International Journal of Neuropsychopharmacology (2011), 14, 261–268.

f CINP 2010 R E V IE W
doi:10.1017/S146114571000115X

Efficacy of escitalopram compared to

citalopram: a meta-analysis

Stuart Montgomery1, Thomas Hansen2 and Siegfried Kasper3

1
University of London, UK
2
H. Lundbeck A/S, Denmark
3
Department of Psychiatry and Psychotherapy, Medical University Vienna, Austria

Abstract
The aim of this review was to assess the clinical relevance of the relative antidepressant efficacy of escit-
alopram and citalopram by meta-analysis. Studies in major depressive disorder (MDD) with both escit-
alopram and citalopram treatment arms were identified. Adult patients had to meet DSM-IV criteria for
MDD. The primary outcome measure was the treatment difference in Montgomery–Asberg Depression
Rating Scale (MADRS) total score at week 8 (or last assessment if <8 wk). Secondary outcome measures
were response (o50 % improvement from baseline) and remission (MADRS f12). A search of the litera-

Downloaded from http://ijnp.oxfordjournals.org/ by guest on January 17, 2017

ture and websites found eight randomized controlled trials (RCTs) and onr naturalistic trial, with a total of
2009 patients (escitalopram, n=995 ; citalopram, n=1014). Escitalopram was significantly more effective
than citalopram in overall treatment effect, with an estimated mean treatment difference of 1.7 points at
week 8 (or last assessment if <8 wk) on the MADRS (95 % CI 0.8–2.6, p=0.0002) (six RCTs used the
MADRS), and in responder rate (8.3 percentage points, 95 % CI 4.4–12.3) (eight RCTs) and remitter rate
(17.6 percentage points, 95 % CI 12.1–23.1) analyses (reported for four RCTs), corresponding to number-
needed-to-treat (NNT) values of 11.9 (p<0.0001) for response and 5.7 (p<0.0001) for remission. The
overall odds ratios were 1.44 (p<0.0003) for response and 1.86 (p<0.0001) for remission, in favour of
escitalopram. In this meta-analysis, the statistically significant superior efficacy of escitalopram compared
to citalopram was shown to be clinically relevant.
Received 18 January 2010 ; Reviewed 10 May 2010 ; Revised 13 July 2010 ; Accepted 19 August 2010 ;
First published online 29 September 2010
Key words : Escitalopram, MADRS, MDD, NNT, response.

Introduction (Boulenger et al. 2006 ; Montgomery & Andersen, 2006 ;

