A 15-Month-Old Child With Recurrent Otitis Media

DR REYNOLDS: Sam is a 15-month-old child who has recurrent otitis media. He lives near Boston with his mother, father, and a healthy older brother, aged 2 years 9 months. Sam is cared for at home while his parents are at work.

Born at term in October 2000 by elective repeat cesarean delivery to a healthy 37-year-old gravida 2, para 2 mother after an uncomplicated pregnancy, Sam had a normal hearing screen at birth. At the age of 5 months he had his first episode of acute otitis media, manifested by spontaneous perforation and draining of pus from the left ear. He was treated with amoxicillin, 80 mg/kg per day, and his eardrum was found intact at a subsequent visit.

In May, June (twice), August, September, and November of 2001, Sam had 6 more episodes of acute otitis without perforation. His pediatrician, Dr T, noted that the diagnoses were made based on pneumatic otoscopy showing pus behind an immobile tympanic membrane when the patient was symptomatic with cold symptoms, irritability, and/or fever. Sam was treated with amoxicillin, azithromycin, and amoxicillin-clavulanate. Revisits showed a normal tympanic membrane between each episode.

In early December 2001, Dr T discussed the relative merits and risks of watchful waiting, prophylactic antibiotics, and tympanostomy tube placement with Sam's parents. At the time of the interview, Mrs G, Sam's mother, reported that he had been taking prophylactic amoxicillin, 250 mg/d, for 1 month and had not had another acute otitis media episode.

Sam's past medical history is significant for a buckle fracture of his right radius and ulna. He has had normal growth and has achieved appropriate developmental milestones. His immunizations are up to date, including pneumococcal conjugate vaccine. His only medication is amoxicillin; he has no known drug allergies. He is not exposed to cigarette smoke. Family history is significant for first cousins on his paternal side with recurrent otitis media.

Physical examination showed a healthy 15-month-old boy. His weight was 22 lb 14 oz (25th percentile), length 31.5 in (50th percentile), and head circumference 19.75 in (95th percentile). His ear examination results were normal by pneumatic otoscopy bilaterally.

Sam has not had a hearing test since birth. His IgG level was borderline low when tested in September 2001 at 288 mg/dL (normal for age, 300-1500 mg/dL); his IgA level was also at the lower limit of normal.

His parents are frustrated by his recurrent ear infections and hope that the prophylactic antibiotics prevent future infections.

Sam really didn't have any medical issues until the ear infections started. I think the initial diagnosis was when he developed a cold. I brought him in and he was a little bit whiny, and the pediatrician looked in his ears and said that he had an ear infection too.

More often than not Sam probably caught the colds from my older son. But when Sam developed a cold, we knew if he had more beyond that because he might start flicking his ear a little bit, or he would wake up in the middle of the night. And he didn't do that when he was feeling well.

My husband and I really were adamantly against the tubes—there's no guarantee that they're going to work, and the other big negative for us was the anesthesia. But I guess there have been many cases where it has worked, and then the patient and the parents don't have to deal with any side effects of the antibiotics. Our concern as far as antibiotics goes is that Sam may develop some sort of resistance.

My son is now taking daily antibiotics. Thus far, it has worked very well. He's had no side effects. He appears to be fine, and so we've been happy with this route.

What I would want to see is some statistics as far as how successful tubes have been, the pros and cons, and the antibiotics. How long should the antibiotics be extended, or what's the usual length of time to ensure that he is not going to get the infections again?

Sam is basically a very healthy baby whose main problem has been frequent ear infections, which started when he was about 5 or 6 months old, and have continued until quite recently. He has had an ear infection about once a month, which of course is more frequently than the average.

Anytime Sam has had ear infections, he has been in a lot of pain, hard to soothe, and had difficulty sleeping, so it produces anguish on his behalf, and also is disruptive to everyone else in the family.

On balance, we leaned toward trying the prophylactic antibiotics first, because despite the numerous and obvious potential side effects, it would avoid surgery, which although a small operation, is still a big thing physically and physiologically, and also an extremely big thing emotionally for parents.

In my experience, parents tend to be mainly happy with ear tubes. I have had some patients in whom the tubes really seem to be a miracle, and there's not another ear infection from then on. In others, the ear infections keep coming.

What is the state of the art in terms of the evidence for outcomes of tubes? I'm aware of much of the literature related to tubes when they're used for otitis media with effusion, but I'm curious what the latest is about tubes for recurrent, acute infections.

