Current Medical Research & Opinion Vol. 27, No.

10, 2011, 2053–2060

0300-7995 Article ST-0175.R1/616192

doi:10.1185/03007995.2011.616192 All rights reserved: reproduction in whole or part not permitted

Original Article
A meta-analysis of the efficacy of quinolone
containing otics in comparison to antibiotic–
steroid combination drugs in the local treatment
of otitis externa

d
Curr Med Res Opin Downloaded from informahealthcare.com by Nyu Medical Center on 10/29/14

ite
l u nl n
na ow io
d,
py us is Lim

se oa
t
er an bu
K
rp c i
fo ers tr
Abstract
aU

so d
R. Mösges
M. Nematian-Samani Background:
rm

D
M. Hellmich The term otitis externa denotes the inflammation of the external auditory canal and can be treated locally in
K. Shah-Hosseini le is al the form of monotherapy or a combination drug.
rin ted m fo

Institute of Medical Statistics, Informatics and

ng or i
For personal use only.

si th rc

Epidemiology, Faculty of Medicine, University of

n

co ed
Cologne, Germany Objective:
1I
t a . Au e

The aim of the present meta-analysis was to compare the efficacy of an antibiotic–steroid combination drug
1

with that of monotherapy. According to current data, a comparable investigation based on network analysis
20

Address for correspondence:

d p ibi m

does not exist.

Prof. Dr. med. Ralph Mösges, Institute of Medical
o

Statistics, Informatics and Epidemiology, Faculty of

la d le ©

Methods:
C

Medicine, University of Cologne, Lindenburger Allee

42, 50931 Cologne, Germany. After systematically searching the PubMed, Medline, Medpilot, Web of Science and Embase electronic
ht
ew p r

Tel.: þ49 221 478 3456; Fax: þ49 221 478 3465; databases, 12 relevant randomized, controlled, clinical studies were identified involving 2682 evaluable
vi e o
an h
ig

Ralph@Moesges.de patients with regard to the cure rate and seven publications with 1251 microbiologically assessable patients.
r o

The collected data were compared directly and indirectly by means of network analysis.
yr
sp ize a

Key words:
di or S
No Cop

y, us

Antibiotics – Glucosteroids – Local treatment – Results:

r

Meta-analysis – Network analysis – Otitis externa – The direct comparison showed a trend towards the superiority of the monotherapy containing quinolone. The
au fo

Quinolones network analysis verified this tendency and demonstrated that pure quinolone drugs can achieve a
significantly higher cure rate (OR: 1.29; 95% CI: 1.06–1.57; p ¼ 0.01) and a significantly superior
Un t

Accepted: 17 August 2011; published online: 15 September 2011

th

Citation: Curr Med Res Opin 2011; 27:2053–60 eradication rate (OR: 1.44; 95% CI: 1.03–2.02; p ¼ 0.03) compared to combination drugs not
containing quinolone. We found substantial heterogeneity (with I2 up to 88.7%) between studies,
presumably due to treatments applied in varying frequency, thus bearing on compliance and outcome.

Conclusion:
With a level Ia evidence, this investigation validates the clinical benefit of quinolones as compared to classic
combination drugs in the local treatment of acute otitis externa.

Introduction
The term otitis externa denotes the inflammation of the external auditory canal,
in the broader sense also the pinna, and can affect the ear in an acute and
chronic manner1.
Statistically, one in ten people suffer at least once in their lives from an
inflammation of the external auditory canal. Most of the time, just one ear is
affected; in 10% both ears become infected2. Besides the typical symptoms such

