You are on page 1of 26

Agomelatine Page 1

London New Drugs Group


APC/DTC Briefing Document

June 2009

AGOMELATINE

Summary
Contents
The drug and the review
Summary 1
Background 5 • Agomelatine is a new antidepressant with selective agonist actions at mela-
Safety 6 tonin receptors and selective antagonist action at serotonin 5HT-2C recep-
Efficacy studies 7 tors. It does not affect the uptake of serotonin, noradrenaline or dopa-
Short term studies 9 mine.
Short term studies with • Agomelatine was launched in the UK in June 2009.
active controls 12
Long term/ prevention
• This review evaluates the evidence supporting agomelatine use in adults
of relapse 13 with major depressive episodes and seeks to define agomelatine’s potential
Absence of discontinuation place in therapy.
symptoms 14
Sexual dysfunction 15 Background
Sleep disorders 17 • NICE Guidance (2004 and draft update 2009) recommends an SSRI antide-
Rating scales and Sleep
pressant (e.g. generic fluoxetine or citalopram) first-line for patients with
assessments 19
Cost per 100,000 moderate to severe depression.
population 20 • Following first-line treatment failure the evidence for sequencing antide-
References 21 pressants is inconclusive, choice should therefore be based on the needs of
Appendices 24 the patient, the profile of the drug and financial considerations.
• New NICE Guidelines on the treatment of depression in adults are antici-
pated in September 2009.

Literature searched
Produced for the • We searched: Medline (agomelatine.af [Limit to: Humans and English Lan-
London New Drugs Group guage]); Embase (agomelatine.af [Limit to: Human and English Language]
by: and [DEPRESSION/dt [Drug Therapy] or *MAJOR DEPRESSION/dt [Drug
Therapy]]); and IDIS ("AGOMELATINE 28160472").
Alexandra Denby,
Regional MI Manager • This was supplemented with information from the EMEA website (Public
Medicines Information Service Assessment Report), NICE guidance, and contact with the manufacturer
Northwick Park Hospital Servier Laboratories.
Middlesex • The study programme was extensive:
HA1 3UJ • 1 short-term placebo-controlled dose ranging study (8 weeks)
Tel: 020 8869 3551 • 2 short-term placebo-controlled studies (6 weeks with optional exten-
Med.info@nwlh.nhs.uk sion for a further 46 weeks)
Further copies of this document • 3 short –term placebo-controlled studies with paroxetine or fluoxetine
are available from URL: active controls (6 weeks and optional extension to 6 months)
www.nelm.nhs.uk
• 2 long-term relapse prevention studies.
• 1 study vs. venlafaxine (6 weeks with optional extension to 6 months),
primary endpoint sleep, secondary endpoint efficacy
• 1 study vs. venlafaxine (12 weeks with optional extension to 6
months), primary endpoint sexual function, secondary endpoint effi-
cacy

THIS IS AN NHS THIS IS AN NHS DOCUMENT


DOCUMENT NOT TONOTBE
TO BE USEDFOR
USED FOR COMMERCIAL
COMMERCIAL AND MARKETING PURPOSES. PURPOSES.
AND MARKETING
PRODUCED TO INFORM LOCAL DECISION-MAKING USING THE BEST AVAILABLE EVIDENCE AT THE TIME OF PUBLICATION.
PRODUCED TO INFORM LOCAL DECISION-MAKING USING THE BEST AVAILABLE EVIDENCE AT THE TIME OF
June 2009 PUBLICATION.
London New Drugs Group APC/DTC Briefing
Agomelatine Page 2

• 1 study vs. sertraline (6 weeks with extension to 6 months), primary endpoint actigraphy,
secondary endpoint efficacy
• 1 open study in patients with major depressive disorder (6 weeks) looking at the effect on
sleep EEG
• 1 study vs. fluoxetine in patients with severe depression (8 weeks with optional extension to
6 months)
• 1 study on sexual disturbance vs. paroxetine.

Efficacy studies
• Antidepressant efficacy is generally determined from 6–8 week RCTs comparing response
against placebo or active comparators, or 6–10 month RCTs comparing relapse rate against pla-
cebo; depression symptoms are evaluated using rating scales.
• All analyses were carried out in the intention-to-treat populations: i.e. analysis was performed
according to the assigned treatment group regardless of protocol deviations and participant
compliance or withdrawal.

Short-term placebo-controlled
• A dose-ranging study (n=711) identified agomelatine 25mg daily as the target dose when com-
pared with agomelatine 1mg and 5mg daily
• In two 6 week placebo controlled RCTs patients with a HAMD score ≥ 22 [moderate to severe
depression] were randomised to either agomelatine (25mg initial for 2 weeks increased to
50mg in non-responders) or placebo; primary endpoint was mean final HAMD score.
• In the first study, 212 patients received agomelatine (n=106) or placebo (n=105). For
the ITT population the between group difference for the mean final HAMD scores was
2.30 (S.E. 1.02), p=0.026.
• In the second study, 238 patients received agomelatine (n=118) or placebo (n=120).
For the ITT population the between group difference for the mean final HAMD scores was
3.44 (S.E. 0.92), p<0.001
• 3 unpublished placebo-controlled studies in which paroxetine or fluoxetine were used as active
controls to validate the study design are reported in the EMEA public assessment report (EPAR).
The efficacy of agomelatine was not directly compared to that of the active controls.
• All 3 studies enrolled patients with moderate to severe depression [HAMD ≥ 22] and fol-
lowed similar methodologies: initial run-in followed by randomisation to agomelatine,
placebo, or active control for 6 weeks with subsequent 18 week extensions.
• The option to increase the agomelatine dose from 25mg to 50mg was not included in
these studies.
• The primary endpoint of difference in HAMD was compared between active control and
placebo, and between agomelatine and placebo. For the first trial, a statistically signifi-
cant difference (p=0.008) was seen between active control and placebo at 6 weeks for
reduction in HAMD score versus placebo, but not for agomelatine and placebo; Neither
the second nor third trial showed statistically significant differences between agomelatine
or the active control and placebo for any comparison at either 6 or 24 weeks, meaning
that no discernable results could be drawn from these studies. A meta-analysis of
agomelatine trials shows that 23.5% of patients require a 50mg dose, which is a reason
why agomelatine may have failed to differentiate from placebo in the first study. The
second and third studies were associated with high placebo response rates.

Active comparator studies


• A 6 week study randomised 313 patients with a HAMD score ≥ 22 to agomelatine 25–50mg
(n=154) or sertraline 50–100mg (n=159) [for each drug the lower initial dose was increased
after 2 weeks for non-improved patients]. The primary endpoint of the study was the efficacy
on rest-activity circadian rhythms. For the secondary endpoint of an improvement in HAMD
score, the difference in scores after 6 weeks was 1.68, in favour of agomelatine (p=0.031).
• There were two studies in patients with moderate to severe depression in which agomelatine
was compared with venlafaxine.