Moore et al. 2005) and is seen in several meta-analyses
Escitalopram has been found to have superior efficacy
(Auquier et al. 2003 ; Lam & Andersen, 2006 ; Kennedy
to other antidepressants in the treatment of major de-
et al. 2006, 2009 ; Lepola et al. 2004 ; Llorca et al. 2005).
pressive disorder (MDD) in a number of individual
The study of Moore et al. (2005) was designed to test
studies and in a series of meta-analyses. Escitalopram
prospectively for a possible difference between 20 mg
has been shown to be superior to paroxetine
escitalopram and 40 mg citalopram in patients with
(Boulenger et al. 2006 ; Kasper et al. 2009) in a pro-
severe depression [Montgomery–Asberg Depression
spective double-blind study and in meta-analyses of
Rating Scale (MADRS ; Montgomery & Åsberg, 1979)
randomized controlled trials (RCTs) comparing escit-
score of o30], and the observed difference of 2.1 was
alopram to commonly prescribed antidepressants
significant and clinically relevant. In patients with
(Kennedy et al. 2006, 2009). The superiority of escit-
very severe depression there is an even more striking
alopram is even more clear-cut in severe depression, as
treatment effect. In the study of Yevtushenko et al.
shown in individual studies in which it was compared
(2007) in severe depression (MADRS o30, except for
to citalopram (the racemic mixture of escitalopram
12 patients) a difference of 3.51 for 10 mg escitalopram
and the R-enantiomer), paroxetine, and venlafaxine
compared to 20 mg citalopram increases to 3.99 in
very severe depression (MADRS o35). The superior-
Address for correspondence : Professor S. Montgomery, PO Box 8751,
London W13 8WH, UK.
ity of escitalopram to citalopram needs to be viewed
Tel. : +44 20 8997 2689 Fax : +44 20 8566 7986 in the context of the paucity of studies showing a sig-
Email : Stuart@samontgomery.co.uk nificant difference between antidepressants. Only
262 S. A. Montgomery et al.
three antidepressants (clomipramine, venlafaxine,
escitalopram) were found to have two or more studies
showing superiority against a standard antidepressant
under conditions of fair comparison (Montgomery
et al. 2007).
This meta-analysis examined published data and
data available on publicly accessible websites from all
studies with both escitalopram and citalopram treat-
ment arms in patients with MDD in order to evaluate
the potential superiority of escitalopram and place it in
the perspective of a clinically relevant difference.
Methods
Study identification
Multiple computer searches using Medline (1966–June
2009), EMBASE (1998–2009), and the Cochrane
Collaboration (1980–June 2009) were conducted.
Key words, including escitalopram AND citalopram
AND (depression or major depressive disorder or
major depressive episode), were specified. Additional
studies in any language were sought in reference lists
of retrieved articles. Unpublished trials were identi-
fied through the Controlled Trials database, www.
lundbecktrials.com, www.forestclinicaltrials.com, and
the National Institute of Health’s Computer Retrieval
of Information on Scientific Projects (CRISP) service
(1972–2005). In addition, the following clinical trial
registration sites were searched : www.lundbecktrials.
com, www.forestclinicaltrials.com, www.clinicaltrials.
gov, www.clinicaltrialresults.org, www.ifpma.org/
clinicaltrials, and www.controlled-trials.com.
Data extraction
For most studies, continuous variables were based
on the full analysis set (FAS), comprising patients
that had been randomized to treatment, had received
at least one dose of study medication, and had at least
one valid post-baseline efficacy assessment. An
analysis was also carried out on the intent-to-treat
(ITT) set, comprising all patients who received at least
one dose of study medication.
Continuous rating scale outcomes [MADRS and 17-item
Hamilton Depression Rating Scale (HAMD17)]
All studies provided information on appropriate
means [most often adjusted, typically from ANCOVAs
using the last-observation-carried-forward (LOCF)
approach], sometimes only in graphs, either at specific
time-points, or change from baseline at specific time-
points/endpoint, and per group sample size. Most
studies provided information on standard deviations
(S.D.) or standard errors of the mean (S.E.M), either di-
rectly or of differences between mean changes from
baseline (S.E.D.) in different treatment arms. In the lat-
ter case, assuming equal variances between compared
groups, the S.D. was estimated from the equation :
2
S:D: =S:E:D:2
(n1
n2 =(n1 +n2 )),
where n1 and n2 are sample sizes for groups 1 and 2,
respectively.
If such measures of variability were not present, they
were imputed from either exact p values or p values
with sufficiently narrow limits (e.g. 0.01<p<0.05 or
p<0.001) from tests comparing two groups. In the
latter case, the upper limit of the p-value interval
was used to impute variability estimates. Because the
difference between the two compared groups was
supplied along with groupwise sample sizes, and as-
Downloaded from http://ijnp.oxfordjournals.org/ by guest on January 17, 2017
suming equal variances between the compared groups,
the S.D. could be calculated by inverting the p value
with the appropriate inverse t distribution.
One study indicated the length of a one-sided
95 % confidence interval (CI) and equating this to
z0.95 . S.E.D.=1.645 . S.E.D., the S.E.M of each change from
baseline was estimated (assuming equal variances be-
tween groups). The use of normal fractile leads to a
higher estimated S.D. than an appropriate t value with
degrees of freedom equalling the sum of the two
groups ’ sample sizes minus 2.
For some studies, the appropriate variability esti-
mates or reasonably exact p values were not provided,
because these studies simply reported p>0.05 or that
the comparison was ‘not significant ’ (which is the
same). For these studies, S.D.s were imputed assuming
p=0.25 and the above inversion method was applied
to this p value. This method produced S.D.s that were
larger than those that could be estimated from graphs
(which are remarkably constant) and will, therefore,
tend to produce estimates attenuated towards the null
hypothesis.
Count or relative frequency outcomes (response
and remission)
In all studies reporting this outcome, a responder was
defined as a binary outcome indicating o50 % change
from baseline. Remission based on the MADRS was
defined as a MADRS score f12, while for HAMD17 it
was a score f7. Of the articles reporting either of these
outcomes, some reported estimates (predictions from
logistic regression analyses) by treatment group, while
others reported crude (i.e. unadjusted) estimates such
as ‘107/157 ’. In the first case, adjusted estimates were
converted to number of patients that, for example,

