Abstract
Objective:
To evaluate the efficacy of tympanomastoidectomy versus parenteral antibiotic therapy for otorrhea as a result of chronic suppurative otitis media (CSOM) without cholesteatoma in the pediatric population.
Methods:
A retrospective review of 221 patients treated for otorrhea at a tertiary academic pediatric hospital was performed to evaluate the impact of tympanomastoidectomy versus parenteral antibiotic therapy on resolution of otorrhea. Inclusion criteria were age 0-18 years, prior treatment with otic and/or oral antibiotic, prior history of tympanostomy tube placement for recurrent otitis media, history of otorrhea, treatment with tympanomastoidectomy or parenteral antibiotic therapy, and follow-up of at least 1 month after intervention. Time to resolution was compared between the two modalities adjusting for age, bilateral ear disease status, and comorbidities using a Cox proportional hazard model.
Results:
Eighty-three ears from 58 children met the inclusion criteria. Ears that initially underwent tympanomastoidectomy had a significantly shorter time to resolution of symptoms (median time to resolution 9 months (95% confidence interval CI: 6.2~14.8) vs. 48.5 months (95% lower CI 9.4, p=0.006). On multivariate analysis, however, only bilateral ear disease status was independently associated with time to resolution of symptoms (hazard ratio 0.4, 95% CI 0.2~0.9, p=0.03). There was no statistically significant difference in the rate of treatment-related complications when comparing tympanomastoidectomy to parenteral antibiotic therapy (p=0.37).
Conclusion:
When adjusting for age, bilateral ear disease status, and comorbidities, there does not appear to be a significant difference in time to resolution of symptoms when comparing parenteral antibiotic therapy to tympanomastoidectomy. An informed discussion regarding risks and benefits of each approach should be employed when deciding on the next step in management for patients with CSOM who have failed more conservative therapies.
Keywords: otorrhea, tympanomastoidectomy, chronic otitis media, pediatric otorrhea, pediatric, chronic ear disease, parenteral antibiotics
1. Introduction
Otitis media (OM) is a leading cause of healthcare visits, antibiotic treatment, and surgery with an estimated 709 million cases worldwide annually.[1, 2] Around 5% of this group will develop chronic suppurative otitis media (CSOM), contributing significantly to the cost and burden of OM.[1] CSOM is defined as chronic inflammation of the middle ear and mastoid mucosa in which the tympanic membrane is not intact, due to either perforation or tympanostomy tube, with resultant otorrhea.[3] It is a leading cause of preventable hearing loss in children and may lead to extra- and intracranial complications, including but not limited to facial paralysis and meningitis.[4] Both otic and oral antibiotics are effective first line approaches to management of CSOM, but in some cases fail to result in resolution of symptoms.[5] Currently, there is no clear consensus regarding the proper management of non-cholesteatomatous CSOM refractory to otic and/or oral antibiotic treatment.
Operative interventions such as tympanomastoidectomy or aggressive medical management with parenteral antibiotic administration are competing schools of thought for treatment of refractory CSOM.[6] Each treatment approach presents its own unique potential for adverse and successful outcomes. Tympanomastoidectomy surgeries risk surgical site infection, facial nerve injury, iatrogenic labyrinthine fistula formation, dural or vascular injury, and the added dangers of general anesthesia.[7] Previous studies have reported a success rate around 74% with rare complications in children treated with tympanomastoidectomy for chronic otorrhea.[8] Compared to operative intervention, parenteral antibiotic therapy has been shown to successfully resolve CSOM in 75-89% of patients with continued long-term resolution in 78% of patients[6, 9]. Parenteral antibiotic treatment generally requires insertion of a central line, with risks that include line infection, adverse medication reactions, and development of antibiotic resistance. Considering the comparable reported outcomes of aggressive surgical and medical management of CSOM, we feel that a direct comparison of these two approaches is a necessary step to develop treatment guidelines. The objective of this study is to compare the efficacy of tympanomastoidectomy versus parenteral antibiotic therapy for treatment of pediatric otorrhea as a result of CSOM.
