PDF Hosted at The Radboud Repository of The Radboud University Nijmegen
PDF Hosted at The Radboud Repository of The Radboud University Nijmegen
PDF Hosted at The Radboud Repository of The Radboud University Nijmegen
Nijmegen
Please be advised that this information was generated on 2016-02-09 and may be subject to
change.
Antibiotics for acute otitis media in children (Review)
This is a reprint of a Cochrane review, prepared and maintained by The Cochrane Collaboration and published in The Cochrane Library
2013, Issue 1
http://www.thecochranelibrary.com
Roderick P Venekamp1 , Sharon Sanders2 , Paul P Glasziou3 , Chris B Del Mar3 , Maroeska M Rovers4
1 Department of Otorhinolaryngology & Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht,
Utrecht, Netherlands. 2 School of Medicine, University of Queensland, Brisbane, Australia. 3 Centre for Research in Evidence Based
Practice, Bond University, Gold Coast, Australia. 4 Department of Operating Rooms, Radboud University Nijmegen Medical Centre,
Nijmegen, Netherlands
Contact address: Roderick P Venekamp, Department of Otorhinolaryngology & Julius Center for Health Sciences and Primary Care,
University Medical Center Utrecht, PO Box 85500, Utrecht, 3508 GA, Netherlands. R.P.Venekamp@umcutrecht.nl.
Citation: Venekamp RP, Sanders S, Glasziou PP, Del Mar CB, Rovers MM. Antibiotics for acute otitis media in children. Cochrane
Database of Systematic Reviews 2013, Issue 1. Art. No.: CD000219. DOI: 10.1002/14651858.CD000219.pub3.
Copyright 2013 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
ABSTRACT
Background
Acute otitis media (AOM) is one of the most common diseases in early infancy and childhood. Antibiotic use for AOM varies from
56% in the Netherlands to 95% in the USA, Canada and Australia.
Objectives
To assess the effects of antibiotics for children with AOM.
Search methods
We searched CENTRAL (2012, Issue 10), MEDLINE (1966 to October week 4, 2012), OLDMEDLINE (1958 to 1965), EMBASE
(January 1990 to November 2012), Current Contents (1966 to November 2012), CINAHL (2008 to November 2012) and LILACS
(2008 to November 2012).
Selection criteria
Randomised controlled trials (RCTs) comparing 1) antimicrobial drugs with placebo and 2) immediate antibiotic treatment with
expectant observation (including delayed antibiotic prescribing) in children with AOM.
Data collection and analysis
Two review authors independently assessed trial quality and extracted data.
Main results
For the review of antibiotics against placebo, 12 RCTs (3317 children and 3854 AOM episodes) from high-income countries were
eligible. However, one trial did not report patient-relevant outcomes, leaving 11 trials with generally low risk of bias. Pain was not
reduced by antibiotics at 24 hours (risk ratio (RR) 0.89; 95% confidence interval (CI) 0.78 to 1.01) but almost a third fewer had residual
pain at two to three days (RR 0.70; 95% CI 0.57 to 0.86; number needed to treat for an additional beneficial outcome (NNTB) 20)
and fewer had pain at four to seven days (RR 0.79; 95% CI 0.66 to 0.95; NNTB 20). When compared with placebo, antibiotics did
not alter the number of abnormal tympanometry findings at either four to six weeks (RR 0.92; 95% CI 0.83 to 1.01) or at three months
Antibiotics for acute otitis media in children (Review) 1
Copyright 2013 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
(RR 0.97; 95% CI 0.76 to 1.24), or the number of AOM recurrences (RR 0.93; 95% CI 0.78 to 1.10). However, antibiotic treatment
did lead to a statistically significant reduction of tympanic membrane perforations (RR 0.37; 95% CI 0.18 to 0.76; NNTB 33) and
halved contralateral AOM episodes (RR 0.49; 95% CI 0.25 to 0.95; NNTB 11) as compared with placebo. Severe complications were
rare and did not differ between children treated with antibiotics and those treated with placebo. Adverse events (such as vomiting,
diarrhoea or rash) occurred more often in children taking antibiotics (RR 1.34; 95% CI 1.16 to 1.55; number needed to treat for an
additional harmful outcome (NNTH) 14). Funnel plots do not suggest publication bias. Individual patient data meta-analysis of a
subset of included trials found antibiotics to be most beneficial in children aged less than two with bilateral AOM, or with both AOM
and otorrhoea.
For the review of immediate antibiotics against expectant observation, five trials (1149 children) were eligible. Four trials (1007 children)
reported outcome data that could be used for this review. From these trials, data from 959 children could be extracted for the meta-
analysis on pain at days three to seven. No difference in pain was detectable at three to seven days (RR 0.75; 95% CI 0.50 to 1.12).
No serious complications occurred in either the antibiotic group or the expectant observation group. Additionally, no difference in
tympanic membrane perforations and AOM recurrence was observed. Immediate antibiotic prescribing was associated with a substantial
increased risk of vomiting, diarrhoea or rash as compared with expectant observation (RR 1.71; 95% CI 1.24 to 2.36).
Authors conclusions
Antibiotic treatment led to a statistically significant reduction of children with AOM experiencing pain at two to seven days compared
with placebo but since most children (82%) settle spontaneously, about 20 children must be treated to prevent one suffering from
ear pain at two to seven days. Additionally, antibiotic treatment led to a statistically significant reduction of tympanic membrane
perforations (NNTB 33) and contralateral AOM episodes (NNTB 11). These benefits must be weighed against the possible harms: for
every 14 children treated with antibiotics, one child experienced an adverse event (such as vomiting, diarrhoea or rash) that would not
have occurred if antibiotics had been withheld. Antibiotics appear to be most useful in children under two years of age with bilateral
AOM, or with both AOM and otorrhoea. For most other children with mild disease, an expectant observational approach seems
justified. We have no trials in populations with higher risks of complications.
Outcomes Illustrative comparative risks* Relative effect Absolute effect NNTB/NNTH No. of participants Quality of the evi-
(95% CI) (95% CI) (95% CI) (95% CI) (studies) dence
(GRADE)
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval (CI)) is based on the
assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI)
CI: confidence interval; RR: risk ratio; RD: risk difference; NNTB: number needed to treat to benefit; NNTH: number needed to treat to harm
Blinding
All included trials in the review of antibiotics against placebo stated Effects of interventions
that they were double-blinded. However, blinding was judged to See: Summary of findings for the main comparison
be adequate in eight of the 11 included trials reporting patient-
relevant outcomes. All five trials comparing immediate antibi-
otics with expectant observation were open label trials. As a con- Pain
sequence, reporting of the childs symptoms by parents was not The combined results of the trials revealed that by 24 hours from
blinded in these trials. However, investigators were blinded in two the start of treatment, 60% of the children had recovered whether
of the five trials (McCormick 2005; Spiro 2006). or not they had placebo or antibiotics. At days two to seven, 82%
of children had spontaneously recovered (pooled control groups).
Antibiotics achieved a 30% relative reduction in the risk of pain
Incomplete outcome data at days two to three (95% CI 14% to 43%) and 21% in the risk
The loss to follow-up was below 5% in six of the 11 trials com- of pain at days four to seven (95% CI 5% to 34%) (Analysis
paring antibiotics with placebo that reported patient-relevant out- 1.1). This means 5% fewer children had pain after two to three
comes. Loss to follow-up was high in three of the 11 trials with days (95% CI -7% to -2%) and four to seven days (95% CI -9%
a total loss to follow-up of 15% (vanBuchem 1981a; vanBuchem to -1%), respectively. Therefore, 20 (95% CI 14 to 40) children
1981b), 7% (Kaleida 1991) and 12% (Damoiseaux 2000), respec- needed to be treated to prevent one child experiencing pain at days
tively. However, one of these trials included all randomised patients two to seven. Plots of the event rate (pain) in the treatment and
in the primary analysis at day four (Damoiseaux 2000). In two of control groups for each study on the various time points (two to
the 11 trials (Halsted 1968; Mygind 1981) the total number of three days and four to seven days) are reported in Figure 3 and
loss to follow-up / exclusions are described but it was unclear from Figure 4. The funnel plot for pain at the various time points did
which treatment group children were excluded. For the review of not reveal asymmetry (Figure 5).
AUTHORS CONCLUSIONS
Implications for practice
Antibiotics produce a small reduction in the number of children ACKNOWLEDGEMENTS
with pain at two to three and four to seven days from assess-
We would like to thank Professor Charles Bridges-Webb for stim-
ment and reduce the number of tympanic membrane perforations.
ulating initial discussions and for constructive advice on the pro-
However, in high-income countries, most cases of AOM spon-
tocol for this review and Professor Steve Berman for helpful com-
taneously remit without complications and the NNTB is 20 for
ments on the draft review. We would like to thank Bruce Arroll
pain at two to three and four to seven days and 33 for tympanic
and Tom Fahey for peer refereeing the 2005 updated review and
membrane perforation. These benefits must be weighed against
Dilip Raghavan, Brian Westerberg, Mark Jones and Peter Morris
the possible harms: for every 14 children treated with antibiotics
for peer refereeing the 2009 updated review. We also thank Dilip
one child experienced an adverse event (such as vomiting, diar-
Raghavan, Brian Westerberg, Teresa Neeman and Peter Morris for
rhoea or rash) that would not have occurred if antibiotics were
commenting on this 2012 updated review.
withheld. Therefore management should emphasise advice about
adequate analgesia and the limited role for antibiotics. Antibiotics Thank you to David McCormick and his colleagues for allowing
are most useful in children under two years of age with bilateral us to access raw study data from the trial comparing immediate
AOM, or with both AOM and otorrhoea. For most other children antibiotics and expectant observation (McCormick 2005).
with mild disease, an expectant observational approach seems jus-
Thank you to Chris Cates for noticing and advising us of an error
tified. Cates has developed an appropriate handout and tested this
in the confidence intervals presented in the review.
together with an optional antibiotic prescription (Cates 1999).
The handout is available at www.nntonline.net/ebm/mainpages/ We gratefully thank Sarah Thorning for her support with the
AOM.asp (accessed 22 November 2012). search strategy and searches.
REFERENCES
References to studies included in this review Halsted 1968 {published data only}
Halsted C, Lepow ML, Balassanian N, Emmerich
Appelman 1991 {published data only}
J, Wolinsky E. Otitis media: clinical observation,
Appelman CL, Claessen JQ, Touw Otten FW, Hordijk
microbiology and evaluation of therapy. American Journal
GJ, de Melker RA. Co-amoxiclav in recurrent acute otitis
of Diseases of Children 1968;115(5):54251.
media: placebo controlled study. BMJ 1991;303:14502.
