Objectives/Hypothesis: The purpose of this study was to analyze the anatomical and audiologic results in more than 1,000 cartilage tympanoplasties that utilized a logical application of several techniques for the management of the difficult ear (cholesteatoma, recurrent perforation, atelectasis). Our hypothesis was that pathology and status of the ossicular chain should dictate the technique used to achieve optimal outcome. Study Design: Retrospective clinical study of patients undergoing cartilage tympanoplasty between July 1994 and July 2001. A computerized otologic database and patient charts were used to obtain the necessary data. Methods: A modification of the perichondrium/cartilage island flap was utilized for tympanic membrane reconstruction in cases of the atelectatic ear, for high-risk perforation in the presence of an intact ossicular chain, and in association with ossiculoplasty when the malleus was absent. A modification of the palisade technique was utilized for TM reconstruction in cases of cholesteatoma and in association with ossiculoplasty when the malleus was present. Hearing results were reported using a four-frequency (500, 1,000, 2,000, 3,000 Hz) pure-tone average air-bone gap (PTA-ABG). The Student t test was used for statistical comparison. Postoperative complications were recorded. Results: During the study period, cartilage was used for TM reconstruction in more than 1,000 patients, of which 712 had sufficient data available for inclusion. Of these, 636 were available for outcomes analysis. In 220 cholesteatoma cases, the average pre- and postoperative PTA-ABGs were 26.5 ± 12.6 dB and 14.6 ± 8.8 dB, respectively (P < .05). Recurrence was seen in 8 cases (3.6%), conductive HL requiring revision in 4 (1.8%), perforation in 3 (1.4%), and postand intraoperative tube insertion in 11 (5.0%) and 18 ears (8.2%), respectively. In 215 cases of high-risk perforation, the average pre- and postoperative PTA-ABGs were 21.7 ± 13.5 dB and 11.9 ± 9.3 dB, respectively (P < .05). Complications included recurrent perforation in 9 ears (4.2%), conductive HL requiring revision in 4 (1.9%), postoperative and intraoperative tube insertion in 4 (1.9%) and 6 ears (2.8%), respectively. In 98 cases of atelectasis, the average pre- and postoperative PTA-ABGs were 20.2 ± 10.9 dB and 14.2 ± 10.2 dB, respectively (P < .05). Complications included 1 perforation (1.0%), conductive loss requiring revision in 2 cases (2.0%), and post- and intraoperative tube insertion in 7 (7.1%) and 12 ears (12%), respectively. In 103 cases to improve hearing (audiologic), the average pre- and postoperative PTA-ABGs were 33.6 ± 9.6 dB and 14.6 ± 10.1 dB, respectively (P < .05). Complications included 1 perforation (1.0%), conductive loss requiring revision in 11 (11%), and post- and intraoperative tube insertion in 6 (5.8%) and 2 (1.9%), respectively. Conclusions: Cartilage tympanoplasty achieves good anatomical and audiologic results when pathology and status of the ossicular chain dictate the technique utilized. Significant hearing improvement was realized in each pathological group. In the atelectatic ear, cartilage allowed us to reconstruct the TM with good anatomic results compared to traditional reconstructions, which have shown high rates of retraction and failure. In cholesteatoma, cartilage tympanoplasty using the palisade technique resulted in precise reconstruction of the TM and helped reduce recurrence. In cases of high-risk perforation, reconstruction with cartilage yielded anatomical and functional results that compared favorably to primary tympanoplasty using traditional techniques. We believe the indications for cartilage tympanoplasty (atelectatic ear, cholesteatoma, high-risk perforation) were validated by these results.
Since the introduction of tympanoplasty in 1952 by Zoellner1 and Wullstein,2 numerous graft materials and methods of placement have been described to reconstruct the tympanic membrane (TM). Skin, fascia, vein, perichondrium, and dura mater have all been used for TM reconstruction.3–9 To date, temporalis fascia and perichondrium remain the most commonly employed materials for closure of TM perforations, and successful reconstruction is anticipated in approximately 90% of primary tympanoplasties.10
In certain situations, however, such as the atelectatic ear, cholesteatoma, and revision tympanoplasty, the results have not been as gratifying. In these situations, fascia and perichondrium have been shown to undergo atrophy and subsequent failure in the postoperative period, regardless of placement technique.11–14 These observations have led to the use of less compliant, more rigid grafting materials for TM reconstruction. Although cartilage is similar to fascia in that it is mesenchymal tissue, its more rigid quality tends to resist resorption and retraction, even in the milieu of continuous eustachian tube dysfunction. It has been shown in both experimental and clinical studies that cartilage is well tolerated by the middle ear, and long-term survival is the norm.15–21
The use of cartilage in middle ear surgery is not a new concept, but the last decade has shown a renewed interest in this material as an alternative to more traditional grafting materials for TM reconstruction. Cartilage was first introduced in middle ear surgery in 195922 and has been described for the limited management of retraction pockets13,19,23–28 and, more recently, for the reconstruction of the tympanic membrane in cases of recurrent perforation, with encouraging results.14,29,30 Despite the thickness of the grafts, the hearing results appear to be good.14,29–33
The purpose of this clinical study was to analyze the anatomical and audiologic results in more than 1,000 cartilage tympanoplasty cases that utilized a logical application of several techniques that were developed or were modified by the author for the management of the difficult ear (cholesteatoma, recurrent perforation, atelectasis). Development and application of these techniques was based on the hypothesis that pathology and status of the ossicular chain should dictate the technique used for cartilage tympanoplasty to optimize outcome.
MATERIALS AND METHODS
The techniques described in this section reflect our current approach, which is based on pathological presentation, for both adult and pediatric patients seen at the University of Arkansas for Medical Sciences and Arkansas Children's Hospital, Little Rock, Arkansas, for cartilage reconstruction of the TM.
