Chapter  12:  Otitis Media with Effusion in Young Children: Clinical Practice Guideline Number 12

Purpose and Scope of the Guideline

The purpose of the Clinical Practice Guideline on Otitis Media with Effusion in Young Children is to assist those who provide health care to young children in identifying and managing otitis media with effusion.

To meet constraints for Guideline development, the Panel focused the Guideline narrowly on management of otitis media with effusion in a "target child" age 1 through 3 years with no craniofacial or neurologic abnormalities or sensory deficits, otherwise healthy except for otitis media with effusion. In most cases, the Panel's findings regarding the effects of bilateral otitis media with effusion on hearing-related development and the timing of diagnostic tests/hearing evaluation and interventions to manage otitis media with effusion are described for this "target child." When possible, Guideline recommendations are broadened to include older children. The Panel also considered costs and availability of health care resources to manage otitis media with effusion in young children.

Organization of the Guideline

This first chapter of the Guideline provides background on otitis media with effusion, including prevalence of otitis media, and discusses the "target" patient for whom Guideline recommendations were formulated. Chapter 2 outlines the results of the Panel's evaluation of the literature on otitis media with effusion. The results are presented in the form of statements of options or recommendations and an algorithm for diagnosis, evaluation, and management. Chapter 3 presents the short-term and long-term outcomes of otitis media with effusion. Diagnostic issues regarding otitis media with effusion in children are addressed in Chapter 4.

Chapter 5 discusses control of environmental risk factors for otitis media with effusion. Chapter 6 reviews pharmaceutical interventions, and Chapter 7 reviews the Panel's findings regarding surgical interventions. Chapter 8 discusses results of the Panel's study of the association of allergy with otitis media with effusion, and Chapter 9 reviews findings regarding other treatments for this condition. Cost impacts of otitis media with effusion in children are presented in Chapter 10. In Chapter 11, research issues identified by the Panel are listed.

Target Population, Settings, and Providers of Care

Following methods recommended by the Agency for Health Care Policy and Research for guideline development, the Otitis Media Guideline Panel defined a "target patient" to narrow the scope of the literature review. The Panel also identified practice settings and providers of care to the target patient.

The target patient is a child age 1 through 3 years with no craniofacial or neurologic abnormalities or sensory deficits, otherwise healthy except for otitis media with effusion.

The Panel developed the Guideline for use in any setting in which children at risk of otitis media with effusion would be identified or treated, including physician offices, outpatient clinics, hospital emergency departments or urgent care centers, and schools or child-care facilities.

The Guideline is intended for use by providers of health care to young children, including primary care and specialist physicians, professional nurses and nurse practitioners, physician assistants, audiologists, speech and language pathologists, child development specialists, and consumers.

Definition and Prevalence of Otitis Media with Effusion

The term otitis media (inflammation of the middle ear) encompasses a number of clinical conditions, each called by a variety of names. Statistics compiled by the National Center for Health Statistics (Schappert, 1992) and the results of longitudinal studies provide information about the incidence of and risk factors for otitis media.

Otitis media is the diagnosis made most frequently at visits by children younger than age 15 years to office-based physicians in the United States. Furthermore, the estimated 24.5 million visits for otitis media made in 1990 represented an increase of nearly 150 percent over the number of such visits made in 1975 (Schappert, 1992). In a large pediatric office practice in Boston, approximately one visit in three made for illness of any kind resulted in the diagnosis of middle ear disease; three-quarters of all followup visits for illness were made for disease of the middle ear (Teele, Klein, Rosner, et al., 1983).

A type of otitis media called otitis media with effusion is characterized by fluid in the middle ear without evidence of ear infection. Other terms used for otitis media with effusion in the literature include:

The Panel did not find scientific evidence regarding the frequency with which the specific diagnosis of otitis media with effusion is made for children visiting ambulatory health care facilities in the United States. For the purposes of guideline development, the Panel estimated that in 25 to 35 percent of cases cited in the National Center for Health Statistics report (Schappert, 1992), the diagnosis "otitis media" represented otitis media with effusion.

The prevalence and natural history of otitis media with effusion have been reported in several research studies. However, much of the data include children in group child-care settings, which are a very select population. Also, the data include different, overlapping age groups. A 2-year study of 103 children age 2 to 6 years in group child care in the United States showed that 53 percent had at least one episode of otitis media with effusion during the first year of the study and 61 percent in the second year; 30 percent of the children had recurrent bouts of otitis media with effusion (Casselbrant, Brostoff, Cantekin, et al., 1985). A study of school children 5 to 12 years old found the incidence of otitis media with effusion to be much lower in children 6 years and older than in younger children (Casselbrant, Brostoff, Cantekin, et al., 1986), in keeping with the decreasing prevalence with increase in age reported by the National Center for Health Statistics (Schappert, 1992).

Otitis media with effusion cleared within 2 months in 80 percent of 103 children age 2 to 6 years in group child care in the United States (Casselbrant, Brostoff, Cantekin, et al., 1985). In a cohort study of 1,439 Dutch children, approximately 60 percent had recovered from the episode of otitis media with effusion without intervention after 3 months (Zielhuis, Straatman, Rach, et al., 1990). The latter clearance rate is cited throughout the Guideline because of the study's large sample size.

Methods for Guideline Development

The Consortium (American Academy of Pediatrics, American Academy of Family Physicians, and American Academy of Otolaryngology -Head and Neck Surgery) received 37 names from 32 organizations and 2 individuals for nomination to the interim panel. After review of the nominations, the Consortium recommended a 15-member interim panel to the Agency for Health Care Policy and Research based on nominees' expertise, geographic location, race, and gender. Subsequently, a notice was published in the December 2, 1991, Federal Register and sent to the 32 organizations already solicited and 8 other organizations. From the resulting 18 additional panel nominations, the Consortium recommended adding 4 additional members to the panel. With Agency for Health Care Policy and Research approval, the Consortium appointed an Otitis Media Guideline Panel consisting of 17 Panel members and 2 Panel co-chairs. The Panel held four 2-day meetings over a 14-month period.

The Panel based their approach to guideline development on Agency for Health Care Policy and Research recommendations and the principles of Eddy (1992). The Panel was introduced to Eddy's principles of evidence-based clinical policy development at its first two meetings. Using these methods, the Panel (1) specified clearly the clinical questions to be addressed by the Guideline, which was achieved by identifying the "target patient" and listing the interventions and outcomes to be explored; (2) developed literature search strategies; (3) examined the literature critically; and (4) developed evidence tables based on the data extracted from the literature review.

The Panel divided into subcommittees to examine the data on interventions for otitis media with effusion. Each subcommittee determined the type of data that would be acceptable for inclusion in the evidence regarding the intervention. The subcommittees examining pharmaceutical or surgical interventions accepted only randomized, blinded, controlled studies for inclusion in the evidence for these interventions. For diagnostic interventions, the subcommittee reviewed the diagnostic gold standard pursuant to deciding which studies to include in the evidence. For control of environmental risk factors, the subcommittee looked for high-quality epidemiologic studies. For speech, language, and developmental outcomes, controlled studies that documented accurate diagnosis of otitis media with effusion in the target population of children were accepted for inclusion in the evidence.

The Panel also appointed subcommittees as necessary to review controversial areas, such as speech and language development and antibiotic treatment of otitis media with effusion. The report of the Subcommittee for Scientific Assessment, which reviewed the data on antibiotic treatment, is included in the Guideline Technical Report. Furthermore, the Panel and contractors guided and assisted those subcontracted to conduct the literature review (Children's Hospital of Pittsburgh), to perform the cost analysis (Lewin-VHI, Inc.), and to provide technical writing services (Children's Hospital of Pittsburgh).

The literature search was conducted using the on-line bibliographic data base of the National Library of Medicine and 10 specialized bibliographic data bases. The data bases searched, the search strategies, and the results are detailed in the Guideline Technical Report. Initial searches resulted in identification of 3,578 bibliographic citations, for which 1,362 abstracts were obtained and evaluated. Based on the initial review, 378 full-text articles were selected for data extraction using a nine-page form developed by the methodologist and Panel. More than 100 additional bibliographic citations were identified through review of literature reference lists, updated on-line searches, requests to professionals, and suggestions from experts. The interim panel chair also contacted numerous professional organizations to solicit unpublished data on this topic. The Guideline includes 158 references. In addition, the Panel sought information about otitis media with effusion by holding an Open Meeting that was announced in the Federal Register. Twelve oral testimonies were presented, eight of which were also written; additionally, twelve written testimonies were provided.

For each article considered for inclusion in the evidence, two different individuals extracted data, thus creating two records for each article in the data base. Because of the controversy regarding otitis media research at the University of Pittsburgh, all articles with a connection to the University of Pittsburgh were reviewed and data extracted by an independent professional. Literature review results were entered into a dBase (Ashton-Tate, Torrance, CA) data base with more than 720 fields. Entries in the data base were then corrected to eliminate any differences between extractions. The Panel reviewed the evidence in the data base and focused attention on randomized, blinded, controlled clinical studies wherever possible, although such evidence was not available in all areas of investigation. The Panel conducted an additional literature search and evaluation in the area of speech, language, and behavioral outcomes of otitis media with effusion. When possible, literature relevant to a given outcome was subjected to meta-analysis using FAST*PRO (Eddy and Hasselblad, 1992) to combine data by either Bayesian or hierarchical Bayesian techniques. To ensure against bias, meta-analysis of the data on antibiotic treatment of otitis media with effusion was reviewed by an independent expert.

During parts of each meeting, the Panel divided into subgroups to expedite the process of reviewing literature and making recommendations on diagnosis and the clinical interventions examined. Special subcommittees worked between Panel meetings to examine certain issues in greater depth, especially those of antibiotic treatment and of long-term outcomes. All deliberations of subgroups were brought back to the full Panel for discussion, revision, and approval. During Panel meetings, sessions were chaired by the co-chairs and the methodologist. The goal was to reach consensus on every decision. When consensus was not possible, votes were taken. If fewer than four Panel members disagreed with the decision, the decision was carried forward without comment. Many topics covered in the Guideline are controversial and remained so after review of the scientific evidence. Some Panel members strongly dissented from the majority opinion on individual recommendations and have written minority comments, included in an appendix to the Guideline Technical Report.

The paragraph below presents the grading system the Panel developed and summarizes the outcome of their deliberations regarding diagnostic and treatment interventions for otitis media with effusion in the target child. Of note is that the final recommendations are at least partially subjective, notwithstanding the level of objective scientific evidence for or against the recommendation. First, judgments about the quality of the science could not be fully objective; members of the Panel differed significantly in their individual judgments on issues such as long-term outcomes and the use of steroid and antibiotic medications, for example, even though all were evaluating the same data base. Second, the Panel occasionally linked weaker scientific evidence with strong theoretical arguments to construct a strong recommendation. An example is the recommendation regarding obtaining a hearing test before myringotomy and insertion of tympanostomy tubes. Each recommendation in the Guideline includes a statement about the basis for the Panel's decision.

Panel Recommendations and Levels of Evidence

Although the Panel chose the target patient as a child age 1 through 3 years, the literature review showed that some of the recommendations could apply to older children. All the recommendations presented here are limited, however, to children with no craniofacial or neurologic abnormalities or sensory deficits (except as specifically noted in the recommendations) and who are healthy except for otitis media with effusion.

Individual recommendations were graded by the Panel as follows:

• Recommendations were made for interventions that the Panel took a stand for or against:

  • Strong recommendations were based on high-quality scientific evidence or, in the absence of high-quality scientific evidence, on strong Panel consensus.

  • Moderate recommendations were based on good-quality scientific evidence or, in the absence of good-quality scientific evidence, on expert opinion of the Panel.

