AlloDerm tympanoplasty of tympanic membrane perforations☆

Materials and methods 

Seventeen adult female chinchillas (Moulton Chinchilla Ranch, Rochester, MN), weighing approximately 400 to 700 g, were chosen for this study. Each chinchilla had bilateral chronic tympanic membrane perforations created as previously described.16, 17, 18 A thermal myringotomy loop (Xomed, Jacksonville, FL) was used to perform a myringectomy of approximately 30% to 50% of the pars tensa creating an inferior central perforation. The mucosal layer of the TM around the perforation was abraded with an otologic hook. Four radial incisions were made in the perforation margin, producing 4-minute flaps that were infolded to coapt the previously roughened mucosal layer.

Postoperatively, the chinchillas were followed for 6 weeks. Otomicrosopic examinations were performed weekly to assess for healing and any signs of infection. After this stage, the animals with mature perforations lacking signs of epithelial regeneration and infection were selected for tympanoplasty.

Using sterile technique, the epithelial margin of the perforation was removed in the usual fashion. A tympanomeatal flap was elevated through a postauricular approach. The middle ear space was examined and then loosely packed with saline soaked Gelfoam (UpJohn Company, Kalamazoo, MI).

The experimental AlloDerm graft (300 μm or 0.3 mm thickness) was cut to size dry, partially rehydrated, and placed as a medial graft with the BMC facing laterally. Fascia grafts were harvested from the lumbar region because the chinchilla has no fascia in the periauricular or temporal region.19 The tympanomeatal flap was returned to its original position, and the external auditory canal was loosely packed with gelfoam. The postauricular and lumbar incisions were closed. Operative time was recorded during the repair of each TM.

At 4 and 10 weeks after the tympanoplasty stage, otomicroscopic examinations were performed. Documentation was made of any presence of a perforation, cholesteatoma, infection, granulation tissue, myringosclerosis, and thickened tympanic membrane.

Euthanasia was performed 10 weeks after the tympanoplasties. The TMs, still attached to the bony tympanic annulus, were harvested and fixed in 10% buffered formalin for 24 hours. After decalcification with formic acid for approximately 4 hours, the samples were routinely processed, embedded in paraffin, sectioned, and stained with hematoxylin and eosin. Multiple levels were available for histologic evaluation. Blinded to the specimens, 2 pathologists examined the tissue documenting any presence of microperforations, epithelial hyperplasia (> 8 cell layers), infiltrates (fibroblasts, lymphocytes, and neutrophils), granulation tissue, myringosclerosis, fibrosis, cholesteatoma, basement membrane complex, and neovascularization (endothelial lined blood vessels). Approximate counts per 5 high-power fields (HPFs) were performed on the cellular infiltrates and blood vessels. The thickness of all grafts was measured with a microscopic micrometer. These features were chosen after reviewing the literature on the histopathology of the TM perforation and replacement membrane.20, 21, 22, 23

Anesthesia for the creation of the TM perforations and all examinations under anesthesia was provided with isoflurane application via a nose cone while each animal was spontaneously ventilating. During the tympanoplasty, anesthesia included dose per body weight intramuscular (IM) ketamine (20 mg/kg), IM acepromazine (0.5 mg/kg), and subcutaneous atropine (0.05 mg/kg) after isoflurane induction and endotracheal intubation. Isoflurane was continued throughout the tympanoplasty.

Buprenorphine, 0.05 mg/kg body weight, was given IM preoperatively and for 48 hours postoperatively for the perforation and tympanoplasty procedures. Analgesia was extended as deemed necessary by the attending veterinarian.

Antibiotic prophylaxis consisted of trimethoprim 40 mg/mL combined with sulfadiazine 200 mg/mL at a dose of 0.125 mL/kg injected IM preoperatively and for 5 days twice daily after both the tympanic membrane perforations and the tympanoplasty procedures. In addition, the same antibiotic therapy in oral suspension form was extended for 3 weeks posttympanoplasty. Topical gentamicin drops were applied each day to the external auditory canal for 1 week before the examination under anesthesia 4 weeks after the tympanoplasty. Euthanasia was achieved by an intracardiac injection of an overdose of phenobarbital.

