Apparent treatment failure of acute (suppurative) otitis media (AOM) is a common dilemma facing the primary care physician. Parents and physicians alike can be frustrated when evidence of infection persists despite what appeared to be an appropriate course of antibiotic therapy. This article explores the possible reasons for antimicrobial treatment failures and suggests effective treatment strategies.
DEFINING CRITERIA FOR TREATMENT OUTCOME
Physical findings from examination of the tympanic membrane (TM) are the most reliable means of judging the treatment outcome of AOM. Treatment can be judged to be completely successful if the TM has a normal color/lucency, is in neutral position, and is freely mobile with both positive and negative insufflation. Nearly half of the patients initially treated for AOM experience complete resolution by the end of a 10-day antibiotic course. In the remaining 50% to 60%, some TM abnormality exists.
Most patients with abnormal TM signs at the end of therapy have signs of otitis media with effusion (OME), previously known as serous otitis media. The distinguishing TM signs of ONJE are a neutral or retracted position and decreased mobility on positive insufflation, but less so with negative insufflation. The altered mobility and retraction of the TM indicate negative pressure in the middle ear. When OME follows treatment of AOM, it does not usually represent persistent infection and should not be considered treatment failure. Otitis media with effusion is usually the first step on the path to complete resolution of AOM and requires antimicrobial therapy only in special cases. However, careful observation is important until the middle ear effusion (MEE) resolves, usually over the next 6 to 12 weeks.
Persistent AOM can be found in 10% to 15% of antibiotic treated patients. The three classic signs of AOM are:
* bulging (positive pressure in the middle ear cavity),
* altered color or lucency of the TM (due to inflammation), and
* decreased mobility of the TM on both positive and negative insufflation (a middle ear filled with fluid allows little mobility).
Bacterial Pathogens in Patients With Previously Untreated and Persistent Acute Otitis Media*
These physical findings are more than 95% predictive of a persistent purulent MEE.1 Despite predicting a purulent MEE, neither these nor any other physical or clinical findings reliably predict whether the purulent MEE requires further antimicrobial treatment. Only aspiration (tympanocentests) and culture of the MEE can confinn sterilization of the purulent MEE after treatment. Thus, clinicians must empirically decide whether to treat the persisting purulent MEE, and if so, with what agent unless they choose to perform tympanocentesis and culture. Knowledge of the microbiology of recently treated AOM compared to previously untreated AOM is important in this decision.
MICROBIOLOGY OF PREVIOUSLY UNTREATED VERSUS PERSISTENT ACUTE OTITIS MEDIA
In previously untreated patients, 65% to 80% of purulent MEE will contain viable organisms, justifying empiric treatment.2,3 Streptococcus pneumoniae, nontypabie (nt) Haemophilus influenzae, and Moraxeua catarrhaUs are the most frequent organisms isolated in previously untreated AOM (Table 1). Nearly all M catarrhalis isolates produce betalactamase. However, the proportion of nt H influenzas that produce beta-lactamase varies geographically4 and that proportion continues to increase in frequency, currently ranging from 15% to 70%. The choice of an empiric first-line antibiotic for previously untreated AOM is based on a knowledge of the variations in the proportions of pathogens causing AOM and their probable susceptibility patterns to various antibiotics in the practitioner's locale.
In contrast, when AOM persists in a patient currently under treatment or recently treated with antibiotics, the percentage of MEE with identifiable pathogens is lower. The proportions of each organism isolated also differ, and more resistant susceptibility patterns are observed when compared to organisms isolated in previously untreated AOM- For example, the incidence of Staphyiococcus aureus isolation in persistent AOM increases as does that of nt H influenzas organisms that produce beta-lactamase. Yet at least half of the organisms isolated from recently treated AOM remain susceptible to the antibiotic just administered.3 Two explanations for persistent AOM with an organism susceptible to the recently completed antibiotic are incomplete compliance in taking the medication on the patient's part and inadequate antibiotic penetration into a positively pressurized middle ear space. Unfortunately, it is difficult to be certain that the antimicrobial agent was administered as prescribed. It is also impossible without tympanocentesis to know whether the patient's pathogen was sensitive to the prescribed antibiotic.
