Middle ear injury through the external auditory canal: a review of 44 cases.

Abstract

We performed a retrospective review of 44 patients with middle ear injury incurred through the external auditory canal. Twenty-two of the 44 patients had presented to our center within 1 month of their injury (early group), and 22 presented later (delayed group); the mean interval from the time of the trauma to presentation was 6 days in the early group and 7 years in the delayed group. The causes of injury were penetrating trauma (70% of cases), thermal insults (20%), and explosive and nonexplosive blasts (9%). Purulent otorrhea, cholesteatoma, and ossicular discontinuity were more common in the delayed group. Otologic surgery was required in 9 early-group patients (41%) and in all 22 delayed-group patients (100%). Two patients in the early group developed a dead eat: The mean pure-tone averages (PTAs) at presentation were 30.7 and 52.2 dB in the early and delayed groups, respectively; after management, the corresponding mean PTAs were 21.0 and 42.5 dB. The respective mean air-bone gaps in the two groups were 14.6 and 28.2 dB at presentation and 8.0 and 17.2 dB after management. We conclude that middle ear injury incurred as a result of trauma sustained through the external auditory canal is associated with considerable morbidity. Patients who present in a delayed fashion have significantly poorer hearing at presentation and after management. Patients who do not develop a dead ear generally derive benefit from reconstruction of the middle ear sound-conduction mechanism.

Introduction

Middle ear injury frequently occurs as a result of trauma sustained directly through the external auditory canal. Penetrating trauma, thermal insults, and explosive and nonexplosive blasts are the usual mechanisms of injury. Penetrating injury is usually the result of an accidental trauma self-inflicted with a slender object such as a cotton-tipped applicator. Thermal injuries are most often secondary to molten slag entering the ear canal during welding. Blast injuries are produced by a rapid positive pressure wave through the external auditory canal; explosive blast injuries occur when the patient is in proximity to the detonation of an explosive device, while nonexplosive blast injuries are caused by a sudden increase in ear canal air pressure, such as occurs with a slap to the ear. (1,2) Middle ear injuries caused by lightning appear to be a form of blast injury, although there may also be a thermal component. (3) Regardless of the mechanism of injury, the tympanic membrane is typically perforated, which results in a conductive hearing loss. Severe trauma may also damage the ossicles, facial nerve, and inner ear structures. Fortunately, most of these injuries respond favorably to treatment. (4)

Patients with middle ear injury characteristically present with audiologic and/or vestibular symptoms. However, the presentation may differ between those patients who present in an early versus delayed fashion. Patients presenting months to years after injury often have chronic suppurative ear disease complicating their original trauma. (5)

The time of presentation notwithstanding, the history, physical examination, and audiogram should provide the physician with adequate information to determine the extent of injury. The ultimate treatment goal for these patients is an intact tympanic membrane and a hearing ear.

In this article, we describe our retrospective study of the presentation and clinical course of 44 patients who experienced middle ear trauma through the external auditory canal. We discuss the mechanisms of injury, subjective complaints, objective findings, and management modalities, and we compare audiologic outcomes in patients who presented in an early versus delayed fashion.

Patients and methods

All patients had presented to a single tertiary care otologic center in Wichita, Kans. Middle ear injury was defined as damage to the tympanic membrane and/or middle ear structures. We identified 51 such patients. Eligibility criteria included (1) a thorough description of the traumatic event, the signs and symptoms at presentation, and any intervention and (2) a complete air- and bone-conduction audiogram obtained at presentation and after any treatment. Seven patients were excluded because the information in their charts was inadequate.

The remaining 44 patients were assigned to one of two groups according to how long after their trauma they presented to our center--within 1 month (early group) or beyond 1 month (delayed group). The early group was made up of 15 males and 7 females (mean age: 28.6 yr; median: 23), and the delayed group was likewise made up of 15 males and 7 females (mean age: 33.9 yr: median: 25). The difference in age between the two groups was not statistically significant (p = 0.4707) and, of course, the sex distribution was identical.