Appelman CL, Claessen JQ, Touw Otten FW, Hordijk Hoberman 2011 {published data only}
GJ, de Melker RA. Severity of inflammation of tympanic Hoberman A, Paradise JL, Rockette HE, Shaikh N, Wald
membrane as predictor of clinical course of recurrent acute ER, Kearney DH, et al.Treatment of acute otitis media
otitis media. BMJ 1993;306:895. in children under 2 years of age. New England Journal of
Claessen JQ, Appelman CL, Touw Otten FW, de Melker Medicine 2011;364(2):10515.
RA, Hordijk GJ. Persistence of middle ear dysfunction after Howie 1972 {published data only}
recurrent acute otitis media. Clinical Otolaryngology 1994; Howie VM, Ploussard JH. Efficacy of fixed combination
19:3540. antibiotics versus separate components in otitis media:
Burke 1991 {published data only} effectiveness of erythromycin estolate, triple sulfonamide,
Burke P, Bain J, Robinson D, Dunleavey J. Acute red ear ampicillin, erythromycin estolate-triple sulfonamide, and
in children: controlled trial of non-antibiotic treatment in placebo in 280 patients with acute otitis media under two
general practice. BMJ 1991;303:55862. and one-half years of age. Clinical Pediatrics 1972;11(4):
Damoiseaux 2000 {published data only} 20514.
Damoiseaux RAMJ, van Balen FAM, Hoes AW, Verheij Kaleida 1991 {published data only}
TJM, de Melker RA. Primary care based randomised, Kaleida PH, Casselhrant ML, Rockette HE, Paradise
double blind trial of amoxicillin versus placebo for acute JL, Bluestone CD, Blatter MM, et al.Amoxicillin or
otitis media in children aged under 2 years. BMJ 2000;320: myringotomy or both for acute otitis media: results of a
3504. randomized clinical trial. Paediatrics 1991;87(4):46674.
Antibiotics for acute otitis media in children (Review) 18
Copyright 2013 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
Laxdal 1970 {published data only} vanBuchem 1981b {published data only}
Laxdal OE, Merida J, Jones RHT. Treatment of acute otitis van Buchem FL, Dunk JHM, vant Hof MA. Therapy of
media: a controlled study of 142 children. Canadian acute otitis media: myringotomy, antibiotics or neither? A
Medical Association Journal 1970;102(3):2638. double-blind study in children. Lancet 1981;2:8837.
Le Saux 2005 {published data only}
References to studies excluded from this review
Le Saux N, Gaboury I, Baird M, Klassen TP, MacCormick
J, Blanchard C, et al.A randomized, double-blind, placebo- Arguedas 2011 {published data only}
controlled noninferiority trial of amoxicillin for clinically Arguedas A, Soley C, Kamicker BJ, Jorgensen DM. Single-
diagnosed acute otitis media in children 6 months to 5 years dose extended-release azithromycin versus a 10-day regimen
of age. Canadian Medical Association Journal 2005;172(3): of amoxicillin/clavulanate for the treatment of children with
33541. acute otitis media. International Journal of Infectious Diseases
Little 2001 {published data only} 2011;15(4):e2408.
Little P, Gould C, Williamson I, Moore M, Warner G, Casey 2012 {published data only}
Dunleavey J. Pragmatic randomised controlled trial of two Casey JR, Block SL, Hedrick J, Almudevar A, Pichichero
prescribing strategies for childhood acute otitis media. BMJ ME. Comparison of amoxicillin/clavulanic acid high dose
2001;322:33642. with cefdinir in the treatment of acute otitis media. Drugs
McCormick 2005 {published data only} 2012;72:19917.
McCormick DP, Chonmaitree T, Pittman C, Saeed K, Chaput 1982 {published data only}
Friedman NR, Uchida T, et al.Nonsevere acute otitis media: Chaput de Saintonge DM, Levine DF, Temple Savage IT,
a clinical trial comparing outcomes of watchful waiting Burgess GW, Sharp J, Mayhew SR, et al.Trial of three-day
versus immediate antibiotic treatment. Pediatrics 2005;115 and ten-day courses of amoxycillin in otitis media. BMJ
(6):145565. Clinical Research Edition 1982;284(6322):107881.
Mygind 1981 {published data only} Engelhard 1989 {published data only}
Mygind N, Meistrup-Larsen K-I, Thomsen J, Thomsen Engelhard D, Strauss N, Jorczak-Sarni L, Cohen D, Sacjs
VF, Josefsson K, Sorensen H. Penicillin in acute otitis TG, Shapiro M. Randomised study of myringotomy,
media: a double-blind placebo-controlled trial. Clinical amoxycillin/clavulanate, or both for acute otitis media in
Otolaryngology 1981;6:513. infants. Lancet 1989;2(8655):1413.
Thomsen J, Meistrup-Larsen KI, Sorensen H, Larsen PK, Liu 2011 {published data only}
Mygind N. Penicillin and acute otitis: short and long Liu P, Fang AF, LaBadie RR, Crownover PH, Arguedas AG.
term results. Annals of Otology, Rhinology and Laryngology, Comparison of azithromycin pharmacokinetics following
Supplement 1980;89:2714. single oral doses of extended-release and immediate-
Neumark 2007 {published data only} release formulations in children with acute otitis media.
Neumark T, Molstad S, Rosen C, Persson L, Torngren A, Antimicrobial Agents and Chemotherapy 2011;55(11):
Brudin L, et al.Evaluation of phenoxymethylpenicillin 50226.
treatment of acute otitis media in children aged 2-16. Ostfeld 1987 {published data only}
Scandinavian Journal of Primary Health Care 2007;25: Ostfeld E, Segal J, Kaufstein M, Gelernter I. Management
16671. of acute otitis media without primary administration of
Spiro 2006 {published data only} systemic antimicrobial agents. Recent advances in otitis
Spiro DM, Tay K, Arnold DH, Dziura J, Baker M, Shapiro media. Proceedings of the Fourth International Symposium.
ED. Wait-and-see prescription for the treatment of acute Toronto: BC Decker, 1987:2359.
otitis media. JAMA 2006;296(10):123541. Rudberg 1954 {published data only}
Thtinen 2011 {published data only} Rudberg RD. Acute otitis media: comparative therapeutic
Thtinen PA, Laine MK, Huovinen P, Jalava J, Ruuskanen results of sulphonamide and penicillin administered in
O, Ruohola A. A placebo-controlled trial of antimicrobial various forms. Acta Otolaryngology 1954;113(Suppl):179.
treatment for acute otitis media. New England Journal of Ruohola 2003 {published data only}
Medicine 2011;364(2):11626. Ruohola A, Heikkinen T, Meurman O, Puhakka T,
Thalin 1985 {published data only} Lindblad N, Ruuskanen O. Antibiotic treatment of acute
Thalin A, Densert O, Larsson A, Lyden E, Ripa T. Is otorrhea through tympanostomy tube: randomized double-
penicillin necessary in the treatment of acute otitis media? blind placebo-controlled study with daily follow-up.
. Proceedings of the International Conference on Acute Pediatrics 2003;111(5):10617.
and Secretory Otitis Media 1985, Jerusalem. Amsterdam: Sarrell 2003 {published data only}
Kugler Publications, 1985:4416. Sarrell EM, Cohen HA, Kahan E. Naturopathic treatment
vanBuchem 1981a {published data only} for ear pain in children. Pediatrics 2003;111(5):5749.
van Buchem FL, Dunk JHM, vant Hof MA. Therapy of Thtinen 2012 {published data only}
acute otitis media: myringotomy, antibiotics or neither? A Thtinen PA, Laine MK, Ruuskanen O, Ruohola A.
double-blind study in children. Lancet 1981;2:8837. Delayed versus immediate antimicrobial treatment for acute
Antibiotics for acute otitis media in children (Review) 19
Copyright 2013 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
otitis media. Pediatric Infectious Disease Journal 2012:[Epub Froom 2001
ahead of print]. Froom J, Culpepper L, Green LA, de Melker RA, Grob P,
vanBuchem 1985 {published data only} Heeren T, et al.A cross-national study of acute otitis media:
van Buchem F, Peeters M, Van t Hof M. Acute otitis media: risk factors, severity, and treatment at initial visit. Report
a new treatment strategy. BMJ 1985;290:10337. from the International Primary Care Network (IPCN) and
the Ambulatory Sentinel Practice Network (ASPN). Journal
Additional references of the American Board of Family Practice 2001;14:40617.
Gates 2002
AAP 2004
Gates GA, Klein JO, Lim DJ, Mogi G, Ogra PL, Pararella
American Academy of Pediatrics, American Academy of
MM, et al.Recent advances in otitis media, 1: definitions,
Family Physicians. Clinical practice guidelines: diagnosis
terminology, and classification of otitis media. Annals of
and management of acute otitis media. Pediatrics 2004;113:
Otology, Rhinology and Laryngology 2002;111:818.
145165.
Akkerman 2005 Higgins 2011
Akkerman AE, Kuyvenhoven MM, van der Wouden JC, Higgins JPT, Green S (editors). Cochrane Handbook
Verhij TJM. Analysis of under-and overprescribing of for Systematic Reviews of Interventions version 5.1.0
antibiotics in acute otitis media in general practice. Journal [updated March 2011]. The Cochrane Collaboration,
of Antimicrobial Chemotherapy 2005;56:56974. 2011. Available from www.cochrane-handbook.org.
Arnold 2005 Klein 1989
Arnold SR, Straus SE. Interventions to improve antibiotic Klein JO. Epidemiology of otitis media. Pediatic Infectious
prescribing in ambulatory care. Cochrane Database Disease Journal 1989;8(Suppl 1):89.
of Systematic Reviews 2005, Issue 4. [DOI: 10.1002/
Koopman 2008
14651858.CD003539]
Koopman L, Hoes AW, Glasziou PP, Appelman CL, Burke
Arola 1990 P, McCormick DP, et al.Antibiotic therapy to prevent
Arola M, Ruuskanen O, Ziegler T, Mertsola J, Nnt- the development of asymptomatic middle ear effusion in
Salonen K, Putto-Laurila A, et al.Clinical role of respiratory children with acute otitis media. Archives of Otolaryngology -
virus infection in acute otitis media. Pediatrics 1990;86: Head and Neck Surgery 2008;134(2):12832.
84855.