THIS IS AN NHS DOCUMENT NOT TO BE USED FOR COMMERCIAL AND MARKETING PURPOSES.
PRODUCED TO INFORM LOCAL DECISION-MAKING USING THE BEST AVAILABLE EVIDENCE AT THE TIME OF PUBLICATION.

June 2009 London New Drugs Group APC/DTC Briefing


Agomelatine Page 3

• In the 12 week study patients were randomised to agomelatine 50mg (n=137) or venla-
faxine 150mg (n=140). The primary endpoint of the study was deterioration in sexual func-
tion and demonstrated statistically significant differences in favour of agomelatine versus
venlafaxine. For the secondary endpoint of the final MADRS score, agomelatine was shown
to be at least as effective as venlafaxine (10.1 vs. 9.8).
• In the 6 week study patients were randomised to either agomelatine (n=165) or venlafaxine
(n=167). The primary endpoint was the effects on sleep variables and demonstrated statis-
tically significant differences in favour of agomelatine versus venlafaxine; final HAMD score
was the secondary endpoint. The final HAMD scores were similar between the groups, indi-
cating similar efficacy for agomelatine and venlafaxine (9.9 vs. 11.0).
• A 8 week study randomised 500 outpatients with HAMD score ≥25 and CGI severity of illness score
≥4 (severely depressed) to agomelatine 25-50mg (n=247) or fluoxetine 20-40mg (n=257) [for
agomelatine the initial lower dose was increased at 2 weeks in non-improved patients, for fluoxet-
ine the initial lower dose was increased at 4 weeks]. The primary endpoint was improvement in
HAMD score from baseline, the difference in scores at week 8 was 1.49, 95% CI=[0.20;2.77]
(p=0.024) in favour of agomelatine.

Long-term relapse-prevention
• Relapse is classified as the appearance of the symptoms of the index episode soon after medica-
tion is stopped, whilst recurrence is the appearance of symptoms in a new episode. Relapse indi-
cates that the treatment duration was too short.
• A study reported in the EPAR had an 8-week open label phase (n=610) during which response to
agomelatine was determined, followed by a 26-week double-blind randomised phase (n=367;
agomelatine 25mg, n=187; placebo, n=180).
• At the end of the 26 week double-blind phase, a significant effect was not demonstrated for
the primary endpoint of relapse (agomelatine 25.9% vs. placebo 23.5%). The placebo re-
lapse rate in this trial was unexpectedly low (approximately half that seen in other studies of
similar design). This may be reflective of some methodological aspects e.g. the broad range
of the severity of depression of patients at inclusion.
• 169 severely depressed patients (n=89 agomelatine, n=80 placebo) (HAMD >25 and CGI-S
≥5) continued in an 18 week extension; at the end of this extension there was a statistically
significant difference in favour of agomelatine (21.3% vs. 31.3%) p=0.046.
• A second (published) study, that included a high proportion of patients with moderate to severe
depression reflecting the types of patient generally seen by psychiatrists, had an 8-10 week open-
label dose determination phase (n=492) followed by a 24 week double-blind phase (n=390)
• For the primary endpoint of relapse, results at 24 weeks were significantly lower (p=0.0001)
for agomelatine (21.7%) than for placebo (46.6%); agomelatine reduced relapse risk by
54% (HR 0.458; 95% CI 0.305–0.60).
• 190 patients (n=106 agomelatine, n=84 placebo) continued in double-blind 20 week exten-
sion. A statistically significant difference in the time to relapse was detected: at 10 months
twice as many patients treated with placebo relapsed (49.9%) compared to those treated
with agomelatine (23.9%) (p<0.0001).
• The three short term studies with active controls all had 18 week extensions. In one study the
relapse rates were lower with agomelatine (14.3%) and fluoxetine (17.8%) than with placebo
(33.3%), p=0.017 and p=0.045 respectively. In the other 2 studies there were no differences
seen between agomelatine or active control and placebo which means no discernable results could
be drawn.

Safety
• Agomelatine does not cause weight gain, has a low risk of sexual dysfunction, low incidence of
gastro-intestinal reactions, absence of discontinuation symptoms, no sedation or daytime drowsi-
ness and an overall incidence rate of adverse events that is similar to placebo.
• Emergent elevations of liver transaminase enzymes more than 3x upper limit of normal (3x ULN)
occur rarely (approx. 1% of patients). Overall incidences of elevations >3xULN in all patients (i.e.
regardless of their transaminase values at baseline) were 1.1% with agomelatine 25mg and
0.72% with placebo (p=not significant). The 0.4% absolute difference with placebo compares to a

THIS IS AN NHS DOCUMENT NOT TO BE USED FOR COMMERCIAL AND MARKETING PURPOSES.
PRODUCED TO INFORM LOCAL DECISION-MAKING USING THE BEST AVAILABLE EVIDENCE AT THE TIME OF PUBLICATION.

June 2009 London New Drugs Group APC/DTC Briefing


Agomelatine Page 4

0.7% absolute difference seen with duloxetine/placebo.


• Agomelatine should not be prescribed in patients with cirrhosis or active liver disease.
• Check liver function tests (LFTs) on initiation; then at around 6, 12 and 24 weeks; and thereafter
when clinically indicated (if transaminases are elevated repeat LFTs within 48 hours; if transami-
nases are >3x ULN, discontinue agomelatine).
• Agomelatine should be used with caution in any patient who drinks substantial quantities of alcohol
or is treated with other medicines associated with a risk of hepatic injury.
• Agomelatine should not be prescribed in patients taking potent CYP1A2 inhibitors, such as fluvox-
amine and ciprofloxacin, as this may result in increased serum levels of agomelatine. Moderate
CYP1A2 inhibitors, such as propranolol, should be used concomitantly with caution.

Critical evaluation of the evidence base

Use of depression rating scales


• HAMD and MADRS measure similar quantifiable elements of depression; CGI measures clinician
rated global improvement and is thus less sensitive
• Rating scales may not capture functional impairment appropriately despite this being an important
determinant of depression severity
• NICE suggest that clinical efficacy is demonstrated by a weighted mean between-group difference
of at least three points or a standardised mean difference of at least 0.5 in HAMD rating score
• EMEA guidance states that acceptable scales for use as the primary endpoint in trials of new anti-
depressants to determine symptomatic improvement include the Hamilton Depression rating scale
of Depression (HAMD) or the Montgomery Asberg Depression Rating Scale (MADRS). The HAMD
scale (17-item scale as recommended) was used for the majority of the agomelatine trials; others
used the MADRS.

The placebo effect in depression studies


• Use of a placebo in an antidepressant trial does not necessarily mean that the patient is untreated.
Patients will have regular visits to their doctor, supportive help and interest in their welfare. In
some trials they will be able to contact the therapist when they need to. Patients in the placebo
arm will receive everything apart from the active drug, and this can constitute a treatment in itself.
It has been estimated that 30% of patients assigned to placebo will respond within six weeks.
• Despite this, SSRIs and serotonin/noradrenaline reuptake inhibitor (SNRI) antidepressants remain
superior to placebo in treating depression regardless of its severity.
• Study population selection (and enrolment of severely depressed patients) is particularly important
for external validity of antidepressant trials.

