Do equivalent doses of escitalopram and citalopram have similar efficacy? A pooled analysis of two positive placebo-controlled studies in major depressive disorder.

Abstract

Escitalopram is the S-enantiomer of citalopram. In this study, we compared the efficacy of equivalent dosages of escitalopram and citalopram in the treatment of moderate to severe major depressive disorder (MDD), based on data from two, pooled, randomized, double-blind, placebo-controlled studies of escitalopram in which citalopram was the active reference. The primary efficacy parameter was the mean change from baseline in the Montgomery Asberg Depression Rating Scale (MADRS) total score. Significant differences in favour of escitalopram were observed for the MADRS [P<0.05, observed cases (OC)/last observation carried forward (LOCF)] and Clinical Global Improvement-Severity of Illness scores (CGI-S; P<0.05, OC/LOCF). Escitalopram separated from placebo at week 1 on the primary efficacy parameter, whereas citalopram first separated from placebo at week 6. An analysis of time to response showed that escitalopram-treated patients responded significantly faster to treatment than citalopram-treated patients (P<0.01). More patients responded to and achieved remission with escitalopram than to citalopram (P<0.05, OC). The HAMD scale was only used in the fixed-dose study, where escitalopram-treated patients had a significant reduction in HAMD-17 total score at week 8 compared to citalopram-treated patients (P<0.05, OC/LOCF). In the pooled subpopulation of severely ill patients (MADRS> or = 30), escitalopram-treated patients showed greater improvement than citalopram-treated patients (P<0.05, LOCF/OC). Escitalopram showed consistently superior efficacy compared to citalopram in the treatment of moderate to severe MDD on all efficacy parameters, and was similarly well tolerated.

Full text

Page 1

Original article149
Do equivalent doses of escitalopram and citalopram
have similar efficacy? A pooled analysis of two positive
placebo-controlled studies in major depressive disorder
Ulla Lepolaa,b, Alan Wadecand Henning Friis Andersend
Escitalopram is the S-enantiomer of citalopram. In this
study, we compared the efficacy of equivalent dosages
of escitalopram and citalopram in the treatment of
moderate to severe major depressive disorder (MDD),
based on data from two, pooled, randomized,
double-blind, placebo-controlled studies of escitalopram
in which citalopram was the active reference. The
primary efficacy parameter was the mean change from
baseline in the Montgomery A˚sberg Depression Rating
Scale (MADRS) total score. Significant differences in
favour of escitalopram were observed for the MADRS
[P<0.05, observed cases (OC)/last observation carried
forward (LOCF)] and Clinical Global Improvement–Severity
of Illness scores (CGI-S; P<0.05, OC/LOCF).
Escitalopram separated from placebo at week 1 on
the primary efficacy parameter, whereas citalopram first
separated from placebo at week 6. An analysis of time
to response showed that escitalopram-treated patients
responded significantly faster to treatment than
citalopram-treated patients (P<0.01). More patients
responded to and achieved remission with escitalopram
than to citalopram (P<0.05, OC). The HAMD scale
was only used in the fixed-dose study, where
escitalopram-treated patients had a significant reduction
in HAMD-17 total score at week 8 compared to
citalopram-treated patients (P<0.05, OC/LOCF). In
the pooled subpopulation of severely ill patients
(MADRSZZ30), escitalopram-treated patients showed
greater improvement than citalopram-treated patients
(P<0.05, LOCF/OC). Escitalopram showed consistently
superior efficacy compared to citalopram in the treatment
of moderate to severe MDD on all efficacy parameters, and
was similarly well tolerated. Int Clin Psychopharmacol
19:149–155? c 2004 Lippincott Williams & Wilkins.
