Large Dosage Amoxicillin/Clavulanate, Compared With Azithromycin, for the Treatment of Bacterial Acute Otitis Media in Children
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
ORIGINAL STUDIES
Large Dosage Amoxicillin/Clavulanate, Compared With
Azithromycin, for the Treatment of Bacterial Acute Otitis
Media in Children
Alejandro Hoberman, MD,* Ron Dagan, MD,† Eugene Leibovitz, MD,† Andres Rosenblut, MD,‡
Candice E. Johnson, MD,§ Anne Huff, PhD,储 Rajesh Bandekar, PhD,储 and Brian Wynne, MD储
Conclusion: Amoxicillin/clavulanate was clinically and bacterio-
Background: A large dosage pediatric formulation of amoxicillin/
logically more effective than azithromycin among children with
clavulanate with an improved pharmacokinetic/pharmacodynamic
bacterial AOM, including cases caused by penicillin-resistant S.
profile was developed to eradicate many penicillin-resistant strains pneumoniae and -lactamase-positive H. influenzae.
of Streptococcus pneumoniae and Haemophilus influenzae (includ-
ing -lactamase-producing strains). Key Words: extra-strength amoxicillin/clavulanate, azithromycin,
Methods: This randomized, investigator-blinded, multicenter trial bacterial acute otitis media, clinical trial
examined treatment of bacterial acute otitis media (AOM) in chil- (Pediatr Infect Dis J 2005;24: 525–532)
dren 6 –30 months of age with amoxicillin/clavulanate (90/6.4 mg/
kg/d in 2 divided doses for 10 days) versus azithromycin (10 mg/kg
for 1 day followed by 5 mg/kg/d for 4 days). Tympanocentesis was
performed at entry for bacteriologic assessment, at the on-therapy
visit (day 4 – 6) to determine bacterial eradication and at any time
before the end-of-therapy visit (day 12–14) if the child was catego-
I n recent years, an important factor affecting the manage-
ment of acute otitis media (AOM) has been the increasing
development of antimicrobial resistance among the common
rized as experiencing clinical failure. Clinical assessments were pathogens that cause AOM, particularly the worldwide prev-
performed at the on-therapy, end-of-therapy and follow-up (day alence of drug-resistant Streptococcus pneumoniae. In 1985,
21–25) visits. ⬃90% of S. pneumoniae strains isolated from patients with
Results: We enrolled 730 children; AOM pathogens were isolated at AOM were susceptible to penicillin (minimal inhibitory con-
baseline for 249 of the amoxicillin/clavulanate group and 245 of the centration 关MIC兴 of ⱕ0.06 g/mL); by 1997, the suscepti-
azithromycin group. For children with AOM pathogens at baseline, bility of S. pneumoniae to penicillin had decreased to 43%.1
clinical success rates at the end-of-therapy visit were 90.5% for Of greater concern has been the appearance of S. pneumoniae
amoxicillin/clavulanate versus 80.9% for azithromycin (P ⬍ 0.01), isolates with resistance to multiple classes of antimicrobial
and those at the on-therapy and follow-up visits were 94.9% versus agents, including macrolides. Among isolates of Haemophi-
88.0% and 80.3% versus 71.1%, respectively (all P ⬍ 0.05). At the lus influenzae and Moraxella catarrhalis, -lactamase pro-
on-therapy visit, pretherapy pathogens were eradicated for 94.2% of duction rates have increased from 16% and 88% of strains
children receiving amoxicillin/clavulanate versus 70.3% of those collected before 1985 to 44% and 96% of those collected in
receiving azithromycin (P ⬍ 0.001). Amoxicillin/clavulanate erad- 1997, respectively.1
icated 96.0% of S. pneumoniae (92.0% of fully penicillin-resistant A large dosage amoxicillin/clavulanate formulation
S. pneumoniae) and 89.7% of H. influenzae (85.7% 关6 of 7 cases兴 of (90/6.4 mg/kg/d; Augmentin ES-600; GlaxoSmithKline,
-lactamase-positive H. influenzae). Corresponding rates for Durham, NC) was approved in the United States for the
azithromycin were 80.4% (54.5%) for S. pneumoniae and 49.1% treatment of children with recurrent and persistent cases of
(100% 关1 of 1 case兴) for H. influenzae (all P ⬍ 0.01 for between- AOM, particularly those for whom S. pneumoniae with penicil-
drug comparisons). lin MICs of ⱕ2 g/mL or -lactamase-positive H. influenzae or
M. catarrhalis are suspected.2–5 Because this formulation and
azithromycin are frequently prescribed for the treatment of
children with this condition, we evaluated their bacteriologic and
Accepted for publication January 12, 2005.
From the *Children’s Hospital of Pittsburgh, Pittsburgh, PA; †Soroka Uni-
clinical efficacy in a randomized clinical trial designed to dem-
versity Medical Center and Ben Gurion University, Beer Sheva, Israel; onstrate statistically significant differences in efficacy between
‡Hospital Santiago del Rio, Santiago, Chile; §Children’s Hospital and treatment groups in children 6 –30 months of age.
University of Colorado School of Medicine, Denver, CO; and 㛳Glaxo-
SmithKline, Collegeville, PA.