Particular considerations for agomelatine


• In the two short-term placebo-controlled studies, statistically significant improvement in symp-
toms of depression was seen with agomelatine compared with placebo.
• When compared with other antidepressants, agomelatine was more efficacious than sertraline and
fluoxetine, and as effective as venlafaxine in treating the symptoms of depression.
• Agomelatine demonstrated significant efficacy in only one of the two placebo-controlled relapse-
prevention studies. In the first study the relapse rate in the placebo arm was unexpectedly low
and may reflect some of the methodological aspects, such as the broad range of the severity of
depression of patients included in the study. A more severely depressed cohort was enrolled in
the second trial, which demonstrated benefit (80% of enrolled patients in the second study vs.
46% in the first study).

• Overall, the data demonstrate a potential role for agomelatine in treating adults with major de-
pressive disorder. The efficacy studies suggest comparable efficacy to currently licensed antide-
pressants.

THIS IS AN NHS DOCUMENT NOT TO BE USED FOR COMMERCIAL AND MARKETING PURPOSES.
PRODUCED TO INFORM LOCAL DECISION-MAKING USING THE BEST AVAILABLE EVIDENCE AT THE TIME OF PUBLICATION.

June 2009 London New Drugs Group APC/DTC Briefing


Agomelatine Page 5

Potential benefits over existing therapy


• As with any new antidepressant, agomelatine’s place in therapy is not yet clear. Agomelatine
may be useful in the case of failure of another antidepressant, either because of lack of effi-
cacy or due to poor tolerability (e.g. unacceptable side effects). It has demonstrated superior
efficacy to Sertraline in one study, fluoxetine in another study and demonstrated benefits over
venlafaxine in two other studies.
• Abrupt withdrawal of agomelatine (after 12 weeks of therapy) has not been shown to be associ-
ated with discontinuation symptoms, which can be problematic with other antidepressants. This
means that no dose tapering is necessary on treatment discontinuation.
• There is suggestion that sexual dysfunction caused by agomelatine is less than that caused by
venlafaxine The pooled analysis of studies using the Arizona Sexual Experience Scale (ASEX)
showed that agomelatine was not associated with sexual dysfunction. In healthy volunteers
agomelatine preserved sexual function in comparison with paroxetine.
• Agomelatine had neutral effect on body weight, heart rate and blood pressure in clinical studies.
• Sleep studies have shown that agomelatine increases sleep efficiency and slow-wave sleep and
has significantly better effects on subjective sleep variables than venlafaxine.
• Agomelatine has a good tolerability profile compared with other classes of antidepressants.

Potential disadvantages over existing therapy


• There is a lack of long term active comparator controlled studies. As such defining a place in
therapy in comparison with the other better established alternatives is not possible. For a new
antidepressant agomelatine has robust data for both short and long term treatment effective-
ness, which will need to be validated in real life clinical practice.

Health economics
• No health economic studies of agomelatine have been published.

Budget impact model: per 100,000 population


• The budget impact model is based on the use of agomelatine as second-line or later treatment.
The incidence of this is estimated at 0.40%, affecting 400 patients per 100,000.
• If agomelatine takes 10% of patients from each second-line treatment, 40 patients would be
treated with it, at a cost of £38.53 per 28 days.
• Over the first year, taking into account a reduction in prescribing costs of the other second line
treatments and the need for liver function tests, the net increase in prescribing costs is esti-
mated to be £5,892 (1.45% increase in spend). This would rise to an increase of £9,285
(2.29% increase in spend) in the second year and £11,607 (2.87% increase in spend) in the
third year, as more patients are prescribed agomelatine.

THIS IS AN NHS DOCUMENT NOT TO BE USED FOR COMMERCIAL AND MARKETING PURPOSES.
PRODUCED TO INFORM LOCAL DECISION-MAKING USING THE BEST AVAILABLE EVIDENCE AT THE TIME OF PUBLICATION.

June 2009 London New Drugs Group APC/DTC Briefing


Agomelatine Page 6

Background No difference between dosing, effectiveness or


safety of agomelatine between older and younger
Agomelatine was launched in the UK in June adults has been reported from trials.5 There are
2009 for the treatment of major depressive no data on the use of agomelatine in children. No
episodes in adults.1 The recommended dose is specific data have been reported on the safety of
25mg, taken at bedtime, which can be in- it during pregnancy, though animal studies have
creased to 50mg after two weeks if there is no reported no risks.5 Agomelatine should be used
improvement in symptoms. Treatment should cautiously in patients with liver impairment, but
continue for at least six months to ensure that its lack of toxicity and broad safety margin sug-
the patient is free of symptoms. No dose ta- gest lower doses could be used with appropriate
pering is required on treatment discontinua- clinical monitoring.5
tion.2
Special populations
Current NICE guidance (Dec 2004)3 and the
draft 2009 Guidance4 both recommend that an Agomelatine is not recommended for use in chil-
SSRI should be used first line in patients with dren and adolescents below 18 years of age and
moderate to severe depression, as they are as should be used with caution in the elderly (≥65
effective as tricyclic antidepressants and less years of age) due to lack of clinical efficacy data.
likely to be discontinued due to side effects. There is limited clinical data on the use of agome-
Generic versions, such as fluoxetine, paroxet- latine in patents with moderate or severe renal
ine, sertraline and citalopram, would be rea- impairment; it should be used with caution in
sonable choices. Following first line treatment such patient.2
failure (lack of response, withdrawal due to
adverse effects) the evidence for treatment Agomelatine should not be used in patients with
sequencing is weak. Choice should therefore hepatic impairment (i.e. cirrhosis or active liver
be based on the profile of the drug, needs of disease).2
the patient and financial considerations. NICE
Guidelines on the treatment of depression in There is no clinical data on the effect of agome-
adults are anticipated in September 2009. latine on pregnancy. Animal studies do not indi-
cate direct or indirect harmful effects to the foe-
Pharmacology tus. Caution should be exercised when prescrib-
ing to pregnant women.2
Agomelatine is a synthetic analogue of the hor-
mone melatonin, and strongly binds to and Safety
stimulates the activity of melatonin MT1 and Agomelatine does not cause weight gain, has a
MT2 receptors, normalising disturbed circadian low risk of sexual dysfunction, low incidence of
rhythms and disrupted sleep-wake cycles.5 gastro-intestinal reactions, absence of discon-
(Melatonin regulates circadian rhythms, includ- tinuation symptoms and an overall incidence rate
ing sleep-wake cycles.) Disturbances in cir- of adverse events that is similar to placebo.6
cadian rhythms have been implemented in the
development of mood disorders.5 Agomelatine In clinical trials, over 3900 patients have received
is also a serotonin-receptor antagonist and agomelatine. The incidence of emergent eleva-
binds to and inhibits the activity of serotonin tions of liver transaminase enzymes more than 3x
5HT2C receptors.5 This antagonism is associ- upper limit of normal (3xULN) was 0.8% in pa-
ated with antidepressant and anti-anxiety ac- tients treated with agomelatine and with normal
tivity, and increases slow-wave sleep. baseline levels and 0.3% those treated with pla-
cebo. In patients treated with agomelatine 50mg,
Agomelatine does not affect the uptake of se- the incidence was 1.3%.6 The incidences in all
rotonin, noradrenaline or dopamine. The inhi- patients (i.e. regardless of their transaminase val-
bition of 5HT2C receptors increases noradrena- ues at baseline) were 1.04% (agomelatine 25mg),
line and dopamine in the frontal cortex and 1.39% (agomelatine 50mg) and 0.72% (placebo)
may contribute to agomelatine’s antidepres- the differences from placebo are not statistically
sant activity.5 Agomelatine has no effect on significant.6
monoamine uptake and no affinity for a, b
adrenergic, histaminergic, cholinergic, dopa- Liver function tests (LFTs) should be performed at
minergic and benzodiazepine receptors.2 the start of treatment then periodically around 6,
12 and 24 weeks and thereafter when clinically
indicated.2 If serum transaminases rise, the LFTs

