Clinical Guidelines And Primary Care J Am Board Fam Pract: first published as 10.3122/jabfm.8.4.305 on 1 July 1995. Downloaded from

Alfred O. Berg, MD, MPH, Series Editor Otitis Media With Effusion In Young Children: Treatment In Search Of A Problem? Lany Culpepper, MD, MPH, andJack Froom, MD

Family physicians and pediatricians see patients We based our review on the published Clinical who have otitis media with effusion (OME), also Practice Guideline, research reports used to derive known as serous otitis or "glue ear," on a daily ba­ the panel recommendations, research reports not sis. OME is a common sequela of acute otitis me­ considered by the panel, one author's (LC) per­ dia, but the timing and appropriate types of inter­ sonal participation on the panel, and clinical vention are subject to debate. Current practice practice experience. Presentation and discussion includes observation only, treatment with antibi­ (by LC) of the guideline at a Hastings Center for otics and decongestants, referral for ear tubes, Ethics meeting further enriched this review. In and even adenoidectomy with or without tonsil­ the following sections we present a summary of lectomy. Both the utility of these interventions the content of the report and conclude with our for resolving OME and the value of resolving critique of the guideline. OME to prevent long-term complications, par­ ticularly compromised development of language Importance of the Problem skills, are uncertain. Otitis media with effusion is one of the most Otitis Media with Effusion in Young Children. common reasons for prescribing antibiotics to Clinical Practice Guideline l addresses these and re­ children and the most common reason for a sur­ lated issues. A consortium of the American Acad­ gical procedure. The number of visits to physi­ emy of Pediatrics, the American Academy of cians for otitis media, including both acute otitis Family Physicians, and the American Academy of media and OME, increased from 9.9 million in Otolaryngology-Head and Neck Surgery under 1975 to 24.5 million in 1990.2 The AHCPR contract with the Agency for Health Care and panel estimated that 25 to 35 percent of these Policy Research (AHCPR) convened an expert cases represented OME. Visits to pediatricians http://www.jabfm.org/ panel to produce the guideline. On the basis of accounted for 56.4 percent of otitis media visits; limited evidence and expert opinion, the panel those to family physicians and general practition­ recommended that antibiotics or insertion of ear ers 30.4 percent. Only 7.2 percent of otitis media tubes (if a hearing deficit exists) be offered to chil­ visits were to otolaryngologists (percentage ob­ dren whose OME has not resolved by 3 months tained from a special analysis at our request by and that children with hearing deficits of 20 dB the National Center for Health Statistics using on 1 October 2021 by guest. Protected copyright. or more receive ear tubes by 4 to 6 months. We National Ambulatory Medical Care Survey data). advise physicians to question both recommenda­ The US Food and Drug Administration found tions. Figure 1 displays the practice algorithm that about 14 percent of all courses of antibiotics published as part of the Clinical Practice Guideline prescribed in the United States were for otitis and incorporates its recommendations. media. In 1986, 44.5 million courses were prescribed for children 10 years old or younger for otitis media, constituting 42 percent of all Submitted, revised, 15 March 1995. antibiotic prescriptions they received. 3 Using From the Depar1:1TIent of Family Medicine, Memorial Hospi­ tal of Rhode Island/Brown University, Pawtucket (LC), and the federal data, one recent estimate is that approxi­ Department of Family Medicine, Health Sciences Center, State mately 800,000 children received 1.3 million University of New York at Stony Brook OF). Address reprint re­ tympanostomy tubes in 1988. Of these, 30 per­ quests to LarryCulpepper, MD, MPH, Department of Family 4 Medicine, Memorial Hospital of Rhode Island/Brown Univer­ cent were replacements. In 1986, 31 million sity, 255 Main Street, Suite 201, Pawnlcket, RI, 02860. visits to physicians were due to otitis media, and

Otitis Media in Young Children 305 J Am Board Fam Pract: first published as 10.3122/jabfm.8.4.305 on 1 July 1995. Downloaded from

Primary care clinician examining an otherwise healthy child age ...------. 1-3 years with no craniofacial or neurologic abnormalities or Note: The asymptomatic patient sensory deficits suspects otitis with fluid in the ear and no media with effusion (OME) signs or symptoms of ear (A, B). infection by definition does not have acute otitis media.

21------%.-----.

Clinician performs pneumatic otoscopy (C).

41------, Is the clinician Clinician may confirm certain of the No -. clinical diagnosis of OME diagnosis of by tympanometry (D). OME?

Does tympanometry Yes confirm the diagnosis of OME?

No http://www.jabfm.org/ Yes 61------...::1::.....-.----, 7~------~L------~ Exit this algorithm to Options for management of this individualized patient patient with OME should include: management appropriate (1) a. observati08 ~E) to the clinical situation. b. Oral antibiotic therapy (F) on 1 October 2021 by guest. Protected copyright. AND (2) Environmental risk factor control counseling (G). ATTENTION Management of the patient at this point in the clinical course should not include: (1) Surgery, including myringotomy with or without tube insertion, tonsillectomy, or adenoidectomo~H) (2) Decongestants and/or antihistamines (I) OR (3) Oral steroid therapy (J).