  • Recommendations not otherwise specified were based on limited scientific evidence or, in the absence of scientific evidence, on expert opinion of the Panel.

• Clinical options were identified as interventions that the Panel failed to find compelling evidence for or against. Clinical options are interventions that a reasonable health care provider might or might not wish to implement in his or her practice.

• No recommendation was made when scientific evidence was lacking and there was no compelling reason to make an expert judgment.

Diagnosis of Otitis Media with Effusion and Evaluation of Hearing

• Pneumatic Otoscopy -- Recommendation: The diagnostic evaluation of suspected otitis media with effusion should include pneumatic otoscopy. Otoscopy alone (without the use of the pneumatic otoscope to test tympanic membrane mobility) is not recommended. [Strong recommendation based on limited scientific evidence and strong Panel consensus.]

• Tympanometry -- Option: Tympanometry may be used as a confirmatory test for otitis media with effusion. [Option based on limited scientific evidence and expert opinion.]

• Acoustic Reflectometry -- No Recommendation: No recommendation is made regarding the use of acoustic reflectometry as a screening or diagnostic test for otitis media with effusion.

• Tuning Fork Tests -- No Recommendation: No recommendation is made regarding the use of tuning fork tests in screening for or diagnosis of otitis media with effusion, except to note that they are inappropriate in the youngest children.

• Hearing Evaluation -- Option: Hearing evaluation may be performed before otitis media with effusion has been present for a total of 3 months. [Option based on insufficient evidence.]

• Hearing Evaluation -- Recommendation: Hearing evaluation is recommended for a child who has had bilateral otitis media with effusion for a total of 3 months. [Recommendation based on limited scientific evidence and expert opinion.]

Management

The Panel decided upon options and made recommendations for management of the otherwise healthy child age 1 through 3 years with no craniofacial or neurologic abnormalities or sensory deficits (except as noted in the recommendations) who enters the management algorithm with otitis media with effusion. Management is discussed for (1) initial presentation, (2) after a total of 3 months of otitis media with effusion, and (3) when otitis media with effusion has continued for a total of 4 to 6 months. These guidelines may be modified to fit the needs of the individual patient.

Algorithm

The notes below are an integral part of the algorithm that follows.

[Algorithm, part 1]
[Algorithm, part 2]
[Algorithm, part 3]

Algorithm 1 Part 1. Algorithm for managing otitis media with effusion in an otherwise healthy child age 1 through 3 years

Algorithm 1 Part 2. Algorithm for managing otitis media with effusion in an otherwise healthy child age 1 through 3 years

Algorithm 1 Part 3. Algorithm for managing otitis media with effusion in an otherwise healthy child age 1 through 3 years

Short-Term Outcomes

Table 1. Outcomes of treating otitis media with effusion (more...)

Table 1. Outcomes of treating otitis media with effusion [1]

Intervention Observation	Benefits [2] Base case.	Harms [2] Base case
Antibiotics	Improved clearance of effusion at 1 month or less, 14.0% (95% CI [3.6%,24.2%]). Possible reduction in future infections.	Nausea, vomiting, diarrhea (2%-32% depending on dose and antibiotic). Cutaneous reactions (<5%). Numerous rare organ system effects, including very rare fatalities. Cost. Possible development of resistant strains of bacteria.
Antibiotics plus steroids	Possible improved clearance at 1 month, 25.1% (95% CI [-1.3%,49.9%]).[3] Possible reduction in future infections.	See antibiotics and steroids separately.
Steroids alone	Possible improved clearance at 1 month, 4.5% (95% CI [-11.7%, 20.6%]). [3]	Possible exacerbation of varicella. Long-term complications not established for low doses. Cost.
Antihistamine/ decongestant	Same as base case.	Drowsiness and/or excitability.[4] Cost.
Intervention Myringotomy with tubes	Immediate clearance of effusion in all children. Improved hearing.	Invasive procedure. Anesthesia risk. Cost. Tympanosclerosis. Otorrhea. Possible restrictions on swimming.
Adenoidectomy	Benefits for young children have not been established.	Invasive procedure. [4] Anesthesia risk. Cost.
Tonsillectomy	Same as base case.	Invasive procedure. [4] Anesthesia risk. Cost.

[1]

The target patient is an otherwise healthy child age 1 through 3 years with no craniofacial or neurologic abnormalities or sensory deficits.

[2]

Outcomes are reported as differences from observation, which is treated as the base case. When possible, meta-analysis was performed to provide a mean and associated confidence interval (CI).

[3]

Difference from base case not statistically significant.

[4]

Risks were not examined in detail because no benefits were identified.

Long-Term Outcomes

The association between congenital or early-onset hearing impairment, particularly sensorineural hearing loss, and impaired speech and language development is well established (Ventry, 1980; Nober, 1967; Matkin, 1968): the earlier the onset and the more severe the hearing impairment, the greater the impairment in development of speech and language. Permanent sensorineural or conductive hearing loss leads to pronounced abnormalities in speech and language acquisition. In contrast, otitis media with effusion can produce mild to moderate conductive hearing impairment, which can fluctuate, remain stable, or alternate with periods of normal hearing. It is not known how many days of effusion and reduced hearing are required before development is adversely affected. Nevertheless, because permanent hearing loss can markedly affect communication, clinicians have hypothesized that the hearing impairment associated with persistent or recurrent episodes of otitis media with effusion at an early age may produce significant speech, language, cognitive, and educational delays and have advocated early, aggressive treatment of otitis media with effusion in young children. Thus, the Panel addressed this issue with particular interest, and rigorously reviewed and analyzed the available literature on hearing-related developmental outcomes of treated and untreated otitis media with effusion.

A search of the literature identified 101 published articles: 71 contained original data and 30 were review, anecdotal, or theoretical articles. None reported randomized, controlled clinical studies. The 71 articles with original data were examined in further detail: all but 35 were excluded from further consideration because: (1) they did not present sufficient developmental data; (2) otitis media with effusion was not documented before age 4; (3) researchers relied entirely on parental report for otitis media with effusion history; (4) they were retrospective studies that focused on children with speech disorders, language disorders, and/or learning disabilities; or (5) children had a biologic condition (such as cleft palate) that increased their risk of otitis media with effusion.

Table 2. Evidence on long-term outcomes of otitis media (more...)

Table 2. Evidence on long-term outcomes of otitis media with effusion in younger children

Developmental domain		Reference	No. of subjects	No. of controls	Design [1]	Test [2]	Age at study assessment (years)	Effects of otitis media with effusion [3]
Speech	Receptive	Gravel and Wallace, 1992	10	13	P,Co	PSI	4	S
	Expressive	Roberts, Burchinal, Koch, et al., 1988	55	0	P,Co	G-F-W	2.5-8	NS
		Teele, Klein, Chase, et al., 1990	194	0	P,Co	G-F-W	7	S
Language	Receptive	Friel-Patti, Finitzo-Hieber, Conti, et al., 1982	35	0	P,Co	REEL, SICD-R	1-2	S
		Friel-Patti and Finitzo, 1990	483	0	P,Co	SICD-R	1-2	NS 18,24m; S 12m
		Gravel and Wallace, 1992	10	13	P,Co	SICD-R	4	NS
		Lous, Fiellau-Nikolajsen, and Jeppesen, 1988	26	26	A,Ca	PPVT	8	NS
		Rach, Zielhuis, and van den Broek, 1988	52	13	P,Co	Reynell	2-4	NS
		Rach, Zielhuis, van Baarle, et al., 1991	84	18	P,Ca	Reynell	4	NS
		Roberts, Burchinal, Davis, et al., 1991	63	0	P,Co	M-Y, BLST, PPVT, CELF	4-6	NS
		Teele, Klein, Rosner, et al., 1984	205	0	P,Co	PPVT ACQ VAQ	3	S S NS
		Teele, Klein, Chase, et al., 1990	194	0	P,Co	PPVT, WUG	7	NS
		Wright, Sell, McConnell, et al., 1988	156	0	P,Co	PPVT, PLS, REEL, Boone	2-4	NS
	Expressive	Friel-Patti, Finitzo-Hieber, Conti, et al., 1982	35	0	P,Co	REEL SICD-E	1-2	S
		Feagans, Sanyal, Henderson, et al., 1987	44	0	P,Co	Paraphrase MLU	5,7	S NS
		Friel-Patti and Finitzo, 1990	483	0	P,Co	SICD-E	1-2	NS 12,24m; S 18m
		Rach, Zielhuis, and van den Broek, 1988	52	13	P,Co	Reynell	2-4	S
		Rach, Zielhuis, van Baarle, et al., 1991	84	18	P,Ca	Reynell	2-4	S
		Roberts, Burchinal, Davis, et al., 1991	63	0	P,Co	CELF-E, MLU	4-6	NS
		Teele, Klein, Chase, et al., 1990	194	0	P,Co	Language Sample WUG, Boston	7	NS
		Wright, Sell, McConnell, et al., 1988	156	0	P,Co	PLS, REEL, Boone	2-4	NS
Other	Intelligence	Roberts, Sanyal, Burchinal, et al., 1986	61	0	P,Co	WPPSI	5,7	S
		Roberts, Burchinal, Collier, et al., 1989	44	0	P,Co	WISC-R	8	NS
	Behavior	Roberts, Burchinal, Collier, et al., 1989	44	0	P,Co	CBI	8	S
		Feagans, Sanyal, Henderson, et al., 1987	44	0	P,Co	SCAN	5,7	S

[1] Abbreviations used for design: A = ambispective study (data collected by history as well as after enrollment in study). Ca = case control (data from children with the problem paired for comparison with data from children without variable). Co = cohort (group of children studied at different points in time). P = prospective study (data collected after enrollment in study).

[2] Abbreviations used for tests: ACQ = Auditory Comprehension Quotient (Zimmerman Preschool Language Scale). BLST = Bankson Language Screening Test. Boone = Boone Infant Speech and Language Development. Boston = Boston Naming Test. CBI = Classroom Behavior Inventory. CELF = Clinical Evaluation of Language Functions. G-F-W = Goldman-Fristoe-Woodcock Test of Articulation. MLU = Mean Length of Utterance. M-Y = Miller-Yoder Language Comprehension Test. PLS=Zimmerman Preschool Language Scale. PPVT=Peabody Picture Vocabulary Text. PSI=Pediatric Sentence Intelligibility Test. REEL=Receptive-Expressive Emergent Language Scale. Reynell=Reynell Developmental Language Scales. SCAN=Behavioral observation of task oriented, affective, and social actions. SICD=Sequenced Inventory of Communication Development. VAQ=Verbal Ability Quotient (Zimmerman Preschool Language Scale). WISC-R=Weschler Intelligence Scales for Children (Revised Versions). WPPSI=Weschler Preschool on a Primary Scale of Intelligence. WUG="WUG" Test.

[3] Abbreviations used for effect: NS=not statistically significant. S=significant (p <= 0.05).

Otoscopy and Pneumatic Otoscopy

The history and physical examination provide vital clues to the presence of otitis media with effusion. However, confirming the diagnosis requires establishing the presence of a middle ear effusion. The principles and usefulness of pneumatic otoscopy for detecting effusion were first reported by Siegle more than a century ago and popularized by Politzer (1909). Pneumatic otoscopy is a two-step procedure. In the first step, the clinician inspects the ear canal and eardrum visually, using a light source and magnification (otoscopy). In the second step, the clinician creates a seal in the ear canal and applies very slight positive and then negative pressure while watching the tympanic membrane to evaluate its mobility (pneumatic otoscopy).