THIS IS AN NHS DOCUMENT NOT TO BE USED FOR COMMERCIAL AND MARKETING PURPOSES.
PRODUCED TO INFORM LOCAL DECISION-MAKING USING THE BEST AVAILABLE EVIDENCE AT THE TIME OF PUBLICATION.

June 2009 London New Drugs Group APC/DTC Briefing


Agomelatine Page 7

should be repeated within 48 hours. Therapy patients assigned to placebo will respond within
should be discontinued if the increase exceeds six weeks.4
3xULN and liver function tests should be per-
formed regularly until they return to normal.2 If The problem for the evaluation of antidepres-
any patient develops symptoms which suggest sants is that the study population is likely to de-
hepatic dysfunction, LFTs should be carried out termine the magnitude of the treatment benefit.
and the decision to continue therapy should be a A poorly selected study population will make a
clinical one, based on laboratory evaluations. If drug look bad; a well selected study population
jaundice develops, treatment must be stopped. will make a drug look better. Since treatment of
Agomelatine should be used with caution in pa- real patients is not subject to the perverse in-
tients who consume a large quantity of alcohol or centives of drug trials, it is often assumed that
who are treated with other drugs that may cause trials that recruit a more severe study popula-
hepatic damage.2 tion corresponds better to everyday practice
than trials that do not. The existence of poorly
Agomelatine is metabolised mainly by cytochrome conducted trials showing minimal benefits can-
P450 1A2 (CYP1A2) (90%) and by CYP2C9 not be ignored, however, and has been used to
(10%). Treatment with agomelatine is contra- case doubt on the efficacy of antidepressants as
indicated in patients who are also taking potent a class.
CYP1A2 inhibitors, such as fluvoxamine and cipro-
floxacin as they can increase the serum levels of Efficacy studies
agomelatine. Use of moderate CYP1A2 inhibitors,
such as oestrogens and propranolol with agome- A number of studies have been carried out as-
latine should be with caution until more experi- sessing the efficacy and safety of agomelatine.
ence has been gained.2 Agomelatine does not These are described in detail in this review, and
induce CYP450 isoenzymes and will not modify their design and main results are summarised in
exposure to medicinal products metabolised by Appendix 1. All patients enrolled had moderate
CYP450.2 to severe depression. The study programme
was extensive:
The placebo response
• 1 short-term placebo-controlled dose ranging
Over the years there has been an increase in the study (8 weeks)7
percentage of patients responding in the placebo • 2 short-term placebo-controlled studies (6
arms of published studies of major depression.3 weeks with optional extension for a further
There have been no studies comparing placebo 46 weeks)8;9
with no treatment in depression, so these patients • 3 short –term placebo-controlled studies
represent several different kinds of treatment ef- with paroxetine or fluoxetine active controls
fect, such as patients who may have had their (6 weeks and optional extension to 6
depression scores inflated to ensure study entry, months)6
patients who were accurately rated but show an
• 2 long-term relapse prevention studies6;10;11
early spontaneous remission and true responders
to placebo.3 The difference in improvements in
• 1 study vs. venlafaxine (6 weeks with op-
depressive symptoms between the drug and the tional extension to 6 months), primary end-
placebo is thought to be greater with increasing point sleep, secondary endpoint efficacy12
degrees of severity of depression. This is often • 1 study vs. venlafaxine (12 weeks with op-
difficult to demonstrate as data is quoted for en- tional extension to 6 months), primary end-
tire groups, rather than individual patients and point sexual function, secondary endpoint
the distribution of the depression rating scores efficacy13
can overlap, thereby diluting the effect.3 • 1 study vs. sertraline (6 weeks with exten-
sion to 6 months), primary endpoint actigra-
Use of a placebo in an antidepressant trial does phy, secondary endpoint efficacy14;15
not necessarily mean that the patient is un- • 1 open study in patients with major depres-
treated.4 Patients will have regular visits to their sive disorder (6 weeks) looking at the effect
doctor, supportive help and interest in their wel- on sleep EEG16
fare. In some trials they will be able to contact • 1 study vs. fluoxetine in patients with severe
the therapist when they need to. Patients in the depression (8 weeks with optional extension
placebo arm will receive everything apart from to 6 months) (conference abstract only)
the active drug, and this can constitute a treat- • 1 study on sexual disturbance vs. paroxeti-
ment in itself. It has been estimated that 30% of ne18

THIS IS AN NHS DOCUMENT NOT TO BE USED FOR COMMERCIAL AND MARKETING PURPOSES.
PRODUCED TO INFORM LOCAL DECISION-MAKING USING THE BEST AVAILABLE EVIDENCE AT THE TIME OF PUBLICATION.

June 2009 London New Drugs Group APC/DTC Briefing


Agomelatine Page 8

sponder rates. The agomelatine/placebo differ-


Guidance from the EMEA19 for the clinical investi- ence in HAMD score was not always greater than
gation of medicinal products for the treatment of 3 points (ref 8 and the active control studies),
depression states that acceptable scales for use the difference considered to be clinically signifi-
as the primary endpoint to determine sympto- cant by NICE. The agomelatine/sertraline HAMD
matic improvement include the Hamilton Depres- score difference was greater than 1.5 points,
sion rating scale of Depression (HAMD, 17-item which is considered clinically significant. The
scale) or the Montgomery Asberg Depression Rat- dose ranging study and the three short term
ing Scale. NICE recommend that where a studies with active controls did not have the op-
weighted mean difference is calculated, a be- tion to increase the agomelatine dose to 50mg,
tween-group (drug/placebo) difference of at least which will results in an underestimate of the effi-
three points (two for treatment-resistant depres- cacy of agomelatine as ~23.5% of patients with
sion) is considered clinically significant for the require a 50mg dose.
HAMD.3 The EMEA also state that the disorder
should be classified according to an internationally Longer term studies will show whether these
acknowledged classification system, preferably effects are sustained over a suitable treatment
the DSM-IV: the clinical trials enrolled patients period of 4-6 months.
who did have depressive disorder according to the
DSM-IV criteria.

