1. health and fitness
2. disease

Otolaryngology–Head and Neck Surgery (2006) 134, S1-S3
COMMENTARY
Commentary: Guidelines and otolaryngology
A Milestone, But Only a Beginning
This supplement represents a milestone in otolaryngology– head and neck surgery, the publication of the first independently produced, evidence-based, clinical practice
guideline by the American Academy of Otolaryngology–
Head and Neck Surgery (AAO-HNS). As such, a few words
of introduction are in order with respect to the importance of
guidelines to clinicians and the specialty.
Guidelines create clout for an organization by showcasing their ability to publish valid documents of high methodological rigor that facilitate evidence-based decisions. In
this respect, the American Academies of Pediatrics, Ophthalmology, and Orthopaedic Surgeons have published 85,
16, and 7 guidelines, respectively, and the American Colleges of Chest Physicians, Obstetricians and Gynecologists,
and Gastroenterology have published 48, 45, and 26.1 An
organization’s ability to influence patient health and public
policy–and to preserve its autonomy– depends largely on the
scope, quality, and accessibility of its guideline repertoire.
Compared with other subject areas, there is paucity of
guidelines in otolaryngology. Of 1559 guidelines indexed
by the National Guideline Clearinghouse (NGC)1 only 40
relate to “otorhinolaryngologic diseases,” and only 1, otitis
media with effusion,2 was sponsored, in part, by the AAO–
HNS. In contrast, over 200 guidelines exist in each category
for diseases of the cardiovascular system, digestive system,
immune system, nervous system, and respiratory tract.
Within the 23 disease/condition categories outlined by the
NGC, only 2 had fewer guidelines than otolaryngology:
parasitic diseases with 17 and animal diseases with 3.
Why Guidelines Are Important
“There are only two sorts of doctors; those who practice
with their brains, and those who practice with their
tongues,” quipped Osler more than a century ago.3 Put
simply, the purpose of guidelines is to encourage more of
the former. Regardless of your expertise, experience, or
confidence in managing acute otitis externa, you will undoubtedly gain added perspective by perusing the guideline
and accompanying meta-analysis. Guidelines are a partial
antidote to Franklin’s adage “Beware the young Doctor &
the old Barber.”4 Whereas there is no substitute for enlight-
ened experience, the intellectual energy that summarizes,
evaluates, and documents scientific knowledge from a wellcrafted guideline will certainly expedite the process.
For science to be cumulative, an intermediate step is
needed to bridge past and future research: synthesis of
existing evidence.5 Guidelines are the pinnacle of evidence
synthesis for defining best practices and guiding health care
policy. Traditional review articles and consensus statements, however, are subjective, inefficient, and often biased
in ways that are not apparent to the reader. Adding levels of
evidence yields the so-called “evidence-based review,”
which is still prone to bias and distortion. Conversely, a
systematic review (meta-analysis) identifies, appraises, and
quantitatively combines studies with special protocols to
reduce bias and quantify outcomes.6 Guidelines begin
where a systematic review ends by considering the role of
values, diversity, patient preference, and risk-harm assessment in the decision-making process, and by using expert
experience to fill gaps in the knowledge base.
Evidence-based guidelines are the springboard for performance measures, a relatively new construct that is being
applied increasingly to determine whether consistent highquality medical care is provided across health care systems.7
Performance measures have far reaching implications because of their roles in maintaining specialty certification and
determining health care reimbursement. An increasing number of health plans and purchasers have adopted pay-forperformance programs, which have questionable validity
unless credible practice guidelines exist to define quality
clinical care.8 Finally, as maintenance of certification becomes more ingrained in the health care paradigm,9 increasing reliance will be placed on guidelines as a benchmark for
quality care.
It’s All in the Methodology
One goal in developing this guideline was to establish
methodology for future AAO–HNSF efforts. To ensure the
validity of our guideline, we adhered to reporting standards
from the Conference on Guideline Standardization
(COGS).10 Scientific evidence was linked to recommendation strength with the use of a paradigm developed by the
American Academy of Pediatrics,11 based on evidence qual-
0194-5998/$32.00 © 2006 American Academy of Otolaryngology–Head and Neck Surgery Foundation, Inc. All rights reserved.
doi:10.1016/j.otohns.2006.02.015
S2
Otolaryngology–Head and Neck Surgery, Vol 134, No 4S, April 2006
ity and the anticipated balance between benefits and harms.
A unique aspect of this classification is acknowledging
exceptional situations where validating studies cannot be
performed, but where there is a clear preponderance of
benefit or harm. We improved on this process by including
an explicit summary of quality, benefits, harm, cost, value
judgments, and the role of patient preferences after the
supporting text for each major policy statement (guideline
recommendation).
Guideline development is a process of consensus and
compromise driven by evidence. Once the scope of the
guideline is determined and clinical questions are defined, a
systematic review is performed to illuminate the existing
evidence base. This review is a springboard for developing
a roster of conceptual issues deemed of practical importance
to clinicians. Consensus and compromise narrow the roster
to a workable agenda of about 7 to 10 key issues, with
additional targeted literature searches as needed. Meta-analyses, if not already available, are conducted to obtain unbiased quantitative estimates of treatment effects to guide
decisions.
During the 7 months devoted to the development of the
Acute Otitis Externa Guideline (ending in November 2005),
our working group met twice with interval electronic review
and feedback on each guideline draft to ensure accuracy of
content and consistency with COGS criteria.10 In September
and November 2005, the Yale Center for Medical Informatics used specialized software12 to appraise adherence of the
draft guideline to methodologic standards, to improve clarity of recommendations, and to predict potential obstacles to
implementation. These steps helped to further maximize the
validity and utility of the resulting guideline.
Some Guideline Caveats
Guidelines are never intended to supersede professional
judgment; rather, they may be viewed as a relative constraint on individual clinician discretion in a particular clinical circumstance. Clinicians should always act and decide
in a way that they believe will best serve their patients’
interests and needs, regardless of guideline recommendations. Guidelines represent the best judgment of a team of
experienced clinicians and methodologists addressing the
scientific evidence for a particular topic.11 They are not
intended to replace clinical judgment or to establish a protocol for all individuals, and may not provide the only
appropriate approach to diagnosis and management.
One pitfall in guideline creation is that a slavish devotion
to evidence can yield a document that lacks practicality and
common sense. Osler3 further observed that “common sense
in medical matters is rare, and is usually in inverse ratio to
the degree of education.” Whereas all guideline authors
possessed a modicum of education, they were also a motley
crew with extensive hands-on experience in diverse practice
settings. Our group represented the fields of otolaryngology– head and neck surgery, pediatrics, family medicine,
infectious disease, internal medicine, emergency medicine,
and medical informatics. The draft guideline also underwent
extensive peer review by leading clinicians and academicians before publication.
Creating and publishing a guideline entails significant
expense, and our expenses were defrayed, in part, by an
unrestricted grant from Alcon Laboratories. Because our
sponsor markets products for acute otitis externa and some
authors had significant contact with the sponsor, the possibility of a conflict of interest exists. To minimize the risk of
conflict we explicitly followed current recommendations13
to 1) disclose relationships before guideline meetings, 2)
decide as a group if any relationships precluded participation of panel members, and 3) include a paragraph of disclosure in the printed guideline. Almost 90% of guideline
authors have some relationship with industry, which does
not prelude participation unless it is an equity holding.13
Preparing for the Future
The cost, complexity, and regional variations of modern
health care have created a pressing need for valid guidelines
to determine health policy. Unless otolaryngologists assume
a proactive role in filling this need, other disciplines will
(continue to) do it for us. As this supplement demonstrates,
the AAO–HNS can produce a guideline that exceeds current
methodologic standards. A challenge for the future is to
sustain momentum by prioritizing needs and developing a
pipeline of ongoing projects. Guidelines, however, do not
arise from divine providence or from simple review and
consensus; commitment to the explicit, evidence-based process briefly described in this commentary is the starting
point for success.
Aristotle once wrote, “By common consent the beginning is more than half the whole task, and throws a flood of
light on many of the aspects of the inquiry.”14 We have a
strong beginning for otolaryngology and guidelines, and the
methodological path is well lit. On behalf of the talented
individuals who helped create this guideline, I invite you to
profit from the insights it provides and from future efforts
that will inevitably follow.
I thank Richard N. Shiffman for his critical review of the
manuscript and helpful suggestions.
Richard M. Rosenfeld, MD, MPH
Brooklyn, NY
January 2006
Dr. Rosenfeld is a prior consultant for Alcon and Daiichi Pharmaceuticals.
REFERENCES
1. National Guideline Clearinghouse. NGC Browse: disease/condition
results. http://www.guideline.gov (accessed 1/2/06).
2. Rosenfeld RM, Culpepper L, Doyle KJ, et al. Clinical practice guideline: otitis media with effusion. Otolaryngol Head Neck Surg 2004;
130:S95–S118.
3. Osler W. Teaching and thinking: the two functions of a medical
school. In: Aequanimitas. Philadelphia: P. Blakiston’s Son & Co;
1904:131.
Rosenfeld
4. Franklin B. Poor Richard, 1733. http://usinfo.state.gov/usa/infousa/
facts/loa/bf1733.htm (accessed 1/5/06).
5. Light RJ, Pillemer DB. Summing up: the science of reviewing
research. Cambridge: Harvard University Press; 1984: p. 3–5.
6. Rosenfeld RM. Meta-analysis. ORL 2004;66:186 –95.
7. Walter LC, Davidowitz NP, Heineken PA, et al. Pitfalls of converting
practice guidelines into quality measures: lessons learned from a VA
performance measure. JAMA 2004;291:2466 –70.
8. Rosenthal MB, Frank RG, Zhonghe L, et al. Early experience with
pay-for-performance: from concept to practice. JAMA 2005;294:
1788 –93.
9. Batmangelich S, Adamowski S. Maintenance of certification in the
United States: a progress report. J Cont Educ Health Prof 2004;24:
134 – 8.
S3
10. Shiffman RN, Shekelle P, Overhage JM, et al. Standardized reporting of clinical practice guidelines: a proposal from the Conference
on Guideline Standardization. Ann Intern Med 2003;139:493– 8.
11. American Academy of Pediatrics Steering Committee on Quality Improvement and Management. Classifying recommendations for clinical practice guidelines. Pediatrics 2004;14:874 –7.
12. Shiffman RN, Karras BT, Agrawal A, et al. GEM: a proposal for a
more comprehensive guideline document model using XML. J Am
Med Informatics Assoc 2000;7:488 –98.
13. Choudry NK, Stelfox HT, Detsky AS. Relationships between authors
of clinical practice guidelines and the pharmaceutical industry. JAMA
2002;287:612–7.
14. Thomson JAK (translator). The ethics of Aristotle. London: Penguin
Books Ltd; 1987: p. 77.
Otolaryngology–Head and Neck Surgery (2006) 134, S4-S23
ORIGINAL RESEARCH
Clinical practice guideline: Acute otitis externa
Richard M. Rosenfeld, MD, MPH, Lance Brown, MD, MPH,
C. Ron Cannon, MD, Rowena J. Dolor, MD, MHS,
Theodore G. Ganiats, MD, Maureen Hannley, PhD,
Phillip Kokemueller, MS, CAE, S. Michael Marcy, MD, Peter S. Roland, MD,
Richard N. Shiffman, MD, MCIS, Sandra S. Stinnett, DrPH
and David L. Witsell, MD, MHS, Brooklyn, New York; Loma Linda, California;
Jackson, Mississippi; Durham, North Carolina; San Diego, California; Dallas, Texas;
New Haven, Connecticut; and Alexandria, Virginia
OBJECTIVE: This guideline provides evidence-based recommendations to manage diffuse acute otitis externa (AOE), defined
as generalized inflammation of the external ear canal, which may
also involve the pinna or tympanic membrane. The primary purpose is to promote appropriate use of oral and topical antimicrobials and to highlight the need for adequate pain relief.
STUDY DESIGN: In creating this guideline, the American
Academy of Otolaryngology–Head and Neck Surgery Foundation
(AAO-HNSF) selected a development group representing the
fields of otolaryngology– head and neck surgery, pediatrics, family
medicine, infectious disease, internal medicine, emergency medicine, and medical informatics. The guideline was created with the
use of an explicit, a priori, evidence-based protocol.
RESULTS: The group made a strong recommendation that
management of AOE should include an assessment of pain, and the
clinician should recommend analgesic treatment based on the
severity of pain. The group made recommendations that clinicians
should: 1) distinguish diffuse AOE from other causes of otalgia,
otorrhea, and inflammation of the ear canal; 2) assess the patient
with diffuse AOE for factors that modify management (nonintact
tympanic membrane, tympanostomy tube, diabetes, immunocompromised state, prior radiotherapy); and 3) use topical preparations
for initial therapy of diffuse, uncomplicated AOE; systemic antimicrobial therapy should not be used unless there is extension
outside of the ear canal or the presence of specific host factors that
would indicate a need for systemic therapy.
The group made additional recommendations that: 4) the choice of
topical antimicrobial therapy of diffuse AOE should be based on
efficacy, low incidence of adverse events, likelihood of adherence
to therapy, and cost; 5) clinicians should inform patients how to
administer topical drops, and when the ear canal is obstructed,
From the Department of Otolaryngology, SUNY Downstate Medical
Center and Long Island College Hospital (RMR); the Departments of
Emergency Medicine and Pediatrics, Loma Linda University Medical Center (LB); the Departments of Otolaryngology and Family Medicine, University of Mississippi School of Medicine (CRC); the Department of
Diagnostic Science, University of Mississippi School of Dentistry (CRC);
the Division of Internal Medicine, Duke University Medical Center (RJD);
the Department of Family and Preventive Medicine, University of California San Diego (TGG); the Center for Vaccine Research, University of
California Los Angeles (SMM); the Department of Otolaryngology, University of Texas Southwestern School of Medicine (PSR); the Center for
Medical Informatics, Yale University School of Medicine (RNS); the
Department of Biostatistics and Bioinformatics, Duke University Medical
Center (SSS); the Division of Otolaryngology, Duke University Medical
Center (DW); and the American Academy of Otolaryngology–Head and
Neck Surgery Foundation (MH, PK).
Conflict of Interest Disclosure: Alcon Laboratories provided an un-
restricted educational grant to the American Academy of Otolaryngology–
Head and Neck Surgery Foundation to create an acute otitis externa (AOE)
performance measure and clinical practice guideline. The sponsor had no
involvement in any aspect of developing the guideline and was unaware of
content until publication. Individual disclosures for group members are:
RM Rosenfeld, past consultant to Alcon Laboratories and Daiichi Pharmaceuticals; and PS Roland, speaking honoraria, departmental consulting
fees for research support from Alcon Laboratories and Daiichi Pharmaceuticals. SM Marcy is a consultant for Medimmune, Merck, SanofiPasteur, and GlaxoSmithKline. No other panel members had disclosures.
Disclosures were made available to the Guideline Development Group for
open discussion, with the conclusion that none of the relationships would
preclude participation.
Reprint requests: Richard M. Rosenfeld, MD, MPH, Department of
Otolaryngology, 339 Hicks Street, Brooklyn, NY 11201-5514.
E-mail address: [email protected].
0194-5998/$32.00 © 2006 American Academy of Otolaryngology–Head and Neck Surgery Foundation, Inc. All rights reserved.
doi:10.1016/j.otohns.2006.02.014
Rosenfeld et al
Clinical Practice Guideline: Acute Otitis Externa
delivery of topical preparations should be enhanced by aural toilet,
placing a wick, or both; 6) when the patient has a tympanostomy
tube or known perforation of the tympanic membrane, the clinician
should prescribe a nonototoxic topical preparation; and 7) if the
patient fails to respond to the initial therapeutic option within 48 to
72 hours, the clinician should reassess the patient to confirm the
diagnosis of diffuse AOE and to exclude other causes of illness.
And finally, the panel compiled a list of research needs based on
limitations of the evidence reviewed.
CONCLUSION: This clinical practice guideline is not intended as a sole source of guidance in evaluating patients with
AOE. Rather, it is designed to assist clinicians by providing an
evidence-based framework for decision-making strategies. It is
not intended to replace clinical judgment or establish a protocol
for all individuals with this condition and may not provide the
only appropriate approach to the diagnosis and management of
this problem.
SIGNIFICANCE: This is the first, explicit, evidence-based clinical practice guideline on acute otitis externa, and the first clinical
practice guideline produced independently by the AAO–HNSF.
© 2006 American Academy of Otolaryngology–Head and Neck
Surgery Foundation, Inc. All rights reserved.
T
he primary purpose of this guideline is to promote
appropriate use of oral and topical antimicrobials
for diffuse acute otitis externa (AOE) and to highlight the
need for adequate pain relief. The target patient is aged 2
years or older with diffuse AOE, defined as generalized
inflammation of the external ear canal, with or without
involvement of the pinna or tympanic membrane. As the
first clinical practice guideline developed independently
by the AAO-HNSF, a secondary purpose was to refine
methods for future efforts. Additional goals were to make
possible an AOE performance measure and to make clinicians
aware of modifying factors that can or may alter management
(eg, diabetes, immunocompromised state, prior radiotherapy,
tympanostomy tube, nonintact tympanic membrane).
This guideline does not apply to children under age 2
years or to patients of any age with chronic or malignant
(progressive necrotizing) otitis externa. AOE is uncommon
before age 2 years, and very limited evidence exists with
respect to treatment or outcomes in this age group. Although
the differential diagnosis of the “draining ear” will be discussed, recommendations for management will be limited to
diffuse AOE, which is almost exclusively a bacterial infection. The following conditions will be briefly discussed but
not considered in detail: furunculosis (localized AOE), otomycosis, herpes zoster oticus (Ramsay Hunt syndrome), and
contact dermatitis.
The guideline is intended for primary care and specialist
clinicians, including otolaryngologists– head and neck surgeons, pediatricians, family physicians, emergency physicians, internists, nurse-practitioners, and physician assistants. The guideline is applicable to any setting in which
children, adolescents, or adults with diffuse AOE would be
identified, monitored, or managed.
S5
INTRODUCTION
A
cute otitis externa (AOE) as discussed in this guideline
is defined as diffuse inflammation of the external ear
canal, which may also involve the pinna or tympanic membrane. A diagnosis of diffuse AOE requires rapid onset
(generally within 48 hours) in the past 3 weeks of symptoms
and signs of ear canal inflammation as detailed in Table 1.
A hallmark sign of diffuse AOE is tenderness of the tragus,
pinna, or both, that is often intense and disproportionate to
what might be expected based on visual inspection.
Also known as “swimmer’s ear” or “tropical ear,” AOE
is one of the most common infections encountered by clinicians. The annual incidence of AOE is between 1:100 and
1:250 of the general population,1,2 with regional variations
based on age and geography; lifetime incidence is up to
10%.3 The direct cost of AOE is unknown, but the ototopical market in the United States is approximately 7.5 million
annual prescriptions with total sales of $310 million (IMS/
Verispan 2004, personal communication). Additional medical costs include physician visits and prescriptions for analgesics and systemic medications, such as antibiotics,
steroids, or both. The indirect costs of AOE have not been
calculated but are likely to be substantial because of severe
and persistent otalgia that limits activities.
AOE is a cellulitis of the ear canal skin and subdermis,
with acute inflammation and variable edema. Nearly all
(98%) AOE in North America is bacterial. The most common pathogens are Pseudomonas aeruginosa (20% to 60%
prevalence) and Staphylococcus aureus (10% to 70% prevalence), often occurring as a polymicrobial infection. Other
pathogens are principally gram negative organisms (other
than P aeruginosa), which cause no more than 2% to 3% of
cases in large clinical series.5-12 Fungal involvement is
distinctly uncommon in primary AOE but may be more
Table 1
Elements of the diagnosis of diffuse acute
otitis externa
1. Rapid onset (generally within 48 hours) in the
past 3 weeks, AND
2. Symptoms of ear canal inflammation that
include:
● otalgia (often severe), itching, or fullness,
● WITH OR WITHOUT hearing loss or jaw pain,*
AND
3. Signs of ear canal inflammation that include:
● tenderness of the tragus, pinna, or both
● OR diffuse ear canal edema, erythema, or
both
● WITH OR WITHOUT otorrhea, regional
lymphadenitis, tympanic membrane erythema,
or cellulitis of the pinna and adjacent skin
*Pain in the ear canal and temporomandibular joint region
intensified by jaw motion.
S6
Otolaryngology–Head and Neck Surgery, Vol 134, No 4S, April 2006
common in chronic otitis externa or after treatment of AOE
with topical, or less often systemic, antibiotics.13
Topical antimicrobials are beneficial for AOE, but oral
antibiotics have limited utility.14 Nonetheless, about 20% to
40% of patients with AOE receive oral antibiotics, often in
addition to topical therapy.2,15,16 The oral antibiotics selected are usually inactive against P aeruginosa and S
aureus, may have undesirable side effects, and, because
they are widely distributed, serve to select out resistant
organisms throughout the body.17,18 Bacterial resistance is
of far less concern with topical antimicrobials, because the
high local concentration of drug in the ear canal will generally eradicate all susceptible organisms plus those with
marginal resistance.4
The cause of AOE is multifactorial. Regular cleaning of
the ear canal removes cerumen, which is an important barrier to moisture and infection.19 Cerumen creates a slightly
acidic pH that inhibits infection (especially by P aeruginosa)
but can be altered by water exposure, aggressive cleaning,
soapy deposits, or alkaline eardrops.20,21 Debris from dermatologic conditions may also encourage infections,6,22 as
can local trauma from attempts at self-cleaning, irrigation,23
and wearing hearing aids.24,25 Other factors such as sweating, allergy, and stress have also been implicated in the
pathogenesis of AOE.26
AOE is more common in regions with warmer climates,
increased humidity, or increased water exposure from
swimming.27,28 Most, but not all, studies have found an
association with water quality (in terms of bacterial load)
and the risk of AOE. The causative organisms are present in
most swimming pools and hot tubs; however, even those
that comply with water quality standards may still contain
AOE pathogens.29-32 Some individuals appear more susceptible to AOE on a genetic basis (those with type A blood
group) and the subspecies of Pseudomonas causing AOE
may be different from those causing other Pseudomonas
infections.33,34
Strategies to prevent AOE are aimed at limiting water
accumulation and moisture retention in the external auditory
canal, and maintaining a healthy skin barrier. No randomized trials have compared the efficacy of different strategies
to prevent AOE. Available reports include case series and
expert opinion that emphasize the prevention of moisture
and water retention in the external auditory canal. Recommendations to prevent AOE include removing obstructing
cerumen; the use of acidifying ear drops shortly before
swimming, after swimming, at bedtime, or all 3 times; the
use of a hair dryer to dry the ear canal; the use of ear plugs
while swimming; and the avoidance of trauma to the external auditory canal.35-38
Variations in managing AOE and the importance of
accurate diagnosis suggest a need for an evidence-based
practice guideline. Failure to distinguish AOE from other
causes of “the draining ear” (eg, chronic external otitis,
malignant otitis externa, middle ear disease) may prolong
morbidity or cause serious complications. Because topical
Table 2
Interventions considered in AOE guideline
development
Diagnosis
Treatment
Prevention
History and physical examination
Otoscopy
Pneumatic otoscopy
Otomicroscopy
Tympanometry
Acoustic reflectometry
Culture
Imaging studies
Audiometry (excluded from guideline)
Aural toilet (suction, dry mopping,
irrigation, removal of obstructing
cerumen or foreign object)
Nonantibiotic (antiseptic or acidifying)
drops
Antibiotic drops
Steroid drops
Oral antibiotics
Analgesics
Complementary and alternative
medicine
Ear canal wick
Biopsy (excluded from guideline)
Surgery (excluded from guideline)
Water precautions
Prophylactic drops
Environmental control (eg, hot tubs)
Avoiding neomycin drops (if allergic)
Addressing allergy to ear molds or
water protector
Addressing underlying dermatitis
Specific preventive measures for
diabetics or immunocompromised
state
therapy is efficacious, systemic antibiotics are often prescribed inappropriately.14,39 When topical therapy is prescribed confusion exists about whether to use an antiseptic,
antibiotic, corticosteroid, or a combination product. Selecting an antibiotic creates additional controversy, particularly
with respect to the role of newer quinolone drops. And
finally, the optimal methods for cleaning the ear canal (aural
toilet) and drug delivery are undefined.
