INTRODUCTION

Top Abstract Introduction Materials and methods Results Discussion References

The three major bacterial species causing acute otitis media (AOM) are Streptococcus pneumoniae, Haemophilus influenzae, and Moraxella catarrhalis. In France their relative prevalences have remained stable over recent years: H. influenzae causes 40 to 50% of cases; S. pneumoniae, 26 to 30% of cases; and M. catarrhalis, only 5 to 10% of cases (15). The emergence of -lactamase-producing isolates of H. influenzae and M. catarrhalis has not substantially modified the management of AOM, since -lactamase inhibitors and -lactamase-resistant cephalosporins are now in use. However, pneumococcus isolates with decreased susceptibilities to penicillin (penicillin-resistant S. pneumoniae [PRSP]) have emerged, and their prevalence in France among cases of AOM has risen from about 20% in 1992 to 65.4% in 1995 (19). Furthermore, the levels of penicillin resistance have progressively increased (MICs for 75% of isolates are now 1 mg/liter). Consequently, the antibiotic therapy currently used for AOM is likely to fail with increasing frequency. Until recently, it proved difficult to correlate the levels of antibiotic resistance of bacteria in middleear fluid with the clinical failures of treatments, since the latter often occurred in children infected with isolates susceptible to the antibiotic administered. Mechanisms unrelated to the bacterial infection might explain some of the persisting clinical signs (age of <18 months, prior history of recurrent otitis media, and S. pneumoniae infection during the winter season) (4, 6, 7, 24). Nevertheless, it has been demonstrated that PRSP isolates are associated with an increased number of treatment failures (2, 10), especially when isolates express high-level resistance to penicillin (17).

Ceftriaxone is a parental expanded-spectrum cephalosporin with pharmacological characteristics (13, 21) and microbial activity particularly adapted to the treatment of AOM unresponsive to previous therapy (5, 11, 15). It has been shown that the concentrations of antibiotic in middle-ear fluid after a single intramuscular injection of 50 mg of ceftriaxone/kg of body weight (in children with documented chronic middle-ear effusion for at least 3 months, requiring insertion of tympanostomy tubes) peaked to 35 mg/liter at 24 h and remained at 19 mg/liter 48 h after the injection, with an estimated half-life of 25 h in middle-ear fluid (21). These results show that the maximum concentration (C_max) of ceftriaxone in middle-ear fluid was 35 to 580 times higher than the MICs at which 90% of the isolates are inhibited (MIC₉₀) for the three major causative pathogens of AOM, including PRSP. The time for which the concentration of ceftriaxone was above the MIC₉₀ was between 100 and >200 h.

The appropriate drug regimen has been determined previously in a phase II pilot study, with a small number of children with AOM: one daily intramuscular injection of 50 mg of ceftriaxone/kg for three consecutive days. In this pilot study, the clinical efficacy was as high as 95% (95% confidence interval [CI], 86 to 100%), including the group of patients infected by S. pneumoniae. The efficacy of this therapeutic regimen needed to be validated. Because there is no reference treatment for AOM with treatment failure, the in vivo kinetic technique recommended by Howie (22) and used by others, such as Dagan et al. (14), is considered the best approach for confirmation of the clinical efficacy of ceftriaxone. Although the correlation between clinical failure and bacteriological failure is not absolute (25), this approach provides evidence for early eradication of bacteria in middle-ear fluid. We used this approach in the present study to test the efficacy of ceftriaxone on PRSP isolates (penicillin MIC  1 mg/liter).

In this multicenter, noncomparative, nonrandomized study, we evaluated the clinical efficacy and safety of ceftriaxone, administered intramuscularly once daily (50 mg/kg/day) for 3 days to resolve AOM in children following clinical failure of previous oral antibiotic therapy. The eradication of S. pneumoniae for which the MIC of penicillin was 1 mg/liter and which was isolated on day 0 from middle-ear fluid was assessed by a second tympanocentesis on day 3, 4, or 5 (day 3 to 5).

(This work was presented in part at the 37th Interscience Conference on Antimicrobial Agents and Chemotherapy, Toronto, Canada, September 1997.)