Short term studies

NICE Guidance recommends that when treating a


patient with moderate-severe depression, the an-
tidepressant dose should be titrated to the recog-
nised therapeutic dose and the efficacy of this
dose assessed over 4-6 weeks.20 If this is effec-
tive, treatment should be continued for a further
4-6 months, or longer in patients with recurrent
depression, at the full treatment dose.3;20 If the
dose is not effective, it should be increased in line
with the schedule recommended in the Summary
of Product Characteristics. If there has been a
partial response after a month, a decision to
switch to another antidepressant can be post-
poned until 6 weeks of therapy. A poorly toler-
ated antidepressant should be switched to an-
other one.3 The following trials were long enough
to assess the initial efficacy of the treatment, but
too short to make any conclusions about whether
agomelatine is effective in maintaining remission.

Randomised, double-blind comparisons versus


placebo are needed to permit adequate evaluation
of efficacy and for distinguishing disease manifes-
tations from adverse reactions.19 The use of a
placebo is controversial and precautions to mini-
mise the impact of the study should be taken:
generally a duration of 6 weeks should be suffi-
cient and a longer duration should be justified.
Three-arm trials including both a placebo and ac-
tive control are recommended.

Nine short term studies have been conducted.


What these short term studies show is that after
six weeks of treatment agomelatine has a favour-
able effect on the symptoms of depression, as
seen by reductions in HAMD scores and high re-

THIS IS AN NHS DOCUMENT NOT TO BE USED FOR COMMERCIAL AND MARKETING PURPOSES.
PRODUCED TO INFORM LOCAL DECISION-MAKING USING THE BEST AVAILABLE EVIDENCE AT THE TIME OF PUBLICATION.

June 2009 London New Drugs Group APC/DTC Briefing


Short term studies

June 2009
Agomelatine

Study Primary
Treatment Results Comments
design endpoint

• Mean HAMD score at baseline indicates that a sufficient


number of patients had severe depression.
Dose- • Response rates: 61.5% • The higher response rate seen with agomelatine sug-
ranging 8- agomelatine 25mg gests it is a good choice for achieving remission, as
week study Agomelatine (83/135); 46.3% placebo short-term studies do not often show significance dif-
Response,
in 711 pa- 1mg, 5mg, (p=0.036); 56.3% paroxet- ferences. Based on the lower total HAMD-D score and
defined as a
tients with 25mg or pla- ine (not significantly better higher percentages of patients responding and in re-
50% or more
a mean cebo. Par- than placebo). mission, this study shows that 25mg is the target dose
improvement
HAMD score oxetine 20mg • Remission rates (HAMD for agomelatine.
in the HAMD
of 27.4 used to vali- <7): 30.4%, agomelatine • No significant difference between agomelatine and pla-
score from
(entry crite- date study 25mg and 25.7%, paroxet- cebo was seen for any adverse event. Headache, anxi-
baseline.
ria of HAMD results. ine groups (p<0.01 and ety, abdominal pain and diarrhoea were the most com-
score p<0.05 vs. placebo), pla- mon side effects with agomelatine. No clinically rele-
≥22).7 cebo (15.4%). vant changes were seen in vital signs, weight, ECG and
biological parameters. The two deaths that occurred
during the study were not related to treatment.

• HAMD scores:
• at endpoint (overall ITT
HAMD final population): 14.1, agome-
score (ITT latine, 16.5 vs. placebo;
Agomelatine
population) • The mean baseline HAMD score was 26. The dose was
25-50mg difference 2.4, p=0.026.
Secondary increased in 36 patients in the agomelatine group and
daily • Group whose dose was in-
efficacy vari- 38 in the placebo group. Seven out of the 12 in the
(n=106) or creased at wk 2: 17.5,
6-week ran- ables: re- placebo group who discontinued did so because of lack
placebo agomelatine vs. 20.4 pla-
domised, sponse to of efficacy, compared to two out of the seven in the
(n=105). cebo; difference 2.9,
double- treatment agomelatine group who discontinued.
Agomelatine p=0.045.
blind study (≥50% re- • The difference in HAMD scores in the overall population
dose in- • severe depression: 14.4,
in adults duction in is below that defined by NICE as clinically significant.
creased to agomelatine vs. 17.3, pla-
with HAMD HAMD score), Results are detailed further in Appendix 1.
50mg after 2 cebo; difference 2.9,

THIS IS AN NHS DOCUMENT NOT TO BE USED FOR COMMERCIAL AND MARKETING PURPOSES.
score ≥22.8 time to first • Dizziness, nasopharyngitis and influenza were more
weeks in p=0.024).

PRODUCED TO INFORM LOCAL DECISION-MAKING USING THE BEST AVAILABLE EVIDENCE AT THE TIME OF PUBLICATION.
response, common in the agomelatine group, whilst headache,
non- • Response rates: 49.1%,
CGI-Severity nausea, fatigue, dry mouth and diarrhoea were more
improved agomelatine vs. 34.3%, pla-
and CGI- common in the placebo group.
patients cebo, p=0.03.
Improvement
scores. • Remission rates: 20.8%,
agomelatine vs. 13.3%, pla-
cebo.

London New Drugs Group APC/DTC Briefing


Page 9
Study Primary
Treatment Results Comments
design endpoint
• The mean baseline HAMD total score was 27.3±2.8.
18/35 patients discontinuing were in the placebo
group; the most common reason was lack of efficacy.

June 2009
HAMD final
At week two, 28 (25.2%) of patients taking agome-
score (ITT
Agomelatine

latine and 55 (48.6%) of patients taking placebo had


population) • HAMD final score: 14.6,
Agomelatine a dose increase.
Secondary agomelatine vs. 18.1 pla-
25-50mg • Symptoms of depression improved in all patients.
efficacy vari- cebo; difference 3.5,
(n=118) or Greater sleep improvements were seen with agome-
ables: re- p=<0.001.
6-week ran- placebo latine, which would be expected from the pharmacol-
sponse to • Response rates: 54.3%,
domised, (n=120). ogical action of the drug. With agomelatine treat-
treatment agomelatine vs. 35.3%,
double-blind Agomelatine ment, early-night and mid-night insomnia were sig-
(≥50% re- placebo, p=0.003.
study in dose in- nificantly improved and there was a trend towards an
duction in • Decreases in CGI-I and
adults with creased to improvement in early-hours insomnia.
HAMD CGI-S scores indicate im-
HAMD score 50mg after 2 • 3.4% of patients treated with agomelatine discontin-
score), time provements in symptoms
≥22.9 weeks in ued due to adverse events compared with 5.8% on
to first re- of depression, with differ-
non- placebo. ~42% of all patients reported AEs, with the
sponse, CGI- ences between the two
improved majority mild to moderate in severity. Serious ad-
Severity and groups in favour of agome-
patients . verse events occurred in six patients, four in the
CGI- latine.
Improve- agomelatine group, but none were considered to be
ment scores. related to study treatment. Agomelatine was not
found to be associated with sexual side effects such
as impotence, ejaculation difficulties or decreased
libido