On otoscopy, the clinician assesses tympanic membrane color, translucency, and resting position (retracted, neutral). The normal tympanic membrane is translucent, with a ground-glass appearance. Usually pearly grey, even a healthy tympanic membrane turns red when the patient cries. Although otoscopes are ubiquitous in clinical settings, few data are available correlating otoscopic findings with the presence of middle ear effusion, and few studies have provided any information about the criteria that suggest effusion. In one study, Finnish researchers sought to correlate tympanic membrane appearance with effusion in children with acute otitis media and otitis media with effusion; the predictive values of visible eardrum characteristics ranged widely (Karma, Penttila, Sipila, et al., 1989).

The pneumatic otoscope, which allows the examiner to observe movement of the tympanic membrane directly, has been widely credited as an advance over otoscopy alone, because pneumatic otoscopy can suggest the presence of effusion even when the appearance of the eardrum gives no indication of middle ear pathology. For this test, crisp movement of the tympanic membrane with slight application of pressure is normal. Thickening of the eardrum causes it to be less mobile. If the tympanic membrane does not move perceptibly with applications of slight positive or negative pressure, a middle ear effusion is highly likely. The examiner should note, however, that almost any eardrum will move if enough pressure is applied. Sometimes application of pressure will reveal an air-fluid level behind the tympanic membrane, and this is diagnostic of a middle ear effusion (Bluestone and Klein, 1990).

When employed under rigorously controlled conditions (i.e., when results of the procedure performed by a clinician who was validated using a prescribed protocol are compared with results of myringotomy performed by another clinician immediately after the procedure), pneumatic otoscopy was found in one study to have a sensitivity of 85 to 90 percent and specificity of 70 to 79 percent (Kaleida and Stool, 1992). In this study, the clinical indications for myringotomy were not clearly described; however, the prevalence of middle ear effusion in the study population must be assumed to be high.

In another study comparing findings at pneumatic otoscopy and at myringotomy, positive and negative predictive values of pneumatic otoscopy for middle ear effusion were found to be 91 percent and 84 percent, respectively (Toner and Mains, 1990). Because, as just discussed, the expectation of middle ear effusion must be very high in a study population to justify myringotomy in every child, the prevalence of otitis media with effusion was probably higher than would be found in a typical clinical setting. Thus, in a typical clinical setting, the positive predictive value of pneumatic otoscopy would be lower and the negative predictive value higher than in the research setting. Another factor that would tend to decrease the positive predictive value and increase the negative predictive value of this test in clinical settings is the lack of calibration of test results against findings at myringotomy.

Tympanometry

Many studies have been conducted to identify correlations between tympanometric findings and the presence of middle ear effusion at myringotomy. Problems have arisen with application of findings to clinical practice, however, because of variations in the way tympanograms have been categorized and the absence of data that would allow calculation of the true sensitivity, specificity, and negative predictive values in typical practice populations.

Early research into the association of tympanometric curves with conditions in the middle ear classified tympanometry results into four or more categories (Jerger, 1970; Bluestone, Beery, and Paradise, 1973; Paradise, Smith, and Bluestone, 1976; Shurin, Pelton, and Finkelstein, 1977; Cantekin, Beery, and Bluestone, 1977). Later research classified tympanogram curves into three categories: Type A (peaked, normal middle ear pressure), Type B (flat), and Type C (peaked, negative middle ear pressure, sometimes further divided into two types according to degree of negative pressure) (Jerger, Anthony, Jerger, et al., 1974; Tos, Poulsen, and Borch, 1978; Fiellau-Nikolajsen, 1980 a, b).

The most recent research has further collapsed the three types into two clinically useful categories: normal (Types A, C, and D, peaked) and abnormal (Type B, flat) (Toner and Mains, 1990; Dempster and MacKenzie, 1991; Babonis, Weir, and Kelly, 1991). In the following discussion, the simpler two-category classification is used. Tables in the original reports that used more complex classification schemes were collapsed into the simpler system to fit diagnostic information from the early studies into a format consistent with current practice.

The positive predictive value for middle ear effusion of a flat (Type B) tympanogram (i.e., the likelihood that an effusion is present if the tympanogram is abnormal) has been found to be between 49 and 99 percent (Fiellau-Nikolajsen, 1980 a, b; Orchik, Morff, and Dunn, 1980; Ben-David, Podoshin, and Fradis, 1981; Babonis, Weir, and Kelly, 1991). In a few studies, tympanometry was performed in children who were scheduled for myringotomy indicated for reasons other than tympanometric findings. A negative predictive value for tympanometry can be estimated from the few patients in these studies who had normal tympanogram results and who underwent myringotomy. In this highly selected population, the negative predictive value of a normal tympanogram (i.e., the likelihood that no effusion is present when the tympanogram is normal) was estimated to be between 64 and 93 percent (Fiellau-Nikolajsen, 1980 a, b; Orchik, Morff, and Dunn, 1980; Ben-David, Podoshin, and Fradis, 1981).

Since tympanometry became clinically available, clinicians have sought to use this diagnostic tool as a screening test for otitis media with effusion (Paradise, Smith, and Bluestone, 1976). Indeed, it has been used as a screening test (without myringotomy for confirmation) by many researchers in the United States and Europe. Longitudinal cohort studies of children followed with tympanometry have provided some of the best data available regarding the natural history of otitis media with effusion. In particular, such studies have shown that otitis media with effusion resolves without treatment in most cases (Tos, Poulsen, and Borch, 1978; Fiellau-Nikolajsen, 1980 a, b; Roland, Finitzo, Friel-Patti, et al., 1989).

Although several articles reviewed by the Panel mention in passing the possibility of using tympanometry as a surrogate for hearing evaluation, the Panel did not find direct evidence in support of this strategy. The clinical utility of tympanometry is limited to measuring the mobility of the tympanic membrane and ossicles and providing an estimate of middle ear pressure. As such, it is neither an appropriate nor a reliable predictor of hearing impairment, except in the most nonspecific sense as a risk marker for the hearing impairment associated with middle ear pathology. Indeed, the positive predictive value for hearing loss of an abnormal tympanogram has been reported to be as low as 49 percent (Dempster and MacKenzie, 1991). Clearly, with a positive predictive value for hearing loss of only 49 percent, tympanometry cannot be used in place of hearing evaluation- half the patients with abnormal tympanometric results could have normal hearing, and theoretically, half with normal results could have abnormal hearing. The negative predictive value, however, is 98 percent (Dempster and MacKenzie, 1991), reflecting the lower prevalence of sensorineural hearing impairment. That is, few children with normal tympanograms have hearing impairment, although those with pure sensorineural pathology and no middle ear involvement do have normal tympanograms. The American Academy of Pediatrics Committee on School Health (1987) specifically recommends against use of tympanometry as the primary screening test for hearing impairment.

Combination of Tympanometry and Pneumatic Otoscopy

Although pneumatic otoscopy and tympanometry essentially measure the same thing -- tympanic membrane mobility -- the strengths of one test can offset the weaknesses of the other. For example, tympanometry gives a quantitative measure of tympanic membrane mobility, whereas pneumatic otoscopy gives only a qualitative measure of this variable. In addition, a false-positive tympanometry result (abnormal tympanogram in the absence of effusion) can be caused by impacted cerumen, a foreign body, tympanic membrane perforation, medial canal stenosis, or improper placement of the instrument tip on the canal wall; performing pneumatic otoscopy before or after tympanometry provides information about ear anatomy and tympanic membrane mobility critical to interpretation of tympanometry results. Accordingly, the Panel recommends pneumatic otoscopy as the primary diagnostic test. Tympanometry may be used as a confirmatory test.

Because of the complementary nature of these two tests, researchers have constructed algorithms for the diagnosis of otitis media with effusion based on combined tympanometric and pneumatic otoscopic findings. Algorithms in widest use include those proposed by Cantekin (1983) and updated by Brostoff and Cantekin (1988). The Panel did not find direct evidence linking the outcome of these algorithms to the presence or absence of middle ear effusion. Le, Daly, Lindgren, et al. (1992) proposed an algorithm in which results of diagnostic testing are validated against findings at myringotomy. None of the research algorithms combining pneumatic otoscopy and tympanometry has been tested for usefulness in clinical practice.

Acoustic Reflectometry

Acoustic reflectometry is a relatively new technique in which a tone sweep is presented to the patient's ear canal and the reflected sound pressure is recorded in decibels. The initial report of this technique stated that in one population of children, in most of whom middle ear disease had been diagnosed by some other method, acoustic reflectometry had high (86 percent) sensitivity and high (76 percent) specificity for otitis media with effusion when correlated with findings at myringotomy (Teele and Teele, 1984). Even higher sensitivity (98.7 percent) and specificity (94.5 percent) were obtained when reflectivity and angle were optimized and tympanometry was used as the reference standard (Combs, 1991). Other studies have not yielded consistent results, with sensitivity of this test ranging from 80 to 87 percent and specificity ranging from only 54 to 70 percent (Lampe, Weir, Spier, et al., 1985; Avery, Gates, and Prihoda, 1986; Babonis, Weir, and Kelly, 1991). One of these studies examined the receiver-operator characteristic of acoustic reflectometry and found no acceptable breakpoint dividing a high probability of effusion from a low probability of effusion; this study did, however, suggest that the absolute value can be less important than documenting a change in a single individual (Avery, Gates, and Prihoda, 1986).

The Panel judged the evidence in support of acoustic reflectometry to be insufficient to make a recommendation regarding use of this test in screening for or diagnosis of otitis media with effusion.

Tuning Fork Tests

The Panel recognized that many experts use tuning fork tests to evaluate hearing and as an indirect means to diagnose otitis media with effusion by assessing conductive hearing loss. However, the Panel found no reports of studies that used other tests as standards against which to judge the sensitivity, specificity, and predictive value of tuning fork tests to identify otitis media with effusion in children. Furthermore, tuning fork tests are not reliable or even appropriate for use in the youngest children covered by this Guideline. A number of methods to test directly for the presence of a middle ear effusion, including pneumatic otoscopy and tympanometry, have been discussed, and better methods are available to assess hearing (see next section).

Hearing Evaluation

A change in hearing threshold is both a clinical outcome and a possible indicator of the presence of otitis media with effusion. Tests that may be used to evaluate hearing (listed in approximate order from those requiring greatest language and communication skills from the patient to those requiring the least skills) include: (1) pure tone threshold audiometry measuring both air and bone conduction, (2) speech reception threshold audiometry, (3) speech awareness threshold audiometry, (4) behavioral observation audiometry, and (5) auditory brainstem response recording and evaluation.

Children who fail hearing screening tests are usually referred to a physician, but the Panel found no direct evidence to estimate the proportion of these children who have otitis media with effusion. The corollary linkage between otitis media with effusion and hearing is better understood: several studies have documented hearing impairment in children with otitis media with effusion diagnosed by tympanometry and pneumatic otoscopy (Fria, Cantekin, and Eichler 1985; Friel-Patti and Finitzo, 1990; Dempster and MacKenzie, 1991). Otitis media with effusion is not always associated with hearing impairment, however. For example, one well-designed study of language learning and hearing in very young children with otitis media with effusion found a correlation of only 0.44 between days with effusion and average hearing over a 12-month period (Friel-Patti and Finitzo, 1990). Days of effusion accounted for less than 20 percent of the observed variation in hearing in this study.

The Panel recognizes that none of the hearing tests found to be effective (with the exception of pure-tone threshold audiometry, which is appropriate only for older children) is routinely available in many primary care settings, especially in rural areas. Furthermore, the Panel recognizes that there are difficulties with the validity, availability, and cost of hearing testing in the youngest children. Nevertheless, for otherwise healthy children age 1 through 3 years, the Panel recommends hearing evaluation after otitis media with effusion has been present bilaterally for 3 months, because of the strong belief that the placement of tympanostomy tubes is not indicated when otitis media with effusion is unaccompanied by bilateral hearing impairment (defined as 20 decibels hearing threshold level or worse in both ears).