Figure 1. Algorithm for managing otitis media with effusion in an otherwise healthy child aged 1 through 3 years. I

306 JABFP July-August 1995 Vo l. 8 0.4 J Am Board Fam Pract: first published as 10.3122/jabfm.8.4.305 on 1 July 1995. Downloaded from

Figure 1 (continued)

Does the patient still have OME 6 weeks after diagnosis by pneumatic otoscopy r------No------~ with optional confirmation by tympanometry?

Yes

9~ ______~ __---- ____--1 ATTENTION Management of this patient at this Management of this patient with OME point should not include: for 6 weeks should include: (1) Surgery. including myringotomy (1) a. Observation with or without tube Insertion, OR tonsillectomy, or adenoidectomy b. Oral antibiotic therapy OR AND (2) Decongestants and/or (2) Environ~ental risk factor control antihisfamines counseling OR http://www.jabfm.org/ AND (3) Oral steroid therapy. (3) Option of hearing evaluation now. on 1 October 2021 by guest. Protected copyright. Does the patient still have OME 3 months after diagnosis by pneumatic otoscopy r------No------~ with optional confirmation by tympanometry?

11r------____~~ Exit this algorithm to individualized patient management appropriate to the clinical situation.

Otitis Media in Young Children 307 J Am Board Fam Pract: first published as 10.3122/jabfm.8.4.305 on 1 July 1995. Downloaded from

Figure 1 (continued)

Refer patient for hearing evaluation. Management of this patient with OME and With unilateral or insignificant hearing loss, 3 or more months after diagnosis with OME should include: No-+ (1) a. Observation OR b. Oral antibiotic therapy AND (2) Environmental risk factor control Yes counseling.

15~ ______~ ______~

Management of this patient with OME and hearing loss, 3 or more months ...------. after diagnosis with OME should include: ATTENTION (1) a. Oral antibiotic therapy Management of this patient OR at this point should not b. Bilateral myringotomy with include: tube placement (1) Tonsillectomy and/or AND adenoidectomy (2) Environmental risk factor control OR counseling. (2) Decongestants and/or antihistamines http://www.jabfm.org/ OR (3) Oral steroid therapy.

Does the j)atient still have OME 4-6 months Exit this algorithm to after diagnosis by individualized patient on 1 October 2021 by guest. Protected copyright. pneumatic otoscopy ">--- No --...... management appropriate with optional to the clinical situation. confirmation by tympan ometry1 18'~------~ Management of this patient with OME for 4-6 months and a history of Yes significant (at least 20 db) bilateral hearing loss should include: (1) Bilateral myringotomy with tube placement AND (2) Environmental risk factor control counseling AND (3) Management appropriate to the clinical situation.

308 JABFP July-August 1995 Vo l. 8 No. 4 the total direct and indirect costs for that year without treatment. The natural history of this have been estimated at $3.5 billion.s Surgical group of children, therefore, is unknown. J Am Board Fam Pract: first published as 10.3122/jabfm.8.4.305 on 1 July 1995. Downloaded from costs for procedures for otitis media exceed $1.2 billion annually. 6 The Guideline Development Process With AHCPR approval, the consortium con­ Prevalence and Epidemiology of Otitis Media vened an interdisciplinary panel selected from with Effusion individuals recommended by specialty organi­ OME is a common experience for children. Cassel- zations. The panel included 3 family physi­ brant, et al. 7 reported that among 103 children aged cians (Alfred O. Berg, MD, MPH, Douglas G. 2 to 6 years participating in group child care and Long, MD, and Larry Culpepper, MD, MPH), monitored for 2 years, 53 percent had at least one 2 pediatricians, 5 otolaryngologists, an infectious episode of OME during the first year of study and disease specialist, a psychologist, a speech-language 61 percent in the second year; 30 percent had re­ pathologist, 3 nurse practitioners, an audiologist, current bouts. OME is present in at least 20 per­ an economist-health policy analyst, and a con­ cent of cases 2 months following an attack of acute sumer. Most are in academic and research posi­ otitis media. l In studies that used pneumatic otos­ tions. Following the first panel meeting, 2 initial copy and tympanometry to detect effusion, higher panelists were replaced because of pharmaceutical rates of effusion at 2 months have been reported. 1,8,9 funding conflicts of interest. Following the second About two-thirds of OME occurs bilaterally.lo meeting, a panel co-chair, Alfred O. Berg, MD, OME often develops as a result of eustachian MPH, was named in addition to the original chair, tube dysfunction preventing normal drainage of Sylvan E. Stool, MD, chairman of the Department middle ear fluid. Such dysfunction is particularly of Otolaryngology at the University of Pittsburgh, common in those of young age because of the the location of the controversial Cantekin Affair. I 2 anatomy of the skull and diameter of the eusta­ The panel held 5 meetings. At the first, they chian tube. Upper respiratory tract infections of narrowed the general topic of otitis media to otitis viral or bacterial origin frequently precede both media with effusion, and the panel received in­ acute otitis media and OME. Children in envi­ structions in evidence-based guideline develop­ ronments associated with frequent upper respira­ ment. This instruction continued at the second tory tract infections, such as day-care settings, are meeting, during which the panel developed its lit­ at particular risk for OME. The panel singled out erature search specifications and defined the tar':' other risk factors including male sex, siblings, get child as aged 1 to 3 years. The third meeting bottle feeding, and supine position feeding. l Pas­ included receipt of public testimony and an initial http://www.jabfm.org/ sive smoking is estimated to be responsible for 8 assessment of the literature. The fourth meeting percent of OME episodes. I I The role of allergy as included subgroup work to assess the validity of a risk factor is undetermined. literature to be included in the panel's deliber­ ations. The final meeting was devoted to reaching Natural History agreement on recommendations, a process com­

