Surgical Management of Stapedo-Vestibular Reparative Granuloma Following Otosclerosis Surgery: A STROBE Retrospective Study

Abstract

Importance

Reparative granuloma (RG) is a rare but severe complication of otosclerosis surgery, presenting with vertigo, tinnitus, and hearing loss. Surgical intervention may offer effective symptom relief.

Objective

To evaluate the effectiveness of surgical management for RG following otosclerosis surgery in alleviating vestibular symptoms and preserving hearing.

Design

Retrospective cohort study (STROBE guidelines).

Setting

Tertiary referral center.

Participants

Nineteen adult patients with histologically confirmed RG and persistent symptoms (vertigo and/or hearing loss) within 3 months of primary stapes surgery.

Intervention

Revision surgery included granuloma resection, prosthesis replacement, and corticosteroid-soaked graft placement as needed.

Main Outcome Measures

Postoperative relief of vestibular symptoms, hearing thresholds [pure-tone average (PTA)], and length of hospital stay.

Results

Symptoms developed on average 3.6 ± 3.9 days postoperatively. Revision surgery resolved vertigo in 18 of 19 patients (95%) by discharge (mean 4.8 ± 1.9 days). Audiometric improvement was limited, with 11 patients retaining serviceable hearing. Preoperative versus postoperative bone conduction PTA showed no significant change (P = .58).

Conclusions

Surgical management effectively alleviates vestibular symptoms associated with RG but has minimal impact on hearing restoration.

Relevance

This study highlights the efficacy of revision surgery for RG in addressing vestibular symptoms. Further multicenter studies are warranted to better understand reparative granuloma and improve treatment strategies.

Graphical Abstract

Key Messages

• Revision surgery for reparative granuloma (RG) provides rapid and effective relief of vestibular symptoms, improving postoperative recovery.

• After revision surgery for removal of RG, only 37% showed improvement of bone conduction hearing >15 dB.

• Early imaging and surgical intervention are crucial for optimal management of this rare complication.

Introduction

Otosclerosis surgery offers excellent results in terms of hearing, but carries a non-negligible risk of sensorineural hearing loss estimated between 0.5% and 3%. ¹

The origin of this sensorineural hearing loss often remains unknown, and several causes have been advanced: intravestibular prosthesis, leakage of perilymphatic fluid, or the occurrence of stapedo-vestibular reparative granuloma (RG).^1,2

Reparative granuloma is a rare and poorly described entity, despite its supposedly significant morbidity. RG has been reported to occur with a postoperative delay ranging from a few days to 6 weeks, and can be responsible for otalgia, tinnitus, severe vertigo, hypoacusis, and even complete deafness. ³ Its pathophysiology is poorly understood, but several factors have been suggested in the literature, including foreign-body reaction due to the prosthesis or other factors,^2,4,5 inflammation caused by the materials used to seal the footplate, ⁶ perilymphatic fistula, ⁶ gastroesophageal reflux, ⁷ autoimmune reaction, infection, ⁸ and surgical technique. ⁸

Management is empirical and can be medical with or without revision surgery, depending on the center. ⁸

This study aimed to evaluate the outcome of surgical management of stapedo-vestibular RG refractory to standard medical treatments, for auditory and vestibular symptoms.

Methods

We retrospectively included adult patients who underwent stapes revision surgery at a tertiary referral care center, for RG complicating the primary surgery for otosclerosis between 2012 and 2020. This study was approved by the Institutional Review Board “IRB 00012801” under the validation number ID “CE_20161027_2_MDL.” The STROBE guidelines were used for reporting. ⁹ Informed consent was obtained from all the participants included in the study.

In our center, the management algorithm for post-stapes surgery sensorineural complications is as follows:

• - The patient is admitted, and intravenous infusion of corticosteroids and antibiotics is administered.

• - Daily monitoring of clinical, vestibular, and audiological status is performed.

• - A Computed Tomography (CT) scan of the temporal bone is performed for each patient, and if rapidly available, a gadolinium-enhanced MRI of the temporal bone is requested.

