Discussion

The first documented attempt at repairing a tympanic membrane perforation was performed in 1640, by Banzer, using a pig's bladder.^{Reference Benzer7} Similar documented attempts were made by Toynbee in 1853 using a rubber disc with silver wire,^{Reference Toynbee8} and by Blake in 1887 using a paper patch.^{Reference Blake9} However, Berthold is credited with performing the first true tympanoplasty, as he applied plaster over the tympanic membrane remnant for de-epithelisation and then applied a skin graft.^{Reference Berthold10} All tympanoplasty techniques are based on the classic concept of the middle-ear transformer mechanism put forward by Helmholtz in 1868.^{Reference Helmholtz11}

Tympanoplasty is broadly divided into overlay and underlay techniques, with each having its own advantages and disadvantages. The over-underlay tympanoplasty technique was introduced as a means of incorporating the advantages of both overlay and underlay techniques, whilst avoiding the disadvantages.^{Reference Kulduk, Dundar, Soy, Guler, Yukkaldiran and Iynen12} It is relatively simple and straightforward, with a good success rate.^{Reference Sarkar1} The handle of the malleus remains an asset in this technique and prevents atelectasis. According to Kartush et al., one vital requirement in this technique, like the overlay method, is the complete removal of epithelium from the handle of the malleus.^{Reference Kartush, Michaelides, Becvarovski and Larouere13} Many consider this impossible to achieve surgically.^{Reference Hough14}

Nejadkazem and colleagues’ study on intratympanic cholesteatoma after tympanoplasty stated that the typical site of these cholesteatomas is near the umbo. They suggested that incomplete removal of epithelium from the umbo is the cause.^{Reference Nejadkazem, Totonchi, Naderpour and Lenarz15} This is not exclusively the result of a lapse by the surgeon, as normal tympanic membrane anatomy plays a crucial role too. The tympanic membrane is intimately connected to the handle of the malleus, especially the umbo, and serves as one of the main supports of the malleus.^{Reference Gulya16} According to Tos, the umbo represents the only site where the handle of the malleus attaches to the tympanic membrane.^{Reference Tos17} Graham et al. also reported that the umbo maintains a uniform distribution of fibrous tissue on its medial and lateral surfaces.^{Reference Graham, Reams and Perkins18} According to the authors, these fibres are intimately attached to the perichondrium surrounding the manubrium over the lower third. Furthermore, optical coherence tomography has also shown the tympanic membrane to have naturally increased thickness (100–120 µm) near the umbo.^{Reference Van der Jeught, Dirckx, Aerts, Bradu, Podoleanu and Buytaert19}

Studies have demonstrated that the experience and skill set of the surgeon determines the ultimate outcome in tympanoplasty, irrespective of other clinical variables.^{Reference El-Seifi and Fouad20} According to Emir et al., the outcomes of tympanoplasty are more unsatisfactory when performed by trainees, and practical guidance from a senior surgeon is essential for better post-operative results.^{Reference Emir, Ceylan, Kizilkaya, Gocmen and Uzunkulaoglu21} It is our opinion that residents, with their limited experience and skills, are more likely to leave residual epithelium over the handle of the malleus in tympanoplasty.

Intentional placement of graft material medial to the handle of the malleus can reduce the chances of epithelial entrapment in tympanoplasty in such situations and avoid unwanted ossicular mobilisation. During the course of this study, the utilisation of other methods were suggested, including using lasers for the removal of tissues over the umbo, to ensure complete de-epithelisation.^{Reference Kartush, Michaelides, Becvarovski and Larouere13} This circumvents the excessive ossicular mobilisation involved in de-epithelisation that may affect ossicular chain stability.^{Reference Mahadevaiah and Parikh6} However, the apparent lack of proper techniques for epithelium identification and the cost factor dissuaded us from it.

When the likelihood of residual epithelium over the umbo is significant, the surgeon can resect the umbo^{Reference Mahadevaiah and Parikh6,Reference Kutluhan, Yalçıner, Güler, Kösemehmetoğlu, Bozdemir and Bilgen22} and conveniently place the graft medial to the handle of the malleus.^{Reference Sanna, Sunose, Mancini, Russo and Taibah5} In our study, 13 specimens (30.95 per cent) had residual epithelium on histopathological examination. Selective removal of the umbo and exteriorisation of the malleus should have sufficiently reduced the likelihood of cholesteatoma in these cases. However, intra-operative assessment was accurate in only two cases. Thus, if the graft was resting over the handle of the malleus in the remaining 11 cases, it could have led to epithelial entrapment. In this study, 29 specimens were negative for epithelium on histopathological examination. However, intra-operative findings were misleading in 14 of these cases, naturally leading to inaccurate positive reporting for epithelium. Therefore, according to the first author, the two cases of correct identification could have occurred unintentionally.

However, our study has some inadequacies. First, the viability of the detected epithelium is unascertained; in other words, whether epithelial pearl or cholesteatoma would have developed in these cases is unknown. There are published studies utilising over-underlay tympanoplasty that did not show any epithelial pearl formation on follow up.^{Reference Khalifa and Khalifa23} Because of the comparatively brief duration of our study, the incidence of epithelial pearl or cholesteatoma in our study group remains unascertained. The second and most significant limitation is the inadequate sample size. The required sample size was difficult for a lone surgeon to realistically achieve in the comparatively short study period. A multicentre meta-analysis is required.