How do you define acute or recurrent otitis media (OM)? How do you make the diagnosis? What is the epidemiology of otitis media? What are its risk factors? What is the bacteriology of and recommended treatment for OM? What workup should be done when a child has recurrent OM? What are the indications, benefits, risks, and evidence available to support treatment of recurrent OM with prophylactic antibiotics or tympanostomy tubes? What evidence supports watchful waiting? What is the recommended treatment for serous OM? Will the new pneumococcal vaccine prevent recurrent OM? What do you recommend for Sam?

DR PARADISE: Otitis media is a general term for middle ear inflammation. Infection is one of the causes of inflammation but not the only cause. The distinction between inflammation and infection is important here, because rational management of OM begins with a determination of whether the inflammation is infective or noninfective in nature.

The continuum of middle ear disease has 2 main components: acute otitis media(AOM) and otitis media with effusion (OME). Otitis media with effusion persists for weeks to months after AOM and consists of noninfective inflammation accompanied by effusion. Otitis media with effusion is also referred to as nonsuppurative, serous, or secretory OM. Otitis media with effusion appears to be the usual first form of OM that young infants experience. In one prospective study, of 742 first episodes of OM occurring during the first year of life, 139 episodes (18.7%) were diagnosed as AOM and 603 episodes (81.3%) as OME.¹ In the same study, of 228 episodes of AOM occurring in the first year of life, 107 (46.9%) developed in infants who had had OME at the preceding visit.¹ Therefore, one wonders whether Sam was found to have middle ear effusion at any visit before his first episode of AOM. Distinguishing between AOM and OME on clinical grounds is usually straightforward, but because each condition may evolve into the other without any surely differentiating physical findings, any schema for distinguishing between them is to some extent arbitrary.²

Accurate diagnosis of either form of OM in infants and young children is often difficult. In AOM, symptoms are highly variable. Most characteristic is ear pain, often manifested by unaccustomed holding or tugging at the ear. Fever, irritability, or other nonspecific symptoms may be present, or there may be no symptoms.³ In OME, symptoms, if any, are usually inapparent, and the associated conductive hearing loss usually goes undetected. On otoscopic examination, the eardrum is often obscured by cerumen, removal of which may be arduous and time-consuming.⁴ Abnormalities of the eardrum may be subtle and difficult to appreciate,⁴ and both underdiagnosis and overdiagnosis are common. Thus, in one study using videotaped images of pneumatic otoscopic examinations, the average proportion of correct diagnoses by pediatricians was 50% (range, 25%-73%) and by otolaryngologists was 73% (range, 48%-88%).⁵ However, other studies of the diagnostic accuracy of pneumatic otoscopy as performed by pediatric clinicians have given more favorable results: one study found sensitivity and specificity of 92% and 83%, respectively.⁶ In another study, diagnostic accuracy ranged from 85% to 97%.⁷ Accordingly, in pediatric practice settings, the gold standard for the diagnosis of OM is the appearance of the eardrum on pneumatic otoscopy. Diagnosing AOM usually begins with determining the presence or absence of middle ear effusion. Physical signs of middle ear effusion consist of at least 2 of 3 eardrum abnormalities: white, yellow, amber, or (rarely) bluish discoloration; opacification other than due to scarring; and most tellingly, decreased or absent mobility. Alternatively, in OME, either air-fluid levels or air bubbles may be visible behind the eardrum.^{2 ,4} Rupture of the eardrum with purulent otorrhea, as in Sam's case, is a sure sign of AOM, but rupture is more the exception than the rule. Rarely, intense erythema of the eardrum alone, in conjunction with unequivocal ear pain, may justify a diagnosis of AOM.

To support a diagnosis of AOM in a child with middle ear effusion, acute inflammation should also be present. This usually is manifested by distinct fullness or bulging of the eardrum, but in occasional cases marked erythema or ear pain that seems clinically important may precede the development of bulging.^{2 ,4 ,8} Mild or moderate erythema of the eardrum does not justify a diagnosis of AOM in a child with middle ear effusion because the erythema may result from vigorous crying or from instrumentation of the external auditory canal. Similarly, associated nonspecific symptoms such as fever, irritability, anorexia, vomiting, and diarrhea do not justify a diagnosis of AOM.