Revision surgery was considered when clinical evolution did not improve, and imaging findings were in favor of perilymphatic fistula, an excessively long intravestibular prosthesis, or an RG.

Patients were included if they presented with histological confirmation of RG after revision surgery for otosclerosis (Figure 1), persistent vertigo and/or hearing loss occurring less than 3 months after primary otosclerosis surgery.

Figure 1.

Histological cross-section of reparative granuloma (HES magnification ×50): fibro-inflammatory lesion composed of numerous vessels with hyalization of their walls and sheets of neutrophils. HES, hematoxylin, eosin, saffron.

Clinical and hearing assessments, primary and revision procedure characteristics, imaging reports (CT and MRI), and postoperative results were collected.

Air conduction (AC) and bone conduction (BC) pure-tone averages (PTA) were calculated as the mean hearing thresholds at 0.5, 1, 2, and 3 kHz (or 4 kHz if 3 kHz was not available), following the guidelines of the American Academy of Otolaryngology-Head and Neck Surgery (AAO-HNS). ¹⁰ Air-bone gap was calculated as BC PTA minus AC PTA.

A value of >110 dB HL for AC and 90 dB HL for BC defined the complete hearing loss. Bone and AC averages before and after primary stapes surgery and after revision were collected from the medical files. Speech recognition thresholds (SRT, defined as the minimum hearing level at 50% recognition of the speech material) were collected after revision surgery.

AC averages were classified according to the recommendations of the International Bureau of Audiophonology (available at https://www.biap.org); normal hearing with PTA above 20 dB HL, mild deafness between 21 and 40 dB HL, moderate deafness between 41 and 70 dB HL, severe deafness between 71 and 90 dB HL and profound deafness above 90 dB HL).

Preoperative status was defined as “preop 1” before primary stapes surgery, “preop 2” before revision surgery, and “postop 2” after revision surgery.

The primary endpoints were the postoperative auditory results and improvement in postoperative vestibular symptomatology, evaluated subjectively by the patient and through the length of postoperative hospital stay.

Mann-Whitney tests were used for statistical analysis because the data distribution was not normal. Statistical significance was defined as a threshold of P < .05. The STROBE guidelines were used for reporting. ⁹

Results

Nineteen patients with an average age of 50 ± 9.3 (31-61) underwent revision surgery for confirmed reparative granuloma between January 2012 and September 2020.

The mean duration of hypoacusis was 8 ± 7 years (0.5-20) before the first intervention. Three patients had already undergone contralateral surgery with good results. Five patients complained of tinnitus before the primary surgery. Preoperative audiometric results showed conductive or mixed hearing loss, with a mean PTA of 50 dB ± 11.4 (30-71.25 dB) in AC and 25 dB ± 9.9 (10-48.75 dB) in BC air-bone gap of 25 dB ± 9.7 (8.8-43.8 dB) on average (Table 1). Four patients presented with mild hearing loss, 14 presented with moderate hearing loss, and 1 severe hearing loss before primary surgery.

Table 1.

Audiometry Results.

