Shelton 2007 the Surgical Learning Curve in Aural Atresia Surg

The LaryngoscopeLippincott Williams & Wilkins, Inc.© 2006 The American Laryngological,Rhinological and Otological Society, Inc.

The Surgical Learning Curve in AuralAtresia Surgery

Nirmal Patel, MBBS (Hons), FRACS; Clough Shelton, MD

Objective: The objective of this retrospective casereview is to examine the effect of surgical learning onhearing outcomes and complications in congenitalaural atresia surgery. Patients: Sixty-four consecu-tive ears (in 60 patients) operated on during the pe-riod of 1994 to 2004 at a tertiary referral center werestudied. Intervention(s): Intervention consisted of au-ral atresiaplasty through an anterior approach by thesame surgeon (C.S.). Main Outcome Measure(s): Hear-ing outcomes and complication rates were comparedbetween four temporally sequential groups of 16 ears.Acceptable hearing and complication rate outcomeswere defined as results comparable to larger series inthe literature. Results: Hearing results, in the shortterm, comparable to larger series were achieved dur-ing the first group of ears (nos. 1–16). A plateau in thelearning curve for short-term hearing outcomes wasachieved after the first two groups, that is, after 32ears. Hearing outcomes, in the long term (>1 year)comparable to larger series, were achieved in the sec-ond group of ears (nos. 16–32). The learning curve forlong-term hearing demonstrated a significant im-provement in outcomes in the final group of 16 earscompared with the first 48 ears. Long-term hearingresults for the final group show closure of the postop-erative air–bone gap to less than 30 dB in 94% of cases.Postoperative complication rates were equivalent tolarger series in the first group of 16 ears and showedno statistically significant difference between thefour groups. There was one patient with sensorineu-ral hearing loss after surgery; there were no anacous-tic ears and no facial palsies in the study group. Con-clusions: A learning curve of at least 32 ears wasrequired to achieve stable short-term hearing results.To achieve stable long-term hearing results requireda learning curve of at least 48 patients in our series.Complication rates remained stable throughout thestudy period. Key Words: Learning curve, congenitalatresia surgery.

Laryngoscope, 117:67–73, 2007

INTRODUCTIONSurgical learning can be broadly viewed as acquiring

technical skills and applying these skills to an anatomicsituation to achieve a desired end result. Anatomic learningin otologic surgery occurs through temporal bone laboratorypractice and operative experience through observing andperforming cases in a variety of pathologic scenarios.

Some literature exists on the surgical learning curveof vestibular schwannoma and stapedectomy surgery.4–9

It may be inferred that learning vestibular schwannomaand stapedectomy surgery is occurring when performingother otologic procedures such as ossiculoplasty and mas-toidectomy with similar anatomy. Congenital aural atre-sia seen in this light presents challenges in that the anat-omy is inconsistent and unique to atresia surgery. No dataare available on the number of cases required to achievestable results in congenital aural atresia surgery.

With this work, we sought to test the hypothesis that alearning curve exists for congenital aural atresia surgery,and stable hearing and complication outcomes can be con-sistently achieved after a period of surgical learning.

MATERIALS AND METHODSAfter University of Utah Institutional Review Board ap-

proval, 160 charts of all patients who presented for aural atresiatreatment from January 1994 to December 2004 were retrospec-tively examined. Sixty-four ears (from 60 patients) with congen-ital aural atresiaplasty were included in the study. Patients wererecruited on a sequential basis from 1994 to 2004. Four studygroups, each with 16 ears, were formed with an equivalent num-ber of atresia grades noted throughout each group. Patients whodid not undergo surgery as a result of facial nerve placement,oval/round window atresia, or who had inadequate audiometricfollow up were excluded.

