Veterinary Dermatology 2006, 17, 316

Review

Blackwell Publishing Ltd

A systematic review and meta-analysis of the efficacy and safety

of cyclosporin for the treatment of atopic dermatitis in dogs
JEAN STEFFAN*, CLAUDE FAVROT and RALF MUELLER
*Novartis Animal Health, Basel, Switzerland Clinic for Small Animal Internal Medicine, Dermatology Unit,
Vetsuisse Faculty, Zrich, Switzerland Medizinische Kleintierklinik, LudwigMaximilians University,
Munich, Germany
(Received 29 July 2005; accepted 10 September 2005)

Abstract The efficacy of cyclosporin A (CsA) for the treatment of canine atopic dermatitis was evaluated based
on the systematic review of prospective clinical trials published between 2001 and 2005. Ten studies with adequate
design characteristics were included. These studies enrolled 799 dogs, 672 (84%) treated with CsA, 160 (20%) with
placebo, 74 (9%) with oral glucocorticoids and 23 (3%) with antihistamines. Treatment duration varied from
2 weeks to 6 months. For safety analysis, data were available from 660 dogs. Lesion scores were improved from
baseline in the range of 30 52%, 53 84% and 5269% after 4, 6 and 16 weeks, respectively. The percentage of
dogs with only mild pruritus rose from 013% at inclusion to 32 59% and 46 90% after 4 and 12 weeks, respectively. In most studies, the frequency of CsA administration could be reduced to every other day in 40% to 50%
of patients after 4 weeks and to twice weekly in 2026% of the dogs after 1216 weeks. Meta-analysis confirmed
highly significant effects of CsA compared to placebo, but none between oral CsA and glucocorticoids. The initial
disease severity, age or body weight of subjects did not influence treatment success. Improvement by more than
50% over baseline of lesion scores was predictive of a better response during treatment maintenance. Vomiting
and soft stools/diarrhoea were the most frequent adverse events seen at least once during the studies. These
occurred in 25% and 15% of subjects, respectively. The frequency of each other type of adverse events was lower
than 2.1%. In summary, the administration of CsA for the treatment of canine AD was found to be as effective
as that of glucocorticoids, and adverse effects were minimal.

IN TRO D U CT I ON
Cyclosporin (ciclosporine, cyclosporin A, CsA) is a
cyclic oligopeptide macrolide that possesses immunomodulating properties from its capability to block
the activity of cytoplasmic calcineurin phosphatase,
thereby preventing the induction of genes encoding for
cytokines and their receptors.1 Consensus is emerging
that, in addition to their repressive effect on cytokine
gene transcription, calcineurin inhibitors also induce
the expression of several genes such as the one encoding transforming growth factor beta.1 As a result of
this unique mode of action, CsA exhibits strong antiallergic effects because of an inhibition of cytokineinduced activation of cells that initiate the cutaneous
immune response (e.g. Langerhans cells and lymphocytes) and also cells that mediate allergic reactions
(i.e. mast cell and eosinophils).
The clinical benefit of oral CsA for the treatment of
atopic dermatitis (AD) in humans was documented
first in 1987.2 A systematic review recently evaluated
the efficacy of oral CsA for the treatment of human
Correspondence: Jean Steffan, Novartis Animal Health, Basel,
Switzerland. E-mail: jean.steffan@ah.novartis.com

AD.3 Ten trials were identified, nine of which were

randomized. When data from three trials were pooled
with respect to itch scores,46 the mean difference in
scores compared to placebo was 15% (95% confidence
interval (CI): 923%).7 Altogether, CsA was found
to be effective for treatment of AD in humans at
dosages varying from 3 to 5 mg kg1 day1, although
how this efficacy compares to that of oral glucocorticoids remains unknown.7 Unfortunately, CsA
exhibits noticeable side effects in humans, and the risk
of adverse drug events and the stress of monitoring for
such events are likely to impact on quality of life.7
The efficacy of oral CsA in reducing skin lesions and
pruritus of canine AD was first reported in a small
open trial in 2001.8 Since then, numerous studies have
been performed in Europe, the USA, Japan and Australia. The objectives of this systematic review were to
evaluate the efficacy and safety of CsA for the treatment
of canine AD using data from all studies. Meta-analyses
(e.g. pooling of data) were performed when study
designs and interventions were homogeneous enough
to warrant such pooling. Whenever applicable, this
review was reported following the recommendations
of the Quality of Reporting of Meta-analyses
(QUOROM) statement.9

2006 Novartis Animal Health Inc. Journal compilation 2006 European Society of Veterinary Dermatology

J Steffan, C Favrot and R Mueller

M ETH O D S
Search strategy
To retrieve clinical trials evaluating the efficacy of oral
CsA in dogs with AD, a combination of information
sources was used. Only trials completed after 1990
were searched, and there was no language restriction.
Three electronic databases were screened to increase
the chance of detecting articles published in either medical
(Medline, ISI Web of Science) or veterinary journals
(ISI Web of Science, CAB Abstracts). All searches were
performed using variations of the following string:
#1 atop* or allerg*
#2 dog or dogs or canine
#3 cyclosporin* or ciclosporin*
#4 #1 and #2 and #3
To locate unpublished clinical trials presented at
international veterinary dermatology meetings, the
authors hand-searched the proceedings of the annual
congresses of the American Academy/College of
Veterinary Dermatology and European Society/
College of Veterinary Dermatology. Books or abstracts
from the four last World Congresses of Veterinary
Dermatology were similarly browsed.
In addition, a message was posted to the VetDerm
Internet list (20 January 2004) to request the identification and sharing of results of relevant clinical trials that
had not been presented or published at the time of searching. Finally, unpublished reports of clinical studies where
CsA was used to treat canine AD were obtained from
Novartis Animal Health (NAH). Individual data sets
were made available by all authors except for one study.10

Selection of studies
To be included in this systematic review, studies had to
satisfy the following criteria:
Study design. Clinical trials had to be prospective and
consist of a minimum of 2 weeks of continuous administration of CsA at oral dosages of 5 mg kg1 day1.
There was no restriction in the total length of the trial
beyond the minimal duration required.
Population. Clinical trials should have enrolled more
than 10 dogs affected with AD, as defined by the recent
American College of Veterinary Dermatology (ACVD)
task force on canine atopic dermatitis as a genetically
predisposed inflammatory and pruritic allergic skin
disease with characteristic clinical features and associated
most commonly with IgE antibodies to environmental
allergens.11 Diagnosis should have been made according
to standard-of-care methods, which could be summarized
by the documentation of clinical signs suggestive of the
disease12 as well as the elimination from consideration
of resembling pruritic inflammatory dermatoses.13

Intervention. Trials should have tested the efficacy of

CsA pro-emulsion concentrate, also known as microemulsified or modified USP CsA.
Principal outcomes. Trials should have reported a minima
an indication of the improvement from baseline of
investigator-assessed lesional or owner-evaluated pruritus
scores using well-defined methods of evaluation of such
outcome measures. Trials should also report adverse
drug events.

Data abstraction
Clinical trials that satisfied inclusion criteria were
reviewed separately by the authors who assessed quality
of study design, details of interventions and outcome
measures. Data were abstracted in tabular form. Results
of the review were compared, and where differences
were noted, they were discussed and reconciled. When
methodology or outcome results were not available, study
authors were contacted to request additional information
and sharing of individual data on enrolled subjects.

Validity assessment
Assessment of study design. The assessment of design
validity was performed by the evaluation of three
parameters, as described previously.14
1 Randomization. Method of generation and concealment of allocation to treatment groups.
2 Masking. Blinding of observers (e.g. investigators)
and participants (e.g. owners) to the treatment
allocation.
3 Loss-to-follow-up. Presence of dropouts and withdrawals and intention-to-treat (ITT) analyses.
These three parameters were graded in accordance
to the recommendations of the Cochrane Skin Group
as adequate, unclear or inadequate.15 When trials
were not randomized, none was the qualifier provided
for the randomization criterion.
Assessment of subject enrollment. For each of the selected
trials, inclusion criteria of subjects were reviewed to determine whether the diagnosis of AD was made according to
standard methods.12,13 The quality of subject selection
was evaluated for each trial as proposed previously:14
1 Well characterized was the term used when only
patients with AD were entered in the study, and
sufficient details on the methods of diagnosis of
AD (clinical signs, rule-out of concurrent or
resembling skin diseases, etc.) were provided to
allow comparison with current standards.13
2 Poorly characterized was used when participant
selection criteria were vague (e.g. uncharacterized
allergic pruritus), or when insufficient details of
the diagnostic work-up were provided.