Supported by GlaxoSmithKline. METHODS
Address for reprints: Alejandro Hoberman, MD, Children’s Hospital of Design. This randomized, investigator-blind, multicenter study
Pittsburgh, 3705 Fifth Avenue, Pittsburgh, PA 15213-2583. Fax 412-
692-5807; E-mail hoberman@chp.edu. was conducted in 34 centers in Bulgaria, Chile, the Dominican
Copyright © 2005 by Lippincott Williams & Wilkins Republic, Guatemala, Israel, Peru (1 center each), Romania (2
ISSN: 0891-3668/05/2406-0525 centers), Latvia (3 centers), Mexico (4 centers) and the United
DOI: 10.1097/01.inf.0000164794.50281.1a States (19 centers) from April 2001 to November 2002. The
The Pediatric Infectious Disease Journal • Volume 24, Number 6, June 2005 525Hoberman et al The Pediatric Infectious Disease Journal • Volume 24, Number 6, June 2005
protocol was approved by national or local ethics committees or infection and symptoms of inflammation after ⱖ72 hours of
the institutional review board for each center. Informed consent therapy, such that additional antibiotic therapy was pre-
was obtained from parents before study entry. scribed for AOM, or if a valid assessment of the clinical
Subjects. Children 6 –30 months of age were eligible for outcome could not be made. This group included children
inclusion in the study if they had AOM, based on the who did not return for assessment at this visit. Similar criteria
presence of purulent otorrhea for ⬍24 hours or middle ear were used for the on-therapy visit. At the follow-up visit,
effusion evidenced by at least 2 of the following tympanic children were classified as experiencing clinical success if
membrane findings: (1) yellow or white discoloration, (2) they demonstrated a persistent clinical cure and as experienc-
opacification or (3) decreased or absent mobility measured ing clinical failure if they had a recurrence of AOM that
with pneumatic otoscopy. Evidence of middle ear effusion required additional antibiotic therapy. This category of clin-
needed to be accompanied by ⱖ1 of the following signs of ical failure also included children categorized as experiencing
inflammation: (1) ear pain (including unaccustomed tugging clinical failure at the on-therapy or end-of-therapy visits and
of the ear), (2) marked erythema or (3) distinct fullness or those who did not return for the assessment.
bulging of the tympanic membrane.5 Bacteriologic response at the on-therapy and end-of-
Children were excluded from the study for the follow- therapy visits was defined as success if there was eradication
ing reasons: weight of ⬎40 kg, otorrhea for ⬎24 hours, of the initial AOM pathogen (with or without the presence of
presence of tympanostomy tubes or anatomic abnormalities a new pathogen), as demonstrated with repeat tympanocen-
associated with prolonged middle ear effusion, serious un- tesis, or a lack of middle ear fluid, as determined through
derlying disease, hypersensitivity to study medications or either observation or tympanocentesis. Children were classi-
concomitant use of medications known to interact with the fied as experiencing bacteriologic failure if the initial patho-
study drugs. We also excluded children who had received gen persisted in the middle ear fluid at the on-therapy visit or
systemic antibiotic therapy within 72 hours, except when before the end-of-therapy visit.
AOM developed while the patient was undergoing prophy- Bacterial Isolation and Susceptibility Testing. Pathogens re-
lactic antimicrobial therapy for AOM. covered from middle ear fluid at local laboratories were sent
Interventions and Assessments. We randomized (1:1) eligible to a central laboratory for confirmation of identification and
children to receive either amoxicillin/clavulanate (90/6.4 mg/ susceptibility testing. Determination of the penicillin, amoxi-
kg/d in 2 divided doses for 10 days) or azithromycin (10 mg/kg cillin, amoxicillin/clavulanate, erythromycin and azithromy-
for 1 day, followed by 5 mg/kg/d for 4 days), and we conducted cin MICs for the 4 protocol-defined AOM pathogens was
baseline (day 1), on-therapy (day 4 – 6), end-of-therapy (day performed with broth microdilution methods, according
12–14) and follow-up (day 21–25) clinical assessments. Efficacy to National Committee for Clinical Laboratory Standards
assessments were conducted for children with bacterial AOM. (NCCLS) guidelines6 and guidelines of the plate manufac-
For bacteriologic assessments, we performed tympanocentesis turer (Trek Diagnostics, Cleveland, OH). MIC results were
for all children with an intact tympanic membrane at the initial interpreted according to NCCLS guidelines,7 and isolates
visit. For children with otorrhea for ⬍24 hours, direct sampling were classified as susceptible, intermediate or resistant to
of otorrhea or tympanocentesis was performed. We considered each drug. Because no NCCLS breakpoints for amoxicillin/
S. pneumoniae, H. influenzae, M. catarrhalis and Streptococcus clavulanate (90/6.4 mg/kg/d) have been assigned, the break-
pyogenes AOM pathogens. A second sample of middle ear fluid point for conventional formulations was used. Isolates of H.
was obtained at the on-therapy visit from children who (1) had influenzae and M. catarrhalis were tested for -lactamase
S. pneumoniae at baseline, (2) or had AOM pathogen(s) at production with Cefinase disks (Becton Dickinson, Sparks,
baseline and either experienced clinical failure or withdrew from MD), according to the manufacturer’s instructions.