Microtia repair was completed before atresia surgery inall cases. The senior author (C.S.) was involved in all surgeries.All atresiaplasties were performed using the anterior approachmethod described by Jahrsdoerfer with modifications sug-gested by De la Cruz.2,10,11 Briefly, an external auditory earcanal was fashioned by drilling just posterior to the root of thezygoma and superior to the temporal mandibular joint. Theatretic bone was removed in the region bordered by the middlefossa dura superiorly, glenoid fossa anteroinferiorly, and mas-toid air cell system posteriorly. The fused incudomalleolar com-plex was identified and bone removed from the structure formobilization. If ossicular fixation was present, then ossicularreconstruction using porous polyethylene partial/total ossicular

From Otolaryngology Head & Neck Surgery, University of Utah,Salt Lake City, Utah, U.S.A.

Editor’s Note: This Manuscript was accepted for publication July 18, 2006.Send correspondence to Clough Shelton, MD, Professor and Chair-

man, Division of Otolaryngology Head & Neck Surgery, University of Utah,Salt Lake City, UT 84132, U.S.A. E-mail: [email protected]

DOI: 10.1097/01.mlg.0000240163.73601.27

Laryngoscope 117: January 2007 Patel and Shelton: Congenital Atresia Surgery Learning Curve

67

prostheses (Medtronic Xomed Corp., Jacksonville, FL) covered byautologous cartilage was performed. The neocanal/tympanic mem-brane was reconstructed using autologous temporalis fascia andsplit-thickness skin (varying from 0.008–0.012 inches in thickness)either piecemeal or with one contiguous piece. Meatoplasties wereperformed in most cases. Canal packing was achieved with Silasticsheeting, Gelfoam, and either 3% Xeroform (Bismuth Tribromo-phenate) or Merocel expandable packing (Medtronic Xomed Corp.).Facial nerve monitoring was performed during all surgeries.

All hearing results were reported in compliance with theAmerican Academy of Otolaryngology–Head and Neck Surgerycommittee on hearing and equilibrium guidelines for evaluationof conductive hearing loss and hearing preservation after acousticneuroma surgery (Fig. 1).12,13 Short-term hearing analysis wasundertaken at 3 months postoperatively in all cases. Long-termhearing outcomes were defined as greater than 1 year postatre-siaplasty. Acceptable hearing and complication outcomes weredefined as those comparable to larger series.1–3,10,14,15

All statistical comparison was made using the nonparamet-ric Kruskal-Wallis analysis of ranks with post hoc Tukey test; theP value was set at �.05 to achieve significance.

RESULTSThere were 26 females and 38 males with a mean age

of 18 (range, 4–47 years; standard deviation [SD] � 13.1).There were 45 right and 19 left ears. No statisticallysignificant difference in atresia severity grades (I to III)was noted between the four groups. The number of patientexclusions was not statistically significantly different be-tween the four groups. Three patients had a craniofacialabnormality associated with the aural atresia, includingGoldenhars, Crouzon’s, and/or hemifacial microsomia.

There were 54 primary and 10 revision surgeries.Revision surgeries were divided equally among the fourgroups. There was no significant difference in ossicularreconstruction numbers between the four study groups.

No prosthesis extrusions were noted during the studyperiod. There was no significant difference in split skinthickness, meatoplasty size, or canal packing materialbetween the four sequential groups.

The mean preoperative air–bone gap (ABG) was 49.5dB (range, 32–66 dB; SD � 11.9). Short-term audiologicfollow up was achieved in all 64 ears. No ABG was com-pletely closed or over closed in the study group. The meanshort-term postoperative ABG was 24.4 (range, 10–48 dB;SD � 8.7). Short-term hearing outcomes equivalent tolarger series were achieved in the first group of 16 ears.Good and serviceable (class A and B) hearing was obtainedin 60 of 64 ears (94%) in the short-term audiologic follow-up. Statistically significant improvement in short-termhearing outcomes were achieved in 32 patients.