Methods used to determine the child's hearing acuity will vary depending on resources available in the community and the child's willingness and ability to participate in testing. Optimally, air- and bone-conduction thresholds can be established for 500, 1,000, 2,000, and 4,000 Hz, and an air-conduction pure tone average can be calculated. This result can then be verified by obtaining a measure of speech sensitivity. Determinations of speech reception threshold and speech awareness threshold are less precise but still acceptable measures of hearing acuity. In the event that facilities or resources are not readily available or the child is unable to cooperate in such testing, the health care provider should use his/her best judgment as to the adequacy of the child's hearing. In these cases, the health care provider should take into consideration whether the child is achieving the appropriate developmental milestones for communication skills.

Infant Feeding Practices

The Panel reviewed literature linking early infant feeding practices to later otitis media, even though the target patient for Guideline development was past the age when a change in infant feeding practices would have much effect in preventing middle ear effusion.

Some studies of breast-feeding versus bottle-feeding showed a several-fold increase in otitis media in bottle-fed compared to breast-fed infants (Saarinen, 1982; Cunningham, 1979; Chandra, 1979; Duncan, Ey, Holberg, et al., 1993; Teele, Klein, and Rosner, 1989; Alho, Koivu, Hartikainen-Sorri, et al., 1990), suggesting that breast-feeding may provide some protection against otitis media. In addition, Eskimo and Inuit studies point to a link between bottle-feeding and an increased incidence of chronic otitis media (Timmermans and Gerson, 1980).

The available literature on infant feeding practices and otitis media is descriptive and focuses on acute otitis media, not on otitis media with effusion. Thus, although studies suggest that bottle-fed infants are two or three times more likely to have episodes of acute otitis media during the first year of life, the proportion who develop otitis media with effusion as a consequence is unknown. Furthermore, the direction of the linkage cannot be determined from available research. For example, perhaps infants with acute otitis media are weaned to the bottle earlier, giving the appearance that bottle-feeding is correlated with acute otitis media when in fact the acute otitis media "caused" bottle-feeding. Finally, the linkage is not firm: several studies have failed to find an association between feeding practices and otitis media (Tainio, Savilahti, Salmenpera, et al., 1988; Paine and Coble, 1982).

Some experts believe that propping a feeding bottle is a practice that may promote the occurrence of otitis media with effusion. The hypothesis is that when a bottle is propped in place, fluid is forced under pressure into the oral cavity, leading to reflux into the middle ear (Bluestone and Klein, 1988). The Panel could locate no research related to bottle propping and otitis media with effusion.

In summary, available research linking infant feeding practices to otitis media with effusion is provocative, but inconclusive. If a causal link is eventually demonstrated, better parent education about infant feeding practices might lead to fewer episodes of otitis media with effusion in later years. By the time a child reaches the age considered in this Guideline, however, the issue is only of historical interest.

Passive Smoking

Evidence that passive smoking may be related to otitis media with effusion has been reported only during the past decade and is limited. Kraemer, Richardson, Weiss, et al. (1983) documented a relative risk of 2.8 for children in households with two or more smokers in an early case-control study. Two other independent early studies documented odds ratios of 1.6 for parental smoking and abnormal tympanometry results or for the risk of surgery for glue ear (Iverson, Birch, Lundqvist, et al., 1985; Black, 1985). In a case-control study, 57 percent of children with otitis media with effusion (cases) had at least one parent who smoked, compared with 39 percent of children without otitis media with effusion (controls) (Hinton, 1989). A more recent case-control study of similar design, however, failed to find the same association (Rowe-Jones and Brockbank, 1992). In a followup study of children with severe chronic otitis media with effusion (glue ear), parental smoking delayed the time to resolution regardless of treatment chosen (Maw and Bawden, 1993).

Two studies found a correlation between higher levels of serum cotinine (a marker for nicotine) and otitis media with effusion. In a randomly chosen sample of Edinburgh primary school children, a direct correlation was found between serum cotinine levels and abnormal tympanograms (Strachan, Jarvis, and Feyerabend, 1989). Another study found a 38 percent higher rate of new episodes of otitis media during the first 3 years of life in children with high levels compared with children with low or undetectable levels of serum cotinine (Etzel, Pattishall, Haley, et al., 1992); the authors estimated that approximately 8 percent of cases of otitis media with effusion and 18 percent of the days with otitis media with effusion may be attributable to passive smoking.

Overall, studies have shown a moderate association between passive smoking and otitis media with effusion. Although proof is lacking that cessation of passive smoking will help prevent otitis media with effusion, the association of passive smoking with otitis media with effusion is consistent across a variety of study designs and is biologically plausible. Thus, the possibility that freedom from passive smoking will decrease a child's risk for and time with otitis media with effusion should be added to the many other reasons for parents and child-care providers to stop smoking.

Child Care

Descriptive research conducted during the past 15 years consistently shows a relationship between otitis media with effusion and child care in group facilities (child-care facilities). This relationship has been observed in the context of research demonstrating an association between many common childhood infections and child-care facility attendance.

One of the earliest population-based studies found a three- to fourfold increase in the prevalence of secretory otitis media in children attending child-care facilities, regardless of season of year (Fiellau-Nikolajsen, 1979). An early Danish study that explored the relationship between otitis media with effusion and a number of factors in the child's history found abnormal tympanograms in 76 percent of children cared for in public child-care facilities versus 69 percent of children cared for at home, a difference that was not statistically significant and not adjusted for other factors (Sorenson, Holm-Jensen, and Tos, 1981). A British case-control study found a relative risk of between 1.47 and 2.00, depending on the control group selected, for otitis media with effusion in young children, adjusted for household smoking, parental working, family history, residence, and type of heating system (Black, 1985). A cohort study in Denmark of children cared for at home or in child-care facilities found that the common cold was followed by secretory otitis media in 56 percent of 1-year-old children cared for at home versus 83 percent cared for in child-care facilities, but that the rate was 20 percent in both groups by the time children attained the age of 5 years (Birch and Elbrond, 1987).

A study of 386 children in Minnesota with recurrent acute otitis media and otitis media with effusion at entry found an odds ratio of 1.2 for child-care facility placement and otitis media with effusion after adjustment for age, age at first episode of otitis, season of birth and enrollment, and antibiotic prophylaxis (Daly, Giebink, Le, et al., 1988). Two population-based European studies found odds ratios of 1.7 and 1.8 for otitis media with effusion in children attending child-care facilities, but variably adjusted for confounding factors (Zielhuis, Heuvelmans-Heinen, Rach, et al., 1989; Alho, Koivu, Hartikainen-Sorri, et al., 1990). A similarly designed Swedish cohort study of 1,306 children found a slightly higher relative risk of 2.6 (Rasmussen, 1993).

All of the studies surveyed suffered from one or more design flaws. For example, many studies were retrospective, and some of the cohort studies were conducted on children already known to be at high risk of otitis media with effusion. More importantly, the cause of otitis media with effusion is multifactorial, and none of the studies adequately adjusted for all known confounding factors.

Despite their shortcomings, these studies point consistently to an association between child-care facility attendance and otitis media with effusion. The relative risks are modest (less than 2.0), but potentially very important given the size of the population base. Furthermore, data recently reported from a cohort study of 2,512 newborns followed for 2 years suggest that removing a child from the group child-care environment might lead to a lower incidence of acute otitis media (Alho, Kilkku, Oja, et al., 1993). More research is needed, however, to establish whether removing the child from a group child-care facility prevents otitis media with effusion, or speeds its resolution.

Antibiotic Therapy

Efficacy of Antibiotics

Antibiotic medications have been considered for treatment of otitis media with effusion because studies have shown that in 27 to 50 percent of cases, middle ear aspirates from children with otitis media with effusion contain bacteria or provide a medium for bacterial pathogens to grow (Liu, Lang, Lim, et al., 1976; Kokko and Tauno, 1976; Sundberg, 1984; Teele, Healy, and Tally, 1980). However, the existence of bacteria in middle ear aspirates is not proof that the pathogens are causing the effusion or that antibiotic therapy will help resolve the effusion. To explore the role that antibiotic therapy might play in managing otitis media with effusion, the Panel examined the literature on the effectiveness of this therapy in clinical situations.

Table 4. Studies considered for inclusion in Guideline (more...)

Table 4. Studies considered for inclusion in Guideline meta-analysis of antibiotics for otitis media with effusion or included in other meta-analyses

Authors	Treatment groups	No. of patients per group	Outcome measure and time of measurement	No. of patients cured per group
Cantekin, McGuire, and Griffith, 1991 (analysis of data reported by Mandel, Rockette, Bluestone, et al., 1987)	Amoxicillin 40 mg/kg/day + decongestant/antihistamine /Amoxicillin alone /Placebo	155/155/150	Otoscopy 4 weeks	52 / 49 / 25
	Same as above	155 / 154 / 150	Tympanometry 4 weeks	26 / 25 / 13
	Same as above	140 / 134 / 131	Audiometry 4 weeks	38 / 46 / 43
Chan, Mandel, Rockette, et al., 1988	Augmentin / Amoxicillin 40 mg/kg/day Amoxicillin each group	56 / 50	Algorithm 10 days	29 / 16
	Same as above	52 / 45	Algorithm 4 weeks	26 / 23
Corwin, Weiner, and Daniels, 1986	Erythromycin 50 mg/kg/day + Sulfisoxazole 150 mg/kg/day 10 days / No treatment	66 / 65	Otoscopy with tympanometric confirmation 1 month	33 / 22
Daly, Giebink, Batalden, et al., 1991	Trimethoprim 8 mg/kg/day + Sulfamethoxazole 40 mg/kg/day / Placebo	21 / 21	Pneumatic otoscopy and tympanometry 2 weeks	5 / 2
deCastro, Jaeger, Martin, et al., 1982	Sulfisoxazole 50 mg/kg/day / Placebo	15 / 15	Otoscopy 1 month	11 / 2
	Same as above	15 / 15	Audiometry 1 month	10 / 1
	Same as above	15 / 15	Tympanometry 1 month	5 / 0
Ernstson and Anari, 1985	Cefaclor 20 mg/kg/day 10 days / No treatment	46 / 45	Otoscopy and tympanometry (A and C1) and normal hearing = cured 2-5 weeks	24 / 5
Giebink, Batalden, Le, et al., 1990	Trimethoprim-Sulfamethoxazole 40 mg/kg/day / No treatment	20 / 19	Algorithm (sensitivity 0.87, specificity 0.74) 4 weeks	10 / 6
Healy, 1984	Trimethoprim 8 mg/kg/day + Sulfamethoxazole 40 mg/kg/day / No treatment	96 / 93	Positive results on otoscopy and type B or C tympanogram = effusion present 4 weeks	56 / 6
Mandel, Rockette, Bluestone, et al., 1987	Amoxicillin 40 mg/kg/day + decongestant/antihistamine / Amoxicillin alone / Placebo	158 / 160 / 156	Algorithm 4 weeks	50 / 46 / 22
Mandel, Rockette, Paradise, et al., 1991	Erythromycin 50 mg/kg/day + Sulfisoxazole 150 mg/kg/day / Cefaclor 40 mg/kg/day / Amoxicillin 40 mg/kg/day / Placebo	80 / 77 / 79 / 78	Algorithm 2 weeks	17 / 17 / 25 / 11
	Same as above	60 / 52 / 54 / 55	Algorithm 4 weeks	20 / 26 / 23 / 20
Marks, Mills, and Shaheen, 1981	Cotrimoxazole / Dimetapp	25 / 26	Flat tympanogram to peaked or 20-decibel audiometric improvement = cured 4-6 weeks	16 / 7
Moller and Dingser, 1990	Erythromycin 50 mg/kg/day 14 days / Placebo	69 / 72	"Air filled" and tympanogram 1 month	12 / 11
Podoshin, Fradis, Ben-David, et al., 1990	Amoxicillin 50 mg/kg/day / Placebo	49 / 37	Tympanometry 2 months	20 / 5
	Same as above	49 / 37	Audiometry 2 months	29 / 21
Schloss, Dempsey, Rishikof, et al., 1988	Erythromycin 50 mg/kg/day + Sulfisoxazole / Placebo Tympanocentesis on all patients	25/27	Pneumatic otoscopy, tympanometry, audiometry 2 weeks	6 / 8
Schwartz and Rodriguez, 1982	Trimethoprim-Sulfamethoxazole 4 mg/kg/day / Placebo	33 / 36	Pneumatic otoscopy and tympanometry 42 days	29 / 32
Sundberg, Eden, Ernstson, et al., 1981	Erythromycin 40-60 mg/kg/day / No treatment	47 / 72	Otoscopy with some tympanometry, myringotomy, and audiometry 2-5 weeks	21 / 11
Thomsen, Sederberg-Olsen, Balle, et al., 1989 [1]	Augmentin 5 mL-7.5 mL three times a day / Placebo	111 / 110	Tympanometry (A, C1) = cured 1 month	69 / 34
Varsano, Volovitz, and Mimouni, 1985	Sulfisoxazole 250-500 mg two times a day / Placebo	19 / 16	Pneumatic otoscopy, 5 weeks	10 / 8

[1] The data reported in this study were the subject of a letter to the editor (Cantekin EI. Antibiotics for secretory otitis media. Arch Otolaryngol Head Neck Surg 1990 May;16(5):626-8).

