When amoxicillin fails

GUIDE FOR PARENTS

By Richard Linsk, MD, PhD, Janet Gilsdorf, MD, and Marci Lesperance, MD

Resistance is only one reason acute otitis media may fail to respond to a first course of amoxicillin. In deciding what to do next, you'll need to take the other possibilities into account as well.

Otitis media is one of the most common conditions diagnosed in pediatric practice and one of the most controversial. One survey of 165 pediatricians reported 147 different sets of diagnostic criteria for acute otitis media (AOM).¹ It would not be surprising if a comparable number of approaches to treatment were being pursued in clinical practice. No one approach is ideal or applicable to all clinical situations. Rather, AOM is a highly variable condition, resistant to clinical study, and unpredictable in any given patient.

Several excellent reviews and commentaries on AOM have been published recently.^2­6 These reviews conclude that antibiotic therapy is indicated for symptomatic AOM, particularly in younger children, and that the agent of choice for initial treatment is amoxicillin. An issue not fully explored in these commentaries is what to do when amoxicillin fails.

As with AOM itself, the diagnosis and management of AOM that does not respond to amoxicillin are complex. No consensus exists as to what constitutes amoxicillin failure. We define it as the persistence or rapid recurrence of otitis media symptoms, such as otalgia, irritability, sleeplessness, anorexia, or fever, together with physical findings of an inflamed tympanic membrane and middle ear effusion, in a patient who has received an appropriate dose of amoxicillin for at least 72 hours. An appropriate dose of amoxicillin is 40 to 60 mg/kg/d divided bid for a child at low risk of colonization with penicillin-nonsusceptible Streptococcus pneumoniae (PNSSP), and 80 to 100 mg/kg/d divided bid for a child at high risk of PNSSP. Risk factors for carriage of PNSSP include recent use of antibiotics, age under 2 years, and group day-care attendance.^7.8

Likely causes of amoxicillin-unresponsive AOM include infection caused by amoxicillin-resistant bacteria, inadequate dosing or absorption of amoxicillin, poor penetration of amoxicillin into the middle ear space, reinfection with a second organism, and AOM caused by viral infection or viral and bacterial co-infection. Finally, symptoms of other illnesses such as viral pharyngitis may lead to discovery of an otherwise asymptomatic middle ear effusion, creating the mistaken impression that the systemic symptoms are caused by amoxicillin-unresponsive AOM.

In dozens of clinical trials with both narrow- and broad-spectrum antibiotics, 5% to 20% of children with an initial diagnosis of AOM have been scored as treatment failures. The exact percentage depends on the population studied and the criteria used. In placebo-controlled studies 70% to 90% of patients improve without therapy within five to seven days. In many of these patients, antibiotics probably would have hastened the resolution of symptoms. An effect of therapy is more likely to be demonstrated in younger children; younger children are also most likely to have antibiotic-unresponsive disease.⁹ Thus, the impact of antibiotic therapy on the initial treatment of AOM is real but marginal. Less clear is the extent to which further therapy or changes in therapy affect the outcome of those 5% to 20% of patients who do not improve initially with antibiotics.

The natural history of AOM

Surprisingly, clinical outcome of antibiotic therapy for AOM correlates poorly with in vitro antibiotic sensitivity of the causative organism. Most experimental evidence demonstrates that clinical outcome is determined chiefly by host and pathogen factors and that antibiotic therapy plays a smaller role. In general, antibiotic therapy has little effect on long-term outcome of AOM, such as clearance of effusion. The high rate of spontaneous resolution and the frequent occurrence of recalcitrant AOM have made it almost impossible to demonstrate any significant difference between different antibiotic therapies in clinical trials.

Although many recent commentaries have focused on the impact of antibiotic resistance on clinical failure, resistant organisms do not account for all of the variability in treatment outcomes. In a study of tympanocentesis cultures from 137 patients with symptomatic, antibiotic-unresponsive AOM, only half of the specimens grew pathogenic bacteria, and more than half of these had persistent bacteria sensitive in vitro to the antibiotic first used (see Table 1).¹⁰ Thus, clinical failure could be attributed to antimicrobial resistance in only about one quarter of the cases. In fact, the pattern of bacteria isolated from persistently symptomatic antibiotic-treated ears was not substantially different from that of untreated patients. Bacteria that have been detected in middle ear cultures from children with persistent AOM include Haemophilus influenzae (both b-lactamase positive and negative), Moraxella catarrhalis, Staphylococcus aureus, group A streptococci, and various gram-negative species.^10­13 By far the most common organism isolated is S pneumoniae. In some series, up to 60% of S pneumoniae middle ear isolates are highly resistant to penicillin (minimum inhibitory concentration greater than or equal to 2 mg/mL).¹⁴ The relative increase in antimicrobial resistance among antibiotic-unresponsive isolates indicates that antimicrobial resistance does contribute to antibiotic failure. If current trends continue, multiple drug-resistant S pneumoniae (DRSP) is likely to be found even more frequently in this population.