! 2011 Informa UK Ltd www.cmrojournal.com A meta-analysis of the local treatment of otitis externa Mösges et al. 2053
 Current Medical Research & Opinion Volume 27, Number 10
as pain, redness, itching and secretion, the clinical picture
can also include conductive hearing loss and even
dizziness3. In addition to viruses and fungi, mainly bacteria
have been named as causative agents – particularly
Staphylococcus aureus and Pseudomonas aeruginosa4.
Drug therapy of otitis externa can be carried out with
various active agents that are usually administered locally
and not systemically. Antiseptic agents, antibiotics, and
glucosteroids rank among the substances used. All of
these remedies are applied either in the form of monother-
apy or a fixed combination drug5. Antiseptics, e.g. acetic
acid, are mostly used together with antibiotics and offer a
broad spectrum of activity against pathogens by lowering
Curr Med Res Opin Downloaded from informahealthcare.com by Nyu Medical Center on 10/29/14
the pH in the ear canal. For years, glucosteroids had the
reputation of primarily reducing the swelling of the ear
canal. Yet, recent studies also ascribe antibacterial and
anti-fungal effects to them in treating otitis externa2.
The number of available studies on steroidal monothera-
pies, however, is too low to be able to make an exact state-
ment1. A specifically effective antibiotic can be prescribed
only if an antibiogram has been made. In practice, this is
initially not the case so that calculated treatment is usually
performed. An antibiotic is chosen that is effective against
both of the pathogens Staphylococcus aureus and
For personal use only.
Pseudomonas aeruginosa. In addition, non-steroidal anti-
inflammatory drugs are commonly used for reducing pain2.
In 2006 Rosenfeld et al. discussed the efficacy of local
antibiotic treatment and in doing so also contrasted mono-
therapies and combination drugs. The results from only
few clinical studies were drawn upon6.
The present systematic review is to investigate the effi-
cacy of the otologic drugs commonly used in the treatment
of otitis externa. Studies in which the efficacy of a fixed
antibiotic–steroidal combination was compared with that
of a monotherapy are then to be analyzed as to whether
clinically relevant differences can be determined.
Materials and methods
Research of the literature
In order to calculate the efficacy of the various drugs, sta-
tistical data from previously published clinical studies were
to be analyzed. Suitable papers were sought systematically.
A catalog of search terms was compiled taking into
account the question at hand, the criteria, and the the-
matic constellations and entered in the MeSH Database via
Pubmed so that other synonyms could be found. These
words were subsequently re-entered in combined form in
the PubMed, Medline, Medpilot, Web of Science, and Embase
electronic databases. The following search terms were
used:
‘Otitis externa’, ‘anti-bacterial agents’, ‘quinolones’, ‘drug
therapy’, ‘local therapy’, ‘therapeutics’, ‘treatment
2054 A meta-analysis of the local treatment of otitis externa Mösges et al.
October 2011
outcome’, ‘therapeutic equivalency’, ‘combination’, ‘com-
bined therapy’, ‘glucocorticoids’, ‘steroids’, ‘instillation,
drug’, ‘ointments’, ‘ear drops’, ‘inflammation’, and
‘efficacy’.
The search was refined further using the terms ‘humans’,
‘clinical trial’, ‘clinical trial phase I’, ‘clinical trial phase II’,
‘clinical trial phase III’, ‘clinical trial phase IV’. No restric-
tions were chosen with respect to the articles’ years of
publication.
In all, 117 articles were found in the systematic search.
Two more studies not listed in the databases could be iden-
tified via cross-references in the technical literature. The
review of the literature was completed in February 2011.
English-language full texts were found for 103 articles.
For 14 studies, only the abstracts were in English, but the
text bodies were written in a foreign language. The full text
could not be found for two studies.
Three independent reviewers evaluated the articles and
extracted the data. The collected data were compared
directly and indirectly by means of network analysis.
Outcome parameters
The improvement of the clinical picture as the ‘overall
cure rate’ on the one hand and the microbiological erad-
ication as the ‘overall eradication rate’ on the other were
uniformly defined as outcome parameters to ensure com-
parability of the studies. The terms ‘cure’ and ‘resolution’
fell into the category ‘success,’ whereas ‘no improvement’
and ‘worsening’ were regarded as ‘failure.’ The statistical
analyses of the clinical cure and of the microbiological
eradication were carried out separately.
Statistical methods
For the overall cure or eradication rate of all studies, the
proportions of the individual studies were transformed
using the Freeman–Tukey transformation7 to be able to
apply the usual methods for the random and fixed effects.
The proportions were depicted in a forest plot with 95%
confidence intervals according to Clopper and Pearson8.
The area of the squares (point estimator of the odds ratio)
is proportional to the patient numbers of the individual
studies.
The indirect comparisons of the single studies were per-
formed based on a random effect model for a meta-analysis
according to DerSimonian and Laird, and a logistic regres-
sion was applied with which the odds ratio – and not the
relative risk – must be incorporated9–11. For this reason,
the odds ratio instead of the relative risk was also used in
the direct comparison in order to depict the effect in a
comparable manner. Figure 1 illustrates the path of the
indirect comparison. The respective graph is labeled
with the study numbers as shown in Table 1. The end
www.cmrojournal.com ! 2011 Informa UK Ltd
 Current Medical Research & Opinion Volume 27, Number 10 October 2011

points identify which active agents were used in the Explanation of terms
studies.
Furthermore, the analysis was stratified depending on In this context, the term ‘monotherapy’ refers to medi-
the study so that each study could be granted a different cine that contains at least one antibiotic or at least one
control group risk. glucosteroid. This means that also a mixture of several
The programs Review-Manager 4.2 and Stata 11.1 were antibiotics still falls into the monotherapy group; this
used for the statistical analysis. definition was applied likewise for glucosteroids.
Consequently, ‘combination drug’ in this context is an
otologic drug that contains both antibiotics as well as
glucosteroids. Antiseptic or antimycotic additives were
Q disregarded.
S3
Antibiotics and glucosteroids were assigned to strength
4
groups and thus classified. The antibiotics were divided
Curr Med Res Opin Downloaded from informahealthcare.com by Nyu Medical Center on 10/29/14