International Clinical Psychopharmacology 2004, 19:149–155
Keywords: Antidepressant, CGI-S, citalopram, escitalopram, HAMD-17,
MADRS, major depressive disorder, SSRI
aKuopion Psykiatripalvelu OY Psychiatric Research, Clinic of Kuopio, Kuopio,
Finland,bDepartment of Psychiatry, Oulu and Helsinki University, Oulu, Helsinki,
Finland,cCPS Ltd, Clinical Research Centre, Glasgow, UK anddH. Lundbeck
A/S, International Clinical Research, Copenhagen, Denmark.
Correspondence and requests for reprints to Henning Friis Andersen,
H. Lundbeck A/S, International Clinical Research, Ottiliavej 9, DK-2500
Copenhagen, Denmark.
Fax: +45 36306771; e-mail: hfa@lundbeck.com
Received 9 October 2003 Accepted 8 January 2004
Introduction
Escitalopram and citalopram are effective antidepres-
sants, with favourable tolerability profiles (Muldoon,
1996; Burke et al., 2002; Wade et al., 2002; Lepola et al.,
2003). Citalopram is a racemate, which consists of equal
amounts of an S- and an R-enantiomer. Pharmacological
studies, performed both in vitro and in vivo, have shown
that the selective serotonin reuptake inhibitor (SSRI)
activity of citalopram resides almost exclusively in the S-
enantiomer, escitalopram (Hyttel et al., 1992), as does the
antidepressant activity (Hogg and Sa ´nchez, 1999). Of the
currently available SSRIs, escitalopram has the highest
selectivity for the human serotonin transporter relative to
the noradrenaline or dopamine transporters (Owens et al.,
2001).
In the clinical development of escitalopram, it was
assumed that the therapeutically active enantiomer
would have the same efficacy as the racemate (citalo-
pram), but at half the dose. However, recent biochemical,
functional and behavioural experiments indicate that
escitalopram has greater efficacy and a faster onset of
action than equivalent doses of citalopram in nonclinical
studies (Mørk et al., 2003; Sa ´nchez et al., 2003a, b). The
lower serotonin reuptake inhibition by citalopram in
these nonclinical studies is assumed to be due to the
inhibition of the effect of escitalopram by R-citalopram,
possibly via an allosteric effect in its interaction with the
serotonin transporter.
The tolerability profile of escitalopram is similar to that of
citalopram, with placebo levels of patients withdrawing
due to adverse events. Indeed, most adverse events
reported by escitalopram-treated patients are mild and
transient (Burke et al., 2002; Lepola et al., 2003), as is the
case for citalopram.
The results from three placebo-controlled clinical trials of
escitalopram in the treatment of major depressive
disorder (MDD) consistently favour escitalopram versus
placebo both in primary care (Wade et al., 2002; Lepola
et al., 2003) and in specialist settings (Burke et al., 2002).
0268-1315 ? c 2004 Lippincott Williams & Wilkins
DOI: 10.1097/01.yic.0000122862.35081.cd
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Page 2

In the two studies in which citalopram was the active
reference, escitalopram separated significantly from
placebo as early as week 1 or 2 on both the primary and
secondary efficacy variables (Burke et al., 2002; Lepola
et al., 2003). Statistically significant separation of citalo-
pram from placebo did not occur until week 6, or later.
Although these studies were not powered to show
differences between two active treatments, there were
indications of advantages for escitalopram versus citalo-
pram, apart from earlier separation from placebo (i.e.
significantly more responders and remitters at endpoint
and a faster time to response) (Lepola et al., 2003).
A previous pooled analysis (Gorman et al., 2002) demon-
strated a statistical superiority of escitalopram over
citalopram. However, that analysis included multiple
doses of both drugs without discriminating between
doses. The present analysis was undertaken to compare
equivalent doses of the two drugs. Thus, 20mg/day
escitalopram was compared with 40mg/day of citalopram
and 10–20mg/day escitalopram was compared with 20–
40mg/day citalopram. For this reason, the data from the
10mg/day escitalopram fixed-dose arm were not in-
cluded. Furthermore, data from a third trial were not
included because neither active drug was statistically
superior to placebo, based on the primary efficacy
endpoint. Inclusion of such a trial is a potential
source of bias because it lacked assay sensitivity (Klein,
2000).