One factor that makes the clinical evaluation of antibiotic therapy difficult is viral infection and co-infection. In a clinical series of 271 patients, bacteria were isolated in 68% of cases of symptomatic AOM. Half of these patients had a concomitant viral infection. Viral infection of the middle ear without bacteria was detected in 12% of patients. No pathogenic cause was identified in 20%. In this study, middle ear effusions were not tested for the presence of atypical organisms.¹⁵ Symptomatic and bacteriologic failure were more common in patients with viral or bacterial co-infection despite treatment with appropriate antibiotics. Younger age was also a risk factor for treatment failure in this study. Viral co-infection appears to interfere with antibiotic penetration into the middle ear.¹⁶ This phenomenon may account in part for the frequent persistence of antibiotic-sensitive bacteria in patients being treated with appropriate antibiotics. If it does, then these bacteria might later respond to the same antibiotics originally administered after the viral or inflammatory process is resolved.

The tendency of many bacteria to clear without antibiotic therapy adds to the difficulties of interpreting the results of trials. In placebo-controlled studies, 50% of H influenzae and 20% of S pneumoniae cleared in two to seven days without antibiotic therapy. Furthermore, symptomatic improvement is an imperfect measure of bacteriological response. In a series of randomized trials using tympanocentesis before and during antibiotic therapy, 10% of patients judged clinical successes had persistent, viable bacteria isolated from the middle ear while 53% of patients judged clinical failures had no remaining bacterial infection.¹⁷ In patients with OME, 70% have viable bacteria without clinical signs of infection.¹⁸ Thus, clinical efficacy does not demonstrate bacteriologic efficacy, and conclusions about antibiotic efficacy based solely on clinical response should be viewed with some suspicion.

Chlamydia pneumoniae is an atypical pathogen commonly associated with upper respiratory infections in children. It differs from the otitis pathogens just described in that it is relatively insensitive to
b-lactams but, for now at least, remains sensitive to macrolides. In fact, C pneumoniae has been recently detected by culture or polymerase chain reaction in up to 8% of middle ear specimens.¹⁹ However, it is also commonly found in nasopharyngeal specimens from well preschoolers.²⁰ In the few cases of C pneumoniae­ associated AOM studied, most infected patients improved without macrolide therapy. Clearly, this area needs further study. For now, we do not believe that the theoretical benefit of action against atypical organisms justifies the use of macrolide antibiotics as routine second-line therapy for amoxicillin-unresponsive AOM.

A treatment strategy

The observations noted above explain some of the frustrations of treating otitis media. With most infectious diseases, one of two approaches can be used to direct antibiotic therapy: specific or empiric therapy. In specific therapy, if the responsible organism can be isolated, the clinician can use its pattern of in vitro antibiotic sensitivities to choose an appropriate narrow-spectrum agent. The natural history of AOM indicates that even in such an ideal situation, clinical success cannot be guaranteed.

Empiric antibiotic therapy is the usual approach to AOM. Most physicians start with amoxicillin because it is well tolerated, low in cost, and effective against likely pathogens. In the event of clinical failure, they choose a second-line antibiotic to "plug the holes" in coverage, often focusing on b-lactamase­producing H influenzae and M catarrhalis. This is a reasonable approach, but again, based on natural history, clinicians should be wary of attributing improvement (or lack of it) to the antibiotic chosen. Nonetheless, it is important to choose wisely and to avoid the temptation to initiate course after course of expensive second-line agents.

The question is which second-line agent to choose. Unfortunately, no randomized trials have been published addressing the treatment of amoxicillin-unresponsive AOM. One noncomparative study demonstrated a satisfactory outcome in 75% of children with persistent S pneumoniae and H influenzae treated with cefprozil.²¹ Interestingly, there was little relationship between clinical outcome and in vitro sensitivity.

It is unclear to what extent data on the treatment of uncomplicated AOM can be applied to clinical cases of antibiotic-unresponsive AOM. Until randomized trials are published, clinicians can base their choice of antibiotic on the known microbiology of amoxicillin-unresponsive AOM and the known in vitro spectrum of antibiotic coverage (see Table 2). Side effects, palatability, convenience of dosing, and cost are also important concerns.