The primary outcome considered in this guideline is
clinical resolution of AOE. Additional outcomes considered
include minimizing the use of ineffective treatments; eradicating pathogens; minimizing recurrence, cost, complications, and adverse events; maximizing the health-related
quality of life of individuals afflicted with AOE; increasing
patient satisfaction40; and permitting the continued use of
necessary hearing aids. The relatively high incidence of
AOE and the diversity of interventions in practice (Table 2)
make AOE an important condition for the use of an up-todate, evidence-based practice guideline.
Rosenfeld et al
Clinical Practice Guideline: Acute Otitis Externa
METHODS
General Methods and Literature Search
To develop an evidence-based clinical practice guideline on
managing AOE, the American Academy of Otolaryngology–Head and Neck Surgery Foundation (AAO-HNSF) assembled a working group who represented the disciplines of
otolaryngology– head and neck surgery, pediatrics, infectious disease, internal medicine, family medicine, emergency medicine, biostatistics, and medical informatics. Several group members had significant prior experience in the
development of clinical practice guidelines.
A MEDLINE search from 1966 through July 2005 was
performed on PubMed with the terms “otitis externa” (MeSH
term) and “swimmer’s ear.” Titles and abstracts unrelated to
AOE were excluded, leaving 240 articles that were collated
under these headings: risk factors (n ⫽ 30), microbiology
(n ⫽ 24), pharmacologic intervention (n ⫽ 118), other
interventions (n ⫽ 17), epidemiology and practice patterns
(n ⫽ 14), potential harms (n ⫽ 30), and otomycosis (n ⫽ 9).
Citations and abstracts were distributed to all group members to assist in formulating and prioritizing evidence-based
statements. Members performed additional targeted MEDLINE searches through September 2005 to supplement the
initial broad search.
A meta-analysis was performed with an a priori protocol41 and a published search strategy for AOE42 to compare
the following topical treatments: antimicrobial vs placebo,
antiseptic vs antimicrobial, quinolone antibiotic vs nonquinolone, steroid-antimicrobial vs antimicrobial, or antimicrobialsteroid vs steroid. Search of MEDLINE from 1966 through
July 2005, without language restrictions, identified 2860 articles, of which 509 were potential randomized trials.43,44 Review of these studies, plus 7 others found in the Cochrane
Database, yielded 43 articles that were assessed by 2 reviewers independently for relevance, study quality,45 and
data extraction. The final data set included 20 articles that
had random allocation, were limited to diffuse AOE (or had
subgroup data), gave results by subjects (or, if results by
ears, 90% or higher must be unilateral disease), and had 2 or
more parallel groups related to the above comparisons.
In a series of conference calls, the working group defined
the scope and objectives of the proposed guideline. During
the 7 months devoted to guideline development ending in
November 2005, the group met twice with interval electronic review and feedback on each guideline draft to ensure
accuracy of content and consistency with standardized criteria for reporting clinical practice guidelines.46 At the first
meeting, all members disclosed relationships with pharmaceutical manufacturers and discussed what impact, if any,
potential conflicts of interest might have on guideline validity.47 The group concluded that none of the relationships
precluded involvement.
In September and November 2005, the Guidelines Review Group of the Yale Center for Medical Informatics used
GEM-COGS,48 the Guideline Implementability Appraisal
(GLIA) and Extractor software, to appraise adherence of the
S7
draft guideline to methodologic standards, to improve clarity of recommendations, and to predict potential obstacles to
implementation. AOE guideline development group members received summary appraisals before their second meeting and modified an advanced draft of the guideline.
The final draft practice guideline underwent extensive
external peer review. Comments were compiled and reviewed by the group chairperson. The recommendations
contained in the practice guideline are based on the best
available published data through September 2005. Where
data are lacking, a combination of clinical experience and
expert consensus was used. A scheduled review process will
occur at 5 years from publication or sooner if new compelling evidence warrants earlier consideration.
Classification of Evidence-Based Statements
Guidelines are intended to reduce inappropriate variations
in clinical care, to produce optimal health outcomes for
patients, and to minimize harm. The evidence-based approach to guideline development requires that the evidence
that supports a policy be identified, appraised, and summarized and that an explicit link between evidence and statements be defined. Evidence-based statements reflect both
the quality of evidence and the balance of benefit and harm
that is anticipated when the statement is followed. The
definitions for evidence-based statements49 are listed in
Tables 3 and 4.
Guidelines are never intended to supersede professional judgment; rather, they may be viewed as a relative
constraint on individual clinician discretion in a particular clinical circumstance. Less frequent variation in practice is expected for a strong recommendation than might
be expected with a recommendation. Options offer the
most opportunity for practice variability.50 Clinicians
should always act and decide in a way that they believe
will best serve their patients’ interests and needs, regardless of guideline recommendations. Guidelines represent
the best judgment of a team of experienced clinicians and
methodologists addressing the scientific evidence for a
particular topic.49
Making recommendations about health practices involves value judgments on the desirability of various outcomes associated with management options. Values applied
by the AOE guideline development group sought to minimize harm, diminish unnecessary and inappropriate therapy,
and reduce the unnecessary use of systemic antibiotics. A
major goal of the working group was to be transparent and
explicit about how values were applied and to document the
process.
AOE GUIDELINE EVIDENCE-BASED
STATEMENTS
Each evidence-based statement is organized in a similar
fashion: evidence-based statement in bold face type, fol-
S8
Otolaryngology–Head and Neck Surgery, Vol 134, No 4S, April 2006
Table 3
Guideline definitions for evidence-based statements
Statement
Definition
Implication
Strong recommendation
A strong recommendation means the benefits of the
recommended approach clearly exceed the harms
(or that the harms clearly exceed the benefits in
the case of a strong negative recommendation)
and that the quality of the supporting evidence is
excellent (Grade A or B).* In some clearly
identified circumstances, strong recommendations
may be made based on lesser evidence when
high-quality evidence is impossible to obtain and
the anticipated benefits strongly outweigh the
harms.
A recommendation means the benefits exceed the
harms (or that the harms exceed the benefits in
the case of a negative recommendation), but the
quality of evidence is not as strong (Grade B or
C).* In some clearly identified circumstances,
recommendations may be made based on lesser
evidence when high-quality evidence is
impossible to obtain and the anticipated benefits
outweigh the harms.
An option means that either the quality of evidence
that exists is suspect (Grade D)* or that well-done
studies (Grade A, B, or C)* show little clear
advantage to one approach versus another.
Clinicians should follow a strong
recommendation unless a
clear and compelling rationale
for an alternative approach is
present.
Recommendation
Option
No recommendation
No recommendation means there is both a lack of
pertinent evidence (Grade D)* and an unclear
balance between benefits and harms.
Clinicians should also generally
follow a recommendation but
should remain alert to new
information and sensitive to
patient preferences.
Clinicians should be flexible in
their decision making
regarding appropriate practice,
although they may set bounds
on alternatives; patient
preference should have a
substantial influencing role.
Clinicians should feel little
constraint in their decision
making and be alert to new
published evidence that
clarifies the balance of benefit
versus harm; patient
preference should have a
substantial influencing role.
*See Table 4 for definition of evidence grades.
lowed by an italicized statement on the strength of the
recommendation. Several paragraphs then discuss the
evidence base that supports the statement; they conclude
with an “evidence profile” of aggregate evidence quality,
benefit-harm assessment, and statement of costs. Finally,
there is an explicit statement of the value judgments, the
role of patient preferences, and a repeat statement of the
strength of the recommendation.
1a. DIFFERENTIAL DIAGNOSIS: Clinicians
should distinguish diffuse AOE from other
causes of otalgia, otorrhea, and
inflammation of the external ear canal.
Recommendation based on observational studies with a
preponderance of benefit over risk.
1b. MODIFYING FACTORS: Clinicians should
assess the patient with diffuse AOE for
factors that modify management (nonintact
tympanic membrane, tympanostomy tube,
diabetes, immunocompromised state, prior
radiotherapy). Recommendation based on
observational studies with a preponderance of benefit
over risk.
Differential Diagnosis. A diagnosis of diffuse AOE requires
rapid onset with signs and symptoms of ear canal inflammation (Table 1). Symptoms of AOE include otalgia (70%),
itching (60%), or fullness (22%), with or without hearing
loss (32%) or ear canal pain on chewing. A hallmark sign of
diffuse AOE is tenderness of the tragus (when pushed),
pinna (when pulled up and back), or both. The tenderness is
Rosenfeld et al
Clinical Practice Guideline: Acute Otitis Externa
Table 4
Evidence quality for grades of evidence
Grade
Evidence quality
A
Well-designed randomized controlled trials
or diagnostic studies performed on a
population similar to the guideline’s
target population
Randomized controlled trials or diagnostic
studies with minor limitations;
overwhelmingly consistent evidence
from observational studies
Observational studies (case control and
cohort design)
Expert opinion, case reports, reasoning
from first principles (bench research or
animal studies)
Exceptional situations where validating
studies cannot be performed and there
is a clear preponderance of benefit over
harm
B
C
D
X
often intense and disproportionate to what might be expected based on visual inspection. Otoscopy will reveal
diffuse ear canal edema, erythema, or both, either with or
without otorrhea or material in the ear canal. Regional
lymphadenitis or cellulitis of the pinna and adjacent skin
may be present in some patients.6,51
AOE can mimic the appearance of acute otitis media
(AOM) because of erythema that involve the tympanic
membrane. Distinguishing AOE from AOM is important
because the latter may require systemic antimicrobials.52
Pneumatic otoscopy will demonstrate good tympanic membrane mobility with AOE but will show absent or limited
mobility with AOM and associated middle-ear effusion.
Similarly, tympanometry will show a normal peaked curve
(type A) with AOE, but a flat tracing (type B) with AOM.
The validity of acoustic reflectometry with AOE is
unknown.
Anything that disrupts the epithelium of the ear canal can
permit invasion by bacteria that cause diffuse AOE. Common predisposing factors for AOE26 are humidity or prolonged exposure to water, dermatologic conditions (eczema,
seborrhea, psoriasis), anatomic abnormalities (narrow canal,
exostoses), trauma or external devices (wax removal, insertion of earplugs, use of hearing aids), and otorrhea caused
by middle-ear disease. AOE may also occur as a result of
ear canal obstruction by impacted cerumen, a foreign object,
or a dermoid or sebaceous cyst. Clinical history should
identify predisposing factors and assess swimming behavior. Other causes of otalgia, otorrhea, and inflammation
should be distinguished from diffuse AOE because management will differ.
Furunculosis is the presence of an infected hair follicle
on the outer third of the ear canal, sometimes referred to as
localized otitis externa. Clinical findings include otalgia,
S9
otorrhea, and localized tenderness. Treatment may include
local heat, incision and drainage, or systemic antibiotics that
cover S aureus, the most common causative agent.
Eczema, seborrhea, and other inflammatory dermatoses that involve the ear canal and surrounding tissues are
relatively common and may predispose to acute infection.
In contrast, contact dermatitis of the ear canal is an
allergic reaction to antigens such as metals (nickel, silver), chemicals (cosmetics, soaps, detergents, shampoo,
hairspray), plastics, rubber, leather, or drugs. Nickel is
the most common contact allergen, affecting around 10%
of women with pierced ears.53-55 Contact allergy also
occurs in some patients who wear hearing aids as a
reaction to the plastics and other chemicals used in hearing aid molds.56,57
Sensitization to topical treatment (secondary contact otitis) can result from prolonged or repeated use of topical
antimicrobials. Many otic preparations (antibiotics and vehicle substances) have been reported to cause sensitization.
Neomycin is the most common substance and causes reactions in about 5% to 15% of patients with chronic external
otitis.58 Patch testing has demonstrated that 13% of normal
volunteers are hypersensitive to neomycin.59 A maculopapular eruption on the conchal bowl and in the ear canal is
consistent with an allergic reaction to a topical agent; an
erythematous streak may extend down the pinna where
drops contact the auricular skin. Management involves removal of the sensitizing agent and application of a topical
steroid.
Viral infections of the external ear, caused by varicella,
measles, or herpesvirus, are rare. Herpes zoster oticus
(Ramsay Hunt syndrome) causes vesicles on the external
ear canal and posterior surface of the auricle, severe otalgia,
facial paralysis or paresis, loss of taste on the anterior
two-thirds of the tongue, and decreased lacrimation on the
involved side.60 Management involves antiviral therapy,
with or without systemic steroid.
Modifying Factors. Key components of the clinical history that can modify management of diffuse AOE include
1) diabetes, HIV infection, or other immunocompromised
states, 2) a history of radiotherapy, and 3) the presence of
tympanostomy tubes or nonintact tympanic membrane.
Malignant (progressive necrotizing) otitis externa is an
aggressive infection that predominantly affects elderly, diabetic, or immunocompromised patients.61 P aeruginosa is
isolated from exudate in the ear canal in more than 90% of
cases. Initial signs and symptoms are those of the initiating
AOE, but untreated disease develops into a skull base osteomyelitis that can invade soft tissue, the middle ear, inner
ear, or brain. Facial nerve paralysis may be an early sign,
with the glossopharyngeal and spinal accessory nerves less
frequently involved. Granulation tissue is classically seen
on the floor of the canal and at the bony-cartilaginous
junction. Clinical diagnosis can be confirmed with a raised
erythrocyte sedimentation rate plus an abnormal computed
S10
Otolaryngology–Head and Neck Surgery, Vol 134, No 4S, April 2006
tomography or magnetic resonance imaging scan61,62; other
imaging modalities include gallium scan, indium-labeled
leukocyte scan, technetium bone scan, and single-photon
emission tomographs (SPECT). Treatment includes systemic antibiotics adequate to cover pseudomonal and staphylococcal infection, including methicillin-resistant S aureus.
Biopsy may be necessary to detect neoplasia if the diagnosis
of malignant otitis externa is uncertain or response to therapy is incomplete.
Otomycosis, or fungal infection of the external ear
canal, is common in tropical countries, humid locations,
after long-term topical antibiotic therapy, and in those
with diabetes, HIV infection, or an immunocompromised
state. Aspergillus species (60% to 90%) and Candida
species (10% to 40%) are often cultured.63 Symptoms
include pruritus and thickened otorrhea, which may be
black, gray, bluish green, yellow, or white. Candidal
otitis externa results in white debris sprouting hyphae,
best seen with an otologic microscope. Aspergillus niger
appears as a moist white plug dotted with black debris
(“wet newspaper”).64 Management may include debridement plus topical antifungal therapy, systemic antifungal
therapy, or both.
Radiotherapy can damage the external ear by causing
acute and late skin reactions that involve the pinna, external
canal, and periauricular region.65 Acute events include erythema, desquamation, or ulceration of the auricle and ear
canal, thus leading to pain and otorrhea. Late skin changes
include atrophy, necrosis or ulceration, external otitis, and
external canal stenosis. Damage to the epithelium of sebaceous and apocrine glands can diminish cerumen secretion.
Management of AOE in patients after radiotherapy may
require systemic antimicrobials.
Middle-ear disease can modify treatment of AOE. Patients with a tympanostomy tube or tympanic membrane
perforation may develop diffuse AOE because of purulent
middle-ear secretions that enter the ear canal. This condition
has been called infectious eczematoid dermatitis because
the skin changes resemble eczema as well as infection.51 As
discussed later in the guideline, clinicians should prescribe
a nonototoxic topical preparation when the tympanic membrane is not intact. Management of the underlying middleear disease may also require systemic antimicrobials, imaging studies, or surgery.
Evidence Profile for 1a: Differential Diagnosis.
●
●
●
●
●
●
Aggregate evidence quality: C, observational studies and
D, expert opinion
Benefit: improved diagnostic accuracy
Harm: none
Cost: none
Benefits-harm assessment: preponderance of benefits
over harm
Value judgments: importance of accurate diagnosis
●
●
Role of patient preferences: none
Policy level: recommendation
Evidence Profile for 1b: Modifying Factors.
●
●
●
●
●
●
●
●
Aggregate evidence quality: C, observational studies
Benefit: optimizing treatment of AOE through appropriate diagnosis and recognition of modifying factors
Harm: none
Cost: additional expense of diagnostic tests or imaging
studies to identify modifying factors
Benefits-harm assessment: preponderance of benefits
over harm
Value judgments: avoiding complications in at-risk patients
Role of patient preferences: none
Policy level: recommendation
2. PAIN MANAGEMENT: The management of
diffuse AOE should include an assessment of
pain. The clinician should recommend
analgesic treatment based on the severity of
pain. Strong recommendation based on well-designed
randomized trials with a preponderance of benefit over
harm.
Pain relief is a major goal in the management of AOE.
Frequent use of analgesics is often necessary to permit
patients to achieve comfort, rest, and to resume normal
activities.66-68 Ongoing assessment of the severity of discomfort is essential for proper management. Use of a faces,69
Oucher,70 or visual analog71 scale may help determine the
level of pain, particularly for children and non-English
speaking patients.
Adequate pain control requires knowing the dose, timing, routes of delivery, and possible adverse effects of an
analgesic.66-68,72 Mild to moderate pain usually responds to
acetaminophen or nonsteroidal anti-inflammatory drugs
given alone or in fixed combination with an opioid (eg,
acetaminophen with codeine, oxycodone, or hydrocodone;
ibuprofen with oxycodone). Administering a nonsteroidal
anti-inflammatory drug during the acute phase of diffuse
AOE significantly reduces pain compared with placebo.73
Convenience, ease of use, and cost make orally administered analgesics the preferred route of administration
whenever possible. Rarely, parenteral analgesia may be
necessary to achieve adequate pain relief in a timely fashion. In all cases, analgesic therapy should be guided by the
recognition that pain is easier to prevent than treat. Thus,
early treatment at an appropriate starting dose is always
indicated. When frequent doses are required to maintain
adequate pain relief, the administration of analgesics at
fixed intervals rather than on a pro re nata (prn) basis may
be more effective. Nonpharmacologic therapies such as heat
or cold, relaxation, and distraction are of unproven value but
have stood the test of time.
Rosenfeld et al
Clinical Practice Guideline: Acute Otitis Externa
Acute analgesia and occasionally procedure-related sedation74 may be required to accomplish adequate aural toilet
in patients with severe inflammation and tenderness of the
canal. In 1 study,75 analgesic cream has been applied to the
ear canal in adults and cooperative children to relieve pain
and anesthetize the external auditory meatus if the tympanic
membrane is intact. Opioids such as fentanyl citrate, morphine sulfate, and hydromorphone hydrochloride are indicated for procedure-related pain and moderate to severe
around-the-clock pain.
Benzocaine otic solution, with or without antipyrine, is
available for topical anesthesia of the ear canal. There are no
clinical trials that show efficacy of these medications in
AOE, and the use of these drops may mask progression of
underlying disease while pain is being suppressed. Topical
benzocaine may cause contact dermatitis that can worsen or
prolong AOE.51 If a topical anesthetic drop is prescribed for
temporary pain relief, the patient should be reexamined
within 48 hours to ensure that AOE has responded appropriately to primary therapy.
The addition of a topical steroid to topical antimicrobial
drops has been shown to hasten pain relief in some randomized
trials,10,76 but others have shown no significant benefits.77,78
Evidence Profile for 2: Pain Management.
●
●
●
●
●
●
●
●
Aggregate evidence quality: B, 1 randomized controlled
trial limited to AOE; consistent, well-designed randomized trials of analgesics for pain relief in general
Benefit: increase patient satisfaction, allows faster return
to normal activities
Harm: adverse effects of analgesics
Cost: direct cost of medication
Benefits-harms assessment: preponderance of benefit
over harm
Value judgments: preeminent role of pain relief as an
outcome when managing AOE
Role of patient preferences: choice of analgesic, degree of
pain tolerance
Policy level: strong recommendation
3. INITIAL THERAPY: Clinicians should use
topical preparations for initial therapy of
diffuse, uncomplicated AOE. Systemic
antimicrobial therapy should not be used
unless there is extension outside the ear
canal or the presence of specific host factors
that would indicate a need for systemic
therapy. Recommendation based on randomized
controlled trials with minor limitations and a
preponderance of benefit over harm.
The recommendation for initial topical therapy applies to
the otherwise healthy patient with diffuse AOE that is not
complicated by osteitis, abscess formation, middle ear disease, or recurrent episodes of infection. Topical therapy
S11
should be supplemented by systemic antibiotics if the affected individual has a condition, especially diabetes that is
associated with markedly increased morbidity, or HIV infection/AIDS with immune deficiency that could impair
host defenses; if the infection has spread beyond the confines of the ear canal into the pinna, skin of the neck or face,
or into deeper tissues such as occurs with malignant external
otitis; or if there is good reason to believe that topical
therapy cannot be delivered effectively (see below Drug
Delivery).2,79
Topical preparations are recommended as initial therapy for diffuse, uncomplicated AOE because of safety,
efficacy over placebo in randomized trials, and excellent
clinical and bacteriologic outcomes in comparative studies. There are no data on the efficacy of systemic therapy
with the use of appropriate antibacterials and stratified by
severity of the infection. Moreover, orally administered
antibiotics have significant adverse effects that include
rashes, vomiting, diarrhea, allergic reactions, altered nasopharyngeal flora, and development of bacterial resistance.18,80-82 Societal consequences include direct transmission of resistant bacterial pathogens in homes and
child care centers.17
Three randomized trials have compared topical antimicrobial vs placebo for treating diffuse AOE.83-85 Metaanalysis of the 2 trials with similar methods83,84 yields a
combined absolute rate difference (RD) of 0.46 based on
89 patients (95% CI, 0.28 to 0.63), which suggests that
only 2 patients needed to be treated (NNT) with topical
antimicrobial to achieve 1 additional cure. Bacteriologic
efficacy (RD, 0.61) was higher than clinical efficacy.
Another trial85 reported significantly less edema and itching 3 days after therapy was initiated, and less edema,
itching, redness, scaling, and weeping 7 days after therapy was initiated. Conversely, another study86 showed no
benefit for an antimicrobial-steroid drop vs placebo, but
patients with chronic otitis externa, otomycosis, and furunculosis were also included.
No randomized, controlled trials have directly compared oral antibiotic therapy with topical therapy. Reviews of survey data, however, show that about 20% to
40% of subjects with AOE receive oral antibiotics, often
in addition to topical antimicrobials.2,15,16 Many of the
oral antibiotics selected are inactive against P aeruginosa
and S aureus, the most common pathogens identified in
cases of AOE. Further, treatment with penicillins, macrolides, or cephalosporins increases disease persistence
(rate ratios, 1.56 to 1.91), and treatment with cephalosporins also increases recurrence (rate ratio, 1.28; 95%
CI, 1.03 to 1.58).2
One additional study directly addresses the use of oral
antibiotics in treating diffuse AOE.87 When patients were
randomized to topical ointment plus oral antibiotic (trimethoprim-sulfamethoxazole) vs topical ointment plus
placebo, there was no significant difference in cure rates
at 2 to 4 days (RD, – 0.01; 95% CI, – 0.21 to 0.18) or at
S12
Otolaryngology–Head and Neck Surgery, Vol 134, No 4S, April 2006
5 to 6 days (RD 0.08; 95% CI, – 0.15 to 0.30). The
ointment (Kenacomb) contained an antifungal, an antibiotic active against gram-negative organisms, an antibiotic
active against gram-positive organisms, and a steroid.