2. Methods
2.1. Study Cohort
After IRB approval, a retrospective chart review of the clinical and demographic data of all pediatric patients treated for otorrhea by four pediatric otolaryngologists at a tertiary academic pediatric hospital from 2010 to 2020 was performed. Two hundred twenty-one children treated for otorrhea were identified during this timeframe. Inclusion criteria were patients aged 0-18, prior treatment with otic and/or oral antibiotic, prior history of tympanostomy tube placement for recurrent otitis media, history of otorrhea with at least 1 physical exam noting otorrhea prior to treatment, treatment with tympanomastoidectomy or parenteral antibiotic therapy, and follow-up of at least 1 month after intervention with documented physical exam by an attending otolaryngologist to note whether otorrhea had resolved. Exclusion criteria included patients with otorrhea who did not undergo either tympanomastoidectomy or parenteral antibiotic therapy as part of their treatment course, lack of follow-up or documented otologic exam noting otorrhea status, or documented cholesteatoma as initial etiology of chronic otorrhea. The analysis was conducted at the ear level rather than patient level, as many patients contributed two ears. Resolution of otorrhea was defined as a clinically documented resolution of otorrhea on at least one otologic exam at least one month after intervention. The surgical treatment group included patients who underwent tympanomastoidectomy as their treatment modality. The medical treatment group included patients who received parenteral antibiotic therapy, defined as intravenous (IV) antibiotic therapy. Since it is common for ears to be treated multiple times and sometimes by both tympanomastoidectomy or parenteral antibiotics, we studied the initial treatment only, and censored the observation at time of the second treatment if resolution had not been achieved by then. A subject was noted to have a comorbidity if there was a known comorbid diagnosis that increases the likelihood of developing otitis media and/or more significant disease (eg Trisomy 21, immunocompromise, etc) recorded in the patient’s chart. Details of the demographic and clinical data of our patient population are described in Table 1.
Table 1:
Demographic and clinical characteristics of patient cohort
| Variable | Surgical (N=54) | Medical (N=29) | P-value |
|---|---|---|---|
| Age - Mean (SD) | 9.1 (4.2) | 3.9 (3.7) | <0.001t |
| Median (IQR) | 8.4 (6.2, 11.8) | 2.3 (1.2, 6.5) | - |
| Range | (1.3, 17.6) | (0.8, 15.2) | - |
| Gender - Female | 23 (42.6%) | 17 (58.6%) | 0.16c |
| Male | 31 (57.4%) | 12 (41.4%) | - |
| Race - Caucasian | 44 (81.5%) | 29 (100%) | 0.15f |
| Asian | 5 (9.3%) | 0 (0%) | - |
| Pacific Islander | 3 (5.6%) | 0 (0%) | - |
| African American | 2 (3.7%) | 0 (0%) | - |
| Laterality – Left | 22 (40.7%) | 14 (48.3%) | 0.51c |
| Right | 32 (59.3%) | 15 (51.7%) | - |
| Bilateral disease status | 27 (50%) | 22 (75.9%) | 0.022c |
| Treating physician - 1 | 30 (55.6%) | 18 (62.1%) | 0.55f |
| 2 | 9 (16.7%) | 2 (6.9%) | - |
| 3 | 9 (16.7%) | 7 (24.1%) | - |
| 4 | 6 (11.1%) | 2 (6.9%) | - |
| Duration of otorrhea prior to first treatment - <1 month | 8 (14.8%) | 5 (17.2%) | 0.67f |
| 1-6 months | 12 (22.2%) | 4 (13.8%) | - |
| >6 months-1yr | 4 (7.4%) | 4 (13.8%) | - |
| >1 yr | 30 (55.6%) | 16 (55.2%) | - |
| Microbiology - H.influenzae | 1 (1.9%) | 3 (10.3%) | 0.52f |
| MRSA | 4 (7.4%) | 2 (6.9%) | - |
| MSSA | 3 (5.6%) | 4 (13.8%) | - |
| No culture | 29 (53.7%) | 13 (44.8%) | - |
| No growth | 3 (5.6%) | 1 (3.4%) | - |
| Other | 5 (9.3%) | 3 (10.3%) | - |
| Pseudomonas | 8 (14.8%) | 2 (6.9%) | - |
| Strep pneumo | 1 (1.9%) | 1 (3.4%) | - |
| Comorbidities | 21 (38.9%) | 10 (34.5%) | 0.69c |
| Resolution | 44 (81.5%) | 10 (34.5%) | * |
| Months to resolution - Median (95% CI) | 9.0(6.2~14.8) | 48.5 (9.4~?) | 0.006l |
| Months of follow un - Median (95% CI) | 44.6 (20.9~?) | 10.4 (9.2~50.3) | 0.006l |
Missing values: None.