were in remission by multiplying the estimated pre-
dictions by the by-treatment FAS sample size. In the
latter case the ‘raw ’ numbers were used. Hence,
the response and remission meta-data consist of both
adjusted and unadjusted figures. The numbers of
patients responding or in remission, together with
by-treatment sample sizes, were extracted from the
studies. The by-treatment ITT sample sizes were used
as denominator in the analyses of responders and re-
mitters, so that patients excluded from the FAS were
treated as non-responders and/or non-remitters.
Preliminary data handling
Continuous rating scale outcomes (MADRS and HAMD17)
For each study, the difference between mean change
from baseline to week 8 (or last assessment for
studies <8 wk) between escitalopram and citalopram
was estimated together with their corresponding
variances/S.E.M. For studies with more than one dose
of escitalopram or citalopram (Burke et al. 2002,
Yevtushenko et al. 2007), only corresponding doses
were used (i.e. 10 mg/d escitalopram vs. 20 mg/d
citalopram, or 20 mg/d escitalopram vs. 40 mg/d
citalopram). The difference in mean change from
baseline and corresponding S.E.M were the by-study
input for the meta-analysis.
Responders and remitters
The responder and remitter rates for each study were
converted into log odds ratios and associated standard
errors. The log odds ratio and associated S.E. was the
by-study input for the meta-analysis.
Number needed to treat (NNT)
The NNT is the inverse of the difference between the
responder (or remitter) probabilities of two treatments.
It was assumed that the responder (or remitter) prob-
ability of each treatment was normally distributed,
with mean as the estimated remission (or response)
probability and standard error as the square root of
[p(1xp)/ITT sample size], where p is the responder (or
remitter) probability of each treatment. With these as-
sumptions, all of the statistical machinery described
below applies to the difference of the responder (or
remitter) probability between treatments. After per-
forming the meta-analysis on the difference between
responder (or remitter) probabilities between treat-
ments and calculating an appropriate confidence in-
terval thereof, the point estimate and the limits of the
confidence intervals were back-transformed to an es-
timated NNT (and associated confidence intervals).
Efficacy of escitalopram compared to citalopram 263
Statistical methods used in the meta-analysis
All meta-analyses based on the above outlined input,
regardless of outcome type, were performed using the
fixed-effects general linear model described in Hedges
& Olkin (1985). All p values and confidence intervals
are based on two-sided considerations with type I
error set to 0.05.
Results
Nine trials were identified that included escitalopram
(n=995) and citalopram (n=1014) treatment arms
(FAS, Table 1). Some treatment arms were omitted
in order to make a fair dose comparison [the 10-mg
escitalopram group (n=119) in study 5 and the 10-mg
citalopram group (n=111) in study 8], since there was
no 20-mg citalopram group in study 5, and no 5-mg
Downloaded from http://ijnp.oxfordjournals.org/ by guest on January 17, 2017
escitalopram group in study 8. The corresponding
numbers in the FAS were escitalopram (n=977) and
citalopram (n=989). All studies were RCTs except
study 9, which was naturalistic. Two studies only used
the HAMD17 (Hamilton, 1960) (studies 4 and 7), and
two studies used both the HAMD17 and the MADRS
(studies 1 and 5). Two studies did not report response
and remission rates (studies 1 and 5). Only one study
was not published as a peer-reviewed article (study 1),
although the data are reported on the Forest website
and these data have been included in previous meta-
analyses.
Treatment difference
The overall mean treatment difference between esci-
talopram (n=813) and citalopram (n=827) was 1.7
(95 % CI 0.8–2.6, p=0.0002) MADRS points for the six
RCTs that used this scale (studies 1, 2, 3, 5, 6, 8) (see
Fig. 1). When the non-RCT (study 9) was also in-
cluded, the overall mean treatment difference between
escitalopram (n=880) and citalopram (n=887) was 1.9
(95 % CI 1.0–2.8, p<0.0001) MADRS points. The two
trials that only used the HAMD17 scale (studies 4 and 7)
had an overall mean treatment difference between
escitalopram (n=97) and citalopram (n=102) of 2.4
(95 % CI x0.7 to 5.4, p=0.1262) points.
Responder
The overall odds ratio for responders was 1.44 (95 %
CI 1.18–1.75, p=0.0003), in favour of escitalopram,
based on the eight RCTs [studies 1–8 ; escitalopram
(Esc)=928, citalopram (Cit)=954, all-patients-treated
set (APTS)]. The responder rates were 72.3 % for
escitalopram and 63.9 % for citalopram, a difference
Table 1. Overview of studies included in the meta-analysis

