1 3

Eur Arch OtorhinolaryngolDOI 10.1007/s00405-014-3007-5

OtOlOgy

Perilymphatic fistulas: can we predict the diagnosis?

Musaed Alzahrani · Raphaelle Fadous · Jean‑Jacque Dufour · Issam Saliba

Received: 30 December 2013 / Accepted: 8 March 2014 © Springer-Verlag Berlin Heidelberg 2014

Keywords Perilymphatic fistula · Fluctuating hearing loss · Vertigo · Stapedectomy · Head trauma · Barotrauma · Dizziness

Introduction

Perilymphatic fistulas (PlF) are defined as an abnormal leak of perilymphatic fluid from the inner to the middle ear, most commonly via either the round or oval window. Since it was first described in 1897 [1], preoperative diagnosis of PlF remains difficult to make with certainty. Many etio-logic factors have been incriminated in causing PlF, such as barotrauma, direct head trauma, acoustic trauma, or after stapedectomy [2–7].

Perilymphatic fistulas can be associated with any combi-nation of hearing loss, vertigo, and tinnitus [8]. A positive fistula sign, defined as the presence of a nystagmus after application of positive pressure to the external auditory canal with a pneumatic otoscope, is a possible indication of a fistula. Nonetheless, it predicts a fistula with varying accuracy, ranging from no predictive value to a sensitivity of 77 % [5, 8].

High-resolution Ct scan can sometimes provide clues to diagnose PlF like pneumolabyrinth, but this test is normal most of the time [9]. Hearing loss is a frequent symptom of the PlF, but is not always present. the audiogram can show any form of hearing loss [5]. It may be fluctuating, as the Fraser test intends to illustrate, but the low sensitivity of this test has limited its use [10].

Albera et al. demonstrated vestibular asymmetry on electronystagmography in 44 % of patients in a surgically proven fistula [3].

Exploratory transmeatal surgery intends to identify a fis-tula either in the oval or the round window [11]. this can be

Abstract this study intends to identify factors that could better predict the diagnosis of perilymphatic fistula (PlF) since exploration surgery is currently the only confirmatory method. this retrospective chart review in a tertiary care center is based on all 71 available patient files operated for a suspicion of PlF between 1983 and 2012. History of pre-disposing factors, clinical findings and investigations were documented pre- and postoperatively. Patients were divided according to intraoperative findings into two groups: group I (fistula negative) and group II (fistula positive). In addi-tion, group II was divided into two subgroups: patients with or without a history of stapedectomy. Both groups were demographically similar. With the exception of history of previous partial stapedectomy (p = 0.04), no statistical dif-ference could be identified in predisposing factors and in clinical findings between the two groups. the evolution of symptoms showed an overall improvement of vestibular symptoms (91 %) and cochlear symptoms (53 %) postop-eratively. Audiograms showed a significant improvement postoperatively in the pure tone audiometry and bone con-duction threshold of group II while the air-bone gap and speech discrimination score did not improve. group I did not show any significant improvement postoperatively in any audiogram parameter. this study failed to identify fac-tors that could better predict the diagnosis of PlF. How-ever, it shows that middle ear exploration with oval and round window obliteration is effective in PlF especially to decrease vestibular symptoms even when fistula is uniden-tified intraoperatively.

M. Alzahrani · R. Fadous · J.-J. Dufour · I. Saliba (*) Division of Otolaryngology—Head and Neck Surgery, Montreal University Hospital Center (CHUM), University of Montreal, 1560, Sherbrooke East Street, Montreal, QC H2l 4M1, Canadae-mail: issam.saliba@umontreal.ca

