Executive Summary

Dale Steele; Gaelen P Adam; Mengyang Di; Christopher Halladay; Ian Pan; Nathan Coppersmith; Ethan M Balk; Thomas A Trikalinos

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Steele D, Adam GP, Di M, et al. Tympanostomy Tubes in Children With Otitis Media [Internet]. Rockville (MD): Agency for Healthcare Research and Quality (US); 2017 May. (Comparative Effectiveness Reviews, No. 185.)

Cover of Tympanostomy Tubes in Children With Otitis Media

Tympanostomy Tubes in Children With Otitis Media [Internet].

Show details

Contents

< Prev Next >

Executive Summary

Background and Objectives

Otitis media is often preceded by a viral upper respiratory tract infection that causes Eustachian tube obstruction, negative middle ear pressure, and accumulation of fluid in the normally air-filled space of the middle ear. Acute otitis media (AOM) is defined as the presence of fluid in the middle ear with signs and symptoms of an acute infection, such as fever and ear pain. Otitis media with effusion (OME) is defined as the presence of fluid in the middle ear behind an intact tympanic membrane without signs and symptoms of an acute infection. ^1,2 OME is defined as chronic OME, if effusion persists for 3 months or longer.¹ Acute otitis media and chronic OME have shared causes. Children with chronic OME are prone to recurrent AOM episodes, and after an AOM episode all children have OME for some time.³ Chronic OME can result in hearing deficits, which put a child at risk for speech and language delays, behavioral changes, and poor academic achievement. Recurrent AOM has been shown to impact quality of life for patients and their caregivers.⁴

Certain children, including those with Down syndrome and cleft palate, have a very high risk for middle ear disease. The American Academy of Otolaryngology–Head and Neck Surgery (AAO-HNS) clinical practice guideline (CPG) identifies a subpopulation of children who may be at increased risk for speech, language, or learning problems from otitis media because of baseline sensory, physical, cognitive, or behavioral factors.^1,5

Myringotomy with TT placement is the most common ambulatory surgery performed on children in the United States⁶, with 667,000 TT placed in children under the age of 15 in 2006.⁷ The proceedings of the National Summit on Overuse, convened in 2012, based on sample of continually enrolled children into a treatment pathways database and a Medicaid database, reported that 2.5 percent of all U.S. children 2 years old and older had TT inserted in 2010.⁸

The effectiveness of TT for chronic OME and recurrent AOM is likely influenced by the many factors that affect the prognosis for middle ear disease in children, including current age, age at first diagnosis, frequency of respiratory tract infections, and day care exposure.⁹

The AAO-HNS CPG recommends that clinicians offer TT to children with recurrent AOM who have middle ear effusion at the time of assessment for tube candidacy, and that clinicians do not perform TT insertion when middle ear effusion is not present.¹

TT placement may result in acute otorrhea in some patients and conversely watchful waiting may result in continued episodes of recurrent AOM, which may include tympanic membrane perforation and otorrhea.

In children with TTs, episodes of otorrhea that reflect acute bacterial infection may be otherwise asymptomatic and less troublesome than AOM episodes in children with intact eardrums.¹⁰ However, otorrhea may be associated with a foul odor, fever, or pain, and it may negatively affect quality of life. Treatment is aimed at eradicating bacterial infection and reducing the duration and severity of symptoms.¹¹

The objectives for this systematic review are to synthesize information on the effectiveness of TT in children with chronic otitis media with effusion and recurrent acute otitis media, summarize the frequency of adverse effects or complications associated with TT placement, synthesize information on the necessity for water precautions in children with TT, and assess the effectiveness of available treatments for otorrhea in children who have TT.

Key Questions

With input from clinical experts during Topic Refinement, and from the public, during a public review period, we developed the following Key Questions (KQs) and study eligibility criteria.

Key Question 1.

For children with chronic otitis media with effusion, what is the effectiveness of TT, compared to watchful waiting, on resolution of middle ear effusion, hearing and vestibular outcomes, quality of life, and other patient-centered outcomes?