Surgical=tympanomastoidectomy versus medical=parenteral antibiotic therapy. MSSA=methicillin sensitive Staphylococcus aureus; MRSA=methicillin resistant Staphylococcus aureus.
Resolution status is analyzed jointly with months to resolution using a log rank test.
T-test,
Chi-squared test,
Fisher’s exact test,
Exact Wilcoxon rank sum test,
Log rank test.
2.2. Statistical Analysis
A descriptive analysis was performed to summarize the patterns of treatment of our study cohort. Baseline variables as well as outcome measures were summarized at ear level stratified by method of first treatment. Four observations received both tympanomastoidectomy and parenteral antibiotics at the first visit, and were classified as having tympanomastoidectomy treatment for our analysis.
Continuous variables were summarized as mean (standard deviation/SD), median (interquartile range/IQR) and range, and compared using a t test (for age) or a Wilcoxon rank sum test (for number of treatments). Categorical variables were summarized as frequency and percent, and compared using chi-squared tests or Fisher’s exact tests. Median time to resolution and 95% confidence intervals (CIs) were estimated using the Kaplan-Meier method, and compared between treatment groups using a log rank test. Ears that had not resolved after first treatment were censored at time of second treatment, or last follow up if there was no subsequent treatment. Follow-up time was also estimated using the Kaplan-Meier method as the time from initial treatment to last follow-up time with censoring at resolution or at the time of the second treatment, whichever came first. Follow-up times were compared between treatment groups using a log rank test. In some cases, we only estimated the lower bound of the 95% confidence interval “95% lower CI” due to few events.
Our analysis focused on initial treatment only because the outcomes of those who underwent multiple treatments were associated with their entire treatment history, making it difficult to assess effects of just our treatments of interest. Furthermore, there were too many possible treatment paths (with too few subjects) to examine outcomes for multiple treatment courses. For these same reasons, complications (which were not aligned with a specific treatment) were descriptively summarized but not compared.
Cox proportional hazards models were used to compare time to resolution between type of first treatment (tympanomastoidectomy vs. parenteral antibiotics) adjusting for age, bilateral status, and comorbidities. These adjustment variables were selected a priori due to their potential influence on disease severity. Because many patients had both ears included in the study, a frailty term for patient was included in the model to account for correlation within patients. Both univariable and multivariable hazard ratios were reported with 95% CIs and p-values. Analyses were done using R version 4.0[10]. Significance was assessed at p=0.05, and all tests were two sided.
3. Results
Eighty-three separate ears from 58 children met the inclusion criteria. Of those, 54 ears underwent tympanomastoidectomy whereas 29 received parenteral antibiotic therapy as their initial treatment. Patients initially treated with parenteral antibiotic were significantly younger, with mean age at time of presentation 3.9 years (3.7) as compared to 9.1 years (SD 4.2) for those treated with tympanomastoidectomy, respectively (p<0.001). Among ears treated with tympanomastoidectony, 57.4% were from boys, compared to 41.4% among those treated with antibiotics. The majority of patients identified as Caucasian (81.5% in the tympanomastoidectomy group and 100% in the antibiotic group). Of the tympanomastoidectomy group, 50% had bilateral disease status compared to 75.9% in the antibiotic group (p=0.022). There was no significant difference in the modality of treatment distribution across the 4 treating physicians (p=0.55). Most patients did not receive a culture during their treatment course in both the surgical and medical groups; of those who did, the most common organism identified was Pseudomonas in the surgical group and methicillin sensitive Staphylococcus aureus in the medical group. The complete demographic and clinical characteristics of the two cohorts are summarized in Table 1.