2006 Novartis Animal Health Inc. Journal compilation 2006 European Society of Veterinary Dermatology

Cyclosporin for canine atopic dermatitis

3 Fairly characterized was used for intermediate
situations.

Assessment of outcome measures

Evaluation of efficacy during treatment induction. The
following six outcome measures were extracted from
all clinical trials:
1 Percentage change from baseline skin lesional
scores assessed by the investigators (clinicians).
2 Percentage of dogs with a 50% reduction
from baseline lesional scores assessed by the
investigators.
3 Percentage change from baseline pruritus scores
evaluated by the owners (if a continuous scale was
used).
4 Percentage of dogs reaching a 50% reduction
from baseline pruritus scores evaluated by the
owners (if a continuous scale was used).
5 Percentage of dogs with treatment identified as
successful using the investigators global assessment of treatment efficacy.
6 Percentage of dogs with treatment identified as
successful (good-to-excellent) using the owners
global assessment of treatment efficacy.
When a multipoint numerical scale was employed
by owners to grade their dogs pruritus levels, the
third and fourth outcome variables were replaced
by percentage of dogs with mild pruritus (e.g. dogs
with pruritus scores 2/5).
Evaluation of efficacy during treatment maintenance.
Four outcome variables were assessed in studies with
longer CsA administration (i.e. > 4 weeks):
1 Percentage change from baseline skin lesional
scores assessed by the investigators (clinicians).
2 Percentage of dogs with mild pruritus, identified
by a score 2/5 or 100/250.
3 Percentage of dogs receiving one-half or one quarter
of the induction dose.
4 Percentage of dogs with treatment identified as
successful (good-to-excellent) using owners and
investigators global assessment of treatment
efficacy.

Quantitative data synthesis and statistics

Intention-to-treat analytical rules. Outcome variables
were analysed using ITT rules, whereby data sets from

all study subjects were kept if at least one follow-up

visit was performed after enrollment. In case of missing
data, the last value observed carried forward procedure was applied. Such rigorous analytic scheme aims
at preventing an overestimation of treatment effect if
many subjects are withdrawn from the experimental
group. When not provided by the authors of the original studies, outcome measures were recalculated using
this rule for ITT analysis. For safety review, all enrolled
cases receiving at least one dose of the tested drugs
were accounted for.
Meta-analysis of randomized controlled trials. Whenever
trials reported data comparing the efficacy of CsA to
that of another intervention (placebo or other positive
control) and when there was homogeneity of study
designs, interventions and end-points, data were
pooled and analysed together using the statistical module of a systematic review software (RevMan 4.2; The
Cochrane Collaboration, Oxford, UK). For metaanalysis of continuous outcome variables (for example,
percentage reduction from baseline scores), the weighted
mean difference with bracketing 95% CI was determined
using the fixed effects model.16 When outcome measures
were not continuous (for example, percentage of dogs
with 50% reduction from baseline scores), the odds
ratios (ORs [Other Responders], with 95% CI) were
calculated using the same model.16
In these meta-analyses comparing the efficacy of
CsA to that of placebo and glucocorticoids, we defined
treatment success as the percentage of dogs achieving
a 50% reduction in lesional scores at the end of a
4- to 6-week induction phase.
Safety data were also extracted and reported for
each study included in this review. Only most commonly occurring adverse drug reactions (ADR) (i.e.
those occurring with a prevalence 0.75% of all treated
cases) were summarized. Statistical analyses were not
used for safety data because of inconsistencies in the
reporting of incidence rates and lack of standardization of terminology.

R E SU LT S
Search and selection of studies
As of 1 March 2004, our literature search identified 13,
19 and 5 articles in Medline, WEBOF Science and
CAB Abstracts databases, respectively. Review of these
published articles yielded six trials that met the inclusion
criteria.8,10,1720
Hand-searching of European and American conference proceedings identified two additional unpublished trials.21,22 Of these two studies, only the first one
met preset selection criteria.21 The second study was
not included after verification of patient data provided
by the author.22 Reasons for exclusion were the presence of variable enrollment criteria of study subjects
(presence of concurrent diseases in some patients), a
progressive increase in the dosage of CsA, the concurrent

2006 Novartis Animal Health Inc. Journal compilation 2006 European Society of Veterinary Dermatology

J Steffan, C Favrot and R Mueller

administration of various dosages of oral antipruritic

drugs to some dogs, the implementation of allergenspecific immunotherapy in some subjects, as well as an
imprecise assessment of treatment efficacy.
Finally, NAH provided the complete report of one
study23 of which only the abstract was available, as well
as two manuscripts accepted for publication in peer
reviewed journals.24,25 When data of the published versions of the articles did not allow adequate evaluation
of the trials, complete reports were obtained from
NAH and re-analysed using the standardized set of
outcome measures. Altogether, 10 clinical trials met
study inclusion criteria.8,10,1721,2325

Validity assessment
Among the selected studies, there were five randomized
controlled trials (RCTs)1719,21,24 and four open studies
reporting the efficacy of CsA without a control group
(Table 1)8,10,20,25 or in parallel comparison with antihistamines.23 Four of the five RCTs were blinded,17
19,24
whereas one used an open design.21 The generation
and concealment of the randomization scheme were
found to be adequate in all four blinded RCTs
(Table 1).1719,24 ITT analytical schemes were used in
all RCTs, whereas in only two of five open trials were all
enrolled subjects accounted for at study end (Table 1).8,20
In all, these 10 selected trials had enrolled 799 dogs with
AD characterized according to current standards.13
Of these study subjects, 672 were treated with
CsA (84%), 74 were prescribed oral glucocorticoids
(15 prednisolone, 59 methylprednisolone) (9%), 23
received a diphenhydramine-containing formulation
(3%) and 160 were given placebo (20%) (Table 1). In
one trial, some dogs received placebo in a first phase,
then they were given CsA.24 Finally, one trial limited its
enrollment to dogs with AD whose signs no longer
responded to 0.52 mg kg1 day1 of (methyl) prednisolone, or for whom glucocorticoid administration
had resulted in unacceptable adverse side effects.10
The duration of CsA administration varied among trials:
one lasted 2 weeks,8 four went on for 6 weeks,17,18,20,23
while five continued for 3 or 410,19,21,24 to 6 months.25
For safety analysis, data were available from 660 dogs
that had received at least one dose of cyclosporin.

Quantitative data synthesis

Efficacy of cyclosporin during treatment induction. Data
from all 10 selected studies8,10,1721,2325 were used to
assess the efficacy of CsA during the first 4 to 6 weeks
of treatment induction (Table 2). When multiple data
sets were reported, only those included in ITT analyses
were taken into account.
Overall Efficacy. The mean or median investigatorassessed percentage reduction from baseline
lesional scores varied between 30% and 52% after
4 weeks10,19,21,24 and between 53% and 84% after 6 weeks
(Table 2).17,18,20,23 In one small open study with a duration of 2 weeks, a reduction in lesional scores of up to
60% was found.8 After 4 to 6 weeks of CsA administration,

the percentage of dogs with a 50% reduction from

baseline lesional scores varied generally between 28%
and 71%, with two extremes of 20% and 95% (Table 2).
The lowest score achieved was in the study enrolling
dogs treated previously with glucocorticoids.10
The mean/median owner-graded percentage reduction from baseline pruritus scores fluctuated between
27% and 34% after 4 weeks10,19,21,24 and varied from
45% to 78% after 6 weeks (Table 2).17,18,20,23 In the first
small open trial that lasted 2 weeks, the improvement
in pruritus scores reached 100%.8
After 4 to 6 weeks of CsA administration, the
percentage of dogs with a 50% reduction from
baseline pruritus scores varied between 35% and 67%
(Table 2).1719,21,25 At the end of 46-weeks induction,
investigators graded the efficacy of this intervention as
successful in 5363% of subjects,18,20,23,24 whereas
owners provided a similar assessment for 4867% of
their pets.18,20,23,24
Efficacy versus placebo (meta-analysis). Two RCTs
reported the efficacy of 4- to 6-week treatment with CsA
compared to placebo (Table 1).18,24 In all, 253 dogs
were enrolled, 160 were treated with placebo and 163
received 5 mg kg1 of CsA once daily. The remaining 30
dogs were treated with 2.5 mg kg1 of CsA once daily.
Finally, in one study, 130 dogs received CsA once the
first placebo-controlled phase was completed.
Using pooled data from these two trials, the
weighted mean difference in percentage improvement
in lesional scores was 48% (95% CI: 4056%), with a
significant and homogeneous treatment effect for
CsA over placebo (Fishers z statistics; P < 0.00001).
After an induction phase of 4 to 6 weeks, the OR of
achieving a 50% reduction in lesional scores was 8.4
(95% CI: 4.615.5) for CsA relative to placebo. Similarly, the OR of achieving a mild pruritus score ( 2 on
a 5-point scale) was 6.9 (95% CI: 4.111.5) for CsA
over placebo. For both of these variables, CsA treatment
effect was found to be significant (z statistics; P < 0.00001)
and homogeneous between studies.
After treatment induction lasting 4 to 6 weeks, the
proportion of dogs achieving a 50% reduction of
lesional scores with CsA was 80/163 (49%) and those
with placebo was 20/160 (13%). Similarly, the proportion of dogs having mild pruritus ( 2/5 pruritus score)
was 98/163 (60%) with CsA and 31/160 (19%) with placebo.
Efficacy versus glucocorticoids (meta-analysis). Two
RCTs compared the efficacy of CsA and oral glucocorticoids (Table 1).17,19 Altogether, 206 dogs with AD
were enrolled in these two trials, 132 (64% of subjects)
received CsA, and 74 (36%) were given oral prednisolone (15 dogs) or methylprednisolone (59 dogs).
When pooling together the results of these two trials,
the weighted mean difference in percentage improvement in lesional scores was only 2% (95% CI: 95%)
in favour of glucocorticoids, a nonsignificant treatment
advantage (Fishers z statistics; P = 0.63). Both studies
yielded homogeneous results.