The presentation pattern was easily discernable. The patients in the early group had presented to us between 1 and 28 days after their trauma (mean: 6 days), while those in the delayed group had presented between 6 months and 53 years after their trauma (mean: 7 yr). At presentation, all patients had undergone a neurotologic office examination with microscopy. Management strategies included observation, office myringoplasty, and middle ear or mastoid surgery. Acute perforations (without other middle ear injury) were observed for spontaneous healing. Paper-patch myringoplasties were performed on perforations with in-turned epithelial edges. Middle ear or mastoid procedures were used to treat persistent perforations, ossicular discontinuity, chronic otitis media with or without cholesteatoma, and perilymph fistula. Patients with suppurative middle ear disease received ototopical antimicrobial treatment prior to surgical intervention. The type of surgical procedures included myringoplasty, tympanoplasty, tympano-ossiculoplasty, tympanomastoidectomy, perilymph fistula repair, and labyrinthectomy.

Outcomes measures included pre- and postmanagement air-conduction pure-tone average (PTA) and air-bone gap (ABG). The PTA was calculated using 500 Hz, 1 kHz, 2 kHz, and 4 kHz: 3 kHz was not recorded for most patients and was therefore not used to calculate the PTA. The ABG was determined by subtracting the PTA bone-conduction value from the PTA air-conduction value. Postoperative bone conduction was used to calculate the postmanagement ABG. Follow-up audiograms were performed between 3 months and 1 year after management. Patients with no audiometric response on follow-up did not have their postmanagement PTA and ABG recorded because of the absence of numerical data: these patients were considered to have a dead ear and were reported as such.

Statistical analyses were used to assess intra- and intergroup differences. Results of both Shapiro-Wilk and Anderson-Darling tests indicated that nonparametric tests should be used for all comparisons. The Wilcoxon signed rank test was used to assess changes within the groups, and the Wilcoxon-Mann-Whitney exact test was used to assess differences between the groups. To assess changes in hearing, Wilcoxon signed rank exact tests were conducted on outcomes measurements (PTA and ABG) in both the early and delayed groups. Wilcoxon-Mann-Whitney tests were also conducted to determine any differences in initial and final hearing measures (PTA and ABG) between the two groups. Finally, Wilcoxon-Mann-Whitney tests were used to determine if one group had more hearing improvement (PTA and ABG) than did the other following management.

This study was approved by our institution's human subject review committee, and it received an institutional review board "exempt" classification.

Results

Mechanisms of injury. Injuries were caused by a variety of events (table 1). Overall, 31 of the 44 patients (70%) experienced an accidental penetrating trauma; this was the most common type of trauma irrespective of presentation time. Twenty-six of the 31 penetrating injuries (84%) were self-inflicted, and 5 (16%) were sustained iatrogenically during foreign-body removal. Penetrating trauma occurred in 17 of the 22 early-group patients (77%) and in 14 delayed-group patients (64%). Nine of the 44 patients (20%) had thermal (slag) injuries--5 in the early group and 4 in the delayed group. All 4 blast injuries (9%) occurred in the delayed group; 2 were caused by nonexplosive blasts (1 compressed-air injury and 1 airbag deployment), and 2 were caused by explosive events (1 fireworks blast and 1 lightning strike).

Presenting signs and symptoms. All 44 patients reported subjective hearing loss at presentation (table 2). Vertigo was more frequently reported in the early group (7 [32%] vs. 2 [9%]). Otorrhea was present during the initial examination in 8 early-group patients (36%) and 13 delayed-group patients (59%). However, the nature of the otorrhea was quite different between the two groups. The early patients had bloody otorrhea from their recent trauma, whereas the delayed group had purulent otorrhea from chronic suppurative ear disease.