Not all treatment failures are symptomatic. Treatment failure, therefore, is detected in two ways. One is by persistence of the clinical symptoms together with TM signs of AOM > 48 to 72 hours after starting treatment. The other is persistence or recurrence of TM signs of AOM at the end of or soon after a 1 0-day course of treatment, regardless of the clinical symptoms and signs. Table 2 lists the clinical presentations that qualify as treatment failures. The clinician should be able to distinguish among these presentations because their management may vary.
CLASSIFICATION OF TREATMENT FAILURE
Early Treatment Failure
Early treatment failure may be defined as persistence of symptoms (eg, irritability, otalgia, fever, decreased appetite, sleep disturbance, and/or purulent otorrhea) 48 to 72 hours after the institution of antibiotic treatment. With effective antibiotic treatment, one expects these clinical Symptoms and signs to wane within 72 hours. When symptoms persist, it is imperative to ensure that the antimicrobial has been administered by the parents and retained by the patient. Noncompliance can take the form of failure to fill the prescription, omission of doses, errors in the amount or timing of the dose, or premature discontinuation of the antibiotic.
If it seems reasonable that the child has received the medication, it may be that inflammation and symptoms persist despite eradication of the bacterial pathogen. Tympanic membrane findings are rarely helpful in deciding whether the MEE is sterile or still infected because some bulging, erythema, and impaired mobility of the TM almost always remain for at least 48 to 72 hours, even with effective antibiotic therapy. Therefore, a decision to alter therapy depends on the severity of symptoms reported by the parents coupled with the clinician's decision that these symptoms are not due to a concomitant viral infection or drug reaction. Even in the best case scenario, antibiotic resistance is present in about half the cases of early treatment failures; therefore, changing to an antimicrobial agent with a broader spectrum of activity is recommended when signs and symptoms of AOM persist longer than 48 to 72 hours.
Clinical Presentation of Acute Otitis Media Treatment Failures
Failure Late in Therapy
Another presentation of treatment failure is the child who initially improves, but signs and symptoms suggestive of AOM recur later on in the 10-day course of treatment. If the classic TM signs of AOM are present in the same or the opposite ear, noncompliance (premature discontinuation or inadequate dosing) of the prescribed antibiotic should be suspected. If fever and irritability recui; but the appearance of the TM is improved, an intervening viral infection or drug reaction should be considered. Altering the antimicrobial therapy is not recommended unless noncompliance or superimposed viral infection are considered unlikely.
Conditions and Situations That Predispose a Child to Become Otitis Prone
Treatment failure may occur when clinical symptoms have improved but TM signs of AOM persist at or soon after completion of a course of antibiotics. An additional course of antimicrobials is recommended despite an almost 50% probability that all viable organisms have been eradicated.3 Tympanic membrane abnormalities can persist despite sterilization of the MEE because eustachian tube dysfunction did not allow drainage of the now sterile middle ear abscess. Repeated viral infections, attendance at a day-care center, and passive smoking have been suggested as contributing factors in cases of persistent AOM. Every effort should be made to shield the child from parental smoking. Congenital abnormalities of the airway or underlying familial eustachian tube abnormalities also may play a role in persistent AOM.
Recrudescence After Cessation of Therapy
A final presentation of treatment failure is a clinical relapse or recurrence of TM signs of AOM, with or without clinical symptoms, shortly after a documented satisfactory outcome at the end of therapy. This may not be uniformly due to treatment failure, but may represent a second infection with a new organism. Distinguishing between these possibilities is impossible if no interval examination confirmed the initial resolution of AOM. Additionally, tympanocentesis and culture of both the first and second effusion are the only means to verify a new organism producing a second infection.
Regardless of whether the occurrence of AOM soon after discontinuing antibiotics is caused by persistence, relapse, or reinfection, one should prescribe a 10-day course of a second-line antimicrobial. Incompletely treated, concomitant sinusitis may be a factor in this clinical presentation. Reinfection also occurs frequently in "otitis prone" children.