Kozyrskyj 2010
Barkai 2009
Kozyrskyj A, Klassen TP, Moffatt M, Harvey K. Short-
Barkai G, Leibovitz E, Givon-Lavi N, Dagan R. Potential
course antibiotics for acute otitis media. Cochrane Database
contribution by nontypable Haemophilus influenzae in
of Systematic Reviews 2010, Issue 9. [DOI: 10.1002/
protracted and recurrent acute otitis media. Pediatric
14651858.CD001095]
Infectious Disease Journal 2009;28:46671.
Berman 1995 Lefebvre 2011
Berman S. Otitis media in developing countries. Pediatrics Lefebvre C, Manheimer E, Glanville J. Chapter 6: Searching
1995;96:12631. for studies. In: Higgins JPT, Green S (editors). Cochrane
Handbook for Systematic Reviews of Interventions Version
Cates 1999
5.1.0 [updated March 2011]. The Cochrane Collaboration,
Cates C. An evidence based approach to reducing antibiotic
2011. Available from www.cochrane-handbook.org.
use in children with acute otitis media: controlled before
and after the study. BMJ 1999;318:7156. [: The handout Little 2006
is available at http://www.cates.cwc.net/] Little P, Moore M, Warner G, Dunleavy J, Williamson
Chao 2008 I. Longer term outcomes from a randomised trial of
Chao JH, Kunkov S, Reyes LB, Lichten S, Crain EF. prescribing strategies in otitis media. British Journal of
Comparison of two approaches to observation therapy for General Practice 2006;56:17682.
acute otitis media in the emergency department. Pediatrics McCormick 2007
2008;121:31352e6. McCormick DP, Chandler SM, Chonmaitree T. Laterality
Chonmaitree 1992 of acute otitis media: different clinical and microbiologic
Chonmaitree T, Owen MJ, Patel JA, Hedgpeth D, Horlick characteristics. Pediatric Infectious Disease Journal 2007;26:
D, Howie VM. Effect of viral respiratory tract infection on 5838.
outcome of acute otitis media. Journal of Pediatrics 1992; Rosenfeld 1994
120:85662. Rosenfeld RM, Vertrees JE, Carr J, Cipolle RJ, Uden DL,
Dickersin 1994 Giebink GS, et al.Clinical efficacy of antimicrobial drugs
Dickersin K, Scherer R, Lefebvre C. Identifying relevant for acute otitis media: meta-analysis of 5400 children from
studies for systematic reviews. BMJ 1994;309(6964): thirty-three randomized trials. Journal of Pediatrics 1994;
128691. 124:35567.
Appelman 1991
Outcomes Primary outcome - treatment failure (i.e. presence of otalgia or fever > 38 C or both
at 3 days)
Assessment by (blinded) general practitioner at 3 days on the presence or absence of fever
(> 38 C) and otalgia and 14 days on the presence or otorrhoea
Assessment by otorhinolaryngologist at 1 month of otoscopy, tympanometry and in
children > 3 years of age an audiogram
Notes
Risk of bias
Blinding of participants and personnel Unclear risk Identical taste and appearance of co-amox-
(performance bias) iclav and placebo not described
Incomplete outcome data (attrition bias) Low risk Loss to follow-up - treatment: N = 3 (4%)
and placebo: N = 2 (4%) due to loss of their
registration forms
Burke 1991
Interventions Tx - amoxicillin 250 mg 3 times daily for 7 days; N = 114 (N = 114 included in analysis
for short-term outcome)
C - matching placebo 3 times daily for 7 days; N = 118 (N = 118 included in analysis
for short-term outcome)
Use of additional medication - analgesics (paracetamol 120 mg/5 mL) for pain as
needed
Outcomes Main outcomes were divided into short-term, middle-term and long-term:
Short-term - (a) duration of symptoms; (b) use of analgesics (assessed by weighing
bottles); (c) clinical signs at 1 week; (d) incidence of complications; (e) treatment failure
(i.e. second-line antibiotics were required)
Middle-term - (a) tympanometry findings at 1 and 3 months
Long-term - (b) number of AOM episodes in 12 months; (b) number of specialist
referrals
Home visits by researcher at day 1, days 4 to 6 and general practitioner visit at day 7
Symptom diary kept by parents for 21 days
Notes It is not clear whether the discharging ears in Table 1 should be included as perforations,
we now included the number of perforations as summarised in Table 2 in our analysis
Risk of bias
Allocation concealment (selection bias) Low risk Randomisation was carried out indepen-
dently of the investigators; randomisation
code was kept sealed and was unknown to
any of the participants in the study
Blinding of participants and personnel Low risk Each bottle was identified only by number
(performance bias)
Incomplete outcome data (attrition bias) Unclear risk Loss to follow-up - not described; all ran-
domised patients included in short-out-
come analysis
Damoiseaux 2000
Interventions Tx - amoxicillin suspension 40 mg/kg/day 3 times daily for 10 days; N = 117 (N = 107
included in analysis for short-term outcome)
C - matching placebo suspension for 10 days; N = 123 (N = 105 included in analysis
for short-term outcome)
Use of additional medication - all children received decongestive nose drops for 7
days; analgesics (paracetamol, children < 1 year: 120 mg suppository, > 1 year: 240 mg
suppository) was allowed
Outcomes Primary outcome - persistent symptoms at day 4: assessed by the doctor and defined as
persistent earache, fever > 38 C, crying, or being irritable. Additionally, prescription of
another antibiotic because of clinical deterioration before the first follow-up visit was to
be considered a persistent symptom
Secondary outcomes - (a) clinical treatment failure at day 11 (i.e. persistent fever,
earache, crying, being irritable, or no improvement of tympanic membrane (including
perforation); (b) duration of fever, pain or crying; (c) mean number of doses analgesics
given; (d) adverse effects mentioned in diaries; (e) percentage of children with middle-
ear effusion at 6 weeks (i.e. combined otoscopy and tympanometry)
Follow-up visits at the GPs clinic were scheduled at day 4 and 11; home visit at 6
weeks by the researcher collecting data of symptoms, referrals and both otoscopy and
tympanometry was performed
Parents were instructed to keep a symptom diary for 10 days
Notes
Risk of bias
Allocation concealment (selection bias) Low risk Randomisation was carried out indepen-
dently of the investigators; randomisation
code was kept in pharmacy of the Univer-
sity Medical Centre Utrecht
Blinding of participants and personnel Low risk Placebo suspension with same taste and ap-
(performance bias) pearance of amoxicillin
Incomplete outcome data (attrition bias) Unclear risk Loss to follow-up/exclusion from analysis
(received other antibiotics or had grommets
inserted) - treatment: N = 10 (9%) and
Halsted 1968
Methods Randomised - yes, pre-determined code which was unknown to physician; method of
random sequence generation unclear
Concealment of allocation - adequate
Double-blind - yes
Intention-to-treat (ITT) - unclear
Loss to follow-up - described, unclear from which treatment group patients were ex-
cluded
Design - parallel
Outcomes Primary outcome - early improvement defined as defervescence and decrease of symp-
toms at 24 to 72 hours
Secondary outcomes - (a) late improvement defined as resolution of symptoms and
normal tympanic membrane at 14 to 18 days, (b) bacteriological cultures
Notes
Risk of bias
Random sequence generation (selection Unclear risk Pre-determined code which was unknown
bias) to physician; method of random sequence
generation unclear
Blinding of participants and personnel Unclear risk Identical taste and appearance of antibi-
(performance bias) otics and placebo not described
Incomplete outcome data (attrition bias) Unclear risk Reasons described, unclear from which
treatment group patients were excluded
Hoberman 2011
to 5)
Use of additional medication - acetaminophen (paracetamol) as needed for symptom
relief
At each visit children were categorised as having met the criteria for either clinical success
or clinical failure
Children who met the criteria for clinical failure were treated with a standardised 10-
day regimen of orally administered amoxicillin (90 mg/kg daily) and cefixime (8 mg/kg
daily)
Outcomes Primary outcomes - (a) time to resolution of symptoms (i.e. time to the first recording
of an AOM-SOS score of 0 or 1 and the time to the second of 2 successive recordings
of that score; (b) symptom burden over time (i.e. mean AOM-SOS score over time each
day for the first 7 days of follow-up and groups weighted mean scores for that period)
Secondary outcomes - (a) clinical failure at day 4 to 5; (b) clinical failure at day 10
to 12; (c) use of acetaminophen (paracetamol); (d) occurrence of adverse events; (e)
nasopharyngeal colonisation rates; (f ) use of healthcare resources (g) relapses
Clinical failure was defined at or before the day 4 to 5 visit as either a lack of substantial
improvement in symptoms, a worsening of signs on otoscopic examination, or both and
at the day 10 to 12 visit as the failure to achieve complete or nearly complete resolutions of
symptoms and otoscopic signs, without regard to the persistence of resolution of middle-
ear effusion. Once a child had met the criteria for clinical failure, he or she remained in
that category for the analysis
Daily symptoms were assessed with the use of a structured interview of 1 of the childs
parents until the first follow-up visit; visits were scheduled at day 4 or 5, day 10 to 12
(end of treatment) and at day 21 to 25
Patients were asked to complete a diary twice a day for 3 days and once a day thereafter
Notes This study did not report pain data that could be used for the review comparing antibi-
otics with placebo
Risk of bias
Random sequence generation (selection Low risk Stratified block randomisation with com-
bias) puter-generated randomisation lists
Allocation concealment (selection bias) Low risk A pharmacist (independent to trial team)
provided masked study medication bottles
with amoxicillin/clavulanate or placebo
Blinding of participants and personnel Low risk Placebo with same taste and appearance of
(performance bias) amoxicillin-clavulanate
Incomplete outcome data (attrition bias) Low risk Children not assessed at day 4 to 5 - treat-
ment: N = 5 (3%) and placebo: N = 5 (3%)
. All randomised patients included in anal-
ysis
Howie 1972
Interventions Tx 1 - erythromycin estolate 125 mg/5 mL - triple sulphonamide suspension 0.5 g/5
mL; N = 80
Tx 2 - ampicillin 250 mg/5 mL; N = 36
Tx 3 - triple sulphonamide suspension 0.5 g/5 mL; N = 23
Tx 4 - erythromycin estolate 125 mg/5 mL; N = 25
C 1 - placebo - equal parts acetaminophen (paracetamol) and chlorpheniramine maleate
syrup; N = 33
C 2 - placebo - 4 parts Kaopectate and 1 part acetaminophen (paracetamol, Tylenol)
plus food colouring; N = 83