the study before the end of therapy. Exceptions included a center Statistical Analyses. We assumed overall clinical success
in Israel, where repeat tympanocentesis was performed at the rates at the end of therapy of 86% for patients treated with
on-therapy visit for all children with AOM pathogens at base- amoxicillin/clavulanate and 75% for patients treated with
line, and a center in the United States, where on-therapy tym- azithromycin.8 Accordingly, a total of 203 children per treat-
panocentesis was performed only for children with non– ment arm were required to detect a difference with 80%
penicillin-susceptible S. pneumoniae (MIC ⱖ 0.12 g/mL) at power at a significance level of 0.05. With the assumptions
baseline. that ⬃70% of all middle ear fluid sample cultures would yield
Outcomes. Compliance was assessed on the basis of medica- a pathogen and 80% of children who had a pathogen would
tion diaries completed by the parents and comparisons of the qualify for the per-protocol population, 203/(0.7 ⫻ 0.8)
amount of medication dispensed with that returned. Clinical resulted in 363 children required per treatment arm. The
response at the end of therapy (day 12–14) among children treatment difference and 95% confidence intervals (CIs) were
with bacterial AOM represented the primary efficacy end- determined for the difference between treatment groups at
point for the study. Children were classified as experiencing each endpoint. Fisher’s exact test or continuity-corrected 2
clinical success if they had lessening or complete resolution test P values were calculated as appropriate.
of otoscopic signs of acute infection and symptoms of in-
flammation, with or without middle ear effusion, such that no
additional antibiotic therapy was needed for AOM. Children RESULTS
were assigned a clinical response of failure if there was an Subjects. We randomized 731 children to amoxicillin/clavu-
inability to resolve or improve the otoscopic signs of acute lanate (368 children) or azithromycin (363 children). One child
526 © 2005 Lippincott Williams & WilkinsThe Pediatric Infectious Disease Journal • Volume 24, Number 6, June 2005 Amoxicillin/Clavulanate for AOM
FIGURE 1. Flow diagram showing
progression of children through the
study and reasons for exclusion
from analyses. Two children receiv-
ing amoxicillin/clavulanate and 4
children receiving azithromycin
who were evaluated at the end-of-
therapy visit were not evaluated at
the on-therapy visit.
in the amoxicillin/clavulanate group did not receive study med- at baseline, ie, 249 children in the amoxicillin/clavulanate
ication; therefore, 730 children were included in the safety group and 245 in the azithromycin group. Nineteen (7.6%)
analysis. The flow of children in the study is shown in Figure 1. children in the amoxicillin/clavulanate group and 38 (15.5%)
The 2 treatment groups were well matched with respect to in the azithromycin group had ⬎1 pathogen. No differences
demographic characteristics and risk factors for AOM (Table 1). between treatment groups were apparent in the distribution of
In general, children in both treatment groups were relatively pathogens at baseline. Overall H. influenzae was the most
young (mean age, 15.1 and 15.2 months for the amoxicillin/ frequently isolated pathogen (121 of 249 关48.6%兴 children in
clavulanate and azithromycin groups, respectively); 87.6% of the amoxicillin/clavulanate group and 124 of 245 关50.6%兴
children were ⱕ24 months of age. Approximately two-thirds of children in the azithromycin group). This trend was observed
the children had siblings, one-third attended day-care centers, in the United States, in Israel and overall. -Lactamase
one-third had AOM and received systemic antimicrobial therapy production was observed in 23.7% of H. influenzae isolates
in the 30 days before enrollment and one-third were exposed to and 96.8% of M. catarrhalis isolates from the combined
tobacco smoke pollution at home. amoxicillin/clavulanate and azithromycin treatment groups.
Bacteriologic Findings. Four hundred ninety-four (67.7%) of Of S. pneumoniae isolates (111 of 249 关44.6%兴 children in the
the 730 children had ⱖ1 protocol-defined pathogen isolated amoxicillin/clavulanate group and 118 of 245 关48.2%兴 chil-
© 2005 Lippincott Williams & Wilkins 527Hoberman et al The Pediatric Infectious Disease Journal • Volume 24, Number 6, June 2005
TABLE 1. Demographic Characteristics and Risk Factors for AOM at Baseline for Children
With Bacterial AOM (per Protocol) and All Children
Evaluable Children With
All Children
Bacterial AOM (per Protocol)
Characteristic/Risk Factor for AOM
A/C AZI A/C AZI
(N ⫽ 200) (N ⫽ 204) (N ⫽ 367) (N ⫽ 363)
Gender (male) 110 (55.0)* 129 (63.2) 196 (53.4) 217 (59.8)
Race
White 87 (43.5) 93 (45.6) 176 (48.0) 179 (49.3)
Black 21 (10.5) 16 (7.8) 41 (11.2) 31 (8.5)
Other† 92 (46.0) 95 (46.6) 150 (40.9) 153 (42.1)
Age (mo)
Mean ⫾ SD 15.1 ⫾ 6.6 15.2 ⫾ 6.6 15.3 ⫾ 6.8 14.9 ⫾ 6.5
Range 6 –32 6 –32 6 –32 5–36
ⱕ24 mo 173 (86.5) 181 (88.7) 317 (86.4) 325 (89.5)
Weight (kg)
Mean ⫾ SD 10.32 ⫾ 2.18 10.29 ⫾ 2.16 10.35 ⫾ 2.30 10.22 ⫾ 2.31
Range 6.6 –17.5 5.9 –17.7 6.6 –29.0 5.4 –29.5
Siblings 136 (68.0) 137 (67.2) 227 (61.9) 243 (66.9)
Attends day care 64 (32.0) 75 (36.8) 127 (34.6) 137 (37.7)
Systemic antimicrobial therapy within 30 d 62 (31.0) 56 (27.5) 109 (29.7) 109 (30.0)
Smoking in household 73 (36.5) 84 (41.2) 137 (37.3) 146 (40.2)
Currently breast-feeding 49 (24.5) 51 (25.0) 71 (19.3) 74 (20.4)
Previously breast-fed 156 (78.0) 161 (78.9) 276 (75.2) 277 (76.3)
AOM within 30 d 62 (31.0) 55 (27.0) 120 (32.7) 113 (31.1)
*Numbers in parentheses, percent.