THE ROLE OF VIRAL COPATHOGENS IN PERSISTENT AOM
Concomitant viral infections have been suggested as a cause of persistent AOM and treatment failure. Epidemiologie and virologie data suggest that respiratory viruses play a role in the pathogenesis of AOM.5'9 A preceding or concurrent viral upper respiratory illness is often observed in children with AOM. Viruses have been isolated in as many as 25% of middle ear aspirates and in as many as 42% of nasopharyngeal secretions from children with AOM. It is probable that the role of viral copathogens has been underestimated because of difficulties with viral detection. The viruses most frequently associated with AOM include respiratory syncytial virus (the most frequent cause of bronchiolitis), parainfluenza viruses (the most frequent causes of croup), influenza virus, and adenovirus; rhinovirus, enterovinises, and coronaviruses are detected less frequently. Prevalence of these viruses depends on the season of the year and geographic location.
Whether concurrent viral infection prolongs the course of AOM or increases the frequency of treatment failure remains controversial. In one Finnish study, no difference was observed in follow-up TM signs, persistence of effusion, or overall outcome regardless of the presence of concurrent viral infections.8 However, respiratory viruses were isolated more frequently from end-of- therapy MEE from patients with persistent AOM.9 Data from Texas5 also demonstrated that bacteria persisted longer in MEE when a virus was also present in pretreatment middle ear cultures. In summary, there is increasing evidence that respiratory viruses play a significant role in AOM treatment failure.
Future antiviral drugs or vaccines may decrease the apparent negative effects of viral agents on the successful outcome of AOM therapy. However, for the present, awareness of potential increased treatment failures could produce more realistic expectations and decreased frustrations for both parents and physicians.
OTHER RISK FACTORS FOR TREATMENT FAILURE
The otitis prone child (Table 3) appears to be one who is at greater risk for frequent treatment failures and relapses than normally expected. The child receiving a second course of antibiotics for persistent AOM also is more likely to experience treatment failure. These increased risks occur even though the middle ear pathogen is susceptible to the prescribed antibiotic and even when compliance is documented.
While immunodeficiency is not a frequent cause of multiple AOM treatment failures, specific abnormalities of the immune system (eg, immunoglobulin deficiencies, neutrophil dysfunctions, and T-cell abnormalities) can present with frequent or persistent AOM. The precise role of allergies in persistent AOM remains controversial, and specific treatment for manifestations of allergic disease does not appear to prevent AOM to any significant degree. Environmental parameters such as attendance at a day-care center and passive smoking have been implicated as risk factors for AOM and may play a role in persistent AOM. Informing the parents that if these risk factors cannot be altered, then their child is likely to have more episodes of AOM and that the episodes will be more resistant to treatment enables the family to know what to expect and perhaps to accept the fact that their child is otitis prone.
CHOICE OF ANTIBIOTICS
The penetration of an antibiotic into the middle ear is often taken for granted. Actually, the middle ear does not have a rich blood supply, and effective concentrations of antibiotic may be difficult to obtain and sustain, especially when the middle ear is filled with "pus under pressure." Some antibiotics have more difficulty penetrating the MEE than others. For example, cephalexin (Keflex, Dista Products Co, Indianapolis, Indiana) exhibits a reasonable spectrum of antimicrobial activity in vitro but is not recommended for the treatment of AOM because of an inherent deficiency in penetrating the MEE.
Subtherapeutic MEE antibiotic concentrations also may be a problem even for drugs that ordinarily penetrate well. For example, the inflammation associated with AOM may produce positive pressure in the middle ear that prevents sufficient antibiotic from reaching the central nidus of organisms in the MEE. Thus, soon after the antimicrobial is discontinued, signs and symptoms of AOM may return even though the clinician chose the "correct" antibiotic empirically. This provides the rationale for repeating a second course of the same antibiotic when treating persistent AOM. However; because half or more of the isolates obtained from persistent AOM by tympanocentesis are resistant to the initial antimicrobial prescribed, it has become a standard of care to change to a second-line, broader spectrum antibiotic. IfAOM persists through a second course of antibiotics, it must be decided whether to alter antibiotic therapy again or to refer the patient for subspecialty evaluation that may include tympanocentesis or myringotomy.
Another factor in the choice of antibiotic after treatment failure is knowledge of proportions of pathogens involved in persistent AOM and the susceptibility patterns for each geographic region. This can be difficult because regional susceptibility patterns change from year to year. A prime example is the prevalence of beta-lactamase-producing nt H influenzae, which ranges between 15% and 70%. Because this information may be difficult to obtain in any given locale, the recommended approach for persistent AOM is to prescribe empirical antimicrobials that are impervious to the beta-lactamases produced by ordinary respiratory pathogens.