Use of additional medication - all children received decongestive nose drops for 7
days; analgesics (paracetamol, children < 1 year: 120 mg suppository, > 1 year: 240 mg
suppository) was allowed
Outcomes Primary outcomes - (a) presence or absence of exudate while on medication; (b) bac-
teriological findings of the exudate when present; no patient-relevant outcomes were
described
At baseline and before treatment was started, the middle-ear exudate was aspirated. The
decision whether to collect exudate on the first repeat visit was made with no knowledge
of the drug regimen to which the patient had been assigned
Notes
Risk of bias
Random sequence generation (selection Unclear risk Method of randomisation not described
bias)
Allocation concealment (selection bias) Low risk Randomisation was performed by a collab-
orating pharmacist
Blinding of participants and personnel Unclear risk Identical taste and appearance of co-amox-
(performance bias) iclav and placebo not described
Incomplete outcome data (attrition bias) Unclear risk Loss to follow-up - not described
Kaleida 1991
Participants N - 536 children (representing 1049 non-severe acute otitis media (AOM) episodes; 980
non-severe AOM episodes included for primary analysis)
Age - between 7 months and 12 years
Setting - secondary care: childrens hospital and a private paediatric practice in Pittsburgh
(USA)
Inclusion criteria - AOM based on presence of middle-ear effusion, as determined
otoscopically, in association with specified symptoms of acute middle-ear infection (fever,
otalgia or irritability), or signs of acute infection (erythema or white opacification, or
both, accompanied by fullness or bulging and impaired mobility), or both
Exclusion criteria - children who recently received antibiotics, who had potential com-
plicating or confounding conditions (e.g. eardrum perforation, asthma, or chronic si-
nusitis)
Baseline characteristics - balanced
Interventions Children were enrolled for a 1-year period. At entry each child was assigned randomly
to a treatment regimen that specified consistent treatments for episodes of non-severe
and severe AOM based on severity of otalgia and the presence of fever (> 39 C orally
or > 39.5 C rectally within the 24-hour period before presentation)
Non-severe AOM episodes were treated with:
Tx - amoxicillin 40 mg/kg/day 3 times daily for 14 days; N = 522 (N = 488 included in
primary analysis)
C - placebo for 14 days; N = 527 (N = 492 included in primary analysis)
Severe AOM episodes in children aged < 2 years were treated with:
Tx 1 - amoxicillin 40 mg/kg/day 3 times daily for 14 days
Tx 2 - amoxicillin 40 mg/kg/day 3 times daily for 14 days and myringotomy
Outcomes Primary outcome - initial treatment failure: in non-severe episodes this was the case
when either otalgia, fever, or both was present more than 24 hours after treatment was
initiated and when 48 hours or more after initial treatment was initiated the childs
temperature reached 38 C orally or 38.5 C rectally or an otalgia score of 6 was
present
Secondary outcomes - (a) recurrent AOM defined as the development of AOM 15 days
or more after the initiation of treatment for a preceding episode (b) new episodes of otitis
media with effusion defined by otoscopy and tympanometry findings
After initial visits, children were followed up by telephone to identify those with persistent
symptoms and children younger than 2 years of age were re-examined within 48 to 72
hours
Follow-up visits were scheduled routinely after 2 and 6 weeks after initial treatment and
monthly thereafter
Notes We included only the non-severe AOM episodes in this review (N = 1049 of which 980
were included for primary analysis); children experiencing non-severe AOM episodes
were randomly allocated to either antibiotics or placebo
Risk of bias
Random sequence generation (selection Unclear risk Method of randomisation not described
bias)
Blinding of participants and personnel Unclear risk Identical taste and appearance of amoxi-
(performance bias) cillin and placebo not described
Interventions Tx 1 - penicillin G 250 mg/m2 /day 4 times daily (approximately 33 mg/kg/day) for at
least 7 days; N = 45
Tx 2 - ampicillin 250 mg/m2 /day 4 times daily (approximately 33 mg/kg/day) for at
least 7 days; N = 49
C - symptomatic therapy (Auralgan ear drops, acetylsalicylic acid, decongestive nose
drops); N = 48
Use of additional medication - children in treatment groups also received symptomatic
therapy as required
Outcomes Primary outcomes - (a) treatment failure (i.e. either deterioration or no improvement
observed at day 7) (b) relapses
Results were evaluated at 7 days, except in cases where the ear inflammation was severe
and the child appeared sufficiently ill (toxic) to warrant further examination 24 to 48
hours after treatment initiation
Children in the control group were subjected to very close scrutiny, especially during the
first 48 hours and particularly when severe involvement was evident (detection bias)
Notes Open-label trial comparing immediate antibiotics (penicillin G and ampicillin) versus
expectant observation
It was unclear whether otalgia played an important role in the definition of treatment
failure
Data on relapses: N = 126 included in analysis, no crude numbers for separate treatment
groups provided
Risk of bias
Random sequence generation (selection Unclear risk Method of randomisation not described
bias)
Blinding of participants and personnel Unclear risk Open-label trial, outcome assessment not
(performance bias) blinded
Incomplete outcome data (attrition bias) Unclear risk Loss to follow-up - not described for short-
term outcome. Loss to follow-up for long-
term outcome (acute otitis media (AOM)
relapses) - N = 16 (11%), no crude numbers
of separate treatment groups provided
Le Saux 2005
Participants N - 531 children (N = 512 children included in analysis; N = 19 were post-hoc excluded
due to inappropriate randomisation (N = 4) or alternative clinical diagnosis (N = 15))
Age - between 6 months and 5 years
Setting - secondary care: emergency department in Ottawa (Canada)
Inclusion criteria - new onset (< 4 days) of symptoms referable to the upper respiratory
tract and either ear pain or fever (> 38 C). In addition, all patients had to have evidence
of middle-ear effusion, defined by 2 of the following signs: opacity, impaired mobility
on the basis of pneumatic otoscopy and redness or bulging (or both) of the tympanic
membrane
Exclusion criteria - antibiotic treatment < 2 weeks prior to randomisation, contraindi-
cation to amoxicillin or penicillin or sensitivity to ibuprofen or aspirin, presence of otor-
rhoea, co-morbid disease such as sinusitis or pneumonia, prior middle-ear surgery, place-
ment of a ventilation tube, history of recurrent acute otitis media (more than 4 episodes
in 12 months), compromised immunity, craniofacial abnormalities, or any chronic or
genetic disorder
Baseline characteristics - balanced
Interventions Tx - amoxicillin suspension (60 mg/kg) 3 times daily for 10 days; N = 258 (N = 253
included in analysis for day 3)
C - matching placebo for 10 days; N = 254 (N = 246 included in analysis for day 3)
Use of additional medication - parents were given a 5-day supply of antipyretic and
analgesic medication in the form of ibuprofen suspension as required for pain or fever
and a 48-hour supply of codeine elixir to be given as required for pain and fever
Notes
Risk of bias
Allocation concealment (selection bias) Low risk Randomisation sequence was kept under
secure conditions and was accessible only
to the trial pharmacist
Blinding of participants and personnel Low risk Placebo was similar with amoxicillin with
(performance bias) regard to appearance and taste and were
dispensed in identical opaque bottles which
were numbered sequentially
Incomplete outcome data (attrition bias) Low risk Loss to follow-up at day 3 - treatment: N
= 5 (2%) and placebo: N = 8 (3%)
Interventions Tx - immediate treatment with antibiotics: amoxicillin syrup 125 mg/5 mL 3 times daily
for 7 days (children who were allergic to amoxicillin received erythromycin 125 mg/5
mL 4 times daily; N = 151 (N = 135 included in analysis)
C - similar antibiotics were prescribed but parents were asked to wait for 72 hours before
considering using the prescription. Parents were instructed that if their child still had
substantial otalgia or fever after 72 hours, had discharge for > 10 days or was not starting
to get better then they should collect the antibiotic prescription that was left at practice;
N = 164 (N = 150 included in analysis)
Use of additional medication - for both groups doctors emphasised the importance of
paracetamol in full doses for relief of pain and fever
Outcomes Primary outcomes - (a) duration of symptoms (i.e. earache, ear discharge, night distur-
bance, crying); (b) daily pain score; (c) episodes of distress; (d) spoons of paracetamol
used; (e) use of antibiotics
Doctors were asked to provide information on days of illness, physical signs and antibiotic
prescribing; parents were asked to complete a daily symptom diary
Notes Open-label trial comparing immediate versus delayed antibiotic prescription (prescrip-
tion provided but advised to fill only if symptoms did not improve or worsened)
Risk of bias
Random sequence generation (selection Unclear risk Method of randomisation not described
bias)
Allocation concealment (selection bias) Low risk Sealed numbered opaque envelopes
Blinding of participants and personnel Unclear risk Open-label trial, outcome assessment not
(performance bias) blinded
Incomplete outcome data (attrition bias) Low risk Loss to follow-up/exclusion from analysis
(intervention ineffective, did not use an-
tibiotics or did not delay) - treatment: N =
16 (12%) and placebo: N = 14 (9%); com-
parison of the baseline information of the
3 types of responders (those who provided
diaries, those who gave information by tele-
phone and those from whom no diary in-
formation could be collected) revealed no
evidence of significant bias between treat-
ment groups or between patients by age or
severity of symptoms
McCormick 2005
Interventions Tx - immediate treatment with antibiotics: oral amoxicillin 90 mg/kg/day twice daily
for 10 days; N = 112 (N = 110 included in analysis at day 12)
C - expectant observation: no immediate antibiotics; N = 111 (N = 108 included in
analysis at day 12)
Children in the control group with AOM failure or recurrence received oral amoxicillin
90 mg/kg/day; children in Tx group with AOM failure or recurrence received amoxicillin-