What factors (such as age, age of onset, duration of effusion, comorbidities, and sociodemographic risk factors) predict which children are likely to benefit most from the intervention?
Does obtaining a hearing test help identify which children are more likely to benefit from the intervention?

Key Question 2.

For children with recurrent acute otitis media, what is the effectiveness of TT, compared to watchful waiting with episodic or prophylactic antibiotic therapy, on the frequency and severity of otitis media, quality of life, and other patient-centered outcomes?

What factors (such as age, age of onset, number of recurrences, presence of persistent middle ear effusion, comorbidities, sociodemographic risk factors, history of complications of acute otitis media, antibiotic allergy or intolerance) predict which children are likely to benefit most from the intervention?

Key Question 3.

What adverse events, surgical complications, and sequelae are associated with inserting TT in children with either chronic otitis media with effusion or recurrent acute otitis media?

Key Question 4.

Do water precautions reduce the incidence of TT otorrhea or affect quality of life?

Key Question 5.

In children with TT otorrhea, what is the comparative effectiveness of topical antibiotic drops versus systemic antibiotics or watchful waiting on duration of otorrhea, quality of life, or need for tube removal?

Analytic Frameworks

The analytic frameworks in Figures A through C describe the specific linkages associating the populations of interest, exposures, modifying factors, and outcomes of interest in the assessment of studies that examine the association between TT placement, intermediate and final health outcomes, and harms (KQs 1, 2 and 3; Figure A); need for water precautions (KQ 4; Figure B); and treatment of otorrhea (KQ 5; Figure C).

Figure A

TT in children with chronic OME or recurrent AOM (Key Questions 1, 2, and 3). OME=otitis media with effusion; AOM=acute otitis media; TT=tympanostomy tubes; KQ=Key Question

This figure considers whether water precautions (key question 4) are associated with a reduced incidence of tympanostomy tube otorrhea or a change in bacterial pathogens cultured from those children who develop otorrhea. Global and otitis-specific child and parental quality of life are considered as the final health outcomes that may be associated with water precautions.

Figure B

Need for water precautions in children with TT (Key Question 4). OME=otitis media with effusion; AOM=acute otitis media; QoL= Quality of Life

Figure C

Treatment of otorrhea in children with TT (Key Question 5). OME=otitis media with effusion; AOM=acute otitis media; QoL= Quality of Life; TT=tympanostomy tube

Methods

The Brown Evidence-based Practice Center (EPC) conducted this review based on a systematic review of the published scientific literature using established methodologies as outlined in the Agency for Healthcare Research and Quality’s (AHRQ) Methods Guide for Comparative Effectiveness Reviews.¹² The PROSPERO protocol registration number is CRD42015029623.

Eligibility Criteria

We use the Population, Intervention, Comparator, Outcomes, and Designs (PICOD) formalism to define the characteristics of the eligible studies for this review.

Populations

For all KQs, studies of children and adolescents from 1 month to 18 years old were eligible. Subpopulations of interest included children at high risk of recurrent AOM or OME, such as children with Down syndrome, cleft palate, other craniofacial anomalies, and primary ciliary dyskinesia; and children at high risk of adverse clinical or developmental outcomes, such as those with preexisting hearing loss, speech and language problems, or developmental disorders. We were also interested in the population of children who have sociodemographic risk factors, such as day care exposure or low socioeconomic status.

For KQ 1, we included studies of children with chronic OME. We preferred the standard definition of effusion that persists for at least three months,¹ but included results based on studies’ alternative definitions if our preferred one was not reported. We excluded children with chronic suppurative otitis media since it is usually associated with a persistently perforated tympanic membrane.

For KQ 2, we included children with recurrent AOM with or without middle ear effusion, defined as three or more well-documented and separate AOM episodes in the past 6 months or at least four well-documented and separate AOM episodes in the past 12 months with at least one in the past 6 months.¹ For studies that did not report the preferred definition, we recorded the study specific definition.