Eur Arch Otorhinolaryngol

1 3

realized in the operating room or as an office-based proce-dure using the otoendoscope, a fast and available diagnostic tool through a small myringotomy incision [12]. Most fre-quently, both windows are obliterated, even in the absence of perilymphatic fluid visualization. Postoperative improve-ment in symptoms led some authors to think that the fistula may be too small to be visualized [13]. gehrkin et al. tried to enhance the visualization of perilymphatic fluid by inject-ing intrathecal fluorescein before surgery. However, fluo-rescein was visible in only 7 % of patients, and is therefore not recommended [14]. Bachmann-Harildstad et al. [15] tested beta-trace protein as a marker for PlF, but showed lit-tle diagnostic sensitivity due to the low available volume of perilymph. However, Ikezono et al. have identified cochlin-tomoprotein as a possible specific marker for PlF. Cochlin-tomoprotein expression was absent in different body fluids, including saliva and serum. It was selectively expressed in perilymphatic fluid collected from stapedectomy, translaby-rinthine, and cochleostomy procedure [16].

the most commonly reported symptoms are nonspe-cific and the few available tests may point to the diagnosis, but cannot confirm its presence with certainty [3, 9, 10]. Exploratory surgery remains the ideal option to confirm the diagnosis [11]. this retrospective chart review intended to identify elements from history of predisposing factors, the clinical findings, and the preoperative tests that could better predict the presence of a PlF, to avoid unnecessary surgical exploration.

Methods

We have performed a retrospective chart review of all avail-able patients at our tertiary care center operated for suspi-cion of PlF. A total number of 71 patients were identified between 1983 and 2012. Of these, 13 patients have been operated for PlF up to five interventions. the parameters listed below were studied for the first exploratory surgery. We divided our population into two groups: groups I (fistula exploration negative) and II (fistula exploration positive). In addition, we stratified group II into two subgroups: (a) patients with or (b) without a history of stapedectomy. the standard technique in our institution after 2003 is a stapedot-omy using a 0.7 mm skeeter drill. the stapedotomy is then sealed with pieces of gelfoam around the prosthesis piston. However, before that date, a partial stapedectomy with poste-rior part of the footplate removal was the regular technique.

Preoperative information

the following predisposing factors were studied: history of cranial trauma, barotrauma, penetrating ear trauma, acous-tic trauma, middle ear disease, previous stapedectomy, and

previous other ear surgeries. Symptoms of vertigo, nausea, vomiting, tinnitus (present or absent), hearing loss, otor-rhea, and the tullio phenomenon were analyzed. Also, the clinical findings of spontaneous nystagmus, nystagmus induced by head shake test or head thrust test, Romberg test, Valsalva sign, and fistula sign were evaluated.

Investigations, when available, included high resolution Ct-scan, looking for pneumolabyrinth, fracture, or penetra-tion of the stapes prosthesis into the oval window. Video-nystagmography (VNg) recordings of spontaneous nystag-mus, positional nystagmus, or a vestibular deficit were also documented. Finally, audiometry was used to classify hear-ing loss as sensorineural (SNHl), conductive, and mixed. the mean values of pre and postoperative bone conduction (BC) and air conduction (AC) thresholds at the frequencies of 500, 1,000, 2,000, and 4,000 Hz were analyzed and used to calculate the pure tone average (PtA) and the air-bone gap (ABg). Because of the possibility of fluctuating hearing loss, the preoperative audiogram values were taken from the worst audiogram in the year preceding the surgery.

Intraoperative information

Intraoperative details, such as the presence or absence of a fistula, and if present, the window at which it occurred were recorded. the patched window and the tissue used were also described.

Postoperative information

the postoperative signs and symptoms recorded were the same as the ones in the preoperative period in order to eval-uate surgical efficacy. the Ct scans were removed from the postoperative evaluation, due to the insufficient number performed.

Statistical analysis

the statistical analysis was performed using SPSS statistics® (version 20). Fisher’s exact test was used to detect statistical difference between both groups according to history, clini-cal findings, and investigations. the evolution of symptoms was evaluated with McNemar’s test. History, clinical find-ings, and investigations were compared between the two subgroups of group II using the Breslow–Day homogeneity test and an odds ratio was estimated using Mantel–Haenszel testing. A p < 0.05 was considered significant.