264
Baseline Withdrawals

S. A. Montgomery et al.
Duration Treatment MADRS due to Completers
Study no. (reference) inclusion RCT Scale Setting Response Remission (dosage) score APTS (FAS) AEs ( %)a ( %)a

1 (Forest Labs, 2005 ; 8 wk Yes MADRS Spec Yes No Esc (10–20)1 28.7¡4.3 125 (124) 8.8 76.8
Anon, 2002) MADRS o22 HAMD17 Cit (20–40)2 28.3¡5.0 123 (119) 4.1 80.5
Pbo 28.8¡5.0
2 (Colonna et al. 2005) 24 wk Yes MADRS GP Yes Yes Esc (10) 29.5¡4.3 175 (165) 5.7 86.3
MADRS o22 Cit (20) 30.2¡4.7 182 (174) 9.9 84.1
3 (Lepola et al. 2003 ; 8 wk Yes MADRS GP Yes Yes Esc (10–20)3 29.0¡4.3 155 (155) 2.6 94.2
Anon, 2002) MADRS o22 Cit (20–40)4 29.2¡4.2 160 (159) 3.8 95.0
Pbo 28.7¡4.0
4 (Lalit et al. 2004) 4 wk Yes HAMD17 Spec Yes Yes Esc (10) 26¡6 69 (69) 1.4 94.2
HAMD17 o18 Cit (20) 25¡5 74 (74) 2.7 95.9
Ser (100)
5 (Burke et al. 2002 ; 8 wk Yes MADRS Spec Yes No Esc (10) 28.0¡4.9 119 (118) 4.2 79.8
Anon, 2002) MADRS o22 Esc (20) 28.9¡4.6 125 (123) 10.4 75.2
HAMD17 Cit (40) 29.2¡4.5 125 (125) 8.8 74.4
Pbo 29.5¡5.0
6 (Moore et al. 2005) 8 wk Yes MADRS GP & Yes Yes Esc (20) 36.3¡4.8 142 (138) 2.9 93.0
MADRS o30 spec Cit (40) 35.7¡5.8 152 (142) 6.3 83.6
7 (Li et al. 2006) 6 wk Yes HAMD17 Spec Yes No Esc (10–20)5 23.7¡4.3 28 (28) – 92.9
HAMD17 o18 Cit (20–40)5 23.8¡3.6 28 (28) – 96.4
8 (Yevtushenko 6 wk Yes MADRS Spec Yes Yes Esc (10) 34.8¡3.5 109 (108) – 99.1
et al. 2007) MADRS o25 Cit (10) 35.4¡3.3 111 (106) – 95.5
Cit (20) 35.7¡3.8 110 (108) 98.2
9 (Lançon et al. 2006)b 8 wk No MADRS Spec Yes Yes Esc (5–20)6 39.6¡6.2 67 (67) 1.5 95.5
MADRS o30 Cit (10–40)7 37.0¡5.8 60 (60) 8.3 86.7

Esc, Escitalopram ; Cit, citalopram ; Pbo, placebo-controlled study ; Ser, sertraline ; RCT, randomized controlled trial, Spec, specialist settings ; n.a., not available ; APTS, all-patients-
treated set ; FAS, full-analysis set ; AEs, adverse events.
a
Values from last visit (LOCF), based on the APTS.
b
Non-randomized design.
1
Mean dose 17.6 mg ; 2 mean dose 35.3 mg ; 3 mean dose 14.0 mg ; 4 mean dose 24.4 mg ; 5 mean dose not available ; 6 mean dose 12.7 mg ; 7 mean dose 28.8 mg.