For KQ 3 and 4, we included studies in children with TT placed for OME or AOM. For KQ 5, we included studies of symptomatic or asymptomatic children with acute TT otorrhea beyond the immediate postoperative period. We defined the immediate postoperative period as 30 days after surgery, but included studies reporting results near that period (e.g., 28 days, 4 weeks).

Interventions/Exposures

For KQs 1, 2 and 3, we considered all studies that included myringtomy with TT placement, with or without adenoidectomy. Tubes were classified as short-term tubes (generally in place for 10–18 months) and long-term tubes (which typically remain in place for several years).

In KQ 4, we distinguished three categories of interventions; avoidance of swimming or head immersion while bathing, canal occlusion methods (e.g. earplugs or headbands), and postexposure prophylaxis using ototopical antibiotics.

KQ 5 compares ototopical preparations, and includes products approved by the U.S. Food and Drug Administration (FDA) (i.e., ofloxacin otic 0.3%, ciprofloxacin 0.3% and dexamethasone 0.1%), and other non–FDA-approved agents, such as hydrocortisone, bacitracin, and colistin.

Comparators

For KQ 1, comparisons of primary interest were watchful waiting or adenoidectomy. Comparators for KQ 2 included watchful waiting, systemic or topical antibiotic therapy for recurrent episodes of AOM, prophylactic antibiotics, and adenoidectomy. KQ 3 did not address comparative harms. In KQ 4, comparators included no water precautions with or without avoidance of higher risk activities (e.g. diving or underwater swimming), and ear plugs or swimming caps. The primary comparators for KQ5 were watchful waiting and oral antibiotic therapy.

Outcomes

For KQs 1 and 2, which address the effectiveness of TT, we considered intermediate outcomes, including the prevalence of middle ear effusion, measures of hearing and vestibular function, such as improved hearing thresholds (audibility), tests of auditory perception and discrimination (clarity), and balance and coordination (vestibular function). For KQ 2, measures of recurrent AOM, including otorrhea were extracted.

Quality of life and patient-centered outcomes were considered, including global and otitis-specific child and parental quality of life, speech and language outcomes, educational achievement, behavioral outcomes such as disobedience, enuresis, or tantrums.

The following outcomes were extracted for KQ 3: Intraoperative and immediate postoperative anesthetic and surgical adverse events, otorrhea beyond the postoperative period, blockage of the tube lumen, granulation tissue, premature extrusion, TT displacement into the middle ear, persistent perforation of the tympanic membrane, myringosclerosis, tympanic membrane atrophy, atelectasis and retraction pockets, worsened hearing thresholds, and other reported (plausibly related to tubes).

Outcomes for KQ 4 included final health and patient-centered outcomes related to child and parental quality of life and intermediate outcomes related to the incidence and duration of otorrhea. Outcomes evaluated relating to KQ 5 (treatment of otorrhea) included global and otitis-specific child and parental quality of life, duration of otorrhea, and need for removal of TT.

Timing

We included studies with any duration of followup.

Setting

We included studies performed in both primary and specialty care settings.

Study Design

We evaluated published, peer-reviewed studies only. For KQs 1, 2, 4, and 5, we included randomized comparative trials and nonrandomized comparative studies, prospective and retrospective where treatment was assigned on a per patient basis. Studies with per ear assignment were excluded (e.g. tubes placed by design in one ear only). For KQ 3, we included prospective surgical single group studies enrolling at least 50 subjects (including arms treated with TT that were part of randomized controlled trials [RCTs] or nonrandomized comparative studies [NRCSs]) and population based retrospective single group studies (registry studies) with at least 1000 subjects.

Searching for the Evidence

We conducted literature searches of all studies in MEDLINE^®, the Cochrane Central Trials Registry and Cochrane Database of Systematic Reviews, Embase^®, and CINAHL^® databases (details in Appendix A of the full report). The last search was run on May 19, 2016. Additionally, we perused the reference lists of published clinical practice guidelines, relevant narrative and systematic reviews, and Scientific Information Packages from manufacturers. Citations were independently screened by two researchers in the open-source, online software Abstrackr (http://abstrackr.cebm.brown.edu/).