Results

Seventy-one patients were included: 17 patients in group I and 54 patients in group II. As shown in table 1, there

Eur Arch Otorhinolaryngol

1 3

was no statistical difference between the two groups when compared for age, sex, or operated side. table 2 demon-strates the predisposing factors, such as: head trauma, barotrauma, acoustic trauma, penetrating trauma, middle ear disease, previous stapedectomy, and other ear surger-ies. History of a previous partial stapedectomy was asso-ciated with higher risk of PlF (p = 0.04). However, the other predisposing factors showed no statistical difference between the two groups. Patients presenting a history of head trauma either had a car accident or a direct blow to the head. Most patients with a history of barotrauma had an explosive mechanism (n = 10), four patients had an implo-sive mechanism, such as during scuba diving (n = 3) or an airplane trip (n = 1), and the three remaining barotrauma patients did not have the details of the incident recorded in the medical chart.

As demonstrated in table 3, multiple symptoms and clinical findings were studied at initial presentation, but none happened to be statistically different between the two groups.

Intraoperative findings

Fistula was identified in 76 % of the patients (n = 54). the majority (66.5 %) were from the oval window (n = 37), while 31.5 % were from the round window (n = 17), and 2 % were from both windows (n = 1).

However, both windows were obliterated in all cases regardless of the presence or absence of fistula. the materials used for obliteration are listed in table 4. Ana-lyzes of the graft material applied for the PlF treatment did not show any difference to decrease the recurrence rate (p = 0.1).

Evolution of symptoms and clinical findings (table 5)

Postoperative improvement of the vestibular symptoms was achieved in 93 % of all patients (both groups combined) while only 53 % had cochlear symptoms improvement.

For group I, vestibular symptoms such as vertigo, nau-sea, and vomiting were resolved completely in all patients (p < 0.05). Also, tinnitus showed a significant improvement postoperatively (p = 0.04). However, although 6 out of 11 patients reported stabilized fluctuation or improved hear-ing, it was not statistically significant (p = 0.07).

In group II, vertigo was improved postoperatively in 91 % of patients (p < 0.05). Vomiting was resolved post-operatively in all cases (p = 0.006). Nausea was improved in all patients postoperatively (p < 0.001), and one patient who did not have nausea preoperatively developed it in the postoperative period. Hearing loss, either fluctuating or constant, was improved postoperatively in 23 of 44 patients (p = 0.04). Similarly, tinnitus showed a statistically signifi-cant improvement in 19 of 35 patients (p = 0.001).

there were an insufficient number of charts mention-ing the tullio phenomenon, the head shake test, head thrust test, Romberg test, tandem test, Valsalva, fistula sign, and VNg to be analyzed. However, spontaneous nystagmus was recorded in 3 patients out of 12 charts. It was resolved in two of them.

When comparing the two subgroups of group II (previ-ous stapedectomy vs. no previous stapedectomy), no sta-tistical difference could be identified by Mantel–Haenszel odds ratio for age, gender, barotrauma, or head trauma. the symptoms of vertigo, nausea, vomiting, tinnitus, and hearing loss had no predictive value on the presence of fis-tula when controlled for the presence or not of previous stapedectomy.

Table 1 Demographics data

n number

Parameter group I (n = 17)

group II (n = 54)

total (n = 71) p value

Age (years) 43.5 ± 8.5 44.4 ± 12.4 44.2 ± 11.5 0.79

gender

Male 35.3 % 48.1 % 45.1 % 0.41

Female 64.7 % 51.9 % 54.9 %

Side

left 64.7 % 66.7 % 66.2 % 1

Right 35.3 % 33.3 % 33.8 %

Table 2 Predisposing factors for perilymphatic fistula

n number of patients

Parameter group I group II total p value

(n = 17) (n = 54) (n = 71)