General Principles. Generally speaking, there are two distinct techniques utilized for cartilage reconstruction of the TM: the perichondrium/cartilage island flap and the palisade technique. Either of these techniques may be utilized, depending on the specific situation, surgical approach, and type of TM pathology. This section will describe each technique in general terms, with decision making and modification of the basic surgical techniques based on the specific situation addressed later.
Cartilage for TM reconstruction is typically harvested from two areas, the tragus and the cymba, depending on the type of reconstruction to be performed. The perichondrium/cartilage island flap is nearly always constructed with cartilage harvested from the tragus. This cartilage is ideal because it is thin, flat, and in sufficient quantities to permit reconstruction of the entire TM. The cartilage, which is typically slightly less than 1 mm thick in most cases, is used as a full-thickness graft. Using the Doppler interferometer, Zahnert et al.34 suggested a slight acoustical benefit could be obtained by thinning the cartilage to 0.5 mm, but this advantage is offset by the unacceptable curling of the graft that occurs when the cartilage is thinned and perichondrium is left attached to one side.
With the palisade technique, cartilage from either the tragus or the cymba can be used for reconstruction. Cartilage from the cymba area of the conchal bowl is used if the surgical approach involves a postauricular incision, and tragal cartilage is used if the approach is transcanal or endaural. The cartilage of the cymba is similar to tragal cartilage in that it has an acceptable thickness of about 1 mm compared to other areas of the concha, which are thicker and irregular. The cymba cartilage is different, however, in that it is curved, making it difficult to create a perichondrium/cartilage island flap suitable for reconstruction of the entire TM.
In summary, tragal cartilage is used for the perichondrium/cartilage island flap, whereas either tragal or cymba cartilage can be used for the palisade technique, depending on the surgical approach.
Perichondrium/Cartilage Island Flap. The general technique of reconstruction using the perichondrium/cartilage island flap begins with harvest of the cartilage from the tragal area.31 An initial cut through skin and cartilage is made on the medial side of the tragus, leaving a 2-mm strip of cartilage in the dome of the tragus for cosmesis (Fig. 1). The cartilage, with attached perichondrium, is dissected medially from the overlying skin and soft tissue by spreading a pair of sharp scissors in a plane that is easily developed superficial to the perichondrium on both sides. At this point, to maximize the length of harvested cartilage, it is necessary to make an inferior cut as low as possible. The cartilage is then grasped and retracted inferiorly, which delivers the superior portion from the incisura area. The superior portion is then dissected out while retracting, which produces a large piece of cartilage, typically 15 mm in length and 10 mm in width in children and somewhat larger in adults.
The perichondrium from the side of the cartilage furthest from the ear canal is dissected, leaving the thinner perichondrium on the reverse side. A perichondrium/cartilage island flap, as described by other authors, is constructed with some notable modifications that allow better fit in the middle ear. Using a round knife, cartilage is removed to produce an eccentrically located disc of cartilage about 7 to 9 mm in diameter for total TM reconstruction. A flap of perichondrium is produced posteriorly that will eventually drape over the posterior canal wall. Next, a complete strip of cartilage 2 mm in width is removed vertically from the center of the cartilage to accommodate the entire malleus handle (Fig. 2). The creation of two cartilage islands in this manner is essential to enable the reconstructed TM to bend and conform to the normal conical shape of the TM. When the ossicular chain is intact, an additional triangular piece of cartilage is removed from the posterior-superior quadrant to accommodate the incus. This excision prevents the lateral displacement of the posterior portion of the cartilage graft that sometimes occurs because of insufficient space between the malleus and incus.
The entire graft is placed in an underlay fashion, with the malleus fitting in the groove and actually pressing down into and conforming to the perichondrium, as shown in Figure 3. The cartilage is placed toward the promontory, with the perichondrium immediately adjacent to the TM remnant, both of which are medial to the malleus. Failure to remove enough cartilage from the center strip will cause the graft to fold up at the center instead of lying flat in the desired position. Likewise, if the strip is insufficient, the cartilage may be displaced medially instead of assuming a more lateral position in the same plane as the malleus.
Gelfoam (Upjohn Laboratories, Kalamazoo, MI) is packed in the middle ear space underneath the anterior annulus to support the graft in this area, and the posterior flap of perichondrium is draped over the posterior canal wall. Middle ear packing is avoided on the promontory and in the vicinity of the ossicular chain. One piece of Gelfoam is placed lateral to the reconstructed TM, and antibiotic ointment is placed in the ear canal (Fig. 4).
Palisade Technique. In this technique, the cartilage is cut into several slices that are subsequently pieced together to reconstruct the TM (Fig. 5), using a method that is a modification of that described by Heermann et al.35 Because of the nature of the reconstruction, it is not necessary to have one large, flat piece of cartilage, and the more curved cymba cartilage is suitable (See Fig. 6). The cymba is harvested from the postauricular incision. A large area of conchal eminence can be exposed by elevating the subcutaneous tissue and postauricular muscle from the conchal perichondrium. The cymba cartilage is the prominent bulge at the superior aspect of the concha (Fig. 7). A circumferential cut the size of the anticipated graft is made through the perichondrium and cartilage but not through the anterior skin. The perichondrium is removed from the postauricular side, and the cartilage with the perichondrium on the anterior aspect is dissected from the skin. This same technique is quite versatile and is used to harvest cartilage for canal wall reconstruction when the retrograde mastoidectomy technique is used for cholesteatoma surgery.