Data Extraction and Data Management

Each study was extracted by one methodologist and confirmed by at least one other methodologist. Data was extracted into customized forms in the Systematic Review Data Repository (SRDR) online system (http://srdr.ahrq.gov). Excluded studies are listed in Appendix B of the full report. Details of included studies are summarized in Appendix C, D and E of the full report.

Assessment of Risk of Bias of Individual Studies

We assessed the methodological quality of each study based on predefined criteria. For RCTs, we used the Cochrane risk of bias tool.¹³ For observational studies, we used relevant questions from the Newcastle Ottawa Scale.¹⁴

Data Synthesis

All included studies were summarized in narrative form and in summary tables that record the important features of the study populations, design, intervention, outcomes, and results.

We performed network meta-analysis of clinical outcomes to compare treatment alternatives across studies for KQs 1 and 5. We also conducted pairwise comparisons by means of random effects meta-analyses of comparative studies. Specific methods and metrics (summary measures) meta-analyzed were chosen based on available, reported study data. When available, these were summarized as odds ratios of categorical outcomes and net change of continuous outcomes (e.g., mean hearing thresholds). Statistical heterogeneity was explored qualitatively. We explored subgroup differences within across studies based on the list of comparisons described in the KQs.

Grading the Strength of Evidence

We graded the strength of evidence (SOE) as per the AHRQ Methods Guide on assessing the strength of evidence.¹⁵

Assessing Applicability

We assessed the direct applicability within and across studies with reference to children with specific comorbidities (Down syndrome, cleft palate, etc.), and whether interventions and comparators are used in current practice.

Results

The literature search yielded 10,129 citations (Figure D). We identified 481 of these as potentially relevant full-text studies, and retrieved them for further evaluation. Overall, 306 full text articles did not meet eligibility criteria (see Appendix B of the full report for a list of rejected articles along with reasons for rejection); thus 184 articles are included in this report.

Figure D

Literature flow diagram. CINAHL = Cumulative Index to Nursing and Allied Health Literature; KQ = Key Question; NRCS = nonrandomized comparative study; RCT = randomized controlled trial; Some publications reported data from the same study. The KQ3 publications (more...)

A trial registry search did not turn up any completed trial that was not already identified through literature searches. As shown in Figure D, the majority of included publications (n=98) related to KQ3, with 50 related to KQ1. There is a relative paucity of studies available for the other KQs.

Key Question 1

We identified 54 publications Of these, there were 29 papers reporting results of 16 RCTs. There were 24 publications reporting 24 NRCSs that assessed the effectiveness of TT in pediatric patients with chronic middle ear effusion. These studies evaluated multiple interventions (TT, TT with adenoidectomy, myringotomy with adenoidectomy, myringtomy alone, adenoidectomy alone, oral antibiotic prophylaxis, and watchful waiting). Two studies included at least some patients with recurrent AOM with or without persistent middle ear effusion.

Randomized Comparative Studies

Hearing thresholds were measured in 16 RCTs. In 10 of these, mean hearing thresholds were reported by arm at various time points. For the network meta-analysis of these RCTs, we classified hearing thresholds obtained at one to three months as “early”. Similarly, hearing thresholds obtained between 12 and 24 months where were classified “late”. Not all studies had interventions at both “early” and “late” time points. Thus, the network of comparators differs for “early” and “late” comparisons. Figure E shows the topology of the network for early hearing thresholds at 1 to 3 months. Such network plots are a visual representation of the evidence base.

A topology map of the network of available comparisons used for meta-analysis of different interventions in children with chronic middle ear effusion. Six classes of intervention are displayed; including Tympanostomy tubes, Tympanostomy tubes and Adenoidectomy, Watchful waiting, Myringotomy, Myringotomy and Adenoidectomy and antibiotic prophylaxis. Of the 15 possible pairwise contrasts, 7 are covered by direct comparisons. Three studies provided direct comparisons between Tympanostomy tubes and watchful waiting. Three studies also allow direct comparisons of Tympanostomy with myringotomy. Several comparisons were directly compared in only a single study: Antibiotic prophylaxis with Tympanostomy, Tympanostomy and Adenoidectomy with Tympanostomy, or Watchful waiting, or Myringtomy and Adenoidectomy.