Downloaded from http://ijnp.oxfordjournals.org/ by guest on January 17, 2017

 Efficacy of escitalopram compared to citalopram 265

Favours Cit Favours Esc Favours Cit Favours Esc

MD-02 MD-02
Colonna et al. Colonna et al.
Lepola et al. Lepola et al.
Lalit et al.* Lalit et al.
Burke et al. Burke et al.
Moore et al. Moore et al.
Li et al.* Li et al.
Yevtushenko et al. Yevtushenko et al.
Lancon et al.* Lancon et al.*

Overall mean Overall mean

MADRS
–4 –2 0 2 4 6 8 10 –20 –10 0 10 20 30 40 50 60
Estimated mean difference from baseline (MADRS or HAMD17): Esc vs. Cit Response (%): Esc minus Cit

Fig. 1. Estimated mean treatment difference in MADRS or
HAMD17 total score at week 8 (or last assessment if <8 wk)
shown with 95 % confidence intervals. Esc, Escitalopram ;
Cit, citalopram. * The studies of Lalit et al. (2004) (4 wk) and
Li et al. (2006) (6 wk) used only the HAMD17, and that of
Fig. 2. Estimated difference in response rates at week 8
(or last assessment if <8 wk) (EscxCit) with 95 % confidence
intervals. Esc, Escitalopram ; Cit, citalopram. Response
is defined as the percentage of patients with o50 %
improvement from baseline (MADRS or HAMD17). * The

Downloaded from http://ijnp.oxfordjournals.org/ by guest on January 17, 2017

Lançon et al. (2006) was a naturalistic study, therefore results naturalistic study of Lançon et al. (2006) was not included
from these studies are not included in the calculation of the in the calculation of the overall mean.
overall mean of 1.7 (95 % CI 0.8–2.6, p=0.0002) MADRS
points.
Favours Cit Favours Esc
Colonna et al.
(EscxCit) of 8.3 percentage points (95 % CI 4.4–12.3).
The corresponding NNT for responder was 11.9 (95 % Lepola et al.
CI 8.1–22.8, p<0.0001). The differences in responder Moore et al.
rates (in percentage points) for each study are shown
Yevtushenko et al.
in Fig. 2.
Lancon et al.*
Remission

Remission was defined as either MADRS f12, or Overall mean

HAMD17 f7 scores. Based on the four RCTs using the
MADRS that reported remission (studies 2, 3, 6, 8 ;
Esc=581, Cit=604, APTS), the overall odds ratio for
remission was 1.86 (95 % CI 1.46–2.36, p<0.0001). The
remission rates were 61.6 % for escitalopram and
44.0 % for citalopram, a difference (Esc – Cit) of 17.6
percentage points (95 % CI 12.1–23.1). The corre-
sponding NNT for remission was 5.7 (95 % CI 4.3–8.3,
p<0.0001). The differences in remission rates (in per-
centage points) for each study are shown in Fig. 3.
–1 0 1 2 3
Remission (log odds ratio) MADRS: Esc vs. Cit
Fig. 3. Estimated difference in remission rates at week 8
(or last assessment if <8 wk) (Esc – Cit) with 95 % confidence
intervals. Esc, Escitalopram ; Cit, citalopram. Remission is
defined as the percentage of patients with MADRS score of
f12. * The naturalistic study of Lançon et al. (2006) and
those of Lalit et al. (2004) and Li et al. (2006) that used only
the HAMD17 are not included in the calculation of the
overall mean.