†
Mostly Hispanic, a few Native American and 1 Middle Eastern.
A/C indicates amoxicillin/clavulanate; AZI, azithromycin.
dren in the azithromycin group), 48.5%, 11.4% and 20.5% Children were classified as experiencing bacteriologic failure
were nonsusceptible to penicillin, amoxicillin and azithromy- if the initial pathogen persisted in the middle ear fluid at the
cin, respectively. Of penicillin-susceptible and -intermediate on-therapy visit or before the end-of-therapy visit.
S. pneumoniae, 4.7% and 17.1%, respectively, were resistant Among the 183 children categorized as experiencing
to azithromycin. Of penicillin-resistant S. pneumoniae bacteriologic success at day 4 – 6, 24.8% (25 of 101 children)
(PRSP) (⬎1.0 g/mL) isolates, 52.2% were also resistant to in the amoxicillin/clavulanate group and 25.6% (21 of 82
azithromycin. children) in the azithromycin group were so categorized
Bacteriologic Efficacy. Determination of bacteriologic re- because of the absence of middle ear effusion, whereas 75.2%
sponses was limited to children from whom pathogens were (76 of 101 children) in the amoxicillin/clavulanate group and
isolated at baseline and for whom the study protocols at the 74.4% (61 of 82 children) in the azithromycin group were so
various study sites called for repeat tympanocentesis on day categorized on the basis of bacteriologic findings for middle
4 – 6. As described previously, repeat tympanocentesis was to ear fluid specimens. Among the children for whom bacteri-
be performed for children who in baseline tympanocentesis ologic responses were determined, amoxicillin/clavulanate,
assessments had the following: (1) in Israel, any of the 4 compared with azithromycin, demonstrated significantly
AOM pathogens isolated; (2) at all other sites except Chil- higher rates of eradication of AOM pathogens (101 of 108
dren’s Hospital of Pittsburgh, any S. pneumoniae isolated; cases 关93.5%兴 versus 82 of 127 cases 关64.6%兴; P ⬍ 0.001), of
(3) at Children’s Hospital of Pittsburgh, any nonsusceptible S. S. pneumoniae (72 of 75 cases 关96.0%兴 versus 74 of 92 cases
pneumoniae isolated. In addition, repeat tympanocentesis was 关80.4%兴; P ⬍ 0.01) and of H. influenzae (29 of 30 cases
performed at all sites in the event of clinical failure occurring 关96.7%兴 versus 18 of 34 cases 关52.9%兴; P ⬍ 0.001). More
before the end-of-therapy visit (day 12–14). However, when pronounced differences between treatment groups, again fa-
repeat tympanocentesis was performed because of clinical voring amoxicillin/clavulanate, were noted in the eradication
failure, rather than because the child met the aforementioned of PRSP (23 of 25 cases 关92.0%兴 versus 12 of 22 cases
criteria at the various sites, results were not included in 关54.5%兴; P ⬍ 0.05) and of azithromycin-resistant S. pneu-
bacteriologic efficacy analyses, to preclude bias that would moniae (14 of 15 cases 关93.3%兴 versus 6 of 18 cases 关33.3%兴;
result from the inclusion only of children who experienced P ⬍ 0.05). Rates of eradication of S. pneumoniae and H.
clinical failure (see below). Therefore, given the particular influenzae according to children’s treatment groups and
design of this study, S. pneumoniae was the pathogen for amoxicillin/clavulanate MICs are presented in Table 2. Of
which bacteriologic results were most frequently available. S. pneumoniae isolates with amoxicillin/clavulanate MICs
Bacteriologic response at the on-therapy and end-of-therapy of ⬍2 g/mL, amoxicillin/clavulanate eradicated 98.0% (51
visits was defined as success if there was eradication of the of 52 isolates) and azithromycin eradicated 87.5% (63 of 72
initial AOM pathogen (with or without the presence of a new isolates) (P ⬍ 0.05). Of S. pneumoniae isolates with amoxi-
pathogen), as demonstrated with repeat tympanocentesis, or cillin/clavulanate MICs of 2 g/mL, amoxicillin/clavulanate
the absence of middle ear fluid in otoscopic examinations. eradicated 92.3% (12 of 13 isolates) and azithromycin erad-
528 © 2005 Lippincott Williams & WilkinsThe Pediatric Infectious Disease Journal • Volume 24, Number 6, June 2005 Amoxicillin/Clavulanate for AOM
TABLE 2. Eradication of Streptococcus pneumoniae and
experiencing clinical failure was noted among those treated
Haemophilus influenzae According to Amoxicillin/
with azithromycin, compared with those treated with amoxi-
Clavulanate MIC and Treatment Group
cillin/clavulanate.