Figure E

Network graph of comparators for early (1 to 3 months) hearing thresholds. The network plot consists of nodes representing the interventions being compared and edges representing the available direct comparisons. The number of studies that include each (more...)

Figure F illustrates the effectiveness of various interventions at 1 to 3 months, compared with watchful waiting. Mean hearing thresholds improved (decreased) by average of 9.1 dB following TT, and by 10 dB following TT with adenoidectomy. Credible intervals for these effects exclude zero. The credible intervals for comparisons between watchful waiting and myringotomy alone, myringotomy with adenoidectomy, and oral antibiotic prophylaxis were wide.

Illustrates the network meta-analysis combining direct and indirect evidence for the effectiveness of various interventions at 1 to 3 months, compared with watchful waiting. Mean hearing levels improved (decreased) by average of 9.1 dB (95% credible interval: −14 to −3.4) following Tympanostomy tubes, and by 10 dB (95% credible interval: −19.0 to −1.9) following Tympanostomy tubes with adenoidectomy. Credible intervals for these effects exclude zero. The decrease in mean hearing levels for Myringotomy compared to watchful waiting was 2.3 dB (95% credible interval −9.2 to 3.9). The credible intervals for comparisons between watchful waiting and myringotomy alone, myringotomy with adenoidectomy, and oral antibiotic prophylaxis were wide and include zero.

Figure F

Early (1 to 3 months) decrease (improvement) in mean hearing thresholds compared with watchful waiting. TT= tympanostomy tubes; CrI=Credible Interval

As shown in Table A, the strategies with the highest probability of being among the three most effective interventions with respect to early improvements in hearing thresholds were TT, TT with adenoidectomy, and myringotomy with adenoidectomy.

Table A

Probabilities (percent) that an intervention is among the three most effective with respect to early hearing thresholds.

Five RCTs reported hearing thresholds at 12 to 24 months. Figure G shows the topology of the network of comparisons at this time interval.

Figure G

Network graph of comparators for late (12 to 24 months) hearing thresholds. The network plot consists of nodes representing the interventions being compared and edges representing the available direct comparisons. The number of studies that include each (more...)

As shown in Figure H, by 12 to 24 months, the mean difference in hearing thresholds for TT alone, compared to watchful waiting was 0 dB (95% CrI [credible interval] −4 to 3). Compared to watchful waiting, myringotomy with adenoidectomy and TT with adenoidectomy have better hearing outcomes by about 4 dB, but the 95 percent credible intervals include zero.

Figure H

Late (12 to 24 months) decrease (improvement) in mean hearing thresholds compared with watchful waiting. TT= tympanostomy tubes; CrI=Credible Interval

As can be seen in Table B, TT with adenoidectomy and myringotomy with adenoidectomy were the two most effective strategies with respect to late hearing thresholds. TT alone, antibiotic prophylaxis, and watchful waiting were among the three least effective ones.

Table B

Probabilities (percent) that an intervention is among the two most effective with respect to late hearing thresholds.

The results for the studies that reported measuring hearing thresholds, but did not report mean hearing thresholds are summarized in the full report.

A network meta-analysis of the mean duration of middle ear effusion is presented in the full report.

Nonrandomized Comparative Studies

The nonrandomized comparative studies (NRCSs) are summarized in the full report. The NRCSs evaluated special populations and are summarized here. Six studies reported results in the populations with comorbidities of interest, including cleft palate/lip and primary ciliary dyskinesia. Three studies (two in cleft palate and one in primary ciliary dyskinesia) compared TT placement with nonsurgical treatment, while one study compared early versus delayed TT in different settings. Two studies assessed the effects of TT and cleft repairing versus cleft repairing alone. Hearing thresholds reported as pure tone averages were reported in four studies. In patients with cleft palate/lip and primary ciliary dyskinesia, respectively, average hearing threshold was lower in TT than non-surgical control, but the difference was not significant. TT in addition to cleft repair improved hearing thresholds with unknown significance. The improvement by early (mean age 3 months) compared to delayed (mean age 40.8 months or not at all in two subjects) TT procedures in patients with cleft palate was marginally significant (P=0.05 for ears with better hearing and P=0.10 for ear with worse hearing). The rate of normal hearing, defined as hearing threshold < 20 dB bilaterally, was significantly higher in TT than control (P <0.05).