clavulanate (90 mg/kg/day of amoxicillin component)
Use of additional medication - all parents received saline nose drops and/or cerumen-
removal drops (if needed), ibuprofen and over-the-counter decongestant/antihistamine
to be given as needed
Outcomes Primary outcomes - (a) parent satisfaction with AOM care; (b) resolution of AOM
symptoms after treatment, including number of doses of symptom medication given;
(c) AOM failure (days 0 to 12) or recurrence (days 13 to 30) defined as attending to
the paediatrician clinic with acute ear symptoms, an abnormal tympanic membrane, or
an AOM severity score higher than that at enrolment; (d) nasopharyngeal carriage of
Streptococcus pneumoniae strains resistant to antibiotics
Secondary outcomes - (a) minor adverse events caused by medication (e.g. allergy,
diarrhoea and candidal infection); (b) serious AOM-related adverse events (e.g. invasive
pneumococcal disease, mastoiditis, bacteraemia, meningitis, perforation of the tympanic
membrane, hospitalisation and emergency ear surgery; (c) parent-child quality-of-life
measures related to AOM
Parents were instructed to complete a symptom diary from day 1 to 10 and a satisfaction
questionnaire on day 12 and day 30; routine follow-up appointments for data collection
were scheduled for day 12 and day 30. Patient-initiated visits was scheduled on request
by the parents for children who seemed to not be responding to treatment
Risk of bias
Blinding of participants and personnel Unclear risk Investigator-blinded study, parents not
(performance bias) blinded
Incomplete outcome data (attrition bias) Low risk Loss to follow-up at day 12 - treatment: N
= 2 (2%) and expectant observation: N = 3
(3%)
Interventions Tx - penicillin 50 mg/mL 4 times daily; children aged 1 to 2 years: 10 mL daily, children
between 3 and 5 years: 20 mL daily, children between 6 and 10 years: 30 mL daily for
7 days; N = ? (N = 72 included in analysis)
C - placebo for 7 days; N = ? (N = 77 included in analysis)
Use of additional medication - acetylsalicylic acid tablets (maximum of 50 mg/kg/day
for 3 days) were supplied as the only supplementary treatment permitted
Outcomes Main outcomes: (a) mean symptoms (i.e. pain, fever) scores; (b) number of analgesic
tables used; (c) contralateral otitis; (d) spontaneous perforation of tympanic membrane;
(e) mean number of days of otorrhoea; (f ) tympanometry results at 1 week, 4 weeks and
3 months
Initial visits were performed at home: otoscopy and bacterial culture from nasopharynx
were performed
Score cards were given to parents
Follow-up visits at hospital at day 2 to 3, day 7, week 4 and week 12. If supplementary
treatment was required at day 2 to 3, then myringotomy was performed. If supplementary
treatment was required at day 7, then amoxicillin was given
Notes
Risk of bias
Random sequence generation (selection Unclear risk Method of randomisation not described
bias)
Blinding of participants and personnel Unclear risk Identical taste and appearance of amoxi-
(performance bias) cillin and placebo not described
Incomplete outcome data (attrition bias) Unclear risk Patients not included in analysis - N =
9 (6%). Reasons described, unclear from
which treatment group patients were ex-
cluded
Neumark 2007
Participants N - 186 children (N = 179 patients were included in analysis; 7 patients were excluded
due to non-compliance with protocol)
Age - between 2 and 16 years
Setting - general practice: 32 healthcare centres and 72 general practitioners in Sweden
Inclusion criteria - acute otitis media (AOM) was based on direct inspection of the
eardrum by pneumatic otoscope or preferably an aural microscope. Findings had to
include a bulging, red eardrum displaying reduced mobility
Exclusion criteria - perforation of the eardrum, chronic ear conditions or impaired
hearing, previous adverse reactions to penicillin, concurrent disease that should be treated
with antibiotics, recurrent AOM (3 or more AOM episodes during the past 6 months)
, children with immunosuppressive conditions, genetic disorders and mental disease or
retardation
Baseline characteristics - balanced
Outcomes Primary outcomes - (a) pain at day 0, 1, 2 and 3 to 7; (b) use of analgesics at day 0, 1,
2, 3, 4 to 7; (c) fever > 38 C at day 0, 1, 2 and 3 to 7; (d) subjective recovery at day 14
and 3 months; (e) perforations at 3 months; (f ) serous otitis media at 3 months
All participants were asked to complete a symptom diary for 7 days; a nurse telephoned
all participants after approximately 14 days to supplement the information in the diary
and to register all acute contacts that had occurred during the first week of treatment; the
final follow-up was performed after 3 months to register perforations and serous otitis
media
Notes Open-label trial comparing immediate antibiotic prescribing versus expectant observa-
tion (no prescription provided but advice on what to do if symptoms did not improve
or worsened)
Risk of bias
Random sequence generation (selection Low risk Internet-based random number generator
bias)
Blinding of participants and personnel Unclear risk Open-label trial, outcome assessment not
(performance bias) blinded
Incomplete outcome data (attrition bias) Unclear risk Patients not included in analysis - N =
7 (4%). Reasons described, unclear from
which treatment group patients were ex-
cluded
Spiro 2006
Outcomes Primary outcome - proportion of each group that filled the prescription for an antibiotic.
This was defined by whether the parent filled the prescription within 3 days of enrolment
and was determined by the response to this question at the interview at day 4 to 6
Secondary outcomes - (a) clinical course of the illness; (b) adverse effects of medications;
(c) days of school or work missed; (d) unscheduled medical visits; (e) comfort of parents
with management of AOM without antibiotics for future episodes
2 trained research assistants blinded to group assignment conducted standardised, struc-
tured telephone interviews with the parents at day 4 to 6, day 11 to 14, day 30 and day
40 after enrolment
Risk of bias
Blinding of participants and personnel Unclear risk Investigator-blinded study, parents not
(performance bias) blinded
Incomplete outcome data (attrition bias) Unclear risk Loss to follow-up at day 4 to 6 treatment:
N = 12 (8%) and expectant observation: N
= 6 (4%)
Methods Randomised - yes, block randomisation, method of random sequence generation not
described
Concealment of allocation - adequate
Double-blind - yes
Intention-to-treat (ITT) - unclear
Loss to follow-up - described
Design - parallel
Interventions Tx - phenoxymethyl penicillin 50 mg/kg/day twice daily for 7 days; N = 159 (N = 159
included in analysis)
C - matching placebo in 2 doses for 7 days; N = 158 (N = 158 included in analysis)
Use of additional medication - all children were given nose drops containing oxymeta-
zoline chloride and, if needed, analgesics (paracetamol)
Notes
Risk of bias
Random sequence generation (selection Unclear risk Block randomisation, method of random
bias) sequence generation not described
Allocation concealment (selection bias) Low risk Randomisation list was kept by the clinical
pharmacologist of the hospital and not dis-
closed to the investigators until the clinical
trial was completed
Blinding of participants and personnel Low risk Placebo with same taste and appearance of
(performance bias) penicillin
Incomplete outcome data (attrition bias) Low risk No children lost to follow-up for primary
analysis
Thtinen 2011
Methods Randomised - yes, computerised random number generator with block length of 10
Concealment of allocation - adequate
Double-blind - yes
Intention-to-treat (ITT) - yes
Loss to follow-up - described
Design - parallel
Outcomes Primary outcome - time to treatment failure (i.e. a composite end point consisting
of 6 independent components: (a) no improvement in overall condition at day 2, (b)
worsening of the childs overall condition at any time, (c) no improvement in otoscopic
signs at day 7, (d) perforation of tympanic membrane at any time, (e) severe infection (e.
g. mastoiditis or pneumonia) necessitating systemic open-label antimicrobial treatment
at any time, (f ) any other reason for stopping the study drug at any time
Secondary outcomes - assessed by study physician - (a) time to the initiation of rescue
treatment; (b) time to development of contralateral AOM; - diary symptom assessment;
(c) resolution of symptoms; (d) use of analgesics
Parents were given a diary to record symptoms, doses of study drugs and any other
medications and adverse events
First visit after enrolment (= day 0) was scheduled at day 2. End-of-treatment visit was
scheduled at day 7
Notes
Risk of bias
Random sequence generation (selection Low risk Computerised random number generator
bias) with block length of 10
Allocation concealment (selection bias) Low risk Concealment of allocation by the pharma-
cist (independent to trial team) by labelling
the identical opaque study drug contain-
ers with allocation numbers; allocation list
was kept at the paediatric infectious disease
ward behind locked doors
Blinding of participants and personnel Low risk Placebo with same taste and appearance of
(performance bias) amoxicillin-clavulanate
Incomplete outcome data (attrition bias) Low risk Loss to follow-up - treatment: N = 1 (1%)
and placebo: N = 2 (1%)
vanBuchem 1981a
Participants N - 202 children (N = 171 children included in analysis; N = 31 were excluded from
the study)
Age - between 2 and 12 years
Setting - both general practice and secondary care: 12 general practitioners in or near
Tilburg (the Netherlands) recruited patients and referred them to 1 of the 3 otorhino-
laryngologists, which excluded those cases where there was disagreement with the diag-
nosis
Inclusion criteria - acute otitis media (AOM) was based on history and clinical picture
(i.e. diffuse redness, bulging of the eardrum, or both)
Exclusion criteria - antibiotic treatment < 2 weeks prior to randomisation, chronic otitis
or otitis media serosa, contraindication for antibiotic treatment
Baseline characteristics - balanced
Interventions Tx - sham myringotomy and amoxicillin 250 mg 3 times daily for 7 days; N = 47
C - sham myringotomy and matching placebo for 7 days; N = 40
Use of additional medication - all participants were allowed to use decongestive nose
drops and analgesic suppositories (i.e. children aged 2 to 7 years: acetylsalicylic acid 50
mg, phenacetin 50 mg, phenobarbitone 15 mg, codeine phosphate 2.5 mg, caffeine 1.