Tympanic membrane perforations were common at presentation irrespective of the mechanism of injury, being present in 39 of the 44 patients (89%). This included all 22 early-group patients (100%) and 17 of 22 delayed-group patients (77%). The delayed group had a tendency toward larger perforations (table 3), and it had more patients with ossicular discontinuity (12 [55%] vs. 4 [18%]).

Cholesteatoma was present in 5 patients in the delayed group (23%); 3 had sustained a penetrating trauma and two had incurred a blast injury. Two patients with penetrating trauma and cholesteatoma had facial nerve involvement. One had cholesteatoma invading the fallopian canal, and the other was found to have a chorda tympani nerve injury at exploration. There were no cholesteatomas or facial nerve injuries in the early group. Two early-group patients and one delayed-group patient with a penetrating injury developed perilymph fistula. These were the only patients in the study who developed a sensorineural component to their hearing loss.

Management. Surgical intervention was required less often for patients who presented in an early fashion (table 4).

Thirteen early presenters were observed to have spontaneous closure of their perforation, including all 5 with a slag injury. Nine early-group patients (41%), all with a penetrating injury, required operative intervention. Five underwent successful paper-patch myringoplasty, and the remaining 4 required middle ear surgery. Two of these patients, both with perilymph fistula, required more than one operative procedure. All early presenters had closure of their perforation at follow-up.

All 22 patients in the delayed group (100%) required middle ear and/or mastoid surgery. Three (14%) of these patients (2 with a penetrating injury and 1 with a blast injury) required more than one operative procedure. Twenty patients in the delayed group (91%) had closure of their tympanic membrane at follow-up. The 2 tympanoplasty failures included 1 case of an explosive blast injury (fireworks) and 1 slag burn. The latter patient had persistent otorrhea.

Perilymph fistula developed in 2 early-group patients and in 1 delayed-group patient with penetrating trauma. Both of the early-group patients presented with severe vertigo, nystagmus, and reduced cochlear function, and both developed a dead ear. These 3 patients were the only patients in our study who developed a sensorineural component to their hearing loss:

* One early-group patient who experienced penetrating trauma from a cotton-tipped applicator demonstrated a moderate mixed hearing loss at presentation. Computed tomography (CT) revealed a complete luxation of the stapes into the vestibule (figure). A fistula repair was accompli shed without removing the stapes from the vestibule by using perichondrium as a seal. Postoperatively, this patient's audiogram revealed a mild to severe down-sloping mixed hearing loss. Four months later, the patient underwent an ossiculoplasty (a malleus-to-oval window graft), but the postoperative audiogram revealed no response. The patient developed a dead ear, although his vestibular symptoms completely resolved.

[FIGURE OMITTED]

* The other early-group patient with a penetrating trauma and perilymph fistula presented with a severe mixed hearing loss and partial luxation of the stapes into the vestibule. His injury had been caused by insertion of a twig. During surgery, the stapes was gently elevated from the vestibule, and the fistula was repaired with perichondrium. A stainless-steel stapes prosthesis was then placed from the incus to the footplate. Postoperatively, the patient had no audiometric response and a dead ear. His vestibular symptoms persisted, and he ultimately required a transcanal labyrinthectomy.

* One delayed-group patient developed a perilymph fistula. This patient had incurred a penetrating injury as a young child (bobby pin). Six years later, he presented to us with a subtotal perforation and severe mixed hearing loss. An incus dislocation and a severely fractured stapes were encountered intraoperatively, and a tympano-ossiculoplasty (malleus-to-oval window) was performed. The perforation closed without significant hearing improvement. A progressive mixed hearing loss developed 1 year later: at that time, a fistula was identified and repaired with perichondrium. The patient's postoperative hearing stabilized, and the mixed hearing loss remained severe.

