Retraction pocket aetiology

The tympanic membrane's elasticity allows for a change of shape and thus volume of the middle air and mastoid cell air system, acting as a natural buffer for pressure changes. When the tympanic membrane's buffer capacity is exceeded, pressures are equalised by the opening of the Eustachian tube. When this system fails, the negative pressure in middle ear leads to prolonged displacement of the tympanic membrane, weakening its elastic fibres and further displacing it towards the middle ear.^{Reference Swarts, Alper, Luntz, Bluestone, Doyle and Ghadiali10,Reference Marchioni, Mattioli, Alicandri-Ciufelli, Molteni, Masoni and Presutti11}

Four different facets of the buffering system can be affected: the Eustachian tube opening, the elasticity or consistency of the tympanic membrane, the mastoid air cell system, and the intratympanic ventilation pathways. Eustachian tube dysfunction has been recognised as a prominent contributor to the progression of a retraction pocket.^{Reference Ramakrishnan, Kotecha and Bowdler4,Reference Steinbach, Pusalkar and Heumann12,Reference Bhide13} Its most common causes are post-nasal space occupying pathology, cartilaginous pathology, tensor veli palatini pathology, and mucosal oedema.^{Reference Ramakrishnan, Kotecha and Bowdler4} Common contributors to the area's mucosal oedema are atopy and allergic rhinitis, nasopharyngeal infections, and gastroesopharyngeal reflux. Lieu et al. showed evidence of pepsin and Helicobacter pylori in the middle-ear fluid of children with otitis media,^{Reference Lieu, Muthappen and Uppaluri14} and Nguyen et al. similarly reported interleukin 4 and eosinophilic infiltration in children with allergic rhinitis.^{Reference Nguyen, Manoukian, Tewfik, Sobol, Joubert and Mazer15}

The aforementioned problems are more prominent in children because of their anatomical differences.^{Reference Ruah, Schachern, Paparella and Zelterman16} As children age, the Eustachian tube develops a higher degree of angle and greater length, making it more resilient to reflux, while the cartilaginous elements stiffen, resisting collapse. Lastly, adenoids tend to regress, reducing the risk of Eustachian obstruction.^{Reference Ramakrishnan, Kotecha and Bowdler4,Reference Ruah, Schachern, Paparella and Zelterman16,Reference Vanneste and Page17}

Paradoxically, Eustachian dysfunction is also associated with a pathologically patent tube – a patulous Eustachian tube. A patulous Eustachian tube is associated with habitual sniffers and people who apply a non-successful Valsalva manoeuvre, leading to repetitive evacuation of the middle ear and a prolonged negative pressure environment, further precipitating chronic otitis media and retraction pocket formation.^{Reference Yoshida, Kobayashi, Takasaki, Takahashi, Ishimaru and Morikawa18}

Middle-ear inflammation is another condition associated with retraction pockets.^{Reference Ruah, Schachern, Paparella and Zelterman16} The episodes of recurrent inflammation lead to atrophy of the lamina propria, while the mucosal inflammation and hypoxic environment alter the gas exchange patterns in the middle ear.^{Reference Sadé19,Reference Danner20} This leads to sustained negative pressure, causing the pliable segments of the tympanic membrane to collapse medially and form retraction pockets.

The abnormal gas exchange in the middle ear associated with a compromised buffer from the mastoid bone contributes to retraction pocket formation.^{Reference Danner20} Impartial pneumatisation or sclerosis of the mastoids, and the inflammation of their mucosa, hampers middle-ear ventilation and leads to negative pressure. According to Boyles law, for a given amount of gas at a specific temperature, pressure multiplied by volume equals a constant; hence, the decreased volume for gas exchange translates as a drop in pressure within the middle ear.^{Reference Sadé, Cinamon, Ar and Siefert21}