In contrast to the findings in AOM, eardrum findings in OME are less distinctive. Erythema and fullness are absent or slight; often the eardrum is retracted. Amber discoloration and opacification may be subtle and difficult to appreciate. As noted earlier, the most telling sign is impaired or absent eardrum mobility. Ancillary diagnostic methods such as tympanometry⁹ (sensitivity, 65%-90%; specificity, 48%-91%^{10 - 11} ) and spectral gradient acoustic reflectometry¹¹ (sensitivity, 67%; specificity, 87%¹¹ ) may be helpful in clarifying questionable otoscopic findings and in serving as validators—or invalidators—of otoscopic diagnoses. However, after a second look, the final arbiter must be the clinician's otoscopic diagnosis.

The definition of recurrent AOM is arbitrary. A criterion used in many studies is the occurrence of 3 episodes in 6 months or 4 episodes in 1 year.¹² Young age and season of the year are the most important risk factors; the peak incidence and prevalence of OM occur during the first 2 years of life, particularly between 6 and 12 months of age.^{7 ,13 - 14} In a study in Pittsburgh, Pa, in which my colleagues and I followed up 2253 infants until 2 years of age as part of a larger, ongoing study of child development in relation to early-life OM (CD/OM study), the proportions of infants developing 1 or more episodes of OM within the first 6, 12, and 24 months of life, respectively, were 48%, 79%, and 91%, and the mean cumulative proportion of days with OM was 20% in the first year of life and 17% in the second.⁷ In general, the earlier in life a child experiences a first episode, the greater the degree of subsequent difficulty as measured by frequency of recurrence, severity, and persistence of middle ear effusion.^{3 ,15 - 16} First occurrence at 5 months of age, as in Sam's case, is not unusual: in the CD/OM study, 24% of suburban infants and 41% of inner-city infants had AOM or OME during their fifth month of life.⁷ Having recurrent episodes of AOM during the summer months, as Sam had, however, is rather unusual; OM is mainly a cold-weather disease.^{17 - 19}

From the CD/OM study,⁷ the principal risk factors for recurrent and persistent OM other than young age and season of the year appear to be low socioeconomic status and exposure to large numbers of other children, whether at home or in day care. For example, the mean cumulative percentage of days with middle ear effusion during the first year of life was 30.0% in 153 infants who had Medicaid health insurance and whose mothers had not completed high school compared with 16.0% in 380 infants who had private health insurance and whose mothers had graduated from college. Corresponding values were 31.1% in 119 infants who were in day care with 5 or more other children compared with 14.6% in 712 infants who had no regular exposure to other children.⁷ Neither low socioeconomic status nor exposure to large numbers of children characterized Sam's case, although he did attend play groups for short periods 3 days each week, and his colds often followed on the heels of his older brother's colds. Less salient risk factors for OM include genetic background, male sex, formula feeding in early infancy, and, perhaps, the degree of exposure to household smoking.⁷

That genetic factors play a role is indicated by the greater degree of concordance for the occurrence of OM between monozygotic twins or triplets than between dizygotic twins or triplets. In a recent study, estimates of discordance for more than 3 episodes of middle ear effusion in the first 2 years of life were 0.04 in monozygotic twins or triplets vs 0.37 in dizygotic twins or triplets (P = .01). Discordance for more than 1 episode of AOM was 0.04 in monozygotic sets vs 0.49 in dizygotic sets (P = .005).²⁰ Sam has first cousins who had recurrent OM, but the same may be true of most children. We do not know whether Sam was breastfed, but we know that he was not exposed to tobacco smoke. Finally, OM is highly prevalent in children with certain congenital abnormalities, notably, cleft palate²¹ and Down syndrome.²²

Pathogenic bacteria can be isolated from middle ear exudate in approximately 75% of cases of AOM; in the remaining cases culture shows either no growth or the presence of nonpathogens.^{13 ,23 - 25} Three pathogens predominate: Streptococcus pneumoniae in approximately 30% to 50% of cases, nontypeable Haemophilus influenzae in approximately 25% to 40%, and Moraxella catarrhalis in approximately 10% to 15%.^{13 ,23 - 25} Respiratory viruses (or virus-derived RNA), especially rhinovirus and respiratory syncytial virus (RSV), also may be found, often in association with pathogenic bacteria.^{26 - 28} Whether viruses alone cause AOM is unknown. However, viruses almost certainly set the stage for bacterial invasion,²⁹ and they may also amplify the inflammatory process and interfere with the resolution of bacterial infection.³⁰

In children with OME, middle ear fluid cultures are usually sterile, but middle ear pathogens are recoverable in approximately 30% of cases.^{31 - 34} In recent studies using polymerase chain reaction assay, bacterial DNA was found in 77% and 95% of cases,^{33 ,35} and viral RNA in 30% of cases.³⁴ Whether these findings constitute evidence of active infection or "fossilized remains" is unclear.³⁶