This study was performed in accordance with the PHS Policy on Humane Care and Use of Laboratory Animals, the NIH Guide for the Care and Use of Laboratory Animals, and the Animal Welfare Act (7 U.S.C. et seq.). The Institutional Animal Care and Use Committee of Madigan Army Medical Center approved the animal use protocol.

Results 

Of the 34 perforations originally made in 17 chinchillas, 22 (65%) perforations were still present at 6 weeks or immediately before the tympanoplasty repair stage. Average perforation size was approximately 30% to 40% of the area of the pars tensa. Eleven perforations were repaired with AlloDerm, and the remaining 11 perforations were repaired with autologous lumbar fascia.

Mean tympanoplasty time for the AlloDerm group was 47 minutes (SD = 11.9). For the fascia group, the mean operative time was 68 minutes (SD = 13.4). A statistically significant difference was found between the surgical times of these 2 groups (P = .001) by using a 2-sample, 2-sided t test not assuming equal variances.

At the study termination, 90% (9 of 10) of the perforations treated with AlloDerm were successfully repaired. The single AlloDerm failure was secondary to a small portion of the edge of the graft displaced laterally through the perforation. Otherwise, the remainder of the AlloDerm graft grossly appeared to be well incorporated into the host TM. All (9 of 9) of the fascia grafted TMs had complete closure of the perforation. Final subject numbers were reduced because of the removal of 2 animals because of postauricular wound complications.

Gross otomicroscopic examinations at 4 and 10 weeks after tympanoplasty showed no apparent differences between the 2 groups. Thinning of both grafts occurred during the postoperative course. Epithelial migration and neovascularization was apparent over the surfaces of both grafts. Prominent blood vessels were present at 4 weeks, but as healing occurred, the vessels became less noticeable (Fig 1A-D)

View Large Image Download to PowerPoint

Fig. 1. Photographs of the inferior one half of the tympanic membrane after tympanoplasty. AlloDerm at (A) 4 weeks and (B) 10 weeks. Fascia at (C) 4 weeks and (D) 10 weeks. Notice the blood vessels overlying the grafts at 4 weeks; the vessels are less noticeable by the 10th week. Graft thinning is also apparent between 4 and 10 weeks.

Histological analysis was performed only on those animals surviving the length of the experiment. There was no presence of microperforations, epithelial hyperplasia, myringosclerosis, fibrosis, or granulation tissue with any of the grafts. Inflammatory infiltrates were absent in 3 AlloDerm and 4 fascia grafts. Five AlloDerm and 3 fascia grafts had less than 50 lymphocytes per 5 HPFs, whereas 2 AlloDerm and 2 fascia grafts showed evidence of 50 to 100 lymphocytes per 5 HPFs. No neutrophils were identified in either of the groups. Fibroblast infiltration was detected in 100% of the AlloDerm grafts, whereas 89% of the fascia grafts showed the presence of fibroblasts. When present, fibroblast population was approximately 200 per 5 HPFs in each graft. Basement membrane complexes were maintained in 80% of the AlloDerm grafts. Eighty-nine percent of the fascia grafts showed the presence of a basement membrane complex. Neovascularization was seen in 80% of the AlloDerm-treated TMs, almost double that found in the fascia grafts (44%). Blood vessel counts for both grafts were estimated at 5 vessels per 5 HPF (Table 2, Fig 2).

Table 2. Findings between experimental and control groups

		AlloDerm	Fascia
	Tympanoplasty time (minutes)*	47	68 P = .001†
	Successful closure of TM perforation (%)	90	100
	Total # of TM's with infiltrates/5 HPF
	Absent	3	4
	<50 lymphocytes	5	3
	50-100 lymphocytes	2	2
	Neutrophils	0	0
	Presence of fibroblasts (%)	100	89
	Presence of basement membrane complex (%)	80	89
	Presence of blood vessels (%)	80	44
	Mean thickness
	Preoperative	300 μm	—
	Postoperative	51 μm	23 μm
	*Tympanoplasty time is for n = 11 in both groups. All other categories contain n = 10 for AlloDerm group and n = 9 for the fascia group. †Student t test.

Abbreviations: TM, tympanic membrane; HPF, high power fields.