25 mg; children aged 8 to 12 years: acetylsalicylic acid 100 mg, phenacetin 100 mg,
phenobarbitone 30 mg, codeine phosphate 5 mg, caffeine 2.5 mg
Outcomes Main outcomes - (a) parent report of pain at day 0, 1 and 7; (b) otoscopic findings at
day 0, 1 and 7; (c) discharge from ear at day 1, 7 and 14; (d) mean temperature at day
0, 1 and 7; (e) AOM relapses at 6 months; (f ) audiogram findings after 4 and 8 weeks
Risk of bias
Random sequence generation (selection Unclear risk Method of randomisation not described
bias)
Blinding of participants and personnel Low risk Sham myringotomy and placebo was simi-
(performance bias) lar with amoxicillin with regard to appear-
ance and taste
Participants N - 202 children (N = 171 children included in analysis; N = 31 were excluded from
the study)
Age - between 2 and 12 years
Setting - both general practice and secondary care: 12 general practitioners in or near
Tilburg (the Netherlands) recruited patients and referred them to 1 of the 3 otorhino-
laryngologists which excluded those cases where there was disagreement with the diag-
nosis
Inclusion criteria - acute otitis media (AOM) was based on history and clinical picture
(i.e. diffuse redness, bulging of the eardrum, or both)
Exclusion criteria - antibiotic treatment < 2 weeks prior to randomisation, chronic otitis
or otitis media serosa, contraindication for antibiotic treatment
Baseline characteristics - balanced
Outcomes Main outcomes - (a) parent report of pain at day 0, 1 and 7, (b) otoscopic findings at
day 0, 1 and 7; (c) discharge from ear at day 1, 7 and 14; (d) mean temperature at day
0, 1 and 7; (e) AOM relapses at 6 months; (f ) audiogram findings after 4 and 8 weeks
Risk of bias
Random sequence generation (selection Unclear risk Method of randomisation not described
bias)
Blinding of participants and personnel Low risk Sham myringotomy and placebo was simi-
(performance bias) lar with amoxicillin with regard to appear-
ance and taste
Arguedas 2011 No comparison of antibiotic to placebo or expectant observation: trial comparing single-dose extended-release
azithromycin versus a 10-day regimen of amoxicillin/clavulanate
Casey 2012 No comparison of antibiotic to placebo or expectant observation: trial comparing amoxicillin/clavulanate high
dose versus cefdinir
Engelhard 1989 No comparison of antibiotic to placebo; the 3 arms were Augmentin, myringotomy, or both
Liu 2011 No comparison of antibiotic to placebo or expectant observation: trial comparing single oral doses azithromycin
of extended-release versus immediate-release formulations
Rudberg 1954 Non-randomised study: assigned randomly based on case-number but then allowed to change groups
Sarrell 2003 No comparison of antibiotic to placebo. Method of randomisation not provided and groups appear to be
unbalanced at baseline
Thtinen 2012 Secondary analysis of placebo-controlled trial. This study included the total group of children allocated to
immediate antimicrobial treatment (N = 161) and a subgroup of children from the placebo group that received
delayed antibiotics (N = 53). As a consequence, comparability of prognosis achieved through randomisation is
violated, producing groups of children that are incomparable which may lead to biased effect estimates
No. of No. of
Outcome or subgroup title studies participants Statistical method Effect size
No. of No. of
Outcome or subgroup title studies participants Statistical method Effect size
1 Pain at 3 to 7 days 4 959 Risk Ratio (M-H, Random, 95% CI) 0.75 [0.50, 1.12]
2 Tympanic membrane perforation 1 179 Risk Ratio (M-H, Fixed, 95% CI) 0.0 [0.0, 0.0]
3 AOM recurrences 1 209 Risk Ratio (M-H, Fixed, 95% CI) 1.41 [0.74, 2.69]
4 Vomiting, diarrhoea or rash 2 550 Risk Ratio (M-H, Fixed, 95% CI) 1.71 [1.24, 2.36]
Outcome: 1 Pain
1 4 to 6 weeks
Appelman 1991 21/51 25/45 5.6 % 0.74 [ 0.49, 1.13 ]
Analysis 1.3. Comparison 1 Antibiotics versus placebo, Outcome 3 Tympanic membrane perforation.
Expectant
observa-
Study or subgroup Antibiotics tion Risk Ratio Weight Risk Ratio
M- M-
H,Random,95% H,Random,95%
n/N n/N CI CI
Little 2001 26/151 44/164 28.2 % 0.64 [ 0.42, 0.99 ]
Expectant
observa-
Study or subgroup Antibiotics tion Risk Ratio Weight Risk Ratio
n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI
Neumark 2007 0/92 0/87 Not estimable
Analysis 2.3. Comparison 2 Immediate antibiotics versus expectant observation, Outcome 3 AOM
recurrences.
Review: Antibiotics for acute otitis media in children
Expectant
observa-
Study or subgroup Antibiotics tion Risk Ratio Weight Risk Ratio
n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI
McCormick 2005 20/109 13/100 100.0 % 1.41 [ 0.74, 2.69 ]
Expectant
observa-
Study or subgroup Antibiotics tion Risk Ratio Weight Risk Ratio
n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI
Little 2001 31/135 22/150 45.4 % 1.57 [ 0.95, 2.57 ]
APPENDICES
Appendix 1. Previous search
Several electronic databases were used to compile relevant published RCTs of antibiotic treatment of AOM in children. The Cochrane
Controlled Trials Register, MEDLINE and Current Contents were searched from 1966 to January 2000 by an expert librarian in
conjunction with one researcher, using combinations of OTITIS MEDIA and a search strategy described by (Dickersin 1994) for
optimally identifying controlled trials. In addition, titles in Index Medicus were checked from 1958 to 1965. The references of all
relevant retrieved trials were checked to identify other articles.
The search was updated in March 2003, and again in July 2008. We searched the Cochrane Central Register of Controlled Trials
(CENTRAL) (The Cochrane Library, 2008, issue 2) which contains the ARI Groups Specialized Register; MEDLINE (1966 to June
week 4 2008); OLDMEDLINE (1958 to 1965); EMBASE (January 1990 to July 2008); and Current Contents (1966 to July 2008).
The bibliographies of relevant articles were checked. A forward search of relevant articles was conducted in Web of Science.
The following search strategy was run on MEDLINE (Ovid) combined with terms from Phase 1 and 2 of the Cochrane highly sensitive
search strategy for identifying reports of RCTs (Lefebvre 2011). Modified terms were used to search the other databases:
Antibiotics for acute otitis media in children (Review) 57
Copyright 2013 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
MEDLINE (Ovid)
#1 exp Otitis Media/
#2 exp Otitis Media with Effusion/
#3 exp Otitis Media, Suppurative/
#4 glue ear.mp.
#5 otitis media.mp.
#6 OME.mp.
#7 AOM.mp.
#8 #1 or #2 or #3 or #4 or #5 or #6 or #7
#9 exp Anti-Bacterial Agents/
#10 exp Drug Therapy/
#11 exp Anti-Infective Agents/
#12 antibiotic$.mp.
#13 #9 or #10 or #11 or #12
#14 #8 and #13
There were no language or publication restrictions.
#3 578 #2 AND #1
Databases=CM, LS Timespan=All Years
Lemmatization=On
#2 528,401 Topic=(random* or placebo* or crossover* or cross over or allocat* or ((doubl* or singl*) NEAR/1 blind*)) OR
Title=(trial)
Databases=CM, LS Timespan=All Years
Lemmatization=On
#1 2,624 Topic=(otitis or glue ear or (middle ear NEAR/3 (infect* or inflam*)) or ome or aom) AND Topic=(antibiotic*
or antibacterial* or antiinfective* or ampicillin* or cephalosporin* or macrolide* or amoxicillin* or amoxycillin* or
penicillin* or cefdinir* or cefpodoxime* or cefuroxime* or azithromycin* or clarithromycin* or erythromycin*)
Databases=CM, LS Timespan=All Years
Lemmatization=On
FEEDBACK
Summary
1. Types of interventions includes surgical procedures versus placebo which are not dealt with in this review and should therefore be
deleted.
2. The authors included only six studies in the analysis but in 1994 another meta-analysis by Rosenfeld and colleagues to which the
authors refer was published which included 33 randomized trials with 5400 children. Were any studies with a no-treatment control
excluded and if so why?
3. The meta-analysis by Rosenfeld is only mentioned in the text; there is no reference to it. How many patients were included in the
meta-analysis?
4. It is stated that trials analysed on an intention to treat basis were preferred. This indicates that other trials were excluded which does
not seem reasonable?
5. The description of the factorial trial is unclear; I suppose the authors excluded all patients who were randomised to myringotomy?
6. In the trial by Laxdal the control group was more closely monitored. The trial therefore violates the principle that all other Traitement
etc. should be the same in the two randomised groups and it should therefore be excluded.
7. The strategy described by Dickersin lacks a publication year and it is not cited in the references.
8. The search was done in August 1994 and the Cochrane review was published in April 1997. The search should therefore have been
updated before publication since Cochrane reviews are meant to be up-to-date.
9. There is no information whether the original authors and the pharmaceutical industry were contacted about additional data including
unpublished trials and trials not registered in Medline. Useful trial data might be expected to be available in books published in
connection with symposia arranged by the drug industry for example.
10. What is quality methodology?
11. The term blinded randomisation should be avoided since it may be confused with blinded treatments; the term concealed allocation
should be used.
12. The elaborated quality assessment scale for the trials does not appear under Results and should therefore be deleted.
13. The authors refer to Rosenfelds meta-analysis when they state that 80% of the children have recovered spontaneously after 24
hours. Since such a percentage refers to untreated patients it raises the question why the authors did not use their own data? If these
data are used in a meta-analysis of the risk difference the NNTB will be 23 not 12 as stated in the Cochrane review.
14. For several of the excluded studies the authors gave no reason for the exclusion.
15. There should be a cross-reference to the authors nearly identical review in the BMJ (24 May 1997).
Reply
The changes made were:
Antibiotics for acute otitis media in children (Review) 60
Copyright 2013 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
1. We updated the search. (see Johansen criticism 7 & 8). No recent trials were found but we recognised that the Appelman trial
qualifies (originally we had thought this was only prevention of recurrent otitis, rather than treatment of acute otitis in children with a
recurrent episode).
2. We have corrected and updated the Relative Risk Reduction and consequent Number-Needed-to-Treat (see Johansen criticism 13).
3. We have separate the four arms of the Van Buchem factorial trial, and treated this as two trials (i.e., two separate strata): (a) without
myringotomy - antibiotics versus placebo (b) with myringotomy - antibiotics versus placebo. (see Johansen criticism 5)
4. As suggested by Andrew Herxheimer, we have added several references including (a) Chris Cates BMJ, and (b) Kozrskyjs meta-
analysis of short versus long duration of antibiotics (rather than just the de Saintonge paper).
5. We have made small text changes in response to Johansens criticisms 5 (description added), 7 (dropped), 10 (- methodological
quality), 11 (- allocation concealment), 13 (corrected in text), 14 (exclusions explained), and 15 (reference added).
6. As we have pointed out to Johansen in the BMJ correspondence, and point out in the discussion here, the Rosenfeld meta-analysis
is largely concerned with comparison between antibiotics. (see Johansen criticism 2 & 3).
Contributors
Helle Krogh Johansen
Peter C. Gtzsche
Summary
This excellent and important review was completed in 1996, and I hope it will soon be updated. It is especially worth noting and
discussing the new study by Christopher Cates (BMJ 13 March 1999, p715-6), who has successfully tried a method in his general
practice of substantially reducing the use of antibiotic in children with acute otitis media. This would considerably strengthen the
implications for practice in the conclusion.
I would like to suggest that in updating this review the objectives be amended and the trial by Chaput de Saintonge et al be added,
because it contributes an important piece of evidence about the duration of amoxicillin therapy. The review concludes that some
children will benefit from antibiotic treatment, and it would be valuable to say (as a result of the Chaput trial) that the evidence
indicates that a 3-day course is no less effective than a 10-day course.
Reply
Chris and I have revised the acute otitis media review. We have made a number of modest changes, though none of these change
the conclusions. However, because a new trial is included weve called it a substantive update.
Contributors
Andrew Herxheimer
Summary
1. I am glad to see this has been updated but the text does not explain what was updated, forcing the reader who wants to know to
compare the previous version with the new one. Is it the sentence referring to Cates 99 [in implics for practice] or other points as
well?