The most compelling argument against the use of oral
antibiotics for diffuse AOE limited to the ear canal is the
efficacy of topical treatments that do not include antibiotics. Effective topical treatments include acetic acid,76,84,88,89 boric acid,77 aluminum acetate,90,91 silver
nitrate,92,93 and an endogenous antiseptic N-chlorotaurine.94 Topical steroids are also effective, as a single
agent,95-97 or in combination with acetic acid,76,88,89 or
an antifungal preparation.98 When the success of these
nonantibiotic therapies is considered, it is likely that for
cases of uncomplicated AOE, oral antibiotics, particularly those with no activity against P aeruginosa or S
aureus, are unnecessary.
An advantage of topical therapy is the very high concentration of antimicrobial that can be delivered to infected tissue, often 100 to 1000 times higher than can be
achieved with systemic therapy. For example a 0.3%
solution of antibiotic (a typical concentration in commercial otic drops) has a concentration of 3000 mcg/mL. Any
organisms known to cause AOE, even those considered “resistant,” will be unlikely to survive contact with this antibiotic
concentration. Because there are between 10 to 20 drops/
mL, depending on the nature of the liquid (solution vs
suspension, viscosity, etc), each dose of 3 to 5 drops contains about 0.5 to 1.5 mg of antibiotic.
Topical therapy avoids prolonged exposure of bacteria
to subtherapeutic concentrations of antibiotic, and may
therefore be less likely than systemic therapy to result in
selective pressure for resistant organisms.4,99 The avoidance of antibiotic exposure of host bacteria resident outside the ear canal, as occurs with systemic therapy, provides a further advantage to the reduction of the selection
of resistant microorganisms. Restrictive use of oral antibiotics for AOE is important because of the increased
resistance among common AOE pathogens, especially S
aureus and P aeruginosa.100,101
Along with prescribing topical antimicrobials, clinicians
should advise patients to resist manipulation to minimize
ear trauma and should discuss issues that pertain to water
restrictions during treatment. The insertion of earplugs or
cotton (with petroleum jelly) before showering or swimming can reduce the introduction of moisture into the ear.
The external auditory canal can be dried after swimming or
bathing with a hair dryer on the lowest setting.
Patients with AOE should preferably abstain from water
sports for 7 to 10 days during treatment. A swimming pool,
as long as prolonged submersion is avoided, can be allowed
in mild cases. Competitive swimmers sometimes return to
competition 2 to 3 days after treatment, or if they use
well-fitting ear plugs, after the pain has resolved.36,102,103
Patients with hearing aids or ear phones should limit insertion until pain and discharge (if present) have subsided.
Evidence Profile for 3: Initial Therapy.
●
●
●
●
●
●
●
●
Aggregate evidence quality: B, randomized controlled
trials with minor limitations; no direct comparisons of
topical vs systemic therapy
Benefit: avoid side effects by not using unnecessary systemic medications, avoid increased disease persistence
rates and disease recurrence rates seen when inappropriate systemic antibiotics are used, reduce antibiotic resistance by avoiding systemic antibiotics, and potential for
increased patient adherence to therapy
Harm: adverse effects of topical antimicrobials
Cost: oral or topical antimicrobials range in cost from a
few dollars to more than $100
Benefits-harms assessment: preponderance of benefit
over harm
Value judgments: desire to decrease the use of ineffective
treatments, societal benefit from avoiding the development of antibiotic resistance
Role of patient preferences: the selection of the specific
therapy may involve patient preferences in this decision
process based on cost, adherence to therapy, and other
factors
Policy level: recommendation
4. TOPICAL THERAPY: The choice of topical
antimicrobial for initial therapy of diffuse
AOE should be based upon efficacy, low
incidence of adverse events, likelihood of
adherence to therapy, and cost.
Recommendation based on randomized trials with some
heterogeneity and a preponderance of benefit over harm.
A variety of topical preparations are approved by the US
Food and Drug Administration (FDA) for treating AOE
(Table 5).104 Most of those currently available in the United
States provide antimicrobial activity through: 1) an antibiotic, which may be an aminoglycoside, polymyxin B, a
quinolone, or a combination of these agents; 2) a steroid,
such as hydrocortisone or dexamethasone; or 3) a low pH
antiseptic, such as aluminum acetate solution or acetic acid.
Efficacy of Topical Therapy. Clinicians should use a topical
drop that is efficacious for diffuse AOE. Efficacy is best
summarized with meta-analysis of randomized controlled
trials, but at the time of this writing (November 2005) there
were no published meta-analyses that concern diffuse AOE.
Therefore, we conducted a systematic review that yielded
20 randomized trials (Table 6)9,10,76-78,83-85,88-91,94-97,105-108
meeting inclusion criteria (see “Methods” section). Two
trials did not report data suitable for statistical pooling.85,96
Full details of the meta-analysis are reported separately in
this supplement,109 but relevant summary data are reported
herein.
The randomized trials in Table 6 have varying methods
and quality. Sample size ranges from 28 to 842 patients;
Rosenfeld et al
Clinical Practice Guideline: Acute Otitis Externa
S13
Table 5
Common topical otic preparations approved by the FDA for treating diffuse AOE
Cost, US$
Active drug(s)
Acetic acid, aluminum acetate
Acetic acid, hydrocortisone
Ciprofloxacin, hydrocortisone
Ciprofloxacin, dexamethasone
Neomycin, polymyxin B, hydrocortisone
Ofloxacin
Trade name
Bottle size, mL
Trade
Generic
Otic Domeboro
VoSol HC
Cipro HC
Ciprodex
Cortisporin Otic
Floxin Otic
60.0
10.0
10.0
7.5
10.0
5.0
31
110
125
125
89
71
22
24
—
—
46
—
Adapted from Fairbanks.104
50% of trials included both children and adults. Two studies88,89 reported outcomes by ears not patients, but were
included in the meta-analysis because 90% of patients had
unilateral disease. Study quality varied, with 50% having a
Jadad quality score less than 345; only 276,96 studies
achieved a maximum quality score of 5. Most (56%) studies
were not double-blind, because of obvious differences in
drug appearance or dosing schedule. The quality of 1
study78 could not be fully assessed because it was an abstract.
Table 6
Randomized controlled trials included in the meta-analysis of AOE treatment
Follow-up,
%
Jadad
score†
No
NS
1
Neo/methylpred vs placebo
Yes
73
‡4
ⱖ4
ns
19-67
Acetic/glyceryl vs placebo
Alum-acetate vs gentamicin
Betamethasone vs oxytet/polymyx/HC
Yes
No
No
100
73
98
‡4
‡4
3
91
120
4-76
18-52
Neo/colistin/HC vs placebo
Cipro vs tobramycin
Wick
No
100
NS
‡2
1
601
102
126
50
181
842
91
ⱖ1
6-73
1-44
8-89
10-82
2-85
22-61
Oflox vs neo/polymyx/HC
Acetic/HC vs neo/colistin/HC
Alum-acetate vs neo/polymyx/HC
NCT vs neo/polymyx/HC
Acetic/HC vs neo/polymyx/HC
Cipro vs cipro/HC vs neo/polymyx/HC
Cipro vs cipro/dex vs neo/polymyx/HC
No
Wick
Yes
Wick
Yes
Yes
No
79
81
93
100
61
83
NS
3
‡3
1
2
‡4
1
1
468
53
54
1-90
11-74
ⱖ18
Cipro/dex vs neo/polymyx/HC
HC butyrate vs oxytet/polymyx/HC
Cipro vs gentamicin
Yes
No
No
85
87
NS
2
2
‡2
28
39
ns
ⱖ18
Yes
Yes
86
67
‡3
‡5
213
ⱖ18
Boric/ethyl vs neo/polymyx/HC
Betamethasone vs betamethasone/
neo
Acetic vs acetic/triamcin vs
neo/polymyx/dex
Yes
93
‡5
Author year, country
N
Age, y
Arnes and Dibb105 1993,
Norway
Cannon and Grunwaldt83
1967, USA
Cannon84 1970, USA
Clayton et al90 1990, UK
Emgard and Helmstrom97
2005, Sweden
Freedman85 1978, USA
Goldenberg et al106 2002,
Israel
Jones et al9 1997, USA
Kime et al88 1978, USA
Lambert91 1981, Cyprus
Neher et al94 2004, Austria
Ordonez et al89 1978, USA
Pistorius et al10 1999, USA
Psifidis et al78 2005,
Greece
Roland et al107 2004, USA
Ruth et al95 1990, Sweden
Sabater et al108 1996,
Spain
Slack771987, UK
Tsikoudas et al96 2002, UK
30
ⱖ18
Cipro vs oxytet/polymyx/HC
40
2-68
43
66
51
van Balen et al76 2003,
The Netherlands
Topical treatment groups
Aural
toilet
Acetic, acetic acid 2%; alum-acetate, aluminum acetate 8%; boric, boric acid 4%; cipro, ciprofloxacin; dex, dexamethasone; ethyl,
ethyl alcohol 25%; HC, hydrocortisone; glyceryl, glyceryl acetate 88%; methylpred, methylprednisolone; NCT, N-chlorotaurine
(antiseptic); neo, neomycin; NS, not stated; oflox, ofloxacin; oxytet, oxytetracycline; polymyx, polymyxin B; triamcin, triamcinolone.
†Indicates a double-blind comparison.
‡Quality score ranging from 1 (lowest) to 5 (highest).
S14
Otolaryngology–Head and Neck Surgery, Vol 134, No 4S, April 2006
Table 7
Summary of meta-analyses of topical antimicrobials for treating acute otitis externa
Treatment group vs control group
outcome*
Antimicrobial vs placebo
1. Clinical cure at 3-10 days
2. Bacteriologic cure
Antiseptic vs antibiotic
3. Clinical cure at 7-10 days
4. Clinical cure at 11-14 days
Quinolone vs nonquinolone antibiotic(s)
5. Clinical cure at 3-4 days
6. Clinical cure at 7-10 days
7. Clinical cure at 14-28 days
8. Improved at 7-10 days
9. Bacteriologic cure
10. Any adverse event
Antimicrobial/steroid vs antimicrobial alone
11. Clinical cure at 7 days
12. Bacteriologic cure
Steroid/antibiotic vs steroid alone
13. Clinical cure at 7-11 days
References
combined
83,84
83,84
76,77,89,94
76,77,88,91
106,107
9,10,78,105,107,108
10,106,107
10,105,107
9,10,78,105,106,107
9,10,107
10,76,78
10,78
95,97
N
Control
rate†
RD (95% CI)‡
P value
89
§112
0.15
0.20
0.46 (0.29, 0.63)
0.61 (0.46, 0.76)
⬍0.001
⬍0.001
318
368
0.65
0.80
0.05 (⫺0.03, 0.12)
0.04 (⫺0.06, 0.13)
476
1475
936
890
980
1330
0.15
0.77
0.83
0.89
0.87
0.15
0.11
0.07
0.04
0.05
0.08
0.002
(⫺0.06, 0.28)
(⫺0.02, 0.16)
(⫺0.01, 0.08)
(⫺0.05, 0.14)
(0.006, 0.16)
(⫺0.07, 0.08)
0.192
0.110
0.145
0.292
0.035
0.963
660
342
0.68
0.93
0.04 (⫺0.08, 0.16)
⫺0.02 (⫺0.15, 0.11)
0.546
0.761
92
0.72
⫺0.20 (⫺0.38⫺0.03)
0.021
0.217
0.468
CI, Confidence interval; RD, absolute rate difference.
*Refer to Table 6 for individual study details.
†Control rate is calculated by simple division of total events by total patients to aid in interpreting the RD.
‡Absolute change in outcomes for treatment vs. control groups, beyond the control group rate, based on random-effects metaanalysis.
§Analysis by ears not patients.
The trials identified by systematic review permit the
following statistical comparisons of topical therapy that are
summarized in Table 7:
●
●
●
●
●
antimicrobial (antibiotic or antiseptic) vs placebo, 2 trials
(discussed in preceding section)
antiseptic vs antimicrobial, 8 trials
quinolone antibiotic vs nonquinolone antibiotic(s), 7 trials
steroid-antimicrobial vs antimicrobial alone, 3 trials
antimicrobial-steroid vs steroid alone, 2 trials
We found no significant differences in clinical outcomes
of AOE (Table 7) for antiseptic vs antimicrobial, quinolone
antibiotic vs nonquinolone antibiotic(s), or steroid-antimicrobial vs antimicrobial alone. Regardless of topical agent
used, about 65% to 90% of patients had clinical resolution
within 7 to 10 days. One potential explanation is that differences among agents may have been missed because of
low statistical power; however, all analyses had a combined
sample size above 300 patients, making low power unlikely.
Further, the magnitude of differences observed between
treatments is very modest, with a maximum rate difference
of 0.11 (NNT of 9 patients). Therefore, even if additional
studies were performed to increase power, it is likely that
differences among treatments remain small or negligible.
The only clinical comparison that achieved statistical
significance in Table 7 is for steroid/antibiotic vs steroid
alone (comparison 13). The reason for inferior outcomes
with steroid/antibiotic is unclear, although in 1 of the stud-
ies97 cited the steroid used for single-agent therapy had high
potency (betamethasone) and the comparator had low potency (hydrocortisone). Conversely, the second study95 used
a low potency steroid (hydrocortisone) in both groups. The
combined analysis with both studies, however, includes
only 92 patients and the broad confidence limits cannot
exclude a trivial effect. Additional studies are needed to
confirm this finding and to increase precision.
Another significant comparison in Table 7 is bacteriologic
efficacy of a quinolone antibiotic vs a nonquinolone antibiotic
(comparison 9); about 87% of patients with AOE have bacteriologic cure after nonquinolone therapy, with an 8% absolute
increase when a quinolone antibiotic is used. The clinical
significance of this modest effect (NNT of 12 patients) is
reduced when we consider that persistent bacteria in the ear
canal after treatment does not necessarily imply persistent
AOE symptoms. Generalizability of bacteriologic results is
also limited because not all patients had positive cultures before treatment and post-treatment cultures were not always
obtained for those who were initially positive.
Although the meta-analysis results suggest minimal or
no difference in clinical or bacteriologic cure rates among
topical agents, some of the more recent studies have shown
significant differences in the rapidity of treatment response
or symptom resolution. For example, the addition of hydrocortisone to ciprofloxacin significantly reduced median ear
pain from 4.7 to 3.8 days,10 and the addition of hydrocortisone to acetic acid reduced median ear pain from 8.0 to 7.0
Rosenfeld et al
Clinical Practice Guideline: Acute Otitis Externa
S15
days.76 Two other studies showed differences in inflammation scores94,107 and 1 showed significantly less itching with
the steroid-containing drop.97
from a few days up to several weeks in published trials. Recent
studies suggest that 7 days of therapy are adequate, at least for
quinolone antibiotics.9,10,107,116
Adverse Events, Adherence To Therapy, and Cost. The lack
of differences in efficacy, on average, among most topical
antimicrobial and steroid preparations (Table 7) suggests
that patient preference and clinician experience are important aspects in selecting therapy. Cost, adherence to therapy,
and adverse effects must also be considered.
Only a few studies9,10,107 report detailed information on
adverse events and show an overall low incidence and comparable rates among treatment groups. The most common
problems are pruritus (about 7%) and site reaction (5%); other
events with an incidence less than 2% include rash, discomfort,
otalgia, dizziness, vertigo, superinfection, and reduced hearing.9,107 None of the randomized trials reported otomycosis
after topical antibiotics, although otomycosis has been described anecdotally after topical ofloxacin therapy for AOE.110
One study10 that compared 2 quinolone preparations to a neomycin product found that only 1.1% of patients had to discontinue the drug because of infection, nausea, or vomiting. Unfortunately, studies of antiseptics have reported limited to no
information with respect to adverse events.
Contact dermatitis is a potential sequela of topical antimicrobial or steroid therapy, but it is rare after a single course of
therapy for diffuse AOE. Three studies9,10,107 have compared
a quinolone drop vs neomycin-polymyxin B-hydrocortisone
drop for diffuse AOE, with no significant difference in adverse
events individually or when combined (Table 7, comparison
10). Conversely, about 30% to 60% of patients with chronic or
eczematous external otitis develop a contact dermatitis, most
often to aminoglycosides such as neomycin and framycetin.58,111-115 No studies are limited specifically to patients with
recurrent AOE, chronic external otitis, or eczematous external
otitis, but it would appear prudent to avoid the use of aminoglycoside drops in these populations.
Remaining factors to consider when prescribing topical
therapy include adherence to therapy and cost. Adherence to
therapy and patient satisfaction are highest when drops are
easy to administer,40 which would entail a less frequent dosing
schedule, shorter duration of therapy, or both. There are no
comparative studies, but drops administered 4 times daily (eg,
neomycin, polymyxin, hydrocortisone) may be less acceptable
to some patients. Cost varies widely among available otic
preparations (Table 5) and range from a few dollars for antiseptics or generic products (eg, neomycin, polymyxin B, hydrocortisone) to more than $100 for quinolones, with or without a steroid.
Dosing schedules for AOE have not been studied systematically, but available data suggest that, at least with quinolone
drops (and perhaps also with the other concentration-dependent drugs like the aminoglycosides), a twice-daily (bid) dose
regimen is adequate. One open label study116 showed good
clinical outcomes when ofloxacin was given once daily. The
optimal duration of therapy has not been determined and varies
Complementary and Alternative Therapies. There are no
data with respect to the efficacy of complementary and
alternative therapies for AOE. Isopropyl (“rubbing”) alcohol, and 5% acetic acid (white vinegar) plus equal parts of
isopropyl alcohol or water, are time honored “home remedies,” but have never been formally evaluated in clinical
trials. The similarity of these preparations to some antiseptic
or acidifying agents that have been studied suggests they
may be effective. For example, most acetic acid preparations have a concentration of 2.5%, which equals a 50%
dilution of white vinegar. Although “tea tree oil” has been
found to be effective in vitro against 71% of organisms
cultured from 52 patients with AOE,117 Pseudomonas was
resistant in 75% of cases, and no controlled efficacy trials
that evaluate this form of therapy have been described.
Ear candles should not be used in treating AOE. Ear candles
have never been shown to be efficacious for AOE but have
been shown to produce harm.118 Obstruction of the ear canal
with paraffin and associated hearing loss and perforation of the
tympanic membrane have been reported.119
Evidence Profile for 4: Topical Therapy.
●
●
●
●
●
●
●
●
Aggregate evidence quality: B, randomized controlled
trials with some heterogeneity
Benefit: effective therapy, appropriate adherence to therapy, and acceptable cost
Harm: low incidence of adverse events
Cost: direct cost of topical antimicrobials or steroids
Benefits-harms assessment: preponderance of benefit
over harm
Value judgments: meta-analysis with cure rates (clinical
and bacteriologic) used as primary measure of efficacy;
heterogeneity among trials considered acceptable for statistical pooling with a random effects model
Role of patient preferences: substantial role for patient
preference in choice of topical therapeutic agent
Policy level: recommendation
5. DRUG DELIVERY: Clinicians should inform
patients how to administer topical drops.
When the ear canal is obstructed, delivery of
topical preparations should be enhanced by
aural toilet, placement of a wick, or both.
Recommendation based on observational studies with a
preponderance of benefit over harm.
For topical treatment to be effective, the drug must be
delivered to infected tissues. Although the majority of patients with uncomplicated AOE will require only topical
medication, for some patients, additional management is
needed to ensure the appropriate drug delivery. Ensuring
S16
Otolaryngology–Head and Neck Surgery, Vol 134, No 4S, April 2006
adequate delivery of the topical medication may require
removal of a foreign body, performance of aural toilet to
remove obstructing debris, placement of a wick to permit
drug delivery through the length of the ear canal, or all
three.
Drug delivery may be impaired by poor adherence to
therapy, poor application (ie, “missing” the ear canal), debris filling the canal, or edema closing the canal. Poor
adherence to therapy and ineffective administration must be
dealt with by providing clear instructions. Self-administration of eardrops is difficult because it must be done by feel.
Only 40% of patients who self-medicate do so appropriately
during the first 3 days,120 often tending to under-medicate.
Adherence to therapy increases significantly when someone
other than the patient applies the drops,121 which makes this
the preferred method of administration when feasible.
Ototopical drops should be applied with the patient lying
down and the affected ear upward. Drops should be run
along the side of the canal until it is filled. The amount
required will vary with the age and size of the patient.
Gentle to-and-fro movement of the pinna is often necessary
to eliminate trapped air and to assure filling, particularly
when a viscous solution is used. The patient should remain
in this position for about 3 to 5 minutes. Use of a timer to
mark the minutes is often helpful to facilitate the cooperation of young children. After the placement of drops, the
canal is best left open to dry to avoid trapped moisture and
infected debris.
The ear canal should be cleared of inflammatory debris,
obstructing cerumen, or any foreign object. There are no
randomized studies of the use of aural toilet in AOE, but
some investigators have proposed that aural toilet by itself
(without antimicrobials) is therapeutic.96 Aural toilet may
be done with a gentle lavage using body-temperature water,
saline solution, or hydrogen peroxide. Alternative methods
of aural toilet include physically removing the obstructing
debris with suction or dry mop (blotting with cotton). Adequate visualization for suctioning may be facilitated by
using an otoscope with an open head or a binocular otologic
microscope.
There are no randomized trials that address the safety of
aural lavage in diabetics or immunocompromised patients
with AOE. Lavage of the ear canal for cerumen impaction
in elderly or diabetic patients, however, has been implicated
as a contributing factor in malignant otitis externa.122-124
The pathophysiology of malignant (necrotizing) otitis externa is poorly understood, but aural water exposure is a
potential iatrogenic factor.61 Patients with risk factors such
as diabetes or immunocompromised state, as well as those
with established malignant otitis externa, may require atraumatic cleaning with aural suctioning under microscopic
guidance.
Clinicians may place a wick in the ear canal if there is
edema that prevents drop entry95 or if most of the tympanic
membrane cannot be visualized.76 The wick should preferably be made of compressed cellulose because it expands
when exposed to moisture, facilitates drug delivery, and
reduces ear canal edema. Alternatively, ribbon gauze can be
used.125 Once a dry wick is placed in the ear canal, some
experts recommend moistening the wick with an aqueous
solution (water, saline solution, aluminum acetate) before
the first application of an otic suspension or a nonaqueous
viscous medication. A wick should not be made of a simple
cotton ball since the cotton can fall apart and become lodged
in the ear canal.
Many treatment studies uniformly use a wick to improve
drug delivery (Table 6), but there are no trials of wick
efficacy. Consequently, the benefit of a wick is questioned
by some clinicians, especially in managing uncomplicated
AOE. However, following first principles, if the anatomy
(narrow or edematous canal) make delivery of the topical
medicine problematic, the use of a wick seems prudent. A
wick is unnecessary once the ear canal edema subsides,
which may occur within 24 hours51 or a few days of topical
therapy. The wick may fall out spontaneously, may be
removed by the patient, or may be removed by a clinician at
a scheduled follow-up visit.
Evidence Profile for 5: Drug Delivery.
●
●
●
●
●
●
●
●
Aggregate evidence quality: C, observational studies and
D, expert opinion
Benefit: improved adherence to therapy and drug delivery
Harm: pain and local trauma caused by inappropriate
aural toilet or wick insertion
Cost: wicks (inexpensive)
Benefits-harms assessment: preponderance of benefit
over harm
Value judgments: none
Role of patient preferences: choice of self-administering
drops vs using assistant
Policy level: recommendation
6. NONINTACT TYMPANIC MEMBRANE.
When the patient has a tympanostomy tube
or known perforation of the tympanic
membrane, the clinician should prescribe a
nonototoxic topical preparation.