Lastly, the air diffusion in the temporal pneumatic system is regulated by the tympanic isthmus.^{Reference Aimi22} The tympanic isthmus boundaries are the tensor tympani anteriorly, the posterior incudal ligament posteriorly, the incus and the head of the malleus laterally, and the attic medially.^{Reference Shirai, Schachern, Schachern, Paparella and Cureoglu23} The tympanic isthmus is the main ventilation path of the epitympanum through its diaphragm, as the Eustachian tube ventilates only the mesotympanic and hypotympanic spaces.^{Reference Kobayashi, Toshima, Yaginuma, Ishidoya, Suetake and Takasaka24–Reference Monsanto, Pauna, Kaya, Hızlı, Kwon and Paparella26} A blockage of the tympanic isthmus can thus lead to extensive middle-ear pathology, and its correlation with cholesteatoma formation has previously been illustrated.^{Reference Monsanto, Pauna, Kaya, Hızlı, Kwon and Paparella26,Reference Miyajima and Honda27}

Almost two-thirds of retraction pockets lie in the pars tensa and one-third are located in its posterosuperior portion.^{Reference Ars28} Ruah et al. explained this occurrence by studying the histopathology of temporal bones, and by demonstrating the persistence of mesenchyme for more than three years in children with otitis media.^{Reference Ruah, Schachern, Paparella and Zelterman16} The affected areas had inadequate elastin maturation, becoming weaker and more pliable when the mesenchyme eventually reabsorbed.^{Reference Ruah, Schachern, Paparella and Zelterman16}

Classification

Tympanic membrane retractions can be categorised as stable or unstable, depending on whether a cholesteatoma has developed. Stable tympanic membrane retractions may be asymptomatic incidentalomas, or associated with conductive hearing loss, recurrent ear discharge or otalgia. Unstable tympanic retractions are caused by the obstructed migration of epithelial cells and keratin entrapment in the middle ear. They are associated with hearing loss, vertigo, facial nerve palsy and intracranial complications.^{Reference Sudhoff and Tos29–Reference Cutajar, Nowghani, Tulsidas-Mahtani and Hamilton31}

Currently, there is no consensus regarding the timing and type of intervention, as we are unable to predict which retractions will progress into cholesteatoma formation, and which will remain stable or resolve.^{Reference Nankivell and Pothier5} The two historical classification systems are the Sadé classification system, used to describe pars tensa pathology,^{Reference Sadé and Sadé32} and the Tos classification system,^{Reference Tos and Poulsen33} which describes pars flaccida pathology (Table 1). Both describe the extension and medialisation of disease, as well as the middle-ear structures’ involvement. Charachon et al. developed a different classification system, highlighting the retraction pockets’ visibility and self-cleaning ability (Table 1).^{Reference Charachon, Barthez and Lejeune34}

Table 1.

Retraction pocket classification systems

TM = tympanic membrane; NA = not applicable

Although there are several other grading systems, each prioritising specific features of retraction pockets as vital, they all have multiple limitations. The principle of classification methods is the correlation between the distinctive features of each grading category and different prognoses and treatment modalities. Unfortunately, there is not a specific treatment modality for each separate stage, and most of the classifications allocate the same treatment alternatives across multiple different stages.^{Reference Alzahrani and Saliba35} In reality, decisions regarding treatment are multifactorial, too complex or numerous to fit under a single classification system. Moreover, there is no consensus on the individual stages of retraction pocket formation correlating to the risk of progression to cholesteatoma.^{Reference Olszewska, Rutkowska and Özgirgin36} Lastly, these grading systems suffer from low intra- and inter-rater reliability,^{Reference Pothier37,Reference Vijayendra, Mahajan, Vijayendra and Ramdass38} requiring otologists to interpret them critically even when used to communicate their results.^{Reference Alzahrani and Saliba35}

Regardless of their inconsistencies, these classification systems remain the sole common tongue for otologists to communicate their findings and keep track of disease progression in each patient. An image-based classification system or the opinions of at least two independent otologists in the classification process, in research, would offer less bias and increased objectivity. Overall, the high number of different classification systems and their relatively limited clinical importance reflects the lack of consensus in the scientific community regarding the optimal management for retraction pockets at each stage.^{Reference Ramakrishnan, Kotecha and Bowdler4} Ultimately, the management decision is guided by balancing the potential benefits and risks of the interventions, taking into consideration the patient's underlying baseline state.