Clinical Efficacy. Among evaluable children with AOM
No. of Isolates Eradicated/Total No. of Isolates* pathogens at baseline, clinical responses at the end-of-therapy
A/C MIC S. pneumoniae H. influenzae
visit (primary efficacy endpoint) were significantly greater for
(g/mL) (n ⫽ 167) (n ⫽ 94) amoxicillin/clavulanate than for azithromycin, ie, 90.5% and
A/C AZI A/C AZI
80.9%, respectively (treatment difference, 9.62; 95% CI
2.37–16.87; P ⬍ 0.01) (Table 4). Significantly greater clinical
ⱕ0.015 2/2 4/4 success rates favoring amoxicillin/clavulanate were also
0.03 31/32 40/43
0.06 3/3 4/5 noted at the on-therapy visit (treatment difference, 6.95; 95%
0.12 6/6 3/6 CI 1.01–12.89) and at the follow-up visit (treatment differ-
0.25
0.5
3/3
4/4
9/10
1/2
2/2
18/20
1/2
19/37
ence, 9.26; 95% CI 0.19 –18.33) (P ⬍ 0.05 for each compar-
1 2/2 2/2 11/13 3/10 ison) (Table 4). For children with H. influenzae isolated at
2 12/13 5/9 3/3 4/4 baseline, marked differences between treatment groups, fa-
4 6/7 5/9 1/1 0/1
8 2/2 1/2 0/1 voring amoxicillin/clavulanate, were apparent in clinical suc-
16 1/1 cess rates at the end-of-therapy visit (day 12–14) (Table 5). A
32
⬎32
trend for greater clinical efficacy with amoxicillin/clavulanate
Total 72/75 (96)† 74/92 (80.4) 35/39 (89.7) 27/55 (49.1) among children with PRSP and azithromycin-resistant S.
*Results are provided for individual pathogens rather than for individual patients; ⬎1 pneumoniae strains was noted, although differences between
pathogen was isolated for 43 children.
†
treatment groups did not reach statistical significance (Table
Numbers in parentheses, percent.
A/C indicates amoxicillin/clavulanate; AZI, azithromycin.
5). When all children with bacterial AOM were included,
differences between treatment groups did not reach statistical
significance (Table 4). Among children for whom no AOM
icated 55.6% (5 of 9 isolates) (P ⫽ 0.16). Of S. pneumoniae pathogens were isolated at baseline, clinical success rates did
isolates with amoxicillin/clavulanate MICs of ⱖ4 g/mL, not differ significantly between the 2 treatment groups, either
amoxicillin/clavulanate eradicated 90.0% (9 of 10 isolates), at the end-of-therapy visit on day 12–14 (81.4% 关96 of 118
including 2 of 2 isolates with MICs of 8 g/mL and 1 isolate children兴 in both groups) or at the follow-up visit on day
with an MIC of 16 g/mL, and azithromycin eradicated 21–25 (68.6% 关81 of 118 children兴 in the amoxicillin/clavu-
54.5% (6 of 11 isolates) (P ⫽ 0.15). lanate group versus 76.3% 关90 of 118 children兴 in the azithro-
Results of repeat tympanocentesis performed when mycin group).
children were categorized as experiencing clinical failure at Relationship Between Bacterial Eradication at the On-
any time up to day 12–14 (end-of-therapy visit) are presented Therapy Visit and Clinical Outcomes at the End-of-
in Table 3. A greater proportion of children categorized as Therapy Visit. Among children from whom S. pneumoniae
and/or H. influenzae were recovered at baseline and who
experienced bacteriologic failure at the on-therapy visit (ie,
TABLE 3. Bacteriologic Outcomes for Children cultures remained positive), clinical failure at the end-of-
Categorized as Experiencing Clinical Failure Before therapy visit occurred for 5 of 7 children (71.4%) treated with
the End-of-Therapy Visit (Day 12–14) amoxicillin/clavulanate and 20 of 49 children (40.8%) treated
with azithromycin. Bacteriologic success at the on-therapy
A/C* AZI visit was associated with clinical success at the end-of-
Pathogen(s)
n % n % therapy visit for 96 of 105 children (91.4%) treated with
amoxicillin/clavulanate and 80 of 89 children (89.9%) treated
Clinical failure 19/200 9.5 39/204 19.1
with azithromycin. These relationships were similar for chil-
All pathogens
Eradication 12 70.6 8 26.7
dren from whom PRSP was recovered at baseline. Therefore,
Persistence 5 29.4 22 73.3 clinical failure at the end-of-therapy visit and bacteriologic
Total 17 100 30 100 failure at the on-therapy visit occurred for 1 of 2 children
Streptococcus pneumoniae (50%) treated with amoxicillin/clavulanate and 5 of 12 chil-
Eradication 7 77.8 5 45.5
Persistence 2 22.2 6 54.5 dren (42%) treated with azithromycin; clinical success at the
Total 9 100 11 100 end-of-therapy visit and bacteriologic success at the on-
Haemophilus influenzae therapy visit occurred for 20 of 22 children (90.9%) treated
Eradication 3 50.0 2 11.1
Persistence 3 50.0 16 88.9
with amoxicillin/clavulanate and 10 of 10 children (100%)
Total 6 100 18 100 treated with azithromycin.