A detailed discussion of the various antibiotics used in treating AOM is beyond the scope of this article but can be found in several recent publications.10'12 A selective review, however, is in order.
Amoxicillin administered orally three times daily remains the initial drug of choice for previously untreated AOM in areas with a low incidence of beta-lactamase-producing organisms.
Trimethoprim-sulfamethoxazole (Bactrim, Roche Laboratories, Nutley, New Jersey and Septra, Burroughs Wellcome, Research Triangle Park, North Carolina) is also popular as a first-line drug or even as a second-line one if amoxicillin was used initially. Its popularity is due to its twice-daily dosing, its low cost, and its reasonable activity against beta-lactamaseproducing nt H influenzae and M catarrhahs. However, complications caused by sulfa drugs and the increasingly observed resistance to this drug by respiratory pathogens must be considered when choosing it.
Doxycycline (Vibramycin, Pfizer Labs Division, New York, New York) also can be used as a first-line or second-line drug if the child is over the age of 8 years. It too is administered twice daily and is also usually effective against beta-lactamase producers.
Broader spectrum, second-line antimicrobials recommended for treating persistent AOM should be resistant to beta-lactamase. The amoxicillinclavulanate combination (Augmentin, Beecham Laboratories, Philadelphia, Pennsylvania), cefuroxime axetil (Ceftin, Alien & Hanburys, Research Triangle Park, North Carolina), or cefixime (Suprax, Lederle Laboratories, Wayne, New Jersey) are good choices. Each of these agents, however, has some drawbacks.
Ceiaclor (Ceclor, EH Lilly and Co, Indianapolis, Indiana) until recently has been a good second-line choice, but currently appears less effective, especially with twice-daily dosing. Its lower efficacy is probably mult ifactorial; decreasing activity against gramnegative organisms with increasing inocula size and increasing resistance of H influenzae and M catarrhaíis are the most important factors. The increasing recognition of serum sickness within 8 weeks of cefaclor use also has decreased enthusiasm for its use.
The erythromycin-sulfa combination (Pediazole, Ross Laboratories, Columbus, Ohio) has a reasonable in vitro spectrum against respiratory pathogens. However, it contains two drugs with notable side effects and requires four-times-a-day dosing. The erythromycin ethyl succhiate component may induce vomiting in many patients when administered in courses lasting longer than 5 days. The complications of the sulfa component (drug rash and Stevens-Johnson syndrome) also must be considered. Noncompltance over a 10-day course with this combination also can be a problem.
Amoxicillin-clavulanate (Augmentin) has proven to be a potent antimicrobial with excellent coverage of respiratory pathogens, including anaerobes and beta-lactamase-producing aerobes. However, like other potent drugs, it must be prescribed and administered in a careful manner. A dose of 30 to 35 mg/kg/day divided in three doses with each dose at least six hours apart is effective. This dosing interval results in fewer patients experiencing diarrhea (< 15%) and discontinuing the drug (<5%) if it is administered on a full stomach followed by 2 oz to 4 02 of fluid. However, up to 50% of patients experience diarrhea, and 10% may need to discontinue the drug if the dose is not calculated precisely, if it is given too soon after a previous dose, or if it is taken on an empty stomach. If the diarrhea is severe, perineal excoriation can occur in infants if the skin is not protected by barrier creams.
Cefuroxime axetil (Ceftin) is a second-generation cephalosporin with excellent in vitro spectrum against middle ear pathogens, but it currently is available only in tablet form, which limits its use in children. In addition, the unpleasant taste of cefuroxime axetil tablets cannot be masked in food or beverages. Crashing the tablets into food can lead to vomiting in nearly half of patients, and this may produce subsequent refusal of any offered medications.
Cefixime (Suprax), a third-generation cephalosporin, is an excellent choice for gram-negative organisms. It has the advantage of once-daily or twice-daily dosing and produces few side effects. However, it is ineffective against staphylococci, which may occur in increased frequency in persistent AOM. Its efficacy against S pneumaniae also has been questioned; breakthrough S pneumoniae bacteremia have been reported during cefixime therapy.