into a quinolone class and a non-quinolone class (amino-

A(+A)
glycosides, polypeptides, and polyketides). The glucoster-
oids, in contrast, were divided corresponding to the
1,2

therapeutic index of topic dermatotherapy into

,4,6

the active ingredient strengths S1 (very weak), S2

,8,1

4,
5,
12 S1/S2
(weak), S3 (strong) and S4 (very strong)12–13: hydrocorti-
0
8

sone 1–2, triamcinolone 2, dexamethasone 2–3, beta-

methasone 3.
,11
7,9 The cure and eradication rates refer as much as possible
Q+S1 8 to the ‘intent-to-treat’ patient number and to the status at
A(+A)+S1/S2
the end of treatment in order to set a comparable standard.
For personal use only.

In some cases, however, this was not possible such that the
A(+A) = At least one antibiotic evaluable patients had to be considered the sample set and
S(x) = Steroid (therapeutic index) the given end results as the cure or eradication rate. In one
Q = Quinolone
case, bilateral infections were counted twice and equated
Figure 1. Network analysis. with the number of patients14.

Table 1. Overview of studies included in the meta-analysis.

Author Design Duration Substance

21
1. Arnes and Dibb 1993 Randomized, controlled,
non-blinded
22
2. Drehobl et al. 2008 Randomized, controlled,
single-blind
3. Emgård and Hellstrom 200523 Randomized, controlled,
non-blinded
24
4. Goldenberg et al. 2002 Randomized, controlled,
non-blinded
5. Gydé et al. 198214 Randomized, controlled,
double-blind
6. Jones et al. 199725 Randomized, controlled,
single-blind
26
7. Kime et al. 1978 Randomized, controlled,
double-blind
8. Pistorius et al. 199927 Randomized, controlled,
non-blinded
28
9. Ruth et al. 1990 Randomized, controlled,
single-blind
10. Schwartz 200629 Randomized, controlled,
single-blind
11. van Balen et al. 200330 Randomized, controlled,
double-blind
12. Wadsten et al. 198531 Randomized, controlled,
single-blind
7 days Ciprofloxacin vs. oxytetracycline–polymyxin
B–hydrocortisone
7 days Ciprofloxacin vs. polymyxin B–neomycin–
hydrocortisone
11 days Betamethasone dipropionate vs. hydrocortisone–
oxytetracycline–hydrochloride–polymyxin B
14 days Ciprofloxacin vs. dexamethasone–oxytetracy-
cline–polymyxin B sulfate–nystatin vs.
tobramycin
14 days Gentamicin vs. colistin–neomycin–hydrocortisone
10 days Ofloxacin vs. neomycin sulfate–polymyxin B sul-
fate–hydrocortisone
11  3 days Hydrocortisone–acetic acid vs. hydrocortisone–
neomycin–colistin
7 days Ciprofloxacin vs. ciprofloxacin–hydrocortisone vs.
polymyxin B–neomycin–hydrocortisone
7 days Hydrocortisone–17-a-butyrate vs. oxytetracy-
cline–hydrocortisone–polymyxin B
7–10 days Ofloxacin vs. neomycin sulfate–polymyxin B sul-
fate–hydrocortisone
at least 7 days Propylene glycol vs. triamcinolone acetic acid vs.
dexamethasone phosphate sodium–neomycin
sulfate–polymyxin B sulfate
7–8 days Framycetin–gramicidin vs. oxytetracycline–
hydrocortisone–polymyxin B

! 2011 Informa UK Ltd www.cmrojournal.com A meta-analysis of the local treatment of otitis externa Mösges et al. 2055
 Current Medical Research & Opinion Volume 27, Number 10 October 2011

Total number of articles or a recent case of acute otitis externa, known allergies to
researched: n = 119 the medicines used, and recent antibiotic or steroid
treatment.