Whether to Treat. Treatment of AOM with antimicrobials has long been routine in the United States and many other countries. However, because the increasing prevalence worldwide of infection caused by multiple drug–resistant S pneumoniae has been linked to antimicrobial use,^{37 - 41} and because the estimated overall rate of spontaneous resolution of AOM is 78% within 4 to 7 days,⁴² some authors have recommended withholding antibiotics entirely in some or all cases of AOM unless symptoms persist or worsen.^{43 - 44} Nonetheless, the policy of initially withholding antimicrobial treatment seems to me injudicious, not only because middle ear infection subsides more promptly and more consistently with antibiotic treatment than without,^{44 - 46} but also because the striking decline in the occurrence of mastoiditis and other suppurative complications of OM over the past half-century appears attributable, at least in part, to the widespread treatment of AOM with antimicrobials.⁴⁷ In particular, I believe that antibiotics should not be withheld initially from children with AOM who are younger than 2 years or who appear to have severe infections, because these are the groups in whom mastoiditis occurs most often,⁴⁷ in whom antibiotic treatment almost certainly gives better results than placebo,^{44 - 45} and in whom recurrent AOM is particularly likely to develop.⁴⁸ I also believe that antibiotics should not be withheld initially from children with AOM who have recent histories of recurrent AOM and are prone to develop further recurrences¹⁵ or from children who may not get satisfactory follow-up. It seems to me that much greater overall reductions in antibiotic use could be achieved, and achieved more rationally, if antibiotics were withheld from children with viral upper respiratory tract infections and children with OME, rather than from children with bona fide AOM. Admittedly, initially withholding treatment in milder cases of AOM in older children will deserve consideration if the prevalence of bacterial resistance continues to escalate.

Choice of Antimicrobial Drugs for First-Line Treatment. Dr T treated most of Sam's episodes of AOM with amoxicillin. Currently, amoxicillin remains the drug of first choice for uncomplicated AOM because of its excellent record of safety, relative efficacy, and low cost.⁴⁹ In particular, amoxicillin is the most efficacious of available oral antimicrobial drugs against both penicillin-susceptible and penicillin-nonsusceptible strains of S pneumoniae.⁴⁹ Increasing the dosage from 40 mg/kg to 80 to 100 mg/kg per day, as Dr T did in treating Sam, will provide efficacy against penicillin-intermediate and some penicillin-resistant strains.⁴⁹ A limitation of amoxicillin stems from its inactivation by the β-lactamases produced by many strains of nontypeable H influenzae and most strains of M catarrhalis. Fortunately, evidence suggests that episodes of AOM caused by these pathogens have overall rates of spontaneous resolution of approximately 50% and 75%, respectively.⁵⁰ For children who are definitely or possibly allergic to β-lactams, an appropriate alternative first-line drug is azithromycin suspension, which is palatable, is administered only once daily, and has an excellent safety record.⁴² For children in whom palatability or convenience of administration are of overriding importance, either azithromycin or cefdinir (discussed later as a second-line drug) would be suitable.⁵¹ Use of a second-line drug for initial treatment would also be appropriate, in my judgment, for children who develop AOM while receiving antimicrobial prophylaxis (also discussed later); for children who are immunocompromised; and for children with severe symptoms who recently had severe, protracted or repeated episodes of AOM.

Duration of First-Line Treatment. Treating children with AOM for the usual 10 days may be unduly long for some children but not long enough for others. Findings in a large observational study suggest that short-course treatment at standard dosage levels, such as 3 or 5 days, may not be adequate in children who have had an episode of AOM within the preceding month.⁵² Additionally, randomized clinical trials indicate that short-course treatment will often prove inadequate in younger children (particularly those younger than 2 years) but may well be adequate in older children.^{53 - 54} My clinical experience is in keeping with those findings and suggests further that treatment is often required for longer than 10 days to eradicate infection in children who are very young or are having repeated or severe episodes.