Management

The progression of tympanic retractions is unpredictable; the majority of grade I and II retraction pockets tend to resolve on their own, but almost one-fifth of cases will deteriorate in grade within five years,^{Reference Charachon, Barthez and Lejeune34} while a few cases will progress into cholesteatoma formation. Treatment modalities include a watch-and-wait approach, medical management and surgical interventions. Currently, there is no consensus regarding the appropriate management of retraction pockets. Although most clinicians support an initial conservative approach, there is no clear cut-off regarding when a surgical solution should be offered or what that should be.^{Reference Nankivell and Pothier5,Reference Saffer, Silva, Peduzzi and Ávila39} It is still debated whether surgical intervention is required for: Sadé grade II/III retraction, to prevent ossicular chain erosion;^{Reference Ramakrishnan, Kotecha and Bowdler4} Tos grade III retraction, to prevent cholesteatoma formation;^{Reference Nankivell and Pothier5} and Sadé grade IV retraction, to improve hearing outcomes.

It is widely accepted that grade I and II retraction pockets benefit more from conservative management. This consists of regular inspections paired with regular aural toilet and medications, for improved nasal and Eustachian tube function. Medications for symptomatic relief in cases of upper respiratory infections or allergic rhinitis include decongestants, antihistamines and steroid sprays.^{Reference Silverstein, Light, Jackson, Rosenberg and Thompson40,Reference Cantekin, Rockette, Bluestone and Beery41} Patients with Eustachian tube dysfunction are better managed using Valsalva manoeuvres and inflation devices.^{Reference Blanshard, Maw and Bawden42}

The surgical treatment of retraction pockets is diverse, with different strategies available depending on the underlying condition and the extent of disease. There are two main categories of treatment: improving middle-ear ventilation and repairing the retraction pocket itself. The targets for improved ventilation may be the Eustachian tube, the mastoids or the middle ear itself. Eustachian tube dilatation or ablation is a viable strategy, in cases of Eustachian dysfunction, while the insertion of ventilation tubes (grommets or T-tubes) remains a mainstay strategy. Although the incidental presence of effusion in the middle ear should not guide decisions on management, its presence may interfere with the diagnosis of other underlying conditions, and thus grommet insertion is advised by some authors.^{Reference Alper and Olszewska43} Nonetheless, the utility of ventilation tubes in tympanic retraction treatment is questioned; in their randomised, controlled trial (RCT), Elsheikh et al. showed no benefit of ventilation tube insertion compared to undertaking only cartilage tympanoplasty in terms of patients’ hearing outcomes.^{Reference Elsheikh, Elsherief and Elsherief44}

Adenoidectomy for patients with persisting adenoidal hypertrophy is an option for Eustachian tube release, although the reported outcomes are inconsistent and the improvement was statistically non-significant.^{Reference Bluestone, Cantekin, Beery and Stool45–Reference Honjo, Tashlma, Mitoma and Hamada47} Some authors recommend adenoidectomy as a first-line treatment for children aged four to eight years with retraction pockets, regardless of the adenoids’ size.^{Reference Alper and Olszewska43} Eustachian tube dilatation is another surgical option that has gained popularity lately; however, there is not yet enough high-level evidence to support its outcomes or safety. A UK-based, cross-sectional survey study showed that most otolaryngologist consultants were reluctant to practise Eustachian dilatation without better evidence, calling for the design of a national audit database and guidelines.^{Reference Koumpa, Moraitis, Bowles and Saunders48}

Simple excision or tympanoplasty with a graft aims to repair the retraction pocket, and can be paired with grommet and/or mastoid exploration.^{Reference Ramakrishnan, Kotecha and Bowdler4,Reference Nankivell and Pothier5,Reference Elsheikh, Elsherief and Elsherief44,Reference Dornhoffer49} The affected section of the tympanic membrane is excised and replaced with a fascial graft harvested from temporalis fascia, or with a perichondrium graft from pinna or tragus cartilage.^{Reference Sharp and Robinson50} This intervention is usually necessary when the thinned tympanic membrane is prolapsing over the promontory or extending anteriorly, having lost its self-cleaning ability and being at risk of keratin debris accumulation. An invasive approach is also necessary in the presence of ossicular chain disruption, even in the absence of cholesteatoma. When tympanoplasty is undertaken, a medial, lateral or cartilage graft may be required, depending on the size of the dehiscence and the degree of underlying Eustachian tube dysfunction.^{Reference Alper and Olszewska43}