In a study of 1439 Dutch children, 60 percent of pleted by mail following the meeting. on 1 October 2021 by guest. Protected copyright. cases of OME cleared without intervention after The panel based its recommendations on a 3 months; 60 percent of the remaining cases re­ combination of scientific evidence and expert solved within another 3 months.7 The panel opinion. When the panel failed to find compel­ adopted this duration of OME as the best esti­ ling evidence for or against interventions, they mate. Although the Dutch study involved large defined clinical options that a reasonable clinician numbers of children with long-term follow-up , might wish to use. A statement of "No recom­ (up to 7 years), it was population based and ex­ mendation" was used when scientific evidence trapolation of findings to individuals seen by was lacking and there was no compelling reason family physicians is unwarranted. Neither the to employ expert judgment. panel nor the authors were able to find large studies of children with acute otitis media treated Content of the Report by primary care physicians who subsequently de­ The contents of the guideline is disseminated in veloped OME and who had long-term follow-up three formats, two for clinicians (Otitis Media with

Otitis Media in Young Children 309 IU/il.l'iOI7 in YOI/Jlg Cbildrcil. ClilliCflI Practice it considered adequate. Met,l-analysis of these J Am Board Fam Pract: first published as 10.3122/jabfm.8.4.305 on 1 July 1995. Downloaded from Guide/ille' and /V/fllwgillp, Otitix !vledill witb Httilxioll 14 studies W,)S not possible hecausc of the diver­ ill }()/fJ1g Cbildren. Quick 1«:/('/'clI(e Guic/e ./fn· sity of measurement instrulllents Llsed and bck ll Clillicimls ) and the third for parents (Middle Ear of uniformity in data. The definitions of OM F Fluid ill Children. Pi/re7lt Guide). 1+ The p,mel di­ varied, and often hearing status assessment was vided the Clinical Pmctice (;'"idclillf: into chapters lacking; instead investigators often lIsed presence that address major issues involved in the diagnosis of effusion as a proxy t()r hearing loss. tn addition, and treatment of ONIE. Additional chapters sum­ investigators used a variety of tests to assess lan­ marize recommendations and discuss costs and guage and speech status at outcome. Of the 14 research issues. studics, only four had a "no treatment" control The interventions considered by thc panel in­ group, the largest of which containcd only 26 clude treatment with antibiotics, antibiotics plus subjects. steroids, steroids alone, antihistamines with or Given the pervasive weakness of the literature, withollt decongestants, and the surgical pro­ it is not surprising that, "In summary, the panel cedures of myringotomy with tympanostomy tubes found that rigorous, methodologically sound insertion, adenoidectomy, and tonsillectomy. The research docs not adequately support or refute guideline also addresses other therapies and the thcory that untreated OME results in speech/ includes recommendations to parents regarding language delays or deficit." It further found that environmental modifications. (I) the OME-related level of hearing loss (if one exists) required to produce language and speech Target Child Defined deficits is unknown; (2) the duration required for The target child is "age 1-3 years with no cranio­ hearing loss to produce such deficits is unknown; facial or neurologic abnormalities or sensory defi­ and (3) whether these deficits are transient or are cits, otherwise healthy except for otitis media long-lasting is unknown. with cffusion." It is this agc group for which thcre Bccause of the weakness of the literature and is concern that impaired hearing as a result of disparate panel opinion, by majority vote the bilateral otitis media with effusion might compro­ panel concluded that "the published data support mise speech and language development. Physi­ the following trends: (I) a weak association be­ cians and parents must use their own judgment in tween otitis media with effusion in early life and assessing the applicability of the guidelinc to abnormal speech and language development of

younger or older children. children younger than age 4 years; and (2) a wcak http://www.jabfm.org/ association between early otitis media with effu­ Clinical Outcomes sion and delay in expressive language develop­ The panel defined short- and long-term out­ ment and behavior (attention) in children over comes .lssociated with either therapy or observa­ 4- years. The effects of OM E on other hearing rc­ tion of OME. Short-term outcomes include lated domains are less clear." The Cli77iClli Pmcticc either clearance or persistence of middle ear effu­ Guidelinc notes that available data "do not show a sion with its impact on hearing; common and rare consistent effect of OME on language and/or on 1 October 2021 by guest. Protected copyright. side effects of antibiotics, steroids, and antihista­ learning once the disease process and its associ­ mine-decongestants; and risks of the surgical pro­ ated hearing loss have resolved," and, "there cedures including anesthetic risks, costs, and seems to be little long-term effect of OME that complications. Symptoms and associated quality­ appears for the first time after age 3." of-life issues were not considered, The long-term outcomes of interest are those Diagnosis and Hearing Evaluation related to impaired hearing and the potential for A strong recommendation of the panel, based on abnormal speech and language development. The limited scientific evidence and strong consensus, panel found only weak scientific evidence, which was that the diagnostic evaluation of suspected did not include any randomized control trials, for otitis media with effusion should include pneu­ a connection between hearing loss due to OME matic otoscopy. "Otoscopy alone (without the use and abnormal speech and hmg'uage development. of pneumatic otoscopy to test tympanic mem­ The panel hased its assessment on the 14- studies brane mobility) is not recommended,"