In addition, this study explores the efficacy of escitalo-
pram versus that of citalopram in the larger cohort by
pooling the data from two published clinical studies.
Methods
Two randomized, multi-centre, placebo-controlled trials
were conducted in outpatients with moderate to severe
MDD in specialist settings in the USA (Burke et al., 2002)
and in primary care in Europe and Canada (Lepola et al.,
2003), and compared the effect of 8 weeks of double-
blind treatment with escitalopram with that of placebo,
with citalopram as the active reference.
Eligible patients were male or female outpatients aged
between 18 and 65 years, who met DSM-IV criteria for a
major depressive episode (American Psychiatric Associa-
tion, 1994), with a minimum score of 22 on the
Montgomery A˚sberg Depression Rating Scale (MADRS)
(Montgomery and A˚sberg, 1979). The Hamilton Depression
scale (HAMD-17) (Hamilton, 1960) was not used in the
flexible-dose study (Lepola et al., 2003), whereas a
minimum HAMD-17 score of 2 on item 1 (depressed
mood) was an additional entry requirement in the fixed-
dose study (Burke et al., 2002).
Patients were excluded if there was evidence of active
suicidal ideation (a score Z5 on item 10 of the MADRS)
or a recent suicide attempt, or if the patient had any
DSM-IV Axis I disorder other than major depression.
Both trials began with a 1-week single-blind placebo,
lead-in period. At the end of the lead-in period, patients
who still met the entry criteria were randomized to
receive 8 weeks of double-blind treatment with active
drug (escitalopram or citalopram) or placebo.
Burke et al. (2002) employed a fixed-dose design, with
three active treatment groups: (i) 10mg/day escitalo-
pram; (ii) 20mg/day escitalopram; and (iii) 40mg/day
Table 1
two studies
Patient disposition, demographics, and mean baseline Montgomery A˚sberg Depression Rating Scale (MADRS) total scores for the
Lepola et al. (2003)Burke et al. (2002)
PBOESC CIT PBOESC CIT
10–20mg20–40mg10mg20mg40mg
Disposition
Patients randomized
Patients treated
Efficacy population (ITT)
Patients completed (%)a
Patients withdrawn (%)a
Reason for withdrawal:
Adverse event(s) (%)
Lack of efficacy (%)
Other (%)
Demographics
Men/women (%)
Mean age, years
Mean weight, kg
Baseline MADRS, mean
No. of severely ill patients (in ITT
population)b
154
154
154
156
155
155
161
160
159
127
122
119
124
119
118
128
125
123
127
125
125
139 (90)
15 (10)
146 (94)
9 (6)
152 (95)
8 (5)
91 (75)
31 (25)
95 (80)
24 (20)
94 (75)
31 (25)
93 (74)
32 (26)
4 (3)
5 (3)
6 (4)
4 (3)
0 (0)
5 (3)
6 (4)
1 (1)
1 (1)
3 (3)
6 (5)
5 (4)
3 (3)
13 (10)
0 (0)
18 (14)
11 (9)
1 (1)
20 (16)22 (18)16 (13)
28/7225/7531/6941/59 29/7133/67 38/62
43
76
28.7
58
43
71
29.0
69
44
74
29.2
69
40
82
29.5
59
41
76
28.0
42
40
80
28.9
51
40
81
29.2
60
PBO, Placebo; CIT, citalopram; ESC, escitalopram.
aExpressed as a percentage of patients treated;
bMADRS Z30 at baseline.
150
International Clinical Psychopharmacology
2004, Vol 19 No 3
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Page 3

citalopram. Patients in the 20mg/day escitalopram group
and in the 40mg/day citalopram group were titrated to
their final dose after 1 week of treatment at half the
assigned dose. No further adjustment of dosage was
permitted. Lepola et al. (2003) employed a flexible-dose
design, with initial doses of 10mg/day for escitalopram
and 20mg/day for citalopram. Based on clinical response,
as measured by CGI-S, the dose could be increased to a
maximum of 20mg/day escitalopram or 40mg/day citalo-
pram. Increases in dose were allowed at week 4 or week 6
and the dose could subsequently be returned to the
initial dose if adverse events emerged.