Discussion
This meta-analysis of all the publicly available
data indicates that escitalopram has superior efficacy
compared to citalopram in the treatment of MDD in
a patient population including both moderate and se-
vere disorder. It includes all available studies in MDD
that included escitalopram and citalopram treatment
arms (head-to-head, with an active control, or placebo-
controlled) and provides further evidence for a
statistically significant efficacy advantage of escitalo-
pram over citalopram.
However, a statistically significant difference be-
tween two treatments may not reflect clinical benefits
that are evident or relevant when treating individual
patients. Separate tests are usually applied to the
data to test if statistical differences are likely to have
clinical relevance. There are several approaches to
266 S. A. Montgomery et al.
determining clinical relevance. In regulatory terms the
most important are the responder analysis, the remit-
ter analysis and the treatment effect (the difference
between two treatments in the improvement from
baseline to endpoint on a standard assessment scale)
(EMEA, 2002).
The criteria used to establish a clinically relevant
difference have almost all been focused on a compari-
son of drug and placebo. Comparing differences
between two active treatments applying the same cri-
terion used to define a clinically relevant difference on
the pivotal scale between active drug and placebo is
very stringent, since this means that the difference to
placebo of the superior treatment must be at least
twice that of the comparator antidepressant. However,
50 % of the difference between active drug and placebo
has also been used as a criterion to indicate probable
clinical relevance when comparing two established
treatments (Montgomery & Möller, 2009).
The studies included in the present meta-analysis
included large regulatory studies sponsored by
H. Lundbeck A/S or Forest Inc., conducted mainly
in Caucasian patients, together with smaller, shorter
studies in different populations, for example in India
(Lalit et al. 2004) and China (Li et al. 2006). Even if no
robust conclusions can be drawn from these smaller
studies individually, the results are consistent with
those from the larger trials.
The naturalistic study of Lançon et al. (2006), which
was also small, has the methodological problems
inherent with this sort of study design that com-
plicate the conclusions that can be drawn from the
study. This is an effectiveness study with an open non-
randomized design and the bias of the investigator in
the choice of treatment and assessment might have
influenced the result. There is some evidence of poss-
ible investigator bias in the allocation of patients to
different treatments, since the escitalopram group
were significantly more severely depressed at baseline
than those on citalopram. This might have been a
chance finding but it might also reflect the clinical
view, which was already current, that escitalopram
was a more effective treatment in severe depression.
All effectiveness studies are open to these sorts of
biases. Nevertheless, for the sake of completeness, the
study was included (Anderson, 2008), but was ex-
cluded from the meta-analysis of the randomised
controlled studies.
The overall treatment difference between escitalo-
pram and citalopram was 1.7 points in favour of esci-
talopram on the MADRS scale, which was statistically
significant (p<0.0001). A direct measure of the clini-
cally relevant difference may be taken from the
difference observed between placebo and citalopram.
On the basis of an analysis of the positive anti-
depressant studies submitted to the FDA, Kirsch et al.
(2002) reported that a difference of around 2 points on
HAMD17 between drug and placebo had been suf-
ficient for regulatory approval. In the case of citalo-
pram this difference was 1.9 points on HAMD17. The
difference between escitalopram and placebo on the
MADRS was found to be 1.9 overall and 2.1 points in
severe depression in the regulatory studies of escita-
lopram (Gorman et al. 2002) which indicates that the
difference of around 2 points would be considered
clinically relevant for both the MADRS and HAMD17.
In a more recent meta-analysis of placebo controlled
studies for all antidepressants, including a range
of other non-regulatory studies, the reported differ-
ence on HAMD17 was 1.8 (Kirsch et al. 2008). Since
the present analysis included a wide range of non-
Downloaded from http://ijnp.oxfordjournals.org/ by guest on January 17, 2017
regulatory studies, a 1.8-point difference between
antidepressant and placebo should be regarded as
clinically relevant for this wider population. It could
therefore be concluded that this difference is clinically
relevant. Since the treatment difference between escit-
alopram and citalopram was 1.7 points, this is also
clinically relevant.
The European Medicines Agency (EMEA) normally
uses a responder analysis to determine whether a
statistically significant difference is also clinically
relevant. A statistically significant advantage on the
responder analysis in favour of an antidepressant
compared with placebo is normally considered by the
Committee for Medicinal Products for Human Use
(CMPH) to be clinically relevant (EMEA, 2002). In a
review of placebo-controlled antidepressant studies
submitted for regulatory approval over a 25-year per-
iod, Melander et al. (2008) report a mean difference
in the percentage of HAMD17 responders to placebo of
16 percentage points (95 % CI 12.0–20.0). The present
analysis of two antidepressants shows a response
advantage of 8.3 percentage points, and a remission
rate advantage of 17.6 percentage points in favour of
escitalopram vs. citalopram. These differences trans-
late into NNTs of approximately 12 for response and
six for remission.
An elegant analysis was carried out by Lam &
Andersen (2006), in which the placebo-controlled data
on citalopram and escitalopram from the same studies
were compared. The treatment effect for citalopram
compared to placebo was constant with increasing
baseline severity measured on the MADRS, but the
effect increased for escitalopram. The advantage for
escitalopram over citalopram became greater and was
more strikingly clinically relevant in treating patients