The first criterion that the second-line agent selected should meet is that it should reliably eliminate b-lactamase­positive H influenzae and M catarrhalis. Reasonable choices include amoxicillin/clavulanate; the second-generation cephalosporin cefuroxime; third-generation cephalosporins such as cefixime, ceftibutin, cefpodoxime, or ceftriaxone; and non­b-lactam antibiotics such as trimethoprim/sulfamethoxasole (TMP/SMX) or the macrolides, azithromycin or clarithromycin. Cefaclor, loracarbef, and cefprozil are somewhat susceptible to degradation by b-lactamase and do not reliably treat b-lactamase­positive H influenzae.²² Neither cefaclor nor loracarbef is effective against S aureus. The only clinical advantage of any of these agents is their pleasant taste. Azithromycin and, to a lesser extent, clarithromycin provide broader coverage and greater convenience than erythromycin/ sulfamethoxasole.

A second criterion is that the agent selected should be effective against S pneumoniae and resistant to b-lactamase, since S pneumoniae is the organism most frequently isolated in amoxicillin-unresponsive AOM, often together with b-lactamase­positive gram-negative species. This exogenous b-lactamase may protect the S pneumoniae by degrading amoxicillin.¹¹ By this criterion, it is reasonable to choose amoxicillin/ clavulanate, cefuroxime, cefpodoxime, or ceftriaxone as a second-line agent. Cefprozil also maintains good activity against intermediate resistant S pneumoniae. Cefixime or ceftibutin would not be a good choice since these agents cannot be counted on to eradicate intermediate-resistant S pneumoniae and S aureus. Approximately 20% of intermediate­penicillin-resistant S pneumoniae are also resistant to macrolides and 40% are resistant to TMP/SMX. Among highly penicillin-resistant isolates, 54% are resistant to macrolides and almost 80% are resistant to TMP/SMX.²³

Third, PNSSP can often be effectively treated with high-dose amoxicillin. Amoxicillin resistance among S pneumoniae is usually caused by altered penicillin-binding proteins rather than by b-lactamase. Increasing the dose of oral amoxicillin (to 60 to 120 mg/kg/d) appears to provide sufficient antibiotic concentration in the middle ear to treat intermediate and, in most cases, high-level penicillin resistance.²⁴ High doses of amoxicillin can produce as favorable a clinical outcome as intravenous cefotaxime.²⁵ Parenteral ceftriaxone may also be effective for most intermediate and many highly resistant S pneumoniae. However, none of the oral cephalosporins provides adequate coverage of highly resistant organisms, which make up 10% to 20% of S pneumoniae isolates. Oral cephalosporins are uniformly inferior to amoxicillin for treatment of penicillin-resistant S pneumoniae.^23,26 Treatment of culture-proven DRSP should be directed by in vitro antibiotic sensitivities.

Based on these facts, we have developed the approach to the management of amoxicillin-unresponsive AOM shown in the flowchart. An integral part of this approach is the use of an adequate dose of amoxicillin for the initial treatment of symptomatic children. For children with significant symptoms and physical signs of ongoing middle ear effusion and inflammation after at least 48 hours of amoxicillin, we recommend the co-administration of amoxicillin and amoxicillin/clavulanate to provide a total dose of 80 to 100 mg/kg/d divided bid of amoxicillin and 10 mg/kg/d of clavulanic acid. We believe this combination is the oral preparation most likely to ensure treatment of the responsible organisms (b-lactamase­positive H influenzae with or without DRSP). Such a preparation of amoxicillin and clavulanic acid will soon be on the market. Since the infections of children receiving second-line drugs are likely to be more resistant to treatment than the infections of children with an initial diagnosis of uncomplicated AOM, we think a full 10-day course of antibiotics is appropriate. In the event of amoxicillin/clavulanate failure, a three-day course of parenteral ceftriaxone would be our next choice. Such a regimen has been shown to result in bacteriologic cure in 100% of patients with antibiotic-unresponsive AOM associated with H influenzae and penicillin-sensitive S pneumoniae, but in only 82% of patients with PNSSP.²⁷ It is unclear whether symptoms resolved in the 34% of patients in this study who had antibiotic-unresponsive AOM and negative cultures. The investigators suggest that a single dose of ceftriaxone (50 mg/kg) might be sufficient in many cases of antibiotic-unresponsive AOM.

Alternative second regimens include oral cefuroxime, cefpodoxime, azithromycin, clarithromycin, and TMP/SMX. Intramuscular ceftriaxone would be a reasonable second-line choice if symptoms are severe or oral compliance is an issue.