2006 Novartis Animal Health Inc. Journal compilation 2006 European Society of Veterinary Dermatology

Table 1. Selection of studies and validity assessment

Fontaine 20018

Iwasaki 200223

Olivry 200217

Olivry 200218

Design
Randomization (allocation generation)
Randomization (allocation concealment)
Masking of outcome assessor
Intention-to-treat analyses

Open
None
None
None
All dogs available
at trials end
Well characterized
14
14
2
Cyclosporin
5.0 mg kg1 q.i.d.

Open, parallel
None
None
None
Not all dogs available
at trials end
Well characterized
115
92
23
6
Cyclosporin
Diphenhydramine
5.0 mg kg1 q.i.d.
2.8 mg kg1 b.i.d.

Blinded RCT
Adequate
Adequate
Adequate
Adequate

blinded RCT
Adequate
Adequate
Adequate
Adequate

Well characterized
30
15
6
Cyclosporin
5.0 mg kg1 q.i.d.

Well characterized

Quality of inclusion of study subjects

No. of dogs with AD entered in trial (total)
No. of dogs with AD treated with CsA
Length of intervention (week)
Pharmacological intervention

91
15

30

31

30
6

Prednisolone
0.5 mg kg1 q.i.d.

cyclosporin
2.5 mg kg1 q.i.d.
5.0 mg kg1 QD

Placebo

Citation (reference)

Steffan 200319

Olivry 200321

Burton 200420

Bensignor 200410

Steffan 200524

Thelen 200525

Design

Blinded RCT

RCT

Open

Open

Open

Randomization (allocation generation)

Randomization (allocation concealment)
Masking of outcome assessor
Intention-to-treat analyses

Adequate
Adequate
Adequate
Adequate

Adequate
Inadequate
Inadequate
Adequate

Quality of inclusion of study subjects

No. of dogs with AD entered in trial (total)
No. of dogs with AD treated with CsA
Length of intervention (week)
Pharmacological intervention

Well characterized
176
117
16
Cyclosporin
5.0 mg kg1 q.i.d.,
e.o.d. with or
or TW

Well characterized
30
30
12
Cyclosporin with
two different
tapering
regimens

None
None
None
All dogs available
at trials end
Well characterized
41
41
6
Cyclosporin
5.0 mg kg1 q.i.d.

None
None
None
Not All dogs available
at trials end
Well characterized
15
15
12
Cyclosporin
5.0 mg kg1

Blinded RCT (phase I);

Open (phase II)
Adequate (phase I only)
Adequate (phase I only)
Adequate (phase I only)
Adequate
Well characterized
262
130
132
4 (phase I); 12 to 16 (phase II)
Placebo then Cyclosporin
cyclosporin* 5.0 mg kg1
q.i.d. (phase I)*

Well characterized
25
25
24
Cyclosporin
5.0 mg kg1 q.i.d.
without food

59
Methylprednisolone
0.51.0 mg kg1
q.i.d., e.o.d.
or TW

Adequate
Adequate
None
Adequate

Cyclosporin for canine atopic dermatitis

Abbreviations: b.i.d., twice daily; e.o.d., every other day; q.i.d., once daily; TW, twice weekly.
*Phase II: all dogs were treated with cyclosporin at 5.0 mg kg1 q.i.d., then e.o.d., then TW.

2006 Novartis Animal Health Inc. Journal compilation 2006 European Society of Veterinary Dermatology

Citation (reference)

J Steffan, C Favrot and R Mueller

After an induction phase of 4 to 6 weeks, the OR of

achieving a 50% reduction in lesional and pruritus
scores were 0.8 (95% CI: 0.41.4) and 0.7 (95% CI:
0.41.3) for CsA over glucocorticoids, respectively. For
both of these variables, CsA treatment effect over
glucocorticoids was found to be nonsignificant
(Fishers z statistics; P = 0.37 and 0.24, respectively).
Both RCTs provided homogeneous results. When combining results of these two studies, the proportion of
dogs achieving a 50% reduction of lesional scores was
58/132 (44%) with CsA and 39/74 (53%) with oral glucocorticoids. After treatment induction lasting 4 to
6 weeks, the proportion of dogs achieving a 50%
reduction of pruritus scores was 50/132 (38%) with
CsA and 36/74 (49%) with oral glucocorticoids.
Efficacy of cyclosporin during maintenance therapy
(3 to 6 months). Five studies evaluated the efficacy
of CsA for long-term treatment of canine AD
(Table 1).10,19,21,24,25 The duration of treatment lasted
3,10,21 419,24 or 6 months.25 After a treatment induction
of 4 weeks, there were 508 dogs available for repeated
evaluation: 449 subjects (88%) received CsA, whereas
59 (12%) were treated with oral methylprednisolone.
One hundred and thirty received placebo during the first
four weeks before being treated with CsA. In three trials,
the overall dose of CsA was reduced by increasing
the intervals between CsA administration from daily to
every other day (EOD) then twice weekly (TW), and
this decrease was based on the clinicians assessment
of the improvement in clinical signs.19,24,25 In one trial,
drug reduction was achieved by either increasing the
intervals between administration or by reducing the
daily dose.21 In the last open study, the dose of CsA was
not decreased during the 3-month follow-up.10
Overall efficacy. After 4 weeks of CsA administration,
the mean/median investigator-graded percentage
reduction from baseline lesional scores varied between
30%10 and 51% (increasing interval group21) (Table 3).
Of note is that the lowest improvement was obtained in
dogs deemed to have failure or unacceptable adverse
effects after glucocorticoid therapy.10 After 8, 12 or 16
weeks of CsA intake, usually with decreasing dose
regimens, the mean/median reduction of lesional
values varied between 53% and 69%, 5079% and 52
69%, respectively (Table 3).10,19,21,24,25
Altogether, the percentage of dogs achieving 50%
reduction from baseline lesional scores increased over
time from 20% to 60% after 4 weeks to 6387% after
1216 weeks (Table 3).10,19,21,24,25 The highest percentage was attained in dogs for which the dose of CsA was
not reduced during maintenance therapy.10
At study onset, the percentage of dogs with mild
pruritus fluctuated between 0% and 13%.19,21,24,25 After
4, 8, 12 and 16 weeks, these percentages increased to
3259%, 5187%, 4690% and 5193%, respectively
(Table 3). 19,21,24,25 Increases in pruritus were noted
in some dogs when the daily dose or frequency of
administration of CsA was reduced.