Thus, either amoxicillin-clavulanate dosed appropriately, cefuroxime axetil, or cefixime as second-line antibiotics are recommended for apparent treatment failure of AOM. Loracarbef and cefpodoxime are both being tested in clinical trials as expanded spectrum oral drugs which, when released, may be third-line drugs for especially recalcitrant AOM.
When the clinician is presented with apparent treatment feilure, noncompliance must be considered first. If this is the problem, the medication should be retnstituted after the parents are counseled. Next, the possibility of a superimposed viral illness also must be considered. When satisfied that these are not the problems, the clinician must consider whether the MEE has created so much positive pressure that antimicrobials cannot completely penetrate the middle ear space. If positive pressure is considered the problem, the same or another antimicrobial should be continued for a second 10-day course, allowing the pressure to decrease with time permitting more complete antibiotic penetration into the MEE.
Alternately, clinicians with appropriate training may elect to relieve the pressure by tympanocentesis or myringotomy. If the clinician decides that it is more likely that the patient has a pathogen resistant to the initial choice of an antimicrobial agent, a second course of a more potent second-line antimicrobial is appropriate. Patients who feil a second course of antimicrobials should receive an alternate second-line antimicrobial or undergo drainage of the middle ear abscess. Those who fail a third course of antimicrobials should be referred to an otolaryngologist for evaluation or surgical intervention. Antimicrobials should be continued until the consultation occurs.
This article has outlined potential clinical presentations for treatment failures as well as choices for second-line antibiotics. The relation of anatomic, environmental, microfaial, and antimicrobial factors in persistent AOM must be considered in order to determine if the clinician should do more than merely prescribe second-line antibiotics. With a greater understanding of the factors involved in apparent treatment failures, rational management decisions can be made, and alternative antibiotics can be prescribed or referral can be used wisely in the treatment of children with persistent AOM.
1. Schwanz RH. Acute otitis media: toward a more precise definition. Clin Pedìorr (Rulo). 1981:20:549-554.
2. Gicbink GS. The microbiology of otitis media. Pediatr. ínjkt DisJ, 1989;8:518-520.
3. Harriso« CJ, Macks MI, Welch DF. Microbiology of recently treated vs untrealtd otitis media. Potion infect Du J. 1985;4:641-646.
4. Doem CV, Jorgenson JH, Thomsberry C, et al. National collaborative study of the prevalence of antimicrobial resistance among clinical isolates of Hoemopbaus m/liieimie. AntimUTob Agenis Chimo*«·. 1988;32:180-185.
5. Chomaitree T, Owen MJ, Howie VM. Respiratory viruses interfere with bactériologie responsi to antibiotic in children with acute otitis media. } injea Dis. 1990;162:546549.
6. Chomaitree T. Viral otitis media. Pediatr Arm. 1990; 19:522-532.
7. Henderson PV1 Collier AM, Sanyal MA, et al. A longitudinal study ci respiratory viruses and bacteria in the etiology of acute oiitis media with effusion. NEn¿]Med. 1982306:1377-1383.
8. Aróla M, fUiuskanen Q Ziegler T, et al. Clinical role of respiratory virus infection in acute otitis media. ftüWicJ. 1990:86:848-855.
9. Aróla M, Ziegler T, Ruuckanen Q Respiratory virus infection a* a cause of prolonged symptoms in acute otitis media. } Peàuar. 1990; 11 6:69 7-701.
10. Le CT, et al. Choosing an antibiotic: efficacy, side effects, and costs. Contemporary ftdiotricj. 1991;8:11-30.
11. BlueMoneCD «al. Management c^otitismedia in infantsand children: cuirentfole of old and new antimicrobial agents. Pediatr Infect DuJ. 1988;7:S129-S136.
12. Biuestone CQ ci al Modern management afoláis medía, Peaíacr Gin North Am. 1989-3&1371-1387.
Bacterial Pathogens in Patients With Previously Untreated and Persistent Acute Otitis Media*
Clinical Presentation of Acute Otitis Media Treatment Failures
Conditions and Situations That Predispose a Child to Become Otitis Prone