Double-
• HAMD final scores: 11.1,
blind, ran-
agomelatine vs. 12.7, • This study has not yet been fully published and there-
domised, 8 Agomelatine
fluoxetine. fore cannot be fully evaluated: information comes
week study 25-50mg Change in
• Change in HAMD score: - from the World Psychiatric Association conference,
in patients (n=247) or HAMD score
17.3, agomelatine, -16.0, April 2009.
with HAMD fluoxetine from base-
fluoxetine. • Doses were increased after 2 weeks (agomelatine) or
score ≥25 & 20-40mg line.
• Between-group difference: 4 weeks (fluoxetine) in non-improved patients.
CGI severity (n=257).
illness score 1.46, p=0.024 in favour of • Patients enrolled in this trial had severe depression.
≥4.17 agomelatine.

Agomelatine • HAMD final scores: 10.3,


Efficacy on agomelatine vs. 12.1, ser-
25-50mg
rest-activity traline; difference 1.68, • This has only been published as a conference poster.
(n=154) or
cycle. p=0.031.
6 week ran- sertraline • Superiority of agomelatine over sertraline was shown

THIS IS AN NHS DOCUMENT NOT TO BE USED FOR COMMERCIAL AND MARKETING PURPOSES.
domised, 50-100mg • Responder rates: 70% after six weeks of treatment.

PRODUCED TO INFORM LOCAL DECISION-MAKING USING THE BEST AVAILABLE EVIDENCE AT THE TIME OF PUBLICATION.
Secondary
double-blind (n=159). agomelatine vs. 61.5% ser- • Most commonly reported adverse events were dry
endpoint:
study in Doses were traline, p=0.119. mouth, headache, diarrhoea and fatigue. More pa-
change in
adults with increased • Between-group difference tients discontinued sertraline (11.3%) than agome-
HAMD score
HAMD score after 2 for CGI-S: 0.28, p=0.043. latine (2.6%).
from base-
≥22.14;15 weeks in • Between-group difference • See Sleep Disorders section for primary endpoint
line of ≥50
non- for CGI-I: 0.29, p=0.023. data.
in ITT popu-

London New Drugs Group APC/DTC Briefing


Page 10

improved • Both CGI differences in fa-


lation
patients . vour of agomelatine.
Study Primary

June 2009
Treatment Results Comments
design endpoint
Agomelatine

Double-
blind, ran-
• There was no significant difference between the treat-
domised 12
ments in the treatment of depression, as shown by the
week study Change in
• MADRS final score: 10.1, reduction in MADRS score and CGI-S and CGI-I score
comparing sexual func-
agomelatine vs. 9.8 venla- changes from baseline. Similar percentages of patients
the sexual Agomelatine tion.
faxine. were responders and remitters in both groups.
side effects 50mg
• Responder rates: 82.5% • 38.1% of patients on venlafaxine had adverse events vs.
of agome- (n=137) or Secondary
agomelatine vs. 79.9% 20.4% on agomelatine. Nausea, headache and upper
latine and venlafaxine endpoint:
venlafaxine. respiratory tract infections were the more common in
venlafaxine 150mg antidepres-
• Remitter rates: 73% both groups. More patients discontinued venlafaxine
in adults (n=140). sant efficacy
agomelatine vs. 66.9% (8.6%, mainly because of gastrointestinal or central
with moder- and toler-
venlafaxine. nervous system side effects) than agomelatine (2.2%).
ate depres- ability
sion • See Sexual dysfunction section for primary endpoint
(MADRS data.
≥20).13
• Agomelatine was as effective as venlafaxine in treating
Agomelatine • HAMD final scores: 9.9 ± depression, in both the overall population and the se-
25mg 6.6, agomelatine vs. 11.0 verely depressed patients.
Random- (n=165) or ± 7.4 venlafaxine. Be-
Leeds Sleep • The time course of improvements in the HAMD score
ised, dou- venlafaxine tween-group difference of
Evaluation was similar for each treatment group and there was no
ble-blind 6 75mg 1.1.
Question- significant difference between the groups at any study
week study (n=167). • Response rates: 76.4%
naire ‘getting visit.
comparing Doses were agomelatine vs. 70.6%
to sleep After an additional 18 weeks more patients taking agome-
the effects increased venlafaxine.
score’ latine (71.6%) compared with those taking venlafaxine
of agome- after 2
• CGI global improvement (66.7%) were still in remission. Bias from the initial fo-
latine and weeks in
Secondary final scores: 1.6±0.7, cus on subjective sleep parameters in favour of agome-
venlafaxine non-
endpoint: agomelatine vs. 1.6±0.8, latine cannot be ruled out. 6
on subject improved
venlafaxine. • Nausea, dizziness and vomiting were more common with
sleep vari- patients
Changes in • venlafaxine than with agomelatine (22.6% vs. 6.0%;
ables. Pa-
HAMD-17 • Severely depressed pa- 9.5% vs. 1.8% and 4.8% vs. 1.2% respectively). Sero-
tients had Patients
score and tients: tonin syndrome occurred in 3% of patients on venla-
HAMD score could con-
CGI- • HAMD final scores: faxine. More patients withdrew due to adverse events in
≥20.12 tinue with an
18-week ex- 11.2±7.0, agomelatine vs. the venlafaxine group (13.2%) vs. the agomelatine

THIS IS AN NHS DOCUMENT NOT TO BE USED FOR COMMERCIAL AND MARKETING PURPOSES.
tension.6 11.2±6.9, venlafaxine. group (4.2%).

PRODUCED TO INFORM LOCAL DECISION-MAKING USING THE BEST AVAILABLE EVIDENCE AT THE TIME OF PUBLICATION.
• See Sleep Disorders section for primary endpoint data.
Hamilton Depression Rating Scale (HAMD) score ≥22 = DSM-IV criteria for major depressive disorder (moderate or severe)
Response to treatment = ≥50% reduction in HAMD score.
Intention-to-treat population: ITT

London New Drugs Group APC/DTC Briefing


Page 11
Short term studies with active controls
In the following studies, fluoxetine and paroxetine were used as the compared to placebo. No direct comparisons between agomelatine
active controls to validate the studies. The main aim of the studies and fluoxetine or paroxetine were performed. Data from these studies
was to confirm the efficacy of the target dose of agomelatine comes from the EMEA CHMP Public Assessment Report.6