At the end of the two largest trials, the owners and

investigators evaluated treatment efficacy (Table 3).19,24
A good-to-excellent global assessment of efficacy
was reported in 6576% of subjects that completed the
study.
In four studies, when clinical signs of AD abated, the
overall dose of CsA was reduced.19,21,24,25 In two studies
in which a large number of dogs were included,19,24
the frequency of successful tapering were very similar.
During the second month of CsA therapy, a reduction
from daily to EOD administration was possible in
3850% of subjects because of 50% reduction in
clinical signs (Table 3). During the third month, such
reduction was achieved in 5058% of dogs, while this
was carried out in 3542% of subjects during the
fourth month of the trials. In dogs whose signs
decreased by 75% after the second month of treatment,
the frequency of CsA administration was decreased to
TW. Such frequency reduction was possible in 1114%
of patients during the third month and in 2126% on
the fourth month. In the two other studies,21,25 fewer
dogs were enrolled, and the distribution of cases according to the dosing regimen was more variable, but the
same pattern was found.
In one RCT, two tapering regimens were compared.21 In one group, 15 dogs received half, then a
quarter of the original daily dosage (DD group) of CsA
(5 mg kg1 once daily) upon achievement of 50% and
75% reduction from baseline lesional scores, respectively. In the other group of 15 dogs, the dose was
tapered by increasing intervals (II group) between
administration of the original dosage (5 mg kg1) EOD
or TW upon achievement of 50% and 75% reduction
from baseline lesional scores, respectively (Table 3).21
During the second month of treatment, halving of the
dose was implemented in 40% and 60% in the DD and
II groups, respectively. Quartering of the dose was
achieved in 40% (DD) and 20% (II) of dogs remaining
in the study during the third month. When comparing
improvement of signs following reduction by daily
dose decreases or administration interval increases,
there were no significant differences between regimens
in either clinical sign improvement, percentage of dogs
achieving dose reduction, adjusted daily drug dose
(e.g. treatment cost) or number of subjects dropping
out of the study.21
Efficacy versus glucocorticoids. Only one RCT compared
the efficacy of maintenance treatment with oral CsA
and methylprednisolone administered at increasing
intervals when clinical signs abated (Table 1).19 In this
trial, the percentage improvement over baseline of
lesional and pruritus scores was comparable and not
significantly different at 4, 8, 12 and 16 weeks after CsA
or methylprednisolone treatment initiation.
After 4 months of treatment, the proportion of
dogs achieving 50% reduction of lesional scores
was 77/117 (66%) with CsA and 34/59 (58%) with oral
methylprednisolone. At that time, the proportion of
dogs achieving 50% reduction of pruritus scores

2006 Novartis Animal Health Inc. Journal compilation 2006 European Society of Veterinary Dermatology

Cyclosporin for canine atopic dermatitis

Table 2. Outcome variables during treatment induction

Citation (reference)

Fontaine 20018

Iwasaki 200223

Olivry 200217

Olivry 200218

Steffan 200319

Induction end time point (weeks)

% of reduction from baseline lesional scores
(investigators)
% of dogs with 50% reduction from baseline
lesional scores (investigators)
% of reduction from baseline pruritus scores (owners)
% of dogs with 50% reduction from
baseline pruritus scores (owners)
% of dogs with mild pruritus (score 2/5)
Week 0
Week 4
% with global assessment of
treatment success (investigators)
% with global assessment of
treatment success (owners)

2
60% (30 7 3 )

6
53% (46 61)

6
58% (4374)

6
67% (53 78)

4
44% (4049)

NA

60%*

60%

71%

42%

100% (68 100)

NA

NA
NA

78% (52 87)

67%

45% (27 63)

48%

34% (2741)
35%

NA
NA
93%

6%
59%
63%

NA
NA
NA

13%
56%
61%

6%
47%
NA

NA

67%*

NA

61%

NA

Citation (reference)

Olivry 200321

Burton 200420

Bensignor 200410

Steffan 200524

Thelen 200525

Induction end time point (weeks)

% of reduction from baseline lesional scores
(investigators)
% of dogs with 50% reduction from
baseline lesional scores (investigators)
% of reduction from baseline pruritus
scores (owners)
% of dogs with 50% reduction from baseline
pruritus scores (owners)
% of dogs with mild pruritus (score 2/5)
Week 0
Week 4
% with global assessment of treatment success
(investigators)
% with global assessment of treatment
success (owners)

4
49% (41 56)

6
84% (79 8 9)

4 (e)
30% (20 40)**

4
45% (37 48)

4
37% (2846)

37%

95%

20%**

45%

28%

33% (27 43)

NA

27%

NA

53% (4264)

50%

NA

NA

NA

44%

NA
NA
NA

7%
90%
53%

NA
NA
NA

6%
41%
59%

0%
32%
NA

NA

48%

NA

48%

NA

Data presented are means or medians (95% CI).

*% of cases in which this parameter was assessed.
Values from group treated with cyclosporin at 5 mg kg1 once daily.
Values from both groups combined at week 4.
Evaluation at week 6.
Values from re-evaluation at week 4.
**Composite score from owner and investigator assessed values.

was 47/117 (40%) with CsA and 25/59 (42%) with oral
methylprednisolone.
Parameters predicting or modifying treatment outcome.
Data from one study24 were used to determine whether
initial disease severity, body weight and age could be
used to predict treatment outcome. There was no
statistically significant correlation between the initial
lesional or pruritus scores and the improvement of these
parameters after 4 weeks. When dogs were grouped
according to their body weight (less than 15 kg, from
15 to 29 kg and more than 29 kg), the response was
almost identical in the induction and maintenance
phase in all subgroups. The response to treatment in
terms of reduction of lesion score and pruritus was
almost identical in young (less than 4 years old), middleaged (46 years old) and old (> 7 years old) dogs.
One study tested the influence of the feeding status at
the time of administration, as was shown to slightly
reduce CsA bioavailability. Whether the drug was
administered or not with concurrent food intake did
not appear to affect outcome (Table 4).25

Results from three of the longest trials19,21,24 were

analysed to assess whether clinical improvement after
1 month of treatment induction could be predictive of
long-term outcome. In these four studies, the dose of
CsA was tapered according to the clinical response as
described previously.
Study subjects were divided according to the magnitude
of improvement after 1 month of CsA administration
at the induction dosage of 5 mg kg1 once daily. They
were classified as high responders (HR) if the lesional
score at week 4 had improved from baseline values by
50% or more. If the improvement in lesional scores
at week 4 was lower than 50%, they were grouped in
the other responders (OR) category. Lesional score
reduction during the first month of treatment was
found to predict the long-term response (Table 4). In
the two studies with the largest numbers of subjects,19,24 the frequency of study withdrawal was lowest
in HR. At study end (4 months), dogs from the HR
group had achieved a higher reduction in lesional
scores (69 75%) than those from the OR group (range
49 63%). The difference between these two groups

2006 Novartis Animal Health Inc. Journal compilation 2006 European Society of Veterinary Dermatology

10

Citation (reference)
% of reduction from baseline lesional scores
(confidence interval)
Week 4
Week 8
Week 12
Week 16

Steffan 200319

DD*
44% (4049)
53% (4759)
50% (4357)
52% (4459)

% of dogs with 50% reduction from baseline

lesional scores (investigations)

4
8
12
43% 66% 63%

% of dogs with mild pruritus

(score 2/5 or 100/250)
Week 0
Week 4
Week 8
Week 12
Week 16
% of dogs receiving reduced Doses during weeks
Half dose = e.o.d. or 2.5 mg kg1 q.i.d.
Quarter dose = TW or 1.25 mg kg1 q.i.d.
% of dogs with /good-to-excellent global
assessment of efficacy as assessed by
investigators and owners

6%
47%
51%
47%
51%
48 812 1216
50% 58% 42%
14% 26%
75%76%

Bensignor 200410

Olivry 200321

16 weeks
66%

With food
52% (4263)
54% (5173)
66% (5274)

NA
4
8
12 weeks 4
8
12 weeks
40% 73% 73%
60% 67% 80%

48 812
40% 33%
40%

Thelen 200525

II*

37% (3356)
69% (5575)
58% (4370)

7%
60%
73%
46%

Steffan 200524

13%
67%
53%
53%
NA
48 812
60% 47%
20%
NA

Abbreviations: e.o.d., every other day; NA, not available; q.i.d., once daily; TW, twice weekly; NA, not available.
*DD, decreasing dosage group; II, increasing interval group.
Data calculated with last value carry forward rule data presented as means or medians (95% CI).