The technique described here differs somewhat from the palisade tympanoplasty of Heermann et al.35 Instead of placing strips of cartilage side by side, an attempt is made to cut one major piece of cartilage in a semi-lunar fashion, which is placed directly on top of the prosthesis, abutting the malleus (Figs. 8A and 8B). This piece of cartilage acts to reconstruct a major portion of the posterior half of the TM and serves as the foundation for the TM cartilage reconstruction. A second, smaller semi-lunar piece is placed between this first piece and the canal wall to reconstruct the scutum precisely (Fig. 8C). This piece is extremely important to prevent cholesteatoma recurrence. Any spaces that result between this cartilage and the canal wall or scutum are filled in with small slivers of cartilage to prevent prosthesis extrusion and recurrent retraction (Fig. 8D). The reconstruction then is covered with the previously harvested perichondrium draped over the posterior canal wall. This technique is favored when ossiculoplasty is performed and the malleus is present and is especially suitable for cholesteatoma surgery. Because the prosthesis is placed prior to the cartilage reconstruction, this technique allows direct visualization and contact of the notched prosthesis to the manubrium handle, which has been shown to provide superior hearing results.36 The prosthesis acts as a scaffolding on which the cartilage is placed, which serves to reconstruct the TM as well as prevent prosthesis extrusion. It likewise allows a precise and “water-tight” fit between the reconstructed TM and the canal wall in the posterior area, where recurrent cholesteatoma is most common. Typically, in these situations, the anterior half of the TM is untouched or is grafted with conventional materials to allow cholesteatoma surveillance and possible intubation in the postoperative period if necessary.
Cartilage Tympanoplasty With Intact Chain. The atelectatic ear. The management of the atelectatic ear remains controversial, and the indications for surgical intervention have been discussed in a previous publication.33 Once the decision is made to intervene surgically, cartilage tympanoplasty using the perichondrium/cartilage island flap, with reconstruction of the entire TM, is advocated. The results with regard to positive graft outcomes and hearing have been reported previously and have been excellent.33 Two technical pitfalls specific for the atelectatic ear with an intact chain are worth mentioning. The first is the medially rotated malleus, which can make insertion of the flap quite difficult. This can be overcome with two techniques. The first is to remove 1 mm of the manubrium at the umbo with the malleus head amputator (malleus “nippers”). This does not affect hearing and allows medial placement of the graft. Attempting to lateralize the malleus with an intact chain should be discouraged because of the possibility of acoustic trauma. The second is to remove a slightly wider strip of cartilage (2 mm) to facilitate the malleus handle. This allows the more medial malleus to indent further into the perichondrium, allowing the cartilage plates to move laterally in the reconstruction and avoiding contact with the promontory. This also allows the anterior island of cartilage more flexibility in positioning, which is necessary to make good contact with the anterior annulus.
Another important technical factor specific to the atelectatic ear is the management of the atrophic TM. After elevation of the atrophic TM off the promontory, it is tempting to insert the cartilage medial to the intact TM. It is important, however, to remove at least a portion of the atrophic TM anterior and posterior to the malleus to ensure that the cartilage flap is incorporated into the reconstructed TM. We have experienced situations where the cartilage can be seen through the reconstructed TM and has fallen away from the TM when this removal of excess skin has not been performed, especially anteriorly.
The High-Risk Perforation. At our institution, the high-risk perforation is considered to be a revision surgery, a perforation anterior to the annulus, a perforation draining at the time of surgery, a perforation larger than 50%, or a bilateral perforation. The perichondrium/cartilage island flap is typically used for these situations if tragal cartilage is available. The perichondrium/cartilage island flap technique described above is used, with the exception that the size of the cartilage is tailored to the size of the perforation. For example, if the perforation is a 50% anterior perforation, the flap is constructed with the posterior island of cartilage removed to avoid the need to modify the normal TM posteriorly, which would otherwise be necessary to facilitate this plate of cartilage. The perichondrium is typically left the same size as described above so that it still extends under the posterior TM and drapes over the canal wall under the tympanomeatal flap for enhanced stability. Gelfoam is placed anteriorly to hold the graft against the annulus, but Gelfoam is avoided posteriorly around the ossicular chain. In revision surgery, fibrosis and disrupted mucosa are frequently seen, and Gelfoam in this milieu should be avoided to alleviate further scarring and optimize hearing results.
Cartilage Tympanoplasty With Ossicular Reconstruction. Reconstruction, or reinforcement, of the TM with cartilage is typically performed in conjunction with ossiculoplasty to prevent prosthesis extrusion, recurrent retractions, and cholesteatoma. While the usual surgical indication is cholesteatoma in these cases, the method of cartilage reconstruction is not dependent so much on the pathology, but on the presence or absence of the malleus manubrium and will be discussed as such.
Malleus Present. When the malleus handle is present, the modification of the palisade technique, as described above, is utilized. This is because of our philosophical approach to ossiculoplasty as well as the typical indication for this technique being cholesteatoma. It has been shown that incorporation of the malleus in ossicular reconstruction provides an acoustic gain, possibly because of its cantenary action in the TM.36 Likewise, the presence of the malleus with an intact anterior malleolar ligament offers improved prosthesis stability by allowing precise length adjustments and ultimate fit, leading to optimal hearing results.37 This has led us to abandon the use of the tragal island flap when the malleus is needed for ossicular reconstruction because of the fact that prior placement of the flap, using the underlay technique as described, obscures the malleus and makes the subsequent ossicular reconstruction less precise. Likewise, it is frequently difficult to carve the cartilage with enough precision so that the island flap fits exactly against the canal wall, which is necessary in cases involving cholesteatoma. We have experienced recurrent cholesteatomas when even a small gap is left between the posterior cartilage and the canal wall. For these reasons, we prefer the palisade technique in this situation.