310 )ABFP July-August 1995 Vol. H No.-t In patients visiting otolaryngologists for whom these exposures makes a clinically important dif­ J Am Board Fam Pract: first published as 10.3122/jabfm.8.4.305 on 1 July 1995. Downloaded from myringotomy is planned, pneumatic otoscopy has ference. Consequently, it proposed as a clinical a sensitivity of85 to 90 percent and a specificity of option, based on limited scientific evidence and 70 to 79 percent. IS Although the sensitivity and strong panel consensus, that "parents should specificity are the same in primary care popula­ be encouraged to control environmental risk tions, the prevalence of OME is lower, so the factors." positive predictive value of pneumatic otoscopy is less. Pharmaceutical Therapies The panel suggested that use of tympanometry The panel concluded that antibiotics confer a 14 as a confirmatory test for OME is a clinical op­ percent improvement within the first month, tion. The positive predictive value of a flat (type based on several meta-analyses that combined B) tympanogram is estimated at between 49 and data from blinded randomized control trials of 99 percent. This value is the likelihood that effu­ antibiotics. Their recommendation, based on sion is present, using myringotomy as the cri­ limited and inconsistent scientific evidence and terion standard. Given a positive predictive value panel consensus, was "use of antibiotic agents is of only 49 percent, less than one-half the children one option for the treatment of a child with otitis with an abnormal tympanogram have a hearing media with efuIsion." They noted that the small loss.16,17 Thus, using a combination of pneumatic improvement 'in resolution of OME must be otoscopy and tympanometry to detect children weighed against the side effects and costs of anti­ with OME and associated hearing loss will yield a biotic therapy. high rate of false positives, because many children The panel found contradictory evidence be­ with abnormal findings have no hearing defi­ tween studies of steroid therapy alone, antibiotic ciency. For this reason the panel recommended, plus steroid therapy, antibiotics alone, and pla­ based on limited scientific evidence and expert cebos. Consequently, based on limited scientific opinion, that hearing evaluation be performed for evidence and majority opinion, the panel con­ any child who has had bilateral OME for a total of cluded that "steroid medications are not recom­ 3 months and that such testing be an option mended for treatment of otitis media with effu­ among children who have OME for a shorter sion in a child of any age." The use of steroids in duration. children recently exposed to varicella can lead to The panel noted also that, especially for young disseminated varicella with severe consequences.

children, there are several methods of evaluating The panel also made a strong recommenda­ http://www.jabfm.org/ hearing and that such evaluation can be technically tion, based on evidence that can be generalized to difficult. In addition, hearing testing might not be a child of any age, that "antihistamine andlor de­ available in many primary care settings, especially congestant agents are not recommended for in rural areas. Nevertheless, the panel recom­ treatment of otitis media with effusion." mended hearing testing because of the "firm be­ lief that placement of tympanostomy tubes is not Surgical Therapies indicated when OME is unaccompanied by bilat­ The panel made three recommendations on in­ on 1 October 2021 by guest. Protected copyright. eral hearing impairment." Of note, the panel sertion of tympanostomy tubes. First, they made found po evidence on which to base its determi­ a strong recommendation, based on evidence that nation of the level of hearing impairment requir­ OME resolves spontaneously in most cases, that ing intervention. It arbitrarily chose 20 dB or "myringotomy with or without insertion of tym­ worse in both ears as the threshold. panostomy tubes should NOT be performed for initial management of otitis media with effusion Control ofEnvironmental Factors in an otherwise healthy child." The panel found that several factors amenable to Second, based on limited scientific evidence change increase the chance that a child will de­ and panel consensus, they recommended that as a velop OME including (1) bottle feeding, particu­ clinical option "antibiotic therapy OR bilateral larly in a supine position; (2) passive smoking; and myringotomy with insertion of tympanostomy (3) enrollment in group child care. The panel tubes ~ay be chosen to manage bilateral otitis found no evidence that intervening to decrease media with effusion that has lasted a total of three