30% (2040)
59% (3964)
79% (5383)
NA
4
8
12 weeks
20% 67% 87%

NA
NA
NA
NA
NA
NA
NA
NA
NA

45% (4148)
58% (5764)
57% (6370)
60% (6673)
4
45

8
68

12
64

12%
58%
64%
63%
64%
48 812 1216
38% 50% 35%
11% 21%
65%69%

16 weeks
68

Without food

37% (2846)
54% (4266)
60% (4971)
64% (5268)
4
8
12
26% 60% 73%

37% (1460)
53% (4074)
62% (4480)
56% (3577)

16 weeks
86%

0%
20%
80%
53%
80%
48 812 1216
7% 47% 13%
0% 13%

4
8
12
30% 60% 80%

0%
50%
60%
70%
60%
48 812 1216
30% 30% 40%
20% 70%
68%

16 weeks
60%

J Steffan, C Favrot and R Mueller

2006 Novartis Animal Health Inc. Journal compilation 2006 European Society of Veterinary Dermatology

Table 3. Outcome variables during treatment maintenance

Cyclosporin for canine atopic dermatitis

11

Table 4. Parameters predicting treatment outcome

Citation
(reference)

No. of dogs at study

end/# dogs at
4 weeks (%)

Mean percentage reduction

from baseline lesional
score at study end (SD)

Mean percentage reduction

from baseline pruritus
score at study end (SD)

% of dogs treated
with each administration
interval at study end

OR

OR

HR

OR

HR

q.i.d.-e.o.d.-TW q.i. d.-e.o.d.-TW

OR
HR

73 (27)
69 (9)*
74 (15)
75 (22)

41 (39)
31 (49)*
38 (38)
40 (44)

49 (34)
20 (66)*
51 (27)
53 (35)

39-41-20

43-43-14
33-37-29

HR

Steffan 200319 46/54 (85%)

Olivry 200321 13/14 (93%)
Steffan 200524

36/42 (90%)
11/15 (73%)

53 (29)
49 (27)*
52 (18)
74/104 (71%) 92/100 (92%) 63 (32)

17-44-39

0-55-45
23-43-33

Abbreviations: e.o.d., every other day; q.i.d., once daily; TW, twice a week; HR, high responders; OR, other responders.
*Decreasing dosage group. Increasing interval group.

Table 5. Adverse reactions observed during the studies

Fontaine Iwasaki Olivry
20018
200223
200217
Total number of cases
Treatment interruption
associated
with adverse events
Vomiting
Diarrhoea/soft stools
Miscellaneous gastrointestinal disorders
Decreased appetite or
anorexia
Otitis
Bacterial skin infection
Malassezia infection
Urinary tract infection
Lethargy/sleepiness
Nodules and cysts
Papillomatosis
Gingival hyperplasia
Lymphadenopathy
Various reproductive
disorders
Neurological disorders

14
0 (0%)

92
5 (5%)

2 (14%) 10 (11%)
13 (14%)

Olivry
200218

15
61
2 (13%) 0 (0%)

1 (7%) 23 (38%)
3 (20%) 6 (10%)
1 (7%)

Steffan
200219

Olivry
200321

117
10 (9%)

30
41
5 (17%) 0 (0%)

43 (37%)
21 (18%)
10 (9%)

1 (1%)
a
a
a

a
a
a

1 (3%)

Burton
200420

4 (10%)
4 (10%)

Bensignor Steffan
200410
200424
15
0 (0%)

5 (33%)

5 (4%)
a
a
a

a
a
a

a
a
a

a
a
a

a
a
a

82 (31%)
53 (20%)

1 (7%)
7 (6%)

1 (7%)
3 (3%)

a
a
a
10 (4%)
6 (2%)
3 (1%)
4 (1%)
6 (2%)
6 (2%)

171 (25%)
100 (15%)
10 (40%) 21 (3%)
14 (2%)
a
a
a

a
a
a
10 (2%)
8 (1%)
10 (2%)
5 (< 1%)
9 (1%)
6 (< 1%)
7 (1%)

7 (6%)
4 (1%)

Total

25
672
3 (12%) 36 (5%)

8 (3%)
a
a
a

1 (7%)

262
11 (4%)

Thelen
200525

1 (4%)

5 (< 1%)

a: Otitis, pyoderma and Malassezia dermatitis were reported in some trials as adverse effects. As these conditions are often associated with atopic
dermatitis and as most of the affected animals have exhibited the same signs before inclusions, these figures were not included in this table.

was significant in all studies (MannWhitney test;

P < 0.05) (Table 4). The mean improvement in pruritus
scores at study end was also higher in the HR group
compared to the OR category, but the difference was
not statistically significant (Table 4). At study end,
dogs from the HR group were treated less frequently
once daily (0 23% of dogs) than those from the OR
group (33 43%). Conversely, between 33% and 45% of
HR dogs received the medication only TW at study end
compared to 14 29% of OR dogs. The distribution of
dosing regimen was significantly different in the two
groups (MannWhitney test; P < 0.05) from either 8
1619 or 812 weeks (Table 4).24
Adverse events. The frequencies of common adverse
events reported in the original studies are reported in
Table 5. Vomiting and gastrointestinal disorders were
observed in all studies with an overall prevalence of 25%
and 18%, respectively. In two studies, the number of
occurrences of gastrointestinal disturbance per treated
dog was recorded.18,24 Between 53% and 71% of affected

dogs vomited between one to three times over a 120day period. Repeated episodes of vomiting were rare.
Similarly, long-lasting (e.g. lasting for more than
7 days) episodes of diarrhoea were considered rare in
both studies. Most of these gastrointestinal signs
were recorded during the first month of treatment.
Discontinuation of CsA because of unacceptable
gastrointestinal side effects was reported by Stefan
and coworkers in 2 out of 262 treated dogs.24 The
occurrence of skin and ear infections was not systematically recorded in all reports. Two studies suggested
that the frequency of skin infection19 or otitis24 decreased
in the first 3 to 4 months during treatment with CsA.
The frequency of other adverse events was lower than
2.1% (Table 5).

D IS C U S S IO N
In this article, the results of 10 clinical trials that
evaluated the efficacy of CsA administration for the

2006 Novartis Animal Health Inc. Journal compilation 2006 European Society of Veterinary Dermatology

12

J Steffan, C Favrot and R Mueller

treatment of canine AD were reviewed. Clinical inferences from this systematic review must be appraised in
light of the sources of potential bias.

Validity
Internal validity. The evaluation of the internal validity
of a systematic review relies on the critical assessment
of four separate sources of bias: selection, detection,
performance and attrition.26
In clinical trials evaluating the efficacy of an active
drug compared to a control medication, selection bias
occurs whenever treatment groups are not comparable
in regard to possible confounding factors.26 To alleviate this source of bias, subjects must be randomized
independently to a treatment group a cornerstone
step in the design of RCTs. The efficacy of randomization depends on the generation of allocation sequences
by truly random methods and from the adequate
concealment of treatment sequences.27 Without an
effective randomization of subjects to a treatment group,
selection bias is likely to occur.26
In this systematic review, 10 trials were kept for
scrutiny and meta-analysis.8,10,1721,2325 Four of 10
studies were RCTs that encompassed more than 500
subjects, approximately two-thirds of the total number
of patients enrolled in all studies.1719,23 One study was
an RCT during the initial treatment phase only.24 Four
RCTs were double blinded,1719,24 whereas treatment
allocation was not masked in the last RCT that evaluated
two dose-reduction schemes.21 The generation and
masking of randomization sequences were found to be
adequate in five and four RCTs, respectively. Overall,
treatment groups appeared to be homogeneous in
signalment and disease severity parameters at trials
onset. Altogether, selection bias is unlikely to have
occurred in the RCTs reviewed herein.
In one open parallel study,23 selection bias occurred
because dogs with mild lesions were treated with the
antihistamine-containing reference product, whereas
dogs with more severe signs were given CsA. Because
of the difference in disease severity at inclusion in the
two treatment groups, these data were not used for
comparing response between these two interventions.
In six studies reviewed herein,8,10,20,21,23,25 the assessment of treatment outcome was not blinded, and these
trials were thus more susceptible to detection bias.26
Indeed, in two of these open studies,8,20 the assessment
of efficacy at the end of treatment induction was higher
than that of the two largest blinded RCTs.19,24
In the trials evaluated in this review, evidence of
performance bias (e.g. one group of dogs was treated
preferentially with medications in addition to that
being evaluated) was not detected.
Attrition bias can occur because of deviations from
protocol and/or loss to follow-up.26 Violation of eligibility
criteria and nonadherence to treatment are possible causes
of deviation from protocol. In most trials included in
this review, occasional deviations from protocol occurred,
and these errors led to the elimination of study sub-