Head trauma 29.4 % (n = 5) 16.7 % (n = 9) 19.7 % (n = 14) 0.3

Barotrauma 35.3 % (n = 6) 20.4 % (n = 11) 23.9 % (n = 17) 0.32

Acoustic trauma 0 % (n = 0) 3.7 % (n = 2) 2.8 % (n = 2) 1

Penetrating trauma 0 % (n = 0) 5.6 % (n = 3) 4.2 % (n = 3) 1

Stapedectomy 64 % (n = 11) 37 % (n = 20) 43.6 % (n = 31) 0.04

Chronic ear disease 0 % (n = 0) 22.2 % (n = 12) 16.9 % (n = 12) 0.06

Other ear surgery 0 0 0 –

Eur Arch Otorhinolaryngol

1 3

Audiometry

Postoperative audiograms in both groups are summarized in table 6. For group I, we did not find any statistically

significant improvement in the BC, AC, or ABg. How-ever, the PtA and BC threshold were improved postopera-tively and were statistically significant (p = 0.01 and 0.03, respectively) in group II. We also evaluated the evolution of

Table 3 Clinical findings at initial presentation of the perilymphatic fistula

+ positif test

Parameter group I n = 17, (%)

group II n = 54, (%)

total n = 71, (%)

Number of charts

p value

Vertigo 82.4 64.8 69.0 71 0.24

Nausea 29.4 25.9 26.8 71 0.76

Vomiting 23.5 13.0 15.5 71 0.44

tinnitus 64.7 66.7 66.2 71 1

Hypoacousis

Constant 52.9 72.2 67.6 71 0.30

Fluctuating 11.8 9.3 9.9

tullio phenomenon 0.0 66.7 50 4 1

Otorrhea 0.0 9.3 7.0 71 0.33

Spontaneous nystagmus 33.3 35.0 34.6 26 1

Head shake test (+) 0.0 50.0 42.9 7 1

Head thrust test (+) 0.0 0.0 0.0 2 N/A

Romberg test (+) 66.7 33.3 40.0 15 0.53

tandem test (+) 0.0 50.0 50.0 2 N/A

Valsalva induced nystagmus (+) 100 40.0 53.8 13 0.19

Fistula sign (+) 50.0 61.9 58.6 29 0.68

Table 4 graft used for obliteration of inner ear windows

Some ears were operated more than one time

Material Number of operated ears (%)

Number of revision surgery (%)

temporalis fascia 34 (45.3) 9 (26.5)

temporalis fascia and bone pate 2 (2.6) 2 (100)

temporalis fascia and muscle 2 (2.6) 0 (0)

Muscle graft 1 (1.3) 0 (0)

Fat graft 22 (29.3) 3 (13.6)

Muscle and fat 1 (1.3) 1 (100)

tragal perichondrium 7 (9.3) 1 (14.3)

gelfoam 4 (5.2) 0 (0)

Venous graft 1 (1.3) 1 (100)

Periosteum and tissel 1(1.3) 1 (100)

Table 5 Evolution of symptoms postoperatively

n number of patients

group group I (fistula negative) (n = 17) group II (fistula positive)

No previous stapedectomy (n = 33) Previous stapedectomy (n = 21)

Symptoms Initial presentation Postoperative p value Initial presentation Postoperative Initial presentation Postoperative p value

Vertigo 14 0 0.001 21 4 14 1 0.001

Nausea 5 0 0.01 25 1 15 0 0.001

Vomiting 4 0 0.03 5 0 2 0 0.006

Hearing loss 11 5 0.07 26 13 17 7 0.001

tinnitus 11 5 0.04 24 10 11 6 0.001

Eur Arch Otorhinolaryngol

1 3

audiograms for each frequency. By comparing single fre-quencies, group II BC and AC had a significant improve-ment in every frequency except 4,000 Hz (Figs. 1, 2). the ABg was also improved although it did not reach a statisti-cal significance.

Discussion

Perilymphatic fistula represents a diagnostic challenge as it can present with any association of hearing loss, tinni-tus, and vertigo [8, 17]. Many factors have been reported to cause PlF, such as head trauma, chronic ear disease, and previous stapedectomy [3, 10, 13]. Excluding previous sta-pedectomy, we were able to identify a predisposing factor in 68 % of the patients (head trauma: 19.7 %, barotrauma: 23.9 %, penetrating trauma: 4.2 %, acoustic trauma: 2.8 %, chronic ear disease: 16.9 %). However, no statistical differ-ence was found between the two groups.