Recently a Centers for Disease Control and Prevention working group published recommendations for management of AOM in the era of pneumococcal resistance.⁶ The CDC group suggested using high-dose amoxicillin together with clavulanate, oral cefuroxime axetil, or IM ceftriaxone in the event of clinical amoxicillin failure. Their recommendations differ from ours in that we recommend delaying the use of IM ceftriaxone until a second-line agent fails. Theoretically, delaying the use of ceftriaxone could delay clinical cure in the 20% of cases of amoxicillin-unresponsive AOM associated with highly resistant PNSSP. According to current data, however, IM ceftriaxone is not demonstrably superior to high- dose amoxicillin for these patients. In the absence of a randomized trial comparing the two agents, we feel it is prudent to delay the use of ceftriaxone and forestall the spread of high-level cephalosporin resistance. A second difference between our recommendations and those of the CDC working group is our preference for high-dose amoxicillin/clavulanate over cefuroxime axetil. Although, cefuroxime appears to be the oral cephalosporin most likely to attain therapeutic levels in middle ear fluid, it is inferior to high-dose amoxicillin for PNSSP.

Many experts would advocate performing tympanocentesis in the case of clinical antibiotic failure. We agree that an attempt to identify causative organisms is preferable to merely giving multiple courses of empiric antibiotics, but it is not practical to expect most practicing pediatricians to perform this procedure in their offices. Furthermore, tympanocentesis is not a panacea for the problem of antibiotic-unresponsive AOM. As explained earlier, in vitro efficacy does not guarantee clinical improvement. (For more information on this option, see "Tympanocentesis: Why, when, how," by Stan L. Block, MD, in the March 1999 issue of Contemporary Pediatrics.)

Tympanocentesis would certainly be indicated for the child at highest risk of recurrent amoxicillin- unresponsive AOM: the infant or toddler in group day care suffering from ceftriaxone-unresponsive, symptomatic AOM. Alternatively, some physicians might opt for early tympanostomy tube placement in lieu of tympanocentesis. Middle ear ventilation procedures offer the advantages of immediate symptomatic relief of otalgia, improved hearing and language acquisition, and decreased need for oral antibiotics. Laser tympanostomy, which can be performed with local anesthesia and results in a tympanic perforation lasting three to four weeks, makes middle ear ventilation even more attractive.²⁸ Obviously, definitive recommendations about the roles of tympanocentesis and tympanostomy in antibiotic-unresponsive AOM require a prospective study comparing the two approaches.

In any case, the aggressiveness of evaluation and treatment should be determined by severity of symptoms and the clinical situation. It is difficult to justify either an invasive procedure such as tympanocentesis or empiric "antibiotic roulette" in a child with minimal symptoms. Adjunctive therapy such as analgesics and nasal decongestants can relieve many of the symptoms parents are most concerned about. Watchful waiting may be the best approach.

No choice is perfect

The combination of high-dose amoxicillin with clavulanic acid comes with several caveats. Amoxicillin/clavulanate treats only 95% of b-lactamase positive H influenzae.²² In addition, 2.5% of H influenzae isolates are b-lactamase negative yet resistant to amoxicillin and second-generation cephalosporins.²² Most of these resistant organisms remain susceptible to oral third-generation cephalosporins. These patterns of resistance are likely to become more common in the future if clinicians continue to choose amoxicillin/ clavulanate and second-generation cephalosporins as first-line agents.

One study showed a 38% bacteriologic failure rate of amoxicillin/clavulanate in children with H influenzae AOM, not associated with in vitro resistance to this agent. However, this was not a comparative study, and it is likely that failure was caused by viral co-infection.²⁹

Perhaps the most significant problem with amoxicillin/clavulanate is its tendency to cause diarrhea. A recent report indicates that the incidence of diarrhea is significantly lower with the new bid formulation and not significantly different from that of other oral antibiotics.³⁰ A physician should discuss proper diaper care with the parents and perhaps recommend an antifungal diaper cream.

Cefuroxime and cefpodoxime treat 95% and 99% of amoxicillin-resistant H influenzae clinical isolates, respectively,²⁹ while cefaclor, loracarbef, and cefprozil treat only 84%, 82%, and 70% of these isolates. Cefixime and ceftibutin each eradicate more than 99% of H influenzae and M catarrhalis organisms. However, both these agents and, to a lesser extent, cefpodoxime are also active, unnecessarily, against gram-negative enteric bacteria.³¹ Inevitably, excessive use of these agents will increase the incidence of gram-negative pathogens resistant to all cephalosporins, such as Salmonella species.

Both cefuroxime and cefpodoxime are relatively unpalatable. They should be given with food and are better tolerated if mixed with chocolate syrup or followed by a flavored beverage such as chocolate milk. Any antibiotic can cause diarrhea.