Recommendation based on reasoning from first principles
and on exceptional circumstances where validating
studies cannot be performed and there is clear
preponderance of benefit over harm.
Special consideration must be given to the individual with
known or suspected perforation of the tympanic membrane.
The external auditory canal, including the tympanic membrane, is lined with keratinizing squamous epithelium, but
the middle ear is lined with mucosa. This mucosa forms the
lateral portion of the round window membrane, which separates the middle-ear space from the fluids of the inner ear.
Antibiotics placed into the middle ear can cross the round
window membrane and reach the inner ear. Ototoxic anti-
Rosenfeld et al
Clinical Practice Guideline: Acute Otitis Externa
biotics delivered into the middle ear space of experimental
animals, including primates, consistently cause severe hearing loss and ototoxic injury to the organ of Corti.126-128
Clinical experience with topical ototoxic antibiotics in patients with tympanic membrane perforation suggests that hearing loss does not occur after a short course of therapy129,130;
however, severe hearing loss has been observed after prolonged or excessive administration of topical drops.130,131 The
validity of these and other clinical reports is limited by retrospective design, incomplete follow-up, and inconsistent audiologic testing. Given the ethical limitations of randomizing
patients with a nonintact tympanic membrane to an ototoxic
drop, it is unlikely that definitive evidence (validating studies)
is forthcoming.
Careful examination of the tympanic membrane will
reveal a perforation in some cases of AOE. The ear canal
and auricle may be so tender or swollen, however, that the
tympanic membrane cannot be visualized without undue
pain or discomfort. If swelling or discomfort do not preclude its use, tympanometry can sometimes be helpful to
establish the presence of an intact tympanic membrane.
When tympanometry shows a normal type A tracing
(peaked curve with normal pressure), the tympanic membrane is assumed to be intact unless there is a reason to
believe it is not (eg, an indwelling tympanostomy tube).
A perforation may be suspected if the patient has a
positive history, unless the most recent examination before
the episode of AOE has verified that the perforation has
closed. Children with tympanostomy tubes are a special
instance within this category. Most tympanostomy tubes
remain in the tympanic membrane for at least 6 to 12
months; therefore a patent tube should be assumed to be
present within the tympanic membrane of any individual
who had it placed less than a year earlier, unless tube
extrusion and subsequent closure of the tympanic membrane have been documented. Individuals who taste substances, presumably medicinals, placed into their ear, or
who can expel air out their ear canal by pinched nose
blowing, can be assumed to have a perforation.
If the tympanic membrane is known or suspected to be
nonintact, topical drops that contain alcohol, have a low pH
(most acidifying/antiseptic agents), or both should be
avoided because of pain and potential ototoxicity. Substances with ototoxic potential (eg, aminoglycosides, alcohol) should not be used when the tympanic membrane is
perforated and the middle ear space is open, because the risk
of ototoxic injury outweighs the benefits compared with
nonototoxic antimicrobials with equal efficacy.132 The only
topical antimicrobials approved by the FDA (December
2005) for middle ear use are ofloxacin and ciprofloxacin/
dexamethasone. Moreover, there is an explicit warning by
the manufacturer that neomycin/polymyxin B/hydrocortisone not be used with a nonintact tympanic membrane:
“WARNINGS. Neomycin can induce permanent sensorineural hearing loss due to cochlear damage,
mainly destruction of hair cells in the organ of Corti.
S17
The risk is greater with prolonged use. Therapy
should be limited to 10 consecutive days (see PRECAUTIONS-General). Patients being treated with
eardrops containing neomycin should be under close
clinical observation. CORTISPORIN Otic Suspension
should not be used in any patient with a perforated
tympanic membrane (emphasis added).”133
AOE can be secondary to acute otitis media. For example, mucopurulent exudate that flows through an acute tympanic membrane perforation from the middle ear can infect
the tissues of the ear canal and creat a secondary otitis
externa. Less commonly, AOE will develop independently
in an ear with acute otitis media (AOM). When AOM exists
together with AOE, the AOM should be treated as an independent disease process according the current guidelines.52
Evidence Profile for 6: Nonintact Tympanic Membrane.
●
●
●
●
●
●
●
●
Aggregate evidence quality: D, reasoning from first principles, and X, exceptional situations where validating
studies cannot be performed
Benefit: avoid pain and hearing loss
Harm: none
Cost: eardrops without ototoxicity are more costly
Benefits-harms assessment: preponderance of benefit
over harm
Value judgments: importance of avoiding iatrogenic hearing loss from a potentially ototoxic topical preparation
when nonototoxic alternatives are available; placing
safety above economic cost
Role of patient preferences: increased cost of nonototoxic
preparation for patients who pay for their own medication
vs risk of ototoxicity in less expensive alternatives
Policy level: recommendation
7. OUTCOME ASSESSMENT: If the patient
fails to respond to the initial therapeutic
option within 48 to 72 hours, the clinician
should reassess the patient to confirm the
diagnosis of diffuse AOE and to exclude
other causes of illness. Recommendation based on
observational studies and a preponderance of benefit over
harm.
Appropriate treatment of uncomplicated AOE should be
followed by symptom improvement (otalgia, itching, fullness) within 48 to 72 hours (Fig 1). In clinical trials that
evaluate patient outcomes of topical treatment with symptom diaries, significant decreases in patient-reported ear
pain are generally seen after 1 day of treatment and most
pain resolves within 4 to 7 days.76,97,116 One prospective
cohort study40 that explored the relationship of patientreported satisfaction with clinical outcomes showed that
symptom relief was the factor most highly associated with
patient satisfaction.
S18
Otolaryngology–Head and Neck Surgery, Vol 134, No 4S, April 2006
Figure 1
Flowchart for managing acute otitis externa.
Rosenfeld et al
Clinical Practice Guideline: Acute Otitis Externa
Initial treatment failure of diffuse AOE may be caused
by an obstructed ear canal, poor adherence to therapy,
misdiagnosis, microbiologic factors, host factors, or contact
sensitivity to eardrops. If topical antimicrobial therapy was
prescribed, the clinician should reassess the patency of the
ear canal to ensure that edema or debris are not impeding
drug delivery. Any obstruction should be addressed with
aural toilet, wick placement, or both (see preceding section,
Drug Delivery), or, if the obstruction cannot be relieved,
systemic therapy is begun with an oral antibiotic that covers
P aeruginosa and S aureus.
The clinician should also assess adherence with therapy.
Patients tend to over-administer ear drops when pain is
greatest and to under-administer as symptoms resolve.40,120
Alternative causes of ear pain and associated otorrhea
should be considered if the patient fails to respond to treatment, though the need for specialist referral is uncommon
(3%) when AOE is treated appropriately.2 Fungi may be
present as a copathogen in some patients with AOE, and
cause persistent infection from overgrowth in the ear canal
if the flora is altered after topical antibacterial therapy.5 A
culture of the ear canal can identify fungi, resistant bacteria,
or unusual causes of infection that require targeted topical
or systemic therapy.
Initial treatment failures that are not related to drug
delivery or microbiologic factors may reflect comorbidity or
misdiagnosis.38,134 Persistent symptoms can be caused by
dermatologic disorders that include dermatitis (atopic, seborrheic, or contact), psoriasis, dermatomycosis, or acne
that involves the external auditory canal. The ear canal and
tympanic membrane should be reexamined to detect an
unrecognized foreign body, perforated tympanic membrane,
or middle ear disease. Patients with severe refractory symptoms should be reassessed for malignant otitis externa or
carcinoma of the external auditory canal, especially if granulation tissue is present.51,135
Contact sensitivity of the external auditory canal can
result in refractory AOE in some patients. Delayed-type
hypersensitivity reactions to topical antiseptic otic preparations are characterized by severe pruritus, skin inflammation, edema of the external auditory canal, and persistent
otorrhea; blisters and vesicles may be present. The allergic
reaction can extend beyond the ear canal to involve the skin
around the ear and the neck. Neomycin-containing eardrops
are most commonly noted to cause contact sensitivity,
which has a 13% to 30% prevalence on patch testing of
patients with chronic otitis externa.134,136,137
Contact sensitivity of the ear canal may also result
from other topical antimicrobials (bacitracin, quinolones,
gentian violet, polymyxin B sulfate), topical steroid preparations (hydrocortisone, triamcinolone), or topical anesthetics (benzocaine alone, or combined with dibucaine
and tetracaine [caine mix]). Preservatives in topical otic
preparations associated with at least a 1% incidence of
contact sensitivity include propylene glycol, thimerosal,
benzalkonium chloride, benzethonium chloride, and
S19
methyl-p-oxybenzoate. Fragrance additives may also
cause similar reactions. Finally, contact sensitivity may
be caused by silicone ear plugs or by hearing-aid molds
that contain silicone or methyl-methacrylate.134,136,137
Evidence Profile for 7: Outcome Assessment.
●
●
●
●
●
●
●
●
Aggregate evidence quality: C, observational studies
Benefit: identify misdiagnosis and potential complications from delayed management; reduce pain
Harm: none
Cost: need for reevaluation by clinician
Benefits-harms assessment: preponderance of benefit
over harm
Value judgments: none
Role of patient preferences: limited
Policy level: recommendation
IMPLEMENTATION CONSIDERATIONS
The complete guideline is published as a supplement to
Otolaryngology–Head and Neck Surgery to facilitate reference and distribution. A full-text version of the guideline will also be accessible free of charge at the www.entnet.org, the AAO–HNSF website. The AAO–HNSF has
also given permission for members of the working group
to have their professional medical societies publish all or
part of the guideline in their journals or in electronic
form. The guideline will be presented to AAO–HNSF
members as a miniseminar at the annual meeting after
publication. Existing brochures and publications by the
AAO–HNSF will be updated to reflect the guideline
recommendations.
Anticipated barriers to application of the recommendations in the guideline include: 1) difficulty of changing
ingrained clinician habits toward prescribing ineffective
systemic therapy for AOE, 2) inability or unwillingness of
some clinicians to perform aural toilet or insert a wick into
the ear canal, and 3) cost of some topical medications,
especially the quinolone products recommended for use
with a nonintact tympanic membrane. The first 2 can be
addressed with educational events and workshops at continuing medical education events. The issue of cost should
become less problematic in the next few years as generic
versions of the quinolone otic drops become available.
The impact of the guideline on clinical practice will be
assessed for otolaryngologists when a performance measure
is developed. As noted above, one purpose of developing
the guideline was to facilitate creation of a performance
measure for maintenance of certification in otolaryngology–
head and neck surgery. The guideline working group did not
specifically discuss measuring impact on clinicians other
than otolaryngologists.
S20
Otolaryngology–Head and Neck Surgery, Vol 134, No 4S, April 2006
RESEARCH NEEDS
1. Clinical trials to determine the efficacy of EMLA
cream (lidocaine 2.5% and prilocaine 2.5%) and other
topical anesthetic solutions for relief of pain caused by
AOE
2. Clinical trials to determine the efficacy of topical steroids for relief of pain caused by AOE
3. Observational studies or clinical trials to determine
optimal time to discontinue water precautions for AOE
4. Studies to assess the utility of drying the ear canal with
a hair dryer or similar device after water exposure to
prevent AOE
5. Increased ability to distinguish treatment failure from
topical sensitivity when a patient with AOE fails to
respond to topical therapy
6. High-quality randomized trials of comparative clinical
efficacy for AOE that use an appropriate randomization
scheme, explicit double-blind protocol, and fully describe dropouts and withdrawals
7. High-quality randomized trials that assess the benefit of
systemic antimicrobial therapy vs topical therapy in
patients stratified by severity of signs and symptoms
8. High-quality randomized trials of comparative clinical
efficacy for AOE that provide clinical outcomes early
in the course of therapy (eg, after 2 to 4 days of
therapy) and compare time to symptom resolution in
addition to categorical responses (eg, cure, improve,
failure) for specific days
9. Comparative clinical trials of “home therapies” for (eg,
vinegar, alcohol) vs antimicrobials for treatment of
AOE
10. Studies to document the effect of pH on outcomes for
topical therapies
11. Define the optimal duration of topical therapy for AOE
and the role of patient preferences
12. Additional clinical trials to define the change in outcomes when a steroid is added to a topical antimicrobial
13. Define with greater precision the indications for aural
toilet and wick placement
14. Determine the efficacy of aural toilet as an independent
factor in treatment of AOE
15. Comparative clinical trials of wick vs no wick in administration of topical therapy
16. Determine the optimal composition and materials for a
wick
17. Define the best methods of teaching clinicians, especially those in primary care settings, how to safely and
effectively perform aural toilet and wick insertion
18. Determine the optimal method to assess tympanic
membrane integrity in patients with AOE (eg, what is
the utility of tympanometry?)
19. Assess the correlation between clinical cure and bacteriologic cure in clinical trials
20. Investigate the importance of bacteriologic cure and
determine the natural history and clinical significance
of bacteriologic failures
21. Assess the role of fungi to determine outcomes
22. Evaluate prevention strategies, including prophylactic
use of vinegar with equal parts of isopropyl (rubbing)
alcohol after swimming
REFERENCES
1. Guthrie RM. Diagnosis and treatment of acute otitis externa: an
interdisciplinary update. Ann Otol Rhinol Laryngol 1999;17:2–23.
2. Rowlands S, Devalia H, Smith C, et al. Otitis externa in UK general
practice: a survey using the UK General Practice Research Database.
Br J Gen Pract 2001;51:533– 8.
3. Raza SA, Denholm SW, Wong JC. An audit of the management of
otitis externa in an ENT casualty clinic. J Laryngol Otol 1995;109:
130 –3.
4. Roland PS, Stroman DW. Microbiology of acute otitis externa. Laryngoscope 2002;112:1166 –77.
5. Dibb WL. Microbial aetiology of otitis externa. J Infect 1991;22:
233–9.
6. Agius AM, Pickles JM, Burch KL. A prospective study of otitis
externa. Clin Otolaryngol 1992;17:150 – 4.
7. Cassisi N, Cohn A, Davidson T, et al. Diffuse otitis externa: clinical and
microbiologic findings in the course of a multicenter study on a new otic
solution. Ann Otol Rhinol Laryngol 1997;86(Suppl 39):1–16.
8. Clark WB, Brook I, Bianki D, et al. Microbiology of otitis externa.
Otolarngol Head Neck Surg 1997;116:23–5.
9. Jones RN, Milazzo J, Seidlin M. Ofloxacin otic solution for treatment
of otitis externa in children and adults. Arch Otolaryngol Head Neck
Surg 1997;123:1193–200.
10. Pistorius B, Westberry K, Drehobl ??, et al. Prospective, randomized,
comparative trial of ciprofloxacin otic drops, with or without hydrocortisone, vs. polymyxin B-neomycin-hydrocortisone otic suspension
in the treatment of acute diffuse otitis externa. Infect Dis Clin Pract
1999;8:387–95.
11. Arshad M, Khan NU, Ali N, et al. Sensitivity and spectrum of
bacterial isolates in infectious otitis externa. J Coll Physicians Surg
Pak 2004;14:146 –9.
12. Manolidis M, Freidman R, Hannley M, et al. Comparative efficacy of
aminoglycoside versus fluoroquinolone topical antibiotic drops. Otolaryngol Head Neck Surg 2004;130(Suppl):S83–S88.
13. Martin TJ, Kerschner JE, Flanary VA. Fungal causes of otitis externa
and tympanostomy tube otorrhea. Int J Pediatr Otorhinolaryngol.
2005 May 28; [Epub ahead of print]; 2005;69:1503-1508.
14. Hajioff D. Otitis externa. Clin Evid 2004;12:755– 63.
15. Halpern MT, Palmer CS, Seidlen M. Treatment patterns for otitis
externa. J Am Board Fam Pract 1999;12:1–7.
16. McCoy SI, Zell ER, Besser RE. Antimicrobial prescribing for otitis
externa in children. Pediatr Infect Dis J 2004;23:181–3.
17. Levy SB. The antibiotic paradox. how the misuse of antibiotic destroys their curative powers. Cambridge, MA: Perseus Publishing;
2002.
18. McCormick AW, Whitney CG, Farley MM, et al. Geographic diversity and temporal trends of antimicrobial resistance in Streptococcus
pneumoniae in the United States. Nat Med 2003;9:424 –30.
19. Nussinovich M, Rimon A, Volovitz B, et al. Cotton-tip applicators as
a leading source of otitis externa. Int J Pediatr Otorhinolaryngol
2004;68:433–5.
20. Goffin F. pH as a factor in the external otitis. N Engl J Med 1963;
268:287–9.
21. Martinez JD, Willis CM, Capper JW. External auditory canal pH in
chronic otitis externa. Clin Otolaryngol 2003;28:320 – 4.
22. Yelland M. Otitis externa in general practice. Med J Austalia 1992;
156:325–30.
23. Blake P, Matthews R, Hornibrook J. When not to syringe an ear. N
Z Med J 1998;111:422– 4.
Rosenfeld et al
Clinical Practice Guideline: Acute Otitis Externa
24. Berry RG, Collymore VA. Otitis externa and facial cellulitis from
Oriental ear cleaners. West J Med 1993;158:536.
25. Brook I, Coolbaugh JC. Changes in the bacterial flora of the external
ear canal from the wearing of occlusive equipment. Laryngoscope
1984;94:963–5.
26. Hirsch BE. Infections of the external ear. Am J Otolaryngol 1992;
13:145–55.
27. Russell JD, Donnelly M, McShane DP, et al. What causes acute otitis
externa? J Laryngol Otol 1993;107:898 –901.
28. Hoadley AW, Knight DE. External otitis among swimmers and nonswimmers. Arch Environ Health 1975; 9:445– 8.
29. Calderon R, Mood EW. An epidemiological assessment of water
quality and swimmers’ ear. Arch Environ Health 1982;37:300 –5.
30. Hansen US. Otitis externa among users of private swimming pools.
Ugeskr Laeger 1997;159:4383– 8.
31. Moore JE, Heaney N, Millar BC, et al. Incidence of Pseudomonas
aeruginosa in recreational and hydrotherapy pools. Commun Dis
Public Health 2002;5:23– 6.
32. Hajjartabar M. Poor-quality water in swimming pools associated with
a substantial risk of otitis externa due to Pseudomonas aeruginosa.
Water Sci Technol 2004;50:63–7.
33. Steuer MK, Hofstadter F, Probster L, et al. Are ABH antigenic
determinants on human outer ear canal epithelium responsible for
Pseudomonas aeruginosa infections? ORL J Otorhinolaryngol Relat
Spec 1995;57:148 –52.
34. Sundstrom J, Jacobson K, Munck-Wikland E, et al. Pseudomonas
aeruginosa in otitis externa: a particular variety of the bacteria? Arch
Otolaryngol Head Neck Surg 1996;122:833– 6.
35. Bojrab DI, Bruderly T, Abdulrazzak Y. Otitis externa. Otolaryngol
Clin North Am 1996;29:761– 82.
36. Nichols AW. Nonorthopaedic problems in the aquatic athlete. Clin
Sport Med 1999;18:395– 411.
37. Raymond L, Spaur WH, Thalmann ED. Prevention of divers’ ear. Br
Med J 1978;1:48.
38. Sander R. Otitis externa: a practical guide to treatment and prevention. Am Fam Physician 2001;63:927–36, 941–2.
39. Hannley MT, Denneny JC III, Holtzer SS. Consensus panel report:
use of ototopical antibiotics in treating 3 common ear diseases.
Otolaryngol Head Neck Surg 2000;122:934 – 40.
40. Shikiar R, Halpern MT, McGann M, et al. The relation of patient
satisfaction with treatment of otitis externa to clinical outcomes:
development of an instrument. Clin Therapeutics 1999;21:1091– 4.
41. Rosenfeld RM. Meta-analysis. ORL 2004;66:186 –95.
42. Kaushik V, Malik T, Saeede SR. Interventions for otitis externa
(protocol). Cochrane Database Syst Rev 2004;2:CD004740.
43. Dickersin K, Scherer R, Lefebvre C. Systematic reviews: identifying
relevant studies of systematic reviews. BMJ 1994;309:1286 –91.
44. Chow TKF, To E, Goodchild CS, et al. A simple, fast, easy method
to identify the evidence base in pain-relief research: validation of a
computer strategy used alone to identify quality randomized controlled trials. Anesth Analg 2004;98:1557– 65.
45. Jadad AR, Moore A, Carroll D, et al. Assessing the quality of reports
of randomized clinical trials: is blinding necessary? Control Clin
Trials 1996;17:1–12.
46. Shiffman RN, Shekelle P, Overhage JM, et al. Standardized reporting
of clinical practice guidelines: a proposal from the conference on
guideline standardization. Ann Intern Med 2003;139:493– 8.
47. Choudhry NK, Stelfox HT, Detsky AS. Relationships between authors of clinical practice guidelines and the pharmaceutical industry.
JAMA 2002;287:612–7.
48. Shiffman RN, Karras BT, Agrawal A, et al. GEM: a proposal for a
more comprehensive guideline document model using XML. J Am
Med Informatics Assoc 2000;7:488 –98.
49. AAP SCQIM (American Academy of Pediatrics Steering Committee
on Quality Improvement and Management). Policy Statement. Classifying recommendations for clinical practice guidelines. Pediatrics
2004;114:874 –7.
S21
50. Eddy DM. A manual for assessing health practices and designing
practice policies: the explicit approach. Philadelphia: American College of Physicians, 1992.
51. Lucente FE, Lawson W, Novick NL. External ear. Philadelphia: WB
Saunders Co; 1995.
52. Lieberthal AS, Ganiats TG, Cox EO, et al. Clinical practice guideline:
American Academy of Pediatrics Subcommittee on Management of
Acute Otitis Media: diagnosis and management of acute otitis media.
Pediatrics 2004;113:1451– 65.
53. Peltonen L. Nickel sensitivity: an actual problem. Int J Dermatol
1981;20:352–3.
54. Rudner EF, Clendenning WE, Epstein E. Epidemiology of contact
dermatitis in North America: 1972. Arch Dermatol 1973;108:537–
40.
55. Larsson-Styme B, Widstrom L. Ear piercing: a cause of nickel allergy
in schoolgirls? Contact Dermatitis 1985;13:268 –93.
56. Meding B, Ringdahl A. Allergic contact dermatitis from the earmoulds of hearing aids. Ear Hear 1992;13:122– 4.
57. Cockerill D. Allergies to ear moulds. Br J Audiol 1987;21:143–5.
58. Smith IM, Kaey DG, Buxton PK. Chronic hypersensitivity in patients
with chronic otitis externa. Clin Otolaryngol 1990;15:155– 8.
59. Schapowal A. Contact dermatitis to antibiotic ear drops is due to
neomycin but not to ciprofloxacin [abstract]. Allergy 2001;56(suppl
68):148.
60. Kuhweide R, Van de Steene V, Vlaminck S et al. Ramsay Hunt
syndrome: pathophysiology of cochleovestibular symptoms. J Laryngol Otol 2002;116:844 – 8.
61. Rubin Grandis J, Branstetter BF 4th, Yu VL. The changing face of
malignant (necrotizing) external otitis: clinical, radiological, and anatomic correlations. Lancet Infect Dis 2004;4:34 –9.
62. Ismail H. Hellier WP, Batty V. Use of magnetic resonance imaging
as the primary imaging modality in the diagnosis and follow-up of
malignant external otitis. J Laryngol Otol 2004;18:576 –9.