The literature search identified 18 reports of controlled studies (or reanalyses of controlled study data) and 2 published meta-analyses (Rosenfield and Post, 1992; Williams, Chalmers, Stange, et al., 1993) of antibiotic therapy for otitis media with effusion (Table 4). The two published meta-analyses differed in the studies they included and how the studies were rated, but both concluded that antibiotics have a small but significantly positive effect on resolution of otitis media with effusion. The Panel was also provided an unpublished meta-analysis by Cantekin and McGuire.

The Panel examined antibiotic therapy data particularly carefully because of a conflict in the literature concerning interpretation of data from a study conducted at the Otitis Media Research Center at the University of Pittsburgh. The initial published analysis of these data was challenged by one of the investigators, who published a different interpretation of the data (Cantekin, McGuire, and Griffith, 1991). Because the Panel included individuals who were affiliated with (by having trained or been employed at) the University of Pittsburgh, and because the University of Pittsburgh was the subcontractor for the Guideline literature review, the Panel took several steps to preserve its impartiality in evaluating evidence regarding antibiotic therapy for otitis media with effusion:

• Members of the Panel who were affiliated with the University of Pittsburgh recused themselves from discussions of antibiotic therapy for otitis media with effusion.

• A subcommittee of the Panel was formed to evaluate the conflicting published and unpublished meta-analyses and to provide written recommendations for management of these reports in development of the Guideline (the complete subcommittee report is contained in the Guideline Technical Report).

• All articles originating at the University of Pittsburgh were reviewed independently by an individual not affiliated with the University of Pittsburgh.

• An independent analyst reviewed studies of antibiotic therapy for otitis media with effusion and also reviewed the results of the Panel's analysis of this topic.

• Evaluations of material to be included in this section of the Guideline were performed by individuals not affiliated with the University of Pittsburgh.

Table 5. Results of eight meta-analyses of the difference (more...)

Table 5. Results of eight meta-analyses of the difference in resolution of otitis media with effusion treated by antibiotic agents and controls as a function of study design

Guideline meta-analysis number	Study types included in meta-analysis	Meta-analysis result [1]	95% Confidence interval
1	Blinded studies with outcomes measured at 10 to 14 days [2]	7.6%	-1% to 14%
2	Blinded studies with outcomes measured at 4 to 6 weeks [3]	18.5%	1.7% to 34.2%
3	All studies with outcomes measured at 10 to 14 days [4]	16.0%	3.7% to 28%
4	All studies with outcomes measured at 4 to 6 weeks [5]	23.4%	9.8% to 36.1%
5	Blinded studies using audiometric outcomes [6]	15.4%	-17.8% to 46.4%
6	Blinded studies using tympanometric outcomes [7]	30.8%	20.1% to 40.7%
7	Blinded studies using otoscopic/ tympanometric outcomes [8]	2.8%	-4.6% to 10.3%
8	All blinded studies [9]	14.0%	3.6% to 24.2%

[1]

Absolute difference in proportion of children experiencing resolution of otitis media with effusion for group treated with antibiotic versus control group.

[2]

Schwartz and Rodriguez, 1982; Daly, Giebink, Batalden, et al., 1991; Schloss, Dempsey, Rishikof, et al., 1988; Mandel, Rockette, Paradise, et al., 1991; Moller and Dingser, 1990.

[3]

Schwartz and Rodriguez, 1982; Thomsen, Sederberg-Olsen, Balle, et al., 1989; Marks, Mills, and Shaheen, 1981; Mandel, Rockette, Paradise, et al., 1991; deCastro, Jaeger, Martin, et al., 1982; Varsano, Volovitz, and Mimouni, 1985.

[4]

Schwartz and Rodriguez, 1982; Daly, Giebink, Batalden, et al., 1991; Ernstson and Anari, 1985; Giebink, Batalden, Le, et al., 1990; Schloss, Dempsey, Rishikof, et al., 1988; Mandel, Rockette, Paradise, et al., 1991; Moller and Dingser, 1990.

[5]

Schwartz and Rodriguez, 1982; Thomsen, Sederberg-Olsen, Balle, et al., 1989; Marks, Mills, Shaheen, et al., 1981; Corwin, Weiner, and Daniels, 1986; Giebink, Batalden, Le, et al., 1990; Healy, 1984; Mandel, Rockette, Paradise, et al., 1991; deCastro, Jaeger, Martin, et al., 1982; Varsano, Volovitz, and Mimouni, 1985.

[6]

Cantekin, McGuire, and Griffith, 1991; Podoshin, Fradis, Ben-David, et al., 1990; deCastro, Jaeger, Martin, et al., 1982.

[7]

Thomsen, Sederberg-Olsen, Balle, et al., 1989; Podoshin, Fradis, Ben-David, et al., 1990; Marks, Mills, and Shaheen, 1981; deCastro, Jaeger, Martin, et al., 1982.

[8]

Daly, Giebink, Batalden, et al., 1991; Schloss, Dempsey, Rishikof, et al., 1988; Mandel,Rockette, Paradise, et al., 1991; Moller and Dingser, 1990.

[9]

Schwartz and Rodriguez, 1982; Thomsen, Sederberg-Olsen, Balle, et al., 1989; Marks, Mills, and Shaheen, 1981; Mandel, Rockette, Paradise, et al., 1991; Daly, Giebink, Batalden, et al., 1991; Schloss, Dempsey, Rishikof, et al., 1988; Moller and Dingser, 1990; Podoshin, Fradis, Ben-David, et al., 1990; deCastro, Jaeger, Martin, et al., 1982; Varsano, Volovitz, and Mimouni, 1985.

The Panel elected to examine all 18 randomized studies and reanalyses of antibiotic therapy for otitis media with effusion, but to concentrate on the 10 studies in which investigators appeared to be effectively blinded to the therapy provided to patients (Table 5, meta-analysis number 8). Antibiotic medications evaluated in these 10 studies included amoxicillin with and without clavulanate potassium, cefaclor, erythromycin with and without sulfisoxazole, sulfisoxazole, and the combination of trimethoprim and sulfamethoxazole.

In addition to evaluating a variety of antibiotic agents, these 10 studies used a variety of methods to diagnose the condition and to assess the effectiveness of the medications in treating otitis media with effusion. Methods to diagnose otitis media with effusion in these studies included tympanometry, otoscopy, hearing evaluation, or a combination of these modalities. Investigators in these 10 studies measured outcomes at various times, from a minimum of 10 days to a maximum of 2 months after the start of antibiotic treatment of otitis media with effusion. These differences in study design made meta-analysis of the data challenging.

In developing this Guideline, the Panel performed eight different meta-analyses, each one using a different subset of the available data. The subsets were created based on (1) time until measurement of outcome, (2) type of test used to evaluate outcome, and (3) blinding. The difference was then calculated between the proportion in the antibiotic group and those in the control group that experienced clearance of otitis media with effusion. The results of the meta-analyses, shown in Table 5, are the absolute differences between rate of clearance of otitis media with effusion with antibiotics and rate of clearance without. That is, if otitis media with effusion cleared in 22 percent of cases with antibiotics and in 19 percent without antibiotics, Table 5 shows a 3 percent difference, with a confidence interval for this difference that depends on the sizes of study populations included in the analysis. A negative difference indicates that the rate of clearance of otitis media with effusion was greater in the group of patients not receiving antibiotic therapy.

Panel subgroup evaluation of reports of antibiotic therapy studies originating at the University of Pittsburgh resulted in exclusion of the data from the two reports (Mandel, Rockette, Bluestone, et al., 1987; Cantekin, McGuire, and Griffith, 1991) for determining the effectiveness of antibiotics on otitis media with effusion, due to difficulties with the studies' methodologies. Data from the second report (Cantekin, McGuire, and Griffith, 1991) were ultimately used in meta-analysis 5 (Table 5) because of a lack of data from other sources, but were excluded from the final results reported in meta-analysis 8.

Results of note in Table 5 are the finding of a much greater difference between the antibiotic therapy and control groups when studies used only tympanometry as an outcome measure for otitis media with effusion compared to use of a combination of otoscopy and tympanometry. This difference is unexplained. In turn, the studies using hearing test results as the outcome measure for otitis media with effusion show a much greater difference between antibiotic and control groups than do studies using a combination of otoscopy and tympanometry. A number of factors can account for this difference: (1) not all young children with otitis media with effusion have a clinically significant hearing loss, (2) hearing evaluation results have a greater degree of inaccuracy in young children than in older children and adults, and (3) the meta-analysis was performed on only a few studies.

The above results showing some improvement, albeit small, in clearance of otitis media with effusion with antibiotic treatment do not answer questions about optimal use of this therapy. For example, what is the advantage to the patient of early clearance of otitis media with effusion? Without knowing this advantage, how can parents and health care providers balance this benefit against the potential adverse effects of antibiotic therapy? What is the overall cost-effectiveness of antibiotic therapy, given that no information is available about the degree to which this therapy can decrease the need for more expensive interventions (hearing evaluation and surgery) after 3 to 6 months of otitis media with effusion? One study (Podoshin, Fradis, Ben-David, et al., 1990) that examined outcomes at 2 months showed a 26.5 percent improvement in clearance of otitis media with effusion when antibiotics were used, but longer term studies are needed to determine whether this effect is simply hastening clearance of effusion in children who would experience resolution without treatment, or promoting clearance in children who would not otherwise experience resolution of the effusion.

Adverse Effects of Antibiotics

Table 6. Adverse effects of antibiotic medications (more...)

Table 6. Adverse effects of antibiotic medications most often prescribed for otitis media in children

Antibiotic	Adverse reactions
Amoxicillin/ ampicillin	Diarrhea, usually mild and dose-related (20-30%). Cutaneous allergic reaction (3-5%). Rarely: hematologic, renal, hepatic effects.
Amoxicillin/ clavulanic acid	Diarrhea (9%). Nausea and vomiting (4%). Skin rashes/urticaria (3%). Rarely: hematologic, renal, hepatic effects (mostly in adults).
Cefaclor	Diarrhea (2.5%). Rarely: hematologic, central nervous system, renal, hepatic, dermatologic reactions; serum sickness (more common with cefaclor than with other antibiotics).
Cotrimoxazole	Skin rashes/urticaria (2%). Nausea, vomiting, diarrhea. Rarely: serious dermatologic reaction (less than 0.1%); hematologic, cardiovascular, central nervous system, endocrine, renal, hepatic, respiratory effects. Rare fatalities due to sulfonamides.
Erythromycin	Gastrointestinal effects (dose-related; 2-2.5% for ethylsuccinate or estolate salt administered to children). Subclinical elevations of liver function enzymes (5%). Rarely: serious adverse effects reported for most organ systems.
Sulfisoxazole	Gastrointestinal effects, uncommon (less than 1%). Rarely: serious hematologic (e.g., blood dyscrasias), dermatologic (e.g., Stevens-Johnson syndrome), neurologic, allergic reactions (less than 0.1%).