When the malleus handle and suspensory ligaments are present, we have found that a Partial Ossicular Replacement Prosthesis (PORP) cut to 2 mm and a Total Ossicular Replacement Prosthesis (TORP) cut to 4.0 to 4.5 mm can be consistently used for precise reconstruction when the notch of the prosthesis is placed just inferior to the insertion of the tensor tympani. After precise ossicular reconstruction is performed, the posterior half of the TM is reconstructed with cartilage pieces as described above. The TM is put together like the pieces of a jigsaw puzzle. The half moon-shaped piece is placed on top of the prosthesis first, abutting the malleus handle, followed by the scutum piece. Because the usual indication for this technique is cholesteatoma, any spaces left between these two plates and the canal wall are reconstructed with slivers of cartilage cut to fit precisely in these areas. This reconstruction is then covered with perichondrium or fascia and the TM remnant, if available. No space is left between the canal wall and reconstructed TM to prevent cholesteatoma or retraction pocket recurrence. The anterior half of the TM is typically not reconstructed with cartilage to allow postoperative surveillance and tube insertion, if necessary.
If reconstruction of the anterior TM is necessary due to pathology, perichondrium is typically used. The technique is the same, but after precise fitting of the prosthesis to the malleus handle, the prosthesis is removed and the perichondrium is placed as an underlay graft. The prosthesis is then re-inserted, with palpation of the malleus handle through the graft to facilitate precise fit. The posterior palisade technique is then performed (Fig. 9).
Malleus Absent. The malleus-absent situation represents one of the most useful indications for cartilage tympanoplasty, but one of the more challenging situations for ossicular reconstruction. Whereas the presence of the malleus makes its possible to perform an ossiculoplasty with an exact fit between two essentially stable, bony platforms and allows the surgeon to build the ossicular reconstruction to the TM, this is obviously not the case if the malleus is absent. However, the cartilage tympanoplasty technique described below has proven useful to alleviate this problem.
The perichondrium/cartilage island flap from tragal cartilage is utilized. Even though the malleus is absent, a similar circular cartilage flap is constructed, again removing the 1- to 2-mm strip of cartilage from the center section to facilitate accurate placement of the ossicular replacement prosthesis. The cartilage is inserted in an underlay technique medial to the anterior TM remnant, with the perichondrium again toward the ear canal. Several pieces of Gelfoam are inserted to support the graft securely to the anterior annulus and the bony ledge just lateral to the supratubal recess. With the anterior portion of the cartilage graft held securely in place, the posterior half is folded out to expose the trailing edge of the anterior piece of cartilage, which acts, in effect, as a neo-malleus (Fig. 10). The distance between the stapes footplate or superstructure and this trailing edge of the anterior cartilage is measured, and the prosthesis is cut to the appropriate length.
We typically use a Dornhoffer TORP® or PORP® for ossicular reconstruction. The head of these prostheses was designed to use with cartilage tympanoplasty techniques as it notches the malleus and broadens posteriorly to shift the center of gravity over the shaft and act as a scaffold for the cartilage reconstruction.37 Other prostheses of similar design have likewise been utilized. The notched portion of the prosthesis is hooked under the trailing edge of the anterior piece of cartilage, much in the same way the malleus would be used in conventional ossiculoplasty, with the shaft placed on the stapes38 (Figs. 11A and 11B). The posterior half of the cartilage is folded back and is supported by the broader posterior head of the prosthesis. This technique allows accurate length measurement and placement of the prosthesis, with direct visualization of the stapes superstructure or footplate. Once the freestanding prosthesis is accurately positioned, supporting Gelfoam is placed to provide extra security during the postsurgical healing phase.
Typically, the packing material of Gelfoam and antibiotic ointment is completely suctioned from the external canal 1–2 weeks after the surgical procedure. Antibiotic steroid-containing drops are used for an additional 2 weeks to clear the ear of residual ointment and Gelfoam, the latter of which can lead to granulation and fibrous tissue formation if inadequately removed from the TM. The adult patient is instructed to begin the Valsalva maneuver, and the pediatric patient to use the Otovent (Invotec International, Jacksonville, FL) three times a day beginning 2–3 weeks after the surgery. A postoperative audiogram is obtained 6–8 weeks later, and the TM is examined. If the hearing result is good and the TM is clear, the ear is examined at 6 months and again at 1 year from the date of surgery. If effusion is present, nasal steroids are added, the Valsalva encouraged, and the ear is examined at 3 months. If the effusion is still present at this time, the ear is intubated. If a total cartilage reconstruction had been performed, tube insertion using traditional techniques can be difficult. In this situation, a CO2 laser myringotomy is preferred, followed by insertion of a soft tube, such as the Goode T-tube (Xomed Surgical Products, Jacksonville, FL).
A computerized otologic database was used to track all patients undergoing reconstruction of the TM with cartilage between July 1994 and July 2001. All surgeries were performed by the author and residents, under direct supervision of the author. Both the database and the individual patient charts were referenced to obtain the necessary data, which included age, sex, preand postoperative audiogram, type of surgery, surgical indication, revision versus primary surgery, length of follow-up, and pathology-specific complications. Hearing results were reported using a four-frequency (500, 1,000, 2,000, 3,000 Hz) pure-tone average air-bone gap (PTA-ABG).39 In calculating the ABG, postoperative bone scores were used, if available. The Student t test was used for statistical comparison. Postoperative complications included poor hearing requiring a second surgery, persistent effusion requiring TM intubation, perforation, retraction, prosthesis extrusion, facial nerve injury, and recurrent cholesteatoma.
During the study period, cartilage was used for TM reconstruction in more than 1,000 patients, of which only 712 had the hospital chart available for inclusion during the period of data assessment. Of these cases, 76 patients had postoperative follow-up with their local physicians, leaving 636 (350 male, 286 female) available for outcomes analysis. The average age of this group was 30 years (range, 4–88 y). The surgeries were performed in 331 pediatric cases (cases in which the patient was less than 21 y of age) and in 305 adult cases. A total of 287 cases (45%) were revision surgeries. The average follow-up time postsurgery was 2.7 years (range, 3 mo to 7 yrs). The overall average preoperative PTA-ABG was 25.7 ± 11.8 dB (SD, standard deviation), and the average postoperative PTA-ABG was 14.1 ± 9.5 dB. This represented a statistically significant improvement in hearing (P < .05). The following postoperative complications were seen: perforation in 14 ears (2.2%), conductive HL requiring revision surgery in 21 ears (3.3%), postoperative tube insertion performed in 28 ears (4.4%), intraoperative tube insertion performed in 38 ears (6.0%), prosthesis extrusion in 3 ears (0.5%). Recurrent cholesteatoma was seen in 8 cases (1.3% of total, 3.6% of cases where cholesteatoma was the primary surgical indication). There were three cases of delayed facial nerve weakness, occurring within 2 weeks of surgery, all of which recovered spontaneously and completely by 4–6 weeks.
