Microbiome and Otic Quinolone Levels Following Tympanoplasty Assessed by Gelatin Sponge Analysis.

OBJECTIVE: To determine if absorbable gelatin sponge (AGS) can be used to assess the posttympanoplasty microbiome and otic antibiotic exposure. STUDY DESIGN: Prospective. SETTING: Tertiary hospital. METHODS: Patients undergoing tympanoplasty were prospectively enrolled. Intraoperatively, AGS was applied to the medial ear canal/tympanic membrane (TM) for 1 minute after canal incision, then saved for analysis. Ear canals were packed with AGS at the end of surgery. Otic ofloxacin was administered until the first postoperative visit, when AGS was collected. Microbial presence was assessed by culture. Ofloxacin levels were assessed by liquid-chromatography mass-spectrometry. RESULTS: Fifty-three patients were included. AGS was collected in 92.9% of patients seen within 21 days compared to 70.8% of those seen at 22 to 35 days. At surgery, AGS yielded bacteria and fungi in 81% and 11%, respectively, including Staphylococcus species (55%) and Pseudomonas species (25%). Postoperatively, AGS yielded bacteria in 71% and fungi in 21% at the meatus, (staphylococci 57% and pseudomonas 25%). TM samples yielded bacteria in 69%, fungi in 6%, staphylococci in 53%, and pseudomonas in 19%. Ofloxacin concentration at the meatus was 248 µg/mL (95% confidence interval [CI]: 119-377) and at the TM was 126 µg/mL (95% CI: 58-194). Ofloxacin-resistant colonies were found in 75% of patients. CONCLUSION: Analysis of AGS is a viable technique for noninvasively studying healing metrics posttympanoplasty, including the microbiome and otic antibiotic exposure. Despite exposure to a high concentration of quinolones, the tympanoplasty wound is far from sterile, which may impact healing outcomes.

Rat Model of External Auditory Canal Stenosis.

HYPOTHESIS: Circumferential electrocautery injury of the rat external auditory canal (EAC) can induce a reproducible animal model of acquired stenosis. BACKGROUND: Acquired EAC stenosis may occur as a result of chronic inflammation or trauma to the EAC skin and is characterized by narrowing of the EAC, retention of debris, and hearing loss. Treatment is surgery but it is often complicated by restenosis. A reliable and inexpensive animal model of EAC stenosis has not been described. There have been no studies correlating the extent of EAC injury with the extent of stenosis. METHODS: Rats received a 25, 50, or 75% circumferential EAC injury with electrocautery. The extent of resulting stenosis was quantified 21 days following injury. The nature of the injury and healing response was assessed with histology. RESULTS: A 25% circumferential injury led to 4 to 34% stenosis (mean, 13%), 50% injury resulted in 43 to 100% stenosis (mean, 73%), and 75% injury resulted in 94 to 100% stenosis (mean=99%, p < 0.0001). The 50% circumferential injury produced 30 to 75% stenosis in five of eight ears, the remainder had >75% stenosis. Wounded ears showed evidence of intact cartilage and epithelium, with increased thickness of the subepithelial layer and localized fibrosis. CONCLUSIONS: Electrocautery injury in the ventral aspect of the rat EAC resulted in reproducible EAC stenosis. This rat model may be useful in studying therapy to prevent acquired EAC stenosis due to acute injury. The correlation of the extent of injury (circumference) with resulting stenosis may inform clinical management of EAC injuries.

The Effects of Steroids on Survival of Mouse and Human Tympanic Membrane Fibroblasts.