Source: Computerized Clinical Information System, 30 June 1993 ed. (MICROMEDEX, Inc., Denver, CO).

Adverse effects of antibiotic therapy range from common to rare and from nuisance to life-threatening. The quality of published evidence linking antibiotic therapy to adverse effects is generally of poor quality, dominated by anecdotes and case reports, mostly in older children and adults. Nevertheless, potentially serious adverse effects -- especially allergic reactions -- do occur with this therapy in young children. The most common adverse effects of antibiotic drug therapy are gastrointestinal, with diarrhea occurring in about 9 percent of children treated with 20-40 mg/kg/day of amoxicillin with clavulanic acid (Table 6). Dermatologic reactions can occur in 3 to 5 percent of cases; severe anaphylactic reactions occur much less frequently; severe hematologic, cardiovascular, central nervous system, endocrine, renal, hepatic, and respiratory adverse effects are more rare still (Table 6). Finally, there is concern that unnecessary use of antibiotics might lead to antimicrobial drug resistance and potentially more serious illness with later episodes of infection.

Steroid Therapy

The literature on use of adrenocorticosteroid medications to treat otitis media with effusion is growing rapidly. The Panel identified 10 controlled studies and one meta-analysis in which results of this therapy were analyzed. In most of these studies, steroid therapy for otitis media with effusion was of short duration (4 to 10 days), and systemic agents were used. The most frequently used steroid regimen was 1 mg/kg/day prednisone for 2 to 4 days followed by dosage tapering.

Antihistamine/Decongestant Therapy

Antihistamine and decongestant medications have been used separately or together for decades to treat otitis media with effusion. The Panel found, however, that the four randomized controlled studies of this therapy failed to show a statistically significant effect of these medications in resolving otitis media with effusion.

The first reported study of antihistamine/decongestant therapy for otitis media with effusion was a three-armed study of decongestant alone (phenylpropanolamine), decongestant with antihistamine (phenylpropanolamine with brompheniramine maleate), and placebo. The results showed weak negative trends for both active treatments against placebo, not reaching statistical significance (Haugeto, Schroeder, and Mair, 1981).

The second published report of antihistamine/decongestant therapy for otitis media with effusion was a double-blinded study of the combination chlorpheniramine maleate and pseudoephedrine hydrochloride in 553 children. The medication showed a very weak (p=0.74) negative tendency to resolve otitis media with effusion in children with unilateral effusion and a very weak (p=0.67) positive tendency to resolve effusion in those with bilateral effusion (Cantekin, Mandel, Bluestone, et al., 1983).

The third study in which antihistamine and decongestant agents were evaluated for otitis media with effusion was a double-blinded study in which antihistamine (chlorpheniramine) alone, decongestant (pseudoephedrine) alone, or placebo was administered. The decongestant showed a very weak positive trend and the antihistamine showed a very weak negative trend toward more effective resolution of otitis media with effusion compared to placebo or the other active agent. The results were not statistically significant, however (Dusdieker, Smith, Booth, et al., 1985).

The last trial the Panel examined in this category was a three-arm, double-blinded study of antibiotic (amoxicillin), antibiotic with antihistamine/decongestant (amoxicillin with chlorpheniramine maleate and pseudoephedrine hydrochloride), and placebo. In this study, addition of antihistamine/decongestant medications to antibiotic therapy showed a weak positive trend for improvement in otitis media with effusion, but the added effect of antihistamine/decongestant therapy was not statistically significant (Mandel, Rockette, Bluestone, et al., 1987).

Meta-analysis of the results reported in these four studies was performed for this Guideline by the hierarchical Bayes method. This meta-analysis showed that the antihistamine/decongestant combination had no effect on resolution of otitis media with effusion (mean, -0.009; 95 percent confidence interval [-0.036, 0.054]).

Antihistamine/decongestant medications can also have adverse effects, including insomnia, drowsiness, behavior changes, changes in blood pressure, and seizures.

Myringotomy with Tubes

Myringotomy with insertion of tympanostomy tubes is most often performed under general anesthesia in an ambulatory surgical center. Some physicians perform this procedure in an office setting with sedation and local anesthesia. Statistics from the National Hospital Discharge Survey for 1987 showed a 71 percent decline between 1977 and 1987 in the rate of myringotomy with or without tube insertions performed as an inpatient procedure, from 216,000 to 62,000 procedures (430 to 118 procedures per 100,000 population) (Derkay, 1993). Balancing this trend, however, is the increasing rate at which many procedures, including myringotomy with or without tube insertion, are performed in free-standing surgical centers. The reported rate for myringotomy with or without tubes performed in such a setting in 1986 was nearly 58,000 cases (Henderson, 1986).

Tympanostomy tubes are available in a myriad of designs, most constructed from plastic and/or metal. Variations in tube design reflect attempts to improve ease of tube insertion or removal and to prolong the period of tube retention in the tympanic membrane. Data comparing outcomes for tubes of various designs are sparse, however, and the Panel assumed that there were no notable differences between available tympanostomy tubes.

Adenoidectomy

Adenoidectomy has been advocated for the treatment of middle ear effusion as an isolated procedure or in combination with tonsillectomy, or with myringotomy and insertion of tympanostomy tubes. This procedure is most often chosen when nasal obstruction, nasal discharge, snoring, and mouth breathing are present. Maw (1983) noted that, historically, adenoidectomy has been recommended when the eustachian tubes are occluded by hypertrophic adenoidal tissue, compromising middle ear ventilation, or when the adenoids appear to be a focus of ascending eustachian tube infection. Absolute size of adenoid tissue, however, is not related to outcome of surgical management of otitis media with effusion, regardless of the type of surgical procedure (Gates, Avery, Cooper, et al., 1989).

Statistics from the National Hospital Discharge Survey show a decline in the frequency with which adenoidectomy is performed in children younger than age 15 years as an inpatient procedure, from 83,000 procedures in 1978 to 14,000 procedures in 1986 (Derkay, 1993). It is likely that the current frequency of adenoidectomy is somewhat higher than this, however, because of the trend for such procedures to be performed increasingly in outpatient facilities of hospitals or in free-standing ambulatory surgical centers: in 1986, one-fifth of tonsillectomy procedures (with or without adenoidectomy) were performed in such outpatient settings. Concurrent with the change in setting for surgery, indications for adenotonsillectomy changed dramatically: in 1978 obstructive sleep apnea was not listed at all as an indication for this surgery, but in 1987 this diagnosis was the indication for adenotonsillar surgery in one-third of patients of all ages.

Tonsillectomy

Evidence of the lack of efficacy of tonsillectomy for otitis media with effusion was inferred from a single randomized controlled study comparing adenotonsillectomy to adenoidectomy (Maw and Herod, 1986). Although tonsillectomy alone was not studied, the results from this study show that adding tonsillectomy to adenoidectomy (and, therefore, tonsillectomy alone) is not justified for the treatment of otitis media with effusion. Furthermore, for the older children in this study, adenotonsillectomy offered no significant advantage over simple adenoidectomy for the treatment of otitis media with effusion. The authors noted that the added morbidity of tonsillectomy with adenoidectomy does not at present justify recommendation for removal of the tonsils in addition to adenoidectomy in children with otitis media with effusion; rather, tonsillectomy in these children should be performed for tonsillar pathology per se and not with regard to the status of the middle ear.

Atopy as a Risk Factor

The Panel found six studies that examined atopy as a risk factor for otitis media with effusion. Because the results of these studies were conflicting and some patients were older than the Guideline target patient, no conclusion could be drawn from the data as a group.

No association between atopy and otitis media with effusion was found in a study reported by Black (1985).

In a Japanese study (Tomonaga, Kurono, and Mogi, 1988), researchers determined that allergic rhinitis was present in 50 percent of 259 patients (mean age 6 years) in whom otitis media with effusion had been diagnosed, and otitis media with effusion was present in 21 percent of 605 patients (mean age 9 years) in whom allergic rhinitis had been diagnosed. Among 108 children (ages 5 to 8 years, mean 6 years) in whom neither condition had been diagnosed, the incidences of allergic rhinitis, otitis media with effusion, and both of these conditions were 17 percent, 6 percent, and 2 percent, respectively.

Another study (Kraemer, Richardson, Weiss, et al., 1983) compared risk factors for persistent middle ear effusion among 76 children admitted for insertion of tympanostomy tubes and 76 controls matched by age, sex, and season of admission for a general surgical procedure. Results of this study showed a nearly fourfold increase in the risk of persistent middle ear effusion in children who had atopic symptoms for more than 15 days per month, although this statistic was uncorrected for a large number (more than 15) of other possible risk factors for otitis media with effusion. When the single risk factor of atopic symptoms for more than 1 day per month was examined, the relative risk for otitis media with effusion was only 0.5.

Bernstein, Lee, Conboy, et al. (1983) reported on a prospective study of the role of hypersensitivity mediated by immunoglobulin E in the occurrence of otitis media with effusion. Subjects for this study were 77 children age 2 to 18 years with recurrent otitis media with effusion who had undergone at least two myringotomies with insertion of tympanostomy tubes. The finding that 32 (42 percent) of the subjects had allergic rhinitis mediated by immunoglobulin E suggests that in patients with allergic rhinitis, otitis media with effusion can result from hypersensitivity mediated by immunoglobulin E. However, the middle ear might be the target organ in only 15 to 20 percent of patients with otitis media with effusion and allergic rhinitis. In this study, the middle ear was the target organ in only 5 (7 percent) of the 77 patients with recurrent otitis media with effusion. This finding, and the fact that more than half (58 percent) of children with recurrent otitis media with effusion in this study did not have allergic rhinitis and were clearly nonallergic, led the authors to conclude that the role of atopy in otitis media with effusion is still unresolved.

In another study, levels of immunoglobulin E were measured in middle ear effusions and serum samples from 58 children age 6 months to 11 years (Boedts, de Groote, and van Vuchelen, 1984). Eight (14 percent) of the 58 patients had markedly elevated serum levels of immunoglobulin E, indicating that an atopic disorder was very likely. Although the authors note that their results and those of other studies do not support allergy as a major factor in the etiology of otitis media with effusion, the results do not rule out the possibility that otitis media with effusion might occur as a complication of allergic rhinitis.

Finally, Bernstein and Reisman (1974) studied allergy in a group of 200 children who had undergone one or more myringotomies with insertion of tympanostomy tubes. They found that 24 percent (46) of the 200 had allergy but that the incidence of allergy was 35 percent among children (n=88) who had undergone more than one myringotomy with insertion of tubes. The authors concluded that the middle ear is not an allergic organ, but that middle ear disease can be secondary to allergy elsewhere in the nasopharynx. However, six patients with recurrent otitis media with effusion and known allergies did not improve with several years of aggressive allergy treatment.

Hyposensitization as a Treatment

Two of the seven articles on allergy reported on the role of hyposensitization in otitis media with effusion (Hurst, 1990; Tomonaga, Kurono, and Mogi, 1988). Both studies were of case-control design, so that the researchers neither randomized subjects nor "blinded" clinicians as to whether patients were receiving hyposensitization treatment.