Choice of Antimicrobial Drugs for Second-Line Treatment. Drugs chosen for second-line treatment should be effective against β-lactamase–producing strains of H influenzae and M catarrhalis as well as against susceptible and most nonsusceptible strains of S pneumoniae. A working group recently assembled by the Centers for Disease Control and Prevention concluded that, based on available evidence, only 3 drugs met that requirement: amoxicillin-clavulanate, cefuroxime axetil, and intramuscular ceftriaxone.⁴⁹ High-dose amoxicillin-clavulanate is particularly well-suited as a second-line drug for treating AOM because high-dose amoxicillin is the most effective orally administered antimicrobial drug against most strains of S pneumoniae⁴⁹ and the addition of clavulanate extends its effective antibacterial spectrum to include β-lactamase–producing bacteria. Unfortunately, amoxicillin-clavulanate remains quite expensive and sometimes causes troublesome diarrhea. The currently available suspension of cefuroxime axetil is not palatable and its acceptability is low.⁵⁵ However, the liquid preparation of cefdinir, which has a spectrum of in vitro efficacy similar to that of cefuroxime,^{56 - 57} is quite palatable.⁵⁵ Ceftriaxone treatment entails both the pain of intramuscular injection and substantial cost, and the injection may need to be repeated once or twice at 2- or 3-day intervals to achieve the desired degree of effectiveness.⁵⁸ Nonetheless, use of ceftriaxone is appropriate in severe cases of AOM when oral treatment is not feasible, after treatment failure using orally administered second-line antibiotics, or when middle ear aspirates yield highly resistant S pneumoniae.^{49 ,58} For children with AOM who have not responded favorably to first-line treatment, I know of no adequate study that has shown clinical superiority of one antibiotic over another. Accordingly, the choice of drugs is best guided by available in vitro, pharmacokinetic, and pharmacodynamic data, and also, because bacterial resistance patterns vary geographically, by personal clinical experience. In children with persistent infection despite seemingly appropriate second-line treatment, microbiological findings in middle ear exudate obtained by tympanocentesis⁵⁹ or by myringotomy may help guide subsequent management. The duration of second-line treatment is best guided by the child's clinical course.

Children with recurrent OM who are otherwise healthy, such as Sam, require no laboratory investigation (except, as noted, microbiological study of middle ear exudate in the event of persistent infection). Sam's IgG level of 288 mg/dL at age 11 months is not low enough to cause concern. A deficiency of IgA is found in some children with recurrent AOM but also is found not infrequently in children without recurrent AOM.⁶⁰ Selective IgG subclass deficiencies may be found in children with recurrent AOM who also have recurrent sinopulmonary infection,⁶¹ but children without infections at sites other than the ear rarely have a readily identifiable immunological deficiency.^{62 - 63} Investigation of these children's serum immunoglobulins is therefore not likely to be productive. Nonetheless, various lines of evidence point to the existence of more subtle immune deficits in many children with recurrent AOM.^{64 - 67}

Unlike laboratory studies, hearing testing should be performed in any child, with or without recurrent AOM, in whom middle ear effusion persists for 3 months or longer without improvement or whose language development or hearing acuity is at issue.⁶⁸ Sam had a normal newborn hearing screen, which provides considerable reassurance (but would not have identified congenital hearing impairment that first becomes detectable after the newborn period⁶⁹ ). More importantly, Sam's ear examination results are now normal, and his otherwise concerned parents have not indicated concern about his hearing. Testing his hearing at this point would therefore seem unnecessary.

Antimicrobial Prophylaxis. Antimicrobial prophylaxis offers variable protection against recurrences of AOM. A meta-analysis of studies using various antimicrobials indicated an overall reduction, with a rate difference of 0.11 episodes per patient-month during the period of prophylaxis, and a rate difference of 0.25 episodes per patient-month in the subgroup of studies in which control children continued to have a relatively high recurrence rate (ie, ≥0.2 episodes per month).⁷⁰ However, because antimicrobial use is a main contributor to bacterial resistance, the risks of sustained prophylaxis now seem to generally outweigh the likely benefits, particularly for children in day care who in any case are at increased risk of colonization with multiple-drug–resistant S pneumoniae. For example, in one study, 24 of 54 children in day care (44%) were colonized with highly or intermediately resistant strains.⁷¹ In another study, day care attendance was associated with an odds ratio for colonization of 2.82.⁷² Prophylaxis may nonetheless be an appropriate option for the child with recurrent AOM who is cared for at home and is usually away from other young children. If prophylaxis is undertaken, the drug of choice should generally be amoxicillin because of its much more favorable efficacy profile than that of sulfonamides against multiple-drug–resistant S pneumoniae,⁴⁰ and the duration should be limited to the cold-weather season at most.