Quality of Life and Patient-Centered Outcomes

Eight studies (five RCTs, three NRCSs, and one that combined both designs) in 12 papers reported on 119 quality of life and patient-centered outcomes (cognitive, language, and behavioral) in 1665 children over multiple time points and arms. These studies reported only 14 outcomes with significant results. In general, the results were not significant and varied in magnitude and direction, even across subscales of the same test.

Only two studies reported specifically on quality of life outcomes: Paradise reported on measures of parental stress at various ages, and Vlastos reported on pediatric health related quality of life. Neither found any significant differences. Full details for all outcomes are in Appendix G of the full report.

Key Question 2

We identified 8 publications, reporting 7 RCTs and 2 NRCSs. The Matilla 2003 paper reported two groups, an RCT which randomly allocated treatment in 137 patients, and a NRCS in which parental choice determined treatment in169 patients. Three RCTs compared TT placement with daily oral antibiotic prophylaxis. Two of these studies included a comparison with placebo and the third compared TT placement with no treatment. The effectiveness of TT alone versus TT with adenoidectomy was evaluable in three studies.

Randomized Comparative Studies

Frequency and Severity of Recurrent Acute Otitis Media

The majority of studies were done prior to widespread use of the conjugate pneumococcal vaccine, in an era where antibiotic resistance was less common, and prophylactic oral antibiotic therapy was more commonly used in clinical practice. Results are summarized by comparison below.

TT Versus Placebo or No Treatment

Gonzalez 1986 reported that in the placebo group three of 20 children had no further episodes of AOM, compared to 12 of 22 in the TT group (P = 0.01, an attack rate of 2.0 in the placebo group, compared to 0.86 in the TT group (P = 0.006). In a post-hoc subgroup comparison of children who had middle ear effusion upon entering the study, attack rate and number of patients who had no further bouts of AOM was significantly better (P < 0.05) in those children without middle ear effusion. However, this group consisted of only 22 patients.

Casselbrant 1992 reported the rate of new episodes per arm was 1.08 in the placebo group versus 1.02 in the TT group (P = 0.25). In the placebo group, 40 percent had no further episodes of AOM, compared to 35 percent in the TT group. In addition, TT placement significantly decreased the percentage of time with AOM compared to placebo (P < 0.001).

Kujala 2012 reported failure rates (defined as at least two episodes of AOM in 2 months, three in 6 months or persistent effusion lasting at least 2 months), percent of children with no recurrent AOM, cumulative number of AOM episodes, and one-year incidence rates. There was an absolute difference in the proportion of failures of −13 percent (95% CI −25 to −01) between the TT and control groups, favoring TT. The one year incidence rate (infections per child per year) was 0.55 (95% CI 0.93 to 0.17) lower in the TT group compared to the control group.

TT Versus Antibiotic Prophylaxis

In the Gonzalez 1986 RCT, 54.5 percent of children in the TT group and 24 percent in the sulfisoxazole prophylaxis group had no recurrent AOM (P = 0.02). The attack rate was 0.86 infections per child in the TT group and 1.4 in prophylaxis group (P = 0.08).

Casselbrant 1992 reported a rate of 0.6 episodes of recurrent AOM per child-year children treated with Amoxicillin and a rate of 1.02 in their TT group (P = 0.001).

El-Sayed found no difference in the treatment outcomes of children treated with trimethoprim/sulfamethoxazole compared to children treated with TT (P = 0.37).