OBJECTIVE: Tympanic membrane (TM) fibroblast cytotoxicity of quinolone ear drops is enhanced by dexamethasone and fluocinolone. Hydrocortisone has not been evaluated. We aimed to assess the effects of these 3 steroids on mouse and human TM fibroblast survival. STUDY DESIGN: In vitro. SETTING: Academic laboratory. SUBJECTS AND METHODS: Mouse and human TM fibroblasts were exposed to hydrocortisone, dexamethasone, or fluocinolone at concentrations in commercial ear drops (1%, 0.1%, or 0.025%, respectively) and at steroid potency equivalents (1%, 0.033%, or 0.0033%, respectively), or dilute ethanol (control), twice within 24 hours or 4 times within 48 hours for 2 hours each time. Cells were observed with phase-contrast microscopy until the cytotoxicity assay was performed. RESULTS: Mouse and human TM fibroblasts treated with any of the steroids had lower survival after 24 and 48 hours compared to control (all P < .0001). After 24 hours, viability of mouse fibroblasts treated with the steroids was not different (P > .05), while treatment with hydrocortisone decreased human TM fibroblast viability (P < .0001). After 48 hours, at concentrations found in ear drops and at equivalent steroid potency, dexamethasone and fluocinolone had similar survival in mouse and human fibroblasts (all P > .05), but hydrocortisone had lower survival in both mouse (P = .02 and P < .0001) and human (P < .0001) fibroblasts. Phase-contrast images mirrored the cytotoxicity findings. CONCLUSION: Steroids found in commercial ear drops reduce survival of mouse and human TM fibroblasts. Hydrocortisone appears to be more cytotoxic than the more potent steroids, dexamethasone and fluocinolone. These findings should be considered when assessing clinical outcomes of ototopical preparations.

Cytotoxicity of Ear Drop Excipients in Human and Mouse Tympanic Membrane Fibroblasts.

OBJECTIVE: Commercial ear drops contain ingredients reported to be inactive. We sought to evaluate such excipients for possible cytotoxicity on human and mouse tympanic membrane (TM) fibroblasts. STUDY DESIGN: Prospective, in vitro. SETTING: Tertiary academic center. SUBJECTS AND METHODS: Mouse and human TM fibroblasts were treated with 1:10 dilutions of benzalkonium chloride (BKC) 0.0025%, 0.006%, or 0.01%; benzyl alcohol 0.9%; polysorbate 80 (PSB) 2.5%; glycerin 2.4%; povidone 0.2%; or water (control), twice within 24 hours or 4 times within 48 hours, for 2 hours each time. Cells were placed back in growth media after the treatments. Cells were observed with phase-contrast microscopy until the cytotoxicity assay was performed. RESULTS: Mouse fibroblasts had lower survival in only the PSB-treated cells compared to the control (P < .0001) after 24 hours. After 48 hours, PSB killed nearly all mouse fibroblasts (P < .0001). BKC decreased fibroblast survival in a dose-dependent manner (P < .001). In human TM fibroblasts, all excipients except povidone and benzyl alcohol after 24 hours and povidone after 48 hours reduced cell survival compared to control (P = .012 to P < .0001). The cytotoxicity of BKC in human TM fibroblasts was also dose dependent (<.0001). PSB was less cytotoxic to human fibroblasts. Phase-contrast images mirrored the cytotoxicity findings. CONCLUSION: Polysorbate 80 and benzalkonium chloride, at concentrations found in commercial ear drops, may be cytotoxic to human and mouse TM fibroblasts. "Inactive" ingredients may need to be considered when evaluating clinical outcomes with commercial ear drops.

Microbiome Analysis of Cholesteatoma by Gene Sequencing.

OBJECTIVE: To compare the microbial flora of cholesteatoma and normal middle ears using gene-based sequencing analysis. STUDY DESIGN: Controlled ex vivo human study. SETTING: Academic, tertiary medical center. SUBJECTS AND METHODS: Brush, swab, and tissue samples were each taken from cholesteatoma matrix and uninvolved tissue in patients with previously untreated, acquired cholesteatoma (n = 19) or middle ear mucosa from patients undergoing cochlear implantation with no history of cholesteatoma or previous middle ear surgery (control; n = 12). DNA was isolated from specimens then 16S rRNA gene sequencing was performed. RESULTS: There was no difference in microbial yield between the sampling methods. Cholesteatoma specimens had lower relative abundance of 14 bacterial species compared with controls including Acidovorax sp., Bacillus sp., Masillia sp., Moraxella osloensis, Phenylobacterium conjunctum, Sphingomonas sp., and Staphylococcus epidermidis (all p < 0.05). Alternaria sp. were present on nearly all the specimens. Alternaria sp. and Cladosporium herbarum (both p ≤ 0.05) were lower in the cholesteatoma compared with control group. There was no difference in the relative abundance of any bacteria or fungi between the cholesteatoma matrix and uninvolved middle ear mucosa. CONCLUSIONS: Microbiome of cholesteatoma matrix is largely similar to adjacent mucosa. This differs from healthy ears. Further study is needed to understand if middle ear microbiome may impact cholesteatoma pathogenesis or treatment.