Audiometry. On average, the early group had significantly better hearing in the injured ear than did the delayed group at both presentation and follow-up (table 5). At presentation, the mean PTAs in the injured ears were 30.7 and 52.2 dB (p = 0.0007), respectively, and the mean ABGs were 14.6 and 28.2 dB (p = 0.0045). At follow-up, the mean PTAs were 21.0 and 42.5 dB (p = 0.0044) and the mean ABGs were 8.0 and 17.2 dB (p = 0.0092). There was no significant difference between the two groups for PTA in the uninjured ear at either presentation (p = 0.3279) or follow-up (p = 0.3823).

Most patients had useful hearing in the injured ear at follow-up regardless of the time of presentation and the mechanism of injury. Notable exceptions were the 2 early-group patients whose penetrating trauma led to perilymph fistula and luxation of the stapes into the vestibule; as mentioned, both ultimately developed a dead ear. Excluding these 2 patients, management led to a significant improvement in mean injured-ear PTA (p = 0.0313) and ABG (p = 0.0313) in the early group. Likewise, the delayed group experienced significant improvement in mean injured-ear PTA (p = 0.0145) and ABG (p = 0.0023) at follow-up. The degree of hearing gain in the two groups was not significantly different in terms of PTA (p = 0.5582) or ABG (p = 0.3573). As expected, no significant change in hearing occurred in the uninjured ear in either group.

Discussion

Ideally, patients with middle ear injury sustained through the external auditory canal are evaluated shortly after the traumatic event. This enables the physician to determine the extent of injury and plan appropriate treatment. However, despite the presence of audiologic and vestibular symptoms, not all patients seek immediate medical attention. Brahe Pedersen and Vendelbo Johanson reviewed cases of middle ear trauma from all causes and found prolonged treatment delays in many instances. (5) Our study validated that finding, as the patients in the delayed group presented an average of 7 years alter their injury.

The reasons why some patients do not seek immediate treatment are unclear. It is possible that some perforations spontaneously heal, and then the ear develops problems down the road. For instance, the development of middle ear cholesteatoma has been reported to be a delayed complication of blast injury. (6) Presumably, epithelial cells are transplanted into the middle ear cleft during the blast. This phenomenon was likely observed in 1 nonexplosive (compressed-air injury) and 1 explosive blast injury (fireworks injury) in the current study. Lack of healthcare access or insurance may also contribute to treatment delay; however, this was not documented for any patients in the current study. Finally, some patients with accidental, self-inflicted injuries may simply be too embarrassed to visit a physician.

Regardless of the mechanism of injury or the time of the initial presentation, patients with traumatic middle ear injuries universally report hearing loss. The hearing loss is usually conductive in nature and the result of a perforation and/or ossicular disruption. Sensorineural hearing losses are less common, and they are typically associated with direct or indirect inner ear trauma. Penetrating missile-type trauma may directly injure the stapes or inner ear and result in perilymph fistula, vertigo, and reduced cochlear function. Luxation of the stapes into the vestibule has been well reported, a finding that is most often associated with a mixed or pure sensory loss. (7,8) Blast injuries may cause disruption of the round or oval window membranes and result in sensorineural loss, but noise-induced temporary or permanent threshold shifts have also been described with these injuries. (1,9) Thermal injuries (slag) may cause a sensorineural loss by superheating the perilymph. (10,11)

Tympanic membrane perforation is common following middle ear trauma through the external auditory canal; perforation was present in 39 of the 44 patients (89%) in our study. Presumably, the remaining 5 patients--all of whom were in the delayed group--experienced a spontaneous closure prior to evaluation. Most acute traumatic perforations have a tendency to heal spontaneously, but the ideal management of these patients is controversial, although it is generally accepted that any in-turned epithelial edges should be returned to their native anatomic position. (4) Lindeman et al found no significant difference in closure rates between patients who were merely observed and those who were treated with paper-patch myringoplasty; they reported an overall closure rate of 94%. (12) Camnitz and Bost treated 50 patients with paper-patch myringoplasty and reported a similar closure rate: 92%. (13) Kristensen reviewed 760 traumatic perforations and reported a spontaneous closure rate of only 80%; he wrote that the mechanism of injury likely influences the rate of spontaneous healing. (14) In our study, all 18 acute, traumatic perforations healed--13 spontaneously and 5 with paper-patch myringoplasty.