Long-term audiologic follow up occurred with 52 of 64ears (81%). The mean long-term postoperative ABG was26.7 dB (range, 9–55 dB; SD � 11.7). Long-term hearingresults equivalent to larger series were achieved in thesecond group of 16 ears. Statistically significant improve-ment in long-term hearing outcomes occurred after 48ears. Good and serviceable (class A and B) hearing wasobtained in 42 of 52 ears (81%) in the long term. Conduc-tive hearing loss outcomes subdivided by sequentialgroups are displayed in Figures 2, 3, and 4. Hearing classoutcomes are displayed in Figures 5, 6, and 7.

Complications were uncommon in this series. Therewas one patient with a postoperative sensorineural hear-ing loss limited to high frequency. There were no cases ofpostoperative facial nerve injury.

Problems with healing occurred in 28 of 64 ears(44%). These included perforation (one ear), myringitis/granulation (eight ears), blunting (nine ears), and lateral-ization (10 ears). Healing problem rates comparable tolarger series were achieved in the third group of patients,although the rate differences between the four groups didnot reach statistical significance. Revisions of the seniorauthor’s own patients were lower after 48 ears, but thisdid not reach statistical significance.

DISCUSSIONBy standardizing reconstruction materials and

methods throughout the study period, we aimed to ex-amine the effect of operative learning on outcomes incongenital aural atresia surgery. Most larger series ofcongenital aural atresia surgeries reported have an an-nual incidence of aural atresia surgeries ranging from3.8 up to 10 ears per year, with the exception of Jahrs-dorfer, who has published an annual surgical atresiarate of 20 ears.2,3,10,14,15 An inference can therefore bemade that most neurotologic fellowship trainees will beinvolved in between eight and 20 atresia surgeries dur-ing their 2-year training period. Furthermore, our datasuggest that approximately 5 years of experience (whenoperating on six ears per year) are required to achieveconsistent short-term hearing outcomes and up to 8years (when operating on six ears per year) of operativeexperience is required to achieve stable long-term hear-ing results. Familiarities with the subtle but significantanatomic variations that occur with atresia surgery, aswell as refinement of surgical technique in exposure and

Fig. 1. The American Academy of Otolaryngology Head and NeckSurgery classification system for hearing after acoustic neuromasurgery. The vertical axis represents the four tone hearing average(500, 1,000, 2,000, and 3,000 Hz) and the horizontal axis is wordrecognition score (%).


68

reconstruction methods, signify the major changes thathave occurred in operative technique during the learn-ing period for the senior author (C.S.).

Over the course of the study period, our techniquehas evolved, instituting many of the suggestions of Teufertand de la Cruz. These include the use of thinner split-

Fig. 2. Overall air–bone gap outcomes comparing preoperative with short- and long-term gaps.

Fig. 3. Short-term air–bone gap learning curve. Short-term air–bone gap outcomes demonstrating a plateau in gaps in the third and fourthgroups, that is, after 32 ears.


69

thickness skin grafts (0.008 inch), creating a meatus 12mm in diameter, and use of Merocel expandable meatalpacking. We now create a tympanic ring as large as pos-sible, removing bone to the limits of the temporal mandib-ular joint anterior inferior, facial nerve posterior, andmiddle fossa superior. When lining the canal, redundant

skin graft is avoided and a Gelfilm disk is used to hold thegraft in the anterior sulcus.

Most objective methods of assessing hearing exam-ine the relative amount of change from preoperativelevels. This method negates the selection bias derivedfrom studying patients with different preoperative lev-

Fig. 4. Long-term air–bone gap learning curve. Long-term air–bone gap outcomes compared as percentages between each group. Note thesignificant difference between the first three groups and last group, illustrating a significant improvement in hearing results in the last group.

Fig. 5. Overall hearing class outcomes comparing preoperative with short- and long-term hearing class results.


70

els. Air– bone gap closures are the recommendedmethod of reporting results for conductive hearing sur-gery. Stapedectomy hearing results are not a good anal-ogy for congenital aural atresia because closure andoverclosure are not realistically achievable in the set-ting of congenital aural atresia (CAA) surgery. Chronicear hearing goals are a better analogy in that the re-construction methods and subsequent anatomy are sim-ilar to CAA surgery.