June 2009
Agomelatine

Study Primary
Treatment Results Comments
design endpoint
• HAMD final score: 14.5±8.2 agomelatine vs. 15.9±
8.6 placebo (not statistically significantly different,
NS). • After six weeks of treatment,
• HAMD final score in fluoxetine group was 2.59 points agomelatine 25mg was not
lower than in the placebo group (p=0.008). significantly better than pla-
Randomised, Agomelatine
cebo in treating depression
double-blind 25mg (n=133) • Responders: 53% agomelatine vs. 47% placebo
but fluoxetine was signifi-
6 week study or fluoxetine (p=NS); 58% fluoxetine vs. 47% placebo (p=NS).
cantly better than placebo.
in adults with 20mg (n=137)
HAMD score or placebo • At the end of the 18 week
• After 18 week extension:
≥22.6 (n=149) extension the rate of relapse
• HAMD: 10.0, agomelatine, 10.7 placebo, 8.4 fluoxet- was significantly lower with
ine. agomelatine (and fluoxetine)
HAMD score • Relapses: 14.3% agomelatine, 33.3% placebo, 17.8% than with placebo.
fluoxetine, (p=0.017 agomelatine vs. placebo;
Secondary p=0.045 fluoxetine vs. placebo).
variables in- • HAMD final score: 13.0±8.0 agomelatine vs. 13.8±8.0 • Agomelatine was not more
cluded CGI, placebo (p=NS). efficacious than placebo after
Randomised, Agomelatine Hamilton Anxi- • HAMD final score in paroxetine group 12.2±8.1. p= six and 18 weeks of treat-
double-blind 25mg (n=142) ety Rating NS vs. placebo. ment.
6 week study or paroxetine Scale, Leeds • Response to treatment and time to remission. • Paroxetine (active control)
in adults with 20mg (n=138) Sleep Evalua- was also not significantly bet-
HAMD score or placebo tion Question- • After 18 week extension: ter than placebo at either
≥22.6 (n=137). naire. • No statistically significant differences in the rates of time point.
relapse and loss of first response when comparing • No discernable results can be
agomelatine or paroxetine with placebo. drawn from the study
• HAMD final score: 12.0±8.2 agomelatine vs. 13.4±8.4 • Agomelatine was not more
Agomelatine placebo (p=NS). efficacious than placebo af-
Randomised, 25mg (n=150)
• HAMD final in fluoxetine group was 0.53 lower than in ter six and 18 weeks of
double-blind or 50mg
the placebo group (p=NS). treatment.
6 week study (n=151) or
• Fluoxetine was also not sig-
in adults with fluoxetine
• After 18 week extension: nificantly better than pla-
HAMD score 20mg (n=148)
• No statistically significant differences in the rates of cebo at either time point.
≥22.6 or placebo

THIS IS AN NHS DOCUMENT NOT TO BE USED FOR COMMERCIAL AND MARKETING PURPOSES.
(n=158). relapse and loss of first response when comparing • No discernable results can

PRODUCED TO INFORM LOCAL DECISION-MAKING USING THE BEST AVAILABLE EVIDENCE AT THE TIME OF PUBLICATION.
agomelatine or fluoxetine with placebo. be drawn from the study
• The option to increase the agomelatine dose from 25mg to 50mg was not included in these studies so some patients may not have been
adequately treated. A meta-analysis of agomelatine trials shows that 23.5% of patients require a 50mg dose, which is a reason why agome-
latine failed to differentiate from placebo.
• About 50-60% of placebo-controlled efficacy studies fail to show a superior efficacy of the antidepressant used to validate the study popula-
tion. One reason for this is the inclusion of patients with insufficient severity, making it difficult to demonstrate a statistically significant

London New Drugs Group APC/DTC Briefing


Page 12

treatment effect compared to placebo.


is the appearance of symptoms in a new episode.17 Relapse indicates
Long term/prevention of relapse studies
that the treatment duration was too short.
There are two relapse prevention studies, undertaken to demonstrate One failed to demonstrate a difference in the time to relapse.23 An ad-
the efficacy of agomelatine in the prevention of relapse after an initial ditional relapse prevention study was finalised after the negative CHMP
opinion in 2006.

June 2009
response. Relapse is classified as the appearance of the symptoms of
the index episode soon after medication is stopped, whilst recurrence
Agomelatine

Primary
Study design Treatment Results Comments
endpoint
HAMD total • Relapse rates: 25.9%, agome-
Agomelatine score W0 latine vs. 23.5%, placebo, p = • No relapse-preventing effect was demonstrated.
25mg for 8 to W8, time not significant. No statistically significant difference in HAMD total
8 week open weeks then, in to relapse • scores between the two groups was seen.
label followed patients with after week • Relapse rates in a post-hoc • Similar results in both groups for the secondary
by a 26 weeks HAMD ≤10, 8 defined analysis (pts with more severe endpoints, measuring severity of illness and im-
double-blind, agomelatine as HAMD depression,46% of the total trial provements in the condition, were seen.
randomised 25mg (n=187) ≥16 or sui- population): 21.3%, agome- • The relapse rate in the placebo arm was unexpect-
phase.6 or placebo cide/suicide latine vs. 31.3% placebo. The edly low and may reflect some of the methodologi-
(n=180) for attempt difference became statistically cal aspects, such as the broad range of the severity
26 weeks. (ITT popu- significant at week 52 of depression of patients included in the study.
lation) (p=0.046).
Agomelatine
25-50mg for • At 24 weeks: • Baseline characteristics of the two treatment
8-10 weeks • Relapse rates: 21.7%, agome- groups were similar. The patients were on aver-
then, in pa- latine vs. 46.6%, placebo, age 43.3 years old, with ~3.6 prior depressive epi-
tients with p<0.0001. sodes, and over 70% were female. The mean
HAMD ≤10, • Risk of relapse reduced by 54% HAMD17 total score at the beginning of the open-
8-10 week Relapse, (hazard ratio 0.458, 95% CI
agomelatine label period was 27 and the mean GCI-S score was
open label defined as 0.305 to 0.60).
25mg 4.9; by the start of the double-blind phase these
study followed HAMD ≥16
(n=141), • Depressive relapse: 20.6%, had fallen to 6 and 1.8 respectively, indicating that
by a 24 week or any
50mg (n=24) agomelatine vs. 41.4%, pla- 8-10 weeks of agomelatine open-label treatment
double-blind, withdrawal
or placebo cebo. had improved depressive symptoms.
randomised, for lack of
(n=174) for • Cumulative relapses: 22.7%, • 390 patients of the 492 originally enrolled entered
phase. Adults efficacy or
24 weeks. agomelatine vs. 50.4%, pla- the 24-week double-blind phase.
enrolled if they suicide/
had HAMD suicide at- cebo, p<0.0001. • Three patients discontinued due to lack of efficacy,
190 pts • all in the agomelatine group. Adverse events were
≥22 and CGI- tempt (ITT
(agomelatine • After a further 20 weeks24 the reason for discontinuation in 4 patients on
S ≥4.10;11 population)
106 and pla- agomelatine compared with one taking placebo.
• Relapse: 23.9%, agomelatine
cebo 84) con- The most commonly reported adverse events were
vs. 49.9%, placebo, p<0.0001.