63. Kaur R. Mittal N, Kakkar M, et al. Otomycosis: a clinicomycologic
study. Ear Nose Throat J 2000;79:606 –9.
64. Ruckenstein MJ. Infections of the external ear. In Cummings CW Jr
(ed). Otolaryngology: Head and Neck Surgery, 4th ed. Philadelphia:
Mosby; 2005: p. 2979-87.
65. Jereczek-Fossa BA, Zarowski A, Milani F, et al. Radiotherapy-induced ear toxicity. Cancer Treat Rev 2003;29:417–30.
66. Schechter N L, Berde CM, Yaster M, eds. Pain in infants, children,
and adolescents. Baltimore, MD: Williams and Wilkins; 1993.
67. Joint Commission on Accreditation of Health Care Organizations.
Pain: current understanding of assessment, management and treatments. National Pharmaceutical Council & JCAHO, 2001. Accessed
8/22/2005 at: www.JCAHO.org/.
68. American Academy of Pediatrics/American Pain Society. The assessment and management of acute pain in infants, children, and adolescents. Pediatrics 2001;108:793–7.
69. Bieri D, Reeve RA, Champion G D, et al The Faces Pain Scale for the
self-assessment of the severity of pain experienced by children:
development, initial validation, and preliminary investigation for
ratio scale properties. Pain 1990;41:139 –50.
70. Beyer JE, Knott CB. Construct validity estimation for the AfricanAmerican and Hispanic versions of the Oucher scale. J Pediatr Nurs
1998;13:20 –31.
71. Powell CV, Kelly A M, Williams A. Determining the minimum
clinically significant difference in visual analog pain score for children. Ann Emerg Med 2001;37:28 –31.
72. Loesser JD, ed. Bonica’s management of pain, 3rd ed. Baltimore,
MD: Lippincott Williams and Wilkins, 2001.
73. Valencia CG, Valencia PG. Potassium diclofenac vs placebo in acute
otitis externa: a double-blind, comparative study [Spanish]. Invest
Med Int 1987;14:56 – 60.
74. American Academy of Pediatrics. Report of the subcommittee on the
management of pain associated with procedures in children with
cancer. Pediatrics 1990;86:826 –31.
S22
Otolaryngology–Head and Neck Surgery, Vol 134, No 4S, April 2006
75. Premachandra DJ. Use of EMLA cream as an analgesic in the
management of painful otitis externa. J Laryngol Otol 1990;104:
887– 8.
76. van Balen FAM, Smit WM, Zuithoff NPA, et al. Clinical efficacy of
three common treatments in acute otitis externa in primary care:
randomised controlled trial. BMJ 2003; 327:1201–3.
77. Slack RWT. A study of three preparations in the treatment of otitis
externa. J Laryngol Otol 1987;101:533–5.
78. Psifidis A, Nikolaidis P, Tsona A, et al. The efficacy and safety of
local ciprofloxacin in patients with external otitis : a randomized
comparative study. Mediterranean J Otol Audiol 2005; 1. Accessed
7/27/2005 at: www.mediotol.org/mjo.htm.
79. Zikk D, Rapoport Y, Redianu C, et al. Oral ofloxacin therapy for
invasive external otitis. Ann Otol Rhinol Laryngol 1991;100:632–7.
80. Doern GV. Antimicrobial resistance with Streptococcus pneumoniae
in the United States. Semin Respir Crit Care Med 2000;21:273– 84.
81. Schrag SJ, McGee L, Whitney CG, et al. Emergence of Streptococcus
pneumoniae with very-high-level resistance to penicillin. Antimicrob
Agents Chemother 2004;48:3016 –23.
82. Pottumarthy S, Fritsche TR, Sader HS, et al. Susceptibility patterns of
Streptococcus pneumoniae isolates in North America (2002-2003):
contemporary in vitro activities of amoxicillin/clavulanate and 15
other antimicrobial agents. Int J Antimicrob Agents 2005;25:282–9.
83. Cannon SJ, Grunwaldt E. Treatment of otitis externa with a topical
steroid-antibiotic combination: a controlled clinical trial. Eye Ear
Nose Throat Monthly 1967;46:1296 –302.
84. Cannon S. External otitis: controlled therapeutic trial. Eye Ear Nose
Throat Monthly 1970;49:186 –9.
85. Freedman R. Versus placebo in treatment of acute otitis externa. Ear
Nose Throat J 1978;57:198 –204.
86. Pedersen CB, Osterhammel D. Otitis externa treated with locacortenvioform ear drops [Danish]. Ugeskrift for Laeger 1971;133:389 –91.
87. Yelland MJ. The efficacy of oral cotrimoxazole in the treatment of
otitis externa in general practice. Med J Aust 1993;158:697–9.
88. Kime CE, Ordonez GE, Updegraff WR, et al. Effective treatment of
acute diffuse otitis externa: II. a controlled comparison of hydrocortisone-acetic acid, nonaqueous and hydrocortisone-neomycin-colistin
otic solutions. Curr Ther Res Clin Exp 1978;23(suppl 5):ss15–ss28.
89. Ordonez GE, Kime CE, Updegraff WR, et al. Effective treatment of
acute diffuse otitis externa: I. a controlled comparison of hydrocortisone-acetic acid, non-aqueous and hydrocortisone-neomycin-polymyxin B otic solutions. Curr Ther Res Clin Exp 1978;23(suppl
5):ss3–ss14.
90. Clayton MI, Osborne JE, Rutherford D, et al. A double-blind, randomized, prospective trial of a topical antiseptic versus a topical
antibiotic in the treatment of otorrhoea. Clin Otolaryngol Allied Sci
1990;15:7–10.
91. Lambert IJ. A comparison of the treatment of otitis externa with
Otosporin and aluminium acetate: a report from a services practice in
Cyprus. J Royal Col Gen Pract 1981;31:291– 4.
92. Smathers CR. Chemical treatment of external otitis. South Med J
1977;70:543–5.
93. van Hasselt P, Gudde H. Randomized controlled trial on the treatment
of otitis externa with one percent silver nitrate gel. J Laryngol Otol
2004;118:93– 6.
94. Neher A, Nagl M, Appenroth E, et al. Acute otitis externa: efficacy
and tolerability of N-chlorotaurine, a novel endogenous antiseptic
agent. Laryngoscope 2004;114:850 – 4.
95. Ruth M, Ekstrom T, Aberg B, et al. A clinical comparison of hydrocortisone butyrate with oxytetracycline/hydrocortisone acetate-polymyxin B in the local treatment of acute external otitis. Eur Arch
Otorhinolaryngol 1990;247:77– 80.
96. Tsikoudas A, Jasser P, England RJ. Are topical antibiotics necessary
in the management of otitis externa? Clin Otolaryngol Allied Sci
2002;27:260 –2.
97. Emgard P, Hellstrom S. A group III steroid solution without antibiotic components: an effective cure for external otitis. J Laryngol Otol
2005;119:342–7.
98. Bak JP, Wagenfeld DJ. Treatment of otitis externa with miconazole
nitrate: a comparative study involving 85 cases. S Afr Med J 1983;
63:562–3.
99. Weber PC, Roland PS, Hannley M, et al. The development of antibiotic resistant organisms with the use of ototopical medications.
Otolaryngol Head Neck Surg 2004;130(suppl):S89 –94.
100. Walshe P, Rowley H, Timon C. A worrying development of otitis
externa. Clin Otolaryngol 2001;26:218 –20.
101. Cantrell HF, Lumbardy CE, Duncanson FP, et al. Declining susceptibility to neomycin and polymyxin B of pathogens in otitis externa
in clinical trials. So Med J 2004;95:465–71.
102. Schelkun PH. Swimmer’s ear: getting patients back in the water.
Physician Sportsmed 1991;19:85–90.
103. Eichel BS. How I manage external otitis in competitive swimmers.
Physician Sportsmed 1986;14:108 –16.
104. Fairbanks DNF. Pocket guide to antimicrobial therapy in otolaryngology– head and neck surgery, 13th edition. Alexandria, VA: American Academy of Otolaryngology–Head and Neck Surgery Foundation; 2005.
105. Arnes E, Dibb WL. Otitis externa: clinical comparison of local
ciprofloxacin versus local oxytetracycline, polymyxin B, hydrocortisone treatment. Curr Med Res Opin 1993;13:182– 6.
106. Goldenberg D, Golz A, Netzer A, et al. The use of otic powder in
the treatment of acute external otitis. Am J Otolaryngol 2002;23:
142–7.
107. Roland PS, Pien FD, Schultz CC, et al. Efficacy and safety of
topical ciprofloxacin/dexamethasone versus neomycin/polymyxin
B/hydrocortisone for otitis externa. Curr Med Res Opin 2004;20:
1175– 83.
108. Sabater F, Maristany M, Mensa J, et al. Prospective double-blind
randomized study of the efficacy and tolerance of topical ciprofloxacin vs topical gentamicin in the treatment of simple chronic otitis
media and diffuse external otitis [Spanish]. Acta Otorrinolaryngol
Esp 1996;47:217–20.
109. Rosenfeld RM, Singer M, Wasserman JM, et al. System review of
topical antimicrobial therapy for acute otitis externa. Otolaryngol
Head Neck Surg 2006;134/4S:S24 –S48.
110. Jackman A, Ward R, April M, et al. Topical antibiotic induced
otomycosis. Int J Pediatr Otorhinolaryngol 2005;69:857– 60.
111. Fraki JE, Kalimo K, Tuohimaa P, et al. Contact allergy to various
components of topical preparations for treatment of external otitis.
Acta Otolaryngol 1985;100:414 – 8.
112. Van Ginkel CJ, Bruintjes TD, Huizing EH. Allergy due to topical
medications in chronic otitis externa and chronic otitis media. Clin
Otolaryngol 1995;20:326 – 8.
113. Hillen U, Geier J, Goos M. Contact allergies in patients with eczema
of the external ear canal [German]. Hautarzt 2000;51:239 – 43.
114. Wilkinson SM, Beck MH. Hypersensitivity to topical corticosteroids
in otitis externa. J Laryngol Otol 1993;107:597–9.
115. Yariktas M, Yildirim M, Doner F, et al. Allergic contact dermatitis
prevalence in patients with eczematous external otitis. Asian Pac J
Allergy Immunol 2004;22:7–10.
116. Torum B, Block SL, Avila H, et al. Efficacy of ofloxacin otic solution
once daily for 7 days in the treatment of otitis externa: a multicenter,
open-label, phase III trial. Clin Ther 2004;26:1046 –54.
117. Farnan TB, McCallum J, Awa A, et al. Tea tree oil: in vitro efficacy
in otitis externa. J Laryngol Otol 2005;119:198 –201.
118. Blakley BW. Coning candles–an alert for otolaryngologists? Ear
Nose Throat J 1996;75:585,588.
119. Seely DR, Quigley SM, Langman AW. Ear candles: efficacy and
safety. Laryngoscope 1996;106:1226 –9.
120. England RJ, Homer JJ, Jasser P, et al. Accuracy of patient
self-medication with topical eardrops. J Laryngol Otol 2000;114:24 –5.
121. Agius AM, Reid AP, Hamilton C. Patient compliance with short-term
topical aural antibiotic therapy. Clin Otolaryngol 1994;19:138 – 41.
122. Rubin J, Yu YL. Malignant external otitis: insight into pathogenesis,
clinical manifestations, diagnosis and therapy. Am J Med 1988;85:
391– 8.
Rosenfeld et al
Clinical Practice Guideline: Acute Otitis Externa
123. Ford GR, Courteney-Harris RG. Another hazard of ear syringing:
malignant external otitis. J Laryngol Otol 1990;104:709 –10.
124. Zikk D, Rapoport Y, Himelfarb MZ. Invasive external otitis after
removal of impacted cerumen by irrigation. N Engl J Med 1991;325:
969 –70.
125. Pond F, McCarthy D, O’Leary S. Randomized trial on the treatment
of oedematous acute otitis externa using ear wicks or ribbon gauze:
clinical outcome and cost. J Laryngol Otol 2002;116:415–9.
126. Jinn TH, Kim PD, Russel PT, et al. Determination of ototoxicity of
common otic drops using isolated cochlear outer hair cells. Laryngoscope 2001;111:2105– 8.
127. Russell PT, Church CA, Hinn TH, et al. Effects of common topical
otic preparations on the morphology of isolated cochlear outer hair
cells. Acta Otolaryngol 2001;121:135–9.
128. Roland PS, Rybak L, Hannley M, et al. Animal ototoxicity of topical
antibiotics and the relevance to clinical treatment of human subjects.
Otolaryngol Head Neck Surg 2004;130(suppl 3):s57–s78.
129. Rakover Y, Keywan K, Rosen G. Safety of topical ear drops containing ototoxic antibiotics. J Otolaryngol 1997;26:194 – 6.
S23
130. Abello P, Vinas JB, Vega J. Topical ototoxicity: review over a 6-year
period [Spanish]. Acta Otorrinolaringol Esp 1998;49:353– 6.
131. Linder TE, Zwicky S, Brandle P. Ototoxicity of ear drops: a clinical
perspective. Am J Otol 1995;16:653–7.
132. Roland PS, Stewart MG, Hannley M, et al. Consensus panel on role
of potentially ototoxic antibiotics for topical middle-ear use: introduction, methodology, and recommendations. Otolaryngol Head
Neck Surgery 2004;130(suppl 3):s51–s56.
133. Monarch Pharmaceuticals. Cortisporin Otic Suspension Sterile package insert. Bristol, TN: Monach Pharmaceuticals, Inc: 2003.
134. Sood S, Strachan DR, Tsikoudis A, et al. Allergic otitis externa. Clin
Otolaryngol Allied Sci 2002;27:233–36.
135. Marzo SJ, Leonetti JP. Invasive fungal and bacterial infections of the
temporal bone. Laryngoscope 2003;113:1503–7.
136. Devos SA, Mulder JJ, van der Valk PG. The relevance of positive
patch test reactions in chronic otitis externa. Contact Dermatitis
2000;42:354 –5.
137. Rutka J. Acute otitis externa: treatment perspectives. Ear Nose Throat
J 2004;83(suppl 4):20 –2.
Otolaryngology–Head and Neck Surgery (2006) 134, S24-S48
ORIGINAL RESEARCH
Systematic review of topical antimicrobial therapy
for acute otitis externa
Richard M. Rosenfeld, MD, MPH, Michael Singer, MD,
Jared M. Wasserman, MD, and Sandra S. Stinnett, DrPH,
Brooklyn, New York; and Durham, North Carolina
OBJECTIVE: To determine the efficacy of topical antimicrobials for acute otitis externa.
STUDY DESIGN: Systematic review and random effects metaanalysis of randomized, controlled trials with parallel groups permitting one or more of the following comparisons: antimicrobial
vs placebo, antiseptic vs antimicrobial, quinolone antibiotic vs
nonquinolone antibiotic, steroid-antimicrobial vs antimicrobial, or
antimicrobial-steroid vs steroid.
RESULTS: Twenty trials met inclusion criteria and 18 had data
suitable for pooling. Topical antimicrobials increased absolute clinical
cure rates over placebo by 46% (95% confidence interval [CI], 29%
to 63%) and bacteriologic cure rates by 61% (95% CI, 46% to 76%).
No significant differences were noted in clinical cure rates for other
comparisons, except that steroid alone increased cure rates by 20%
compared with steroid plus antibiotic (95% CI, 3% to 38%). Quinolone drops increased bacteriologic cure rates by 8% compared with
nonquinolone antibiotics (95% CI, 1% to 16%), but had statistically
equivalent rates of clinical cure and adverse events.
CONCLUSION: Topical antimicrobial is highly effective for
acute otitis externa with clinical cure rates of 65% to 80%
within 10 days of therapy. Minor differences were noted in
comparative efficacy, but broad confidence limits containing
small effect sizes make these of questionable clinical
significance.
SIGNIFICANCE: Summary estimates from the 13 meta-analyses can be used to facilitate evidence-based management recommendations and clinical practice guideline development.
© 2006 American Academy of Otolaryngology–Head and Neck
Surgery Foundation, Inc. All rights reserved.
From the Department of Otolaryngology (Drs Rosenfeld, Singer, and
Wasserman), State University of New York Downstate Medical Center and
The Long Island College Hospital, Brooklyn, NY, and the Department of
Biostatistics and Bioinformatics (Dr Stinnett), Duke University, Durham, NC.
Dr. Rosenfeld is a prior consultant for Alcon and Daiichi Pharmaceuticals.
A
cute otitis externa (AOE) is a diffuse inflammation of
the external ear canal. Also known as “swimmer’s ear”
or “tropical ear,” AOE has a lifetime incidence of 10% and
accounts for 7.5 million annual ototopical prescriptions in
the United States. The direct costs of treating AOE include
physician visits, analgesics, and systemic medications, such
as antibiotics, steroids, or both. Indirect costs have not been
calculated, but are likely to be substantial because of severe
otalgia that limits activities.1
There is no consensus with respect to optimal management of AOE. Despite the efficacy of topical therapy, systemic antibiotics are often prescribed inappropriately.2
When topical therapy is prescribed, confusion exists about
whether to use an antiseptic, antibiotic, corticosteroid, or a
combination product.3 The selection of an antibiotic creates
additional controversy, particularly with respect to the role
of newer quinolone drops. No systematic reviews of relevant clinical trials exist to facilitate decision making, although one protocol had been published and was in progress
at the time of this writing (December 2005).4
Systematic review is a form of literature review in which
studies are systematically assembled, appraised, and combined with the use of explicit and reproducible methods to
reduce bias.5 We performed a systematic review of topical
antimicrobial therapy for AOE as part of a multidisciplinary, evidence-based, clinical practice guideline created
by the American Academy of Otolaryngology–Head and
Neck Surgery Foundation (AAO-HNSF).1 Our goal was to
identify relevant randomized controlled trials (RCTs) and
Reprint requests: Richard M. Rosenfeld, MD, MPH, Department of
Otolaryngology, 339 Hicks Street, Brooklyn, NY 11201-5514.
E-mail address: [email protected].
0194-5998/$32.00 © 2006 American Academy of Otolaryngology–Head and Neck Surgery Foundation, Inc. All rights reserved.
doi:10.1016/j.otohns.2006.02.013
Rosenfeld et al
Systematic Review of Topical Antimicrobial . . .
Table 1
MEDLINE search strategy for identifying acute otitis
externa articles*
Number
Search history
Results
1
2
3
4
5
6
7
8
9
exp otitis externa/
exp otitis/
otitis.tw.
inflamm$.tw.
infect$.tw.
or/2-5
exp ear, external/
(extern$ adj1 ear).tw.
(extern$ adj1 auditory adj
canal).tw.
pinna$.tw.
or/7-10
6 and 11
“swimmer’s ear.”tw.
or/1,12-13
1,436
17,141
12,768
243,665
663,353
884,321
10,615
1,609
1,276
10
11
12
13
14
1,622
12,882
1,844
10
2,860
*Search performed 7/27/2005 with strategy adapted from
Kaushik et al.4
derive summary estimates of effect size by statistically
pooling data from similar studies. The estimates of effect
size would be used in developing major guideline statements for managing AOE.6
S25
nonquinolone antibiotic; steroid-antimicrobial vs antimicrobial; and antimicrobial-steroid vs steroid.
Translators assisted in analyzing 5 articles in Spanish,
Italian, Russian, and Danish. Quality of the remaining 20
articles10-29 (Fig 1) was assessed with the Jadad scale,
which awards a maximum of 5 points based on randomization, masking, withdrawals, and dropouts.30 The Jadad scale
was used for descriptive purposes and to see how study
results varied by quality score.
The 2 independent reviewers (MS, JMW) abstracted data
for the 20 articles with a standardized form. Descriptive
information included the definition of AOE, inclusion criteria, exclusion criteria, sample demographics, frequency
and method of aural toilet, use of a wick for drug delivery,
and patient compliance with treatment. Quantitative information included number of subjects at trial start, number of
withdrawals by group, adverse events by group, clinical
outcomes by group, and bacteriologic outcomes by group.
Any disagreement or inconsistency among data obtained by
the reviewers for a given article was resolved after mutual
discussion with a third reviewer (RMR).
Clinical outcomes were defined as “cured” (absence of
all presenting signs and symptoms of diffuse AOE) or
“improved” (partial or complete relief of presenting signs
METHODS
A systematic review was performed with an a priori protocol to reduce bias and to maximize the quality of reported
results.7,8 An electronic MEDLINE search (Table 1) from
1966 through July 2005 for AOE articles was performed
with the use of a search strategy adapted from a Cochrane
protocol.4 The resulting set of 2860 articles was limited to
509 with a maximally sensitive strategy to find clinical trials
suitable for meta-analysis.9 Electronic search of the Cochrane Registry of Clinical Trials with “otitis externa OR
external otitis OR swimmer’s ear” identified 64 studies, of
which 7 were unique. CINAHL search from 1982 through
July 2005 did not identify any unique studies. Titles and
abstracts of the initial data set were scanned for parallel
group RCTs of topical therapy for diffuse AOE. Articles
were excluded if they were not about a clinical trial, had
only a single treatment group, or dealt with otorrhea caused
by conditions other than diffuse AOE (eg, otomycosis, tympanostomy tube otorrhea, middle-ear disease, eczematous or
malignant otitis externa). The remaining initial data set
contained 43 articles.
Two independent reviewers (MS, JMW) assessed the
initial data set for articles that were limited to diffuse AOE
(or had subgroup data for diffuse AOE) and had 2 or more
parallel treatment groups that permitted 1 or more of the
following topical drug comparisons: antimicrobial vs placebo; antiseptic vs antimicrobial; quinolone antibiotic vs
Figure 1 Flowchart of article selection for the final data set
used in meta-analysis.
S26
Otolaryngology–Head and Neck Surgery, Vol 134, No 4S, April 2006
and symptoms). Binary outcomes were emphasized (eg,
cured vs not cured, improved vs not improved, bacteriologic
cure vs failure), but continuous outcomes were recorded
when present (eg, mean days of otorrhea, mean days of
otalgia). Clinical binary outcomes were recorded by time
point, and were anticipated to be roughly combinable as
early response at 3 to 4 days, end-of-therapy response at 7
to 13 days, and test-of-cure at 14 to 21 days. Final time
points for the combined analysis would be determined based
on data availability for each specific comparison. Bacteriologic response was not recorded by time point because no
study gave more than 1 outcome.
Data from individual studies were combined (pooled)
whenever results were available from 2 or more source
articles for a particular endpoint and outcome time. Although we initially planned on using the intent-to-treat
(ITT) denominator in all analyses, we instead used the
per-protocol denominator because only 4 studies17,22,24,29
reported ITT data. The unit of analysis was patients, but 2
antiseptic studies18,21 reporting outcomes by ears were included because 90% or more of subjects had unilateral
AOE. Two other studies15,28 could not be combined because
they reported only time to symptom resolution and did not
report any binary clinical or bacteriologic outcomes.
If a study contained more than 2 parallel treatment
groups, only the 2 groups most relevant to the hypothesis
being tested were used. For example, a study29 with the
treatment groups (A) acetic acid, (B) acetic acid ⫹ triamcinolone, and (C) neomycin ⫹ polymyxin B ⫹ dexamethasone could be used to test the hypotheses antiseptic vs
antimicrobial (B vs C) and steroid-antimicrobial vs antimicrobial (B vs A). The first comparison is made with groups
B vs C instead of B vs. A, because both groups contain a
steroid. Even though the steroid is not exactly the same in
both groups, this is more relevant to the hypothesis being
tested than a comparison in which only 1 group had a
steroid.