Effects of Common Ear Drops on Tympanic Membrane Healing in Rats.

Objective Ciprofloxacin + dexamethasone ear drops have been associated with higher rates of tympanic membrane perforations than ofloxacin. This was thought to be an effect of the steroid; however, ciprofloxacin (sans steroid) has been found to be more toxic to tympanic membrane fibroblasts than ofloxacin in vitro. This study aimed to compare the effect of these agents on tympanic membrane healing in vivo. Study Design Controlled animal study. Setting Academic research laboratory. Methods Perforations were created in 54 rats with a carbon dioxide laser. Rats were randomized to 6 groups (9/group), with 1 ear receiving ofloxacin, ciprofloxacin, dexamethasone, ofloxacin + dexamethasone, ciprofloxacin + dexamethasone, or neomycin, and the contralateral ear receiving saline twice daily for 10 days. Healing was assessed over 40 days. Results Ofloxacin did not delay healing relative to saline. All other treatments delayed healing relative to ofloxacin at day 10 ( P < .0001). Dexamethasone and ofloxacin + dexamethasone delayed healing up to day 14 ( P < .0001). Neomycin and ciprofloxacin + dexamethasone further delayed healing up to day 28 ( P = .009) and day 35 ( P = .043), respectively. All eardrums healed by day 10 with ofloxacin, day 20 with ciprofloxacin, day 28 for dexamethasone, and day 35 for ofloxacin + dexamethasone. At day 40, 2 of 9 ciprofloxacin + dexamethasone-treated eardrums had not healed. All saline-treated eardrums in the ofloxacin, ciprofloxacin, dexamethasone, and ofloxacin + dexamethasone groups were healed by day 14, but this was delayed to day 20 in the ciprofloxacin + dexamethasone group ( P = .007). Conclusions Ototopical quinolones delay rat tympanic membrane healing in a drug-specific manner, with ciprofloxacin having a greater impact than ofloxacin. This effect is potentiated by steroids.

Cytotoxicity of Ciprofloxacin and Steroids in Mouse Tympanic Membrane Fibroblasts.

Objective Ciprofloxacin, commonly given as eardrops, has been shown to adversely affect tympanic membrane fibroblasts. Dexamethasone potentiates this effect. A newly available eardrop contains ciprofloxacin and fluocinolone, a more potent steroid. We evaluated the cytotoxic effects of this preparation on mouse tympanic membrane fibroblasts. Study Design Prospective, in vitro. Setting Academic laboratory. Subjects and Methods In experiment 1, fibroblasts were exposed to 1:10 dilutions of commercially available 0.3% ofloxacin, 0.3% ciprofloxacin, 0.3% ciprofloxacin + 0.1% dexamethasone, 0.3% ciprofloxacin + 0.025% fluocinolone, or dilute hydrochloric acid (control), twice within 24 hours. In experiment 2, cells were also treated with the dilutions of the pure form of dexamethasone 0.1% or fluocinolone 0.025%, alone and in combination with ofloxacin or ciprofloxacin. Cells were exposed to the solutions for 2 hours each time and were placed back in growth media after the treatments. Cells were observed with phase-contrast microscope until the cytotoxicity assay was performed. Results Survival of fibroblasts treated with ofloxacin was not different from the control. Fibroblasts treated with ciprofloxacin, ciprofloxacin + dexamethasone, or ciprofloxacin + fluocinolone had much lower survival (all P < .0001). Cells treated with ciprofloxacin + fluocinolone had lower survival than ciprofloxacin ( P < .0001) and ciprofloxacin + dexamethasone ( P = .0001). Steroids alone also decreased fibroblast survival compared to control ( P < .0001). The combination of dexamethasone or fluocinolone with ciprofloxacin, but not ofloxacin, further decreased fibroblast survival ( P < .0001). Phase-contrast images mirrored the cytotoxicity findings. Conclusion Tympanic membrane fibroblast cytotoxicity of ciprofloxacin is potentiated by corticosteroids. This effect may be deleterious when treating a healing perforation but beneficial when treating granulation tissue on the tympanic membrane.

Mitochondria-Targeted Antioxidant Mitoquinone Reduces Cisplatin-Induced Ototoxicity in Guinea Pigs.