A key difference between patients who present early and late is the presence of suppurative ear disease at presentation in the latter group. Brahe Pedersen and Vendelbo Johanson observed that patients with a middle ear injury who sought treatment late often had chronic otitis media and poor hearing results. (5) Our findings were in concordance, as more than one-half (59%) of our delayed group had purulent otorrhea at presentation. Moreover, nearly one-quarter (23%) of the delayed group had cholesteatoma. Finally, ossicular discontinuity was much more common in the delayed group (55 vs. 18%). These findings likely contributed to the poorer hearing at presentation in the delayed group.

Penetrating trauma has been reported to be the most common cause of middle ear injury through the external auditory canal. (4) This was true in our study regardless of the time of presentation. These injuries are usually accidental and self-inflicted, and they frequently occur when patients use a slender object to scratch an itch or clean the ear canal. Perforations are the rule, and ossicular discontinuity is common. Penetrating trauma is the most common cause of luxation of the stapes into the vestibule and resultant perilymph fistula. (7) CT may support the diagnosis and reveal displacement of the stapes or pneumolabyrinth in these patients. (8) However, direct trauma to the round or oval window may cause a perilymph fistula without stapes luxation. (15) In our study, both patients with luxation of the stapes and all 3 with a perilymph fistula had penetrating injuries.

Slag burns often cause thermal injuries to the external canal, tympanic membrane, and middle ear structures during welding. Facial nerve paralysis as a result of a direct thermal trauma has also been described. (10,11) The temperature of molten slag can reach 1,000[degrees]C; cauterization of the tympanic membrane remnant can result. (16) A cauterized perforation may heal slowly and be prone to infection. Frenkiel and Alberti reported that outcomes of treatment--be it medical or surgical--are not as good in patients with thermal injury as they are in patients with other types of injury. (17) The poorer outcomes may be attributable to devascularization and a poor supply of nutrition to the recipient graft site. In many cases, however, slag trauma causes only a minor perforation and may be ignored or unnoticed by the patient. (18) In our study, all 5 patients with acute perforations caused by slag experienced spontaneous healing; however, 1 delayed-group patient had a failed tympanoplasty and persistent otorrhea. Slag trauma can be easily avoided by simply wearing ear protection.

Blast injuries through the external auditory canal may be explosive or nonexplosive. Both types are caused by a rapid positive pressure wave exerted through the external auditory canal. (1,2) Tympanic membrane perforations are common, and significant trauma to the middle and inner ear structures may occur. Both types of blast injury have been associated with noise-induced sensory hearing loss, and at times round window membrane rupture. (1,9,19) Explosive injuries are more severe; the severity of the middle ear injury appears to be proportional to the patient's proximity to the blast. (20) Nonexplosive blast trauma is more common, occurring most often during human altercations, sports, and recreational water activities. As previously noted, blast injuries may result in delayed cholesteatoma formation.

Lightning trauma to the middle ear is rare. The rapid expansion of the surrounding air during a lightning strike causes an explosive blast trauma that often results in unilateral or bilateral perforation. (3) Lightning may reach temperatures of 20,000[degrees]C, causing thermal injury, as well, but the blast is believed to be responsible for most of the damage. (3) Facial nerve paralysis and sensorineural hearing loss are less common, and most of the perforations heal uneventfully with myringoplasty or tympanoplasty.

The management goals for patients with middle ear injury are an intact tympanic membrane and a hearing ear. The former is achieved in most patients, although those with welding-slag and explosive blast injuries appear to have poorer rates of closure. Unfortunately, some patients are not as fortunate when it comes to hearing. Regardless of the cause, middle ear trauma through the external auditory canal may result in permanent end-organ damage and in some cases a dead ear. Patients with a perilymph fistula frequently present with cochlear hearing loss; despite immediate fistula repair, the hearing outcome is often poor. Severe penetrating trauma to the stapes and round window also carries a poor hearing prognosis.