Outcomes by Type of Surgery Performed
Type I reconstruction was performed in 319 cases, of which 226 were available for review. Of these, 64 (28%) were revision cases. The preoperative PTA-ABG was 16.1 ± 11.0 dB. The postoperative PTA-ABG was 11.3 ± 9.2 dB. This difference was statistically significant (P < .05). Ossicular reconstruction was performed with the stapes superstructure present (PORP) in 499 cases, of which 252 were available for review and 109 (43%) were revisions. The average preoperative PTA-ABG was 26.7 ± 12.5 dB, and the postoperative PTA-ABG was 14.5 ± 8.7 dB. This difference was statistically significant (P < .05). Ossicular reconstruction was performed with no stapes superstructure (TORP) in 282 cases, of which 158 were available for review. Of these, 114 (72%) were revision cases. The average preoperative PTA-ABG was 34.4 ± 11.9 dB, and the postoperative PTA-ABG was 16.6 ± 10.5 dB, which was statistically significant (P < .05). Mastoidectomy was performed as part of the procedure in 236 cases (37%).
Outcomes by Surgical Indication
Cholesteatoma. Of the 636 cases included for assessment, cholesteatoma was the primary surgical indication in 220 (35%). Of these, 87 (40% of 220) were revisions. A total of 100 were performed in children and 120 in adults. The average follow-up time postsurgery was 4 years and 11 months, but the average chart follow-up (average time after surgery at which the last chart entry containing an audiogram was made) was 11 months. The average preoperative PTA-ABG was 26.5 ± 12.6 dB, and the average postoperative value was 14.6 ± 8.8 dB, which represented a statistically significant improvement in hearing (P < .05). Recurrent cholesteatoma was seen in 8 cases (3.6%), conductive HL requiring revision in 4 cases (1.8%, 3 because of a displaced prostheses and 1 because of incus necrosis), perforation in 3 cases (1.4%), postoperative tube insertion in 11 cases (5.0%), and intraoperative tube insertion in 18 (8.2%). Of the recurrent perforations, all occurred in a portion of the TM not reconstructed with cartilage. Two of these perforations persisted after extrusion of a tube placed intraoperatively in the anterior normal eardrum.
Perforation. Perforation was the indication for surgery in 215 cases (34%), with 129 surgeries performed in children and 86 in adults. The surgery was a revision in 100 cases (47%). The average time since surgery was 2 years and 11 months, and the average chart follow-up was 12 months. The average preoperative PTA-ABG was 21.7 ± 13.5 dB while the postoperative value was 11.9 ± 9.3 dB, which was significantly different (P < .05). Complications included recurrent perforation in 9 ears (4.2% of 215 cases, 6 in children and 3 in adults, 4 of which were revisions), conductive HL requiring revision in 4 ears (1.9%, 1 displaced prosthesis, 1 extruded prosthesis, 1 incus necrosis, 1 fibrosis), postoperative tube insertion in 4 ears (1.9%), and intraoperative tube insertion in 6 ears (2.8%). Of the recurrent perforations, 6 occurred in the portion of the TM not reconstructed with cartilage, usually in the small space between the cartilage and annulus anteriorly. Of these 6 cases, 4 occurred spontaneously in the postoperative period, and 2 persisted after tube extrusion. Two ears showed complete dissolution of the cartilage graft. Both of these were in children, with one occurring within 3 months of surgery and one occurring 1 year after the surgery. Both ears had been initially reconstructed with perichondrium/cartilage island flaps, and both were later successfully repaired using the cartilage palisade technique. No perforations were seen acutely (<6 wk from surgery).
Atelectasis. An atelectatic eardrum was the surgical indication in 98 cases (15%). Of these, 21 (21% of 98) were revision cases. Sixty-six were performed in children and 32 in adults, with an average time since surgery of 3 years and a chart follow-up of 15 months. The average preoperative PTA-ABG was 20.2 ± 10.9 dB, and the average postoperative value was 14.2 ± 10.2 dB. This difference was significant (P < .05). Complications included 1 perforation (1.0%), conductive loss requiring revision in 2 cases (2.0%, 1 prosthesis extruded, 1 displaced), postoperative tube insertion in 7 (7.1%), and intraoperative insertion in 12 (12%).
Audiologic. This indication refers to situations where the goal of the surgery was to improve the hearing. A total of 103 audiologic cases (16%) were performed. Most of these (79 ears, 77% of 103) represented patients undergoing revision surgeries, many of whom were referred to our institution for further reconstructive surgery after the primary pathology had been dealt with by the referring physician. Frequent examples included canal wall-down surgery without reconstruction, second-stage canal wall-up surgery for cholesteatoma, and surgery for chronic otitis media with poor hearing. A total of 36 were performed in pediatric patients and 67 in adult patients, with an average time since surgery of 2 years and 7 months and a chart follow-up of 10 months. The average preoperative PTA-ABG was 33.6 ± 9.6 dB, and the average postoperative value was 14.6 ± 10.1 dB, which was a statistically significant difference (P < .05). Complications included 1 perforation (1.0%), conductive loss requiring revision surgery in 11 (11%), postoperative tube insertion in 6 (5.8%), and intraoperative tube insertion in 2 (1.9%). One case (1.0%) was complicated by prosthesis extrusion. Recurrent conductive HL was comparatively high in this group (11 patients, 11%). Further analysis revealed that 5 of these 11 patients had prosthesis dislocation at the time of revision surgery. Of these, all represented situations in which the cartilage island flap was used in a malleus-absent situation. Three of the eleven cases had stapes fixation requiring a staged stapedectomy, and 2 had fibrosis.
