (Fig. 1).Fig. 1

Precision high-pressure syringe driver

Three devices were assessed: the 3.0 × 20 mm TubaVent BET catheter (Spiggle & Theis Medizintechnik GmbH, Overath, Germany), the 6.0 × 15 mm EverCross angioplasty catheter (Covidien/Medtronic Ltd., Dublin, Ireland), and the 6.5 × 15 mm Sterling Monorail, angioplasty catheter (Boston Scientific, Marlborough, MA), the latter two acting as a similarly sized and structurally comparable alternative to the AERA BET device (Acclarent Inc.) which was unavailable at the time of the study. Prior to first use, each new catheter was inflated to working pressure once with no data collected, in order to unfold the packaged balloon. Preliminary studies confirmed that after the device’s first use, subsequent inflations were indistinguishable in terms of the pressure-volume relationship during inflation. Balloon diameter was measured three times using Vernier callipers placed across both the proximal and distal end of the parallel segment of the balloon, and a mean diameter value calculated for each of four inflation pressures (2.5, 5.0, 7.5 and 10.0 bar). Balloon overall length including the tapered sections (measured at 10 bar inflation pressure) was also measured with callipers three times, with a mean calculated.

BET was performed in human cadavers, and in simplified ET structures made from model materials. Adult fresh-frozen cadaveric heads were thawed overnight, after which they have tissue properties close to those found in vivo. To explore ET deformation during BET, model materials with purely plastic or elastic properties were selected. Silicone tubing of inner diameter (ID) 2 mm and outer diameter (OD) 4 mm was used as an elastic model of the ET, with detergent (Fairy, Procter and Gamble) used to lubricate the balloon within the tube. White plasticine (Newplast, Newclay Products) formed into tubes of ID 2 mm and OD 5 mm using a custom mould was used as a plastic model of the ET.

Three protocols were used in different specimens. Each Eustachian tube was only used once for a single protocol.

(1) The balloon catheter was inflated from 0 to 10 bar before immediate deflation to 0 bar, forming an inflation-deflation cycle. The cycle was initially performed outside the cadaver specimen, with the balloon unconstrained. The BET catheter was then inserted into the cadaver or model material ET, in the cadaver using a 45° angled introducer guided via a 30° endoscope. The catheter was slowly advanced into the tube, stopping to reposition if any resistance was felt. Once the catheter was fully inserted, three 10-bar inflation-deflation cycles were performed, with approximately 1 min between each, without moving the catheter between cycles. This protocol was used for both 3- and 6-mm balloons in both the cadavers and the two models. For this protocol, most cadavers received BET on one side/ear, with the other side untouched. In most cases, each cadaver had a 3.0-mm diameter balloon inflated in one ET, and a 6.0- or 6.5-mm balloon inflated on the other side, generating paired data.

(2) As in protocol 1, a 10-bar inflation-deflation cycle was performed with the balloon unconstrained, and the catheter then inserted into the ET. The balloon was first inflated to 2.5 bar, before immediate deflation. The 2.5-bar inflation-deflation cycle was repeated twice further. Without moving the balloon, further inflation-deflation cycles to pressures of 5, 7.5 and 10 bar were used, with cycles at each pressure repeated three times. This protocol was only used in the human cadavers. Each cadaver had a 3.0-mm-diameter balloon inflated in one ear ET, and a 6.5-mm balloon inflated on the other side.

(3) Following an unconstrained 10-bar inflation-deflation cycle, further cycles were performed at different ET insertion depths. Using paint marks on the balloons, the central, parallel sections of the 3.0 × 20 mm balloons were inserted to depths of 10, 15 and 20 mm from the nasopharyngeal ostium, while the 6.0-mm balloons were inserted to 10 and 15 mm. During initial inflations, the balloons therefore protruded, unconstrained, within the nasopharynx. As with protocol 2, this protocol was only used in human cadavers, and in all cases 3.0- and 6.0-mm balloons were used on opposite sides.

Data recorded for all inflation-deflation cycles were analysed in Matlab (MathWorks, Inc.). For each cycle, inflation pressure was plotted against injected volume, tending to form a loop due to the hysteresis of the system.

The area under the inflation part of the pressure-volume loop represents the energy required to inflate the balloon against resistance. If the balloon is unconstrained, this energy is the work done on the balloon’s structure alone, whereas during BET, it is the work done on the balloon and the ET. The area under the deflation part of the loop is the energy returned by the tissues and balloon as they return to their original shape. The area within the hysteresis loop therefore represents the energy dissipated into the balloon and ET. To calculate the energy dissipated into the ET during plastic deformation, the area within the 2nd cycle pressure-volume loop, representing elastic deformation alone, was subtracted from the area within the 1st cycle loop (Supplementary Figure A).

The energy transferred to the ET soft tissues through plastic deformation was considered to approximate the degree of plastic deformation occurring during BET, and was compared between the two balloon sizes.