Most middle ear trauma patients do not develop a dead ear, and they generally derive benefit when the middle ear sound-conduction mechanism is restored. However, patients who delay seeking treatment do appear to have poorer hearing. In our study, the patients in the delayed group had significantly poorer mean PTAs and ABGs at the initial evaluation and after treatment. Contributing factors likely include the higher incidences of chronic suppurative ear disease, cholesteatoma, and ossicular discontinuity. As a group, these patients appear to require more extensive otologic surgery than do patients who present in an early fashion. Still, regardless of presentation time, most patients who experience middle ear trauma through the external auditory canal do have useful hearing following management.

References

(1.) Berger G, Finkelstein Y, Harell M. Non-explosive blast injury of the ear. J Laryngol Otol 1994;108:395-8.

(2.) Sprem N, Branica S, Dawidowsky K. Tympanoplasty after war blast lesions of the eardrum: Retrospective study. Croat Med J 2001;42:642-5.

(3.) Gluncic I, Roje Z, Gluncic V, Poljak K. Ear injuries caused by lightning: Report of 18 cases. J Laryngol Otol 2001;115:4-8.

(4.) Bellucci RJ. Traumatic injuries of the middle ear. Otolaryngol Clin North Am 1983;16:633-50.

(5.) Brahe Pedersen C, Vendelbo Johanson L. Traumatic lesions of the middle ear: Aetiology and results of treatment. Clin Otolaryngol Allied Sci 1986; 11:93-7.

(6.) Kronenberg J, Ben-Shoshan J, Modan M, Leventon G. Blast injury and cholesteatoma. Am J Otol 1988;9:127-30.

(7.) Yamasoba T, Amagai N, Karino S. Traumatic luxation of the stapes into the vestibule. Otolaryngol Head Neck Surg 2003; 129: 287-90.

(8.) Herman P, Guichard JR Van den Abbeele T, et al. Traumatic luxation of the stapes evidenced by high-resolution CT. AJNR Am J Neuroradiol 1996;17:1242-4.

(9.) Teter DI, Newell RC, Aspinall KB. Audiometric configurations associated with blast trauma. Laryngoscope 1970:80:1122-32.

(10.) Stage J, Vinding T. Metal spark perforation of the tympanic membrane with deafness and facial paralysis. J Laryngol Otol 1986; 100:699-700.

(11.) Panosian MS, Wayman JW, Dutcher PO Jr. Facial nerve paralysis from slag injury to the ear. Arch Otolaryngol Head Neck Surg 1993:119:548-50.

(12.) Lindeman PS, Edstrom S, Granstrom G. et al. Acute traumatic tympanic membrane perforations: Cover or observe? Arch Otolaryngol Head Neck Surg 1987;113:1285-7.

(13.) Camnitz P, Bost WS. Traumatic perforations of the tympanic membrane: Early closure with paper tape patching. Otolaryngol Head Neck Surg 1985:93:220-3.

(14.) Kristensen S. Spontaneous healing of traumatic tympanic membrane perforations in man: A century of experience. J Laryngol Otol 1992; 106: 1037-50.

(15.) Strohm M. Traumatic lesions of the round window membrane and perilymph fistula. Ann Otol Rhinol Laryngol 1986:35:177-247.

(16.) Fisher EW. Gardiner Q. Tympanic membrane injury in welders: Is prevention neglected? J Soc Occup Med 1991:41:86-8.

(17.) Frenkiel S, Alberti PW. Traumatic thermal injuries of the middle ear. J Otolaryngol 1977;6:17-22.

(18.) Kecht B. [On the so-called sweat bead injuries of the ear]. Wien Med Wochenschr 1966:116:572-4.