jects. In three studies, owner compliance was assessed

by measuring CsA blood levels and /or careful examination of owners prescription records.18,19,24
Loss to follow-up refers to the inavailability of study
subjects because they refuse to participate (e.g. drop out),
are unreachable at some end points or adverse events
or medical issues warrant their withdrawal for health
reasons.26 To decrease attrition bias caused by heavy loss
to follow-up, ITT analytical schemes are employed.28
In the five RCTs scrutinized in this review,1719,21,24
all study dogs were included in ITT analyses that were
reported solely or in addition to the evaluation of per
protocol data sets limited to subjects completing the
trial. In only two of the open studies were some dogs
unaccounted for at the end of the study.10,23 Pooling all
trials together, the attrition of subjects only totaled 21
of 672 dogs (3%), a minute amount that is unlikely to
add noticeable overestimation of CsA treatment effect.
In summary, the internal validity of this systematic
review is affected by the nonrandomized design of 5 of
10 trials, the selection prejudice in one trial (15% of
total subject number) and the possible detection bias
in five open studies (22% of total subject number).
Clinical inference based on the results of this systematic
review must take the parameters above in consideration.
External validity. All studies included in this review
were completed after 1999, and the quality of enrollment of subjects was found to be high, an improvement
compared to that of earlier clinical trials evaluating
other interventions (reviewed in Olivry and Mueller14).
All studies limited their selection of subjects to dogs
diagnosed with AD according to standard methods.12,13
In general, the criteria for diagnosis of AD included the
presence of suggestive clinical signs as well as the elimination from consideration of resembling or coexisting
pruritic skin diseases. In most studies, documentation
of the existence of allergen hypersensitivity was also
required. The majority of dogs enrolled in the various
studies had been referred to veterinary dermatologists
working at private or public institutions. However,
review of signalment and disease characteristics at the
time of enrollment did not identify any reasons to
believe that the dogs selected were not representative of
the population of atopic dogs seen by both generalist
and specialist veterinarians.
All trials selected in this review evaluated the effect
of CsA and other inactive or active controls on a constellation of signs and symptoms relevant to the extent
and severity of canine AD.14 Both skin lesion and
pruritus reduction were taken into consideration. Remarkably, 7 of 10 studies used the second version of
the canine atopic dermatitis extent and severity index
(CADESI) scoring system, therefore providing a comparable evaluation of skin lesions.1719,21,2325 Advantages
and disadvantages of such a scoring system have been
discussed previously.14 In the two largest trials, owners
and investigators also reported a global assessment of
treatment efficacy, a subjective measure that is likely to
represent the owners satisfaction with CsA-induced

2006 Novartis Animal Health Inc. Journal compilation 2006 European Society of Veterinary Dermatology

Cyclosporin for canine atopic dermatitis

alleviation of signs and increase in the pets quality of
life. Of note is that most trials discussed herein were of
relative short duration, with only four studies lasting
3 months or longer.19,21,24,25 Unfortunately, such short
trials do not provide any applicable information on the
long-term efficacy of CsA for treatment of canine AD,
a chronic recurrent disease likely to last many years.
Overall, the nature of enrolled subjects, their disease
parameters and the outcome measures evaluated were
felt to be compatible with the extrapolation of the
results of this systematic review to the short- to mediumterm treatment of the entire canine atopic population.

Publication bias
During the process of systematic review, publication
bias can occur because of the type of studies being selected,
the existence of any language barriers or inadequate
reporting of outcome variables.26
In this systematic review of the efficacy of CsA for the
treatment of canine AD, publication bias was felt to be
negligible because of (i) the inclusion of all clinical
trials identified in the three major medical and veterinary databases, (ii) the addition of unpublished trials
identified via a widely used internet list and the handsearching of major dermatology meetings abstract
booklets and (iii) the lack of restriction in languages in
which studies were written. One could argue that the
elimination of one study reporting negative results22
might be an obvious proof of publication bias. Unfortunately, multiple confounding factors were identified
in that report after scrutiny of individual subject information provided by the author. Indeed, the enrollment
of dogs with concurrent pruritic diseases and the addition of oral glucocorticoids to the treatment regimen of
several dogs were felt likely to confuse the evaluation
of the real efficacy of CsA therapy. Such issues were
believed to represent sensible causes for not including
this retrospective study in the present evaluation.
Finally, all trials included herein were initiated after
1997, and the quality of outcome variable selection
and reporting was found to be substantially higher
than that identified in the previous systematic review of
pharmacological treatment of canine AD.14 When
information on study design, enrollment or attrition of
subjects or outcome measures was found to be insufficient, study authors were contacted to provide case
details. In all but one instance, the requested information was provided.
In summary, it is felt that, if present, publication bias
was likely to be minimal, and that it should not influence
the generalization of results of this systematic review.

Implications for practice

The examination of the results of this systematic review
offers useful information on the use of CsA for treatment of canine AD. After the administration of CsA at
5 mg kg1 once daily for 4 to 6 weeks to dogs with AD
(induction phase), the extent and severity of skin
lesions can be expected to decrease by at least 40% and
that of pruritus by a minimum of 30%. At that time,

13

between one- and two-thirds of patients will experience

an improvement by half or greater in the severity of
their signs. These beneficial effects of CsA might be
lower in dogs with previous glucocorticoid treatment.
After the induction phase, the continuous administration of CsA should lead to a further reduction in
clinical signs of AD, eventually reaching a plateau
between 2 and 4 months after treatment initiation. The
magnitude of the beneficial effect should reach 50
70% reduction in skin lesions. Up to 85% of dogs could
experience a halving of the extent and severity of their
skin lesions, and approximately half of the subjects
could be expected to have mild pruritus.
After a beneficial treatment effect is obtained, the
administration of CsA may be decreased. During
maintenance therapy, approximately one-third to onehalf of the dogs will receive half of the original dose,
and one of four to five patients will receive one-fourth
of the induction dose. The dose reduction scheme most
often tested was an increase in the interval between two
doses, with EOD administration implemented upon
reaching 50% reduction from baseline skin lesions
and TW administration initiated upon attaining the
75% lesion improvement benchmark. Cost reduction by
halving and quartering the daily dose, upon reaching
the benchmarks stated previously, is not expected to
lead to substantial differences in treatment efficacy
compared to increasing interval protocols. Of note is
that CsA dose reduction causes worsening of signs in
some patients, but there is no evidence that signs of AD
will fail to improve again upon re-increase of the dose.
Response to therapy after 4 weeks appears to be predictive of treatment efficacy after 1216 weeks. Dogs in
whom lesion severity and extent improve by more than
50% after 4 weeks of CsA administration are more
likely to be maintained on treatment than those with
improvement of lower magnitude. They are also more
likely to have greater improvement in skin lesions, and
they are more likely to be given the drug TW than once
daily. However, further improvement in pruritus
should not be expected after dose tapering in dogs with
good initial improvement.
Oral CsA is more effective than placebo for reducing
signs of AD in dogs. In contrast, the efficacy of CsA
and orally administered glucocorticoids (prednisolone/
methylprednisolone) appears to be comparable, both
in terms of improvement of clinical signs and percentage
of dogs achieving half of lesional and pruritus scores.
At the dosage of 5 mg kg1, CsA appears to be a safe
short-term treatment of dogs with atopic dermatitis.
Gastrointestinal adverse reactions are the most common adverse events and are probably drug-related reactions. They are usually mild and do not need treatment.
They very rarely require discontinuation of this drug.
Other adverse events have been reported but they are
uncommon. Because they are infrequent, causality
assessment is difficult to make. When drug related, most
of these reactions also appear to be dose dependent,
and they resolve with dose reduction or treatment
discontinuation.

2006 Novartis Animal Health Inc. Journal compilation 2006 European Society of Veterinary Dermatology

14

J Steffan, C Favrot and R Mueller

Implications for research

Limitations

This systematic review identified several questions that

need to be answered with clinical trials. For example, as
AD is a chronic disease of many years duration, the
effect (and risk) of long-term treatment needs to be
assessed. Trials evaluating the efficacy and harm of
treatment with CsA with decreasing dose regimens for
periods covering 6 to 12 months, as well as for the
second year of administration, would be desirable. A
randomized controlled trial design may be inappropriate
because of the length of such trials. Open cohort studies
may be better suited. One such study has recently been
published.29 In addition, the evolution of canine AD
years after discontinuation of CsA administration must
be assessed further.
CsA is likely to be administered as second-tier pharmacotherapy in dogs with AD because of its high cost
and, after standard-of-care treatment with oral glucocorticoids has failed or led to severe adverse drug
events. The efficacy of CsA in dogs with AD refractory
to oral glucocorticoids must be confirmed, and such
effect must be differentiated from the treatment with
CsA of dogs with unacceptable side effects of previous
glucocorticoid therapy.
Finally, a potential synergism between oral CsA
administration and the topical application of medicated
shampoos, glucocorticoid or calcineurin inhibitor
formulations must be evaluated. Both benefit versus
cost and benefit versus risk analyses should be
assessed.
Because several drugs (e.g. ketoconazole, erythromycin) increase blood concentration of CsA,30 the
addition of such drugs to a treatment protocol could
lead to substantial reduction in doses of CsA needed
for optimal effect, thereby decreasing overall treatment
cost. Such cost reduction (the degree of which depends
on the size of the dog and the country of practice) has
been shown previously for dogs with perianal fistulas
treated with oral ketoconazole and CsA.31,32 Combination protocols must be evaluated carefully, and the
impact of the addition of these drugs on the benefit,
risk and cost of CsA treatment must be assessed. Dose
adjustment and cyclosporin blood monitoring may be
required, thereby hampering the benefit of treatment
cost reduction.
The response to cyclosporin therapy was not
influenced by body weight, but breed influence could
not be assessed because of the large number of breeds
included in the studies. Insufficient numbers were
available to conduct subpopulation analysis. Further
studies may be required to evaluate if there are breed
differences in response.
Quality of life is generally used in human clinical
studies to assess treatment benefit for the patient. This
has not yet been used in trials on canine dermatological
diseases. Quality of life parameters, such as loss
of sleep, pain, lethargy/depression or behavioural
changes should be considered in future studies, if such
parameters can be scored objectively enough by pet
owners.