Tympanostomy Tubes. Tympanostomy tube insertion is effective (as long as tubes remain in place and patent) in reducing recurrence rates of AOM both in children with persistent OME^{73 - 74} and in children like Sam who have recurrent AOM but not persistent OME.¹² However, in each of these studies the reduction was offset by the occurrence of episodes of purulent otorrhea through the tympanostomy tubes. For example, in one of the studies, a rate difference of 0.72 episodes of AOM per patient-year was found during the first year after tube insertion, but when episodes of tube otorrhea were included, the overall rate difference was 0.42 episodes per patient-year.⁷⁴ On the other hand, episodes of tube otorrhea are usually unaccompanied by pain or fever and are generally less troublesome than episodes of AOM. Furthermore, anecdotal evidence and personal clinical experience suggest that in children more severely affected than many of those enrolled in the studies cited, tube insertion offers a greater degree of benefit than that reported.

Benefits and Costs of Tubes. The benefits of tympanostomy tubes include the medical and nonmedical costs saved by preventing recurrent episodes of OM. Nonmedical costs stem mainly from lost work, but also from day care, babysitting, and transportation. For an episode that involved only one visit, recent estimates of average medical and nonmedical costs were $132 and $131, respectively, increasing to $331 and $458 for an episode that involved more than 1 visit.⁷⁵ Potential savings from episodes prevented by tube placement must be balanced against the operation's cost, currently about $3000 in Pittsburgh, and certain risks.^{76 - 79} Common complications include obstruction of the tube lumen, secondary infection with tube otorrhea, and premature extrusion. In children receiving tubes in the CD/OM study, tube otorrhea occurred at least once within 12 months in 75% of the children and within 18 months in 83%.⁷⁹ Sequelae following tubal extrusion are also not uncommon. In various reported series, residual perforation of the eardrum has been noted in about 2% of ears; tympanosclerosis in 30% to 50% of ears; atrophic scarring, atelectasis, retraction, or retraction pocket in 25% to 50% of ears; and cholesteatoma in 0.7% of ears.^{78 ,80 - 81} Whether hearing loss may develop or be exaggerated as a late-in-life consequence of tube insertion remains speculative. Recurrence of OM following the extrusion of tubes is common, especially in younger children. In our Pittsburgh CD/OM study,⁸² 340 children received tubes during their first 3 years of life. Of these, 243 (71.5%) developed a recurrence of OM within 2 years after tubal extrusion, 26 (7.6%) developed no recurrence within 2 years after extrusion, and 71 (20.9%) either had tubes in place when last examined or developed no recurrence during postextrusion follow-up periods of less than 2 years.

Although Sam did not have persistent OME, many children do, and its management has generated much controversy.⁶⁸ Except in the occasional child who seems troubled by the usual mild to moderate conductive hearing loss (in the CD/OM study cited earlier, hearing loss was present in approximately half of instances in which children had unilateral effusion and in approximately three quarters of instances in which they had bilaterial effusion⁸² ), the main reason to consider treatment for OME is to prevent known or possible complications and sequelae.

Complications and Sequelae of OME. A common complication of OME is the development of AOM. As noted earlier, in the CD/OM study, AOM often developed in infants who had had OME at the preceding visit.¹ Infrequent sequelae of OME include pathological middle ear changes, including tympanic membrane retraction pockets, adhesive otitis, ossicular discontinuity, and cholesteatoma,^{83 - 84} and perhaps, cochlear damage with resulting sensorineural hearing loss.⁸⁵ The extent to which such sequelae result from OME specifically, rather than from repeated or chronic middle ear infection, seems unclear. That persistent early-life OME can cause lasting impairments of speech, language, cognitive, or psychosocial development has been reported by many investigators, but supporting evidence has been weak.⁸⁶ Results in the CD/OM study, involving 6350 healthy infants enrolled within the first 2 months of life and monitored regularly for middle ear effusion, suggest that effusion during the first 3 years of life that persists continuously for 90 days bilaterally or 135 days unilaterally, or intermittently for specified proportions of longer periods, poses no measurable developmental risk at the age of 3 years.⁸²