Statistical pooling was done with a random-effects
model of meta-analysis that assumes a population (distribution) of true effect sizes with each source article representing 1 member of this population.31 Under this model, results
are expected to vary from study to study, with differences
caused by experimental error and differences in populations
(between-study variability). Because of this additional variability, the 95% confidence interval (CI) for the pooled
result is wider (less precise) than for a fixed-effect model.
Statistical analysis was performed with the Comprehensive Meta-Analysis,32 which weights study results by the
inverse of variance and calculates a random effects estimate
of the combined effect and 95% CI. A test of heterogeneity
is performed with the Q statistic to evaluate constancy of
effect across strata. Significant heterogeneity exists if P ⬍
0.05, although the test has low power and important variations may be present even with a nonsignificant result.33 For
this reason, the random effects model is used regardless of
the test of heterogeneity, although test results are still stated
and explored. Lastly, the Q statistic was used to calculate I2,
which ranges from 0% to 100% and describes the percentage of total variation across studies caused by heterogeneity
(25% is low, 50% moderate, 75% high).34
RESULTS
Descriptive characteristics of the 20 randomized trials included in the final data set are summarized in Table 2. Year
of publication ranged from 1967 to 2005, with 50% of
studies published after 1994 and 25% in 2002 or later. As
noted above, 1 study23 was reported in abstract form only,
and 2 studies15,28 met inclusion criteria but did not contain
data suitable for statistical pooling.
Most studies were conducted in the United States (40%),
United Kingdom (12%), or Sweden (10%). Other countries
with 1 study each were Norway, Israel, Cyprus, Austria,
Greece, Spain, and The Netherlands. The study setting was
specialist office (otolaryngologist) for 12 (60%), primary
care for 3 (15%), both locations for 3 (15%), and not
specified for 2 studies. Most studies (70%) did not mention
financial support, but 4 (20%) were supported by industry
and 2 (10%) had nonindustry funding. Funded studies were
more likely to be published recently (P ⫽ 0.005) and to
have a larger sample size (P ⫽ 0.05).
Sample size ranged from 28 to 842 subjects, with a
median of 79 and interquartile range of 123. There was no
correlation between year of publication and sample size.
Most studies (55%) contained children and adults, but none
were limited to only children. One study17 reported outcomes separately for children and adults, but we combined
these during analysis for consistency with other studies.
When children were included, the minimum age was 1 year
in 3 studies, 2 years in 2 studies, 4 years in 2 studies, and 6
years or older in 4 studies.
Disease definition criteria for included trials are listed in
Table 3. Almost 50% of studies, including some that were
recently published, did not explicitly define AOE. The maximum allowable duration of signs and symptoms was stated
in 5 (20%) studies, ranging from 2 to 4 weeks. Inclusion
criteria tended to be brief, but exclusion criteria were explicit and extensive in many studies. Most studies (55%)
excluded patients with a tympanic membrane perforation or
nonintact tympanic membrane.
Methodologic characteristics of included trials are shown
in Table 4. A double-blind protocol was used in 50% of
studies overall, but in none of the 4 industry-funded studies.
The Jadad quality scores ranged from 1 (lowest) to 5 (highest), with a median of 2.5. A low rating (score of 2 or less)
was achieved in 50% of studies, and only 2 studies achieved
the highest rating of 5. There was no correlation between
Jadad score and year of publication.
Aural toilet was explicitly mentioned in 9 trials, wicks
were placed in 3 trials, and 8 trials did not mention any
specific measures to clean the ear canal or facilitate drug
Rosenfeld et al
Systematic Review of Topical Antimicrobial . . .
S27
Table 2
Descriptive characteristics of randomized trials included in the final data set
Author year
Arnes and Dibb10
1993
Cannon and
Grunwaldt11 1967
Cannon12 1970
Clayton et al13 1990
Emgard and
Hellstrom14 2005
Freedman15 1978
Goldenberg et al16
2002
Jones et al17 1997
Kime et al18 1978
Lambert19 1981
Neher et al20 2004
Ordonez et al21 1978
Pistorius et al22 1999
Psifidis et al23 2005
Roland et al24 2004
Ruth et al25 1990
Sabater et al26 1996
Slack27 1987
Tsikoudas et al28 2002
van Balen et al29 2003
Country
Setting†
Funding
N
Age, y
Groups used in meta-analyses
Norway
2
None
30
ⱖ18
Cipro vs oxytet/polymyx/HC
USA
2
None
40
2-68
Neo/methylpred vs placebo
USA
UK
Sweden
2
2
2
None
None
Industry
43
66
51
ⱖ4
NS
19-67
Acetic/glyceryl vs placebo
Alum-acetate vs gentamicin
Betamethasone vs oxytet/polymyx/HC
USA
Israel
2
2
None
None
91
120
4-76
18-52
Neo/colistin/HC vs placebo
Cipro vs tobramycin
Industry
None
None
Nonindustry
None
Industry
NS
Industry
None
None
None
None
Nonindustry
601
102
126
50
181
842
91
468
53
54
28
39
213
ⱖ1
6-73
1-44
8-89
10-82
2-85
22-61
1-90
11-74
ⱖ18
NS
ⱖ18
ⱖ18
Oflox vs neo/polymyx/HC
Acetic/HC vs neo/colistin/HC
Alum-acetate vs neo/polymyx/HC
NCT vs neo/polymyx/HC
Acetic/HC vs neo/polymyx/HC
Cipro vs cipro/HC vs neo/polymyx/HC
Cipro vs cipro/dex vs neo/polymyx/HC
Cipro/dex vs neo/polymyx/HC
HC butyrate vs oxytet/polymyx/HC
Cipro vs gentamicin
Boric/ethyl vs neo/polymyx/HC
Betamethasone vs betamethasone/neo
Acetic vs acetic/triamcin vs neo/
polymyx/dex
USA
USA
Cyprus
Austria
USA
USA
Greece
USA
Sweden
Spain
UK
UK
Netherlands
1
NS
1
2
NS
3
3
2
3
2
2
2
1
Acetic, acetic acid 2%; alum-acetate, aluminum acetate 8%; boric, boric acid 4%; cipro, ciprofloxacin; dex, dexamethasone; ethyl,
ethyl alcohol 25%; HC, hydrocortisone; glyceryl, glyceryl acetate 88%; methylpred, methylprednisolone; NCT, N-chlorotaurine
(antiseptic); neo, neomycin; NS, not stated; oflox, ofloxacin; oxytet, oxytetracycline; polymyx, polymyxin B; triamcin, triamcinolone
†1, Primary care; 2, specialist; 3, primary care and specialist.
delivery. Most trials (55%) did not assess adherence to
therapy. All trials except 1 administered topical therapy for
at least 7 days, and most (55%) advised 7 to 11 days of
treatment. Only 2 trials reported less than 70% follow-up,
but follow-up was not mentioned in 4 trials.
Data were pooled from 18 randomized trials to perform
13 meta-analyses in 5 categories; these are summarized in
Table 5. A complete listing of all data abstracted from
included studies is provided in the Appendix. All analyses
were performed with patients as the unit analysis, except for
number 2 in Table 5 (bacteriologic cure, antimicrobial vs
placebo). Two studies18,21 in the antiseptic vs antibiotic
comparisons (numbers 3 and 4 in Table 5) also reported
outcomes by ears, but were included in the analyses because
less than 10% of patients had bilateral AOE.
est, the homogeneous results (I2 ⫽ 0%) and large effect size
support a clinically significant effect. The 95% CI for the
clinical cure rate is consistent with a number needed to treat
(NNT) of 1.5 to 3.5 patients. The study by Cannon12 included
5 cases of AOE that recurred within 9 months and 1 patient
with bilateral AOE.
Freedman15 compared topical neomycin/colistin/hydrocortisone to topical placebo for AOE and found less severe
edema and itching with active therapy at day 3 (P ⬍ 0.05),
and less severe edema, itching, redness, scaling, and weeping at day 7 (P ⬍ 0.05). Results were only reported with
symptom severity scales, without any information on cure
vs failure that could be used in meta-analysis.
Antimicrobial vs Placebo
Topical antiseptic and topical antibiotic achieved comparable clinical cure rates for AOE at 7 to 14 days (Table 5,
numbers 3 and 4; Figs 4 and 5). Most studies (4 of 6) were
double blind, of good quality (4/6 had a Jadad score ⱖ3),
and used aural toilet or placed a wick for drug delivery.
Antiseptics used in the meta-analyses were acetic acid/
hydrocortisone (2 studies), acetic acid/triamcinolone (1),
boric acid/ethyl alcohol (1), aluminum acetate (1), and Nchlorotaurine (1). The antibiotic comparator was neomycin
Topical antimicrobial increased absolute clinical cure rates of
AOE by 46% and bacteriologic cure rates by 61% compared
with placebo (Table 5, numbers 1 and 2; Figs 2 and 3). The 2
studies combined for this analysis had the same first author,
were double blind, of high quality (Jadad score 4), and used
aural toilet. In 1 study, the treatment group received neomycin/
methylprednisolone, and in the other they received acetic acid/
glyceryl triacetate. Although the combined sample size is mod-
Antiseptic vs Antibiotic
S28
Otolaryngology–Head and Neck Surgery, Vol 134, No 4S, April 2006
Table 3
Disease definition criteria for randomized trials included in the final data set
Author year
Definition of AOE
Inclusion criteria
Arnes and Dibb10
1993
Pain/discharge plus
characteristic
clinical signs
AOE as defined
Cannon and
Grunwaldt11 1967
Cannon12 1970
NS
Clayton et al13 1990
NS
Mild, moderate or severe
otitis externa
Moderate or severe external
otitis with exudate
Otorrhea from otitis externa
Emgard and
Hellstrom14 2005
NS
Total ear canal score of 3.0
or higher (maximum 10.0)
based on ear canal color,
swelling, effusion, and
extent of redness outside
the ear canal
Freedman15 1978
Goldenberg et al16
2002
NS
Painful condition from
secondary infection
of macerated skin
and subcutaneous
tissues of external
auditory canal
Clinically diagnosed
AOE
Clinical diagnosis of AOE
AOE diagnosed by an
otolaryngologist
Kime et al18 1978
Diffuse AOE with
edema, discharge,
pain, pruritus, or
erythema
Primary or recurrent diffuse
AOE
Lambert19 1981
Inflammation of the
ear canal, with or
without discharge,
plus tragal
tenderness
NS
NS
Jones et al17 1997
Neher et al20 2004
NS
Unilateral or bilateral AOE
with purulent or
mucopurulent discharge;
maximum duration of 2
weeks
Otitis externa diagnosed by
otolaryngologist
Exclusion criteria
Systemic antibiotic use, pregnancy,
fungal infection, perforated TM,
history of middle-ear surgery,
allergy to quinolone antibiotic
NS
Mild external otitis without exudate
Topical or systemic antimicrobials
in past 3 weeks, fungal otitis,
drug sensitivity to agents used,
acute middle ear disease,
cholesteatoma
Treated for otitis externa in past 30
days; anti-inflammatory,
antibiotic, or antifungal
medications for concomitant
illness; perforated TM; hearing
aid; need for ear canal packing;
mastoiditis; poor motivation;
psychosis or intellectual
impairment; pregnancy; breast
feeding; planned pregnancy
NS
Prior systemic or topical treatment,
sensitivity to study drug, TM
perforation
TM perforation within 6 months,
chronic otitis externa, seborrheic
dermatitis, fungal infection,
invasive otitis externa,
antimicrobial within 14 days,
quinolone within 30 days, overthe-counter treatment in past 36
hours, allergy to study drugs,
hepatitis, immunosuppression,
pregnant, lactating
Perforated TM, otitis media,
furunculosis, mastoiditis, history
of chronic otitis externa, medical
condition requiring systemic
antibiotic or steroid, known drug
sensitivity, herpes simplex virus,
varicella, of vaccinia
Middle-ear pathology
Malignant otitis externa, topical
treatment with other drugs,
systemic antibiotic or steroid use,
pregnancy, participation in
another study
Rosenfeld et al
Systematic Review of Topical Antimicrobial . . .
S29
Table 3
(Continued)
Author year
Definition of AOE
Inclusion criteria
Ordonez et al21
1978
Diffuse AOE
Primary or recurrent AOE
complicated by
inflammation
Pistorius et al22
1999
Edema of the external
auditory canal,
tenderness with
pinna movement,
otalgia
AOE as defined plus
symptoms within 2 days
of study entry; maximum
duration of 3 weeks
Psifidis et al23 2005
NS
Roland et al24 2004
Diffuse cellulitis and
bacterial infection of
the external
auditory meatus
that may involve
underlying
structures
Maximum duration of 3
weeks
Mild, moderate, or severe
AOE as defined;
maximum duration of 4
weeks
Ruth et al25 1990
NS
Sabater et al26 1996
Otalgia with tragal
tenderness and
otoscopic signs of
otitis externa
(hyperemia or
edema of ear canal,
otorrhea with intact
TM)
NS
Slack27 1987
Tsikoudas et al28
2002
Edematous ear canal
with moist keratin
debris
van Balen et al29
2003
Redness or swelling
of the external
auditory canal or
debris within the
canal plus pain,
itching, otorrhea,
hearing loss, or a
stuffy feeling
Mild or moderate AOE;
maximum duration of 2
weeks
Diffuse otitis externa
Diagnosis of otitis externa
and no treatment for at
least 2 weeks
Clinical diagnosis of otitis
externa
AOE as defined; maximum
duration of 3 weeks
AOE, Acute otitis externa; NS, not specified; TM, tympanic membrane.
Exclusion criteria
Perforated TM, furunculosis,
mastoiditis, chronic otitis externa,
medical conditions requiring
systemic antibiotic or steroid,
drug sensitivity, herpes simplex
virus, vaccinia, varicella
Perforated TM, acute otitis media,
malignant otitis externa,
dermatitis, AOE in past 30 days,
fungal otitis, furuncles,
mastoiditis, stenosis, exostosis,
ear tumors, underlying disease,
neutropenia, HIV,
immunocompromise, pregnancy,
lactation, allergy to study drugs,
previous study enrollment
NS
Nonintact TM, acute or chronic
otitis media, post-tympanostomy
otorrhea, malignant otitis
externa, overt fungal or viral ear
infection, congenital ear
anomalies, obstructive exostoses,
mastoid or other suppurative
non-infectious ear disorders,
seborrheic dermatitis of ear
canal, immunosuppression, renal
failure, hepatitis, diabetes,
pregnancy, lactation, receiving
other otic therapy
Fever, perichondritis, or otitis
externa within preceding month
Pregnant, lactating, allergy to study
drug, renal or hepatic
insufficiency, antibiotic less than
7 days prior to entry, nonintact
TM
Previous mastoid surgery,
perforated TM
Allergy to neomycin, ear canal
edema severe enough to prevent
use of otic drops, middle ear
disease
Otitis externa more than 3 weeks,
furunculosis, acute otitis media,
perforated TM, perichondritis,
fever, allergy to study drug, prior
study entry, treated for otitis
externa in past month, pregnant
S30
Otolaryngology–Head and Neck Surgery, Vol 134, No 4S, April 2006
Table 4
Methodologic characteristics of randomized trials included in the final data set
Author year
Arnes and Dibb10 1993
Cannon and
Grunwaldt11 1967
Cannon12 1970
Clayton et al13 1990
Emgardand
Hellstrom14 2005
Freedman15 1978
Goldenberg et al16
2002
Jones et al17 1997
Kime et al18 1978
Lambert19 1981
Neher et al20 2004
Ordonez et al21 1978
Pistorius et al22 1999
Psifidis et al23 2005
Roland et al24 2004
Ruth et al25 1990
Sabater et al26 1996
Slack27 1987
Tsikoudas et al28 2002
van Balen et al29 2003
Jadad
score†
Aural
toilet
Check for
compliance
Days of
therapy
Outcome
day(s)
Follow-up, %
Single
1
No
No
7
8
NS
Double
Double
Double
4
4
4
Yes
Yes
No
No
No
Yes
10
3-7
21
10
3-7
21
73
100
73
Single
Double
3
2
No
Wick
No
Yes
11
21
4, 11
3, 7, 21
98
100
Single
Single
Double
Single
Single
Double
Single
Single
Single
Single
Double
Double
Double
Double
1
3
3
1
2
4
1
1
2
2
2
3
5
5
No
No
Wick
Yes
Wick
Yes
Yes
No
Yes
No
No
Yes
Yes
Yes
Yes
Yes
Yes
No
No
Yes
Yes
No
Yes
No
No
No
Yes
No
14
10
8-14
14
9
8-14
7
7
7
7
8
7-21
11
7-21
3-4, 14
10
11
14
Daily,1-9
10
7, 14-28
7
3, 8, 18
14
8
7, 14, 21
11
7, 14, 21
NS
79
81
93
100
61
83
NS
85
87
NS
86
67
93
Blinding
NS, not stated
†Quality score ranging from 1 (lowest) to 5 (highest)
in all studies, combined with polymyxin B/hydrocortisone
(4 studies), colistin/hydrocortisone (1), or polymyxin
B/dexamethasone (1). The analyses performed at 7 to 10
days and 11 to 14 days contained homogeneous studies (I2
⬍ 25%), gave consistent results, and did not suggest low
statistical power.
Quinolone Antibiotic vs Nonquinolone
Antibiotic
Topical quinolone antibiotic and topical nonquinolone antibiotic achieved comparable clinical cure rates for AOE at
3 to 4 days, 7 to 10 days, and 14 to 28 days (Table 5,
numbers 5 to 7; Figs 6, 7, and 8) and comparable clinical
improvement rates at 7 to 10 days (Table 5, number 8, Fig
9). All studies but 1 (6 of 7) were single blind, all but 1 had
low methodologic quality (Jadad score ⱕ2), and only 2 used
aural toilet. Quinolones used in the meta-analyses were
ofloxacin (1 study), ciprofloxacin alone (3), or ciprofloxacin
combined with dexamethasone (2) or hydrocortisone (1).
The antibiotic comparators used were gentamicin (1 study),
tobramycin (1), or polymyxin/hydrocortisone combined
with neomycin (3) or oxytetracycline (2). None of the comparisons were statistically significant, and the magnitude of
the rate differences (0.05 to 0.11) suggest small clinical
effects.
Sensitivity analysis was performed to assess the impact
of model choice (random-effects vs fixed-effects) on out-
comes in the analyses with statistically homogeneous studies (Figs 6 and 8). The only analysis for which this affected
statistical significance was for clinical cure at 3 to 4 days
(Fig 6). Using a fixed-effects model, the absolute rate difference (RD) was 0.065 with a 95% CI of 0.009 to 0.122
(P ⫽ 0.024). Although statistically significant, the NNT of
15 patients suggests a small clinical effect. Further, the I2 of
62% (Table 3) indicates moderate heterogeneity (despite a
nonsignificant P value), which makes a random effects
model more appropriate. Clinical cures at 14 to 28 days (Fig
8) remained nonsignificant when reanalyzed with a fixedeffects model.
Topical quinolone therapy increased absolute bacteriologic cure rates of AOE by 8.0% over nonquinolone antibiotic therapy (Table 5, number 9, Fig 10). This result,
however, was highly influenced by 1 study10 with a small
sample size and an RD at least 2 times higher than all other
studies. When this study is excluded from the meta-analysis
in Figure 10, the RD decreases to 0.056 (95% CI, – 0.006,
0.119), which is no longer statistically significant (P ⫽
0.079). The remaining studies are heterogeneous (P ⫽
0.021), but even when a fixed-effects model is used the
pooled RD is not statistically significant without the small
study by Arnes and Dibb.10
Additional sensitivity analysis of bacteriologic cure rates
was performed to assess modifying factors. Industry-funded
studies (n ⫽ 3) had an RD of 0.045 that was significantly
Rosenfeld et al
Systematic Review of Topical Antimicrobial . . .
S31
Table 5
Summary of meta-analyses of topical antimicrobials for treating acute otitis externa
Treatment group vs control group
Outcome: studies combined*
References
combined
N
Control
rate†
RD (95% CI)‡
P value
I2,%¶
Antimicrobial vs. placebo
1. Clinical cure at 3-10 days
2. Bacteriologic cure
11,12
11,12§
89
112§
0.15
0.20
0.46 (0.29, 0.63)
0.61 (0.46, 0.76)
⬍0.001
⬍0.001
0
0
Antiseptic vs. antibiotic
3. Clinical cure at 7-10 days
4. Clinical cure at 11-14 days
20,21,27,29
18,19,27,29
318
368
0.65
0.80
0.05 (⫺0.03, 0.12)
0.04 (⫺0.06, 0.13)
0.217
0.468
0
21
Quinolone vs. nonquinolone
antibiotic(s)
5. Clinical cure at 3-4 days
6. Clinical cure at 7-10 days
7. Clinical cure at 14-28 days
8. Improved at 7-10 days
9. Bacteriologic cure
10. Any adverse event
16,24
10,17,22,23,24,26
16,22,24
10,22,24
10,16,17,22,23,24
17,22,24
476
1475
936
890
980
1330
0.15
0.77
0.83
0.89
0.87
0.15
0.11
0.07
0.04
0.05
0.08
0.002
(⫺0.06, 0.28)
(⫺0.02, 0.16)
(⫺0.01, 0.08)
(⫺0.05, 0.14)
(0.006, 0.16)
(⫺0.07, 0.08)
0.192
0.110
0.145
0.292
0.035
0.963
62
70
40
69
74
72
22,23,29
22,23
660
342
0.68
0.93
0.04 (⫺0.08, 0.16)
⫺0.02 (⫺0.15, 0.11)
0.546
0.761
69
54
14,25
92
0.72
⫺0.20 (⫺0.38⫺0.03)
0.021
0
Antimicrobial/steroid vs
antimicrobial alone
11. Clinical cure at 7 days
12. Bacteriologic cure
Steroid/antibiotic vs steroid alone
13. Clinical cure at 7-11 days
CI, confidence interval; RD, absolute rate difference.
*Refer to Tables 2 and 3 for individual study details; data from the treatment and control groups are displayed in Figs 2-14.
†Control rate is calculated by simple division of total events by total patients to aid in interpreting the RD.
‡Absolute change in outcomes for treatment vs control groups, beyond the control group rate, based on random-effects metaanalysis.
¶Test for heterogeneity among studies combined; ranges from 0% (no heterogeneity) to 100% (maximum heterogeneity).
§Analysis by ears, not patients.
less than nonfunded studies (n ⫽ 3) with an RD of 0.191
(P ⫽ 0.021, ANOVA). One higher quality study (Jadad
score 3 or higher) had an RD of 0.012 that was significantly
less than lower quality studies (n ⫽ 5) with an RD of 0.101
(P ⬍ 0.001, ANOVA). Studies that performed aural toilet (n
⫽ 2) had an RD of 0.079 that was significantly less than
studies without toilet (n ⫽ 4) with an RD of 0.120 (P ⫽
0.007, ANOVA). The 95% CIs were broad for all of these
analyses, reflecting the small number of overall studies
combined.
Three studies that compared adverse events showed no
overall combined difference between a quinolone preparation and neomycin/polymyxin B/hydrocortisone (Table 5,
number 10; Fig 11). The most common events reported
were pruritus (about 7%) and site reaction (5%); other
events with an incidence less than 2% included rash, discomfort, otalgia, dizziness, vertigo, superinfection, and reduced hearing.17,24 In 1 study,22 only 1.1% of patients had
to discontinue topical drops because of infection, nausea, or
vomiting.
Figure 2 Random effects meta-analysis of topical antimicrobial (treated) vs topical placebo (control) for clinical cure at 3 to 10 days of
acute otitis externa. Forest plot shows absolute rate difference and 95% confidence intervals for individual studies (squares proportional to
study weight) and combined result (diamond). The combined result is highly significant (number need to treat of 2) and the confidence
interval suggests good precision. The test for heterogeneity among studies is not significant (P ⫽ 0.96).