Evaluations were conducted after 1, 2, 4, 6 and 8 weeks
(Burke et al., 2002) or after 1, 2, 3, 4, 6 and 8 weeks of
double-blind treatment (Lepola et al., 2003). Efficacy
assessments included the MADRS and the Clinical
Global Impression–Improvement and Severity of Illness
scales (CGI-I and CGI-S) (Guy, 1976). In the fixed-dose
study, HAMD-24 was also rated (Burke et al., 2002).
The primary efficacy outcome measure in both trials was
the change from baseline in MADRS total score at
endpoint using the last observation carried forward
(LOCF). The efficacy analyses were based upon the
pooled intent-to-treat (ITT) population, which com-
prised all patients from both studies who received at least
one dose of double-blind study product and who had at
least one valid post-baseline MADRS assessment, with
the exception of those in the 10mg/day escitalopram
fixed-dose arm (Burke et al., 2002), to compare equivalent
doses of escitalopram and citalopram [20 versus 40mg/
day for Burke et al. (2002) and 10–20 versus 20–40mg/day
for Lepola et al. (2003)].
The rating scale scores were analysed using analyses of
covariance, while the remission and response rates were
analysed using logistic regression. In both types of
analyses, treatment group and centre were used as factors
and baseline efficacy score as a covariate. For all variables,
analyses were conducted using both LOCF and observed
cases (OC). Time to response was analysed using
parametric survival analysis, including factors for centre
and treatment.
Results
Patient demographics
Table 1 presents the patient disposition, demographics,
and mean baseline MADRS total score for the two
studies. There were no significant differences between
treatment groups in sex, mean age and weight, or mean
MADRS total score at baseline.
Efficacy
Table 2 presents the adjusted mean changes from
baseline in MADRS total score and CGI-S score after
8 weeks of treatment with placebo, escitalopram, or
citalopram for the two studies separately and for the
pooled analysis (OC and LOCF). Escitalopram was
statistically superior to placebo in both studies. The
escitalopram group had a consistently numerically greater
decrease in MADRS total score and CGI-S score
compared to the citalopram group, with the difference
approaching significance (P<0.10) in one or both studies
(Table 2).
When the data from both studies were pooled (Fig. 1),
the mean decrease from baseline in MADRS total score
Table 2
A˚sberg Depression Rating Scale (MADRS) total score and Clinical
Global Improvement–Severity of Illness (CGI-S) score after 8
weeks of treatment with placebo, citalopram, or escitalopram in
each study and in the pooled analysis (OC and LOCF)
Adjusted mean changes from baseline in Montgomery
MADRSCGI-S
n
OCLOCF OC LOCF
Lepola et al. (2003)
PBO
CIT
ESC
Burke et al. (2002)
PBO
CIT40
ESC10
ESC20
Pooled analysis
PBO
CIT
ESC
154
159
155
–13.6
–14.6
–15.9**
–12.2
–13.7
–15.1**
–1.56
–1.68
–1.90*
–1.42
–1.58
–1.80*,w
119
125
118
123
–10.4
–13.8**
–14.9***
–16.2***,w–14.1***,w–1.66***
–9.4
–12.1*
–13.4***
–0.98
–1.38*
–1.58***
–0.86
–1.18*
–1.40***
–1.42***,w
273
284
278
–12.6
–14.5**
–16.2***,z–15.0***,z–1.86***,z–1.71***,z
–11.3
–13.4**
–1.39
–1.61*
–1.25
–1.48*
OC, Observed cases; LOCF, last observation carried forward; PBO, placebo;
CIT, citalopram; ESC, escitalopram.
*P<0.05;**P<0.01;***P<0.001 versus placebo;wP<0.10;zP<0.05 versus
citalopram.