Otitis Media in Young Children 311 J Am Board Fam Pract: first published as 10.3122/jabfm.8.4.305 on 1 July 1995. Downloaded from months in :111 otherwise healthy child age 1-3 allergy management as ,) treatment for otitis years who has a bilateral hearing deficit (defined media with effusion." Literature on the role of ,11- as 20 decibels hearing threshold level or less in lergy hyposensitization as a treatment for OME is the hetter-hearing ear)." In this matter, the Uilli­ limited and flawed hy weak rese:m:h designs. wI Pl"f{cticl' G"ideline and the published algorithm conflict with the discllssion in the Quick Retemm! Other Therapies Guide .If)/" Cliniciflns. \Vhile the ClilliCflI Practice Similarly, because of lack of scientific evidence, Guidelille and the algorithm indicate that at 3 the panel concluded "No recolllmendation is Illonths the clinician must either LIse antibiotics or made regarding other therapies (chiropractic, insert tubes, the Quick R~fe"(,lIcc G{(ide states, "Ob­ holistic, naturopathic, tr,l(litional or indigenous, servation OR antibiotic therapy are treatment op­ homeopathic) for the treatment of otitis media tions for children with effusion that has been with effusion in the otherwise healthy child age present less than 4- to () months and at any time in 1 to 3 years." children without a 20-decihcl hearing threshold level or worse in the better-hearing ear." Cost Impacts Third, it made ,\ moderate recommendation, Through a series of extrapolatjons based on sev­ based on limited evidence and strong consensus, eral large data sets, a contractor employed by the that "hilateralmyringotol11Y with insertion of panel estimated the 199] costs of treating OME, tympanostomy tubes is recommended to manage if the guideline had been implemented for all bilateral otitis media with effusion that has lasted 2 -year-old children. For this cohort, total costs a total of 4- to 6 months in an otherwise healthy would have been $246.6 million, of which $118.8 child age 1 through 3 years who has a bilateral million was for medical care during the first hearing deficit." 3 months of OME, $80.3 million for surgical '1ympanosclerosis occurs in 51 percent of pa­ treatments, $16.8 million for pharmaceutical tients following the initial insertion of tubes.IS,I') costs, ,mel $77.1 million t()r indirect costs. Additional complications include persistent otor­ rhea (in 13 percent) and more rarely, granuloma Research Issues t()J"Illation, cholesteotoll1a, and permanent tym­ The panel named priority issues for research. panic membrane perforation. The panel noted These issues include the natural history of OME; that up to 30 percent of children receiving one set approaches to hearing evaluation; control of envi­ http://www.jabfm.org/ of tubes will receive a second set within 5 years ronmental risk factors; the role of antibiotics, and that structural changes, such as Haccidit:y, re­ steroids, allergy treatment and other therapies; traction, and tympanosclerosis, occur at increased and the impact of surgery, including on symptoms rates with repe,\t surgery. The panel did not in­ and quality of life. vestigate the long-term effects, if any, on hearing later in childhood of having ear tubes as a young Critique child. The CliniCflI Pmctice Guideline can help physicians on 1 October 2021 by guest. Protected copyright. The panel reviewed additional forms of surgery become aware of critical issues and approaches and, based on limited scientific evidence and that might be useful for the care of their patients. strong panel consenSLIS, recommended that ad­ There are areas, however, in which physicians enoidectomy not be performed for OME in the should carefully examine the information pre­ absence of specific adenoid disease. The panel sented and consult additional sources, because ,lis!) recommended that "tonsillectomy should not they could come to conclusions different from he performed, either ,llone or with adenoidec­ those reached by the panel. tomy, for the treatment of otitis media with effusion ina child of any age." Critical Issues Perhaps most important is the controversy sur­ Allergy and Hyposensitization rounding the hypothesized relation between effu­ BecaLIse of insufficient evidence clarifying the re­ sion-related hearing deficit and compromised lation between allergy and OME, the panel con­ speech and Lmguage development. As the f,TUide­ cluded, "No recommendation is made regarding line notes, such an association has yet to be