Fistula was identified in 76 % of our patients. the majority of these fistulas were localized to the oval window. this might be explained by the high number of patients presenting with barotrauma and previous stapedectomy. We believe that absent PlF in group I was a false nega-tive finding. this is because of the remarkable postopera-tive improvement; especially, the vestibular symptoms. It is probably a microfistula that is too small to be visualized or a fluctuating fistula that disappears in the lying posi-tion during surgery. In addition, sometimes it is confused with middle ear exudate and local anesthetics. Fitzgeralds et al. [18] reported a negative exploration in 78 % of 197 patients. they obliterated both windows in all patients and reported postoperative improvement of both vestibular (87 %) and auditory symptoms (40 %). Such findings led to redefine the diagnosis of PlF as postoperative symptomatic improvement rather than an intraoperative finding [18].

Barotrauma may cause PlF either by an “implosive” mechanism, described as being an increase in middle ear pressure, or an “explosive” mechanism, which is an increased pressure in the perilymph compartment [19].

Our series demonstrates a history of barotrauma in 23.9 % of patients, which is higher than what was previously reported in the literature [8]. lehrer et al. showed that 51 % of their patients with confirmed fistulas had a history of head trauma [7]. this proportion is much higher than the 19.7 % we obtained form our series. Our result is more consistent with the results of Seltzer et al., in which 23 % of patients had a history of trauma: 11 % of them being head trauma, 9 % barotrauma, and 2 % a direct trauma to the ear [7]. Acoustic trauma and erosion caused by a cho-lesteatoma can also lead to PlF [4, 5, 10]. Acoustic trauma was incriminated in causing a PlF in 1 % of patients in the series by Seltzer et al., a proportion similar to ours [8].

Table 6 Hearing level preoperatively and postoperatively in both groups

PTA pure tone average, BC bone conduction, ABG air-bone gap, SDS speech discrimination score, SD standard deviation

group Hearing level Preoperatively Postoperatively p value

Mean SD Mean SD

group I (fistula negative)

PtA (dB Hl) 48.5 30.4 57.5 40.3 0.26

BC threshold (dB Hl) 34.5 23.8 39.5 30.3 0.35

ABg (dB Hl) 13.9 13.7 17.9 21.8 0.34

SDS in % 95 10 90 20 0.39

group II (fistula positive)

PtA (dB Hl) 61.8 26.3 46.6 25.3 0.01

BC threshold (dB Hl) 41.1 23.5 31.6 20.9 0.03

ABg (dB Hl) 21 12.5 17.1 14.7 0.12

SDS in % 78.2 29.5 86.8 14.2 0.31

Fig. 1 It shows the hearing thresholds by frequency of the positive fistula group preoperatively and postoperatively as well

Eur Arch Otorhinolaryngol

1 3

Stapedectomy surgery can cause a fistula immediately postoperatively or up to many years after the surgery [2, 3]. In a series from Iowa, 26 % of patients had a previous oto-logic surgery, including stapedectomy, making it their most frequent predisposing factor [17]. In our series, previous stapedectomy was noted in 31 out of 71 patients (43 %). It is the most common predisposing factor for fistula in group II (37 %).

In our series, 69 % of all 71 patients presented with vertigo. It was improved in 100 % of group I and in 91 % of group II patients, which is comparable to the literature [20, 21]. McCabe et al. demonstrated in their series that 82 % of patients had vestibular complaints [8, 17]. they also reported that 83 % of patients with PlF had auditory symptoms, with fluctuating hearing in 36 % of their sub-jects [17]. We found fluctuating hearing loss in 9.9 % of our patients. Bussière et al. reported also 10 % of fluctuat-ing hearing loss [10]. Many other studies report fluctuating audition as having an important predictive value [8, 10, 17]. tinnitus, present in 66.2 % of our patients, was improved in 36 % of subjects. the rate of preoperative tinnitus var-ies according to various studies, ranging from 63 to 80 % [21]. Postoperative hearing loss improvement is also vari-able, ranging from 25 to 73 % [7, 17, 20–22]. In our series, 53 % noticed improved hearing and stabilized fluctuation. No postoperative SNHl loss was noted, however, in the

fistula negative groups, we noticed a slight deterioration in the PtA, BC, and ABg postoperatively, although this was not statistically significant (table 6).