TT Versus TT and Adenoidectomy

An RCT by Mattila 2003 found no difference in cumulative hazard of recurrent AOM or in efficacy, defined as one minus the adjusted relative risk in randomized and nonrandomized comparisons of children who underwent TT with adenoidectomy compared with TT alone.

In the Kujala 2012 study, there was no significant difference in the TT with adenoidectomy group compared to the TT only group in the number of failures (absolute difference −5%, 95% CI −16 to 6, P = 0.37), in the time to failure (P = 0.29) or to the first AOM (P = 0.6), or in the proportion of children with no AOM episodes (absolute difference 1%, CI −13 to 15, P =1.0).

A subsequent 2005 RCT, which enrolled 217 children, 162 of whom had recurrent AOM, again found no differences in the mean number of otitis media episodes overall or in the subgroup of children with recurrent AOM at enrollment.

Quality of Life Outcomes

Although Kujala 2014 found that insertion of TT tubes, with or without adenoidectomy, significantly reduced the risk of recurrent AOM, a subsequent publication from the same trial examining quality of life outcomes at study entry, 4 months and 12 months found no differences in overall ear-related quality of life (Otitis Media-6 questionnaire [OM-6]), or for the subscales of caregiver concern, emotional distress, physical suffering, activity limitations, hearing loss, or speech impairment between surgically treated and no surgery groups.

Nonrandomized Comparative Studies

In a cross-sectional study, Grindler 2014 reported both disease-specific quality of life outcomes, utilizing OM-6 score, and health related quality of life, using the PedsWL Infant Impact Module, in 1208 patients. The OM-6 score was higher (reflecting worse otitis specific quality of life) in children in otolaryngology practices who had been recommended to undergo TT placement than in children with prior TT placement.

Key Question 2a

There are no prospective planned comparisons evaluating whether the presence of middle ear effusion (at time of surgical evaluation) modifies the effectiveness of TT placement for recurrent AOM. Gonzalez 1986 report a retrospective subgroup comparison based on the presence or absence of middle ear effusion at initial evaluation and conclude that the attack rate, as well as the number of patients who had no further bouts of AOM, was significantly better (p < 0.05) in those children without middle ear effusion. However, this group consisted of only 22 patients. Two studies specifically excluded patients with middle ear effusion at time of surgical evaluation.

Casselbrant 1992 analyzed data with a multivariable Poisson model, and concluded that TT reduced the number of episodes of AOM/otorrhea only in those subjects whose episodes of AOM occurred year round. In their model, younger age and white race were significantly associated with higher rates of recurrent AOM, but there treatment by age and treatment by race interactions were not found.

Key Question 3

We extracted data on the occurrence of 11 adverse events from 85 cohorts and from RCTs and NRCSs included in KQs 1 and 2. The adverse events considered were: perioperative complications, otorrhea, tube blockage, granulation tissue formation, premature extrusion, displacement of the TT into the middle ear space, persistent perforation, myringosclerosis (tympanosclerosis), presence of atrophy, atelectasis or retraction, cholesteatoma and long term hearing loss. We did not consider other adverse events, such as antibiotic resistance, gastrointestinal side effects of antibiotics or pain related to ear drops. The number of publications reporting each event, and the median (with 25^th and 75^th percentiles) percent of patients and ears are summarized in Table C.

Table C

Median percentage of patients and ears with adverse events associated with TT placement.

See Appendix G of the full report for adverse event details by study, and for study specific details, including design, recruitment period, tube type(s) used, age, proportion with recurrent AOM, followup time, and study specific definitions. In general, the study specific definitions of adverse events are poorly reported and/or highly variable between studies.

Key Question 4

We identified 11 publications which reported 2 RCTs and 9 NRCSs, which evaluate a range of interventions, from complete water restriction (e.g., no swimming or head immersion while bathing), physical protection while swimming (utilizing ear plugs or bathing caps), postexposure prophylactic ear drops, avoidance of high risk activities (e.g., diving), to completely unrestricted exposure to water. All studies compared either no-swimming or ear plugs with a second group of swimmers with or without post-exposure antibiotic ear drops.