312 JABFP July-Aub'llst 1995 Vol. H No.4 proved or expressed in quantitative terms. If it ex­ For other groups of children there might be J Am Board Fam Pract: first published as 10.3122/jabfm.8.4.305 on 1 July 1995. Downloaded from ists, it is also necessary to determine the duration additional clinical concerns and outcomes of and magnitude of the required hearing loss and interest to the physician and family that influence whether it needs to be intermittent or constant. treatment. Similarly the age during which effusion-related A third issue highlighted by the panel is the dis­ hearing loss is associated with impaired develop­ tinction between OME and hearing loss. The ment of speech and language has not been ascer­ panel provides a valuable service by distinguish­ tained. Finally, whether hypothesized deficits are ing these two conditions. Persistent effusion often transient, last through childhood, or are perma­ occurs with little or no hearing deficit. Conse­ nent is unknown. Although not explored by the quently, placement of tympanostomy mbes based panel, this controversy can be viewed as part of on effusion alone will subject great numbers the larger debate about whether there are limited of children to needless surgery. The amount of temporal windows of developmental opportunity effusion and consequent hearing deficit can for children that, if not used, permanently com­ wax and wanej therefore, a single test result promise human potential. below threshold for intervention could be false The absence of compelling published studies reassurance. The converse is also truej a single that document an association between OME and abnormal hearing test is insufficient evidence that abnormal speecp and language development does the child has a persistent hearing deficit and is not prove that a causal association does not exist possibly at risk for abnormal speech and language or that such a relation is unimportant. Several development. studies use the presence of effusion as a surrogate marker for hearing loss, which might result in the The Clinical Approach inclusion of a large number of children whose The clinical approach outlined by the panel is of hearing is either normal or only mildly impaired. some value to physicians, even if the specific steps Consequently, analyses that fail to show an associa­ are questionable. The panel took into account the tion might be flawed because only a minority natural history of OMEj even with no interven­ of children in the sample have a serious OME­ tion 60 percent of children clear their effusion related hearing deficit with impaired speech and during every 3-month period of continued obser­ language development. vation. 7 The panel recommended against tympa­ Another critical issue involves the narrow defi­ nostomy tube insertion at the time OME is diag­ nition of the patient and the relevant outcomes. nosed. Instead, it recommended at least 3 months http://www.jabfm.org/ The target child as defined by the panel is a young of observation before consideration of ear mbes child with a persistent middle ear effusion and and suggested that observation for a total of 4 to no intervening recurrent episodes of acute otitis 6 months is appropriate. Second, the panel rec­ media. This case is not the otitis-prone child ommended a hearing evaluation before insertion who seeks care for frequent repeat episodes of tympanostomy tubes. Because a great number of acute otitis for whom OME might be an im­ of children with OME will have normal or only on 1 October 2021 by guest. Protected copyright. portant risk factor for recurrent infection. In­ mildly impaired hearing, such testing frequently stead, the target child is an "otherwise healthy," will prevent needless surgery. This recommenda­ often asymptomatic child with effusion, which is tion is particularly important in light of the recent usually detected either at follow-up for an episode finding that 59 percent of surgeries for tube of acute otitis media at a preventive health main­ placement are for equivocal or inappropriate indi­ tenance visit or as an incidental finding upon ex­ cations.20 The panel's recommended timing for a amination for other illness. The target child of hearing test is flawed. They recommended hear­ the guideline also is not the older child with ing assessment if bilateral effusion has persisted asymptomatic effusion detected serendipitously for 3 months. Because the purpose of a hearing or following referral from school for hearing loss test is to decide whether insertion of ear tubes is found on a screening examination. The major warranted, testing should be done immediately concern for the targeted child is the possibility before a decision to insert tubes. A test done at that hearing loss during the first 2 to 3 years of life 3 months is of no value if referral for ear tube in­ might affect speech and language development. sertion is delayed until 4 to 6 months.

Otitis Media in Young Children 313 Areas ofContrOllersy siderable difficulty for families. Second, for chil­ J Am Board Fam Pract: first published as 10.3122/jabfm.8.4.305 on 1 July 1995. Downloaded from Perhaps one of the important insights gained dren who have had OME for at least 3 months, from the Clillical Practice Guideline is that there is the Pm·ent Guide recommends 'VElking steps limited scientific evidence on which to base clini­ to prevent middle e;lr Huid (especially keeping cal care of children with OME. The panel based your child away from cigarette smoke)." This ad­ virtually all recommendations on expert opinion vice suggests a basis of scientific evidence. VVhile with limited or no scientific evidence. Such is the secondhand smoke has been implicated as a risk case not only fiJr recommendations related to an­ factor for OME, there has been no investigation tibiotics and tympanostomy tubes placement, but of its role in persistence of OME and no evidence even for whether hearing deficits secondary to that removing this exposure will alter outcomes. OME are related to impaired speech and lan­ The two presentations of this advice in the sec­ guage development. In many instances the rec­ tion on treatment could lead to family conflict or ommendations represented an action by majority feelings of guilt, especially if tbe child subse­ vote of a p,lI1el with divergent opinions, rather quently receives tympanostomy wbes. than a consensus reached through discussion. The panel had only one meeting at which it ar­ Areas of Clinical Uncertainty rived at most of its conclusions and was not subse­ There are at least two areas for which the Cliniwl quently assembled to discuss the draft of the final Practice Guideline does not fully discuss informa­ report. (Most panelists did not see the final ver­ tion critical to a clinician's treatment decisions. sion of the guideline publications before its public release.) Evidence Regarding Outcomes Foll07ving Antibiotic In several areas, the guideline is a product of Treatment ofOME the dichotomy between an interventionist ap­ Although the panel did discuss it, the CliniCfli proach to prevent a possible unproved problem Practice Guidelille does not highlight consistent and a prudent approach, which shuns interven­ evidence that there is no difference between chil­ tions of unproved benefit, especially when those dren who receive antibiotics and those who re­ entail considerable cost and possible adverse ef­ ceive placebo as measured by persistent effusion fects. These conflicting views and the lack of true or hearing deficit 2 to 4 months later. Tbis equiva­ consensus are reflected in the language of the lence was shown not only by meta-analyses con­ Clinical Practice Guideline. For instance, given the ducted for the panel, but has been reported by

limited evidence of the efficacy of antibiotics, in others conducting similar analyses.21 Thus the http://www.jabfm.org/ the chapter on drug treatments the panel pre­ limited (14 percent, or lout of 6 children treated) sented antibiotics as an option rather than a benefit obtained for the 1 month highlighted in recommendation; however, in the chapter on sur­ the Clinical Practice Guideline is transient. Given gical treatments and in the algorithm, it recom­ the cost and real potential for side effects, this ad­ mended that antibiotics be given to all children ditional evidence that any effect of antibiotics