Moraxella catarrhalis Treatment Compliance. Rates of compliance with study med-
Eradication 2 100 1 100 ications were high in both treatment groups (amoxicillin/
Total 2 100 1 100
clavulanate, 85.8%; azithromycin, 91.3%). Clinical success
*Two children in the amoxicillin/clavulanate group and 9 children in the azithromycin
group did not undergo repeat tympanocentesis.
rates at the end of therapy for children who were noncom-
A/C indicates amoxicillin/clavulanate; AZI, azithromycin. pliant with study medication were 38.5% (20 of 52 children)
© 2005 Lippincott Williams & Wilkins 529Hoberman et al The Pediatric Infectious Disease Journal • Volume 24, Number 6, June 2005
TABLE 4. Clinical Efficacy of Amoxicillin/Clavulanate and Azithromycin According to Study Visit
Treatment
Population A/C (%) AZI (%) P
Difference
Evaluable children with bacterial AOM
On-therapy visit (d 4 – 6) 94.9 (188/198)* 88.0 (176/200)* 6.95 (1.01–12.89)† ⬍0.05
End-of-therapy visit (d 12–14) 90.5 (181/200) 80.9 (165/204) 9.62 (2.37–16.87) ⬍0.01
Follow-up visit (d 21–25) 80.3 (155/193) 71.1 (135/190) 9.26 (0.19 –18.33) ⬍0.05
All children with bacterial AOM
On-therapy visit (d 4 – 6) 89.6 (223/249) 83.3 (204/245) 6.29 (⫺0.13 to 12.72) NS
End-of-therapy visit (d 12–14) 80.7 (201/249) 73.5 (180/245) 7.25 (⫺0.54 to 15.05) NS
Follow-up visit (d 21–25) 68.3 (170/249) 62.0 (152/245) 623 (⫺2.56 to 15.02) NS
*Numbers in parentheses, n/N.
†
Numbers in parentheses, 95% CI.
A/C indicates amoxicillin/clavulanate; AZI, azithromycin; NS, not significant.
TABLE 5. Clinical Efficacy of Amoxicillin/Clavulanate and Azithromycin at the End-of-Therapy Visit
(Day 12–14) Among Children With Haemophilus influenzae or Streptococcus pneumoniae Isolated at
Baseline
Clinical Efficacy (%)
Pathogen P Treatment Difference 95% CI
A/C AZI
H. influenzae 91.6 (87/95)* 73.7 (70/95) ⬍0.01 17.9 6.37–29.42
S. pneumoniae 89.8 (79/88) 86.0 (86/100) 0.57 3.8 ⫺6.59 –14.13
Penicillin-resistant 92.0 (23/25) 78.3 (18/23) 0.24 13.7 ⫺10.37–37.84
Azithromycin-resistant 88.2 (15/17) 70.0 (14/20) 0.25 18.2 ⫺12.46 – 48.93
*Numbers in parentheses, n/N.
A/C indicates amoxicillin/clavulanate; AZI, azithromycin.
and 40.6% (13 of 32 children) for children randomized to
TABLE 6. Adverse Events and Withdrawals From the
amoxicillin/clavulanate and azithromycin, respectively.
Study According to Treatment Group
Safety. Adverse events and withdrawals from the study ac-
cording to treatment group are shown in Table 6. A/C AZI
Safety Outcome (N ⫽ 367) (N ⫽ 363) P
DISCUSSION n % n %
Among children with AOM, an antimicrobial agent
Adverse events
with relatively low antibacterial efficacy may result in similar At least 1 adverse event 139 37.9 128 35.3 NS
clinical outcomes, compared with those that result from using Fever 38 10.4 37 10.2 NS
Diarrhea 21 5.7 13 3.6 NS
an antimicrobial agent with high antibacterial efficacy.9,10
Withdrawal from the
However, rates of bacterial eradication are helpful in discrim- study
inating among agents. Accordingly, there is a growing accep- Adverse event 21 5.7 7 1.9 ⬍0.05
tance of the need to take pharmacokinetic/pharmacodynamic Insufficient therapeutic 13 3.5 35 9.6 ⬍0.01
effect
measures into consideration when choosing antimicrobial Protocol violation 35 9.5 21 5.8 NS
therapy for children with AOM. The best predictors of effi- Lost to follow-up 21 5.7 13 3.6 NS
cacy are the time during which an antimicrobial agent main- Other 6 1.6 13 3.6 NS
Overall 96 26.2 89 24.5 NS
tains serum concentrations above the MIC of the pathogen for
-lactam antimicrobial agents and the ratio of the area under A/C indicates amoxicillin/clavulanate; AZI, azithromycin; NS, not significant.
the concentration-time curve to the MIC for macrolide/aza-
lide antimicrobial agents.11–13 For amoxicillin, a serum con-
centration greater than the MIC for 40 –50% of the dosing strains and most H. influenzae strains would not be eradicated
interval is predictive of 80 – 85% bacterial eradication in at a rate higher than that observed without antibiotics.
AOM.11 Amoxicillin/clavulanate (90 and 6.4 mg/kg/d of the Previous studies among children with AOM showed a
amoxicillin and clavulanate components, respectively) was clear relationship between early (day 4 –5) eradication of
developed to sustain amoxicillin/clavulanate concentrations infecting pathogens from the middle ear fluid and positive
in middle ear fluid at levels that would predict eradication of clinical outcomes.14,15 The results of the current study are
PRSP with amoxicillin/clavulanate MICs of ⱕ4 g/mL. consistent with previous reports. There was a strong correla-
Applying the area under the concentration-time curve/MIC tion between early bacteriologic success at the on-therapy
model, the prediction for azithromycin, at the dosage used in visit (day 4 – 6) and clinical success at the end-of-therapy visit
this study, would be that macrolide-resistant S. pneumoniae (day 12–14) in the combined treatment groups for both S.