View Large Image Download to PowerPoint

Fig. 2. Tympanic membrane with AlloDerm graft (H&E stain, original magnification × 200). Lateral surface at top and medial surface at bottom of photograph. Disorderly arranged collagen fibers occupy the medial two thirds of the TM. Fibroblast infiltration and neovascularization is appreciated, as well as epithelial and mucosal cell coverage on opposing surfaces of the graft.

Graft thickness of the AlloDerm before implantation was 300 μm. At study end, the mean thickness of the AlloDerm was 51 μm or 17% of its original thickness. Fascia grafted TMs were thinner with an average thickness of 23 μm.

Discussion 

The ideal grafting material in tympanoplasty has long been debated. This material should be reliable, have a high success rate and low donor site morbidity, and be obtained in the same operative field. The graft should serve to provide a scaffold by which the epidermal layer can proceed to close the defect. This includes the ability to be infiltrated by mesenchymal cells and undergo neovascularization while a condensation of fibrous elements forms between the migrating epithelium and middle ear mucosa.

Autografts, allografts, xenografts, and alloplasts have all been used with varying success rates in tympanoplasty. The most common graft used today is autologous temporalis fascia. When performing a postauricular tympanoplasty, it can be obtained through the same incision with little added morbidity to the patient. However, in transcanal approaches to tympanoplasty or in revision cases in which the graft has been previously harvested, the graft must be obtained through an additional surgical incision or extended dissection with its associated surgical morbidity.

AlloDerm has been examined previously as a substitute grafting material in both animal and human studies. McFeely et al14 and Laidlaw et al15 used the same animal model as we did to compare grafting with AlloDerm to autologous temporalis fascia and rice paper, respectively. Both performed medial grafting with the allograft placed through the TM perforation. McFeely et al14 showed a perforation closure rate of 80% with AlloDerm and 90% with periauricular fascia. Laidlaw et al15 reported complete closure with AlloDerm in 78% of the subjects compared with 66% with rice paper.

Our study showed slightly better success rates of perforation closure at 90% with AlloDerm and 100% with autologous lumbar fascia. We found no useful fascia in the chinchilla's periauricular region. Therefore, we harvested lumbar fascia, which is strikingly similar to human temporalis fascia in its thickness as well as harvesting and handling properties. This fascia, along with placing all grafts medially through a tympanomeatal flap, may have produced more successful repairs in both the control and experimental groups. Because rice paper is not of connective tissue origin and because it was used as a lateral graft may have led to more failures in Laidlaw et al's study.15 More subject numbers would be necessary to determine a statistically significant difference in perforation repair rates between AlloDerm and autologous tissue grafting in this animal model.

Better perforation closure rates have been reported in human studies after AlloDerm tympanoplasty of chronic TM perforations.12, 13 Also, these reports provide audiological information after at least 6 months of follow-up. Youssef inserted the allograft through the perforation placing it on the medial surface of the TM. Twenty-seven of 30 (90%) adult patients had complete closure of their perforations with an ABG of less than 10 dB in 24 (80%) of those patients. Benecke13 compared AlloDerm and temporalis fascia tympanoplasty with 20 subjects in each group. He reports 100% take of all grafts and average closure of the ABG to approximately 5 dB in both groups. In the AlloDerm group, 17 were revision cases using the lateral grafting technique, whereas the remaining 3 were primary underlay repairs.

Gross inspection of our grafted TMs showed no apparent differences between AlloDerm and fascia. The AlloDerm repaired TMs showed no overt evidence of rejection or extrusion. The allograft appeared to integrate well. Epithelial migration over the lateral surface was apparent by the observation of blood vessels overlying the grafts. Thinning of the allografts was appreciated, but contraction or fragmentation did not occur. The thinning provided some translucency that may have allowed visualization of a cholesteatoma if one developed. These observations are similar to previous reports in the chinchilla model.14, 15

On histopathologic examination, the majority of our experimental AlloDerm grafts showed either no inflammation or only a mild chronic inflammatory response as witnessed by the lymphocytic infiltration. No acute inflammation was found. The inflammatory response was similar in the control fascia group. We discovered no microscopically fibrotic or sclerotic TMs or epithelial hyperplasia. These findings contrast those of McFeely et al, examining only approximately 25% of their allograft TMs, who reported specimens with a “brisk inflammatory response” and “thickened with epithelial...proliferation.”14 Similar to our results, Laidlaw et al15 reported a low incidence of inflammation and fibrosis, but more than one third of their samples were lost because of technical difficulties during the harvest of the TMs.