In the two RCTs, Goldstein 2005 reported a slightly higher average rate of otorrhea per month in children who did not wear ear plugs (mean 0.10 episodes/month, compared to a mean of 0.07; P = 0.05) in a Poisson regression model adjusting for compliance. Parker 1994 reported identical mean otorrhea rates in nonswimmers and swimmers. Table D summarizes the occurrence of one or more episodes of otorrhea in the RCTs.

Table D

RCTs: Water precautions—one or more episodes of otorrhea.

The forest plot in Figure I, summarizes the results of a random effects meta-analysis from the NRCSs only with separate summary estimates for ear plugs and avoidance of swimming. The summary odds ratio comparing ear plugs versus no precautions of having one or more episodes of otorrhea was 1.70 (95% CI 0.95 to 3.06). The odds ratio for nonswimming compared to no precautions was 1.52 (95% CI 0.71 to 3.25). It is notable that events rates in the RCTs are systematically higher in both control and intervention arms in the RCTs compared with event rates in NRCSs. A possible explanation is more complete ascertainment of outcomes in RCTs.

Illustrates the effectiveness of nonrandomized comparative studies of ear plugs or nonswimming compared to no precautions. The summary odds ratio for the comparison of ear plugs was 1.7, with a 95 percent confidence interval ranging from 0.95 to 3.06. For the nonswimming vs. no precautions, the summary odds ratio was 1.52 (95% confidence interval 0.71 to 3.25).

Figure I

Nonrandomized comparative studies only, children with one or more episodes of otorrhea. CI= Confidence Interval; NRCS= Nonrandomized Comparative Study; OR = Odds ratio (values > 1 favor ‘no precautions’ arms; values < 1 (more...)

There appears to be a statistically nonsignificant trend in the NRCSs, which favor no ear plugs and no precautions. This trend may reflect a possible bias (e.g. patients who chose to swim may be less likely to report minor degrees of otorrhea).

Overall, aside from the small reduction in mean number of episodes of otorrhea found in the Goldstein RCT, the available evidence does not support the conclusion that either ear plugs or avoidance of swimming reduces the risk of otorrhea related to swimming.

Key Question 5

We identified 12 papers, representing 11 studies, reporting 10 RCTs and 1 NRCS, with a total of 1811 patients analyzed (1405 in RCTs and 406 in NRCSs) that assessed the effectiveness of various interventions to treat TT otorrhea. The studies reported multiple comparisons, including oral antibiotics (amoxicillin and amoxicillin/clavulanate), various antibiotic drops and antibiotic-glucocorticoid drops, oral corticosteroids, and combinations. Several studies had a watchful waiting or placebo arm.

Two studies were excluded from our meta-analysis, the NRCS by Dohar where specific treatments used in the historical practice group and concurrent practice group were not reported and a study which compared an antibiotic-glucocorticoid topical drop containing neomycin sulfate, polymyxin B sulfate and hydrocortisone with a topical spray formulation containing neomycin sulfate and dexamethasone. The network of available comparisons is shown in Figure J.

A topology map of the network of available comparisons used for meta-analysis of different interventions for treatment of Tympanostomy tube otorrhea. Six classes of intervention are displayed; Watchful waiting or Placebo, Oral antibiotic, topical antibiotic, topical antibiotic-glucocorticoid, oral antibiotic and oral glucocorticoid and topical antibiotic-glucocorticoid and oral antibiotic. Of the 15 possible pairwise contrasts, 8 are covered by direct comparisons. Three studies provide direct comparisons between oral antibiotics and Watchful waiting or Placebo. Two studies contrast oral antibiotics with topical antibiotic treatments and two studies contrast topical antibiotic-glucocorticoid drops with oral antibiotics. A single study directly compares Watchful waiting or placebo with topical antibiotics and with topical antibiotic-glucocorticoid treatments.

Figure J

Network of treatment comparisons (RCTs). RCT= Randomized Control Trial; PO=Oral The network plot consists of nodes representing the interventions being compared and edges representing the available direct comparisons. The number of studies that include (more...)