with increasing severity of depression. These data
confirm that escitalopram has different properties as
an antidepressant compared to citalopram.
Based on treatment difference, and NNTs derived
from response and remission rates, the statistically
significant superiority of escitalopram vs. citalopram
can be considered as clinically relevant.
Acknowledgements
The authors thank D. J. Simpson (H. Lundbeck A/S)
for technical assistance in the preparation of the
manuscript. The authors are entirely responsible for
the scientific content of this article.
Statement of Interest
Thomas Hansen is an employee of H. Lundbeck A/S.
Stuart Montgomery has received consulting fees or
honoraria from AstraZeneca, Bionevia, Bristol–Myers
Squibb, GlaxoSmithKline, Johnson & Johnson, Lilly,
H. Lundbeck A/S, Merck & Co. Inc., M’s Science,
Merz Pharmaceuticals, Neurim Pharmaceuticals,
Otsuka, Pfizer Inc., Pierre Fabre, Roche Pharmaceu-
ticals, Sanofi-Aventis, Sepracor Inc., Servier Labora-
tories, Synosis, Takeda, Theracos, Transcept, UBC,
Xytis, Wyeth. Siegfried Kasper received grants/
research support, consulting fees and honoraria with-
in the last 3 years from AstraZeneca, Bristol–Myers
Squibb, CSC, Eli Lilly, GlaxoSmithKline, Janssen
Pharmaceutica, Lundbeck, MSD, Novartis, Organon,
Pierre Fabre, Pfizer, Schwabe, Sepracor, Servier and
Wyeth.
References
Anderson IM (2008). Commentary on ‘ Re-evaluation of
the efficacy and tolerability of venlafaxine versus SSRI :
meta-analysis’ by Weinmann et al. Psychopharmacology 196,
521–522.
Anon (2002). Läkemedelsmonografier. Cipralex
(escitalopram) [in Swedish]. Info från Läkemedelsverket 13,
71–72.
Auquier P, Robitail S, Llorca P-M, Rive B (2003).
Comparison of escitalopram and citalopram efficacy : a
meta-analysis. International Journal of Psychiatry in Clinical
Practice 7, 259–268.
Boulenger JP, Huusom AKT, Florea I, Baekdal T,
Sarchiapone M (2006). A comparative study of the efficacy
of long-term treatment with escitalopram and paroxetine
in severely depressed patients. Current Medical Research
and Opinion 22, 1331–1341.
Burke WJ, Gergel I, Bose A (2002). Fixed-dose trial of the
single isomer SSRI escitalopram in depressed outpatients.
Journal of Clinical Psychiatry 63, 331–336.
Efficacy of escitalopram compared to citalopram 267
Colonna L, Andersen HF, Reines EH (2005). A randomised,
double-blind, 24-week study of escitalopram (10 mg/day)
versus citalopram (20 mg/day) in primary care patients
with major depressive disorder. Current Medical Research
and Opinion 21, 1659–1668.
EMEA (2002). Committee for Proprietary Medicinal Products
(CPMP). Note for guidance on clinical investigation of
medicinal products in the treatment of depression. CPMP/
EWP/518/97, Rev. 1. European Medicines Agency.
Forest Laboratories (2005). http://www.forestclinicaltrials.
com/CTR/CTRController/CTRViewPdf?_file_id=scsr/
SCSR_SCT-MD-02_final.pdf (not peer-reviewed).
Accessed 4 January 2010.
Gorman J, Korotzer A, Su G (2002). Efficacy comparison of
escitalopram and citalopram in the treatment of major
depressive disorder : pooled analysis of placebo-controlled
trials. CNS Spectrums 7 (Suppl. 1), 40–44.
Hamilton MA (1960). A rating scale for depression. Journal
of Neurological and Neurosurgical Psychiatry 23, 56–62.
Hedges LV, Olkin I (1985). Statistical Methods for
Downloaded from http://ijnp.oxfordjournals.org/ by guest on January 17, 2017
Meta-analysis. New York : Academic Press, p. 369.
Kasper S, Baldwin D, Larsson Lönn S, Boulenger JP (2009).
Superiority of escitalopram to paroxetine in the treatment
of depression. European Neuropsychopharmacology 19, 229–237.
Kennedy SH, Andersen HF, Lam R (2006). Efficacy of
escitalopram in the treatment of major depressive disorder
compared with conventional selective serotonin reuptake
inhibitors and venlafaxine XR : a meta-analysis. Journal of
Psychiatry and Neuroscience 31, 122–131.