	MATERIALS AND METHODS

Top Abstract Introduction Materials and methods Results Discussion References

Patients. This multicenter, noncomparative, nonrandomized study was approved by the Ethical Committee on 18 December 1995. The study was conducted in the Paris Region by 30 independent ear, nose, and throat specialists. One informed consent was obtained from the parents of the children upon entry into the study, and another was obtained if a second tympanocentesis was performed. Subjects included were children who had failed to respond to oral antibiotic treatment for AOM. Upon their inclusion in the study, samples were collected by tympanocentesis for bacteriological tests to determine the pathogens responsible for the AOM. Patients infected by a PRSP isolate (penicillin MIC  1 µg/ml) were identified and scheduled to undergo a second tympanocentesis on day 3 to 5 for bacteriological testing if middle-ear effusion persisted. This MIC was chosen for ethical reasons, to avoid a second tympanocentesis for patients with pneumococci for which penicillin MICs were low.

Criteria of inclusion and exclusion. This study included children of both sexes aged 5 months to 5 years, 10 months, who underwent outpatient treatment. They presented unilateral or bilateral AOM which failed to respond to oral antibiotic therapy. The children included had received treatment for at least 3 full days or had discontinued the antibiotic course less than 48 h before inclusion. Failure was evidenced by general symptoms which led to consultation: fever, otalgia, or nonspecific signs of AOM (irritability and/or night awakening and/or inconsolable crying). It was confirmed by otoscopic signs: a red, thickened, and bulging tympanic membrane, sometimes perforated and releasing pus (otorrhea), according to the schema proposed by Paradise (27). A middle-ear fluid specimen was collected from each patient included in the study, for bacteriological examination. Thirteen patients were excluded from analysis of the clinical response for the following reasons: antibiotic treatment for intercurrent diseases before day 10 (four patients), noninclusion of tympanocentesis in the protocol for a normal tympanic membrane (one patient), entry into the study 3 days before the first injection of ceftriaxone (one patient), prior antibiotic treatment lasting fewer than 3 days (one patient), tympanostomy tubes and otorrhea for more than 72 h upon entry into the study (one patient), ceftriaxone overdose (one patient), previous antibiotic therapy stopped 12 days before inclusion in the study (one patient), withdrawal of consent at step 2 (one patient), adenoidectomy at day 4 (one patient), and loss to follow-up before day 10 (one patient).

Study design. Four visits were scheduled. At the first, on day 0, ear fluid was sampled and treatment was begun. The second visit, on day 3 to 5, took place at least 12 h after the last ceftriaxone injection; a second tympanocentesis was performed if a pneumococcus isolate for which the penicillin MIC was 1 mg/liter had been obtained on day 0 and if the middle-ear effusion persisted on day 3 to 5. The clinical efficacy of the drug regimen was evaluated on day 10 to 12, and a late follow-up visit was scheduled for day 28 to 42 to evaluate the clinical response. Each patient was monitored by the same physician throughout the study.