The use of cartilage is experiencing a renaissance in ear surgery because it appears to offer an extremely reliable method for reconstruction of the tympanic membrane in cases of advanced middle ear pathology and eustachian tube dysfunction. Many of the surgeries performed in this series would not have been attempted had it not been for the availability of cartilage tympanoplasty techniques. Like any new treatment, however, it is necessary to validate indications, assess outcomes, and recognize any complications. The purpose of this study was to examine function and hearing results in more than 1,000 patients undergoing cartilage tympanoplasty for a difficult ear (cholesteatoma, recurrent perforation, atelectasis).
Conceptually, one might anticipate a significant conductive HL with a TM that is rigid and thick, a fact that has hampered the routine acceptance of cartilage as a grafting material. However, there is no evidence in the literature to support the notion that cartilage is associated with a detrimental impact on hearing. A publication comparing perichondrium and cartilage in revision type I tympanoplasty showed no difference in hearing between the two groups, with both exhibiting an ABG of less than 10 dB.31 Another comparison study was performed by Gerber et al., in which cartilage was compared to fascia in a frequency-specific manner, and again no significant difference was seen.32 Graft take was 100% for the cartilage groups in both studies. Duckert et al. reported 97% graft take with cartilage and excellent hearing results.30 In their series, closure of the ABG to within 10 dB was achieved in 87% of type I tympanoplasties, 73% of type III (PORP) tympanoplasties, and 70% of type III (TORP) tympanoplasties. Milewski, reporting on his large series of cartilage tympanoplasties, of which 197 were type I, achieved drum closure in 92%, with an average ABG of less than 30 dB in 92%.14 Amadee et al. reported on 52 cases of cartilage tympanoplasty using the palisade technique, of which 18% were type I.29 TM closure was universal, with an average ABG of 4 dB. Adkins reported 55 cases of cartilage tympanoplasties for retractions.28 Graft take was universal, with less than a 10-dB conductive loss.
The results reported here for 636 evaluable ears showed the average PTA-ABG was 11.3 dB for type I tympanoplasties, 14.5 dB for type III (PORP), and 16.6 dB for type III (TORP). Considering that almost half of these cases were revisions, a factor that has shown to impart a negative impact on hearing, the results compare well with those using more traditional techniques. Given the postoperative appearance of the TM after cartilage tympanoplasty (Fig. 4), it is admittedly surprising that no HL is incurred. There is no satisfactory explanation for this phenomenon, other than that the dictum of “form follows function” appears not to apply to cartilage tympanoplasty.
A second technical issue regarding cartilage tympanoplasty concerns cartilage thickness. An elegant study using the laser Doppler interferometer and cadaver cartilage conducted by Zahnert at al. demonstrated that the ideal acoustical thickness of cartilage should be about 0.5 mm, compared to the full thickness at harvest of 0.7–1 mm.34 While these findings are not disputed, the techniques described herein utilize full-thickness grafts. The reason is technical. Typically, perichondrium is left attached to one side of the cartilage for stability. When the perichondrium/cartilage island flap is thinned, the graft curls unacceptably toward the side with the perichondrium, making placement difficult and less exact. Full-thickness grafts are believed, in our experience, to give more precise reconstruction, a critical issue in cholesteatoma surgery. Therefore, because our hearing results have been good, we recommend full-thickness grafts for cartilage reconstruction of the eardrum.
When analyzing the results of this study, follow-up time was reported as both the time from surgery and as the time after surgery at which the last chart entry containing an audiogram was made. Thus, the average time after surgery was generally several years, and the last follow-up with audiometric data was generally 11–14 months. This is because of three factors. First and foremost concerns the managed care climate in the United States. It has become somewhat difficult to follow these patients, especially those without complaints, for a long period of time because of pressure to encourage follow-up with the patient's primary care physician. This is an unfortunate fact that will continue to limit long-term, longitudinal assessment of this type of patient, especially where audiometric data are concerned. The next two factors deal with the practice setting. Our institution is a tertiary referral center; thus, patients frequently come from a great distance and wish to have follow-up closer to home. Lastly, at routine follow-up, if the patient has no hearing complaint, an audiogram is usually not obtained to maintain a cost-effective practice. As a result of these factors, it is very difficult in this series to comment on specific long-term results with regard to hearing. Adkins has reported that the hearing remained stable or improved in his series,28 and that is the opinion of the author. However, that conclusion cannot be confirmed from the data reported herein.
The indications for cartilage reconstruction of the TM used in our study appear to be appropriate, as evidenced by our results with regard to graft acceptance and hearing outcome. In this series of patients, cholesteatoma represents one of the most controversial but important pathological conditions where cartilage is used. The primary purpose of cholesteatoma surgery is to eradicate disease and provide a safe, hearing ear. The magnitude of the controversy with regard to optimal surgical management is beyond the scope of this discussion, but cartilage should arguably be involved in each technique. To appreciate the utility of cartilage in TM reinforcement of the eardrum, one must understand the nature of cholesteatoma recurrence. The results for canal wall-up surgery reported in the literature show tremendous variability in recurrence rates. Sheehy and Robinson40 initially reported recurrence rates of 5%, and Glasscock41 reported a 14% recurrence rate. At the other extreme, Mazzoni et al.42 and Cody et al.43 reported recurrence rates of 33% and 50%, respectively.