Parallels may be drawn between hearing results inunilateral congenital atresia surgery and vestibularschwannoma surgery. In both cases, the hearing in thecontralateral ear is typically normal and the goal is toprovide a hearing level in the operative ear to result ina useful binaural affect.16 The binaural affect results insound direction cues allowing better hearing perfor-mance in the presence of background noise. The practi-cal implication for children with congenital aural atre-

Fig. 6. Short-term functional hearing class outcomes demonstrating a plateau in class outcomes in the third and fourth groups (that is, after 32 ears).

Fig. 7. Long-term functional hearing class outcomes compared as percentages between each group. Note the significant difference betweenthe first three groups and last group, illustrating a significant improvement in hearing results in the last group.


71

sia is better hearing performance in a noisy classroomsituation. In this article, as well as examining hearingoutcomes for conductive hearing loss surgery, we soughtto apply the guidelines of the American Academy ofOtolaryngology–Head and Neck Surgery for reportingthe results of hearing preservation after acoustic neu-roma surgery13 to hearing results after correction ofcongenital aural atresia to determine serviceability ofatresiaplasty hearing outcomes. What is considered agood hearing outcome in patients with bilateral auralatresia, however, is different in that any serviceablehearing is a functional improvement on the premorbidsituation.

Class A hearing is defined as a speech receptionthreshold (SRT) of �30 dB and speech discriminationscore (SDS) of �70%. The rationale for choosing this figureis that a patient with an SRT below 30 is consideredsocially handicapped and a candidate for a hearing aid. Aclass A result, with the patient reporting binaural hear-ing, improvement in difficult acoustic environments andwith sound localization should be considered a successful“useful” outcome. Although restricted by the limitations ofa retrospective series, we were able to demonstrate audio-logically useful improvement in the short term for 74% ofpatients. Furthermore, class A and B (good and service-able) outcomes were achieved in 94% of patients in theshort term. When examined in terms of conductive hear-ing loss surgery and ABG closure, we achieved 73% ofpatients in the short term with an ABG closure of �30 dB,a figure considered a successful outcome in most largeraural atresia series. A valid critique of hearing resultanalysis in patients with aural atresia is what exactlyconstitutes useful hearing in these patients, especiallywhen a very good low morbidity alternative, bone-anchored hearing aid (BAHA), is available.17–19 This issuewas not addressed in our work, because patients withBAHA were excluded from the study group; however, wetypically use BAHAs in patients with unreconstructableatresias or patients who choose it as a method ofrehabilitation.

Typically, hearing results tend to gradually declineto a stable plateau in all forms of hearing reconstruc-tion. This has been well described in the CAA literature,and our data within its limitations support this find-ing.1–3 On average, class A hearing results wereachieved in 43% of patients in the long term and class Aand B outcomes in 58% of patients. ABG closures to �30dB in the long term were achieved in 55% of patients.Long-term hearing rates compare favorably with largerseries in the last two groups of 16 ears. Examining thedata with respect to long-term audiologically usefulhearing in unilateral aural atresia for the entire group,38% of patients achieved class A hearing with no re-quirement for a hearing aid. Long-term hearing resultsfor the final group of patients show closure of the ABG�30 dB in 94% and class A hearing in 56%. No patientsin our series experienced significant morbidity, and anypatient with a poor long-term hearing outcome couldstill be postoperatively fit with a conventional air con-duction hearing aid.

Healing problems in our series occurred in 44% ofpatients. Although there was a trend toward the last 32ears having a lower rate of healing problems, this didnot reach statistical significance. Of note, the last twogroups of 16 patients had a shorter follow-up periodthan the first two groups. Most of the healing compli-cations after atresia surgery such as lateralization andblunting can occur up to 3 years after the operation.Therefore, despite the trend toward statistical signifi-cance, these data could experience selection bias, and infact, the four groups may be very similar in terms ofpostoperative healing outcomes.