In the study reported by Hurst (1990), 20 patients between age 3 and 11 years who had a history of otitis media with effusion were offered hyposensitization; those who refused (3 patients) served as controls. In this study, 65 percent of the patients who chose (and complied with) allergy treatment remained disease-free for more than 3 years. During episodes of noncompliance with the allergy treatment, otitis media with effusion returned.

In the Japanese study (Tomonaga, Kurono, and Mogi, 1988), 12 of 41 children with allergic rhinitis who had also had tympanostomy tubes inserted to manage associated otitis media with effusion underwent hyposensitization therapy. In these children, the mean time from extrusion of the tympanostomy tube to recurrence of otitis media with effusion was more than 11 months, compared to a mean time to recurrence in the untreated 29 children of less than 3 months.

Because patients in these studies were older than the target patient for the Guideline, and because the study designs and numbers of patients studied were inadequate, the Panel could not estimate the effectiveness of hyposensitization therapy for otitis media with effusion in children age 1 through 3 years. Furthermore, anaphylactic reactions to hyposensitization have been reported in as many as 6 percent to 11 percent of patients (Ostergaard, Kaad, and Kristensen, 1986; Mosbech, Dirksen, Dreborg, et al., 1990). For some patients, this significant risk of anaphylaxis might well outweigh the possible benefits of hyposensitization for the treatment of otitis media with effusion.

Although the seven reports on the topic of allergy and otitis media with effusion propose plausible theories concerning a connection between allergic rhinitis, eustachian tube physiology, and the development of otitis media, the data available do not fully support or elucidate the nature of the connection. Thus, the Panel could not make any statement regarding the impact of allergy or its treatment on the incidence or prevalence of otitis media with effusion in children.

Natural History

It is characteristic of health care providers in the United States to intervene for otitis media with effusion, but the Panel was impressed by data suggesting that otitis media with effusion usually follows a benign course without treatment. Thus, although such a study might be difficult to implement because it runs counter to prevailing attitudes, research to document the natural course of otitis media with effusion is essential to the identification of factors that separate children at risk for serious consequences of this condition -- and who therefore require treatment- from those who will do well if otitis media with effusion is not treated.

A key issue in research into the natural history of otitis media with effusion is the putative association of this condition with long-term adverse effects on speech and language development, learning, and behavior. There is general agreement that not treating a significant and/or long-lasting hearing impairment associated with otitis media with effusion could cause serious disability. However, little is known about the effects that fluctuating hearing loss can have on the long-term outcomes for otherwise normal children. Studies must be conducted to identify the combinations of patient characteristics and hearing loss that indicate risk of long-term problems. Factors that must be studied include:

• Child's age at onset of the hearing impairment.

• Duration of the hearing impairment.

• Severity of the hearing impairment.

• Child's social environment.

• Effect of fluctuation of hearing.

High-quality descriptive research studies are essential to definition of the proximal and long-term effects of hearing loss on speech, language, learning, and behavioral outcomes. Some of the linkages in the causal chain can be examined in children with hearing loss from causes other than otitis media with effusion.

Diagnosis

No data have been reported on the use of otoscopy and pneumatic otoscopy in typical practice settings, although these two tests are the foundation upon which diagnostic and treatment decisions are made by many, if not most, clinicians. Similarly, tympanometry is used by many clinicians, alone or in combination with otoscopy, to evaluate for otitis media, and yet the methods for and results of performing tympanometry in typical practice settings, especially in primary care, have not been documented. The Panel recommends that:

• Diagnostic criteria be clarified for use of otoscopy, pneumatic otoscopy, and tympanometry in typical practice settings, particularly primary care.

• A diagnostic algorithm combining tympanometry and pneumatic otoscopy -- similar to that used in research -- be developed for use in clinical practice settings.

• The correlation of hearing evaluation results with otitis media with effusion be examined in large study populations representative of those cared for in a typical clinical practice.

Control of Environmental Risk Factors

Available research suggests that environmental factors are associated with otitis media with effusion, but the data are descriptive and study results are open to criticism because of the research methods employed. Studies of impeccable design are needed on the associations between factors such as infant feeding practices, cigarette smoke, allergens, and child-care situations and the incidence, duration, complications, and other characteristics of otitis media and otitis media with effusion. Once associations have been identified, intervention studies should be conducted to determine the nature of the linkage(s) and whether controlling for the environmental factor(s) results in improved outcome for children with otitis media with effusion.

Intervention

Clinical studies must be conducted to determine whether treatment of any kind improves the outcome of otitis media with effusion in children. The studies must be designed:

• To involve enough children for the results to be statistically commanding.

• To involve patient populations similar to those seen in typical practice.

• To control adequately for multiple confounding variables so that the results are clinically meaningful.

The best studies to meet these needs would be multicenter. In addition, they must be long-term, because the most important clinical outcomes of otitis media with effusion in children (speech, language, learning, behavior) may not be known until years after the intervention. For this reason, any recommendations, such as those made by the Panel in this Guideline, must be considered provisional until the results of such high-quality, long-term research confirm or refute their value.

Panel Biosketches

• Sylvan E. Stool, MD, Co-Chair

• Professor of Otolaryngology and Pediatrics

• University of Pittsburgh School of Medicine

• Director of Education, Department of Otolaryngology

• Children's Hospital of Pittsburgh

• Alfred O. Berg, MD, MPH, Co-Chair

• Professor and Associate Chair, Department of Family Medicine

• University of Washington, Seattle

• Stephen Berman, MD

• Professor of Pediatrics

• University of Colorado School of Medicine

• Attending Pediatrician,

• Children's Hospital

• Denver, Colorado

• Cynthia J. Carney

• Consumer

• Washington, District of Columbia

• James R. Cooley, MD, FAAP

• Senior Consultant in Pediatrics, Clinical Quality Management

• Harvard Community Health Plan,

• Brookline, Massachusetts

• Larry Culpepper, MD, MPH

• Professor of Family Medicine

• Brown University/Memorial Hospital of Rhode Island

• Roland D. Eavey, MD, FAAP, FACS

• Associate Professor of Otology and Laryngology

• Harvard Medical School

• Director, Pediatric Otolaryngology Service

• Massachusetts Eye and Ear Infirmary

• Lynne V. Feagans, PhD

• Professor of Human Development

• College of Health and Human Development

• The Pennsylvania State University

• Terese Finitzo, PhD, CCC/A

• Administrative Director, Neuroscience Center

• Methodist Medical Center,

• Dallas, Texas

• Ellen M. Friedman, MD

• Associate Professor of Otolaryngology and Pediatrics

• Baylor College of Medicine,

• Texas

• Janice A. Goertz, RN, CPNP, BS

• Certified Pediatric Nurse Practitioner

• Pediatric Centers of Southwestern Michigan, PC

• Allan J. Goldstein, MD, FACS

• Director, Medical Affairs,

• Battle Creek Health System,

• Michigan

• Quality Assurance,

• American Academy of Otolaryngology-Head and Neck Surgery

• Kenneth M. Grundfast, MD

• Chairman, Department of Otolaryngology-Head and Neck Surgery

• Children's National Medical Center

• Douglas G. Long, MD

• Family Physician,

• Boothbay, Maine

• Loretta L. Macconi, RN, MSN, CRNP

• School Nurse Practitioner,

• School District of Philadelphia

• LoBirtha Melton, RN, BSN, MPA

• Director, Women & Children Services

• Alameda County Medical Center,

• Oakland, California

• Joanne Erwick Roberts, PhD, CCC-SLP & AUD

• Research Associate Professor of Speech and Hearing Sciences

• University of North Carolina at Chapel Hill

• Jessie L. Sherrod, MD, MPH, FAAP

• Assistant Professor of Pediatrics and Infectious Disease

• Charles R. Drew University of Medicine and Science

• Clinical Assistant Professor of Pediatrics

• University of California at Los Angeles

• Jane E. Sisk, PhD

• Professor, School of Public Health

• Columbia University

Consultants and Subcontractors ^[1]

Panel Staff

• Hanan S. Bell, PhD

• American Academy of Family Physicians

• Kansas City, Missouri

• Methodologist

• Karin Bierstein, JD, MPH

• (until May 1993)

• Maureen T. Hannley, PhD

• (from May 1993)

• American Academy of Otolaryngology-Head and Neck Surgery

• Alexandria, Virginia

• Health Policy Coordinators

• Carla T. Herrerias, MPH

• Project Manager

• Robert H. Sebring, PhD

• Principal Investigator

• American Academy of Pediatrics

• Elk Grove Village, Illinois

AHCPR Staff

• Jean R. Slutsky,PA, MSPH

• Project Officer

• Randie A. Siegel

• Managing Editor

• Rose Calbert Findley,MUP

• Public Affairs Specialist

Pilot Reviewers ^[2]

• Neil H. Brooks, MD

• Rockville Family Physicians

• Rockville, Connecticut

• James Feist, MD

• Practicing Pediatrician

• Bozeman, Montana

• Patrick E. Killeen, MS, PA-C

• Co-chair, Public Education Committee

• Connecticut Academy of Physician Assistants

• Southbury, Connecticut

• R. Dee Legako, MD

• Family Physician

• Edmond, Oklahoma

• Rodney P. Lusk, MD

• Associate Professor,

• Washington University School of Medicine

• Surgeon-in-Chief, Pediatric Otolaryngology

• St. Louis Childrens Hospital

• St. Louis, Missouri

• Carol A. Rudy, MPH, ARNP, CPNP

• Rockwood Clinic

• Spokane, Washington

• Peggye Wells, MD

• Children's Clinic of Clarksdale

• Clarksdale, Mississippi

• Timothy W. Wheeler, MD

• Kaiser Permanente

• Fontana, California

Peer Reviewers ²

• Herman Abromowitz, MD

• Private Practice

• Dayton, Ohio

• Gerson H. Aronovitz, MD

• Clinical Assistant Professor of Pediatrics

• Emory University College of Medicine

• Atlanta, Georgia

• Cheston M. Berlin, Jr, MD

• University Professor of Pediatrics

• Professor of Pharmacology

• College of Medicine

• The Pennsylvania State University

• Hershey, Pennsylvania

• Joel Bernstein, MD, PhD, PC

• Williamsville, New York

• Nicholas Black, MD

• Reader in Public Health Medicine

• London School of Hygiene and Tropical Medicine

• London, England

• Mary Ann Block, DO

• Assistant Professor

• Texas College of Osteopathic Medicine

• Fort Worth, Texas

• Stanley Block, MD, FAAP

• Kentucky Pediatric Research, Inc.

• Bardstown, Kentucky

• ^[2]Being listed in this section does not necessarily imply endorsement of the Guideline products.

• Jeffrey L. Blumer, PhD, MD

• Professor of Pediatrics and Pharmacology

• School of Medicine Case Western Reserve University

• Cleveland, Ohio

• Francine Boullosa, PA

• Kaiser Permanente NW Region

• American Academy of Physician Assistants

• Salem, Oregon

• Patrick E. Brookhouser, MD

• Director,

• Boys Town National Research Hospital

• Omaha, Nebraska

• Orval E. Brown, MD

• Chairman, Division of Pediatric Otolaryngology

• Children's Medical Center of Dallas

• Dallas, Texas

• Erdem I. Cantekin, PhD

• Professor of Otolaryngology

• School of Medicine

• University of Pittsburgh

• Pittsburgh, Pennsylvania

• Susan A. Carlin, MD

• Assistant Professor of Pediatrics

• Case Western Reserve University

• Metro Health Medical Center

• Cleveland, Ohio

• Stephen A. Chartrand, MD

• Associate Professor of Pediatrics and Medical Microbiology

• School of Medicine

• Creighton University

• Omaha, Nebraska

• Van G. Chauvin, MSN, MHS, RNCS, CSN

• Family Nurse Practitioner Preceptor

• Dallas Independent School District

• Dallas, Texas

• Evelyn Cherow, MA

• Director, Audiology Division

• American Speech and Hearing Association

• Rockville, Maryland

• Albert M. Collier, MD

• Professor of Pediatrics

• Chief, Pediatric Infectious Diseases Division

• School of Medicine University of North Carolina at Chapel Hill

• Robert M. Corwin, MD, FAAP

• Pediatric Associates

• Camilius, New York

• William G. Crook, MD, FAAP

• International Health Foundation, Inc.