Bacteriological studies. Pus was drawn from the ear by needle suction, either directly in cases of purulent otorrhea or immediately following a tympanocentesis, and was placed in Portagerm transport medium (bioMérieux, Marcy-l'Etoile, France). It was mailed with 24 h by special delivery to the Microbiology Department of the Hôpital Necker-Enfants Malades. A qualitative bacteriological analysis was performed immediately upon the arrival of the sample at the laboratory. Aliquots (100 µl) of the sample were plated onto each of the following media (and incubated under the conditions described): 5% sheep blood agar (bioMérieux) and Isovitalex (bioMérieux) chocolate agar plate (10% CO₂ at 37°C for 24 to 48 h); blood agar (anaerobiosis for 48 h at 37°C); enriching medium for S. pneumoniae (Trypticase broth; Diagnostics Pasteur, Marne-la-Coquette, France) containing 20 µg of gentamicin/ml (in air at 37°C for 4 to 18 h), followed by plating on 5% sheep blood agar). Each isolate of H. influenzae or M. catarrhalis was identified as previously described (26) and tested for -lactamase production by using nitrocefin disks (Cefinase; bioMérieux). Each isolate of S. pneumoniae was identified as previously described (26), and abnormal susceptibility to -lactam antibiotics was detected by using the E-test (0.02 to 32 mg/liter) (AB Biodisk, Solna, Switzerland) to determine the MIC of penicillin for S. pneumoniae. S. pneumoniae isolates were defined by the level of penicillin susceptibility as susceptible (MIC < 0.1 mg/liter), intermediate (MIC = 0.1 to 1 mg/liter), and resistant (MIC > 1 mg/liter). MICs were confirmed by the reference Mueller-Hinton agar dilution method (bioMérieux) (with 5% fresh lysed horse blood for S. pneumoniae) using serial dilutions of each antibiotic tested: penicillin, amoxicillin, and ceftriaxone for S. pneumoniae and amoxicillin, amoxicillin-clavulanate, and ceftriaxone for H. influenzae and M. catarrhalis. The susceptibility values were defined according to the standards of the Comité de l'Antibiogramme of the Société Française de Microbiologie (12) (penicillin and ceftriaxone breakpoints are 0.06 and 1 mg/liter and 0.05 and 2 mg/liter, respectively). Each S. pneumoniae and H. influenzae isolate was serogrouped by using immune serum from the Statens Seruminstitut (Copenhagen, Denmark). Bacterial isolates were frozen at 80°C for further investigations if necessary. Investigators were informed by fax whenever a pneumococcus isolate for which the penicillin MIC was 1 µg/ml, based on the E-test value, was obtained, so that they could perform a second tympanocentesis on day 3 to 5 as appropriate.

Study drug. Ceftriaxone (Ro 13-9904) presented as powder vials containing 1 g for parenteral use. Once dissolved, the drug volume injected should have been 0.2 ml/kg (i.e., an antibiotic concentration of 250 mg/ml). Ceftriaxone was injected by the investigator or a nurse as one intramuscular injection of 50 mg/kg/day (maximum daily dose, 1 g) for 3 days.

Evaluation criteria. Cases were considered clinical failures on day 10 to 12 if patients had persistent AOM with a bulging tympanic membrane or otorrhea, underwent an extra tympanocentesis, not scheduled in the protocol, between day 0 and day 10 to 12, or had received another antibiotic. Cures included all other cases. On day 28 to 42, cases considered clinical cures on day 10 to 12 were classified as follows: recurrence, intercurrent infection justifying another antibiotic therapy, cure, or secretory otitis media, which is considered compatible with AOM cure. Bacteriological eradication was defined as the absence of pneumococcus, as assessed by culture, from the second tympanocentesis sample, which was collected on day 3 to 5. Eradication of S. pneumoniae isolates for which penicillin MICs were 1 mg/liter was the main evaluation criterion for the efficacy of treatment in the population with AOM. Intramuscular injections of ceftriaxone were performed by nurses who carefully monitored compliance with the antibiotic treatment.

Statistical analysis. The number of patients to be included in the study was calculated by using epidemiological data for AOM (30 to 40% due to S. pneumoniae) and the frequency (70%) of PRSP isolates (3). Satisfactory estimation of the bacteriological eradication rate of these pneumococci by ceftriaxone required at least 25 patients infected with PRSP isolates (penicillin MIC  1 mg/liter), and thus inclusion of approximately 200 patients with persistent AOM. The analysis was descriptive; qualitative variables are given as numbers of cases and percentages, and quantitative variables are given as numbers of subjects, means and standard deviations, minima, maxima, and medians. Data analyses were performed with SAS Institute statistical software. The statistical independence of PRSP (penicillin MIC  1 mg/liter) upon inclusion in the study was verified between the subgroup of patients with a clinical cure outcome at day 10 to 12 and the subgroup of patients with a clinical failure outcome at day 10 to 12 by using the same tests. For qualitative variables, clinical cure rates on day 10 to 12 and on day 28 to 42 were given with their 95% CIs, calculated by using the binomial probability law. Rates of bacteriological eradication on day 3 to 5 were calculated with their 95% CIs. An independent committee of experts verified the inclusion and exclusion criteria for every patient. In addition, every questionable case was reevaluated individually at the end of the study.