Objective To determine if mitoquinone (MitoQ) attenuates cisplatin-induced hearing loss in guinea pigs. Study Design Prospective and controlled animal study. Setting Academic, tertiary medical center. Subjects and Methods Guinea pigs were injected subcutaneously with either 5 mg/kg MitoQ (n = 9) or normal saline (control, n = 9) for 7 days and 1 hour before receiving a single dose of 10 mg/kg cisplatin. Auditory brainstem response thresholds were measured before MitoQ or saline administration and 3 to 4 days after cisplatin administration. Results Auditory brainstem response threshold shifts after cisplatin treatment were smaller by 28 to 47 dB in guinea pigs injected with MitoQ compared with those in the control group at all tested frequencies (4, 8, 16, and 24 kHz, P = .0002 to .04). Scanning electron microscopy of cochlear hair cells showed less outer hair cell loss and damage in the MitoQ group. Conclusion MitoQ reduced cisplatin-induced hearing loss in guinea pigs. MitoQ appears worthy of further investigation as a means of preventing cisplatin ototoxicity in humans.

Microflora of Retained Intracochlear Electrodes from Infected Cochlear Implants.

Objectives Cochlear implant infections may be refractory to medical management and require device removal with subsequent reimplantation. During device removal, the intracochlear electrode array is commonly left in place to prevent obliteration of the cochlear lumen. If the electrode is colonized with pathogens, this risks contaminating the replacement implant. In this study, we compare the microorganisms detected on infected cochlear implants against those on the retained electrode using culture and microbial gene-sequencing techniques. Study Design Prospective single-cohort study. Setting Tertiary medical center. Subjects and Methods Six patients with refractory cochlear implant infections had the receiver-stimulator and extracochlear electrode removed to facilitate treatment of the infection. The intracochlear electrode was removed at (delayed) reimplantation. Implant specimens were analyzed by microbial culture and 16S DNA gene sequencing. Results Staphylococcus aureus was the organism most commonly identified. None of the 6 patients' intracochlear electrodes yielded microbes by culture. Two intracochlear electrodes revealed bacterial species, and 1 revealed fungal species by gene sequencing. There was no correlation between the microbes on the infected extracochlear implants and the retained intracochlear electrodes. All subjects underwent reimplantation after resolution of their infections. One of 6 subjects developed a second infection after reimplantation, with S aureus in the primary and secondary infections. Conclusions The intracochlear electrodes of infected cochlear implants carry a low microbial burden. Preserving intracochlear electrodes upon removal of infected cochlear implants appears to carry a low risk of contaminating a replacement cochlear implant.

Collagen and α-Tubulin of Mouse Tympanic Membrane Fibroblasts Treated with Quinolones and Aminoglycosides.

Objective To assess collagen and α-tubulin levels of mouse tympanic membrane fibroblasts treated with quinolone and aminoglycoside antibiotics at concentrations found in eardrops. Study Design Prospective controlled cell culture study. Setting Academic tertiary medical center. Subjects Mouse tympanic membrane fibroblasts. Methods In experiment 1, fibroblasts were treated with the following for 24 or 48 hours: phosphate-buffered saline (negative control), dilute hydrochloric acid (positive control), 0.5% gatifloxacin, or commercially available 0.3% ciprofloxacin, 0.3% ciprofloxacin + 0.1% dexamethasone, 0.3% ofloxacin, 0.5% moxifloxacin, 0.3% gentamicin, or 3.5 mg/mL of neomycin + polymyxin B sulfate + hydrocortisone. In experiment 2, cells were treated with the pure form of gatifloxacin, gentamicin, ofloxacin, or ciprofloxacin. Cells were observed with phase-contrast microscope until harvested. Proteins were extracted for Western blotting with antibodies against collagen α1 type I (collagen 1A1) and α-tubulin, and for densitometry to quantify levels. Results Collagen and tubulin levels in fibroblasts treated with ofloxacin, moxifloxacin, gatifloxacin, or gentamicin for 24 hours were not different from the saline control. Fibroblasts treated with neomycin + polymyxin B + hydrocortisone, ciprofloxacin + dexamethasone, or ciprofloxacin for 24 hours had lower collagen 1A1 and α-tubulin levels (all P < .001) than the negative control. After 48 hours, fibroblasts treated with neomycin + polymyxin B sulfate + hydrocortisone, ciprofloxacin + dexamethasone, ciprofloxacin, or moxifloxacin had lower collagen 1A1 ( P ≤ .007) and α-tubulin ( P < .0001; except ciprofloxacin, P = .033) as compared with control. In experiment 2, only cells treated with ciprofloxacin had lower collagen 1A1 and α-tubulin levels and cell viability (all P < .0001) than control. Cytotoxicity assay and phase-contrast images mirrored the protein findings. Conclusion The adverse impact of topical antibiotic exposure on tympanic membrane collagen and tubulin protein levels is drug specific. This may be important for selection of ototopical therapy.