According to Pickuth et al., a pneumolabyrinth seen on Ct scan is the most predictive of a fistula, but this sign is not present in all cases [9]. they also report fluid col-lection in the middle ear and displacement of stapes pros-thesis as being signs of a possible fistula [9]. In our chart review, a major limitation is the insufficient number of charts recording investigations and the evolution of signs and symptoms, making them impossible to interpret. Vide-onystagmography could help directing toward a PlF, as Albera et al. demonstrated by obtaining a vestibular asym-metry in 44 % of patients with a surgically proven fistula [3]. However, this evidence is not consistent in literature, as other authors could not identify a correlation between VNg findings and the presence of a fistula [8, 21]. this test was reported in an insufficient number of charts in our series to be compared.

Management of PlF is controversial. Spontaneous heal-ing of PlF might be achieved with bed rest and sympto-matic treatment [23]. However, inner ear changes may become irreversible due to a collapsing endolymphatic compartment with progressive cochlear and vestibular symptoms. Early exploration seems wise in this condition to obliterate the leaking windows [20, 23]. Park et al. found an auditory improvement in 57 % as compared to 33 % in suspected barotrauma-induced PlF if explored within 10 days [20]. Exploration tympanotomy in our series was very beneficial for both groups. It was more effective for vestibular symptoms (91 %) than cochlear symptoms (53 %). these findings are suggestive that obliterating both windows can help avoid missing an intermittent PlF or a small quantity of fluid in cases of negative exploration. Postoperative improvement in the fistula negative group (100 % of vertigo and 53 % of hearing loss) is suggestive of the presence of a micro-fistula that is unidentified during middle ear exploration.

Different grafting materials have been utilized for PlF repair. temporalis fascia, fat grat, tragal perichondrium, areolar tissue, and gelfoam have reported with successful results [8, 17, 24–26]. to our knowledge, no reported ran-domized study analyzes the graft material in the treatment of PlF and the postoperative result. In our series, 13/71 patients (18 %) had single or multiple revision surger-ies for persistent PlF, the material utilized in such cases are detailed in table 4. No statistical difference was found between the graft material applied.

In the absence of a predictive factors of PlF, since the consequences are major as compared to the minor or absent surgical complications, we suggest obliteration of both inner ear windows in all cases, where PlF is suspected and where other differential diagnoses are ruled out.

Fig. 2 It shows the hearing thresholds by frequency of the negative fistula group preoperatively and postoperatively as well

Eur Arch Otorhinolaryngol

1 3

Conclusion

this study shows that exploration tympanotomy with oval and round window obliteration is effective in PlF. How-ever, as many previous studies, it failed to identify factors that could better predict the diagnosis of PlF. A high index of suspicion; especially, in the presence of predisposing factors, is appropriate to perform a middle ear exploration associated with inner ear windows obliteration even in the absence of intraoperative identified PlF. Upon this study, obliteration of inner ear windows is more beneficial than harmful.

Acknowledgments No funding received for this work from any of the following organizations: NIH, Welcome trust, HHMI or other.

Conflict of interest None.