not receiving ear tubes by 3 1110nths. lasts for only 1 month should be included in a on 1 October 2021 by guest. Protected copyright. The Parent Gllide provides a major disservice in physician's decision of whether to use them. one area. £t recommends environmental modifi­ cations with little regard for whether parents can H C/tri17g Level Tbresboldfor [17terventio17 implement the recommendations. There is no ev­ A second area of clinical uncertainty, not explored idence that sllch modifications will work; yet they by the panel, is the need to determine the magni­ might produce anxiety and guilt among parents nnle of hearing impairment that presumably re­ who did not breast feed, who smoke, or who must quires intervention. Without discussion of the is­ work and place their child in day care. The PIlTe1lf sues involved, the panel adopted a 20-dB level of Guide contains two recommendations that con­ hearing loss in the better ear as the threshold for flict with both the panel discussions and available intervention. All cbildren with OME for 4 to 6 data. First, the Pllre1lt Guide advises parents to months whose hearing loss is above this threshold "Try to keep your child away from playmates who are to receive tympanostomy tubes. are sick" to prevent OME. There is no evidence Yet this threshold is arbitrary and based on no that such isolation is useful; yet it can cause con- evidence. Its adoption encourages a very aggres-

314 JABFP July-August 1995 Vol. 8 No.4 the air conduction threshold was 5 dB; for those

sive approach to placing ear tubes. To put this J Am Board Fam Pract: first published as 10.3122/jabfm.8.4.305 on 1 July 1995. Downloaded from level in perspective, 25 to 30 dB is the level of the with OME treated without surgery, 7.8 dB; and whispered voice and 50 to 60 dB is the level of av­ for those that received car tuhcs, 12.0 dB. Thus erage speech. Vowel sounds are louder, conso­ ear tubes themselves might result in an average nants are softer. Twenty-five to 40 dB is classified long-term 5-dB loss, possihly caused by tyl11pano­ as mild hearing loss; 41 to 55 dB as moderate loss. sclerosis or other tympanic membrane changes. 22 For young children with permanent hearing loss, otolaryngologists might not recommend use of a Summary hearing aid unless the level ofloss in the better ear Ultimately, the physician making decisions re­ exceeds 30 to 35 dB. The panel's arbitrary choice garding the care of a child with OME must realize of a level 10 dB less as the criterion upon which to that evidence required for rational care of chil­ recommend surgery is unwarranted. dren with OME is not yet available. We do not Determination of the appropriate threshold for know the level or duration of OME-related hear­ action is of substantial practical importance. Few ing deficit, if any, that leads to clinically important studies have reported the distribution of hearing speech and language development impairment, acuity in children with OM£. Available studies, nor do we know the age and other characteristics however, all have reported hearing acuity for of children vulnerable to such problems. If such most children to be in the range of no or mild loss developmental problems do occur, we do not (25 to 40 dB). For example, a study by Fria, et al. lo know whether they are transient or permanent. used speech awareness threshold for infants 24 The panel included a disproportionate number of months or younger and speech reception thresh­ otolaryngologist members, whose recommenda­ old and pure tone audiology for children older tions might have been driven by an intervention­ than 24 months. They found that hearing acuity ist mind-set. Were the panel to have been com­ among children with OME did not change sig­ posed predominately of generalists and those with nificantly with duration of OME. Using the dis­ expertise in research methods, rather than repre­ tribution reported by Fria, et al., the panel's sentatives of disciplines that gain from an inter­ threshold will result in at least 56 percent of in­ ventionist approach, it is likely that the limited fants and 50 percent of children with OME per­ scientific information available on OME would sisting for 4 to 6 months being referred for sur­ have resulted in a different set of recommenda­ gery. If, instead, they used a 25 -dB referral rate tions. A careful reading of the Clinical Practice threshold, 41 percent of infants and 36 percent of Guideline, supplemented by additional sources, http://www.jabfm.org/ children would be referred; for a 30-dB threshold indicates that a less aggressive approach is reason­ this drops to 26 percent and 23 percent, respec­ able. The physician caring for a child with OME tively. For a 35 -dB threshold, referral for surgery might find that symptoms, associated quality-of­ decreases to 18 percent and 14 percent, and for a life issues, and family preferences (issues not con­ 40-dB level (the upper limit of "mild loss"), it de­ sidered by the panel) all appropriately play an im­ creases to 11 percent and 4 percent, respectively. portant role in determining treatment and that Thus, a shift upward of only 10 dB would de­ for the asymptomatic child with OME, no treat­ on 1 October 2021 by guest. Protected copyright. crease the number of children receiving ear tubes ment might be the most appropriate decision. by more than 50 percent. An increase of 15 dB re­ duces by more than two-thirds the number of References 1. Stool SE, Berg AO, Berman S, Carney C), Cooley children receiving ear tubes. JR, Culpepper L, et al. Otitis media with effusion in Fria, et al.lO studied children examined at an young children. Clinical practice guideline. Number ear, nose, and throat clinic and found that the av­ 12. AHCPR Publication No. 94-0622. Rockville, erage hearing level in the better ear was 27.5 dB. MD: Agency for Health Care Policy and Research, In comparison, a large Dutch community study of Public Health Services, US Department of Health children with OME, aged 2 to 4 years, reported and Human Services, July 1994. 2. Schappert SM. National Ambulatory Medical Clre an average hearing level of 20 dB (range 5 to 45 Survey: 1991, summary. Advance data. Vital and dB). Of note, this latter study assessed hearing on Health Statistics. Hyattsville, MD: US Dept. of 1 follow-up at age 7 /2 to 8 years. For those whose Health and Human Services, Public Health Service, ears had been free of OME at original assessment, 1992.