No.	Preoperative 1			Preoperative 2					Postoperative 2
	BC PTA	AC PTA	ABG	First prosthesis type/length—diameter (mm)		BC PTA	AC PTA	Revision comments	BC PTA	Preop 2-postop 2 BC PTA difference	AC PTA	ABG	SRT	Follow-up duration	New prosthesis type/length—diameter (mm)
1	39	60	21	Teflon	4.5-0.6	>90	110	C/EL	>90	0	110	20	>120	14M	NA	NA
2	19	63	44	Teflon	4.5-0.6	>90	110	C/IL	>90	0	110	20	>120	8M	Teflon	4.5-0.6
3	26	49	23	Teflon	4.5-0.6	49	58	EL	68	19	68	0	95	44M	Titanium	4.5-0.5
4	25	60	35	Teflon	4.5-0.6	56	88	C/IL	44	−12	74	30	90	5M	Teflon	4.25-0.6
5	29	41	13	Teflon	4.5-0.6	>90	110	DL/EL	31	−59	31	0	25	64M	Teflon	4.0-0.6
6	18	36	19	Teflon	4.5-0.6	79	110	C/LP/EL	>90	11	110	20	>120	5M	Teflon	4.25-0.6
7	26	38	11	Teflon	4.5-0.6	61	61	LP/EL	45	−16	49	4	55	56M	Teflon	4.25-0.6
8	13	43	30	Teflon	4.5-0.6	45	45	LP/EL	61	16	69	8	85	24M	Teflon	4.5-0.6
9	19	63	44	Teflon	4.5-0.6	63	91	IL	>90	27	110	20	>120	4M	Titanium	4.5-0.5
10	49	58	9	Teflon	4.5-0.6	54	55	IL	38	−16	38	0	40	51M	Titanium	4.5-0.5
11	30	53	23	Teflon	4.5-0.6	>90	110	DL/IL	41	−49	48	6	60	4M	Teflon	4.5-0.6
12	26	59	33	Teflon	4.5-0.6	56	80	IL	>90	34	110	20	>120	17M	Titanium	4.5-0.5
13	16	41	25	Titanium	4.5-0.5	43	66	IL	43	0	53	10	50	2M	Teflon	4.5-0.6
14	20	46	26	Teflon	4.5-0.6	59	90	EL	40	−19	41	1	55	12M	Titanium	4.5-0.5
15	30	55	25	Teflon	4.5-0.6	70	110	IL	50	−20	57	7	65	41M	Titanium	4.5-0.5
16	15	30	15	Teflon	4.5-0.6	50	65	EL	15	−35	15	0	35	3M	Titanium	4.5-0.5
17	10	35	25	Teflon	4.5-0.6	15	43	IL	>90	75	110	20	>120	5M	Titanium	4.5-0.5
18	40	71	31	Teflon	4.5-0.6	65	75	C/IL	70	5	103	33	>120	3M	Titanium	4.5-0.5
19	23	50	27	Teflon	4.5-0.6	70	103	IL	70	0	110	40	>120	1M	Titanium	4.5-0.5

Abbreviations: ABG, air-bone gap; AC, air conduction; BC, bone conduction; C, Conversion from stapedotomy to stapedectomy; DL, dry Labyrinth; EL, extralabyrinthine reparative granuloma; IL, intralabyrinthine reparative granuloma; LP, long piston; NA, Not Applicable; PTA, pure-tone average; SRT, speech recognition threshold.

Bold text indicates serviceable hearing.

Ten senior surgeons from 5 different hospitals performed the primary procedure under general anesthesia.

This consisted of a calibrated stapedotomy in 12 cases, 6 with and 6 without laser, and a partial posterior stapedectomy with graft placement in 7 cases (6 with temporal fascia and 1 with a vein interposition). The prosthesis was made of Teflon in 18 patients and titanium in 1 patient.

During these procedures, powdered gloves were not used, no absorbable material was introduced into the middle ear (gelatin sponge), and no incidents of perilymph aspiration were reported. However, in 2 cases, unusual peri-operative bleeding was reported despite reverse Trendelenburg position and satisfactory blood pressure control.

Early discharge from postoperative care was authorized for only 10 patients; the other 9 suffered from severe postoperative vertigo that prevented them from returning home.

Symptoms appeared on average 3.6 days ± 3.9 (0-14 days) after surgery, requiring readmission for those who had returned home. Symptom onset was spontaneous in 17 patients. One patient reported that blowing his nose strongly had triggered the symptoms, and 1 patient reported that heavy lifting had been the triggering factor.

Sixteen patients presented with hearing loss and vestibular symptoms, 2 patients had vestibular symptoms without hearing loss, and 1 patient had hearing loss without vestibular symptoms.

For vertigo assessment, a clinical examination and video-nystagmoscopy were performed for each patient. Contralateral nystagmus was observed in 15 patients and homolateral nystagmus in three. No nystagmus was observed in one of the patients, who did not present with vestibular symptoms.