The Adenoid Microbiome in Recurrent Acute Otitis Media and Obstructive Sleep Apnea.

OBJECTIVE: To compare the microbial flora of adenoids from patients with recurrent acute otitis media (AOM) and patients with obstructive sleep apnea (OSA). MATERIALS AND METHODS: This study was prospective and controlled. Adenoids were obtained from children undergoing adenoidectomy for recurrent AOM (n=7) or OSA (n=13). Specimens were processed for total deoxyribonucleic acid (DNA) isolation. 16s DNA 454-pyrosequencing was performed on AOM (n=5) and OSA (n=5) specimens. All specimens were analyzed by real-time polymerase chain reaction for the quantification of the oral commensal bacteria, Streptococcus salivarius. RESULTS: All adenoid specimens had evidence of microbes. Haemophilus influenzae, Moraxella catarrhalis, Streptococcus pneumoniae, Pseudomonas aeruginosa, and Staphylococcus aureus were among the dominant species in all samples. S. salivarius, Prevotella sp. and Terrahaemophilus aromaticivorans were more common on adenoids from OSA patients (p<0.05). Bradyrhizobium sp. was more common on adenoids from patients with recurrent AOM (p<0.05). The microbial profiles associated with recurrent AOM were different from, but overlapped with OSA. S. salivarius quantified by real-time PCR was not different between the two groups. CONCLUSION: Microbes are present on all adenoid specimens, though the microbial profile differs between recurrent AOM and OSA. The clinical significance of these differences remains to be determined.

Ciprofloxacin Decreases Collagen in Mouse Tympanic Membrane Fibroblasts.

OBJECTIVES: To determine how collagen production by tympanic membrane fibroblasts is affected by ciprofloxacin at levels found in eardrops. STUDY DESIGN: Prospective, controlled, and blinded cell culture study. SETTING: Academic tertiary medical center. SUBJECTS: Cell culture of mouse fibroblasts. METHODS: A primary fibroblast culture was established from mouse tympanic membranes. Fibroblasts were cultured until they were 75% confluent, then treated with dilute hydrochloric acid (control) or ciprofloxacin (0.01% or 0.3%) for 24 or 72 hours for Western blotting and for 24 or 48 hours for cytotoxicity assay. Cells were observed with phase-contrast microscope. Western blotting was performed for collagen type 1 α1 (collagen 1A1) and α-tubulin. RESULTS: Fibroblasts treated with 0.01% and 0.3% ciprofloxacin for 24 hours had lower levels of collagen 1A1 (P = .0005 and P < .0001, respectively) and α-tubulin (both P < .0001) than control fibroblasts. Collagen 1A1 and α-tubulin levels were lower in fibroblasts treated with 0.3% than with 0.01% ciprofloxacin (P = .02 and P = .014). After 72 hours, 0.3% ciprofloxacin completely eliminated collagen 1A1 and α-tubulin (P < .001). Cells treated with 0.01% ciprofloxacin for 72 hours also had lower collagen 1A1 (P < .0001) and α-tubulin (P = .005) as compared with the control. Seventy-two-hour incubation in 0.01% or 0.3% ciprofloxacin resulted in lower levels of collagen 1A1 (P = .009 and P < .0001, respectively) and α-tubulin (P = .007 and P < .0001, respectively) than 24-hour incubation. Cytotoxicity assay and phase-contrast microscopy mirrored these findings. CONCLUSIONS: Treatment of tympanic membrane fibroblasts with 0.3% ciprofloxacin, as found in eardrops, reduces fibroblast viability and collagen and α-tubulin protein levels. These findings could explain tympanic membrane healing problems associated with quinolone eardrops.