Introduction

The repair of an eardrum perforation has been the milestone operation in otology, since the first surgical attempts in the field of ear surgery. The tympanoplasty is a well-described procedure, widely performed all over the world, aiming to reconstruct the eardrum and contribute to a well-aerated, healthy, and hearing middle ear. The myringoplasty (type I tympanoplasty) in particular was first described by Berthold in 1878, and since then numerous surgical techniques have been developed, and various graft materials have been used for repairing the tympanic membrane defect [1]. Indeed, skin, fascia, vein, perichondrium, and dura mater have all been employed in tympanic membrane reconstruction [2–8]; however, temporalis fascia represents the most widely used grafting material [9].

The situation gets more complex, and failure rates are considered higher in cases of Eustachian tube dysfunction, retraction pocket, adhesive otitis media, and subtotal or total perforation. Therefore, graft materials more rigid than fascia (i.e., cartilage), and more resistant to infection, resorption, and retraction have been proposed as more appropriate for tympanic membrane reconstruction [10–13]. However, the increased thickness, stiffness, and mass of cartilage [14] may negatively influence the integration of the graft and the hearing results.

The aim of the present study was to assess the existing evidence in favor of or against cartilage type I tympanoplasty in comparison with temporalis muscle fascia (TMF) myringoplasty, with regard to graft integration rates and hearing results. The respective complications from the use of these grafting materials in type I tympanoplasties will also be explored.

Materials and methods

An extensive search of the literature was performed in Medline, Scopus, EMBASE, and CINAHL up to February 2012, having as primary end-points the comparison of hearing results and graft integration rates in patients who had undergone type I tympanoplasty using either cartilage, or TMF for the reconstruction of the tympanic membrane. The number of studies initially selected was 111.

Using this framework of results, the retrieved studies were critically appraised, according to evidence-based guidelines for the categorization of medical studies (Tables 1, 2, 3) [9, 14–25]. Language restrictions limited the included literature to English-speaking articles. Forty studies continued to meet the defined criteria, and were further analyzed.

Table 1 Levels of evidence regarding the primary research question in studies that investigate the results of a treatment (http://www.cebm.net/index.aspx?o=1025)
Full size table
Table 2 Strength of recommendation by category of evidence for guideline development [40]
Full size table
Table 3 Cartilage versus temporalis muscle fascia in type I tympanoplasty
Full size table

During the search the keywords “tympanic”, “membrane”, “perforation”, “graft”, “success”, “hearing”, “gap”, “tympanoplasty”, “myringoplasty”, “cartilage”, and “fascia” were utilized. The keywords “tympanoplasty”, “myringoplasty”, “cartilage”, and “fascia” were considered primary, and were either combined to each of the other keywords individually, or used in groups of three. In addition, reference lists from the retrieved articles were manually searched.

Patients with history of ossicular discontinuity, ossiculoplasty, cholesteatoma, previous ear surgery, or syndromes affecting the status of the middle ear were excluded.

The meta-analysis of data was carried out in the StatsDirect statistical software, and the Random-Effects Model was used to assess the pooled proportion of success and the pooled proportion of re-perforation in the cartilage and fascia groups. Statistical importance was accepted at the level of 0.05.

Results

Among the 40 analyzed studies, four represented prospective randomized studies, two were prospective studies, 14 were retrospective comparative studies, and 15 retrospective studies. There were also two systematic reviews and three books.

Eighteen studies directly compared cartilage and fascia in type I tympanoplasties [9, 14–30]. Among these studies, two were incorporated in a larger patient series by the same principle author, and were not included in the analysis of pooled data to avoid double-counting of the operations [26, 27]. Three more studies also included patients with cholesteatoma [28–30]. In the absence of clear-cut data referring only to patients with type I tympanoplasty without cholesteatoma, these studies were also not used in the analysis of pooled data. Finally, one study exclusively included patients with revision type I tympanoplasties, and was further excluded to avoid sample heterogeneity [14].

From the remaining 12 studies, one represented level I, one level II, and ten level III evidence. The total number of treated patients was 1,286. Cartilage reconstruction was used in 536 type I tympanoplasties, whereas TMF in 750. The mean graft integration rate in the cartilage group was 92.4 % (95 % CI 87.8–96.0) and in the temporalis fascia group 84.3 % (95 % CI 76.9–90.5). The difference proved statistically significant (p < 0.05).

With regard to the functional outcomes of the operations, two level III studies showed a significant difference between the pre- and postoperative air-bone gap closure, in favor of the cartilage grafting materials, and an additional level III study improved results in the cartilage group in cases of subtotal and total TM perforation, and better postoperative air-bone gap closure with fascia in cases of central perforation.

The majority of treated patients were followed up for over 1 year. The rates of re-perforations were 7.6 (95 % CI 4.03–12.2) and 15.5 % (95 % CI 8.9–23.6) for cartilage and fascia, respectively (p < 0.05). The rest of the reported complications are summarized in Table 4.

Table 4 Complications of type 1 tympanoplasty
Full size table

Discussion

TMF is widely used for the reconstruction of tympanic membrane perforations, with generally satisfying results. The fascia is flexible and has more or less the same thickness as a tympanic membrane (when properly prepared) [26]. It is also easily accessible, available in sufficient size, and can be trimmed to the desired dimensions. However, TMF is composed of irregularly arranged elastic fibers and fibrous connective tissue. Hence, it may demonstrate radical and unpredictable changes in shape, shrinking, or even thickening postoperatively [31].