The mean BC hearing threshold was 67.8 dB ± 22.8 (42.5-110 dB) before revision surgery (“preop 2”). Otoscopy, with reservations about its interpretation in the early postoperative period, revealed a purplish-red appearance centered on the posterior superior quadrant of the eardrum.

All patients received first-line antibiotics and corticosteroids treatment prior to revision surgery, for a mean duration of 10.7 days ± 7.4 (3-30 days).

The CT scan performed systematically showed an opacity suggestive of a granuloma in every case (Figures 2 and 3). In 3 cases, excessive piston penetration >1.5 mm was observed, and in 3 patients, persistent pneumolabyrinth (at >5 days postoperatively) was observed.

Figure 2.

Temporal bone CT scans at day 4 of a left platinectomy. (A) Opacity encompassing the prosthesis, filling the oval fossa (arrow), (B) extending into the mesotympanum, (C) encompassing the incus. (D) Presence of an image of aerial density in the vestibule indicating a pneumolabyrinth (dotted arrow).

Figure 3.

CT scans at day 5 of left platinectomy, oblique axial (A, B), coronal (C): opacity encompassing the prosthesis and filling the oval window, extending inferiorly and posteroinferiorly.

Contrast-enhanced MRI (without “hydrops” protocol) was performed in 5 cases in the acute period. Suspicion of RG was described in 1 case (Figures 3 and 4), while other reports described non-specific inflammatory tissue and inflammation, or hematoma. In the absence of improvement, and despite well-managed medical treatment, all patients underwent revision surgery.

Figure 4.

Same patient as Figure 2. CT scans (A) and MRI, T1 gradient echo without (B) and with (C) contrast injection, and T2 spin echo (D). Intravestibular extension of granuloma is visualized (dotted arrow).

All revision surgeries were performed in our referral center by 4 senior surgeons, under general anesthesia with an average delay of 11.6 ± 14.3 days (2-52) after the initial surgery.

Peroperatively, the granuloma invaded the labyrinthine cavity in 11 patients, and remained extralabyrinthine in 8 patients (Figure 5). Samples were systematically collected for histological confirmation.

Figure 5.

Preoperative view of revision surgery of left ear. (1) Reparative granuloma. (2) Incus long-process. (3) Teflon stapes prosthesis.

Surgery consistently included removal of the prosthesis, resection of the granuloma, and replacement with a new piston (8 in Teflon and 10 in titanium) except in 1 patient in whom replacement of the prosthesis was not performed (details in Table 1).

Of the 12 stapedotomy procedures, 5 required conversions to stapedectomy. Of the 7 patients who underwent primary stapedectomy, all grafts were removed and replaced with another corticosteroid-soaked temporal fascia graft. On each revision, corticosteroid instillations were introduced into the middle ear to reduce inflammation. Nine patients experienced immediate relief of their vertiginous symptoms, and a further 9 showed significant improvement in early postop with an average postoperative discharge time of 4.8 ± 1.9 days (2-10). One patient complained of persistent instability and required vestibular rehabilitation therapy to resolve the symptoms. No patient presented a recurrence of RG after revision surgery.

The hearing results are presented in Table 1, with the last available hearing assessment performed on average at 19 months ± 21 (1-64M) postop. The median BC PTA after revision surgery was 55 dB (IQR 41-90) versus 60 dB (IQR 50-79; P = .58) before revision surgery (preoperative 2). The median AC PTA after revision surgery was 68.5 dB (IQR 48-110). Eleven patients had measurable SRT on vocal audiometry at the follow-up evaluation.

Extralabyrinthine and intralabyrinthine granulomas presented with comparable BC thresholds before revision (median 60 dB; IQR = 49.5-54.5 for extralabyrinthine RGs vs 63 dB; IQR = 54.5-70; P = .93). After revision surgery, the median BC was 53 dB (IQR 35.5-79.0) for extralabyrinthine RG versus 70 dB (IQR 43.3-90; P = .36) for RGs with intralabyrinthine portion.