Fig. 1
15
0
02468
20
25
30
Week
Estimated Mean MADRS Total Score
PBO n=273
CIT n=284
ESC n=278
LOCF
***
***
***
**
*
**
***#
**
**
#
#
Estimated mean MADRS total scores during 8 weeks of treatment with
placebo, citalopram, or escitalopram. Pooled results from both studies
[ITT, OC and LOCF (endpoint)]. *Pr0.05; **Pr0.01; ***Pr0.001
versus placebo; #Pr0.05 versus citalopram.
Do escitalopram and citalopram have similar efficacy? Lepola et al.
151
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Page 4

after 8 weeks of treatment was statistically significantly
greater in the escitalopram group than in the citalopram
group, yielding estimated differences of 1.77 and 1.62
(OC and LOCF, P=0.015 and P=0.026, respectively)
(Table 2 and Fig. 2). Escitalopram showed statistically
significant separation from placebo at week 1 and
every visit thereafter, whereas citalopram first showed
statistically significant separation from placebo at week 6
(Fig. 1). The clinical relevance of the decrease in
MADRS total score was reflected in the change from
baseline in the CGI-S scores for the pooled population.
Patients in the escitalopram group had a significant
decrease in mean CGI-S score compared to those in the
citalopram group, in both the OC and LOCF data sets
(P=0.013 and P=0.015, respectively) (Table 2 and
Fig. 3).
A statistically significantly greater proportion of patients
responded to escitalopram than to citalopram (defined as
Z50% decrease from baseline in MADRS total score)
(P=0.009, OC; P=0.033, LOCF). An analysis of time to
response showed that escitalopram-treated patients
responded significantly faster than citalopram-treated
patients (P=0.0057). In addition, a significantly greater
proportion of escitalopram-treated patients than citalo-
pram-treated patients achieved remission (MADRS total
score r12) (P=0.044, OC; P=0.123, LOCF) (Table 3
and Fig. 4). Based on the LOCF data, 7.9% more patients
were responders to escitalopram than to citalopram at
week 8, while 5.6% more patients were in remission. This
means that for every 13 patients treated with escitalo-
pram or citalopram, one more patient will be a responder
if treated with escitalopram whereas, for remission, it is
one patient for every 18 patients treated. Based on OC
data, 10.0% more patients were responders to escitalo-
pram than to citalopram at week 8, while 7.3% more
patients were in remission. These numbers correspond to
one additional responder for every 10 patients treated,
and one additional patient achieving remission for every
14 patients treated with escitalopram.
In the fixed-dose study, escitalopram-treated patients
had a significant reduction in HAMD-17 total score
compared to the citalopram-treated patients (estimated
differences 1.90; P=0.024, OC; 1.62; P=0.034, LOCF)
(Fig. 5).
The analysis of the pooled results in the severely ill
patients (baseline MADRS Z30) revealed a significantly
greater decrease in MADRS total score in the escitalo-
pram group than the citalopram group (estimated
difference 2.92; P=0.029, OC; estimated difference
2.67; P=0.043, LOCF) (Fig. 6). The difference between
escitalopram and citalopram in the severely ill was so
Fig. 2
0
1
2
3
4
OCLOCF
Estimated Difference from Placebo in
Mean MADRS Total Score
ESC n=278
CIT n=284
**
*** #
*** #
**
Estimated difference from placebo in mean MADRS total scores after
8 weeks of treatment with either citalopram or escitalopram. Pooled
results from two studies (ITT, OC and LOCF). **P<0.01;
***P<0.001 versus placebo; #P<0.05 versus citalopram.
Fig. 3
0
0.2
0.4
0.6
OC
LOCF
Estimated Difference from Placebo in
Mean CGI-S Score
ESC n=278
CIT n=284
*
*
*** #*** #
Estimated difference from placebo in mean CGI-S scores after 8 weeks
of treatment with either citalopram, or escitalopram. Pooled results from
both studies (ITT, OC and LOCF). *P<0.05; ***P<0.001 versus
placebo; #P<0.05 versus citalopram.