2. There are embarrassingly many typos in the refs to excluded and additional studies: Chaput de SaintoNGE, amoxyciillin, author
not in bold in the first few additional refs, below that several authors names begin in lower case when they should all begin with a
capital.
3. It is implied that no comcrit was received before the final submission date for CL99 issue 3. Is this true? I think I sent one early
this year.
CONFLICT OF INTEREST: None.
Reply
Excluded and additional references have been corrected and completed.
Contributors
Andrew Herxheimer
Summary
1. The new study also reported diarrhoea and rashes. Shouldnt it be included in this outcome (side effects) also?
2. I think the methods used for calculating the NNTB should be made explicit.
3. The new trial is important because it looks at a sub-group who were believed to be a greater risk of poor outcomes. In EBM OM
Rosenfeld and Bluestone review the study inclusion criteria and state that the meta-analysis most likely can be applied to children 2
years of age or older with non severe AOM, and most likely cannot be applied to infants with severe symptoms. This study provides
the best evidence that the conclusions of the meta-analysis do appear to apply to this group. Perhaps this point needs to be emphasised
(the peak incidence of AOM is 9 months).
4. I think the comment that 80% resolve spontaneously within 2 to 7 days is now slightly misleading as about 70% of the control
children were clinical failures in this new study.
5. The entry in the table characteristics of included studies should be consistent with previous entries.
6. Some typographical errors and inconsistent spelling.
Reply
Thank you for your comments and suggestions.
The Absolute risk difference was used to calculate the NNTB in this systematic review. This has now been stated in the Results
section of the review. A comment regarding the application of the conclusions to infants with severe symptoms has been added to the
discussion section. The 70% incidence of clinical failure in the Damoiseaux, 2000 study have been included and typographical errors
and inconsistencies have been corrected.
Contributors
Peter Morris
Summary
The second graph (comparison of outcome Abnormal Tympanometry) has wrong labels on the X-axis.
It says antibiotics better (left) and placebo worse (right). The second should probably be placebo better.
The other graphs are correctly labelled.
I certify that I have no affiliations with or involvement in any organisation or entity with a direct financial interest in the subject matter
of my
criticisms.
Reply
The label on the x-axis has been corrected and now reads Placebo better.
Contributors
Johannes C van der Wouden
Summary
I agree with other commentators that this is a very good and important review. However, I would like some more clarity concerning
one statement in your conclusions: Antibiotic treatment may play an important role in reducing the risk of mastoiditis in populations
where it is more common.
What is the basis for this statement? In the included studies with more than 2000 children only one mastoiditis case occurred in a
patient in a penicillin treated group. In the review you mention two articles concerning the mastoiditis. Firstly, the study of Rudberg
(1954), which was excluded since it was not properly randomised. Even if it were, the rate of 17 % of mastoiditis cases is in these times
highly unlikely, as is shown in the included studies. The second article by Berman (1995) is a literature review, where only the available
literature concerning developing countries were reviewed. The goal of this review was to determine the extent to which otitis media
impacts mortality and morbidity in developing countries, not to study the effect of antibiotics on (acute) otitis media or mastoiditis.
In neither of these studies evidence is shown that antibiotic treatment reduces the risk of mastoiditis, certainly not in developed
countries. Since I think the rest of the review is excellent, I wonder if you could explain to me the reasons for including this statement
in the conclusions.
I certify that I have no affiliations with or involvement in any organisation or entity with a direct financial interest in the subject matter
of my criticisms.
Reply
Dear Markus,
We included the caveat about mastoiditis because we, and the reviewers, were concerned about misinterpretation of the results in
situations with high rates of mastoiditis. We were mindful that an absence of evidence is not equal to evidence of absence. Since the
trials we analysed did not include high rates of mastoiditis, we can use them as the sole basis. Given that we have two weaker pieces of
evidence:
1. The trials do show a modest reduction in other infective complications
2. The excluded Rudberg trial did show dramatic effects that we dont think explicable from the potential biases of that study.
Prudence would then suggest that antibiotics are advisable if there is a substantial risk of mastoiditis,
Regards,
Paul Glasziou
I certify that I have no affiliations with or involvement in any organisation or entity with a direct financial interest in the subject matter
of my criticisms.
Antibiotics for acute otitis media in children (Review) 63
Copyright 2013 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
Contributors
Markus Oei (ENT surgeon)
Summary
I am a bit troubled by the way the conclusions of this review are written. By combining results of treatment at Days 2 to 7 in arriving
at a NNTB of 15 one is going to underestimate treatment benefit after 2 days. In your abstract though you say the ARR is 7% and
NNTB 15 for some pain after two days. This is simply not correct. If one carefully looks at trials that record pain at the end of day 2
the ARR is in fact 20% giving a NNTB of 5. Clearly acute otitis media is an acute condition and the main benefit of antibiotics is pain
control and symptom relief. If this is measured at the end of 2 days the benefits are greater than one would surmise just from reading
the review. It would be absurd to do a review of pain relief for biliary colic treated with pethidine and measuring the outcome 7 days
later. For acute conditions symptom control in the first few days should be the outcome of interest. NNTB are meaningless unless
giving a time period at which they apply. I think the review needs correcting. This is not just of academic interest but of direct relevance
to parents and doctors faced with a child with AOM in pain. Unfortunately your review gets quoted uncritically and invariably the
NNTB of 15 is given for symptom control after 2 days. I am currently trying to correct a brochure produced here in New Zealand
for GPs to give to parents of children with AOM and it uncritically repeats this misleading information. If you want to comment on
symptom control after Day 2 DO NOT pool it with data from Day 7 or later!
I certify that I have no affiliations with or involvement in any organisation or entity with a direct financial interest in the subject matter
of my criticisms.
Reply
Thankyou for your comment. We agree that we should be clearer about the time frame to which the ARR 7% and NNTB 15 applies.
With the availability of results of the individual patient data meta-analysis (Rovers 2006) we are able to obtain a clearer indication of
the recovery pattern over time. We have reported this in the text and included an extra figure.
Contributors
Paul Corwin
Summary
Summary
Feedback: This is a comment on two of the meta-analyses in the Cochrane Review, Glasziou et al. (2004). These analyses are for the
outcomes Vomiting, Diarrhea or Rash and Contralateral AOM.
1) Vomiting, Diarrhea or Rash
First we consider the meta-analysis relating possible adverse effects of treatment. In Glasziou et al. (2004), this is done using the
composite outcome Vomiting, Diarrhea or Rash. The data used for this meta-analysis are reproduced in the table below.
Outcome: Vomiting, Diarrhea or Rash
Study Treatment Control
Thalin et al. (1985) 1/159 1/158
Burke et al. (1991) 53/114 36/116
Mygind et al. (1981) 3/72 1/77
Damoiseaux et al. (2000) 20 12
We noted five major problems with this meta-analysis. The first relates to clinical heterogeneity. This was manifested in variations in
terms of the types of adverse effects recorded, who recorded them (parent or physician) and the time period over which they were
Antibiotics for acute otitis media in children (Review) 64
Copyright 2013 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
recorded (from 3-4 days to 21 days). In Thalin et al. (1985), the effects were recorded by an ENT physician on days 3-4 or days 8-10.
In Burke et al. (1991), they were recorded by a parent in a 21-day diary. In Mygind et al. (1981), it was done with 7 day parental score
card. And in Damoiseaux et al. (2000), this was done by a physician on day 4 and day 11.
Another related problem is the use of the outcome Vomiting, Diarrhea or Rash as an entity. Vomiting is only reported in Burke et
al. (1991). It is not clear whether it was not observed, or observed but not reported in the other studies. Also, in Burkeat al. (1991),
as noted, such effects were recorded over a 21-day period while the maximum recording period for the other studies was 11 days. The
totals then gave a much higher weight to Burke et al. (1991) than may be appropriate.
A third problem is possible double or triple counting with the use of the composite outcome. For Burke et al. (1991), the group
numerator is the sum of the cases for each effect. A number of children may well have had two or three of these effects at the same time.
A fourth problem is also with the numbers used. Damoiseaux et al. (2000) gives two sets of numbers for de novo diarrhoea, for day
4 and for day 11. Glasziou et al. (2004) uses the day 4 numbers only. The reason for this choice is not clear. It may be better to use the
sums of the numbers for the two days (provided this does not involve double counting.)
Further, the group denominators used for Burke et al. (1991) are perhaps not what they should be. In this study, the adverse effects
were recorded by parents. Only 220 (treatment = 107, control = 113) out of a total of 232 (treatment = 114, control = 118) diaries
were collected. Using the total group size in the numerator (also done in Burke at al. (1991)) is thus not appropriate.
Finally, it is not clear if the numbers for adverse effects in Burke et al. (1991) and Damoiseaux et al. (2000) included the cases known
or suspected to have dropped out of the study due to an adverse effect.
In our view, this meta-analysis should be modified as follows: First, do not use the data on vomiting until it is reported in at least one
other study. Second, do not use a composite adverse effect outcome. Instead, perform separate meta-analyses for diarrhoea and rash.
Third, for Damoiseaux et al. (2000), use the total numbers for day 4 and day 11, with the above noted qualification in mind. Fourth,
for Burke et al. (1991) change the denominators as noted above. Finally, include drop outs due to side effects in the meta-analyses. The
table below gives the possible numerators to be used for these meta-analysis.
Separated Data on Side Effects
Vomiting Diarrhea Rash
Study T C T C T C
Thalin et al. (1985) ? ? 0 0 1 1
Burke et al. (1991)+ 20 14 24 16 16 9
Mygind et al. (1981) ? ? 2 1 1/2? 0
Damoiseaux et al. (2000)*,+ ? ? 20 12 0 3
Damoiseaux et al. (2000)? ? ? 34 22 0 3
Note: ? Unclear if vomiting not observed or not reported.
Note: ? = 2 if a dropout was not counted; else = 1.
* Day 4; ? Day 4 and Day 11; + unclear if dropouts counted.
2) Contralateral AOM
The occurrence of contralateral AOM, as is made clear in Glasziou et al. (2004), is relevant for only the cases with unilateral AOM at
the outset. This numbers in the table below are used for the meta-analysis of this outcome in Glasziou et al. (2004).
Outcome: Contralateral AOM
Study Treatment Control
Thalin et al. (1985) 4/159 17/158
Burke et al. (1991) 29/98 33/102
Mygind et al. (1981) 2/72 6/77
Overall 35/329 56/337
The first problem is clinical heterogeneity, as noted in the table below. The issues in that respect are similar to those stated for the meta-
analysis of adverse effect.