One major argument against the use of macrolide/azalide antibiotics for first- or second-line treatment of otitis media is that macrolide resistance among S pneumoniae isolates is becoming increasingly common. Only about 80% of intermediate resistant PNSSP and 50% of highly resistant PNSSP remain sensitive to this class of antibiotics.²³ Therefore, high-dose amoxicillin is more likely to be effective against penicillin-resistant S pneumoniae than macrolide antibiotics. High levels of macrolide use on antibiotic- resistant group A streptococci have been shown to be associated with the development of group A streptococcal macrolide resistance.³² Macrolide resistance among H influenzae appears to be significantly less common.²² Azithromycin seems to be particularly effective in selecting macrolide resistance. In one study, a single dose of azithromycin reduced S pneumoniae colonization from 68% to 29% over two weeks in 79 Australian aboriginal children.³³ At the same time the fraction of isolates that was macrolide resistant jumped from 1/54 to 6/11. Two months later 27% of children tested were colonized with macrolide-resistant pneumococci. These observations contrast with those of a comparative study of azithromycin and amoxicillin in which no change in colonization was observed after six months of either agent.³⁴ The difference between these two studies most likely reflects differences in the rate of antibiotic resistance and horizontal bacterial transmission in the two populations. Given the increasing frequency of macrolide resistance among penicillin-resistant pneumococci, the inevitability of continued b-lactam use among children, and the availability of high-dose amoxicillin with clavulanate, it seems reasonable to reserve macrolides for situations in which they represent the best or only alternative. Such situations might include AOM in a patient allergic to penicillin and cephalosporin, AOM with atypical pneumonia, AOM with diarrhea, and AOM caused by macrolide-sensitive bacteria. Based on pharmacokinetics, tolerability, palatability, cost, and convenience, azithromycin is the macrolide/azalide of choice, although clarithromycin appears to be somewhat more active against PRSP. Both of these agents tend to be concentrated in infected middle-ear fluid, and it is unclear to what extent pathogenic bacteria are exposed to these intracellular antibiotics. In all cases, treatment choice should take into account the child's day-care status, age, and risk of carrying resistant bacteria.

TMP/SMX remains a good second-line choice because of its low cost, tolerability, and limited impact on the selection of b-lactam­resistant flora. However, clinicians should be aware that it is not very effective against PNSSP.

Combatting the myth of amoxicillin "immunity"

A major complication of amoxicillin failure is that it convinces many parents that "amoxicillin doesn't work for my child anymore." Once a patient has failed to respond to amoxicillin therapy, it is tempting to label that patient "amoxicillin unresponsive" and to exclude amoxicillin from the armamentarium available for future episodes of AOM. However, as reviewed above, it is inappropriate to attribute all episodes of amoxicillin-unresponsive AOM to resistance. In most cases, the bacteria isolated from these patients are sensitive to amoxicillin. The impact of initial antibiotic choice on outcome is minimal. One population-based study found a 12% to 13% rate of antibiotic failure as measured by a prescription for a second course of antibiotics within three weeks.³⁵ The failure rate was similar regardless of the agent first prescribed, but the cost of the prescription was five- to 20-fold higher for the second-line agent. So, despite occasional failure, amoxicillin should remain the first choice for pediatric patients with AOM.

Understandably, some parents interpret the clinical failure of an antibiotic that their child has taken repeatedly as a form of "immunity" to amoxicillin, similar to the immunity to disease that childhood immunizations convey. If a clinical entity of "amoxicillin immunity" existed, the only plausible mechanism would be prolonged colonization with amoxicillin-resistant bacteria. Several studies have tracked the prevalence and persistence of nasopharyngeal colonization with amoxicillin-resistant bacteria. In most cases, colonization with resistant bacteria is associated with recent exposure to antibiotics; in the absence of antibiotics, these bacteria appear to clear spontaneously over several months.³⁶ Clinical failure with amoxicillin in previous years should not be taken as an indication to choose a second-line agent for further otitis episodes.

It is tempting to consider choosing a broad-spectrum agent such as amoxicillin/clavulanate or a second-generation cephalosporin as first-line therapy for children who have a previous history of amoxicillin failure or live in communities with high rates of b-lactamase­ positive H influenzae.³⁷ Supporting such a practice is concern that the repeated choice of amoxicillin for children with amoxicillin-unresponsive otitis media might erode parental confidence in the physician.³⁷ Parent education about otitis etiology and treatment will help to counter parental reservations about the repeated use of amoxicillin. In particular, we stress that S pneumoniae is a common otitis pathogen, and that high-dose amoxicillin works best against this organism.

If a parent thinks we would be able to cure their child if only we could choose the correct antibiotic, disappointment is inevitable. An example of this can be seen in the recent movie "As Good As It Gets," in which the lead character complains that her child's asthma would be cured if only the intern in the ER would choose "the right antibiotic." As clinicians we know that the "right antibiotic" will not cure asthma, or even an ear infection. Nonetheless, in a recent survey, 30% of parents reported that they have requested a specific antibiotic from their pediatrician, and 30% of practitioners report that they often change their choice of antibiotic in response to parental request.³⁸ One way to counteract these trends is to educate parents about otitis and antibiotic resistance (see "Guide for Parents: Understanding ear infections"). Reassurance about the favorable natural history of AOM and the recommendation of adequate analgesia are also helpful in relieving parental distress.