Exclusion of irrelevant articles

after review of the
fulltext/abstract: n = 95
Articles not included because
they were reviews: n = 6
Thematically relevant
Curr Med Res Opin Downloaded from informahealthcare.com by Nyu Medical Center on 10/29/14
Statistical evaluation
For the patients treated, the average cure rate is 73.6%
(95% CI: 0.72–0.75) assuming the fixed effects model or
77.8% (95% CI: 0.73–0.82) for the random effects model
(Figure 3). With respect to microbiological eradication,
the rates are 84.6% (fixed effects model, 95% CI: 0.83–
0.87) or 84.5% (random effects model, 95% CI: 0.78–0.90)
(Figure 4). The heterogeneity of the overall cure rate is

studies: n = 18 I2 ¼ 86% or for the overall eradication rate I2 ¼ 88.7%

Studies not included because
the fulltext was not available:
n=2
Studies not included because
of heterogeneous outcome-
measures: n = 4
Inclusion of relevant
For personal use only.
(Figures 3 and 4).
The following analogies could be drawn via direct and
indirect comparisons, thus enabling further analyses to be
conducted.
The direct comparison of non-quinolone combinations
(A þ A þ S1/S2) with pure quinolone drugs shows a ten-
dency towards the superiority of quinolones (odds ratio
A þ A þ S1/S2: 0.72; 95% CI: 0.50–1.04; p ¼ 0.08)
(Figure 5). Here, 1024 patients applied quinolone and

studies in the systematic
review: n = 12
Figure 2. Study inclusions and exclusions during the reviewing process.
Results
Analysis of the literature
The search resulted in 119 hits. Three independent
reviewers then analyzed them, with 95 clinical studies
being assessed as having inapplicable topics and six reviews
in the search results being discarded. Eighteen studies were
rated as having suitable topics. Later in the procedure,
however, four of these studies could not be included in
the survey, since their results were not expressed in cure
or eradication rate percentages; they used symptom scores
as outcome parameter or had inconsistent definitions of
the cure rate15–18. Two other studies had relevant topics
for the meta-analysis after assessing their abstracts. Access
to the full texts, however, was denied so that these studies
could not be incorporated into the analysis19–20.
Thus, twelve randomized clinical studies remained for
the statistical analysis (Figure 2)14,21–31.
Patients
Pooling of the studies resulted in a total number of 2682
patients for the analysis of the cure rate and 1251 for the
eradication rate. Children and adults were included in the
analysis. In nearly all studies, exclusion criteria for
the participants were a perforated eardrum, a necrotizing
985 study participants used the combination drug. The
situation is similar for the eradication rate: the odds ratio
for the combination drugs is 0.71 (95% CI: 0.43–1.18;
p ¼ 0.19) (Figure 6). Here, too, the patients were distrib-
uted evenly (Q: n ¼ 485; combination; n ¼ 460).
An indirect comparison (network analysis using logistic
regression) of the cure rate of 2562 patients and the erad-
ication rate of 1251 patients could be carried out
(Figure 1). The probability of cure (odds ratio of Q: 1.29;
95% CI: 1.06–1.57; p ¼ 0.01) and of eradication (odds
ratio of Q: 1.44; 95% CI: 1.03–2.02; p ¼ 0.03) is signifi-
cantly higher for pure quinolone drugs compared to com-
bination medication (Table 2).
Discussion
‘A lot helps a lot’ was a maxim of practical medicine for a
long time. When treating infectious diseases, however, the
aimless application of various active ingredients can result
in the development of resistances and undesirable side
effects. Modern medicine makes focused use of active sub-
stances, e.g., adapted exactly to the spectrum of pathogens.
In this systematic review with a meta-analysis on the basis
of published randomized, controlled, clinical studies of
acute otitis externa, quinolone-containing monotherapies
perform – as expected – significantly better in terms of
both the cure and the eradication rate than classic antibi-
otic–steroid combination drugs. This is verified for the first
time utilizing network analysis and level Ia evidence.
Currently available studies show that fluoroquinolones,
led by ciprofloxacin, unlike aminoglycosides do not cause

2056 A meta-analysis of the local treatment of otitis externa Mösges et al. www.cmrojournal.com ! 2011 Informa UK Ltd
 Current Medical Research & Opinion Volume 27, Number 10 October 2011

Otitis externa - Cure

Study ID Treatment #Events/#Total Estimate (95% CI)

Gydé MC, 1982 A(+A) 10/10 1.000 (0.692-1.000)
Wadsten CJ, 1985 A(+A) 21/26 0.808 (0.606-0.934)
Goldenberg D, 2002 A(+A) 40/40 1.000 (0.912-1.000)
Kime CE, 1978 A+A+S1/S2 36/48 0.750 (0.604-0.864)
Gydé MC, 1982 A+A+S1/S2 11/13 0.846 (0.546-0.981)
Wadsten CJ, 1985 A+A+S1/S2 22/29 0.759 (0.565-0.897)
Ruth M, 1990 A+A+S1/S2 17/23 0.739 (0.516-0.898)
Arnes E, 1993 A+A+S1/S2 5/14 0.357 (0.128-0.649)
Jones RN, 1997 A+A+S1/S2 227/300 0.757 (0.704-0.804)
Pistorius B, 1999 A+A+S1/S2 167/227 0.736 (0.673-0.792)
Goldenberg D, 2002 A+A+S1/S2 40/40 1.000 (0.912-1.000)
van Balen FA, 2003 A+A+S1/S2 60/73 0.822 (0.715-0.902)
Emgård P, 2005 A+A+S1/S2 10/24 0.417 (0.221-0.634)
Schwartz RH, 2006 A+A+S1/S2 61/95 0.642 (0.537-0.738)
Drehobl M, 2008 A+A+S1/S2 183/309 0.592 (0.535-0.648)
Curr Med Res Opin Downloaded from informahealthcare.com by Nyu Medical Center on 10/29/14