	RESULTS

Top Abstract Introduction Materials and methods Results Discussion References

Clinical efficacy. Between 9 January and 16 July 1996, 187 patients were enrolled in the study. Each investigator enrolled one to six patients (mean, 6; median, 3 [95% CI, 1.26]). Only one patient did not receive ceftriaxone because the parents withdrew consent. One hundred eighty-six patients were included and treated for 3 days with ceftriaxone. Of these, 13 (7.0%) were excluded from analysis of the clinical response. The mean age of the children evaluated for the clinical efficacy criteria was 17.7 ± 13.2 months (median, 13 months; range, 5 to 70 months).

Clinical efficacy according to the bacteriological data. Middle-ear fluid sample cultures were sterile for 60 of 186 patients (32.3%) and positive for 126 patients (67.7%) at the time of enrollment in the study. Samples from three patients were contaminated. Among the 123 patients with positive cultures, 59 (46.8%) were infected with S. pneumoniae, 56 (44.4%) with H. influenzae, and 8 (8.8%) with M. catarrhalis. S. pneumoniae and H. influenzae isolates were associated with another species in 26.9 and 31.1% of the samples, respectively. H. influenzae was associated with S. pneumoniae in 68.4% of the samples. Among the 67 pneumococcus isolates, 77.6% were PRSP isolates (penicillin MIC  0.125 µg/ml, as determined by the agar dilution method) and were isolated in 46 patients. Serogroups 23, 14, and 19 were the most prevalent, accounting for 25.4% (93.8% PRSP), 25.4% (93.8% PRSP), and 25.8% (73.3% PRSP) of isolates, respectively, and included 79% of all the PRSP isolates. For 43 of the S. pneumoniae isolates (36 patients), the penicillin MIC was 1 mg/liter; of these, 31 isolates (26 patients) were highly resistant to penicillin (MIC > 1 mg/liter) and none were resistant to ceftriaxone (MIC₉₀ = 1 mg/liter). Among the 61 H. influenzae isolates, 49.2% were -lactamase producers and were isolated in samples from 27 patients. The bacteriological data and the risk factors have been reported in detail elsewhere (18).

Eradication of PRSP isolates. Bacterial eradication was evaluated for 36 patients infected by S. pneumoniae isolates for which the penicillin MICs were 1 mg/liter. Nine patients were thereafter excluded from the evaluation of bacteriological eradication (three due to bacterial contamination, two due to unproductive tympanocentesis, and four because the second sample was not done). All nine were cured by day 10 to 12, and seven patients were still considered cured by day 28 to 42 (one was considered to have a recurrence and the other had an intercurrent infection. Among the 27 remaining patients, 24 had no pneumococci in the sample collected on day 3 to 5 (88.9% [95% CI, 70.8 to 97.7%]). No eradication of S. pneumoniae was observed for three patients, whose cases were considered bacteriological failures. Two of these patients were evaluated as clinically cured on day 10 to 12, although each was infected with a resistant isolate (penicillin MIC = 2 mg/liter; ceftriaxone MIC = 1 mg/liter), one belonging to serogroup 23 and the other to serogroup 14. The third patient was infected by a serogroup 19 isolate for which penicillin and ceftriaxone MICs were estimated at 1 and 0.5 mg/liter, respectively. This child had otoscopic signs of AOM associated with spontaneous otorrhea by day 10, and the case was considered a clinical and bacteriological failure. By day 28 to 42, 18 (78.3%) of the 23 patients with early clinical cures remained free of symptoms of AOM. One patient (4.3%) infected by day 0 with a serogroup 9 penicillin-resistant isolate (MIC = 2 mg/liter) which was eradicated by day 3 to 5 had secretory otitis media. The other four patients (17.4%) were considered to have recurrences.