Patients with recurrent AOM might benefit from influenza and pneumococcal immunization, and their parents may respond to a discussion of risk factors such as cigarette smoke, pacifiers, and group day care. Most parents are more willing to accept a watchful approach to AOM therapy once they understand the risks that "antibiotic roulette" poses to a child in day care.

The repeated use of broad-spectrum antibiotics in infants and children with recurrent ear infections virtually assures the eventual acquisition of multiresistant organisms. Fortunately, we can assure parents that each episode will eventually resolve and that future episodes will become less and less frequent.

Achieving a balance

Obviously, the management of AOM is complex, and no one strategy is ideal for all patients. Clinicians need to balance the uncertain clinical benefit of antibiotic therapy with the known costs of antibiotic use including allergic and gastrointestinal side effects, compliance problems, and the selection of antibiotic resistant bacteria in the patient and community. In our opinion, clinicians can safely reduce the number of antibiotic prescriptions written by directing antibiotic therapy toward patients with symptoms of middle ear pain and younger patients at higher risk of complications, by prescribing adequate does of amoxicillin, and by following an orderly progression of second-line agents chosen to eradicate likely organisms.

Understanding ear infections

Many children get at least one ear infection during their early years, and some children get ear infections repeatedly. In fact, ear infections are responsible for more sick-child visits to pediatricians than any other childhood illness. This guide will help you understand what happens during an ear infection, why these infections develop, and how they are treated.

What is otitis media?

When your pediatrician says that your child has otitis media, it means that the child has fluid in the middle ear (see drawings). When the fluid becomes infected, the condition is called acute otitis media. An uninfected, painless buildup of fluid in the middle ear is otitis media with effusion.

What causes ear infections?

An ear infection is usually a complication of a cold.
It occurs when a cold causes swelling and blockage of the eustachian tube, so that fluid builds up in the middle ear and becomes infected. Infants and toddlers have shorter and more angled eustachian tubes than older children; this makes them more prone to otitis media. Most children outgrow the tendency to develop ear infections by the time they are 4 years old.

Breastfeeding an infant lowers the risk of ear infections. Certain other things increase the risk. These include:

• Exposure to frequent viral infections in day care

• Exposure to secondhand cigarette smoke

• Use of pacifiers

• A family history of frequent ear infections.

How do you know when your child has an ear infection?

The most common symptom of an ear infection is ear pain. While an older child can tell you that his ear hurts, a younger child may simply act irritable and cry more than usual. Lying down, feeding, and sucking cause painful pressure changes in the middle ear, so the child may eat less than the normal amount or have trouble sleeping. Other symptoms are fever, loss of balance, difficulty hearing, and yellow or blood-spotted drainage from the child's ear. If your child tugs at her ear, it may not be due to an ear infection. Ear pain or irritability can also be caused by a sore throat, teething, or swimmer's ear.

How are ear infections treated?

Acute otitis media is usually treated with an antibiotic. Without an antibiotic, most infections (about 70%) will get better in a few days. But even with an antibiotic, some infections (about 10%) will not get better in a short time. Unfortunately, it is impossible to know which children will and won't get better. So pediatricians usually prescribe an antibiotic, and a number of children (about 20%) will benefit.

For an uncomplicated ear infection, the antibiotic most often prescribed is amoxicillin. It treats the most common bacterial cause of ear infections--a bacteria called Streptococcus pneumoniae--better than most other antibiotics. If symptoms continue, your pediatrician may prescribe some other type of antibiotic that works on less common bacteria, such as amoxicillin-resistant Haemophilus influenzae. This doesn't mean that your child is "immune" to amoxicillin, only that this particular infection has not responded to it. Amoxicillin is still the best antibiotic to try first if your child gets another ear infection in the future.

Make sure that your child takes the medicine your pediatrician prescribes for the length of time advised by the doctor. Even when symptoms have gone away, some weakened bacteria may still be present.

You can give your child acetaminophen (such as Tylenol) or ibuprofen (such as Advil, Motrin, or PediaCare Fever) for pain and fever as needed. Children and adolescents should not take aspirin because it increases their risk of contracting Reye syndrome. To help relieve pain, apply warm compresses or a hot water bottle to the area around the ear. Be sure the compress or bottle is not too hot. Saline nose drops and a humidifier may help relieve nasal congestion.