530 © 2005 Lippincott Williams & WilkinsThe Pediatric Infectious Disease Journal • Volume 24, Number 6, June 2005 Amoxicillin/Clavulanate for AOM
pneumoniae (126 of 137 children, 92.0%) and H. influenzae design; (2) a low rate of bacterial AOM at baseline (55%); (3)
(50 of 57 children, 87.7%), whereas the corresponding values fewer children (⬃65%) ⬍24 months of age, ie, those most
for children without bacterial eradication were considerably likely to experience treatment failure;15 (4) use of day 28 –32
lower (65.4% and 46.7% for S. pneumoniae and H. influen- visit results as the primary outcome; and (5) performance of
zae, respectively). Viewed alternatively, our data indicate repeat tympanocentesis for only 7 of 64 children categorized
that, for AOM caused by either S. pneumoniae or H. influ- as experiencing clinical failure (mostly at one study site).
enzae, failure to eradicate pathogens from the middle ear fluid Despite these limitations and differences in study design,
within 4 – 6 days increased the risk of clinical failure by outcomes for children with AOM caused by H. influenzae
4.3-fold (for S. pneumoniae, 35.6% versus 8.0%; for H. (azithromycin, 67%; amoxicillin/clavulanate, 81%; 95% CI
influenzae, 53.3% versus 12.3%), confirming the experience of the difference, ⫺34.7% to ⫺6.8%) were similar to those
of others that antimicrobial agents with greater bacteriologic reported previously8 and to those in our study.
efficacy result in higher clinical cure rates at the end of We conclude that amoxicillin/clavulanate was effective
treatment. However, it must be noted that many children for in eradicating AOM pathogens and that this led to high
whom cultures remained positive at day 4 – 6 eventually clinical cure rates. Amoxicillin/clavulanate was significantly
experienced cures. more effective than azithromycin bacteriologically and clin-
As in other clinical trials, H. influenzae was a fre- ically. The relatively high bacteriologic and clinical efficacy
quently isolated pathogen.2,8,16 Azithromycin failed to erad- of amoxicillin/clavulanate, together with its excellent safety
icate H. influenzae for 47.0% of children, a finding consistent profile, makes it a preferred antimicrobial choice for the
with both the previously reported low bacteriologic efficacy treatment of children with recurrent or persistent AOM,
of azithromycin against this pathogen and the pharmacoki- particularly those at risk for infection with nonsusceptible
netic/pharmacodynamic model.8,17,18 Peric et al19 recently S. pneumoniae and/or -lactamase-positive H. influenzae or
demonstrated that most H. influenzae strains have an intrinsic M. catarrhalis.
macrolide-efflux mechanism, which is not overcome by dos-
age increases within tolerable limits.20 Therefore, despite the ACKNOWLEDGMENTS
in vitro findings that most H. influenzae strains are suscepti- We acknowledge the contributions of Richard Morales
ble to azithromycin (according to the NCCLS definition of and Tracy Henrick, GlaxoSmithKline Biometrics and Data
having a MIC of ⬍8 g/mL), the results of clinical studies Science Group, and John Breton, GlaxoSmithKline Clinical
with bacteriologic outcomes demonstrate that almost all H. Research, for their contributions to the conduct of the study
influenzae isolates are, in fact, resistant to azithromycin, at and the development of the manuscript. We also thank Jack L.
least with the concentrations and dynamics achieved with the Paradise, MD, and Ellen R. Wald, MD, for their advice on
currently recommended dosages. the study design and for their careful review of the manu-
Children ⬍2 years of age are more likely than older script.
children to have AOM caused by resistant pathogens, such as
PRSP, and accordingly are more likely to experience clinical REFERENCES
failure.14,15,21–23 The overall prevalence of PRSP (26.6% of 1. Jacobs MR. Increasing antibiotic resistance among otitis media patho-
gens and their susceptibility to oral agents based on pharmacodynamic
all S. pneumoniae isolates) in the present study, in which 86% parameters. Pediatr Infect Dis J. 2000;19(suppl):S47–S55.
of children were ⬍2 years of age, is consistent with these 2. Dagan R, Hoberman A, Johnson CE, et al. Bacteriologic and clinical
observations. Despite a relatively high prevalence of PRSP, efficacy of high dose amoxicillin/clavulanate in children with acute otitis
amoxicillin/clavulanate eradicated S. pneumoniae in most media. Pediatr Infect Dis J. 2001;20:829 – 837.
3. Dowell SF, Butler JC, Giebink GS, et al. Acute otitis media: manage-
instances (bacteriologic success rate of 96%). More than ment and surveillance in an era of pneumococcal resistance: a report
one-half of the PRSP isolates were also resistant to macro- from the Drug-Resistant Streptococcus pneumoniae Therapeutic Work-
lides. This level of penicillin-macrolide cross-resistance is ing Group. Pediatr Infect Dis J. 1999;18:1–9.
similar to that reported previously in a surveillance study of 4. American Academy of Pediatrics, Subcommittee on Management of
Acute Otitis Media. Diagnosis and management of acute otitis media.
Western Europe and the United States between 1998 and Pediatrics. 2004;113:1451–1465.
2000.24 In contrast to the high PRSP eradication rates 5. Hoberman A, Marchant CD, Kaplan SL, Feldman S. Treatment of acute
achieved with amoxicillin/clavulanate (92.0%), azithromycin otitis media consensus recommendations. Clin Pediatr. 2002;41:373–
eradicated 54.5% of PRSP isolates (P ⬍ 0.05). 390.