When reporting results with cholesteatoma surgery, it is essential to discuss both recurrence and residual rates. Residual disease is that which is left behind after initial surgery and is determined by the method of cholesteatoma exposure and removal. Seventy-five percent of all residual disease is believed to manifest by 2 years after surgery.44 Recurrent cholesteatoma, on the other hand, is caused by recurrent retractions from the reconstructed TM and is dictated by the method of reconstruction and presence of continued eustachian tube dysfunction, a problem that is difficult to assess preoperatively. One must also keep in mind that “second-look” surgeries 1 year after primary surgery affect only the rate of residual disease, and these 2-stage surgeries are still associated with a 10–12% recurrence rate due to retractions in the reconstructed eardrum when using traditional reconstruction techniques.45,46
We use a technique of partial canal wall removal for cholesteatoma extirpation followed by cartilage reconstruction. Long-term results using these techniques have been reported elsewhere and were good with regard to functional outcome.47 For TM reconstruction, the cartilage palisade technique is preferred for reasons discussed earlier. Most cases involving cholesteatoma require ossicular reconstruction, and a previous study has shown superior hearing results when the malleus is incorporated into the reconstruction.36 The palisade technique allows precise placement of the prosthesis against the malleus, with the prosthesis acting as a scaffold to support the cartilage posteriorly. This technique allows excellent approximation between the canal wall and TM reconstruction, creating a “water-tight” fit, which has greatly reduced recurrent disease. Our recurrence rate in this series was quite low, but the follow-up should ideally be longer to draw any significant conclusions with regard to recurrence. However, a study with a smaller number of ears (75 cases) and longer follow-up (average, 4 y) showed a recurrence rate of less than 10%, which is acceptable and considerably better than results obtained using traditional techniques without cartilage reconstruction of the TM.47,48
With this technique, we prefer to leave the anterior portion of the TM without cartilage for observation and possible tube placement should this be necessary in the postoperative period, which was the case in about 5% of the ears in our series. For reasons that are not totally understood, the posterior portion of the TM tends to be much more prone to retraction than the anterior portion, so leaving this area without cartilage has not proven to be a problem with recurrent cholesteatoma.49,50 One serious disadvantage of cartilage in this scenario is that it creates an opaque TM posteriorly, which could potentially hide residual disease. This is a problem that should be recognized, and surgical discretion should be used. If major disruption of the cholesteatoma sac occurs at extirpation, one must consider the advisability of performing a second-look surgery at a later date. However, this also applies to cholesteatoma surgery in general, not just in cases where cartilage is used in the reconstruction. One must recognize the fact that most residual disease occurs in the epitympanum, an area that is hidden by the bony canal wall and scutum when canal wall-up surgery of any type is performed.51 Posterior cartilage TM reconstruction does probably delay the diagnosis of residual cholesteatoma, but in our experience, the disease will become manifest either anteriorly or as a recurrence of a conductive HL. We have had no major complications as a result of this delay in diagnosis, and this approach has been supported by others.52,53
The high-risk perforation is a situation where cartilage has proven extremely valuable. A high-risk perforation is loosely defined as one that has been shown to be associated with increased failure rates using traditional techniques. Revision tympanoplasty has been shown by numerous authors to be a risk factor for subsequent failure in graft acceptance and hearing results.54–56 Nearly half of the perforation group in our series were revision cases, and, of the 9 recurrent perforations after cartilage reconstruction, 4 were revisions. Drainage at the time of surgery is considered by many to represent a negative prognostic factor.54,55 Bellucci's classification system delineates four prognostic groups based on the degree of drainage.57 Albu et al.,55 Black,54 and others have likewise shown this to be statistically relevant. While every attempt is made to dry an ear prior to surgical intervention in our practice, we do not consider this a prerequisite for tympanoplasty. Cartilage is used in this situation and has proven to yield successful results. Of the recurrent perforations, only 2 occurred in ears that were draining at the time of surgery. Cartilage is likewise used for reconstruction when the size of the perforation is larger than 50%. In our experience, the larger perforations fair less well, and this is in full agreement with findings reported by Booth,58 Albu et al.,55 and Jurovitzki and Sadé.59 Three recurrent perforations in our series had subtotal perforations as a surgical indication.
Age as a prognostic factor in cartilage tympanoplasty of high-risk perforations is somewhat confusing. In previous studies, a young patient age (<18 y) was not found, in and of itself, to have a negative influence on success.31,33 There is, however, an association between young age and other significant factors, such as bilateral ear disease and drainage at the time of surgery, which are associated with immature tubal function. Overall, the number of cases in this series is split nearly evenly between adults and children. However, in the high-risk perforation group, almost two-thirds were in children, substantiating the association between young age and other factors (bilateral ear disease, drainage) felt to be significant. Likewise, two-thirds (6/9) of the recurrent perforations were in children. Our general approach to pediatric patients is to avoid repairing the TM during the otitis-prone years (<3 y). If the contralateral ear is normal, routine tympanoplasty is performed at age 4.60 If the contralateral ear is abnormal at this time, adenoidectomy is considered and tympanoplasty is generally deferred until age 7.61–63 If contralateral disease is still present at this time, cartilage tympanoplasty is performed on the worse ear as a perforation in the contralateral ear has been shown to be associated with a high risk for failure.64
The results in the high-risk perforation group have been satisfactory, but are somewhat worse than the cartilage group as a whole. Overall, this group represents 34% (215/636) of the total cases, and 64% (9/14) of the recurrent perforations. Interestingly, 6 of these 9 perforations (67%) occurred in a portion of the TM not reconstructed with cartilage. It has been our policy to reconstruct only that portion of the TM containing the perforation and leave the TM remnant in place. Based on the findings of this study, that policy may need to be re-assessed, with consideration given to more aggressive use of cartilage in this situation.