Unlike fascia, cartilage demonstrates higher mechanical stability [32], considerable stiffness, and slower metabolism, and can therefore be considered a reliable grafting material [10, 18, 33]. Cartilage has a constant shape, it is firmer than fascia, lacks fibrous tissue [10], but shows high concentration of the highly resistant protein elastin [23]. These features help the postoperative dimensions of the graft to remain the same, and cover large perforations with stability. Moreover, at least in theory, cartilage grafting may prevent retraction pockets [21] and re-perforations, which may follow episodes of acute otitis media. Finally, harvesting cartilage graft is not more difficult than fascia, whether it is taken from the concha or the tragus [34–36].

However, concerns had been previously expressed that the rigid nature of the cartilage may theoretically impede with the sound-conducive properties of the tympanic membrane [14, 23, 37]. Indeed, Zahnert et al. [38] suggested that the ideal acoustic thickness of cartilage should be approximately 0.5 mm, instead of the standard full thickness cartilage graft (0.7–1 mm thick), to achieve optimal hearing results. However, thinning the cartilage makes the reconstruction process more difficult due to the most probable twisting of the cartilage. Hence, precise placement of the cartilage pieces, as well as a reduction in the number of palisades is required to successfully apply this technique [20]. Atef et al. [39] also concluded that slicing the cartilage to half its normal thickness added to the technical difficulties of the procedure without making a significant difference to the hearing gain, after analyzing the effect of cartilage disc thickness on hearing results following perichondrium–cartilage island flap tympanoplasty.

The present study, taking into account the results of over 1,000 patients and applying strict inclusion criteria, demonstrated that the graft integration rate is higher in the cartilage compared with the temporalis fascia group in type I tympanoplasty (p < 0.05; Figs. 1, 2). Indeed, 9 out of 12 analyzed studies reported a success rate of over 90 % in the cartilage patient group (Fig. 3). In contrast, most fascia patient groups had lower success rates, and only three fascia groups reported a success rate of over 90 % (Fig. 3). It should be noted that the majority of the analyzed studies were level III, with only one study representing level II, and one study representing level I evidence. The results from the level I study were suggestive of at least non-inferiority of cartilage compared with fascia in type I tympanoplasties, whereas the quality of evidence from the remaining studies allows us to adopt a grade C strength of recommendation regarding the effectiveness of cartilage versus fascia in type I tympanoplasties (Table 2).

Fig. 1
figure 1

Raw data for the success rate of cartilage versus temporalis fascia grafting in type 1 tympanoplasty (patient series)

Full size image
Fig. 2
figure 2

Proportion meta-analysis plot for the success rate of cartilage versus temporalis fascia grafting in type 1 tympanoplasty (patient series). Temporalis fascia tympanoplasties are depicted on the right, and cartilage tympanoplasties on the left side of the graph

Full size image
Fig. 3
figure 3

Success rates in type 1 tympanoplasty by grafting material

Full size image

It should also be noted that two level III studies showed a significant difference between the pre- and postoperative air-bone gap closure, in favor of the cartilage grafting materials. Yetiser et al. [9] showed a statistically different postoperative air-bone gap of 14.2 ± 7.7 dB in the cartilage group as compared to 19.7 ± 12 dB in the fascia group (p = 0.008). Similar results were reported by Onal et al. [17], who found a mean postoperative air-bone gap of 12.08 ± 6.71 dB for the fascia group, and 9.33 ± 4.74 dB for the cartilage group. The difference was statistically significant (p = 0.027), even though the respective air-bone gap between the two graft materials did not differ pre-operatively (p = 0.572). However, the better hearing outcomes in these two studies should be weighed against the respective results of nine other studies, which did not report any statistically significant differences between the two methods of reconstruction, and an additional study, which showed improved results in the cartilage group in cases of subtotal and total tympanic membrane perforation, and better postoperative air-bone gap closure with fascia in cases of central perforation. Furthermore, most studies report statistically significant difference between the pre- and postoperative air conduction thresholds within but not between the cartilage and fascia groups (Table 5). The aforementioned discrepancies preclude us from drawing definite conclusions regarding the strength of the respective recommendations.

Table 5 Audiometric results in type I tympanoplasty (cartilage vs. temporalis muscle fascia)
Full size table

Previous studies had also reported a graft integration rate of 90–95 % for the first year after a type I tympanoplasty, and a rate of re-perforation of 10–15 % over the next 3–10 years [17, 18]. However, the results of the present study suggest that the relatively high rates of re-perforation may only apply for fascia myringoplasties (15.5 %), as the respective results in the cartilage group were found significantly lower (7.6 %, p < 0.05). Among the other complications of type I tympanoplasty, retraction pockets, otitis media with effusion, anterior blunting, and graft lateralization are usually surgically managed, whereas most of the rest are minor and can be dealt with conservatively (Table 4).

It should be mentioned that despite the calls for a wider use of cartilage in type I tympanoplasties [16, 18, 20], this grafting material is primarily used in cases of Eustachian tube dysfunction, adhesive otitis media, and subtotal perforation. Hence, there can be a selection bias when forming the respective comparison groups. The strict selection criteria in the present study limited the possibility of such bias, because all included patients had been randomly assigned to either grafting material, according to the detailed analysis of the materials and methods section of each included study. Based on the available data, the utilization of cartilage for the reconstruction of the tympanic membrane in type I tympanoplasties can, thus, be recommended.

Conclusion

The use of cartilage in type I tympanoplasty is associated with higher graft integration rates as compared to fascia reconstructions (grade C strength of recommendation). In addition, the obtained audiometric results appear to be at least comparable, and the rate of re-perforation is lower.

Although cartilage is primarily used as grafting material in cases of Eustachian tube dysfunction, adhesive otitis media, and subtotal perforation in everyday surgical practice, a wider utilization for the reconstruction of the tympanic membrane in myringoplasties can be considered.