Arnes E, 1993 Q 14/16 0.875 (0.617-0.984)

Jones RN, 1997 Q 225/301 0.748 (0.694-0.796)
Pistorius B, 1999 Q 179/236 0.758 (0.699-0.812)
Goldenberg D, 2002 Q 40/40 1.000 (0.912-1.000)
Schwartz RH, 2006 Q 88/113 0.779 (0.691-0.851)
Drehobl M, 2008 Q 214/318 0.673 (0.618-0.724)
Pistorius B, 1999 Q+S1 165/233 0.708 (0.645-0.766)
Kime CE, 1978 S1/S2 34/44 0.773 (0.622-0.885)
Ruth M, 1990 S1/S2 20/23 0.870 (0.664-0.972)
van Balen FA, 2003 S1/S2 46/61 0.754 (0.627-0.855)
Emgård P, 2005 S3 20/26 0.769 (0.564-0.910)

Fixed effect 0.736 (0.720-0.753)

Random effects 0.778 (0.730-0.823)
For personal use only.

0.0 0.2 0.4 0.6 0.8 1.0

Proportion

Quantifying heterogeneity: I2 = 86% (80.7%; 89.9%)

Figure 3. Cure rate.

ototoxic side effects29,32. Furthermore, fluoroquinolones
possess very high in vitro activity against gram-negative
pathogens, e.g. Pseudomonas aeruginosa as the main path-
ogen in otitis externa33. Thus, gyrase inhibitors appear to
be first-line treatment for otitis externa.
Pistorius et al. verify a statistically significantly
shorter time-to-end of ear pain when a glucosteroid
from the active substance group 1 (hydrocortisone) is
added to ciprofloxacin (median time-to-end of pain:
3.79 days vs. 4.67 days; p ¼ 0.039)27. In none of the
studies included in the meta-analysis was a quinolone
used in combination with an effective glucosteroid
having the strength 3 or 4. It would certainly be inter-
esting to investigate the performance of a combination
drug consisting of a quinolone and such a glucosteroid
compared with a pure quinolone drug. Roland et al.
have conducted studies in which ciprofloxacin and dexa-
methasone were used and which showed clinical superi-
ority to a polymyxin–neomycin–hydrocortisone
combination (clinical cure: 90.9% vs. 83.9%;
p ¼ 0.0375), yet this does not allow conclusions to be
drawn about the contribution of dexamethasone to quin-
olone34. It therefore continues to be difficult to quantify
the contribution of glucosteroids to healing success.
Steroids are known to act in an anti-inflammatory and
antiphlogistic manner. Their effect does not appear until
after a few days and becomes noticeable in the reduction
of redness and swelling of the skin and mucous mem-
branes2,35. Consequently, pressure-induced pain
diminishes. This parameter, however, is not reflected
in the outcome parameters of eradication and cure.
The parameter ‘swelling’ is not included in the microbi-
ological eradication, and it is expressed only indirectly
in the cure rate through pain reduction. This could
cause the limitation of our results to the disadvantage
of the steroid combination drugs. In addition, the
number of available studies on steroid monotherapies
is small1. Emgård et al. confirmed the significantly
greater improvement of symptoms concerning itching
on a VAS scale under betamethasone treatment as
opposed to the combination of hydrocortisone–oxytetra-
cycline–polymyxin (p50.01). However, this must be
seen in the context of the small number of patients
(n ¼ 51)23. Placebo-controlled, three-arm studies would
have been desirable in order to demonstrate the steroid
effect. Yet, for ethical reasons such studies are out of the
question and are not compatible with the Declaration of
Helsinki36.