For cases that initially underwent tympanomastoidectomy, the median time to resolution was 9.0 months (95% CI 6.2~14.8) with a median follow-up of 44.6 months (95% lower CI 20.9). For cases that initially treated with antibiotics, the median time to resolution was 48.5 months (95% lower CI 9.4) with a median follow-up of 10.4 months (95% CI 9.2~50.3). There was a statistically significant difference between the tympanomastoidectomy and antibiotic group in both time to resolution (antibiotic group was longer, p=0.006) and follow-up time (tympanomastoidectomy was longer, p=0.006).
Cox proportional hazard models were used to compare time to resolution of symptoms. Age at initial treatment, bilateral status, and comorbidities were included as covariates. Without adjusting for any other variables, type of initial treatment, age, and bilateral status were significantly associated with time to resolution. However, in the multivariable model, only bilateral status was independently associated with a longer time to resolution of symptoms (hazard ratio 0.4, 95% CI 0.2~0.9, p=0.03). See Table 2 for the univariate and multivariate Cox proportional hazard models.
Table 2.
Univariate and multivariate Cox proportional hazards models for comparing time to resolution
| Variable | HR1 (95% CI) | P1 value | HR2 (95% CI) | P2 value |
|---|---|---|---|---|
| Initial surgical treatment | 2.6 (1.1,6.6) | 0.037 | 1.2 (0.4,3.8) | 0.71 |
| Age | 1.1 (1.0,1.2) | 0.027 | 1.1 (1.0,1.2) | 0.10 |
| Bilateral | 0.3 (0.1,0.8) | 0.012 | 0.4 (0.2,0.9) | 0.030 |
| Comorbidities | 1.4 (0.5,3.6) | 0.51 | 1.6 (0.7,3.7) | 0.30 |
Univariate analysis,
Multivariate analysis.
A total of 19 patients (23%) had a treatment-related complication. Of those who underwent tympanomastoidectomy, 14/54 (25.9%) had a treatment-related complication. These included surgical site infection, persistent tympanic membrane perforation, subsequent cholesteatoma, and jugular bulb diverticulum rupture. Of those who underwent treatment with antibiotics, 5/29 (17.2%) had a treatment-related complication. These included line infection and persistent nausea/vomiting. See Table 3 for a summary of the treatment-related complications.
Table 3.
Summary of treatment-related complications across all provided treatments, at ear level.
| Complication? | Sub type | N | % |
|---|---|---|---|
| None 64 (77%) | |||
| Surgical 14 (17%) | Surgical site infection | 5 | 6 |
| cholesteatoma | 3 | 3.6 | |
| persistent tympanic membrane perforation | 3 | 3.6 | |
| cholesteatoma, persistent tympanic membrane perforation | 1 | 1.2 | |
| Surgical site infection, jugular bulb diverticulum rupture | 1 | 1.2 | |
| Surgical site infection, persistent tympanic membrane perforation | 1 | 1.2 | |
| Medical 5 (6%) | PICC line infection | 3 | 3.6 |
| Nausea/vomiting | 2 | 2.4 |
4. Discussion
To examine the efficacy of aggressive medical versus surgical management, we reviewed data from patients who underwent treatment with parenteral antibiotic therapy and/or tympanomastoidectomy. In the unadjusted comparison, patients who underwent tympanomastoidectomy as the initial next step in therapy after more conservative initial therapy were more likely to have resolution of symptoms in a shorter period of time compared to those who received parenteral antibiotic therapy. Using the univariate cox proportional hazard model, initial surgical treatment and older age were both significantly associated with a shorter time to resolution, while bilateral disease status was significantly associated with a longer time to resolution. When treatment type, age, bilateral status and comorbidities were all included in the multivariate model, this effect was largely lost and only bilateral status was independently associated with a longer time to resolution. It should be noted, however, that the cohort who received parenteral antibiotic therapy as an initial treatment modality were more likely to have bilateral disease and be younger in age at time of initial treatment. Thus, it is difficult to delineate exactly what is contributing to the observed difference in time to resolution. It is interesting to note that the presence of comorbidities that would increase the likelihood of developing otitis media and/or more significant disease were not significantly associated with a longer time to resolution on either univariate or multivariate analysis. We did not compare number of treatments to resolution status as we focused our analysis on the initial treatment modality. Multiple treatments would inherently correspond to a longer time to resolution.
