1)    Pediatric myringoplasties have an early take rate of up to 100% but a late failure rate of 30% over two years.

2)    Adult myringoplasties have an early take rate of approximately 95% and a late failure rate of 8%.

3)    Lateral facial graft has a high take rate but also a high complication rate (cholesteoma, canal stenosis blunting of the anterior canal wall).

4)    Medial facial graft has a lower take rate but a lower complication rate.

5)    Eustachian tube function approaches normal at age 7 to 8 years.
 

The first method used to treat tympanic membrane perforations was the use of prosthetic patching.  Materials used included tubes of ivory covered with pig's bladders (Banzer, 1640), cotton wool pledgets (Yearsley, 1863), vulcanized rubber disks (Toynbee, 1857) and paper (Blake 1887).  The next advance was thc use of silver nitrate (Roosa, 1876; Politzer, 1894) or tricholoracetic acid (Okuneff, 1895) cautery to stimulate thc margins of the perforation and obtain closure of the tympanic membrane.  The rate of closure using this technique increased after Dunlap (1917,1947) reported that repeated treatments at specific intervals are usually necessary.  Joynt (1919) combined the use of cautery and paper patching.  This latter technique was recently reported by Derlacki (1953) and Wright (1956).  However, this technique still would only close 69% to 88% of the perforations and require an average of 10 to 14.6 treatments.

Repair of a perforation using tissue was successfully accomplished as early as 1887 by Benhold, who coined thc term "Myringoplastik".  The drum and promontory was deepithelizcd and a skin graft was laid over thc drum and onto the promontory.  This technique was not widely used until it was popularized in the mid 1950's by Wullstein (1952) and Zollner (1955).  House (1953) reported his experience with skin grafting and noted a case where the graphs pulled off of the promontory giving rise to a freely mobile tympanic membrane.  Use of a skin graft (House 1960), canal skin graft (House,1961), or pedical flap (Sooy, 1956) applied solely to thc ear drum was soon advocated.  However, long term closure rates were only 77% (Sheehy, 1967) to 84% (Sheehy, 1980).  Other grafting materials were searched for.  Temporalis fascia was reported to have an excellent take rate due to its low metabolic requirements by Storrs (1960).  Closure rates using this material positioned with an overlay technique have been reported as high as 97% (Sheey, 1980).

The overlay technique is associated with problems of blunting of the anterior sulcus, lateralization of the graft (Sheey, 1980; Glasscock, 1973, Booth, 1974), and epithelial pearl formation (Glasscock, 1973).   Placing the graft under the perforation avoids these problems but is technically more difficult.  The underlay technique was first reported by Shea (1960) and Tabb (1960) using vein as a grafting material.  Storrs (1961) described utilizing temporalis fascia and Glasscock (1973,1982) refined the technique.  Management of the anterior canal wall hump has always been a problem with thc underlay technique.  Several solutions are described including the use of a postauricular approach, a second canal nap based anterior 3rd lateral and finally making a large inferior canal wall nap as described by Glasscock (1982). to expose the anterior canal bone and remove it with a drill.

Plugging of a small perforation with fat has also been successful in restoring an intact tympanic membrane.  Take rates approximate 80% with success in children reported by Gross et. al. (1989).

Despite the evolution of myringoplasty to a reliable and effective technique in closing a tympanic membrane perforation.  Its indications and use in children is still debated.  It is the purpose of this communication to report the author's experience, using a facial graft underlay approach and to review the literature concerning myringoplasty in children.

Methodology
A total of 54 ears in 50 patients underwent myringoplasty using a facial graft underlay approach.  A single surgeon, the author, performed all operations. Charts were review for the size of perforation, rate of closure, age of patient, hearing result and status of aural discharge at presentation and surgery. Children underwent myringoplasty when they had an opposite ear with an intact ear drum which had been free from otitis media for at least one year.  In the case of bilateral perforation, the child was either 8 years of age or had residual perforations from myringotomy tubes and at least 3 years had elapsed between tube insertion and attempt at repair. Children were selected for this later category if the perforations were large and a significant hearing loss was present.

Results
The results are shown in Table 1.  The total take rate for all subjects was 82%. Several patient were high risk for failures. procedures were performed as revisions. One patient, age 12, who was, immunocompromised, had a delayed failure at 9 months on his 5th revision and an early failure on his 6th.  His first 5 procedures were done at a very young age.  In an additional patient, age 15 yrs, developed a pinhole perforation 8 months postoperatively which was closed with paper patching.  A third patient, age 45, had a delayed failure at 8 months. This patient had lupus, and originally had a posterior inferior perforation.  A 5% anterior superior perforation developed along with necrosis of the malleus, and extrusion of thc incus through thc drum and exposed bone in thc posterior superior canal. One month later, all bone involvement spontaneously healed by the perforation persisted.  Another adult patient, from Africa, had necrosis of the entire tympanomeatal flap.  The initial perforation was inferior and closure was uneventful.  This patient returned to Africa after 60 days of follow-up.  At this time no perforation was present, but necrotic tissue still covered the drum and posterior canal.  This patient is counted as a failure. 

The overall take rate in adults was 85% and in children was 79%.  Five out of seven failures in children were delayed compared to one out of three (the patient with lupus) in adults.  The failure rate in children 0 - 8 years was less than that found in the 9 - 17 yr olds (p < 0.05 Chi Square).  Three children age 4 years of age had closure of their perforation with follow-up of 329 to 542 days.  Nine patients had wet ears 1 month before surgery, including two ears wet at the time of surgery.  Two patients which had a dry ear at the time of surgery had a graft failure.  One of these patients was an immunocompromised child undergoing a 5th revision.