Table 3
A˚sberg Depression Rating Scale (MADRS) total score] or in
remission (MADRS total score r12) at week 8: pooled analysis
of the two studies
Patients responding [Z50% decrease in Montgomery
PlaceboEscitalopram Citalopram
Responders (%)
OC
LOCF
Remitters (%)
OC
LOCF
40.7
36.6
63.6***,z
56.8***,w
53.6**
48.9**
37.2
33.7
51.6***,w
46.4***
44.3
40.8*
OC, Observed cases; LOCF, last observation carried forward.
*P<0.05;**P<0.01;***P<0.001 versus placebo;wP<0.05;zP<0.01 versus
citalopram.
152
International Clinical Psychopharmacology
2004, Vol 19 No 3
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Page 5

great that it was also statistically significant for the fixed-
dose study alone (estimated difference 4.24, 20mg/day
escitalopram versus 40mg/day citalopram; P=0.045, OC;
P=0.125, LOCF).
Tolerability
Both citalopram and escitalopram were well tolerated,
with similar adverse event profiles. Only headache
(placebo 20%, escitalopram 16% and citalopram 19%)
and nausea [placebo 8%, escitalopram 16% (P<0.05
versus placebo) and citalopram 18% (P<0.05 versus
placebo)] were reported by Z10% of the patients in
any treatment group in the pooled analysis. The overall
rates of withdrawal were similar between treatment
groups within the two studies, although they were
generally higher in Burke et al. (2002) than in Lepola
et al. (2003) (Table 1). For patients in the flexible-dose
study, the withdrawal rate due to adverse events was
similar between the treatment groups and at the
same level as that of the placebo group (Table 1) (Lepola
et al., 2003). For patients in the fixed-dose study,
the withdrawal rates due to adverse events were similar
in the 20mg/day escitalopram group and the 40mg/
day citalopram group (10% and 9%, respectively) (Burke
et al., 2002).
Discussion
The results of the present pooled analyses demonstrated
a clinically relevant superiority of escitalopram over
citalopram. This difference was supported by the higher
response and remission rates and CGI scores for
escitalopram. These findings are particularly noteworthy
in light of the current literature showing not only
difficulties in separating efficacy between antidepres-
sants, but also between established antidepressants and
placebo (Khan et al., 2002).
Fig. 4
0
5
10
15
20
25
30
Response Remission
Percent Difference from Placebo
ESC n=278
CIT n=284
**
*** ##
*** #
Difference from placebo in the proportions of patients responding
(Z50% decrease in MADRS total score) and in remission (MADRS
total score r12) after 8 weeks of treatment with either citalopram or
escitalopram. Pooled results from both studies (ITT, OC). **P<0.01;
***P<0.001 versus placebo; #P<0.05; ##P<0.01 versus
citalopram.
Fig. 5
0
1
2
3
4
OCLOCF
Estimated Difference from Placebo in
Mean HAMD-17 Score
ESC n=123
CIT n=125
*** #
*** #
*
Estimated difference from placebo in mean HAMD-17 scores after 8
weeks of treatment with either citalopram or escitalopram in the fixed-
dose study (ITT; OC and LOCF). *Pr0.05; ***Pr0.001 versus
placebo; #Pr0.05 versus citalopram.
Fig. 6
15
00
2468
20
25
30
Week
Estimated Mean MADRS Total Score
PBO n=117
CIT n=129
ESC n=120
***#
***#
**
*
*
**#
LOCF
Estimated mean MADRS total scores for severely ill patients (MADRS
Z30 at baseline) during 8 weeks of treatment with placebo,
citalopram, or escitalopram. Pooled results from both studies [ITT, OC
and LOCF (endpoint)]. *Pr0.05; **Pr0.01; ***Pr0.001 versus
placebo; #Pr0.05 versus citalopram.
Do escitalopram and citalopram have similar efficacy? Lepola et al.
153
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.