S32
Otolaryngology–Head and Neck Surgery, Vol 134, No 4S, April 2006
Figure 3 Random effects meta-analysis of topical antimicrobial (treated) vs topical placebo (control) for bacteriologic cure of acute otitis
externa. The combined absolute rate difference (diamond) is highly significant (number needed to treat 1.6) and the confidence interval
suggests good precision. The test for heterogeneity among studies is not significant (P ⫽ 0.68). Sample size is slightly higher than Figure
2 because the analysis is by ears, not patients.
Antimicrobial/steroid vs Antimicrobial Alone
Topical antimicrobial/steroid and topic antimicrobial alone
achieved comparable clinical and bacteriologic cure rates
for AOE at 7 days (Table 5, numbers 11 and 12; Figs 12 and
13). Most studies (2 of 3) were single blind, of low quality
(2 of 3 had Jadad score of 1), and performed aural toilet (2
of 3). Antimicrobial/steroid combinations used in the metaanalyses were ciprofloxacin/hydrocortisone, ciprofloxacin/
dexamethasone, and acetic acid/triamcinolone. The antibiotic comparator in all studies was the same antimicrobial
without the steroid. The analyses of clinical and bacterio-
logic cure had moderate heterogeneity (I2, 54% to 69%),
gave consistent results, and did not suggest low statistical
power.
Tsikoudas et al28 compared betamethasone/neomycin to
betamethasone alone for AOE and found no difference at
day 11 in symptom scores (P ⫽ 0.16) or observer assessment scores (P ⫽ 0.30). The small sample size (39 patients)
cannot exclude the possibility that a difference was missed
because of low statistical power. Results were reported only
with symptom scores, without any information on cure vs
failure that could be used in meta-analysis.
Figure 4 Random effects meta-analysis of topical antiseptic (treated) vs topical antibiotic (control) for clinical cure at 7 to 10 days of
acute otitis externa. The combined absolute rate difference (diamond) is not significant, and the confidence interval does not suggest a
meaningful effect was missed. The test for heterogeneity among studies is not significant (P ⫽ 0.88).
Figure 5 Random effects meta-analysis of topical antiseptic (treated) vs topical antibiotic (control) for clinical cure at 11 to 14 days of
acute otitis externa. The combined absolute rate difference (diamond) is not significant and the confidence interval does not suggest a
meaningful effect was missed. The test for heterogeneity among studies is not significant (P ⫽ 0.28).
Rosenfeld et al
Systematic Review of Topical Antimicrobial . . .
S33
Figure 6 Random effects meta-analysis of topical quinolone antibiotic (treated) vs topical nonquinolone antibiotic (control) for clinical
cure at 3 to 4 days of acute otitis externa. The combined absolute rate difference (diamond) is not significant, but the upper limit of the
confidence interval (0.28) is clinically important and suggests a meaningful effect may have been missed (low power). The test for
heterogeneity among studies is not significant (P ⫽ 0.11).
Steroid/antibiotic vs Steroid Alone
Topical steroid alone increased absolute clinical cure rates
of AOE by 20% at 7 to 11 days compared with topical
steroid/antibiotic combination therapy (Table 5, number 13;
Fig 14). Studies were single blind (2 of 2), of intermediate
quality (Jadad score 2 to 3), and used aural toilet. Steroids
used in the meta-analyses were betamethasone (1 study) and
hydrocortisone butyrate (1). The antibiotic/steroid comparator was oxytetracycline/polymyxin B/hydrocortisone in
both studies. Although the overall effect is statistically significant, the 95% CI is broad and the lower limit approaches
zero (0.03). Similarly, the 95% CI for the NNT (5 to 33
patients) cannot exclude a trivial effect.
DISCUSSION
Topical antimicrobial therapy is highly effective for AOE
(Table 5), but the choice of topical antimicrobial has a
minimal impact on rates of clinical and bacteriologic cure.
Without treatment only 15% of patients with AOE have
Figure 7 Random effects meta-analysis of topical quinolone antibiotic (treated) vs topical nonquinolone antibiotic (control) for clinical
cure at 7 to 10 days of acute otitis externa. The combined absolute rate difference (diamond) is not significant and the confidence interval
does not suggest a meaningful effect was missed. The test for heterogeneity among studies is significant (P ⬍ 0.01).
Figure 8 Random effects meta-analysis of topical quinolone antibiotic (treated) vs topical nonquinolone antibiotic (control) for clinical
cure at 14 to 28 days of acute otitis externa. The combined absolute rate difference (diamond) is not significant and the confidence interval
does not suggest a meaningful effect was missed. The test for heterogeneity among studies is not significant (P ⫽ 0.19).
S34
Otolaryngology–Head and Neck Surgery, Vol 134, No 4S, April 2006
Figure 9 Random effects meta-analysis of topical quinolone antibiotic (treated) vs topical nonquinolone antibiotic (control) for clinical
improvement at 14 to 28 days of acute otitis externa. Combined result (diamond) indicates no significant impact of antibiotic therapy and
the confidence interval does not suggest a meaningful difference was missed. The test for heterogeneity among studies is significant (P ⫽
0.04).
clinical cure within 10 days, rising to about 65% to 80%
when topical antimicrobial, steroid, or both are administered. Comparable clinical outcomes occur with antiseptic
vs antibiotic, quinolone vs nonquinolone antibiotic, and
antimicrobial vs antimicrobial plus steroid; steroid alone
had better outcomes than steroid plus antibiotic. The incidence of bacteriologic cure tends to exceed the clinical
response, with about 80% to 95% bacteriologic efficacy at
the test of cure visit. Quinolones have slightly better bacteriologic efficacy than nonquinolone antibiotics.
The only clinical comparison with statistical significance
was steroid/antibiotic vs steroid alone (Table 5, number 13;
Fig 14). The reason why antibiotic-containing agents decreased absolute cures by 20% (RD, – 0.20) is unclear;
although in 1 study,14 the steroid used for single-agent
therapy had high potency (betamethasone) and the comparator had low potency (hydrocortisone). Conversely, the second study25 used a low potency steroid (hydrocortisone) in
both groups. The combined analysis with both studies, however, includes only 92 patients and the broad confidence
Figure 10 Random effects meta-analysis of topical quinolone antibiotic (treated) vs topical nonquinolone antibiotic (control) for
bacteriologic cure of acute otitis externa. The combined absolute rate difference (diamond) indicates a significant benefit of quinolone
therapy (number needed to treat of 12), but the lower limit of the confidence interval approaches zero. The test for heterogeneity among
studies is significant (P ⫽ 0.01).
Figure 11 Random effects meta-analysis of topical quinolone antibiotic (treated) vs topical nonquinolone antibiotic (control) for adverse
events when treating acute otitis externa. The combined absolute rate difference (diamond) indicates no difference in adverse event rates.
The test for heterogeneity among studies is significant (P ⫽ 0.03).
Rosenfeld et al
Systematic Review of Topical Antimicrobial . . .
S35
Figure 12 Random effects meta-analysis of topical antimicrobial/steroid combination (treated) vs topical antimicrobial alone (control)
for clinical cure at 7 days of acute otitis externa. The combined absolute rate difference (diamond) is not significant and the confidence
interval does not suggest a meaningful effect was missed. The test for heterogeneity among studies is significant (P ⫽ 0.04).
limits cannot exclude a trivial effect (– 0.03). Additional
studies are needed to confirm this finding and to increase
precision.
Another significant comparison concerns bacteriologic
efficacy (Table 5, number 9; Fig. 10); about 87% of patients
with AOE have bacteriologic cure after nonquinolone therapy, with an 8% absolute increase (95% CI, 1% to 16%)
when a quinolone antibiotic is used. The clinical significance of this modest effect (NNT of 12 patients) is reduced
when considering that persistent bacteria in the ear canal
after treatment does not necessarily imply persistent AOE
symptoms or clinical failure. Generalizability of bacteriologic results is also limited because the analysis includes
only patients with a positive baseline culture who returned
post-treatment. When applied to the broader population of
all patients with AOE not just those with baseline bacterial
growth, the effect size is further reduced.
Although the meta-analysis results suggest minimal or
no difference in cure rates among topical agents, some of
the more recent studies have shown significant differences
in the rapidity of treatment response or symptom resolution.
For example, adding hydrocortisone to acetic acid significantly reduced median ear pain from 8.0 to 7.0 days,29 and
adding hydrocortisone to ciprofloxacin reduced median ear
pain from 4.7 to 3.8 days compared with neomycin/polymyxin B/hydrocortisone (but did not reduce pain compared
with ciprofloxacin alone).22 Similarly, when compared with
neomycin/polymyxin B/hydrocortisone, antiseptic alone
significantly reduced median symptom duration from 11.1
to 9.4 days,19 and mean days of inflammation from 7.4 to
5.6.20
The validity of meta-analysis results depends largely on
quality of the included trials. Trial quality can be graded
with the Jadad score that awards 1 point each for random-
Figure 13 Random effects meta-analysis of topical antimicrobial/steroid combination (treated) vs topical antimicrobial alone (control)
for bacteriologic cure of acute otitis externa. The combined absolute rate difference (diamond) is not significant and the confidence interval
does not suggest a meaningful effect was missed. The test for heterogeneity among studies is not significant (P ⫽ 0.14).
Figure 14 Random effects meta-analysis of topical steroid/antibiotic combination (treated) vs topical steroid alone (control) for clinical
cure at 7 to 11 days of acute otitis externa. The combined absolute rate difference (diamond) indicates a significant benefit of steroid alone
(number needed to treat of 5), but the lower limit of the confidence interval approaches zero. The test for heterogeneity among studies is
not significant (P ⫽ 0.32).
S36
Otolaryngology–Head and Neck Surgery, Vol 134, No 4S, April 2006
ization, double blinding, and a statement to describe withdrawals and dropouts.30 Up to 2 additional points may be
added or subtracted based on the adequacy of randomization
and double blinding. Only 50% of included trials achieved
a Jadad score of 3 or higher, which is the traditional cut-off
for “good” quality in systematic reviews. Similarly, double
blinding was present in 50% of overall studies and only
15% of the quinolone vs nonquinolone comparisons. Our
pooled results most likely overestimate efficacy, because
single blind trials exaggerate effect sizes by 17% and trials
with inadequate or unclear allocation concealment exaggerate effects by 30% to 41%.35
Randomized trials are most valid when they analyze
results by intention-to-treat (ITT), which maintains treatment groups that are similar apart from random variation.36
An ITT analysis includes all patients in the groups to which
they were originally assigned, regardless of whether they
completed treatment, deviated from the protocol, or withdrew from the study. Because only 4 trials17,22,24,29 explicitly stated the ITT denominators, we instead used the perprotocol denominators in the meta-analyses. A per-protocol
analysis, however, can potentially yield a biased treatment
effect.36 Use of the ITT denominators would have reduced
the pooled effect size in all related meta-analyses.
Generalizability is another important issue to consider
when interpreting meta-analysis results. Common exclusion
criteria in the trials were systemic antibiotic use, cellulitis,
perichondritis, furunculosis, fungal infection, herpes simplex virus, nonintact tympanic membrane, middle-ear disease, chronic otitis externa, seborrheic dermatitis, immunosuppression, and malignant otitis externa. Consequently,
results may not apply if 1 or more of these factors exist.
Some of the source articles did not specify exclusion criteria
or mentioned them only briefly (Table 3), which raises the
possibility that patients with some of these conditions (eg,
perforated tympanic membrane, middle ear disease) may
have contributed to the outcomes. The lack of an ITT
analysis also reduces generalizability to “real world” situations where compliance, dosing adequacy, follow-up, and
outcome assessment vary greatly. Last, bacteriologic results
are especially difficult to apply because patients contributing to the analysis represent only 30% to 75% of those
originally enrolled in large studies.17,22,24
Adverse events were infrequent with topical therapy and
rarely required stopping the drug. Only a few studies,17,22,24
however, reported detailed information on adverse events,
which is a major deficit in the source articles. A comparison of
adverse events was possible for the quinolone vs nonquinolone
trials (Table 5, number 10; Fig 11), and the pooled analysis
showed comparable incidence rates. Studies were heterogeneous, however, and one trial showed slightly fewer events
with ciprofloxacin/dexamethasone than neomycin/polymyxin
B/hydrocortisone (RD, – 0.055, number need to harm 18).
Contact dermatitis from neomycin or an aminoglycoside was
not reported in any trials. No hearing loss was reported, but
most trials explicitly excluded patients with a nonintact tym-
panic membrane. If a patient with a tympanic membrane perforation or tympanostomy tube develops AOE, a nonototoxic
drop is preferred for treatment.37
The lack of significant differences in all clinical comparisons in Table 5 (except number 13) raises the possibility of
type II statistical error or low power. All of the “negative”
analyses included over 300 patients, and the upper limit of the
95% CI for the RD exceeds 0.16 in only one comparison
(number 5). An RD of 0.16 equals an NNT of 7.0, which
suggests that at most a very modest clinical effect was missed.
For the comparison of clinical cure at 3 to 4 days for quinolone
vs nonquinolone (Table 5, number 5), the maximum RD consistent with the data is 0.28, which may be clinically meaningful (NNT 4). Additional studies of early clinical outcomes
are needed to increase precision of the RD for this comparison.
Systematic review results are limited by the research
protocol used, quality of the original studies, adequacy of
data reporting, and the impact of study diversity on pooled
estimates.38 We used an a priori protocol with established
methods to minimize bias in selecting articles, abstracting
data, and publishing results.7,8 The quality of many original
studies was low, which may have positively biased outcomes as noted previously. Because all clinical comparisons
except 1 were nonsignificant, we do not consider this a
major problem. All studies except 2 reported data suitable
for statistical pooling,15,28 but many studies reported symptom severity scores in addition to rates of cure vs failure.
Unfortunately, these scores could not be statistically combined because of heterogeneity and incomplete reporting
(eg, no standard deviation). Severity scores included overall
symptoms,12,28 pain,14,15 tenderness,17,24 itching,14,15 discharge,15,24 edema,15,24 redness,15 scaling,15 inflammation,24 insomnia,14 missed work,14 analgesic consumption,14 and patient satisfaction.17
Heterogeneity was found in many of the analyses (8 of 13
in Table 5 had I2 ⬎ 25%), and was partially controlled with the
use of endpoints of cure vs failure at specific time intervals and
the use of a random effects model to combine data. Of most
concern were variations in study quality as reflected by the
Jadad scores. Only 2 trials achieved a maximal quality score,
and only 50% were double blind. Future trials would benefit
from clear inclusion criteria, explicit randomization schemes,
explicit protocols for double blinding, full description of dropouts and withdrawals, and multiple outcome measures that
include clinical response (especially in the early phase of
therapy) and bacteriologic outcomes. Intention to treat analysis
should be used to derive the most valid and generalizable
estimates of effect size. More explicit reporting of adverse
events is also advised. Adhering to the Consolidated Standards
of Reporting Trials (CONSORT) statement would address
these concerns.39
Evidence-based medicine rests on a triad of published
research, provider experience, and patient preference.40 The
largely negative results of our meta-analysis suggest that
decisions that concern choice of topical therapy for AOE
could be based primarily on the latter 2 components, given
Rosenfeld et al
Systematic Review of Topical Antimicrobial . . .
the paucity of high-quality published trials. The absence of
efficacy differences we observed suggests a need for properly conducted and reported randomized trials. Nonetheless,
the estimates of effect size obtained should help future
trialists develop research protocols, and the diversity found
among trials should help illuminate best practices for future
endeavors. Our results can also provide policy makers and
guideline developers with the statistical raw material to
which values and consensus can be added in formulating
evidence-based management recommendations.
We thank Maureen Hannley, PhD, for helping to identify articles and prepare evidence tables.
REFERENCES
1. Rosenfeld RM, Brown L, Cannon CR, et al. AAO-HNSF clinical
practice guideline on acute otitis externa. Otolaryngol Head Neck Surg
2006;134/4S:S4-S23.
2. Hannley MT, Denneny JC III, Holtzer SS. Consensus panel report. Use
of ototopical antibiotics in treating 3 common ear diseases. Otolaryngol Head Neck Surg 2000;122:934 – 40.
3. Hajioff D. Otitis externa. Clin Evid 2004;12:755– 63.
4. Kaushik V, Malik T, Saeede SR. Interventions for otitis externa (protocol). The Cochrane Database of Systematic Reviews. 2004; Issue 2:
Art No: CD004740.
5. Egger M, Smith GD, Phillips AN. Meta-analysis: principles and procedures. BMJ 1997;315:1533–7.
6. Freemantle N, Mason J, Eccles M. Deriving treatment recommendations from evidence within randomized trials: the role and limitations
of meta-analysis. Int J Technol Assess Health Care 1999;15:304 –15.
7. Rosenfeld RM. Meta-analysis. ORL 2004;66:186 –95.
8. Moher D, Cook DJ, Eastwood S, et al. Improving the quality of reports
of meta-analyses of randomized controlled trials: the QUOROM statement. Lancet 1999;354:1896 –900.
9. Dickersin K, Scherer R, Lefebvre C. Systematic reviews: identifying
relevant studies for systematic reviews. BMJ 1994;309:1286 –91.
10. Arnes E, Dibb WL. Otitis externa: clinical comparison of local ciprofloxacin versus local oxytetracycline, polymyxin B, hydrocortisone
treatment. Curr Med Res Opin 1993;13:182– 6.
11. Cannon SJ, Grunwaldt E. Treatment of otitis externa with a topical
steroid-antibiotic combination: a controlled clinical trial. Eye Ear Nose
Throat Monthly 1967;46:1296 –302.
12. Cannon S. External otitis: controlled therapeutic trial. Eye Ear Nose
Throat Monthly 1970;49:186 –9.
13. Clayton MI, Osborne JE, Rutherford D, et al. A double-blind, randomized, prospective trial of a topical antiseptic versus a topical
antibiotic in the treatment of otorrhoea. Clin Otolaryngol Allied Sci
1990;15:7–10.
14. Emgard P, Hellstrom S. A group III steroid solution without antibiotic
components: an effective cure for external otitis. J Laryngol Otol
2005;119:342–7.
15. Freedman R. Versus placebo in treatment of acute otitis externa. Ear
Nose Throat J 1978;57:198 –204.
16. Goldenberg D, Golz A, Netzer A, et al. The use of otic powder in the
treatment of acute external otitis. Am J Otolaryngol 2002;23:142–7.
17. Jones RN, Milazzo J, Seidlin M. Ofloxacin otic solution for treatment
of otitis externa in children and adults. Arch Otolaryngol Head Neck
Surg 1997;123:1193–200.
18. Kime CE, Ordonez GE, Updegraff WR, et al. Effective treatment of
acute diffuse otitis externa: II. A controlled comparison of hydrocortisone-acetic acid, nonaqueous and hydrocortisone-neomycin-colistin
otic solutions. Curr Ther Res Clin Exp 1978; 23(suppl 5):ss15–ss28.
S37
19. Lambert IJ. A comparison of the treatment of otitis externa with
Otosporin and aluminium acetate: a report from a services practice in
Cyprus. J Royal Col Gen Pract 1981;31:291– 4.
20. Neher A, Nagl M, Appenroth E, et al. Acute otitis externa: efficacy and
tolerability of N-chlorotaurine, a novel endogenous antiseptic agent.
Laryngoscope 2004;114:850 – 4.
21. Ordonez GE, Kime CE, Updegraff WR, et al. Effective treatment of acute
diffuse otitis externa: I. A controlled comparison of hydrocortisone-acetic
acid, nonaqueous and hydrocortisone-neomycin-polymyxin B otic solutions. Curr Ther Res Clin Exp 1978;23(suppl 5):ss3–ss14.
22. Pistorius B, Westberry K, Drehobl, et al. Prospective, randomized,
comparative trial of ciprofloxacin otic drops, with or without hydrocortisone, vs polymyxin B-neomycin-hydrocortisone otic suspension
in the treatment of acute diffuse otitis externa. Infect Dis Clin Pract
1999;8:387–95.
23. Psifidis A, Nikolaidis P, Tsona A, et al. The efficacy and safety of local
ciprofloxacin in patients with external otitis : a randomized comparative study. Mediterranean J Otol Audiol 2005;1. Accessed 7/27/2005
at: www.mediotol.org/mjo.htm.
24. Roland PS, Pien FD, Schultz CC, et al. Efficacy and safety of topical
ciprofloxacin/dexamethasone versus neomycin/polymyxin B/hydrocortisone for otitis externa. Curr Med Res Opin 2004c;20:1175– 83.
25. Ruth M, Ekstrom T, Aberg B, et al. A clinical comparison of hydrocortisone butyrate with oxytetracycline/hydrocortisone acetate-polymyxin B in the local treatment of acute external otitis. Eur Arch
Otorhinolaryngol 1990;247:77– 80.
26. Sabater F, Maristany M, Mensa J, et al. Prospective double-blind
randomized study of the efficacy and tolerance of topical ciprofloxacin
vs. topical gentamicin in the treatment of simple chronic otitis media
and diffuse external otitis (Spanish). Acta Otorrinolaryngol Esp 1996;
47:217–20.
27. Slack RWT. A study of 3 preparations in the treatment of otitis
externa. J Laryngol Otol 1987;101:533–5.
28. Tsikoudas A, Jasser P, England RJ. Are topical antibiotics necessary in
the management of otitis externa? Clin Otolaryngol Allied Sci 2002;
27:260 –2.
29. van Balen FAM, Smit WM, Zuithoff NPA, et al. Clinical efficacy of
three common treatments in acute otitis externa in primary care:
randomised controlled trial. BMJ 2003;327:1201–3.
30. Jadad AR, Moore A, Carroll D, et al. Assessing the quality of reports
of randomized clinical trials: is blinding necessary? Control Clin Trials
1996;17:1–12.
31. DerSimonian R, Laird N. Meta-analysis in clinical trials. Controlled
Clin Trials 1986;17:177– 88.
32. Borenstein M, Rothstein H. Comprehensive meta-analysis: a computer
program for research synthesis (v. 1.0.25). Englewood, NJ: Biostat
Inc; 2000.
33. Sutton AJ, Abrams KR, Jones DR. An illustrated guide to the methods
of meta-analysis. J Eval Clin Pract 2001;7:135– 48.
34. Higgins JPT, Thompson SG, Deeks JJ, et al. Measuring inconsistency
in meta-analysis. BMJ 2003;327:557– 60.
35. Schulz KF, Chalmers I, Hayes RJ, et al. Empirical evidence of bias:
dimensions of methodological quality associated with estimates of
treatment effects in controlled trials. JAMA 1995;273:
408 –12.
36. Hollis S, Campbell F. What is meant by intention to treat analysis?
survey of published randomized controlled trials. BMJ 1999;319:
670 – 4.
37. Hannley MT, Denneny JC III, Holtzer SS. Consensus panel report. use
of ototopical antibiotics in treating 3 common ear diseases. Otolaryngol Head Neck Surg 2000;122:934 – 40.
38. Ioannidis JPA, Lau J. Pooling research results: benefits and limitations
of meta-analysis. J Qual Improvement 1999;25:462–9.
39. Moher D, Schulz KF, Altman D. The CONSORT statement; revised
recommendations for improving the quality of reports of parallelgroup randomized trials. JAMA 2001;285:1987–91.
40. Sackett DL, Rosenberg WMC, Gray JAM, et al. Evidence-based medicine: what it is and what it isn’t. BMJ 1996;312:71–2.