Kennedy SH, Andersen HF, Thase ME (2009). Escitalopram
in the treatment of major depressive disorder : a meta-
analysis. Current Medical Research and Opinion 25, 161–175.
Kirsch I, Deacon BJ, Huedo-Medina TB, Scoboria A, et al.
(2008). Initial severity and antidepressant benefits : a
meta-analysis of data submitted to the Food and Drug
Administration. PLoS Medicine. Published online :
26 February 2008. doi :10.1371/journal.pubmed.0050045.
Kirsch I, Moore TJ, Scoboria A, Nicholls SS (2002). The
emperor’s new drugs : an analysis of antidepressant
medication data submitted to the US Food and Drug
Administration. Prevention & Treatment 5, Article 23.
Lalit V, Prakash A, Rajendra H, Anukant M, et al. (2004).
Escitalopram versus citalopram and sertraline : a
double-blind controlled, multi-centric trial in Indian
patients with unipolar depression. Indian Journal of
Psychiatry 46, 333–341.
Lam RW, Andersen HF (2006). The influence of baseline
severity on efficacy of escitalopram and citalopram in the
treatment of major depressive disorder : an extended
analysis. Pharmacopsychiatry 39, 180–184.
Lançon C, Sapin C, Note I, Farisse J (2006). Comparison of
escitalopram and citalopram in outpatients with severe
major depressive disorder : a prospective, naturalistic,
8-week study. International Journal of Psychiatry in Clinical
Practice 10, 131–137.
Lepola U, Wade A, Andersen HF (2004). Do equivalent doses
of escitalopram and citalopram have similar efficacy? A
pooled analysis of two positive placebo-controlled studies
268 S. A. Montgomery et al.
in major depressive disorder. International Clinical
Psychopharmacology 19, 149–155.
Lepola UM, Loft H, Reines EH (2003). Escitalopram
(10–20 mg/day) is effective and well tolerated in a
placebo-controlled study in depression in primary care.
International Clinical Psychopharmacology 18, 211–217.
Li J, Shen WW, Liu Y, Xu L, et al. (2006). The effectiveness
and safety of escitalopram in the treatment of major
depression : a randomized double-blind active-drug
controlled trial [in Chinese]. Chinese Journal of Evidence-
based Medicine [Zhongguo xunzheng yixue zazhi] 6, 552–556.
Llorca P-M, Azorin J-M, Despiegel N, Verpillat P (2005).
Efficacy of escitalopram in patients with severe
depression : a pooled analysis. International Journal of
Clinical Practice 59, 268–275.
Melander H, Salmonson T, Abadie E, van Zwieten-Boot B
(2008). A regulatory apologia – a review of
placebo-controlled studies in regulatory submissions
of new-generation antidepressants. European
Neuropsychopharmacology 18, 623–627.
Montgomery SA, Andersen HF (2006). Escitalopram versus
venlafaxine XR in the treatment of depression. International
Clinical Psychopharmacology 21, 297–309.
Montgomery SA, Åsberg M (1979). A new depression scale
designed to be sensitive to change. British Journal of
Psychiatry 134, 382–389.
Montgomery SA, Baldwin DS, Blier P, Fineberg NA, et al.
(2007). Which antidepressants have demonstrated superior
efficacy? A review of the evidence. International Clinical
Psychopharmacology 22, 323–329.
Montgomery SA, Möller HJ (2009). Is the significant
superiority of escitalopram compared with other
antidepressants clinically relevant? International Clinical
Psychopharmacology 24, 111–118.
Moore N, Verdoux H, Fantino B (2005).
Prospective, multicentre, randomized, double-blind
study of the efficacy of escitalopram versus citalopram
in outpatient treatment of major depressive
disorder. International Clinical Psychopharmacology 20,
131–137.
Yevtushenko VY, Belous AI, Yevtushenko YG,
Gusinin SE, et al. (2007). Efficacy and tolerability of
escitalopram versus citalopram in major depressive
Downloaded from http://ijnp.oxfordjournals.org/ by guest on January 17, 2017
disorder : a 6-week, multicenter, prospective, randomized,
double-blind, active-controlled study in adult outpatients.
Clinical Therapeutics 29, 2319–2332.