Interestingly, previously published literature has reported rates of infection resolution up to 89% with parenteral antibiotic therapy[11]. In our cohort, we observed a much lower rate of success with initial parenteral antibiotic therapy alone. This may be related to prior reported cohorts having undergone multiple treatment modalities prior to parenteral antibiotic therapy, longer treatment or follow-up periods, or inherent differences in the group of patients who were initially selected for parenteral antibiotic therapy, including eustachian tube anatomy, extent or severity of disease, or inadequate coverage of pathogenic bacterial species.
CSOM can be a difficult entity to treat if first-line otic and/or oral antibiotic therapy is not effective in resolving otorrhea. Options for more aggressive medical and surgical management include parenteral antibiotic therapy versus tympanomastoidectomy[6, 12, 13]. Whether medical or surgical management is chosen as the next step in treatment depends on several factors, including but not limited to microbiology, success of previous therapies, and physician preference. When deciding what the most appropriate next step in management of CSOM should be in patients who have failed otic and/or oral antibiotic therapy, one must consider potential complications. Apart from the risks of general anesthesia and continued postoperative drainage, serious complications from tympanomastoidectomy include but are not limited to iatrogenic labyrinthine fistula, hearing loss, facial nerve injury, dural injury and vascular injury[14]. Our cohort did include 1 case of intraoperative jugular bulb diverticulum rupture that was able to be controlled without any noted postoperative deficits. Parenteral antibiotic therapy includes the risk of line infections, adverse medication reactions, and other drug-specific toxicities. The risks and benefits of each treatment approach should be comprehensively reviewed with families and shared decision-making should be employed prior to proceeding with the next step in management. While our cohort may initially suggest that tympanomastoidectomy is the more effective approach to resolution of CSOM, this data should be interpreted cautiously, as this effect is lost on multivariate analysis. As such, it would still be justifiable to start with the nonoperative parenteral antibiotic therapy given previously published success rates, especially in younger patients with bilateral disease, and reserve tympanomastoidectomy for patients who continue to have evidence of disease at follow-up.
The primary limitations of our study were its retrospective study design and relatively small sample size. CSOM is a particularly difficult entity to study as there is currently no protocolized way to decide which patients are more appropriate for each treatment arm. Prior to requiring more aggressive medical or surgical management, patients with CSOM are often treated with different therapy regimens based on physician preference, and success of therapy may be affected by variable adherence to treatment regimens. As such, there may be inherent differences in severity and extent of disease which would be difficult to determine on physical exam.
5. Conclusion
Our study suggests that while tympanomastoidectomy generally has a shorter time to resolution of symptoms compared to parenteral antibiotic therapy, there is not a clear predictor for efficacy of tympanomastoidectomy versus parenteral antibiotic therapy for pediatric patients with CSOM who have failed previous otic and/or oral antibiotic therapy. A discussion regarding risks and benefits of each treatment modality and shared decision-making should be employed prior to proceeding with next steps in medical and surgical management.
Source of Financial Support or Funding:
This work was supported by the University of Utah Population Health Research (PHR) Foundation, with funding in part from the National Center for Research Resources and the National Center for Advancing Translational Sciences, National Institutes of Health, through Grant UL1TR002538.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Conflicts of Interest: None.