Separate to the pressure-volume studies, some cadaver ETs were histologically examined after BET. ETs were flushed bilaterally with saline, before 2 ml of Alcian blue stain (aqueous solution) was injected into each ET using a soft-tipped syringe, and then confirmed to enter the middle ear. Next a 3.0 × 20 mm or 6.0 × 15 mm balloon catheter was inserted into the left or right ET, and a single 10-bar inflation-deflation cycle was performed. The contralateral ET was not treated, acting as a control. ETs were then thoroughly irrigated with saline to remove residual stain.

ETs were harvested via a trans-oral approach, having removed the soft palette. Blunt and sharp dissections were used to free the ET from the skull base and surrounding soft tissue structures, taking care not to crush the specimen. The cartilaginous ET was transected free as close to the bony portion as was technically possible. Specimens were fixed in 10% neutral buffered formalin, and cut transversely at 4-mm intervals for standard histological processing. Sections were prepared with haematoxylin and eosin stains.

Comparing the cadaver data to the model material data suggests that the ET undergoes both plastic and elastic deformation during BET, and it was found to consistently occur in all ears with both the 3.0- and 6.0/6.5-mm-diameter balloons.

The finding of plastic deformation of the ET during BET indicates that changes are occurring within the ET structure. Cadaver studies provided the first evidence as to what these changes may be: McCoul et al. endoscopically examined the ET after dilatation, and found linear fissuring with or without submucosal disruption in nearly 60% of ears [7]. Poe et al. found partial-thickness mucosal tears in some BET-treated ETs, but a full-thickness tear in only one [6]. In contrast, Ockermann et al. observed ‘microtears’ within the cartilage, while the mucosa and Ostmann’s fat pad were undamaged [5]. In common with these findings, we identified both mucosal tears and cracks in the apex region of the ET cartilage following BET. The absence of trauma in control specimens, and the blue stain within the mucosal and cartilage tears, suggested that the histological findings were genuine, having occurred during BET.

A single study has reported in vivo histological findings relating to BET: Kivekas et al. found diffuse crush injury and shearing of the mucosa and submucosa following BET, and in three biopsies taken 5 to 12 weeks postoperatively, restoration of largely healthy ciliated pseudo-columnar epithelium was found. It was hypothesised that BET removes irreversibly injured or inflamed cell lining, promoting recovery with normal tissue. This correlates with endoscopic findings in patients where visually scored mucosal inflammation at the nasopharyngeal orifice appears to improve after BET [11, 12]. In our study, it is likely that crushing of the mucosa and submucosa occurred, but it was felt that this could not reliably be assessed in the ET specimens, mainly due to the potential effects of manipulation of the tissue during harvest.

This study demonstrates that balloon size, inflation pressure and balloon insertion depth all impact on the changes induced in the ET during BET, and these are therefore likely to be important variables that determine the clinical effectiveness of the intervention. These technical details have typically been overlooked when comparing or combining BET clinical data, but they may underlie some of the variation seen in outcomes [2], and they will be important considerations during future development of the technique.

Figure 4 demonstrates a significantly greater dissipation of energy during plastic deformation with the 6.0/6.5-mm balloons, a trend also seen in Fig. 6, suggesting that more plastic deformation occurred with the wider device. Given the findings, with the wider balloons the same tissue must be displaced or crushed more, and/or more tissue must be involved, possibly due to involvement of deeper structures such as cartilage or muscle.

The cartilaginous ET targeted for dilation with BET is conical in shape, narrowing from a relatively wide nasopharyngeal ostium to the narrowest part of the ET close to the bony-cartilaginous junction. It is perhaps not surprising therefore that plastic deformation of the ET appears to increase with greater depth from the ET nasopharyngeal ostium. Figure 8 shows the balloon dimensions in relation to the narrowest part of the ET, which in adults is located 20.5 ± 4.2 mm from the pharyngeal orifice [13]. It is noticeable that due to the tapering design of the balloon ends (a feature also seen in the AERA balloon not tested), the 20-mm-long balloon does not have an effect on a longer section of the ET than the 15-mm balloon. If the balloons are fully inserted as instructed by the BET device manufactures using either the TubaVent introducer or the AERA insertion marker, during inflation the full diameter of the balloon should sit within or close to the narrowest point of the ET.Fig. 8

Scale diagram of the 3 × 20 mm TubaVent (TV) balloon overlying the 6 × 15 mm EverCross (EC) balloon (6.5 × 15 mm Sterling Monorail balloon not shown on this figure for clarity). Full insertion of both TubaVent and AERA devices is intended during clinical use. The red dashed line indicates the mean point of the narrowest part of the ET based on histological studies

The majority of the plastic deformation occurred at relatively low inflation pressures, at or below 5 bar, and while deformation appeared to continue up to 10 bar with all balloons tested, this was less marked. BET and angioplasty balloons are designed to be non-compliant, meaning a designated size is reached at relatively low pressure, with minimal size increase with additional pressure. Our measurements confirmed this to be the case with the devices used in this study, and this may explain why higher pressures cause limited additional plastic deformation. In the absence of reported significant pressure-related complications, and with some ongoing changes up to 10 bar, currently this appears a reasonable target inflation pressure. However, BET can be performed under local anaesthesia [14], and lower-pressure inflation could be considered in these cases to reduce patient discomfort.