• Jackson, Tennessee

• Sidney T. Dana, MD

• Clinical Professor of Otolaryngology

• Upstate Medical Center

• State University of New York

• Chair, Department of Otolaryngology

• Community General Hospital

• Syracuse, New York

• Ronald W. Deskin, MD

• Associate Professor of Otolaryngology

• University of Texas

• Chief, Division of Pediatric Otolaryngology

• University of Texas

• Medical Branch at Galveston

• Galveston, Texas

• Amelia Drake, MD

• Department of Pediatrics

• Division of Infectious Disease

• University of North Carolina

• Chapel Hill, North Carolina

• Burris Duncan, MD, FAAP

• Professor of Pediatrics

• School of Medicine

• University of Arizona

• Tucson, Arizona

• Bruce Edgren, MD

• Director of Clinical Programs

• Professor of Pharmacy

• University of Minnesota

• Minneapolis, Minnesota

• Jeanne N. Farrell, Med, CCC-A

• Clinical Audiologist

• St. Francis Hospital and Medical Center

• Hartford, Connecticut

• Philip Fireman, MD

• Professor of Pediatrics

• School of Medicine

• University of Pittsburgh

• Director of Allergy/Immunology

• Children's Hospital of Pittsburgh

• Pittsburgh, Pennsylvania

• Nick Freemantle

• Lecturer in Health Policy and Research

• Nuffield Institute for Health

• University of Leeds

• United Kingdom

• Sandy Friel-Patti, PhD

• Associate Professor

• Callier Center for Communication Disorders

• University of Texas

• Dallas, Texas

• Jack Froom, MD

• Professor of Family Medicine

• State University of New York

• Stony Brook, New York

• Viola M. Frymann, DO, FAAO

• Director,

• Osteopathic Center for Children

• La Jolla, California

• Diane E. Fuquay, MD

• Children's Hospital and Medical Center

• Seattle, Washington

• George A. Gates, MD

• Professor and Vice Chair Department of Otolaryngology- Head and Neck Surgery

• School of Medicine

• Washington University

• St. Louis, Missouri

• Michael J. Gerardi, MD, FAAP

• Director, Pediatric Emergency Services

• St. Barnabas Medical Center

• Livingston, New Jersey

• G. Scott Giebink, MD

• Professor of Pediatrics and Otolaryngology

• School of Medicine University of Minnesota

• Minneapolis, Minnesota

• Harold Goodman, DO

• Director,

• Center for Osteopathic Care

• Silver Spring, Maryland

• Judith S. Gravel, PhD

• Director of Audiology

• Associate Professor of Otolaryngology

• Albert Einstein College of Medicine and the Montefiore Medical Center

• Bronx, New York

• Mark P. Haggard

• Director, Institute of Hearing Research

• Medical Research Council, United Kingdom

• Nottingham, United Kingdom

• Barbara M. Harley, MD

• Director, Division of Ambulatory Pediatrics

• Allegheny General Hospital

• Pittsburgh, Pennsylvania

• Gerald Healy, MD

• Department of Otolaryngology

• Children's Hospital Medical Center

• Boston, Massachusetts

• James A. Hedrick, MD

• Physicians to Children and Adolescents

• Bardstown, Kentucky

• Frederick W. Henderson, MD

• Professor of Pediatrics

• University of North Carolina at Chapel Hill

• Thomas J. Herr, MD, FAAP

• Pediatric Group Associates, SC

• Moline, Illinois

• J. Patrick Hieber, MD

• Clinical Pediatric Associates

• Dallas, Texas

• Lisa S. Honigfeld, PhD

• Consumer

• Hartford, Connecticut

• Virgil M. Howie, MD

• Professor of Pediatrics

• University of Texas Medical Branch

• Galveston, Texas

• Kathryn R. Jimmerson, RN, DSN, CPNP

• Assistant Professor of Pediatric Primary Health

• School of Nursing

• University of Alabama

• Birmingham, Alabama

• Lawrence Kleinman, MD

• Children's Hospital

• Boston, Massachusetts

• Michael L. LeFevre, MD

• Associate Professor of Family and Community Medicine

• University of Missouri

• Columbia, Missouri

• James D. Legler, MD

• Assistant Professor of Family Practice

• University of Texas Health Science Center

• San Antonio, Texas

• Martin Levinson, MD

• Chairman, Department of Pediatrics

• Sinai Hospital

• Detroit, Michigan

• Georgia P. MacDonough, RN, MA, SNP

• Faculty Associate

• College of Nursing

• Arizona State University

• Tempe, Arizona

• Ellen Mandel, MD

• Assistant Professor of Pediatrics and Otolaryngology

• School of Medicine

• University of Pittsburgh

• Pittsburgh, Pennsylvania

• Colin Marchant, MD

• Associate Professor of Pediatrics

• Division of Infectious Disease

• New England Medical Center

• Boston, Massachusetts

• Kenneth E. Matthews, MD, FAAP

• Bryan, Texas

• Richard A. Maw, MS, FRCS

• Consultant Otolaryngologist

• Bristol Royal Infirmary

• United Kingdom

• Paula Menyuk, PhD

• Professor of Developmental Studies

• School of Education

• Boston University

• Boston, Massachusetts

• Philip G. Miller, MD

• Olson Pediatric Clinic

• Lake Oswego, Oregon

• Robert E. Mines, MD

• North Oakland Pediatric Medical Group

• Oakland, California

• Mary Anne Morelli, DO

• Osteopathic Center for Children

• College of Osteopathic Medicine of the Pacific

• La Jolla, California

• Robert F. Nease, Jr, MD

• Assistant Professor of Community and Family Medicine

• Dartmouth Medical School

• Hanover, New Hampshire

• John D. Nelson, MD

• Professor of Pediatrics

• University of Texas

• Southwestern Medical Center

• Dallas, Texas

• Jerry L. Northern, PhD

• Head, Audiology Division

• Professor of Otolaryngology

• University of Colorado Health Sciences Center

• Denver, Colorado

• Gregory F. O'Brien, MD

• President,

• Cleveland Ear, Nose, Throat and Facial Surgery Group Inc.

• Cleveland, Ohio

• Lawrence C. Pakula, MD, FAAP

• Timonium, Maryland

• Jack L. Paradise, MD

• Professor of Pediatrics

• University of Pittsburgh

• School of Medicine

• Pittsburgh, Pennsylvania

• James Christopher Post, MD

• Assistant Professor of Otolaryngology

• Division of Pediatric Otolaryngology

• School of Medicine

• University of Pittsburgh

• Pittsburgh, Pennsylvania

• William P. Potsic, MD

• Director, Pediatric Otolaryngology

• Children's Hospital of Philadelphia

• Professor of Otolaryngology/ Head and Neck Surgery

• University of Pennsylvania

• Philadelphia, Pennsylvania

• Ioana Razi, MD

• Pediatrician

• Washington, District of Columbia

• Bonnie L. Renfro, MD, FAAP, FAAOA

• Albuquerque, New Mexico

• Beth Richardson, MSN, RN, CPNP

• Assistant Professor

• Indiana University

• Pediatric Nurse Practitioner

• Methodist Hospital

• Indianapolis, Indiana

• William J. Rodriguez, MD, PhD

• Chairman, Pediatric Infectious Diseases

• Children's Hospital

• Professor of Pediatrics

• Medical School

• George Washington University

• Washington, District of Columbia

• Rita A. Rooney, RN, MPH

• Health Education Specialist

• Asthma and Allergy Foundation of America

• Washington, District of Columbia

• Richard A. Rosen, MD

• Clinical Assistant Professor of Pediatrics

• Brown University

• Harvard Community Health Plan of New England

• Lincoln, Rhode Island

• Richard M. Rosenfeld, MD, MPH

• Assistant Professor and Director

• Division of Pediatric Otolaryngology

• State University of New York Health Science Center at Brooklyn and Long Island College Hospital

• New York

• Jackson Roush, PhD

• Associate Professor and Director Division of Speech and Hearing Sciences

• Medical School University of North Carolina at Chapel Hill

• Patricia Rowell, PhD, RN

• Senior Policy Fellow

• Department of Practice, Economics and Policy

• American Nurses Association

• Washington, District of Columbia

• Robert J. Ruben, MD

• Professor and Chair Department of Otolaryngology

• Albert Einstein College of Medicine

• Montefiore Medical Center

• Bronx, New York

• Lt. Col. Robert Edward Sayers, MD

• United States Air Force Medical Corps

• Developmental Pediatrics

• William Beaumont Army Medical Center

• El Paso, Texas

• David L. Schriger, MD, MPH

• Assistant Professor

• Emergency Medical Center

• School of Medicine

• University of California at Los Angeles

• Nadine C. Schwab, RN, PNP, MPH

• Education Consultant for Health Services

• Connecticut Department of Education

• Orange, Connecticut

• Paul Shurin, MD

• Associate Professor of Clinical Pediatrics

• Montefiore Hospital

• Bronx, New York

• Jeffrey H. Silber, MD, PhD

• Assistant Professor of Pediatrics School of Medicine

• The University of Pennsylvania

• Philadelphia, Pennsylvania

• Grayce Sivigny, RNC, CPNP, CSNP

• Montana Director

• National Association of School Nurses

• Big Timber, Montana

• David S. Smith, MD

• Director, Family Practice Residency Program

• St. Michael Hospital

• Assistant Professor of Family and Community Medicine

• Medical College of Wisconsin

• Milwaukee, Wisconsin

• David T. Tayloe, Jr, MD, FAAP

• Goldsboro Pediatrics, PA

• Goldsboro, North Carolina

• David W. Teele, MD

• Professor of Pediatrics

• Christchurch School of Medicine

• New Zealand

• Professor of Pediatrics

• School of Medicine

• Boston University

• Boston, Massachusetts

• R. Edward Thompson

• Chief Executive Officer and Executive Director

• The EAR Foundation

• Nashville, Tennessee

• Hugh H. Tilson, MD, DrPH

• Chairman, Health Outcomes Work Group

• Pharmaceutical Manufacturers Association

• N. Wendell Todd, MD

• Associate Professor of Surgery (Otolaryngology) and Pediatrics

• School of Medicine

• Emory University

• Atlanta, Georgia

• Ernesto C. Torres, MD

• Frederick, Maryland

• Betty R. Vohr, MD

• Medical Director,

• Rhode Island Hearing Assessment Project

• Director,

• Neonatal Follow-up Clinic

• Associate Professor

• School of Medicine

• Brown University

• Providence, Rhode Island

• Howard Weinblatt, MD, FAAP

• Ann Arbor, Michigan

• Richard J. Wiet, MD, FACS

• Professor of Clinical Otolaryngology

• Division of Otology, Neurotology, Skull Base Surgery

• Northwestern Memorial Hospital Medical School

• Northwestern University

• Chicago, Illinois

• Ram Yogev, MD

• Professor of Pediatrics

• Medical School Northwestern University

• The Children's Memorial Hospital

• Chicago, Illinois

• Delosa A. Young, MD, FAAP

• Quality Resource Manager

• Department of Pediatrics

• Kaiser Permanente,

• CAP Medical Group

• Reston, Virginia