! 2011 Informa UK Ltd www.cmrojournal.com A meta-analysis of the local treatment of otitis externa Mösges et al. 2057
 Current Medical Research & Opinion Volume 27, Number 10 October 2011

Otitis externa - Eradication

Study ID Treatment #Events/#Total Estimate (95% CI)

Goldenberg D, 2002 A(+A) 32/40 0.800 (0.644-0.909)

Kime CE, 1978 A+A+S1/S2 42/48 0.875 (0.748-0.953)

Arnes E, 1993 A+A+S1/S2 7/14 0.500 (0.230-0.770)

Jones RN, 1997 A+A+S1/S2 97/103 0.942 (0.878-0.978)

Pistorius B, 1999 A+A+S1/S2 118/135 0.874 (0.806-0.925)

Goldenberg D, 2002 A+A+S1/S2 37/37 1.000 (0.905-1.000)

Schwartz RH, 2006 A+A+S1/S2 23/34 0.676 (0.495-0.826)

Drehobl M, 2008 A+A+S1/S2 115/174 0.661 (0.585-0.731)

Arnes E, 1993 Q 15/16 0.938 (0.698-0.998)

Curr Med Res Opin Downloaded from informahealthcare.com by Nyu Medical Center on 10/29/14

Jones RN, 1997 Q 85/93 0.914 (0.838-0.962)

Pistorius B, 1999 Q 135/146 0.925 (0.869-0.962)

Schwartz RH, 2006 Q 39/56 0.696 (0.559-0.812)

Drehobl M, 2008 Q 128/174 0.736 (0.664-0.799)

Pistorius B, 1999 Q+S1 130/137 0.949 (0.898-0.979)

Kime CE, 1978 S1/S2 39/44 0.886 (0.754-0.962)

Fixed effect 0.846 (0.825-0.865)

Random effects 0.845 (0.777-0.902)

For personal use only.

0.0 0.2 0.4 0.6 0.8 1.0

Proportion

Quantifying heterogeneity:
I 2 = 88.7% (83.1%; 92.5%)

Figure 4. Eradication rate.

Review: Otitis externa

Comparison: 01 A+A+S1/S2 vs. Q
Outcome: 01 Cure

Study A+A+S1/S2 Q OR (random) Weight OR (random)

or sub-category n/N n/N 95% CI % 95% CI

Arnes E 5/14 14/16 3.61 0.08 [0.01, 0.50]

Jones RN 227/300 225/301 26.18 1.05 [0.73, 1.52]
Pistorius B 167/227 179/236 24.37 0.89 [0.58, 1.35]
Goldenberg D 40/40 40/40 Not estimable
Schwartz RH 61/95 88/113 18.01 0.51 [0.28, 0.94]
Drehobl M 183/309 214/318 27.83 0.71 [0.51, 0.98]

Total (95% CI) 985 1024 100.00 0.72 [0.50, 1.04]

Total events: 683 (A+A+S1/S2), 760 (Q)
Test for heterogeneity: Chi² = 10.98, df = 4 (P = 0.03), I² = 63.6%
Test for overall effect: Z = 1.73 (P = 0.08)

0.1 0.2 0.5 1 2 5 10

Favours Q Favours A+A+S1/S2

Figure 5. Direct comparison: cure rate (classic combination therapy vs. quinolones).

When interpreting the results, it is important to know
whether the observed differences in the individual studies
are considered to be coincidental and how great the vari-
ability is among the studies. Reasons for heterogeneity are
possibly due to the various treatments which influence the
frequency of medicine application and ultimately also
have an effect on patient compliance and treatment suc-
cess. Fluoroquinolones usually have a lower rate of
application.
Systematic errors could be due to a publication bias on
the one hand or the different blinding methods of the
various studies on the other. These errors have an impact

2058 A meta-analysis of the local treatment of otitis externa Mösges et al. www.cmrojournal.com ! 2011 Informa UK Ltd
 Current Medical Research & Opinion Volume 27, Number 10 October 2011

Review: Otitis externa

Comparison: 01 A+A+S1/S2 vs. Q
Outcome: 02 Eradication

Study A+A+S1/S2 Q OR (random) Weight OR (random)

or sub-category n/N n/N 95% CI % 95% CI

Arnes E 7/14 15/16 4.59 0.07 [0.01, 0.65]

Jones RN 97/103 85/93 15.40 1.52 [0.51, 4.56]
Pistorius B 118/135 135/146 23.18 0.57 [0.25, 1.26]
Schwartz RH 23/34 39/56 19.62 0.91 [0.36, 2.28]
Drehobl M 115/174 128/174 37.21 0.70 [0.44, 1.11]

Total (95% CI) 460 485 100.00 0.71 [0.43, 1.18]

Total events: 360 (A+A+S1/S2), 402 (Q)
Test for heterogeneity: Chi² = 6.61, df = 4 (P = 0.16), I² = 39.5%
Test for overall effect: Z = 1.31 (P = 0.19)

0.1 0.2 0.5 1 2 5 10

Favours Q Favours A+A+S1/S2

Figure 6. Direct comparison: eradication rate (classic combination therapy vs. quinolones).
Curr Med Res Opin Downloaded from informahealthcare.com by Nyu Medical Center on 10/29/14