Plastic deformation was only seen on the first inflation, with no further changes occurring during the second or third inflation. This suggests that one inflation is sufficient, and the multiple inflations proposed by some clinicians are unlikely to add clinical benefit, and therefore may unnecessarily prolong the procedure.

Although clinical studies of BET have been of limited size and quality, BET appears to provide at least a short-term symptomatic improvement in patients with OETD [2]. Clinical data have also shown increased passive ET opening with Valsalva [12, 15, 16], a marked improvement in active ET opening with swallowing [17], and normalisation of tubomanometry R values [18, 19]. Based on the experimental findings, the mechanisms for these clinical improvements can be hypothesised.

If mucosa, cartilage or even muscle is crushed, the ET lumen may become more open at rest, or open wider during paratubal muscle contraction. According to Poiseuille’s law, the wider lumen would reduce the resistance to air flow, possibly improving ET ventilatory function.

An alternative, or possibly complementary, explanation for the experimental findings is that BET alters the stiffness (compliance) of part or all of the ET, through the crushing of soft tissues, or the mucosal tears and cartilage cracks as seen histologically. Of note, the cracks occurred in a region of the ET rich in elastin fibres that is thought to act as a hinge during opening [20]. A number of studies have suggested that ET compliance is an important determinant for its function, with low compliance (high stiffness) associated with OETD and middle ear effusion in adults [21–23]. Modelling also suggests that an increase in ET compliance may enhance active ET opening, with mucosal stiffness of greater influence than muscle function or cartilage stiffness [24, 25]. The relationship between compliance and ET function may be more complex in children, however [21, 26, 27].

The findings from this study could be consistent with an increase in lumen diameter following BET, or with an increase in ET compliance. The relative contribution of these acute structural and mechanical changes to ET function following BET in vivo remains unclear, and post-operative swelling is also likely to occur soon after BET possibly affecting ET mechanical properties.

A small number of studies have found that the improvement in ET function after BET is not immediate, but emerges after a period of 6 to 8 weeks [18, 28]. This timescale suggests that a process of healing or scarring is occurring to improve the function.

Looking to analogues of BET in other tissues, Modi et al. performed an in vivo assessment of balloon dilation of the subglottic airway in a rabbit model [29], and while differences in structure exist, both the subglottis and ET have a cartilaginous skeleton lined by respiratory epithelium and mucosa. Thirty days following dilation, they found regeneration of a normal epithelial lining, with submucosal fibrosis present in many specimens, particularly where larger balloon sizes had been used. It may be that submucosal fibrosis occurs following BET, possibly stiffening the ET.

In cases of otitis media, the ET mucosa is thickened with numerous glands, and the underlying submucosa contains large numbers of macrophages, plasma cells and lymphocytes [30]. The crushing of this tissue and restoration of healthy mucosa as described by Kivekas et al. has a number of possible effects. Healthy mucosa will be thinner, facilitating ET opening, and the normal function of cilia and mucus production may aid the clearance of secretions from the middle ear [31]. In addition, ET mucosa generates a surfactant that reduces surface tension and acts as an anti-adhesive, reducing the ET opening pressure [32]. There is evidence that ET surfactant may be reduced in OME [32, 33], and the restored healthy mucosa may secrete more surfactant, thus improving ET opening.

ET cartilage displays features of both fibrous and elastic tissue types [20], but the hinge portion in which the cracks were found following BET is most similar to the elastic cartilage of the pinna and epiglottis. It is likely that there will be some healing of these cracks, with comparable elastic cartilage elsewhere tending to repair slowly with a dominance of type 1 collagen, and loss of normal tissue architecture that may affect ET compliance [34].

This work represents the largest study of BET in cadavers, though sample sizes were necessarily limited. A striking feature of the findings was the consistency between different specimens, with every ear demonstrating some degree of plastic and elastic deformation, and a statistically significant difference in energy dissipation found between balloon sizes. However, there were an inadequate number of specimens to enable a link to be established between histologically identified trauma and the results of pressure-volume loops. All cadaver specimens used were adults, typically of older age, and as changes may occur in Eustachian tube cartilage with ageing (such as focal calcification) it is possible that results could differ in young adults.

The main limitation of this cadaver-based work was the inability to study the effect of tissue healing and remodelling following BET, or the clinical impact of the mechanical changes. Following further development, the high-pressure syringe driver could be used during BET in vivo to allow quantification of the energy dissipation, and therefore an assessment of the plastic deformation occurring during the first inflation. These results could then be correlated with clinical outcome measures at follow-up intervals, to determine if the extent of deformation measured has a clinical correlate. If the two were found to be related, deformation could be maximised through experimental modification of the device (size, shape, materials) or technique (insertion pressure and duration of inflation) in cadavers. It should be noted that in theory if balloon dilation was performed excessively, a permanently open or patulous Eustachian tube may result, which is a condition equally as problematic as OETD. While this has not been reported clinically using the balloon sizes or techniques tested in this study, it is another area that requires clinical correlation when designing future devices and techniques.