CONCLUSIONBy standardizing materials and method through-

out the study period, some evidence for the developmentof a learning curve with respect to hearing outcomes incongenital aural atresia surgery has been demon-strated. These data provide a partial answer to thestudy question of whether predictable surgical out-comes can be achieved in the setting of the particularlyunique anatomic characteristics found in patients withaural atresia.

AcknowledgmentsThe authors thank Susan Thiebault, PhD, for her

assistance in statistical analysis.

BIBLIOGRAPHY1. Lambert PR. Congenital aural atresia: stability of surgical

results. Laryngoscope 1998;108:1801–1805.2. De la Cruz A, Teufert KB. Congenital aural atresia surgery:

long-term results. Otolaryngol Head Neck Surg 2003;129:121–127.

3. Teufert KB, De la Cruz A. Advances in congenital auralatresia surgery: effects on outcome. Otolaryngol Head NeckSurg 2004;131:263–270.

4. Hughes GB. The learning curve in stapes surgery. Laryngo-scope 1991;101:1280–1284.

5. Buchman CA, Chen DA, Flannagan P, et al. The learningcurve for acoustic tumor surgery. Laryngoscope 1996;106:1406–1411.

6. Puls T. Stapes surgery: results when performing a moderatenumber of stapedectomies. Acta Otorhinolaryngol Belg1997;51:23–25.

7. Welling DB, Slater PW, Thomas RD, et al. The learning curvein vestibular schwannoma surgery. Am J Otol 1999;20:644–648.

8. Kanzaki J, Inoue Y, Ogawa K. The learning curve in post-operative hearing results in vestibular schwannoma sur-gery. Auris Nasus Larynx 2001;28:209–213.

9. Elsmore AJ, Mendoza ND. The operative learning curve forvestibular schwannoma excision via the retrosigmoid ap-proach. Br J Neurosurg 2002;16:448–455.

10. Jahrsdoerfer RA, Yeakley JW, Aguilar EA, et al. Gradingsystem for the selection of patients with congenital auralatresia. Am J Otol 1992;13:6–12.

11. Jahrsdoerfer RA, Kesser BW. Issues on aural atresia forthe facial plastic surgeon. Facial Plast Surg 1995;11:274 –277.

12. Committee on Hearing and Equilibrium guidelines for theevaluation of results of treatment of conductive hearingloss. American Academy of Otolaryngology–Head andNeck Surgery Foundation, Inc. Otolaryngol Head NeckSurg 1995;113:186–187.


72

13. Committee on Hearing and Equilibrium guidelines for theevaluation of hearing preservation in acoustic neuroma (ves-tibular schwannoma). American Academy of Otolaryngology–Head and Neck Surgery Foundation, Inc. OtolaryngolHead Neck Surg 1995;113:179–180.

14. Shih L, Crabtree JA. Long-term surgical results for con-genital aural atresia. Laryngoscope 1993;103:1097–1102.

15. Chandrasekhar SS, De la Cruz A, Garrido E. Surgery ofcongenital aural atresia. Am J Otol 1995;16:713–717.

16. Shelton C, Brackmann DE, House WF, et al. Acoustic tumor

surgery. Prognostic factors in hearing conversation. ArchOtolaryngol Head Neck Surg 1989;115:1213–1216.

17. Wazen JJ, Caruso M, Tjellstrom A. Long-term results withthe titanium bone-anchored hearing aid: the US experi-ence. Am J Otol 1998;19:737–741.

18. Lustig LR, Arts HA, Brackmann DE, et al. Hearing rehabil-itation using the BAHA bone-anchored hearing aid: resultsin 40 patients. Otol Neurotol 2001;22:328–334.

19. Wazen JJ, Spitzer J, Ghossaini SN, et al. Results of thebone-anchored hearing aid in unilateral hearing loss. La-ryngoscope 2001;111:955–958.


73

Documents

Shelton 2007 the Surgical Learning Curve in Aural Atresia Surg