CMRO peer reviewers may have received honoraria for their

Table 2. Indirect comparison: cure rate and eradication rate (quinolones vs.
classic combination therapy).
review work. The peer reviewers on this manuscript have dis-
closed that they have no relevant financial relationships.
Q. vs. A(þA) þ S1/S2 Number OR 95% CI P4|z|
of subjects Acknowledgements
Ralph Mösges and Mehregan Nematian-Samani contributed
Cure rate 2562 1.29 1.06–1.57 0.012 equally to this manuscript.
Eradication rate 1251 1.44 1.03–2.02 0.034
We thank Gena Kittel BA for assisting in the preparation of
this manuscript.
For personal use only.

on the quality of the individual studies and can lead to an

overestimation or underestimation of the actual treatment
effect.
Conclusion
By using network analysis, this investigation demonstrates
Ia at the level of evidence that quinolone monotherapy is
significantly superior to non-quinolone combination drugs
consisting of older antibiotics and steroids with respect to
bacterial eradication (OR: 1.44; 95% CI: 1.03–2.02;
p ¼ 0.03) and the cure (OR: 1.29; 95% CI: 1.06–1.57;
p ¼ 0.01) of acute otitis externa.
Transparency
Declaration of funding
This project was financed with university funding and an unrest-
ricted research grant provided by Alcon Pharma GmbH,
Germany. The sponsor took no influence on the preparation of
the manuscript.
Declaration of financial/other relationships
R.M. has disclosed that he has received funding for his research
from Alcon Pharma GmbH, Germany, and has acted as a scien-
tific consultant for Alcon Inc., USA. The other authors have no
conflicting financial or other interests to report. The present
systematic review was performed in the context of a university
research project.
References
1. Hajioff D, Mackeith S. Otitis externa. Clin Evid 2008;06:510
2. Neher A, Nagl M, Scholtz AW. Otitis externa: etiology, diagnostic and therapy.
HNO 2008;56:1067-79; quiz 1080
3. Halpern MT, Palmer CS, Seidlin M. Treatment patterns for otitis externa. J Am
Board Fam Pract 1999;12:1-7
4. Neher A, Nagl M, Appenroth E, et al. Acute otitis externa: efficacy and toler-
ability of N-chlorotaurine, a novel endogenous antiseptic agent. Laryngoscope
2004;114:850-4
5. Osguthorpe JD, Nielsen DR. Otitis externa: review and clinical update.
Am Fam Physician 2006;74:1510-16
6. Rosenfeld RM, Singer M, Wasserman JM, et al. Systematic review of topical
antimicrobial therapy for acute otitis externa. Otolaryngol Head Neck Surg
2006;134(4 Suppl):S24-48
7. Mosteller F, Youtz C. Tables of Freeman–Tukey transformations for binomial
and Poisson distributions. Biometrika 1961;48:433
8. Clopper C, Pearson ES. The use of confidence or fiducial limits illustrated in
the case of the binomial. Biometrika 1934;48:433-40
9. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials
1986;7:177-87
10. Lumley T. Network meta-analysis for indirect treatment comparisons. Stat
Med 2002;21:2313-24
11. Glenny AM, Altman AD, Song F, et al. Indirect comparisons of competing
interventions. Health Technol Assess 2005;9:1-148
12. Luger T, Loske KD, Elsner P, et al. Topische Dermatotherapie mit
Glukokortikoiden – Therapeutischer Index: Deutsche Dermatologische
Gesellschaft. JDDG 2004;2:629-34
13. Stiftung Warentest: Neurodermitis, Kontaktekzem. Available at: http://
www.test.de/themen/gesundheit-kosmetik/medikamente/vom_arzt/a_
haut_haare/a_ekzem_neurodermitis/a_ekzem_neurodermitis/bespr.med/
[Last accessed 22 May 2011]
14. Gydé MC, Norris D, Kavalec EC. The weeping ear – clinical re-evaluation of
treatment. J Int Med Res 1982;10:333-40