Overuse of antibiotics

The overuse of antibiotics can create problems. Antibiotics may kill "healthy" bacteria in the digestive system as well as the bacteria responsible for the ear infection, causing diarrhea. The absence of healthy bacteria may make it easier for disease-causing bacteria to take hold in the body. Also, when antibiotics are used often, bacteria that are resistant to their effects have a chance to develop. Infections caused by resistant bacteria can be very difficult to treat.

Your doctor will decide what is the best treatment for your child. He will not prescribe an antibiotic if he thinks your child's infection is caused by a virus, since antibiotics have no effect against viruses. He may not prescribe one if your child has otitis media with effusion (which usually does not cause pain) rather than acute otitis media. In otitis media with effusion, the fluid in the middle ear is usually not infected. Middle ear fluid may remain for weeks even after proper treatment of an ear infection.

If you don't understand why your doctor is not prescribing an antibiotic, ask her, but please do not urge her to give antibiotics to your child unnecessarily. Parents and doctors need to work together to prevent the further development of resistance.

When should you call your doctor?

Your child should feel better 48 to 72 hours after starting to take an antibiotic. If such symptoms as fever, pain, and irritability are still present after three days of medication, call your doctor.

Your pediatrician will want your child to come in for one or more follow-up examinations until all signs of inflammation and infection have disappeared. Be sure to keep all your appointments.

DR. LINSK is Clinical Instructor, Division of General Pediatrics, University of Michigan, Ann Arbor, MI. He has a research grant from Glaxo Wellcome Inc.

DR. GILSDORF is Professor of Pediatrics and Communicable Diseases, and Director, Division of Pediatric Infectious Diseases at the same institution.

DR. LESPERANCE is Assistant Professor, Department of OtolaryngologyÏHead and Neck Surgery at the same institution.

Acknowledgments:
The authors would like to thank their colleagues and patients' parents for many helpful discussions about the management of acute otitis media. We also thank Dr. Gary Doern for reviewing the manuscript.

REFERENCES

1. Hayden G: Acute suppurative otitis media in children. Diversity of clinical diagnostic criteria. Clin Pediatr 1981;20:99

2. Dowell S, Marcy S, Phillips W, et al: Otitis media-principles of judicious use of antimicrobial agents. Pediatrics 1998;101:S165

3. Rosenfeld R: An evidence-based approach to treating otitis media. Pediatr Clin North Am 1996;43:1165

4. Conrad D: Should acute otitis media ever be treated with antibiotics? Pediatr Ann 1998;27:66

5. Chartrand S, Pong A: Acute otitis media in the 1990s: The impact of antibiotic resistance. Pediatr Ann 1998; 27:86

6. Dowell S, Butler J, Giebink G, et al: Acute otitis media: Management and surveillance in an era of pneumococcal resistance--a report from the drug-resistant Streptococcus pneumoniae therapeutic working group. Pediatr Infect Dis J 1999;18:1

7. Duchin J, Breiman R, Diamond A, et al: High prevalence of multidrug resistant Streptococcus pneumoniae among children in a rural Kentucky community. Pediatr Infect Dis J 1995;14:745

8. Block S, Harrison C, Hedrick J, et al: Penicillin-resistant Streptococcus pneumoniae in acute otitis media: Risk factors, susceptibility patterns, and antimicrobial management. Pediatr Infect Dis J 1995;14:751

9. Rosenfeld R, Vertrees J, Carr J, et al: Clinical efficacy of antimicrobial drugs for acute otitis media: Meta-analysis of 5,400 children from thirty-three randomized trials. J Pediatr 1994;124:355

10. Pichichero M, Pichichero C: Persistent acute otitis media: I. causative pathogens. Pediatr Infect Dis J 1995; 14:178

11. Brook I, Yocum P: Bacteriology and b-lactamase activity in ear aspirates of acute otitis media that failed amoxicillin therapy. Pediatr Infect Dis J 1995;14:805

12. Leibovitz E, Raiz S, Piglanski L, et al: Resistance pattern of middle ear fluid isolates in acute otitis media recently treated with antibiotics. Pediatr Infect Dis J 1998;17(6):463

13. Harrison C, Marks M, Welch D: Microbiology of recently treated acute otitis media compared with previously untreated acute otitis media. Pediatr Infect Dis J 1985;4:641

14. Gehanno P, N'Guyen L, Derriennic M, et al: Pathogens isolated during treatment failures in otitis. Pediatr Infect Dis J 1998;17:885

15. Chonmaitree T, Owen M, Patel J, et al: Effect of viral respiratory tract infection on outcome of acute otitis media. J Pediatr 1992;120:856

16. Canafax D, Yuan Z, Chonmaitree T, et al: Amoxicillin middle ear penetration and pharmacokinetics in children with acute otitis media. Pediatr Infect Dis J 1998;17:149