When a retraction pocket becomes complicated with cholesteatoma (Figure 2), its radical excision is required. The technique employed depends on the disease extension, the structures involved and the patient's hearing.^{Reference Zhang, Chen, Sun and Zheng51–Reference Duckert, Makielski and Helms53} Although the treatment of cholesteatoma exceeds the purpose of the current review, there are conflicted opinions regarding the risk of leaving residual disease in the area of the anterior epitympanum in cases of conservative resection.^{Reference Williams54,Reference Steward, Choo and Pensak55} The general consensus is that an atticotomy with osteoplastic reconstruction could be undertaken, instead of a canal wall up or canal wall down technique.^{Reference Duckert, Makielski and Helms53} In cases where the malleoincudal joint is free from the disease, the ossicular chain can be spared, leading to a reduction in associated hearing morbidity in that patient group.^{Reference Zhang, Chen, Sun and Zheng51}

Fig. 2.

Retraction pocket complicated with cholesteatoma, managed with posterior cartilage graft and laser tympanoplasty: (a) pre-operatively and (b) 12 months post-operatively.

While the treatment of an unstable retraction pocket necessarily involves surgical excision of the accumulated cholesteatoma, the balance of potential harm or benefit in the case of stable retraction pockets is blurred. Retraction pockets tend to recur, at an average rate of 35 per cent in cases of simple excision,^{Reference Sharp and Robinson50} with other authors reporting similar outcomes that are not affected by retraction grade or patient age.^{Reference Srinivasan, Banhegyi, O'Sullivan and Sherman56} There is no strong evidence supporting any of the treatment modalities; the latest Cochrane review, by Nankivell and Pothier, identified only 2 RCTs with a total population of 71 participants.^{Reference Nankivell and Pothier5} The large number of different classification systems available reflects this lack of robust evidence supporting a specific treatment modality over its alternatives.

The most important question yet to be answered is whether retraction pocket treatment can prevent their progression in terms of grade and cholesteatoma formation. Currently, there is no high-quality evidence supporting such practise; an RCT by Barbara revealed a non-statistically significant benefit of cartilage tympanoplasty compared to a watch-and-wait approach in terms of either the progression of retraction pockets to a worse stage or of cholesteatoma formation.^{Reference Barbara57}

Although surgical management options offer a radical solution compared to the conservative alternatives for tympanic retractions, they carry peri- and post-operative risks. According to Luntz et al., post-operatively the tympanic membrane was normally aerated in two-thirds of cases and was improved in 19 per cent of cases, while in 20 per cent of the cases the tympanic membrane remained unchanged or even deteriorated.^{Reference Luntz, Eisman and Sadé58} Moreover, the authors did not find a statistically significant benefit of cartilage graft over observation in terms of either disease progression or hearing outcomes.^{Reference Nankivell and Pothier5} Overall, the traditional mainstay surgical therapy options are invasive, with increased risks of morbidity, and with no high-quality evidence regarding their effectiveness in terms of disease progression, hearing outcomes or recurrence rates.

Novel surgical techniques

The relatively low benefits offered by surgical intervention, and the increased morbidity, highlight the need for less invasive alternatives.^{Reference Saunders59} We will illustrate three novel techniques.

Glue tympanoplasty can be applied under local anaesthesia, in non-atelectatic grade I–III retractions. The approach is via a 0- or 30-degree endoscope. The retraction pocket is lateralised and lifted out with suction. Fibrin tissue glue is applied over it and it is glued onto the tympanic membrane surface. Fibrin tissue glue has been used in endoscopic myringoplasty cases, with excellent hearing outcomes and perforation closure rates.^{Reference Usami, Iijima, Fujita and Takumi8,Reference Li, Liang, Cheng, Zhang, Ren and Sheng60} At the moment, no published large series have reported on the outcomes of glue tympanoplasty, but our experience so far has shown improved hearing outcomes and tympanic membrane stability, with no recurrence of the retraction pocket within a 12-month follow-up period (Figure 3). Currently, there are no long-term follow-up data for these patients, thus we cannot assess the longevity of this procedure's outcomes.