Antimicrobial Treatment. Antimicrobials have definite but limited efficacy in resolving OME, presumably because they help eradicate nasopharyngeal infection, inapparent middle ear infection, or both. A meta-analysis of randomized controlled trials, in which a successful outcome was defined as complete resolution of effusion in all affected ears at the first posttreatment assessment, indicated an average difference in rates of successful outcome of 23%, ranging from 14% in populations with OME of relatively short duration and a high natural cure rate to 31% in populations with persistent OME and a low natural cure rate.⁸⁷ However, the benefit of antimicrobial treatment is often short-lived; in one trial, among participants whose effusion had cleared within 4 weeks, approximately half had a recurrence of effusion within the subsequent 3 months.⁸⁸ For this reason, and because of concern about bacterial resistance, routine antimicrobial treatment of OME, recommended as an appropriate option in official guidelines issued in 1994,⁶⁸ soon no longer seemed reasonable,⁸⁹ and has been specifically recommended against in the most recent consensus review.⁹⁰ However, in my judgment, antimicrobial treatment is indicated for children with OME who have purulent nasal discharge persisting without improvement for 2 weeks or longer, suggesting that the children have an associated bacterial upper respiratory tract infection, most often sinusitis,⁹¹ or in children in whom, because of the persistence of OME, tympanostomy tube insertion is being considered. In the latter circumstance, in my experience, OME may sometimes resolve after a 2- to 4-week course of second-line antibiotic treatment, an observation in keeping with the more favorable treatment effect of antimicrobials in children whose OME is long-standing.⁸⁷

Tympanostomy Tube Insertion. Once undertaken, tympanostomy tube insertion is effective in reducing children's proportion of time with middle ear effusion during immediately succeeding periods. For example, in the clinical trial component of the CD/OM study, in which children were randomized to receive tubes either promptly or after a specified period, the mean proportion of time with middle ear effusion during the first 12 months postrandomization was 29% in the early-treatment group compared with 48% in the late-treatment group, and the percentage of children in the late-treatment group who had effusion more than 50% of the time was approximately 3 times the percentage in the early-treatment group.⁸² However, tubes are effective only as long as they remain in place and patent, and their tenure, while varying from weeks to months, averages about 1 year.⁷⁹

Tympanostomy tube insertion is the appropriate treatment for a child with persistent OME who is discomfited by the associated hearing loss; who has an unusual but possibly related symptom, such as otalgia, tinnitus, or disturbance of balance; or whose tympanic membrane shows a deep retraction pocket.⁹² In children without those conditions, given the uncertainties concerning possible consequences of long-standing OME and the risks and cost of tube insertion, the point at which the operation should be considered remains debatable. Factors militating in favor of tube insertion would include a child's questionable or suboptimal language development for age, an unfavorable history regarding the frequency and severity of episodes of AOM, a history of adverse reaction to multiple antimicrobial drugs, and regular exposure to large numbers of other children. Factors militating against tube insertion would include good language skills, normal or near-normal auditory acuity, limitation of effusion to one ear, a history relatively free of AOM, and, of course, parental reluctance concerning surgical intervention. Taking these variables importantly into account, a reasonable approach, in my judgment, would be to consider tube insertion after 6 to 12 months of continuous bilateral OME or 9 to 18 months of continuous unilateral OME, provided that second-line antibiotic treatment just before the time of decision had proven ineffective. Importantly, because even persistent OME usually clears spontaneously during the summer, watchful waiting through at least 1 summer season is advisable in all children with OME who are otherwise well.

Recurrence of OM After Tympanostomy Tube Insertion. In children who have already undergone tube insertion and who subsequently develop troublesome OM, adenoidectomy is likely to afford a measure of benefit. For example, in our randomized clinical trial of adenoidectomy, during the first and second years of follow-up, respectively, children who received adenoidectomy had 47% and 37% less time with OM than control children and 28% and 35% fewer episodes of AOM.⁹³

Sam received heptavalent pneumococcal conjugate vaccine in timely fashion. In 2 recent studies the vaccine reduced the number of episodes of AOM overall by only 7% and 6%, respectively.^{94 - 95} However, in 1 of the studies,⁹⁴ the vaccine reduced the number of episodes by 9% to 23% in children with histories of frequent recurrences and reduced the number of children undergoing tympanostomy tube insertion by 20%.⁹⁴

Importantly, the serotypes represented in the vaccine have thus far generally been the ones most likely to exhibit antibiotic resistance.^{96 - 97} Whether the vaccine serotypes currently prevalent will undergo mutation, whether pneumococcal AOM will be caused increasingly by nonvaccine serotypes, and whether other organisms will move in to take the place of S pneumoniae as the dominant AOM pathogen can only be determined with time.