Meta-analysis of safety parameters requires a standardization of techniques. A precise methodology has

been designed for efficacy assessment, but no methodology is available for safety parameters. Limitations
come from the definition of adverse events, standardization of terms and measurement methods.
An adverse event is defined as any observation in
animals that is unfavourable and unintended and occurs
after the use of the investigational veterinary product,
whether or not considered to be product related
(VICH Guideline on Good Clinical Practices CVMP/
VICH/595/98). When the unfavourable or unintended
event is associated with the normal course of the treated
disease, it is difficult to assess if the event is treatment
or disease related. In the case of AD, bacterial skin
infections, Malassezia infections or otitis are very common. They may or may not be considered as adverse
drug events when occurring during the course of the
treatment. Some investigators may have recorded these
events as adverse, whereas others may not have assessed
these events as drug related.
In case of chronic conditions such as canine AD,
extended observation periods are needed to assess
long-term adverse effects of any intervention. Patients
may develop signs of intercurrent diseases that evolve
over time before a final diagnosis can be made. In such
cases, several signs are listed for a single individual,
actually describing a single disease entity. For example,
the lethargy observed in several dogs may have resulted
from intercurrent disease.
Finally, the report of the incidence of adverse drug
events only partially reflects the detrimental effects of
the intervention on patients. Indeed, adverse events
occurring once or repeatedly through the period of
medication are usually similarly accounted for, although
the medical and economical consequences of single
and multiple events are greatly different.

C O N C LU S IO N S
Altogether, this systematic review and meta-analysis of
10 clinical trials enrolling approximately 800 dogs with
AD provides strong evidence for the efficacy of oral
CsA for reducing skin lesions and pruritus in dogs with
AD. Treatment effect with CsA is comparable to that
of the oral glucocorticoids prednisolone and methylprednisolone. Risk appears to be minimal for treatment
lasting less than 6 months.

D ISC LO SU R E O F P O T E N T IA L
C O N F L IC T O F IN T E R E ST
Jean Steffan is employed as International Project
Leader by Novartis Animal Health in Basel, Switzerland.
Claude Favrot has received lecturing honorarium
and research support from Novartis Animal Health,
but otherwise reports no conflict of interest.

2006 Novartis Animal Health Inc. Journal compilation 2006 European Society of Veterinary Dermatology

Cyclosporin for canine atopic dermatitis

Ralf Mueller has been consulting for Novartis
Animal Health since 2000, and this company provided
funding for a clinical trial at Colorado State University.
The trial was conducted at that location after Dr Muellers
departure to Ludwig-Maximilians-University in Munich.

ACKN OWLEDGE ME NT S
The authors thank Dr Thierry Olivry from North
Carolina State University for his extensive assistance in
writing this manuscript and collating the data, and
Wolfgang Seewald and Gnther Strehlau from Novartis
Animal Health in Basel for their help in the performance
of statistical analyses required for this review.

REFEREN CE S
1. Mascarell L, Truffa-Bachi P. New aspects of cyclosporin
a mode of action: from gene silencing to gene up-regulation.
Mini Reviews in Medicinal Chemistry 2003; 3: 205
14.
2. van Joost T, Stolz E, Heule F. Efficacy of low-dose
cyclosporine in severe atopic skin disease. Archives of
Dermatology 1987; 123: 166 7.
3. Hoare C, Li Wan Po A, Williams H. Systematic review
of treatments for atopic eczema. Health and Technology
Assessment 2000; 4: 1191.
4. Wahlgren CF, Scheynius A, Hagermark O. Antipruritic
effect of oral cyclosporin A in atopic dermatitis. Acta
Dermato-Venereologica 1990; 70: 323 9.
5. Sowden JM, Berth-Jones J, Ross JS et al. Double-blind,
controlled, crossover study of cyclosporin in adults with
severe refractory atopic dermatitis. Lancet 1991; 338:
137 40.
6. Munro CS, Levell NJ, Shuster S et al. Maintenance treatment with cyclosporin in atopic eczema. British Journal
of Dermatology 1994; 130: 376 80.
7. Williams H, Thomas K, Smethurst D et al. Atopic
eczema. In: Williams H, Rzany B. eds. Evidence-Based
Dermatology. London, UK: BMJ Books, 2003: 144 218.
8. Fontaine J, Olivry T. Treatment of canine atopic dermatitis with cyclosporine: a pilot clinical study. Veterinary
Record 2001; 148: 662 3.
9. Moher D, Cook DJ, Eastwood S et al. Improving the
quality of reports of meta-analyses of randomised controlled trials: the QUOROM statement. Lancet 1999;
354: 1896 900.
10. Bensignor E, Guagure E. Utilisation de la ciclosporine
pour le traitement des formes rebelles de dermatite
atopique canine [Use of cyclosporine for treatment of
refractory canine atopic dermatitis]. Pratique Medicale
et Chirurgicale de L Animal de Compagnie 2004; 39: 15
19.
11. Olivry T, DeBoer DJ, Griffin CE et al. The ACVD task
force on canine atopic dermatitis: forewords and lexicon.
Veterinary Immunology and Immunopathology 2001;
81: 143 6.
12. Griffin CE, Deboer DJ. The ACVD task force on canine
atopic dermatitis (XIV): clinical manifestations of canine
atopic dermatitis. Veterinary Immunology and Immunopathology 2001; 81: 255 69.

15

13. DeBoer DJ, Hillier A. The ACVD task force on canine

atopic dermatitis (XV): fundamental concepts in clinical
diagnosis. Veterinary Immunology and Immunopathology
2001; 81: 2716.
14. Olivry T, Mueller RS. Evidence-based veterinary dermatology: a systematic review of the pharmacotherapy of
canine atopic dermatitis. Veterinary Dermatology 2003;
14: 12146.
15. Leonard T, Williams H, Delamere F. Cochrane skin
group. The Cochrane Library, vol issue 4. Oxford, UK:
Update Software, 2002.
16. Koretz RL. Methods of meta-analysis: an analysis. Current Opinion in Clinical Nutrition and Metabolic Care
2002; 5: 46774.
17. Olivry T, Rivierre C, Jackson HA et al. Cyclosporine
decreases skin lesions and pruritus in dogs with
atopic dermatitis: a blinded randomized prednisolonecontrolled trial. Veterinary Dermatology 2002; 13: 77
87.
18. Olivry T, Steffan J, Fisch RD et al. Randomized controlled trial of the efficacy of cyclosporine in the treatment of atopic dermatitis in dogs. Journal of the
American Veterinary Medical Association 2002; 221:
3707.
19. Steffan J, Alexander D, Brovedani F et al. Comparison
of cyclosporine A with methylprednisolone for treatment
of canine atopic dermatitis: a parallel blinded randomized controlled trial. Veterinary Dermatology 2003;
14: 1122.
20. Burton G, Burrows A, Walker R et al. Efficacy of
cyclosporine in the treatment of atopic dermatitis in
dogs: combined results from two veterinary dermatology
referral centres. Australian Veterinary Journal 2004; 82:
68185.
21. Olivry T, Rivierre C, Murphy KM et al. Maintenance
treatment of canine atopic dermatitis with cyclosporine:
decreasing dosages or increasing intervals? (Abstract).
Veterinary Dermatology 2003; 14: 220.
22. Bloom PB. A prospective study of micro-emulsified
cyclosporine A for the treatment of atopic dermatitis in
24 dogs: the Michigan experience (Abstract). Veterinary
Dermatology 2003; 14: 210.
23. Iwasaki T, Nagata M, Masuda K et al. The clinical trial
of cyclosporin in the treatment of canine atopic dermatitis:
a non-blinded randomized antihistamine-controlled trial
(Abstract). Proceedings of the 23rd Annual Meeting of
the Japanese Society of Clinical Veterinary Medicine.
Osaka, Japan 2002, 904.
24. Steffan J, Parks C, Seewald W et al. Clinical trial evaluating the efficacy and safety of cyclosporine in dogs with
atopic dermatitis. Journal of the American Veterinary
Medical Association 2005; 226: 185563.
25. Thelen A, Mueller R, Linek M et al. Administration of
cyclosporine with food does not influence the clinical
response in atopic dogs. Veterinary Record, in press.
26. Jni P, Altman DG, Egger M. Systematic reviews in
health care: assessing the quality of controlled clinical
trials. British Medical Journal 2001; 323: 426.
27. Altman DG. Randomisation: essential for reducing bias.
British Medical Journal 1999; 302: 14812.
28. Guyatt G, Cook D, Devereaux PJ et al. Therapy. In:
Rennie D ed. Users Guides to the Medical Literature
Essentials of Evidence-Based Clinical Practice. Chicago,
IL: AMA Press, 2002: 81120.
29. Radowicz SN, Power HT. Long-term use of cyclosporine