Discussion

Stapes surgery is associated with an inherent labyrinthine risk, and the occurrence of RG is one of the suspected etiologies. To our knowledge, this study represents the largest RG series in the literature. Surgical revision provides effective management of vertigo, with fast improvement in the majority of patients. There was no significant improvement or degradation in hearing after revision. However, it is noteworthy that 11 of 19 patients retained serviceable hearing following surgical revision.

Although described in 1962 by Haris and Weis, RG remains a misunderstood entity among otologists worldwide. Its incidence is estimated to be between 0.07% and 1%, which corresponds to an exaggeration of the normal healing process.^8,11,12 Its pathophysiological process, which is probably multifactorial, is poorly understood. One of these hypotheses is based on the cerebrospinal fluid (CSF) leakage model, as contact of the CSF with the middle ear mucosa causes a genuine granuloma. Based on this model, a greater presence of perilymphatic fluid (identical in composition to CSF) within the middle ear in cases of stapedectomy or perilymphatic fistula could also be associated with a greater number of granulomas. ⁶

In our study, symptoms may have arisen very early after the intervention, in line with the study by Watts et al, but contrary to some descriptions.^{11 -15}

Otoscopy may reveal a purplish mass in the retro-tympanum, which may be difficult to analyze and differentiate from a simple postoperative hematoma; however, in the particular case of granuloma, it can be centered on the postero-superior part of the eardrum.^{12 -16}

The presence of pneumolabyrinth should be interpreted with caution. Only a large or delayed pneumolabyrinth associated with clinical signs is correlated with a high risk of perilymph fistula, which should therefore lead to revision surgery. ¹⁷ Regarding the excessive intravestibular penetration of a piston, usually defined as a shaft penetrating more than 1 mm into the vestibule, Vandevoort et al pointed out that the length of a titanium prosthesis may be overestimated by up to 0.5 mm. ¹⁷ Their study found no correlation between piston penetration and sensorineural symptomatology, a result confirmed by Fang et al.^17,18

The choice to change the piston material was left to the surgeon’s discretion. Replacing the prosthesis with a different material in cases of granuloma progressively became routine practice. This approach stemmed from the uncertain pathophysiology of reparative granuloma and the possibility of a material-related reaction. However, no recurrences were observed when a Teflon prosthesis was replaced with another Teflon prosthesis. At a certain point during the inclusion period, surgeons tended to reduce the length of the new prosthesis, in case excessive penetration had been involved, even if not perceptible on CT or intraoperatively. However, subjectively, this did not appear to have any impact, and the center’s protocol was then to retain a standard prosthesis size if this appeared to be suitable from the surgeon’s point of view.

In our series, a systematic CT scan seemed to be more sensitive for RG detection than T1 MRI gadolinium, although no gold-standard imaging exists. However, MRI could be helpful in assessing intravestibular extension of the RG. ¹⁹

Despite the lack of sensitivity in our study, MRI can nevertheless suggest the presence of RG. (intermediate signal on T1, hypointense on T2 and contrast-enhanced).^{19 -21} It is possible that improvements in the definition and democratization of post-contrast 4-hour-delayed MRI protocols may give rise to more exploitable results. We believe that MRI should not delay surgical management.

Our study has several limitations. First, it was a retrospective study analyzing a limited number of patients in a single center. The absence of a comparator is a significant limit to interpretation, particularly subjective symptoms, prone to spontaneous relief, such as vestibular symptoms. We had no possibility of a reliable control group, since no patient with a strong suspicion of granuloma refused the operation and therefore only underwent medical treatment. This problem is all the more pronounced when the diagnosis is histological rather than radiological. The evaluation of vertigo symptoms was not quantified using a questionnaire, and the patient’s subjective assessment and length of stay were the only data available. Advances in imaging technology may make it possible to reliably diagnose RG, enabling comparisons between different therapeutic approaches. Given the rarity of this complication, a large-scale, controlled multicenter study would is necessary to provide a larger number of patients. The lack of statistical power may have impeded the comparison of extra and intralabyrinthine granulomas.

Conclusion

Reparative granuloma is an exceptional but disastrous complication of stapes surgery. This study suggests that, at the very least, surgical management can lead to fast and effective resolution of vestibular symptoms.