The atelectatic ear represents the first and most well described situation in which cartilage techniques have been utilized. Whereas perichondrium and fascia have been shown to undergo atrophy after use in this situation,11 numerous reports have established the efficacy of cartilage in TM reconstruction of the atelectatic ear.19,25,26 The techniques of cartilage tympanoplasty are not as controversial as the indications for surgery itself. Much of the confusion associated with this disorder stems from a poor understanding of the underlying pathophysiologic conditions that ultimately lead to changes in the TM, resulting in atrophy, diffuse or local retractions, and cholesteatoma formation.65,66 The controversy is augmented by the fact that, early in the course of the disease, and even in the presence of incus necrosis, HL is frequently minimal, and patients are mostly asymptomatic. A grading system that is a modification of Sadé's67 has been presented to develop a treatment algorithm for this condition.33 Once a decision is made to proceed with surgery (grade III with adhesions and grade IV), the perichondrium/cartilage island flap is generally used to reconstruct the TM.
Two-thirds of the patients in our atelectasis series were pediatric, and the results with regard to hearing and graft acceptance were good. Only one perforation at the edge of the cartilage next to the anterior sulcus was seen, and only one prosthesis extrusion occurred. One interesting point is that 7% of this group received a postoperative ventilation tube, which is not significantly higher than the other groups, but 12% received an intraoperative tube through the cartilage, which is higher. This may represent a bias in surgical decision-making as opposed to a true need for ventilation. Because the exact mechanism of eustachian tube dysfunction leading to this condition (occlusion, hypofunction, hyperfunction) is poorly understood, we tend to perform cartilage tympanoplasty with prophylactic tube insertion at the time of surgery. However, because our experience with the atelectatic ear has been good, with very few complications, we are performing fewer primary intubations, and a ventilation tube is placed postoperatively if indicated by the presence of an effusion.
The use of cartilage tympanoplasty, while encouraging, also provides some specific challenges, the first of which is learning a new surgical technique. The only real difference in this technique and traditional underlay procedures is harvesting and shaping the cartilage graft. This has not been a significant challenge in resident training, especially with the perichondrium/cartilage island flap. The palisade technique, while not difficult, does require some practice to efficiently cut the cartilage pieces to fit with minimal trial and error. While these techniques do have learning curves, both methods have been taught to lower-level residents at our institution without difficulty.
The second major challenge, and one that continues to present some difficulty, represents the inability to predict, preoperatively, eustachian tube function. This has been the subject of numerous articles, and the general consensus continues to be that no specific exam finding or test reliably predicts tubal function.68,69 As outlined in our Results section, close scrutiny has been given to the need to intubate the TM reconstructed with cartilage. We generally try to avoid primary tube insertion because most of the cases in our series (90%) did not require it and 4 of the 14 (28%) perforations in this series resulted after tube extrusion or removal. Likewise, postoperative intubation after cartilage reconstruction can be very challenging and frequently requires a second operation. It is necessary to cut out a small piece of cartilage and to use a soft t-tube, such as a Goode t-tube. Because of the thickness of the cartilage graft, it is not possible to use a grommet type tube in this situation.
We have tried to solve this problem using two new techniques. The first is to use the CO2 laser under topical anesthesia (phenol) to vaporize a 1- to 2-mm hole in the cartilage. This has been successful in the outpatient setting in the few patients in whom it has been used to date. The second technique is to place the t-tube through the cartilage graft before insertion of the graft at surgery in patients who are considered very high risk for ongoing eustachian dysfunction (see Fig. 12). This technique was initially designed for patients with Down syndrome, but has proven useful for long-lasting intubation in patients with other craniofacial abnormalities and tubal dysfunction secondary to treated malignancies of the head and neck.70 Certainly, until some reliable test of tubal function is discovered, a finite percent of patients, 5% in our experience, will require postoperative tube insertion. On a positive note, however, while eustachian tube dysfunction was felt to be the underlying cause of the pathological changes in the majority of the ears for which cartilage tympanoplasty was used, 90% had successful surgery without the need for secondary ventilation.
In summary, cartilage is proving to be a very effective material for the reconstruction of the TM in cases of advanced middle ear pathology. It is it particularly useful for the management of the atelectatic ear, cholesteatoma, and the high-risk perforation (wet, recurrent, large, and/or anterior). Functional results in each pathological group in our series were good, with a statistically significant improvement in hearing appreciated in each.
Both the palisade and perichondrium/cartilage island flap are utilized depending on the pathological condition of the middle ear and status of the ossicular chain. The palisade technique is preferred in cases of cholesteatoma and when ossicular reconstruction is needed in the malleus-present situation. The island flap is preferred in the management of the atelectatic ear, the high-risk perforation, and when ossicular reconstruction is performed in the malleus-absent situation. Suitable cartilage can be harvested from both the cymba area of the concha and the tragus, depending on the surgical approach and technique used. Tragal cartilage is preferred for the island flap while cymba cartilage is more suitable for the palisade technique. Thinning the cartilage is not needed for good functional outcome. While the need for postoperative tube insertion is relatively rare, it can prove to be difficult when the entire TM is reconstructed with cartilage, emphasizing the need to optimize tubal function and continue research to better predict outcome based on preoperative parameters.
Results in our patient series indicate that these cartilage tympanoplasty techniques achieve good anatomical and audiologic results when pathology and status of the ossicular chain dictate the specific technique utilized. A statistically significant improvement in hearing was realized in each pathological group. We believe the indications for cartilage tympanoplasty have been validated by these results. In the atelectatic ear, cartilage allowed us to reconstruct the TM with good anatomic results compared to traditional reconstructions, which have shown high rates of retraction and failure. In cases of cholesteatoma, cartilage tympanoplasty using the palisade technique resulted in precise reconstruction of the TM and helped reduce rates of recurrence. Finally, in the high-risk perforation, reconstruction with cartilage yielded anatomical and functional results that compare favorably to primary tympanoplasty using traditional techniques.
It should be noted, however, that the results of this study represent relatively short-term follow-up, and only the test of time will substantiate these preliminary findings.