Fig. 3.

Retraction pocket managed with fibrin glue: (a) pre-operatively and (b) 12 months post-operatively.

The application of laser in tympanoplasty was first described by Goode in cadaveric material. The author reported that laser use led to the architectural remodelling of the middle fibrous layer and the fusion of tympanic membrane collagen fibres.^{Reference Goode61} The histology of the atelectatic tympanic membrane suggests a thinning of the collagen layer, along with obliteration of the lamina propria.^{Reference Stenfeldt, Johansson and Hellström3} The laser beam contracts the tympanic membrane, by heating the collagen fibres, leading to their fusion and remodelling. Different wavelengths have been tried, including those of holmium, carbon dioxide and potassium-titanyl-phosphate (KTP) lasers.^{Reference Ryan and Kaylie6,Reference Brawner, Saunders and Berryhill9,Reference Kurokawa and Goode62,Reference Kauvar and Geronemus63} Usual complications include microscopic or macroscopic perforations, carbonisation, and coagulation. In animal models, the threshold for observed complications was 8 kW/cm² power density. Most authors tend to avoid using a beam more concentrated than 12 W/cm².^{Reference Ryan and Kaylie6,Reference Brawner, Saunders and Berryhill9,Reference Goode61,Reference Kurokawa and Goode62,Reference Nair, Gildener-Leapman and Parnes64}

Although no study so far has compared different types of laser, in our practice we prefer the KTP laser, as it inflicts less thermal damage in the deeper middle-ear structures.^{Reference Kamalski, Verdaasdonk, De Boorder, Vincent, Trabelzini and Grolman65} Defocused KTP laser beams are applied across the surface of the retraction pocket in bursts of 200 ms, at 1 W. The long-term effects of laser application on the tympanic membrane are related to re-epithelialisation of the affected area and the stimulation of fibroblasts, to produce new collagen.^{Reference Kauvar and Geronemus63,Reference Ostrowski and Bojrab66}

Ryan and Kaylie showed that laser tympanoplasty is feasible through a flexible hand-held laser, operating with a direct line-of-sight.^{Reference Ryan and Kaylie6} We followed this principle in our practice, pairing the accuracy of a hand-held laser with the better visualisation achieved by an endoscope. This allows us to accurately apply laser in otherwise difficult-to-reach areas of the tympanic membrane. The technique of endoscopic laser tympanoplasty described offers the flexibility with all the safety precautions for the best patient outcome, and can be applied in an out-patient department setting. Laser tympanoplasty can be undertaken under local anaesthesia in patients who are co-operative, while general anaesthetic remains an alternative option depending on patients' tolerance and preference.

Latest developments in micro-instrumentation have resulted in micro-endoscopes, which are the next step in the evolution of otoendoscopy, and have recently been replacing microscope-otoscopy.^{Reference Marchioni, Mattioli, Alicandri-Ciufelli, Molteni, Masoni and Presutti11,Reference Anschuetz, Presutti, Marchioni, Bonali, Wimmer and Villari67} So far, the application of micro-endoscopes is limited to the delivery of topical medications.^{Reference Mood and Daniel68} The next step is the design of potent micro-endoscopes for the direct investigation of middle-ear ventilation pathways through grommets. Although currently in the early testing phase, micro-lasers are being developed, promising the exploration of and procedures within the middle ear, without mobilising the tympanic membrane. So far, microexploration of the middle ear has been achieved via transnasal, trans-Eustachian access using a robotic endoscope.^{Reference Fichera, Dillon, Zhang, Godage, Siebold and Hartley7,Reference Gafford, Freeman, Fichera, Noble, Labadie and Webster69} Currently there are no series published, but the cadaveric and three-dimensional models demonstrate an efficient and minimally invasive approach to the middle ear, allowing for the manipulation and excision of local pathology, without compromising the tympanic membrane or ossicular integrity.