Fibroblast infiltration and neovascularization was found in more AlloDerm grafts than fascia grafts. The significance of this finding is unknown. The AlloDerm grafts microscopically appeared to have a more unorganized and less dense collagen fiber arrangement than fascia, which may have allowed improved mesenchymal cell infiltration. Basement membrane complexes were maintained in 80% of the AlloDerm grafts. Interestingly, all but 1 fascia graft showed formation of a BMC. All AlloDerm grafts had the BMC placed laterally. Both the lateral and medial surfaces were re-epithelialized and remucosalized, respectively, in all grafts. This supports earlier determinations that orientation of the AlloDerm's BMC is unnecessary when healing is required on both sides of the graft.11, 14 All of these histologic findings show that AlloDerm is well integrated into the host tympanic membrane and does not differ substantially from autologous fascia.

Graft thickness in the AlloDerm group reduced to approximately 17% of its original 300 μm (0.3 mm) thickness. With facial soft-tissue augmentation, AlloDerm has been shown to lose it's bulk by approximately 15% to 70% with more loss noticed in highly mobile areas like the lips.24 The chinchilla TM is not a highly mobile structure; therefore, it is more likely that natural resorption has led to AlloDerm thinning. The final average thickness of 51 μm for AlloDerm in our study is comparable to the thickness of normal (35-104 μm) and spontaneously healed (14-46 μm) chinchilla TMs and normal human TMs (30-90 μm).17, 20, 25

AlloDerm sizes ranging from 0.08 mm to 0.3 mm in thickness have been reported in the literature.12, 13, 14, 15 We chose the 0.3-mm AlloDerm because LifeCell Corporation could guarantee total uniform sheet thickness before implantation for purposes of comparing with postoperative measurements. Youseff12 also used the 0.3 mm thickness AlloDerm obtaining perforation closure in 90% of his patients. However, Benecke13 repaired his perforations with a thinner AlloDerm graft (0.15 mm-0.3 mm) achieving higher rates of perforation closure and better audiological results. The previous animal studies used even thinner grafts ranging from 0.08 mm to 0.127 mm and showed only approximately 80% success in closing TM perforations. The ideal thickness has yet to be determined, but the 0.3-mm AlloDerm displays excellent integration in our study. Nonetheless, resorption will occur and final graft thickness will approximate the natural thickness of the native TM.

We found the 0.3-mm AlloDerm very similar to fascia in its handling characteristics. Our initial tympanoplasty involved rehydrating the AlloDerm over 10 minutes as described by the package insert. Rehydrated AlloDerm tended to be difficult to handle because it became fairly adherent when rolled on itself. We decided to moisten the AlloDerm enough to allow removal of the paper backing while maintaining some rigidity similar to dried fascia. The saline from the gel foam placed in the middle ear space and external auditory canal technically reconstituted the graft. Our study shows that the partially rehydrated AlloDerm is comparable to autologous fascia on clinical and histologic examinations.

Conclusion 

Our study shows that AlloDerm is as successful as autologous fascia at closing chronic tympanic membrane perforations in the chinchilla model. Clinical and histologic evaluations reveal excellent integration of the AlloDerm into the host TM. Mild to no inflammatory response was appreciated with AlloDerm microscopically, which was very similar to autologous fascia. The use of AlloDerm significantly reduced total operative time.

Advantages of AlloDerm tympanic membrane grafting in humans could include (1) avoiding a donor site incision during transcanal tympanoplasty, thus reducing pain and discomfort, scar formation, and the potential for hematoma; (2) reducing surgical time and operative costs; and (3) using the graft in revision cases in which no useful temporalis fascia can be located. Further research will be necessary to investigate Alloderm's reperforation rate, resistance to infection, response to ototopical medications, reaction to eustachian tube dysfunction, and hearing outcomes in long-term follow-up studies.