S38
Otolaryngology–Head and Neck Surgery, Vol 134, No 4S, April 2006
APPENDIX
Study 1
Data abstraction for Arnes and Dibb10
Sample size
Method of administration
Drug category or categories
Days of therapy
Amount
Dosing frequency
Adverse events
Clinical outcomes by patients, n/N (%)
Complete success 8d:
Partial success 8d:
Unsuccessful 8d:
Indeterminate 8d:
Bacteriologic outcomes by patients, n/N (%)
Eradication
Persistence
Superinfection
Ciprofloxacin 0.2%
Oxytetracycline 0.5%,
polymyxin B 10,000 IU/mL,
hydrocortisone 1.5%
16
Topical drops
Quinolone antibiotic
7
2-3 Drops
BID
NS
14
Topical drops
Nonquinolone antibiotic
7
2-3 Drops
BID
NS
14/16 (87.5)
2/16 (12.5)
0/16
0/16
5/14
4/14
4/14
1/14
(35.7)
(28.6)
(28.6)
(7.1)
15/16 (93.8)
1/16 (6.2)
1/16 (6.2)
7/14 (50.0)
7/14 (50.0)
0/14
BID, Twice a day; NS, not stated.
Study 2
Data abstraction for Cannon and Grunwaldt11
Sample size
Method of administration
Drug category or categories
Days of therapy
Amount
Dosing frequency
Adverse events
Clinical outcomes by patients, n/N (%)
Good response
Fair response
No response
Worse
Bacteriologic outcomes by ears, n/N (%)
Improved
Not improved
TID, Three times a day; NS, not stated.
Neomycin 0.13%,
methylprednisolone 0.5%
Placebo
20
Topical drops
Nonquinolone antibiotic, steroid
10
4 drops
TID
NS
20
Topical drops
Placebo
10
4 drops
TID
NS
11/20
5/20
2/20
2/20
(55.0)
(25.0)
(10.0)
(10.0)
19/23 (82.6)
4/23 (17.4)
2/20
2/20
7/20
9/20
(10.0)
(10.0)
(35.0)
(45.0)
4/22 (18.2)
18/22 (81.8)
Rosenfeld et al
Systematic Review of Topical Antimicrobial . . .
S39
Study 3
Data abstraction for Cannon12
Acetic acid 2%,
glyceryl triacetate 88%
Sample size*
Method of administration
Drug category or categories
Days of therapy
Amount
Dosing frequency
Adverse events
Clinical outcomes ears, n/N (%)
Complete response
Partial response
Ineffective
Bacteriologic outcomes by ears, n/N (%)
Cure
Failure
Other
Reduction in total symptom score (9-point scale)
23 Ears
Topical drops
Antiseptic
3-7
4 Drops
QID
NS
Placebo
26 Ears
Topical drops
Placebo
3-7
4 Drops
QID
NS
15/23 (65.2)
5/23 (21.7)
3/23 (13.0)
5/26 (19.2)
4/26 (15.4)
17/26 (65.4)
31/39 (79.5)
8/39 (20.5)
6/28 (21.4)
22/28 (78.6)
5.6
0.0 (P ⬍ .005)
QID, Four times a day; NS, not stated.
*98% of enrolled subjects had unilateral acute otitis externa.
Study 4
Data abstraction for Clayton et al13
Sample size
Method of administration
Drug category or categories
Days of therapy
Amount
Dosing frequency
Adverse events
Clinical outcomes by patients, n/N (%)
Improved at 21 days
Not improved at 21 days
Bacteriologic outcomes
TID, Three times a day; NS, not stated.
*Concentration not specified.
Aluminum acetate 8%
Gentamicin*
25
Topical drops
Antiseptic
21
5 Drops
TID
NS
41
Topical drops
Nonquinolone abx
21
5 Drops
TID
NS
18/25 (72.0)
7/25 (28.0)
NS
31/41 (75.6)
10/41 (24.4)
NS
S40
Otolaryngology–Head and Neck Surgery, Vol 134, No 4S, April 2006
Study 5
Data abstraction for Emgard and Hellstrom14
Betamethasone
dipropionate .05%
Oxytetracycline 1%,
polymyxin B 10,000 IU/mL,
hydrocortisone 0.5%
Sample size
Method of administration
Drug category or categories
26
Topical drops
Steroid
Days of therapy
Amount
Dosing frequency
Adverse events, n/N (%)
1. Any adverse event
2. Severe adverse event
3. Life-threatening adverse event
Clinical outcomes by patients, n/N (%)
1. Cured on day 11 and relapse-free for 6 months
Success
Failure
2. Therapeutic response day 4
Excellent
Good
Fair
Treatment failure
3. Therapeutic response day 11
Excellent
Good
Fair
Treatment failure
4. Overall condition per otolaryngologist, day 4
No evident symptoms
Mild symptoms
Symptoms present
Severe symptoms present
Bacteriologic outcomes by patients, n/N (%)
No growth, day 11
Other
1. Itching during acute phase, VAS rating
2. Pain during acute phase, VAS rating
3. Sleep loss during acute phase, VAS rating
4. Ability to work during acute phase, VAS rating
11
1 Unit dose pipette
BID
25
Topical drops
Nonquinolone antibiotic,
steroid
11
4 Drops
TID
10/26 (38.5)
1/26 (3.8)
0/25
16/25 (64.0)
3/25 (12.0)
0/25
14/24 (58.3)
10/24 (41.7)
6/22 (27.3)
16/22 (72.7)
6/26
16/26
4/26
0/26
0/26
20/26
5/26
1/26
0/26
0/24
14/24
10/24
3/24
0/24
10/24
9/24
2/24
3/24
(41.7)
(37.5)
(8.3)
(12.5)
1/24
13/24
8/24
2/24
(4.2)
(54.2)
(33.3)
(8.3)
5. Paracetamol tablets during acute phase
(23.1)
(61.5)
(15.4)
(76.9)
(19.2)
(3.8)
5/26 (19.2)
20/26 (76.9)
1/26 (3.8)
0/26
13/22 (59.1)
5/18 (27.8)
P ⬍ 0.01, favors group 1
P ⫽ 0.72
P ⫽ 0.69
P ⫽ 0.23
1.8
(58.3)
(41.7)
(12.5)
5.2
Rosenfeld et al
Systematic Review of Topical Antimicrobial . . .
S41
Study 6
Data abstraction for Freedman15
Neomycin 0.33%, colistin sulfate
0.3%, hydrocortisone 1.0%
Sample size
Method of administration
Drug category or categories
Days of therapy
Amount
Dosing frequency
Adverse events
Clinical outcomes by patients
Bacteriologic outcomes
Other
1. Symptom severity scales day 3
Edema
Itching
Redness
Scaling
Weeping
Pain
2. Symptom severity scales day 7
Edema
Itching
Redness
Scaling
Weeping
Pain
47
Topical drops
Nonquinolone antibiotic, steroid
21
4 Drops
TID
NS
No binary outcomes reported
Reported only for specific bacteria,
not by ears or patients
P
P
P
P
P
P
⬍
⬍
⫽
⫽
⫽
⫽
0.05 favoring group 1
0.05 favoring group 1
NS
NS
NS
NS
P
P
P
P
P
P
⬍
⬍
⬍
⬍
⬍
⫽
0.05
0.05
0.05
0.05
0.05
NS
favoring
favoring
favoring
favoring
favoring
group
group
group
group
group
Placebo
44
Topical drops
Placebo
21
4 Drops
TID
NS
1
1
1
1
1
TID, Three times a day; NS, not stated
Study 7
Data abstraction for Goldenberg et al16
Sample size
Method of administration
Drug category or categories
Days of therapy
Amount
Dosing frequency
Adverse events
Clinical outcomes by patients, n/N (%)
1. Clinical resolution, day 3-4
2. Clinical resolution, day 14
Bacteriologic outcomes by patients, n/N (%)
(%) Bacteriologic cure, day 14
Dexamethasone 10 mg,
oxytetracycline 90,000 U,
polymyxin B 100,000 U,
nystatin 1,000,000 U
Ciprofloxacin 0.3%
40
Topical powder
Nonquinolone antibiotic,
steroid, antifungal
14
Metered-dose powder
BID
NS
40
Topical drops
Quinolone
antibiotic
14
NS
TID
NS
40
Topical drops
Nonquinolone
antibiotic
14
NS
TID
NS
34/40 (85.0)
40/40 (100.0)
31/40 (77.5)
40/40 (100.0)
22/40 (55.0)
40/40 (100.0)
37/37 (100.0)
30/33 (90.9)
32/40 (80.0)
BID, Twice a day; TID, three times a day; NS, not stated.
Tobramycin 0.3%
S42
Otolaryngology–Head and Neck Surgery, Vol 134, No 4S, April 2006
Study 8
Data abstraction for Jones et al17
Neomycin 0.35%, polymyxin
B 10,000 IU/mL,
hydrocortisone 1.0%
Ofloxacin 0.3%
Sample size*
Method of administration
Drug category or categories
Days of therapy
Amount
Dosing frequency
Adverse events, n/N (%)
Any adverse event
Clinical outcomes by patients, n/N (%)
Clinical cure at 10 days
Bacteriologic outcomes by patients at 10
days, n/N (%)
Eradication
Persistence
Recurrence
Other
1. Symptom scores for tenderness
2. Symptom scores for secretion/exudate
3. Patient satisfaction scores
301 Intent-to-treat
242 Per protocol
Topical drops
Quinolone antibiotic
10
5 Drops (children) – 10
drops (adults)
BID
300 Intent-to-treat
232 Per protocol
Topical drops
Nonquinolone antibiotic, steroid
10
3 Drops (children) – 4 drops
(adults)
QID
29/242 (12.0)
24/232 (10.2)
215/242 (88.8)
206/232 (88.8)
91/93 (97.8)
1/93 (1.1)
1/93 (1.1)
102/103 (99.0)
1/103 (1.0)
0/103
P ⫽ NS, adults & children
P ⫽ NS, adults & children
P ⫽ NS, adults & children
BID, Twice a day; TID, three times a day; NS, not stated.
*Data reported originally for children and adults separately, but combined in the table.
Study 9
Data abstraction for Kime et al18
Acetic acid 2.0%,
hydrocortisone 1.0%
Sample size*
Method of administration
Drug category or categories
Days of therapy
Amount
Dosing frequency
Adverse events (n/N)
Clinical outcomes by ears, n/N (%)
1. Clinical cure at 11 days
2. Improved at 11 days
Bacteriologic outcomes by ears, n/N (%)
Bacteriologic cure at 11 days
Neomycin 0.33%, colistin sulfate 0.3%,
hydrocortisone 1.0%
44 Ears
Topical drops
Antiseptic, steroid
8-14
5 Drops
QID
NS
48 Ears
Topical drops
Nonquinolone antiibotic, steroid
8-14
5 Drops
QID
NS
34/44 (77.3)
44/44 (100.0)
36/48 (75.0)
48/48 (100.0)
39/44 (88.6)
42/48 (87.5)
QID, Four times a day; NS, not stated.
*90% of enrolled subjects had unilateral acute otitis externa.
Rosenfeld et al
Systematic Review of Topical Antimicrobial . . .
S43
Study 10
Data abstraction for Lambert19
Aluminum acetate*
Sample size
Method of administration
Drug category or categories
65
Topical drops
Antiseptic
Days of therapy
Amount
Dosing frequency
Adverse events, n/N (%)
Stinging
Clinical outcomes by patients, n/N (%)
Cured at 14 days
Bacteriologic outcomes by patients, n/N (%)
Other
Mean days until complete symptom
resolution (SD)
14
NS
TID
3/65 (4.6)
Neomycin 3,400 IU/ml,
polymyxin, 10,000 IU/mL,
hydrocortisone 1.0%
61
Topical drops
Nonquinolone antibiotic,
steroid
14
NS
TID
0/61
59/65 (90.8)
NS
49/61 (80.3)
NS
9.36 (3.36), n ⫽ 28
11.06 (4.6), n ⫽ 31
TID, Three times a day; NS, not stated.
*Concentration not specified.
Study 11
Data abstraction for Neher et al20
Sample size
Method of administration
Drug category or categories
Days of therapy
Amount (eg, drops)
Dosing frequency
Adverse events, n/N
Clinical outcomes by patients, n/N (%)
1. Successful outcome at 9 days
2. Complete resolution of inflammation at 5 days
Bacteriologic outcomes
Other
Mean days to resolution of inflammation (SD)
QD, Every day; NS, not stated.
N-chlorotaurine 1.0%
Neomycin 3,400 IU/ml,
polymyxin, 10,000 IU/mL,
hydrocortisone 1.0%
25
Topical drops
Antiseptic
9
NS (applied to wick)
QD
0/25
25
Topical drops
Nonquinolone abx, steroid
9
NS (applied to wick)
QD
0/25
25/25 (100.0)
13/25 (52.0)
NS
24/25 (96.0)
4/25 (16.0)
NS
5.6 (1.6)
7.4 (1.6)
S44
Otolaryngology–Head and Neck Surgery, Vol 134, No 4S, April 2006
Study 12
Data abstraction for Ordonez et al21
Acetic acid 2.0%,
hydrocortisone 1.0%
Neomycin 0.35%, polymyxin
B 10,000 IU/mL,
hydrocortisone 1.0%
Sample size*
Method of administration
57 Ears
Topical drops
Drug category or categories
Days of therapy
Amount
Dosing frequency
Adverse events, n/N (%)
Clinical outcomes by patients, n/N (%)
1. Clinical cure at 10 days
2. Improved at 10 days
Bacteriologic outcomes by patients, n/N (%)
Bacteriologic cure at 10 days
Antiseptic, steroid
8-14
5 Drops
QID
1/57 (1.8)
61 Ears
Topical drops
Nonquinolone antibiotic,
steroid
8-14
5 Drops
QID
1/61 (1.6)
49/57 (86.0)
57/57 (100.0)
50/61 (82.0)
59/61 (96.7)
45/57 (78.9)
48/59 (81.4)
QID, Four times a day.
*93% of enrolled subjects had unilateral acute otitis externa.
Study 13
Data abstraction for Pistorius et al22
Ciprofloxacin 0.2%
Sample size
Method of administration
Drug category or categories
Days of therapy
Amount
Dosing frequency
Adverse events, n/N (%)
1. Any adverse event
2. Discontinuation of study
drug
Clinical outcomes by patients,
n/N (%)
1. Resolution day 7
2. Improvement day 7
3. Resolution day 14-28
4. Relapse day 14-28
Bacteriologic outcomes by
patients, n/N (%)
Eradication day 7
Other
Median days to end of ear
pain
285 Intent-to-treat
239 Per protocol
Topical drops
Quinolone antibiotic
7
3 Drops
BID
66/236 (28.0)
1/236 (0.4)
179/236
219/236
200/236
6/236
(75.8)
(92.8)
(84.7)
(2.5)
Ciprofloxacin 0.2%,
hydrocortisone 1.0%
282 Intent-to-treat
236 Per protocol
Topical drops
Quinolone antibiotic,
steroid
7
3 Drops
BID
70/233 (30.0)
4/233 (1.7)
165/233
210/233
195/233
4/233
(70.8)
(90.1)
(83.7)
(1.7)
Neomycin 0.35%, poly-myxin
B 10,000 IU/mL,
hydrocortisone 1.0%
275 Intent-to-treat
228 Per protocol
Topical drops
Nonquinolone antibiotic,
steroid
7
3-4 Drops
TID
55/227 (24.2)
3/227 (1.3)
167/227
198/227
178/227
7/227
(73.6)
(87.2)
(78.4)
(3.1)
135/146 (92.5)
130/137 (94.9)
118/135 (87.4)
4.7
3.8 (P⫽0.039, 2 vs 1)
4.1 (P⫽NS, 3 vs 1 or 2)
BID, Twice a day; TID, three times a day; NS, not stated.
Rosenfeld et al
Systematic Review of Topical Antimicrobial . . .
S45
Study 14
Data abstraction for Psifidis et al23
Neomycin 0.35%, polymyxin B 10,000 IU/mL,
hydrocortisone 1.0%
Sample size
Method of administration
Drug category or categories
Days of therapy
Amount
Dosing frequency
Adverse events
Clinical outcome by patients,
n/N (%)
Complete resolution at 7
days
Bacteriologic outcomes by
patients, n/N (%)
1. Eradication
2. Superinfection
Ciprofloxacin 0.2%,
hydrocortisone 1.0%
Ciprofloxacin 0.2%
32
Topical drops
Nonquinolone antibiotic,
steroid
7
3 Days
TID
NS
29
Topical drops
Quinolone antibiotic,
steroid
7
3 Days
BID
NS
7
3 Days
BID
NS
27/32 (84.4)
29/29 (100.0)
29/30 (96.7)
23/32 (71.9)
5/32 (15.6)
24/29 (82.8)
3/29 (10.3)
28/30 (93.3)
0/30
BID, Twice a day; TID, three times a day; NS, not stated.
30
Topical drops
Quinolone antibiotic
S46
Otolaryngology–Head and Neck Surgery, Vol 134, No 4S, April 2006
Study 15
Data abstraction for Roland et al24
Ciprofloxacin 0.3%,
dexamethasone 0.1%
Sample size
Method of administration
Drug category or categories
Days of therapy
Amount
Dosing frequency
Adverse events, n/N (%)
Any adverse event
Clinical outcomes by patients, n/N (%)
1. Cured at day 3
2. Cured at day 8
3. Cured at day 18
4. Improved at day 3
5. Improved at day 8
6. Improved at day 18
Bacteriologic outcomes by patients, n/N (%)
Eradication day 18
Outcomes for culture-positive patients only:
1. Clinical cure at 3d
2. Clinical cure at 8d
3. Clinical cure at 18d
4. Inflammation at 18d
5. Edema at 18d
6. Tenderness at 18d
7. Discharge at 18d
232 Intent-to-treat
197 Per protocol
?? Culture positive*
Topical drops
Quinolone antibiotic,
steroid
7
3-4 Drops
BID
6/197 (3.0)
25/197
143/197
179/197
180/197
186/197
186/197
(12.7)
(72.6)
(90.9)
(91.4)
(94.4)
(94.4)
162/171* (94.7)
P
P
P
P
P
P
⫽
⫽
⫽
⫽
⫽
⫽
Neomycin 0.35%, polymyxin
B 10,000 IU/mL,
hydrocortisone 1.0%
236 Intent-to-treat
199 Per protocol
?? Culture positive*
Topical drops
Nonquinolone antibiotic,
steroid
7
3-4 Drops
TID
17/199 (8.5)
15/199
134/199
167/199
172/199
187/199
183/199
(7.5)
(67.3)
(83.9)
(86.4)
(94.0)
(92.0)
153/177* (86.4)
0.0375, favors 1
NS
0.0321, favors 1
0.0268, favors 1
NS
NS P ⫽ NS
BID, Twice a day; TID, three times a day; NS, not stated.
*Denominator calculated from Table 2 in the source article.
Study 16
Data abstraction for Ruth et al25
Oxytetracycline 1.0%, polymyxin
B 10,000 IU/mL, hydrocortisone 0.5%
Sample size
Method of administration
Drug category or categories
Days of therapy
Amount
Dosing frequency
Adverse events, n/N (%)
Any adverse event
Clinical outcomes by patients, n/N (%)
Success at 14 days
Bacteriologic outcomes
Hydrocortisone-17-alpha
butyrate 0.1%
23 Per protocol
Topical drops
Nonquinolone antibiotic, steroid
7
3-5 Drops
TID
23 Per protocol
Topical drops
Steroid
7
3-5 Drops
TID
10/23 (43.5)
8/23 (34.8)
17/23 (73.9)
NS
20/23 (87.0)
NS
Rosenfeld et al
Systematic Review of Topical Antimicrobial . . .
S47
Study 17
Data abstraction for Sabater et al26
Sample size
Method of administration
Drug category or categories
Days of therapy
Amount
Dosing frequency
Adverse events, n/N (%)
Clinical outcomes by patients, n/N (%)
Success at 8 days
Bacteriologic outcomes
Ciprofloxacin 0.5%
Gentamicin 0.3%
30
Topical drops
Quinolone antibiotic
8
5 Drops
TID
0/30
24
Topical drops
Nonquinolone antibiotic
8
5 Drops
TID
0/24
26/30 (86.7)
NS
19/24 (79.2)
NS
TID, Three times a day; NS, not stated.
Study 18
Data abstraction for Slack27
Neomycin 0.5%, polymyxin B
10,000 IU/mL,
hydrocortisone 1.0%
Econazole 1.0%, polymyxin B
15,000 IU/mL, fluocinolone
acetonide 0.1%
7-21
2 Drops
QID
7
Topical drops
Nonquinolone antibiotic,
steroid
7-21
2 Drops
QID
8
Topical drops
Antifungal, nonquinolone
antibiotic, steroid
7-21
2 Drops
QID
4/9 (44.4)
0/7
4/8 (50.0)
6/9 (66.7)
9/9 (100.0)
9/9 (100.0)
NS
3/7 (42.9)
6/7 (85.7)
7/7 (100.0)
NS
6/8 (75.0)
7/8 (87.5)
8/8 (100.0)
NS
Boric acid 4%,
ethyl alcohol 25%
Sample size
Method of administration
Drug category or
categories
Days of therapy
Amount
Dosing frequency
Adverse events, n/N (%)
Burning/stinging
Clinical outcomes by
patients, n/N (%)
Cure within 7 days
Cure with 14 days
Cure within 21 days
Bacteriologic outcomes
9
Topical drops
Antiseptic
QID, Four times a day; NS, not stated.
S48
Otolaryngology–Head and Neck Surgery, Vol 134, No 4S, April 2006
Study 19
Data abstraction for Tsikoudas et al28
Sample size
Method of administration
Drug category or categories
Days of therapy
Amount
Dosing frequency
Adverse events (n/N)
Clinical outcomes by patients
Bacteriologic outcomes
Other
Patient symptom scores day 11
Observer assessment scores day 11
Betamethasone
dipropionate 0.1%
Betamethasone dipropionate
0.1%, neomycin 0.5%
17
Topical drops
Steroid
11
3 Drops
NS
NS
No binary outcome data
NS
22
Topical drops
Steroid, nonquinolone antibiotic
11
3 Drops
NS
NS
NS
P ⫽ 0.164
P ⫽ 0.300
NS, not stated.
Study 20
Data abstraction for van Balen et al29
Acetic acid 0.7%
Sample size
Method of administration
Drug category or categories
Days of therapy
Amount
Dosing frequency
Adverse events, n/N (%)
Discontinued study drug
Burning, pain, or itching
Clinical outcomes by patients,
n/N (%)
1. Cured at 7 days
2. Cured at 14 days
3. Cured at 21 days
Bacteriologic outcomes
Other
Median days to symptom
resolution (95% CI)
Acetic acid 0.7%,
triamcinolone 0.1%
71 Intent-to-treat
65 Per protocol
Topical drops
63 Intent-to-treat
61 Per protocol
Topical drops
Antiseptic
7-21
3 Drops
TID
Antiseptic, steroid
7-21
3 Drops
TID
2/65 (3.1)
P ⫽ NS among
groups
1/61 (1.6)
Neomycin 0.5%,
poly-myxin B 10,000 IU/mL,
dexamethasone 0.07%
79 Intent-to-treat
73 Per protocol
Topical drops
Nonquinolone antibiotic,
steroid
7-21
3 Drops
TID
0/73
19/65 (29.2)
37/65 (56.9)
40/65 (61.5)
NS
P ⬍ 0.001 among
groups
29/61 (47.5)
46/61 (75.4)
54/61 (88.5)
NS
31/73 (42.5)
60/73 (82.2)
63/73 (86.3)
NS
8.0 (7.0-9.0)
7.0 (5.8-8.3)
6.0 (5.1-6.9)
TID,Three times a day; NS, not stated.