6. National Committee for Clinical Laboratory Standards. Performance
A recent report by Arrieta et al25 compared a larger Standards for Antimicrobial Susceptibility Testing, 10th Informational
dosage of azithromycin (20 mg/kg/d administered once daily Supplement. Document M100-S10. Wayne, PA: National Committee for
for 3 days) with amoxicillin/clavulanate (90/6.4 mg/kg/d in 2 Clinical Laboratory Standards; 2000.
divided doses daily). The authors concluded that azithromy- 7. National Committee for Clinical Laboratory Standards. Performance
Standards for Antimicrobial Susceptibility Testing, 12th Informational
cin resulted in improved clinical outcomes at day 28 –32 and Supplement. Document M100-S11. Wayne, PA: National Committee for
that clinical efficacies for the 2 antimicrobial agents were Clinical Laboratory Standards; 2002.
comparable for microbiologically evaluable children. There 8. Dagan R, Johnson CE, McLinn S, et al. Bacteriologic and clinical
were substantial differences in study design between the efficacy of amoxicillin/clavulanate vs. azithromycin in acute otitis me-
dia. Pediatr Infect Dis J. 2000;19:95–104.
study by Arrieta et al25 and the current study. Methodologic 9. Marchant CD, Carlin SA, Johnson CE, Shurin PA. Measuring the
limitations of the study by Arrieta et al25 included (1) a comparative efficacy of antibacterial agents for acute otitis media: the
sample size too small for a single-tympanocentesis study “Pollyanna phenomenon.” J Pediatr. 1992;120:72–77.
© 2005 Lippincott Williams & Wilkins 531Hoberman et al The Pediatric Infectious Disease Journal • Volume 24, Number 6, June 2005
10. Dagan R, Klugman K, Craig WA, Baquero F. Evidence to support the 18. Howie VM, Ploussard JH. Efficacy of fixed combination antibiotics
rationale that bacterial eradication in respiratory tract infection is an versus separate components in otitis media: effectiveness of erythromy-
important aim of antimicrobial therapy. J Antimicrob Chemother. 2001; cin estolate, triple sulfonomide, ampicillin, erythromycin estolate-triple
47:129 –140. sulfonomide, and placebo in 280 patients with acute otitis media under
11. Craig WA, Andes D. Pharmacokinetics and pharmacodynamics of two and one-half years of age. Clin Pediatr (Phila). 1972;11:205–214.
antibiotics in otitis media. Pediatr Infect Dis J. 1996;15:944 –948. 19. Peric M, Bozdogan B, Jacobs MR, Appelbaum PC. Effects of an efflux
12. Craig WA. Pharmacokinetic/pharmacodynamic parameters: rationale for mechanism and ribosomal mutations on macrolide susceptibility of
antibacterial dosing of mice and men. Clin Infect Dis. 1998;26:1–12. Haemophilus influenzae clinical isolates. Antimicrob Agents Chemother.
13. Craig WA, Kiem S, Andes DR. Free drug 24-hr AUC/MIC is the PK/PD 2003;47:1017–1022.
target that correlates with in vivo efficacy of macrolides, azalides, 20. Jacobs MR. Building in efficacy: developing solutions to combat
ketolides and clindamycin. In: 42nd Interscience Conference on Anti- drug-resistant Streptococcus pneumoniae. Clin Microbiol Infect.
microbial Agents and Chemotherapy, San Diego, CA, 2002. Washing- 2003;9(suppl 1):27.
21. Giebink G, Batalden P, Russ J, Chap T. Cefaclor versus amoxicillin in
ton, DC: American Society for Microbiology; 2002. Abstract A-1264.
treatment of acute otitis media. Am J Dis Child. 1984;138:287–292.
14. Dagan R, Leibovitz E, Greenberg D, Yagupsky P, Fliss DM, Leiberman
22. Harsten G, Prellen K, Heldrup J, Kalm O, Kornfalt R. Treatment failure
A. Early eradication of pathogens from middle ear fluid during antibiotic
in acute otitis media. Acta Otolaryngol. 1989;108:253–258.
treatment of acute otitis media is associated with improved clinical 23. Hoberman A, Paradise JL, Burch DJ, et al. Equivalent efficacy and
outcome. Pediatr Infect Dis J. 1998;17:776 –782. reduced occurrence of diarrhea from a new formulation of amoxicillin/
15. Carlin SA, Marchant CD, Shurin PA, Johnson CE, Super DA, Rehmus clavulanate potassium (Augmentin) for treatment of acute otitis media in
JM. Host factors and early therapeutic response in acute otitis media. children. Pediatr Infect Dis J. 1997;16:463– 470.
J Pediatr. 1991;118:178 –183. 24. Jacobs MR, Felmingham D, Appelbaum PC, Grüneberg RN. The Alex-
16. Barnett ED, Klein JO. The problem of resistant bacteria for the man- ander Project 1998 –2000: susceptibility of pathogens isolated from
agement of acute otitis media. Pediatr Clin North Am. 1995;42:509 – community-acquired respiratory tract infection to commonly used anti-
517. microbial agents. J Antimicrob Chemother. 2003;52:229 –246.
17. Dagan R, Leibovitz E, Fliss DM, et al. Bacteriologic efficacies of oral 25. Arrieta A, Arguedas A, Fernandez P, et al. High-dose azithromycin
azithromycin and oral cefaclor in treatment of acute otitis media in versus high-dose amoxicillin-clavulanate for treatment of children with
infants and young children. Antimicrob Agents Chemother. 2000;44: recurrent or persistent acute otitis media. Antimicrob Agents Chemother.
43–50. 2003;47:3179 –3186.
532 © 2005 Lippincott Williams & WilkinsYou can also read