Clinical Heterogeneity: Contralateral AOM
Study Time Period Evaluator(s)
Thalin et al. (1985) day 8-10 or day 30 ENT Physician
Burke et al. (1991) 21 days Parent
Mygind et al. (1981) 1 week Physician
A further problem with this meta-analysis is the denominators used. Consider this issue for each study.
Thalin et al. (1985): The denominators in Glasziou et al. (2004) include unilateral and bilateral cases. Only 82% of the episodes were
unilateral at the start but the breakdown by group is not given in the paper. We obtained adjusted denominators as follows. Treatment:
0.82?159 = 130; Control: 0.82?158 = 130. The bias now remains the same but the precision level is now corrected.
Antibiotics for acute otitis media in children (Review) 65
Copyright 2013 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
Burke et al. (1991): The denominators represent the total unilateral cases for each group. The study authors used these denominators.
Completed 21-day diaries, the source of data on contralateral otitis, were, however, available only for 107 (of 114) in the treatment
group and 113 (of 118) in the control group. So either one assumes that only the bilateral cases had missing diaries (which is unlikely)
or that the rate of missingness in each group was not affected by laterality. In the latter case, the adjusted denominators are: Treatment:
(98?107)/114 = 92; Control: (102?113)/118 = 98. The level of bias remains unknown but the precision level is possibly better.
Mygind et al. (1991): The denominators used include unilateral and bilateral cases. But there were 8 bilateral cases in the placebo
group and 14 in the control group. So the appropriate denominators are Treatment: 72 - 8 = 64; Control: 77 - 14 = 65. The bias and
precision levels are now corrected.
The appropriately adjusted data for this meta analysis are given below.
Contralateral AOM: Adjusted Data
Study Treatment Control
Thalin et al. (1985) 4/130 17/130
Burke et al. (1991) 29/92 33/98
Mygind et al. (1981) 2/64 6/65
Overall 35/286 56/294
References
1. Burke P, Bain J, Robinson D and Dunleavey J (1991) Acute red ear in children: Controlled trial of non-antibiotic treatment in
general practice, British Medical Journal, 303, 558?562.
2. Damoiseaux RAMJ, van Balen FAM, Hoes AW, Verheij TJM and de Melker RA (2000) Primary care based randomised, double
blind trial of amoxicillin versus placebo for acute otitis media in children aged under 2 years, British Medical Journal, 320: 330?334.
3. Glasziou PP, Del Mar CB, HayemMand Sanders SL (2004) Antibiotics for acute otitis media in children, Cochrane Database of
Systematic Reviews, 2004; (1): CD000219. Art. No: CD000219, DOI: 10.1002/14651858.CD000219.pub2 (21pages)
4. Mygind N, Meistrup-Larsen K-I, Thomsen J, Thomsen VF, Josefsson K and Sorenson H (1981) Penicillin in acute otitis media: a
double-blind placebo-controlled trial, Clinical Otolaryngology, 6: 5?13.
5. Thalin A, Densert O, Larsson A, Lyden E and Ripa T (1986) Is penicillin necessary in the treatment of acute otitis media? In:
Proceedings of the International Conference on Acute and Secretory Otitis Media, Amsterdam, The Netherlands, Kegler Publications,
pages 441?446.
Submitter agrees with default conflict of interest statement:
I certify that I have no affiliations with or involvement in any organization or entity with a financial interest in the subject matter of
my feedback.
Reply
1) We acknowledge the variation in methods of collecting and recording information on adverse events and in the types of adverse
events reported in the included trials. We contend however, that considering vomiting, diarrhoea or rash as an entity is justified by the
easier interpretation it provides . Though the events are biologically very different, they are of similar seriousness; irritating and difficult
to manage but minor in nature. Also, as pointed out in the above comments, dividing the adverse events into each type would not be
helpful as they are infrequently reported (i.e. vomiting is only reported in one study). We recognise that lumping the adverse events
together is a crude approach but believe the benefits in continuing to do so outweigh the drawbacks. In the discussion section of this
update we have made reference to the results of the individual patient data meta analysis (Rovers 2006) (which included a subset [n =
6 ] of the trials included in this review [n = 10]) which reports separately on the frequency of diarrhoea and rash in the treatment and
control groups. We appreciate your consideration and suggestions related to the inclusion of drop outs due to side effects in the Burke
and Damoiseaux studies. Corrections to the data have been incorporated.
2) Thankyou for pointing out the numerical errors in the meta analysis of contralateral AOM. We have corrected the analysis as
suggested. This results in a minor changed to the pooled random effects OR (OR 0.44 95% CI 0.16, 1.26 versus 0.45 95% CI 0.16,
1.23) with antibiotics appearing to reduce contralateral AOM though the effect was not significant with the random effects model.
Contributors
Karim F. Hirji, D.Sc
Peter C. Gtzsche
Summary
The title and conclusion of the review need revising as it is just reviewing the effect of penicillin family antibiotic on the AOM and
not other antibiotics. It is suggesting to changed the title to Usage of penicillin family Antibiotics for acute otitis media in children.
Warm regards.
P.S: The only included trials were too old and they just used the publish data:
Halsted 1968 ampicillin 100 mg/kg/day or phenethicillin 30 mg/kg/day plus sulphisoxazole 150 mg/kg/day
Howie 1973 one of erythromycin, ampicillin, or triple sulphonamide plus erythromycin
Submitter agrees with default conflict of interest statement: I certify that I have no affiliations with or involvement in any organization
or entity with a financial interest in the subject matter of my feedback.
Reply
The title is our intention. However, as you point out, it just so happens that most (but not all) antibiotics trialled against placebo for
acute otitis media were from the penicillin group. Moreover more trials might be undertaken using non-penicillin antibiotics. So it is
appropriate to retain the original title.
Chris Del Mar, 19 June, 2012
Contributors
Amirkambiz Hamedanizadeh, Medical Doctor
WHATS NEW
Last assessed as up-to-date: 8 November 2012.
8 November 2012 New citation required but conclusions have not The general conclusions and recommendations regard-
changed ing the effectiveness of antibiotics on pain and adverse
events remained unchanged. Antibiotic treatment lead
to a statistically significant reduction of children with
AOM experiencing pain at two to seven days com-
pared with placebo, but since most children (82%) set-
tle spontaneously, about 20 children must be treated to
prevent one suffering from ear pain at two to three and
four to seven days. (In the previous version the number
needed to treat to benefit (NNTB) was 16). However,
in this updated review antibiotic treatment appeared
to have a statistically significant beneficial effect on the
number of tympanic membrane perforations (risk ratio
(RR) 0.37, 95% confidence interval (CI) 0.18 to 0.76;
NNTB 33) and contralateral acute otitis media (AOM)
episodes (RR 0.49, 95% CI 0.25 to 0.95, NNTB 11)
as compared with placebo
For every 14 children treated with antibiotics one child
8 November 2012 New search has been performed A new review author joined the team to update this re-
view. We updated the searches in November 2012. Two
new trials were identified for the review of antibiotics
against placebo (Hoberman 2011; Thtinen 2011).
These studies included children < 35 months of age and
provided data on pain (Thtinen 2011), contralateral
otitis, late recurrences (Hoberman 2011), perforation
and adverse events (Hoberman 2011, Thtinen 2011)
.
The Laxdal 1970 trial has been removed from the re-
view of antibiotics against placebo and added to the
review of immediate antibiotics versus expectant obser-
vation
No new trials were identified for the review of imme-
diate antibiotics compared with expectant observation.
Furthermore, we did not identify ongoing trials
In this updated review, we now provide outcome data
for pain at 24 hours, two to three days and four to
seven days (in earlier versions outcome data for pain
were presented at 24 hours and two to seven days)
HISTORY
Protocol first published: Issue 1, 1995
Review first published: Issue 3, 1996
2 September 2009 Amended 95% confidence intervals for the outcome pain at 2-7
days and adverse events stated in the abstract and body
of the review corrected
2 July 2008 New search has been performed The search was updated in July 2008. Four new trials
were identified and included in the review (Le Saux
2005, Spiro 2006, Neumark 2007 and McCormick
2005). One of these trials (Le Saux 2005) compared
antibiotics with placebo. For the outcome pain at 24
hours and 2 to 7 days, inclusion of this trial did not
alter the overall conclusions of the primary analy-
sis. The three other new trials (Spiro 2006, Neumark
2007, McCormick 2005) compared immediate antibi-
otics with various observational approaches. One of
the new trials compared immediate antibiotics with
delayed prescribing (Spiro 2006). The other trials
(McCormick 2005 and Neumark 2007) compared im-
mediate antibiotics with watchful waiting, in which
no prescription was supplied but advise on when to
seek treatment was provided. Outcome data on pain
at 3 to 7 days from these trials were analysed with data
from another trial of immediate versus delayed pre-
scription (Little 2001). In earlier versions of the re-
view data from the Little (Little 2001) trial had been
included in a sensitivity analysis. In this update, data
from the four trials comparing immediate versus ob-
servational management strategies have been included
in the main analysis. Information on subgroups of chil-
dren who are most likely to benefit from treatment
with antibiotics, obtained from a meta-analysis of in-
dividual patient data has been included in this review
(Rovers 2006). Methods of the IPD meta-analysis,
conducted by two authors on this review (and others)
are also included. Survival curves from the IPD meta-
analysis showing the pattern of recovery from acute
otitis media over time has been included as an extra
figure. Two ongoing trials comparing antibiotics with
placebo in children < 35 months have been identified
18 February 2005 Feedback has been incorporated Feedback and reply added.
20 November 2000 Feedback has been incorporated Feedback comments and replies added.
3 February 2000 New citation required and conclusions have changed Conclusions changed.
CONTRIBUTIONS OF AUTHORS
Chris Del Mar (CDM) and Paul Glasziou (PG) prepared the original version of the review.
Sharon Sanders (SLS) conducted searches, identified studies, extracted data and prepared manuscript for the updated reviews in 2003,
2007 and 2008.
Maroeska Rovers (MMR) participated in the 2007 update by providing data and information from the individual patient data meta-
analysis that has been included in this update.
Roderick Venekamp (RPV) conducted searches, identified studies, extracted data and prepared manuscript for the updated review in
2012.
PG, CDM, MMR, SLS and RPV have reviewed and provided comment on the updated version of the review.
DECLARATIONS OF INTEREST
None noted.
INDEX TERMS
Medical Subject Headings (MeSH)
Acute Disease; Age Factors; Anti-Bacterial Agents [adverse effects; therapeutic use]; Earache [drug therapy]; Otitis Media [ drug
therapy; prevention & control]; Randomized Controlled Trials as Topic; Secondary Prevention; Tympanic Membrane Perforation [drug
therapy]