References

1. Ruben RJ, yankelowitz SM (1989) Spontaneous perilymphatic fistula in children. Am J Otol 10(3):198–207

2. Friedland DR, Wackym PA (1999) A critical appraisal of spon-taneous perilymphatic fistulas of the inner ear. Am J Otol 20(2):261–276 (discussion 276–279)

3. Albera R, Canale A, lacilla M, Cavalot Al, Ferrero V (2004) Delayed vertigo after stapes surgery. laryngoscope 114(5):860–862

4. Kung B, Sataloff Rt (2006) Noise-induced perilymph fistula. Ear Nose throat J 85(4):240–241, 245–246

5. Maitland Cg (2001) Perilymphatic fistula. Curr Neurol Neurosci Rep 1(5):486–491

6. Pullen FW 2nd (1992) Perilymphatic fistula induced by baro-trauma. Am J Otol 13(3):270–272

7. lehrer JF, Rubin RC, Poole DC, Hubbard JH, Wille R, Jacobs gB (1984) Perilymphatic fistula—a definitive and curable cause of vertigo following head trauma. West J Med 141(1):57–60

8. Seltzer S, McCabe BF (1986) Perilymph fistula: the Iowa experi-ence. laryngoscope 96(1):37–49

9. Pickuth D, Brandt S, Berghaus A, Spielmann RP, Heywang-Köb-runner SH (2000) Vertigo after stapes surgery: the role of high resolution Ct. Br J Radiol 73(873):1021–1023

10. Bussières R, Portmann D, Noyon P (2003) When to suspect a perilymphatic fistula? Rev laryngol Otol Rhinol 124(4):259–264

11. Al Felasi M, Pierre g, Mondain M, Uziel A, Venail F (2011) Peri-lymphatic fistula of the round window. Eur Ann Otorhinolaryngol Head Neck Dis 128(3):139–141

12. Wall C 3rd, Rauch SD (1995) Perilymph fistula pathophysiology. Otolaryngol Head Neck Surg 112(1):145–153

13. Portmann D, Souza lF, Bussières R, Noyon P (2005) Validation of a clinical scale for the diagnosis of perilymphatic fistula. Rev laryngol Otol Rhinol 126(4):243–248

14. gehrking E, Wisst F, Remmert S, Sommer K (2002) Intraopera-tive assessment of perilymphatic fistulas with intrathecal admin-istration of fluorescein. laryngoscope 112(9):1614–1618

15. Bachmann-Harildstad g, Stenklev NC, Myrvoll E, Jablonski g, Klingenberg O (2011) β-trace protein as a diagnostic marker for perilymphatic fluid fistula: a prospective controlled pilot study to test a sample collection technique. Otol Neurotol 32(1):7–10

16. Ikezono t, Shindo S, Sekiguchi S (2009) Cochlin-tomopro-tein: a novel perilymph-specific protein and a potential marker for the diagnosis of perilymphatic fistula. Audiol Neurootol 14(5):338–344

17. McCabe BF (1989) Perilymph fistula: the Iowa experience to date. Am J Otol 10(4):262

18. Fitzgerald DC, getson P, Brasseux CO (1997) Perilymphatic fis-tula: a Washington, DC, experience. Ann Otol Rhinol laryngol 106(10 Pt 1):830–837

19. Singleton gt, goodhill V (1997) Sudden deafness and round window rupture. (laryngoscope 1971; 81(9):1462–74). laryngo-scope 107(5):577–579

20. Park gy, Byun H, Moon IJ, Hong SH, Cho y-S, Chung W-H (2012) Effects of early surgical exploration in suspected barotrau-matic perilymph fistulas. Clin Exp Otorhinolaryngol 5(2):74–80

21. Black FO, Pesznecker S, Norton t et al (1992) Surgical manage-ment of perilymphatic fistulas: a Portland experience. Am J Otol 13(3):254–262

22. Anon JB, Miller gW (1985) Perilymph fistula. South Med J 78(12):1454–1457

23. Nomura y (1994) Perilymph fistula: concept, diagnosis and man-agement. Acta Otolaryngol Suppl 514:52–54

24. Fitzgerald DC (1996) Perilymphatic fistula in teens and young adults: emphasis on preexisting sensorineural hearing loss. Am J Otol 17(3):397–400

25. gunesh RP, Huber AM (2003) traumatic perilymphatic fistula. Ann Otol Rhinol laryngol 112(3):221–222

26. Weider DJ (1992) treatment and management of perilymphatic fistula: a New Hampshire experience. Am J Otol 13(2):158–166