Safety and compliance. All 186 patients included in the study and treated received the three scheduled injections (50 mg/kg/day). There was no case of discontinuation of treatment due to adverse events, particularly injection site reactions, spontaneously reported by the investigator. Eleven (12.6%) of the adverse events had doubtful, possible, or probable links with the study drug according to the physicians: diarrhea (seven patients), otorrhea (two patients), rash (one patient), and headache one patient. Only one reaction at the injection site was reported (hematoma). One hundred eighty-seven patients were evaluated for compliance. Only one patient did not receive the ceftriaxone injections, because the parents withdrew consent. Therefore, 186 patients were included and treated, and all patients received three intramuscular injections of ceftriaxone. No treatment was discontinued.

	DISCUSSION

Top Abstract Introduction Materials and methods Results Discussion References

This study was designed to evaluate the efficacy of ceftriaxone for treatment of AOM in children following clinical failure of oral antibiotic therapy. We used the dosage recommended for childhood infections (50 mg/kg/day), which had previously been found successful in cases of AOM (1, 8, 9, 20, 23, 29), during a 3-day course of once-daily intramuscular injection defined by a pilot clinical trial performed with a range of doses and with a small number of patients. Among the 173 evaluable patients, 145 (83.8%) were cured. The clinical efficacy by day 28 to 42 was 81.2% for S. pneumoniae infections, 69.4% for H. influenzae infections, and 100% for M. catarrhalis infections. Because at present there is no reference comparator for this indication, no comparison of our data could be performed. The relative prevalences of various pathogens were consistent with previous data found for AOM after clinical failure of antibiotic therapy (10). Bacterial cultures from 60 patients (32.3%) were negative, a rate lower than those previously found for patients with clinical failure of treatment for AOM (~50%) (10, 28). Negative cultures might correspond to patients with frequent recurrent episodes who have recently been treated with antibiotic therapy and who display inflammatory reactions without detectable bacteria. This might be due to inaccessible bacteria replicating at low levels and present in small numbers in host tissues. For this group, too, ceftriaxone was successful (75.5% efficacy).

The penicillin MIC₅₀ and MIC₉₀ for S. pneumoniae isolates were, respectively, 1 and 2 mg/liter; those of ceftriaxone were 0.5 and 1 mg/liter, and those of amoxicillin were both 1 mg/liter. The ceftriaxone MIC₅₀ and MIC₉₀ for H. influenzae -lactamase producers and nonproducers were identical, 0.0075, mg/liter, and the MIC₅₀ and MIC₉₀ of amoxicillin-clavulanic acid were 0.25 and 0.5 mg/liter, respectively. The ceftriaxone MIC₅₀ and MIC₉₀ for S. pneumoniae were similar to those reported in the literature (13, 16), i.e., 0.5 mg/liter (for isolates with intermediate penicillin susceptibility, the MIC was 0.1 to 1 mg/liter) and 1 mg/liter (for penicillin-resistant isolates, the MIC was >1 mg/liter), respectively. As expected, the rate of PRSP isolates was very high in our study (77.6%). Thus, we could demonstrate the remarkable efficacy of ceftriaxone in vivo by performing a second paracentesis by day 3 to 5 on patients with S. pneumoniae isolates for which MICs were 1 mg/liter. Among the 27 evaluable patients, eradication was obtained at a rate estimated at 88.9%. This bacterial elimination was associated with clinical success in these patients (96.2% by day 8 to 10 and 78.3% by day 28 to 42). These results are important in view of the growing incidence of clinical failures in AOM treatment (10), associated with the increasing rate of PRSP (14, 17). There is, therefore, a need to identify by paracentesis the bacterial pathogens responsible for clinical failures of AOM treatment before starting a new antibiotic treatment. In conclusion, whatever the etiological agent and the antibiotic previously prescribed, the clinical and bacteriological results obtained in this noncomparative study show that ceftriaxone is of benefit for patients with AOM for whom previous oral treatment has failed. Ceftriaxone is particularly useful for treatment of AOM caused by PRSP, which is frequently responsible for the failure of a first-intent treatment, as confirmed by our study. Our data indicate that ceftriaxone at the dose of 50 mg/kg/day, injected intramuscularly once daily for 3 days, should be recommended for the treatment of AOM in children failing to respond clinically to a previous antibiotic therapy, particularly when PRSP isolates are involved.