References:
- [1].Monasta L, Ronfani L, Marchetti F, et al. , Burden of disease caused by otitis media: Systematic review and global estimates, PLoS One. 7 (2012) 4 e36226. 10.1371/journal.pone.0036226. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [2].Schilder AGM, Marom T, Bhutta MF, et al. , Panel 7: Otitis media: Treatment and complications, Otolaryngology–Head and Neck Surgery. 156 (2017) 4_suppl S88–S105. 10.1177/0194599816633697. [DOI] [PubMed] [Google Scholar]
- [3].Briton BH, Otolaryngology, 3rd edition (4 volume set). Edited by michael m. Paparella, donald a. Shumrick, jack I. Gluckman, william I. Meyerhoff. Wb saunders, harcourt brace jovanovich, inc., Philadelphia, Pennsylvania, 1991, 3, 696 pp., $450.00, Head & Neck. 13 (1991) 6 564–564. 10.1002/hed.2880130619. [DOI] [Google Scholar]
- [4].Verhoeff M, van der Veen EL, Rovers MM, Sanders EA, Schilder AG, Chronic suppurative otitis media: A review, Int J Pediatr Otorhinolaryngol. 70 (2006) 11–12. 10.1016/j.ijporl.2005.08.021. [DOI] [PubMed] [Google Scholar]
- [5].Bluestone CD, Current management of chronic suppurative otitis media in infants and children, Pediatr Infect Dis J. 7 (1988) 11 Suppl S137–140. 10.1097/00006454-198811001-00003. [DOI] [PubMed] [Google Scholar]
- [6].Kenna MA, Rosane BA, Bluestone CD Medical management of chronic suppurative otitis media without cholesteatoma in children--update 1992, Am J Otol. 14 (1993) 5 469–473. [DOI] [PubMed] [Google Scholar]
- [7].Mittal R, Lisi CV, Gerring R, et al. , Current concepts in the pathogenesis and treatment of chronic suppurative otitis media, J Med Microbiol. 64 (2015) 10 1103–1116. 10.1099/jmm.0.000155. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [8].Vartiainen E, Results of surgical treatment for chronic noncholesteatomatous otitis media in the pediatric population, Int J Pediatr Otorhinolaryngol. 24(1992)3 209–216. 10.1016/0165-5876(92)90018-k. [DOI] [PubMed] [Google Scholar]
- [9].Fliss DM, Dagan R, Houri Z, Leiberman A, Medical management of chronic suppurative otitis media without cholesteatoma in children, J Pediatr. 116 (1990) 6 991–996. 10.1016/s0022-3476(05)80666-3. [DOI] [PubMed] [Google Scholar]
- [10].Team RC. A language and environment for statistical computing. R foundation for statistical computing. Vienna, Austria; 2021. [Google Scholar]
- [11].Kenna MA, Bluestone CD, Reilly JS, Lusk RP, Medical management of chronic suppurative otitis media without cholesteatoma in children, 96 (1986) 2 146–151. 10.1288/00005537-198602000-00004. [DOI] [PubMed] [Google Scholar]
- [12].Vartiainen E, Kansanen M, Tympanomastoidectomy for chronic otitis media without cholesteatoma, Otolaryngol Head Neck Surg. 106 (1992) 3 230–234. 10.1177/019459989210600304. [DOI] [PubMed] [Google Scholar]
- [13].Merchant SN, Wang P-C, Jang C-H, et al. , Efficacy of tympanomastoid surgery for control of infection in active chronic otitis media, Laryngoscope. 107 (1997) 7 872–877. 10.1097/00005537-199707000-00007. [DOI] [PubMed] [Google Scholar]
- [14].Wiet R, Harvey S, Littlefield P, Complications of surgery for chronic otitis media, in: Brackmann D, Shelton C, Arriaga M (Eds.), Otologic surgery, 4th ed. Elsevier, Philadelphia, PA, 2016. [Google Scholar]