From the taped interview with Sam's mother, I gained the impression that, although concerned about Sam's ear infections, she wasn't feeling overwhelmed or desperate. In referring to Sam's behavior during his episodes she used terms such as "a little bit whiny," "flicking his ear a little bit," and "would wake up in the middle of the night." She did not refer to inconsolable crying or high fever or other possible manifestations of severe infection. On the other hand, Dr T's comments about Sam referred to "a lot of pain, hard to soothe, had difficulty sleeping . . . anguish on his behalf and also disruptive to everyone else in the family." If after questioning Sam's parents directly it seemed to me that Sam's episodes were not extremely troublesome, and that his parents could weather the storm reasonably well, I would recommend what I consider the safest of the 3 options posed by Dr T, namely, antibiotic treatment of individual episodes as they occur. This option avoids the risks both of prophylaxis and of tubes. If, on the other hand, parental endurance, patience, and equanimity seemed severely taxed, I would first try amoxicillin prophylaxis in light of Sam's limited exposure to other children and reserve tube insertion as a last resort. In either case, I would recommend that Sam receive influenza vaccine, since several studies have found a limited protective effect of the vaccine against recurrent AOM.^{98 - 101}

A PHYSICIAN: Could you comment on the conflicting data about long-term consequences of OME with regard to language development and learning disabilities?

DR PARADISE: Many studies have found relationships between persistent early-life OM and later impairment of speech, language, cognitive, or psychological development, whereas other studies have not.⁸⁶ All of the studies were associational in nature, ie, they considered whether children with more OM early in life ended up with less favorable developmental outcomes, or conversely, whether children with less favorable developmental outcomes had had more OM early in life. Setting aside methodological limitations of many of the studies, the problem remains that associational research does not fully address the issue of causality¹⁰² ; many factors might conceivably predispose children both to OM and to developmental impairment. To address causality directly, we incorporated a randomized clinical trial in the CD/OM study. Of the 6350 infants initially enrolled, we randomized 429 with persistent middle ear effusion during their first 3 years of life to have tympanostomy tubes inserted either promptly or, if effusion persisted, after 6 or 9 months depending on whether the effusion was bilateral or unilateral. At age 3 years there were no differences between the early- and the late-treatment group in measures of speech, language, cognition, or psychological development.⁸² Those results, combined with data from other children in the study, suggest strongly, as I noted earlier, that antecedent middle ear effusion within the duration limits we studied poses no developmental risk at the age of 3 years. The study is ongoing, and we expect to report further results in the children at later ages.

Two other randomized trials have been undertaken since ours began.^{103 - 104} To the extent that the trials were similar to ours, their results were similar, but in those trials, among other limitations, developmental testing was considerably less comprehensive and follow-up was shorter.

A PHYSICIAN: I have found that many parents are receptive to holding off on antibiotics. I give them a prescription but say try to wait, and that works very well.

DR PARADISE: I am reluctant to withhold antibiotics initially from children with bona fide AOM. Although, admittedly, most children will recover without treatment, my experience is in keeping with the evidence cited earlier^{44 - 46} that children get better faster and more surely with than without antibiotics. It seems to me that the bulk of antibiotic misuse regarding OM consists in treating and retreating children who have OME rather than AOM.

A PHYSICIAN: Otolaryngologists have exerted pressure on pediatricians to avoid prophylactic antibiotics. A study some years ago suggested that using antibiotic prophylaxis only during colds could stop ear infections in a selected population. Do you have any comment on that practice?

DR PARADISE: I believe I know that study,¹⁰⁵ but in 2 other studies the benefit was equivocal.^{106 - 107} So if I were going to prescribe prophylaxis—which I would consider only in children who are not exposed to large numbers of other children—I would opt for antibiotics continuously, both because of those study results and because, in my experience, many ear infections begin before there is other evidence of upper respiratory tract infection.

A PHYSICIAN: I am intrigued about the duration of antibiotic treatment. You mentioned that you could treat the older kids for a shorter time. Is the variable here permeability of the antibiotic or sterilization of the fluid?

DR PARADISE: I think the main variables are immunologic response and eustachian tube structure and function. The pharmacodynamics probably do not differ much between younger and older children. Older children are more resistant to infection in general. Their immune systems are better developed¹⁰⁸ and they generally have less nasopharyngeal lymphoid tissue.¹⁰⁹ Also, their eustachian tubes are less horizontally disposed and less compliant.¹¹⁰ Finally, in older children, who are upright more of the time, gravity may help promote drainage from the middle ear cavity.

MRS G, SAM'S MOM: To date, the antibiotics that we have been using for my son have been very effective and we have been very happy with that and have weathered through the winter months very well. But how long do we continue? Do we go until his second birthday when we say well, it looks like the chances of him getting recurring ear infections are less, or where do we go from here?

DR PARADISE: I would stop prophylaxis around the first of June because we know that the incidence of AOM drops dramatically during the summer months.^{17 - 19}