! 2011 Informa UK Ltd www.cmrojournal.com A meta-analysis of the local treatment of otitis externa Mösges et al. 2059
 Current Medical Research & Opinion Volume 27, Number 10 October 2011
15. Abelardo E, Pope L, Rajkumar K, et al. A double-blind randomised clinical trial
of the treatment of otitis externa using topical steroid alone versus topical
steroid-antibiotic therapy. Eur Arch Otorhinolaryngol 2009;266:41-5
16. Mosges R, Domrose CM, Loffler J. Topical treatment of acute otitis externa:
clinical comparison of an antibiotics ointment alone or in combination with
hydrocortisone acetate. Eur Arch Otorhinolaryngol 2007;264:1087-94
17. Mosges R, Schroder T, Baues CM, et al. Dexamethasone phosphate in anti-
biotic ear drops for the treatment of acute bacterial otitis externa. Curr Med
Res Opin 2008;24:2339-47
18. Tsikoudas A, Jasser P, England RJ. Are topical antibiotics necessary in the
management of otitis externa? Clin Otolaryngol Allied Sci 2002;27:260-2
19. Lidlholdt T GP, Kehrl W, Leibermann A. Otitis externa treated by topical
antibiotics. Otolaryngology – Head and Neck Surgery 1997;117:116
20. Psifidis A, Nikolaidis P, Tsona A, et al. The efficacy and safety of local cip-
rofloxacin in patients with external otitis: a randomized study. Mediterr J Otol
2005;1:20-3
Curr Med Res Opin Downloaded from informahealthcare.com by Nyu Medical Center on 10/29/14
21. Arnes E, Dibb WL. Otitis externa: clinical comparison of local ciprofloxacin
versus local oxytetracycline, polymyxin B, hydrocortisone combination treat-
ment. Curr Med Res Opin 1993;13:182-6
22. Drehobl M, Guerrero JL, Lacarte PR, et al. Comparison of efficacy and safety
of ciprofloxacin otic solution 0.2% versus polymyxin B–neomycin–hydrocor-
tisone in the treatment of acute diffuse otitis externa. Curr Med Res Opin
2008;24:3531-42
23. Emgård P, Hellstrom S. A group III steroid solution without antibiotic compo-
nents: an effective cure for external otitis. J Laryngol Otol 2005;119:342-7
24. 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
25. Jones RN, Milazzo J, Seidlin M. Ofloxacin otic solution for treatment of otitis
For personal use only.
externa in children and adults. Arch Otolaryngol Head Neck Surg
1997;123:1193-200
26. Kime CE, Ordonez GE, Updegraff WR, et al. Effective treatment of acute,
diffuse otitis-externa. 2. Controlled comparison of hydrocortisone–acetic
acid, non-aqueous and hydrocortisone–neomycin–colistin otic solutions.
Curr Ther Res Clin E 1978;23:Ss15-Ss28
2060 A meta-analysis of the local treatment of otitis externa Mösges et al.
27. Pistorius B, Westberry K, Drehobl M, et al. Prospective, randomized,
comparative trial of ciprofloxacin otic drops, with or without hydrocorti-
sone, vs. polymyxin B–neomycin–hydrocortisone otic suspension in
the treatment of acute diffuse otitis externa. Infect Dis Clin Prac 1999;
8:387-95
28. Ruth M, Ekstrom T, Aberg B, et al. A clinical comparison of hydrocorti-
sone butyrate with oxytetracycline/hydrocortisone acetate–polymyxin B in
the local treatment of acute external otitis. Eur Arch Otorhinolaryngol 1990;
247:77-80
29. Schwartz RH. Once-daily ofloxacin otic solution versus neomycin sulfate/
polymyxin B sulfate/hydrocortisone otic suspension four times a day: a mul-
ticenter, randomized, evaluator-blinded trial to compare the efficacy, safety,
and pain relief in pediatric patients with otitis externa. Curr Med Res Opin
2006;22:1725-36
30. van Balen FA, Smit WM, Zuithoff NP, et al. Clinical efficacy of three common
treatments in acute otitis externa in primary care: randomised controlled trial.
BMJ 2003;327:1201-5
31. Wadsten CJ, Bertilsson CA, Sieradzki H, et al. A randomized clinical trial of two
topical preparations (framycitin/gramicidin and oxytetracycline/hydrocorti-
sone with polymyxin B) in the treatment of external otitis. Arch
Otorhinolaryngol 1985;242:135-9
32. Roland PS, Belcher BP, Bettis R, et al. A single topical agent is clinically
equivalent to the combination of topical and oral antibiotic treatment for otitis
externa. Am J Otolaryngol 2008;29:255-61
33. Madhusudhan KT, Counts C, Lody C, et al. Comparative in vitro activity of
three fluoroquinolones against clinical isolates by E test. Chemotherapy
2003;49:184-8
34. Roland PS, Pien FD, Schultz CC, et al. Efficacy and safety of topical cipro-
floxacin/dexamethasone versus neomycin/polymyxin B/hydrocortisone for
otitis externa. Curr Med Res Opin 2004;20:1175-83
35. Sander R. Otitis externa: a practical guide to treatment and prevention. Am
Fam Physician 2001;63:927-36
36. WMA Declaration of Helsinki – Ethical Principles for Medical Research
Involving Human Subjects. Seoul: WMA General Assembly, 2008
www.cmrojournal.com ! 2011 Informa UK Ltd