2006 Novartis Animal Health Inc. Journal compilation 2006 European Society of Veterinary Dermatology

16

J Steffan, C Favrot and R Mueller

in the treatment of canine atopic dermatitis. Veterinary

Dermatology 2005; 16: 81 6.
30. Guagure E, Steffan J, Olivry T. Cyclosporin A: a new
drug in the field of canine dermatology. Veterinary
Dermatology 2004; 15: 61 74.
31. Mouatt JG. Cyclosporin and ketoconazole interaction

for treatment of perianal fistulas in the dog. Australian

Veterinary Journal 2002; 80: 20711.
32. Patricelli AJ, Hardie RJ, McAnulty JE. Cyclosporine
and ketoconazole for the treatment of perianal fistulas
in dogs. Journal of the American Veterinary Medical
Association 2002; 220: 100916.

Rsum Lefficacit de la cyclosporine A (CsA) pour le traitement de la dermatite atopique canine a t value
en se basant sur une revue systmatique des essais cliniques prospectifs publis entre 2001 et 2005. Dix tudes
avec un protocole correct ont t inclues. Ces tudes concernaient 799 chiens, 672 (84%) traits avec la CsA, 160
(20%) avec un placebo, 74 (9%) avec des glucococorticoides oraux, et 23 (3%) avec des antihistaminiques. La dure
de traitement a vari de deux semaines six mois. Pour les donnes de scurit, les donnes de 660 chiens ont t
tudies. Les scores lsionnels ont t amliors de 3052%, 5384% et 5269% aprs quatre, six et 16 semaines
respectivement. Le pourcentage de chien prurit lger est pass de 013% linclusion 3259%, et 4690% aprs
quatre et 12 semaines respectivement. Dans la plupart des tudes, la frquence dadministration de la CsA a pu
tre rduite tous les deux jours chez 40% 50% des patients aprs quatre semaines et deux fois par semaine
chez 20 26% des chiens aprs 12 16 semaines. Une mta-analyse a confirm les effets hautement significatifs
de la CsA en comparaison avec le placebo, mais pas de diffrence entre la CsA et les glucocorticoides. La svrit
de la maladie, lge ou le poids des animaux trauts ninfluenait pas le taux de succs. Une amlioration de plus
de 50% tait prdicitive dune meilleure rponse pendant le traitement dentretien. Des vomissements et des
diarrhes/selles molles taient les effets secondaires les plus frquents. Ils apparaissaient chez 25% et 15% des
animaux traits respectivement. La frquence des autres types deffets secondaires observs tait infrieure 2.1%.
En rsum, ladministration de CsA pour le traitement de la DA canine est aussi efficace que les glucocorticoides
et les effets secondaires sont minimes.
Resumen Se evalu la eficacia de cyclosporina A para el tratamiento de dermatitis atpica canina, basndonos
en una revisin sistemtica de pruebas clnicas prospectivas publicadas entre 2001 y 2005. Se incluyeron diez
estudios que posean adecuadas caractersticas en su desarrollo experimental. Estos diez estudios analizaron
799 perros, 672 (84%) tratados con ciclosporina A, 160 (20%) tratados con placebo, 74 (9%) tratados con
glucocorticoides por va oral, y 23 (3%) tratados con antihistamnicos. La duracion del tratamiento oscil entre
dos semanas y seis meses. El anlisis de las complicaciones derivadas del tratamiento estuvo disponible en 660
perros. El valor asignado a las lesiones mejor con respecto al valor inicial en un rango del 3052%, 5394% y
5269% tras cuatro, seis y 16 semanas, respectivamente. El porcentaje de perros slo con prurito minimo ascendi
desde 013% al inicio hasta 3259%, y 4690% tras cuatro y doce semanas, respectivamente. En la mayora de
los estudios la frecuencia de la administracin de ciclosporina A pudo reducirse a das alternos en un 40 a un
50% de los pacientes tras cuatro semanas, y a dos veces por semana en un 2026% de los perros tras 1216 semanas.
El anlisis estadstico confirm efectos altamente significativos de la ciclosporina A comparada con placebo, pero
no entre ciclosporina A oral y glucocorticoides. Ni la intensidad inicial de la enfermedad, ni la edad, ni el peso
corporal influyeron en el xito de la respuesta. Una mejora de un 50% sobre el valor inicial de las lesiones fue
pronostico de una mejor respuesta en la fase de mantenimiento. Vmitos y heces blandas/diarrea fueron los efectos
adversos mas frecuentes observados al menos una vez durante las pruebas. Estos signos ocurrieron en un 25%
y un 15% de los perros, respectivamente. La frecuencia de cada uno de los otros efectos adversos fue menos de
un 2.1%. En resumen, la administracin de ciclosporina A para el tratamiento de dermatitis atpica canina fue
tan efectiva como la administracin de glucocorticoides y con mnimos efectos adversos.
Zusammenfassung Die Wirksamkeit von Cyclosporin A (CsA) fr die Behandlung von caniner atopischer Dermatitis wurde, basierend auf einer systematischen Review prospektiver klinischer Studien, die zwischen 2001 und
2005 publiziert worden sind, evaluiert. Zehn Studien mit passenden Design Charakteristika wurden einbezogen.
Diese Studien protokollierten 799 Hunde, von denen 672 (84%) mit CsA, 160 (20%) mit Plazebo, 74 (9%) mit
oralen Glukokortikoiden und 23 (3%) mit Antihistaminen behandelt worden waren. Die Behandlungsdauer
variierte von zwei Wochen bis zu sechs Monaten. Fr die Sicherheitsanalyse standen Daten von 660 Hunden zur
Verfgung. Die sogenannten Lesion scores verbesserten sich von den Ausgangswerten in einem Bereich von
3052%, 5384% und 5269% nach vier, sechs bzw. sechzehn Wochen. Der Prozentsatz der Hunde mit nur mildem
Juckreiz stieg von 013% bei der Aufnahme bis zu 3250% nach vier bzw. 4690% nach 12 Wochen. In den meisten
Studien konnte die Hufigkeit der Verabreichung von CsA bei 4050% der Patienten nach vier Wochen auf jeden
zweiten Tag reduziert werden und auf zweimal wchentlich bei 2026% der Hunde nach 12 bis 16 Wochen. Eine
Meta-Analyse besttigte einen hochsignifikanten Unterschied von CsA im Vergleich zu Plazebo, aber keinen
zwischen oralem CsA und Glukokortikoiden. Der ursprngliche Erkrankungsgrad, das Lebensalter oder das
Krpergewicht der Individuen beeinflussten den Behandlungserfolg nicht. Eine Verbesserung von mehr als 50%
im Vergleich zu den ursprnglichen Lesion scores ermglichte die Vorhersage einer besseren Reaktion whrend
der Erhaltungstherapie. Erbrechen und weicher Kot/Durchfall waren die hufigsten Nebenwirkungen, die mindestens einmal whrend der Studien gesehen wurden. Diese kamen bei 25% bzw. bei 15% der Individuen vor. Die
Hufigkeit von allen anderen Arten der Nebenwirkungen war niedriger als 2.1%. Zusammenfassend wurde die
Gabe von CsA fr die Behandlung von caniner AD fr genauso effektiv befunden wie die Gabe von Glukokortikoiden, wobei die Nebenwirkungen minimal waren.
2006 Novartis Animal Health Inc. Journal compilation 2006 European Society of Veterinary Dermatology