Otitis Media in Young Children ) 15 3. 'leele DW, Klein.10, Rosner B. Epidemiology ofoti­ for clinicians. AIICPR Publication No. ()4-0023. J Am Board Fam Pract: first published as 10.3122/jabfm.8.4.305 on 1 July 1995. Downloaded from tis Ill~di;l during the first se\'~n years of life in chil­ Rockville, M D: Agency for llealth Care Policy and dren in greater Boston: a prospective cohort study. Research, Public I Iealth Services, US Departm~nt of J Infect Dis JC)W); 160:H3-0..J. . llealth and Human Services, July 1994. ..J.. Brig'ht RA, Moore Riv!, ./eng LL, Sharkness C\1, H. Stool SE, Berg AO, Berman S, Carney C.J, Cooley tLl1nhurger SE, I familton PM. The prevalence of JR, Culpepper t, et <11. Middle ear fluid in young tyl1lP~lI1ostol1ly tubes in children in the United children. Parent guide. AHCPR Publication No. 04- States, 10HH. AmJ Public Health 1993; H3(7): 1026-1\. 062-+. Rockville, i\ID: Agency for Health Care Policy 5. Stool SE, Field tvlj. The impact of otitis media. Pe­ and Research, Public Health Services, US Depart­ dian'ic Tnfec Dis .11989; H( 1 Suppl):S 11-4. ment of Health and Human Services, July 1<)<)4. o. (;rllndfast K, Carney Cj. Otitis media in our society 15. Kaleida PH, Stool SE. Assessment of otoscopists' today. In: Crundfast K, Carney CJ. Ear infections in accuracy regarding middle-ear effusion. Am J Dis your child. I Jollywood, FL: Compact Books, 19H7 :0- Child 1902; 146:433-5. 19. 16. Maw AR. Using tympanoll1etry to detect glue ear in 7. CasselhranrML, BrostoffLM, Cantekin El, Flaherty general practice: overreliance will lead to overtreat­ MR, Doyle \VJ, Bluestone CD, et al. Otitis media ment. BMJ 1002; 304:67-R with effusion in preschool children. Laryngoscope 17. \'an I:lalcn FAM, de Melker RA. Validation of a port­ I 0H5; l)S:..J.2H-36. able tympanometcr for use in primary «lre. Tnt J 8. Kaleicla PI!, Casselbrant ML, Rockette HE, Para­ Pediatr Otorhinolaryngol 1<)<)4; 29:219-25. dise JL, Bluestone CD, Blatter M"'\Il, et al. Amoxicil­ 18. Lindholclt T. Ventilation tubes in secretory otitis me­ lin or myringotomy or hoth for acute otitis media: dia. A randomized, controlled study of the course, the results of a randomized clinical trial. Pediatrics 1<)91; complications, and the sequelae of ventilation tubes. H7:..J.(j(j-N. Acta OtoJaryngol (Stockh) I 9H3; 398(Suppl): 1-2H. <). Hendrickse "'TA, Kusmiesz H, Shelton S, Nelson]D. 19. ,'vIaw AR. Development of tympanosclerosis in chil­ Five vs. ten days of therapy for acute otitis media. Pe­ dren with otitis media with effusion and ventilation diatr Infect Dis J I <)8H; 7: 17 -23. tubes.J Laryngol Oto11991; 105:614-7. 10. Fria T], Cantekin EI, Eichler .TA. Hearing acuity of 20. Kleinman LC, Kosecoff], Dubois R\"~ Brook RH. children with otitis media with effusion. Arch Oto­ The medical appropriateness of tympanostorny bryngol I 9H5; 111:1 0-6. tuhes proposed for children younger than 10 years in II. Etzel RA, Pattishall EN, Haley NJ, Fletcher RH, the United States.JAA1A 199..J.; 171:1250-5. Henderson F\V. Passive smoking and middle ear ef­ 21. ",Tilliams RL, Chalmers TC, Strange KC, Chalmers fusion among children in day care. Pediatrics 1902; F'T~ Bowlin SJ. Use of antibiotics in preventing re­ <)0:22H-32. current acute otitis media and in treating otitis media 12. Rennie D. The Cantekin affair. JAhIA 1<)91; with effusion. A meta-analytic attempt to resolve the 266:3333-7. brouhaha. JA,\1A 1993; no: 1344-51.

13. Stool SE, Berg AO, Berman S, Carney C], Cooley 22. Zielhuis GA, Rach GH, van den Broek P. The natu­ http://www.jabfm.org/ JR, Culpepper L, et a!. c~Ianaging otitis media with ral course of otitis media with effusion in preschool effusion in young children. Quick reference guide children. Eur Arch Otolaryngol1990; 247:215-21. on 1 October 2021 by guest. Protected copyright.

316 JABFP July-August 1995 Vol. 81':0.4