(19.) Wurtele R Traumatic rupture of the eardrum with round window fistula. J Otolaryngol 1981;10:309-12.

(20.) Miller IS, McGahey D, Law K. The otologic consequences of the Omagh bomb disaster. Otolaryngol Head Neck Surg 2002;126: 127-8.

John M. Lasak, MD; Mark Van Ess, DO; Thomas C. Kryzer, MD; Richard J. Cummings, MD

From the Wichita (Kans.) Ear Clinic (Dr. Lasak, Dr. Kryzer, and Dr. Cummings); the Department of Surgery (Dr. Lasak and Dr. Kryzer) and the Department of Pediatrics (Dr. Lasak), University of Kansas School of Medicine-Wichita; and the Department of Surgery, Via Christi Riverside Medical Center, Wichita (Dr. Van Ess).

Reprint requests: John M. Lasak, MD, The Wichita Ear Clinic, 9350 E. Central Ave., Wichita, KS 67206. Phone: [phone]; fax: [phone]; e-mail: [e-mail]

Table 1. Mechanisms of injury

Early group             n (%)      Delayed group              n (%)

Cotton-tipped           10 (45)    Bobby pin                   4 (18)
  applicator
Welding slag             5 (23)    Welding slag                4 (18)
Rat-tail comb            2 (9)     Foreign-body removal        4 (18)
Key                      1 (4)     Blast*                      4 (18)
Bicycle wheel spoke      1 (4)     Cotton-tipped applicator    4 (14)
Artist's brush handle    1 (4)     Coat hanger                 1 (4)
Twig                     1 (4)     Bicycle wheel spoke         1 (4)
Foreign-body removal     1 (4)     Quill feather               1 (4)

Total                   22 (100)                              22 (100)

* Blast injuries included 1 compressed-air injury, 1 airbag-deployment
injury, 1 fireworks injury, and I lightning-strike injury.

Table 2. Findings at presentation

                          Early group   Delayed group
                             n (%)          n (%)

Subjective hearing loss    22 (100)       22 (100)
Vertigo                     7 (32)         2 (9)
Tinnitus                   10 (45)         9 (41)
Otorrhea                    8 (36)        13 (59)
Perforation                22 (100)       17 (77)
Ossicular injury            4 (18)        12 (55)
Tympanosclerosis            2 (9)         10 (45)
Cholesteatoma               0              5 (23)
Facial nerve injury         0              2 (9)
Nystagmus                   2 (9)          0
Perilymph fistula           2 (9)          1 (5)

Table 3. Type of perforation

                           Early group   Delayed group
                              n (%)          n (%)

Small central                13 (59)         2 (9)
Subtotal                      8 (36)          12
Total/near total              1 (5)            3
None                            0              5

Table 4. Management

                           Early group   Delayed group
                              n (%)          n (%)

Observation                  13 (59)         0
Surgery *                     9 (41)        22 (100)
Paper patch                   5 (23)         0
Tympanoplasty                 0             12 (55)
Tympano-ossiculoplasty        4 (18)         9 (41)
Tympanomastoidectomy            0            3 (14)
Labyrinthectomy               1 (5)          0
Fistula repair                2 (9)          1 (5)

* 2 early-group patients (9%) and 3 late-group patients (14%)
required more than one trip to the operating room.

Table 5. Hearing data at presentation and follow-up

                       Early group           Delayed group
                           (dB)                  (dB)           p Value

Presentation
Mean (range) PTA   30.7 (6.25 to 78.75)     52.2 (25 to 88)     0.0007
Median PTA                 28.0                  48.0
Mean (range) ABG      14.6 (0 to 30)       28.2 (0 to 51.3)     0.0045

Follow-up
Mean (range) PTA     21.0 (6.3 to 50)     42.5 (2.5 to 107.5)   0.0044
Median PTA                 16.0                  34.0
Mean (range) ABG      8.0 (0 to 30)        17.2 (0 to 56.3)     0.0092