17. Carlin S, Marchant C, Shurin P, et al: Host factors and early therapeutic response in acute otitis media. J Pediatr 1991;118:178

18. Bluestone C, Stephenson J, Martin L: Ten-year review of otitis media pathogens. Pediatr Infect Dis J 1992;11:S7

19. Block S, Hammerschlag M, Hedrick J, et al: Chlamydia pneumoniae in acute otitis media. Pediatr Infect Dis J 1997;16:858

20. Normann E, Gnarpe J, Gnarpe H, et al: Chlamydia pneumoniae in children attending day-care centers in Gavle, Sweden. Pediatr Infect Dis J 1998;17:474

21. Pichichero M, McLinn S, Aronovitz G, et al: Cefprozil treatment of persistent and recurrent acute otitis media. Pediatr Infect Dis J 1997;16:471

22. Doern G, Bruggemann A, Pierce G, et al: Antibiotic resistance among clinical isolates of Haemophilus influenzae in the United States in 1994 and 1995 and detection of b-lactamase strains resistant to amoxicillin-clavulanate: Results of a national multicenter surveillance study. Antimicrob Agents Chemother 1997;41:292

23. Doern G, Bruggemann A, Holley H, et al: Antibiotic resistance of Streptococcus pneumoniae recovered from outpatients in the United States during the winter months of 1994 and 1995: Results of a 30-center national surveillance study. Antimicrob Agents Chemother 1996;40:1208

24. Seikel K, Shelton S, McCracken G: Middle ear concentrations of amoxicillin after large doses in children with acute otitis media. Pediatr Infect Dis J 1997;16:710

25. Roger G, Carles P, Pangon B, et al: Management of acute otitis media caused by resistant pneumococci. Pediatr Infect Dis J 1998;17:631

26. Dagan R, Abramson O, Leibovitz E, et al: Impaired bacteriological response to oral cephalosporins in acute otitis media caused by pneumococci with intermediate resistance to penicillin. Pediatr Infect Dis J 1996;15:980

27. Leibovitz E, Piglanski L, Raiz S, et al: Bacteriologic efficacy of a three-day intramuscular ceftriaxone regimen in nonresponsive acute otitis media. Pediatr Infect Dis J 1998;17:1126

28. Silverstein H, Kuhn J, Choo D, et al: Laser-assisted tympanostomy. Laryngoscope 1996;106:1067

29. Patel J, Reisner B, Vizirinia N, et al: Bacteriologic failure of amoxicillin-clavulanate in treatment of acute otitis media caused by non-typable Haemophilus influenzae. J Pediatr 1995;126:799

30. Bottenfield G, Burch D, Hedrick J, et al: Safety and tolerability of a new formulation (90 mg/kg/d divided every 12 h) of amoxicillin/clavulanate (Augmentin) in the empiric treatment of pediatric acute otitis media caused by drug-resistant Streptococcus pneumoniae. Ped Infect Dis J 1998;17:963

31. Wiedemann B, Luhmer E, Zuhlsdorf M: In vitro activity of cefpodoxime and 10 other cephalosporins against gram-positive cocci, enterobacteriaceae, and Pseudomonas aeruginosa, including b-lactamase producers. Infection 1991;19:363

32. Seppala H, Klaukka T, Vuopio-Varkila J, et al: The effects of changes in the consumption of macrolide antibiotics on erythromycin resistance in group A streptococci in Finland. N Engl J Med 1997;337:441

33. Leach P, Shelby-James T, Mayo N, et al: A prospective study of the impact of community-based azithromycin treatment of trachoma on carriage and resistance of Streptococcus pneumoniae. Clin Infect Dis 1997;24:356

34. Marchiso P, Principi N, Sala E, et al: Comparative study of once weekly azithromycin and once daily amoxicillin treatments in prevention of recurrent acute otitis media in children. Antimicrob Agents Chemother 1996;40:2732

35. Berman S, Bryns P, Bandy J, et al: Otitis media-related antibiotic prescribing patterns, outcomes, and expenditures in a pediatric Medicaid population. Pediatrics 1997;100:585

36. Brook I, Gober A: Monthly changes in the rate of recovery of penicillin-resistant organisms from children. Pediatr Infect Dis J 1997;16:255

37. Pichichero M: Assessing the treatment alternatives for acute otitis media. Pediatr Infect Dis J 1994;13:S27

38. Palmer D, Bauchner H: Parents' and physicians' views on antibiotics. Pediatrics 1997 [http://www.pediatrics. org/cgi/content/full/99/6/e6]

Richard Linsk,Marci Lesperance,Janet Gilsdorf. When amoxicillin fails.

Contemporary Pediatrics

1999;10:67.