THIS IS AN NHS DOCUMENT NOT TO BE USED FOR COMMERCIAL AND MARKETING PURPOSES.
tinued with a headache (10.3% in the agomelatine group and
In the more severely depressed

PRODUCED TO INFORM LOCAL DECISION-MAKING USING THE BEST AVAILABLE EVIDENCE AT THE TIME OF PUBLICATION.
further 20 7.5% in the placebo group), nasopharyngitis (6%
patients, this risk was reduced
weeks of vs. 9.8%) and back pain (6.1% vs. 3.4%).
by 58.6% (p<0.0001).
treatment.
Hamilton Depression Rating Scale (HAMD) score ≥22 = DSM-IV criteria for major depressive disorder (moderate or severe).
There is one fundamental difference between the two trials. In the second trial, a larger proportion of patients had severe depression (80%
vs. 46%). The results of the second trial were accepted by the CHMP as a demonstration of maintenance of efficacy, probably because of

London New Drugs Group APC/DTC Briefing


Page 13

this larger pool of severely depressed patients.6


Absence of discontinuation symptoms

June 2009
Agomelatine

All antidepressants have the potential to cause withdrawal symp- Symptoms include nausea, vomiting, headaches, loss of appetite,
toms. The Maudsley Prescribing Guidelines recommend that when agitation, sweating, worsening of symptoms and weakness. These
one has been taken for 6 weeks or longer, it should not be usually occur within the first few days of discontinuing treatment
stopped abruptly unless a serious adverse event has occured.18 and are generally mild to moderate, but in some patients may be
severe.26

Primary
Study design Treatment Results Comments
endpoint

• Patients had a MADRS score of 18-27,


which indicates moderate to severe de-
• No statistically significant pression.
Agomelatine difference in the number of • Abrupt discontinuation of agomelatine did
Double-blind 12 25mg emergent discontinuation not cause any discontinuation symptoms.
week study in pa- (n=167) or symptoms was seen be- • It should be noted that after 12 weeks of
tients with moderate paroxetine tween patients discontinu- treatment it is recommended that the
depression to assess 20mg Total num- ing agomelatine and those paroxetine dose is reduced over a number
the effects of abrupt (n=168). ber of continuing with agome- of weeks, or even months and not
discontinuation of Paroxetine emergent latine. stopped abruptly, in order to avoid dis-
agomelatine. Pts was used to DESS • After the first 7-day discon- continuation symptoms.19;27 These re-
with sustained re- validate the symptoms tinuation period, a statisti- sults support the fact that abrupt with-
mission re- study design. in first and cally significant difference in drawal of paroxetine is associated with a
randomised to either second 7- the number of emergent significant increase in discontinuation
placebo (to assess Agomelatine: day periods discontinuation symptoms symptoms. The main symptoms reported
discontinuation 88 re- of the dis- was seen between patients more frequently in patients discontinuing
symptoms) or active randomised, continuatio discontinuing paroxetine paroxetine were insomnia, dreaming,
(to act as control) 27 to placebo. n period. (7.3±7.1) and those con- muscle aches, dizziness, runny nose,
for a further 2 weeks Paroxetine: tinuing with paroxetine chills, nausea and diarrhoea (p<0.05).
(‘discontinuation 104 re- (3.5±4.1), p<0.001. No • The study was not designed to investigate
period’).28 randomised, statistically significant dif- the persistence of the discontinuation
43 to placebo. ference was seen after the symptoms beyond 2 weeks. Discontinua-
second period. tion symptoms would be expected to ap-
pear during the first week and decline
during the second week.

THIS IS AN NHS DOCUMENT NOT TO BE USED FOR COMMERCIAL AND MARKETING PURPOSES.
PRODUCED TO INFORM LOCAL DECISION-MAKING USING THE BEST AVAILABLE EVIDENCE AT THE TIME OF PUBLICATION.
DESS: Discontinuation Emergent Signs and Symptoms check list, a clinician rated instrument which covers 43 signs and symptoms
associated with discontinuation or interruption of antidepressant treatment. The limitation to the DESS is that some symptoms could
occur even if no discontinuation has taken place.

London New Drugs Group APC/DTC Briefing


Page 14
Sexual dysfunction

June 2009
Sexual dysfunction is a side effect that has been attributed to antide- therapy. The effects of dopamine and noradrenaline on sexual func-
Agomelatine

pressant use, and can be problematic, especially leading to problems tioning could explain the diminished feelings of desire and arousal
with long-term treatment compliance.29 About one third of patients experienced by patients with depression. Dopamine influences de-
(36%) find antidepressant-induced sexual dysfunction to be an unac- sire and motivates sexual behaviour, whilst noradrenaline stimulates
ceptable side effect, constituting grounds for discontinuation of arousal. Serotonin systems have a negative effect on arousal.27

Primary
Study design Treatment Results Comments
endpoint
Percentage of SAR population
with: • The differences were maintained when the re-
• Deterioration in total scores: sults were analysed separately for men and
7.3% agomelatine vs. 15.7% women (but note the small number of men in the
venlafaxine. trial). Treatment-emergent sexual dysfunction
• Reduced sex drive/desire: was significantly less prevalent among sexually-
Percentage
3.6% agomelatine vs. 19.4% active patients who received agomelatine com-
Double-blind, of patients
venlafaxine, p=0.007. pared with venlafaxine. The main differences
randomised in the SAR
were in reduced desire and orgasm, which re-
12 weeks group re- • Deterioration in orgasm in
flects the pharmacological differences between
study com- porting a women: 4.3% agomelatine
the two treatments. Venlafaxine inhibits
paring the Agome- deteriora- vs. 21.2% venlafaxine,
noradrenaline, dopamine and serotonin, whilst
sexual side latine tion of at p<0.0001.
agomelatine is a melatonin agonist and also in-
effects of 50mg least one • Trend towards more favour-
creases noradrenaline and dopamine. It should
agomelatine (n=137) or point from able scores with agomelatine
be noted that, as for other antidepressant trials,
and venla- venlafaxine baseline in with other domain and total
no sexually related adverse events were re-
faxine in 150mg the Sex FX scores.
corded, but instead they were reported via indi-
adults with (n=140). scale.
rect questioning.
moderate de- Percentage of SA-FS population
•38.1% of patients on venlafaxine had adverse
pression Secondary with deterioration in:
events vs. 20.4% on agomelatine. Nausea, head-
(MADRS endpoint: • Desire: 6% agomelatine vs.
ache and upper respiratory tract infections were
≥20).13 changes in 16.4% venlafaxine.
the more common in both groups. More patients
depression. • Orgasm: 9.1% agomelatine discontinued venlafaxine (8.6%, mainly because
vs. 18.5% venlafaxine of gastrointestinal or central nervous system side
• Total Sex FX score: 8.2% effects) than agomelatine (2.2%).
agomelatine vs. 15.2% venla- •Seen Short Term studies section for data on de-
faxine. pression.

THIS IS AN NHS DOCUMENT NOT TO BE USED FOR COMMERCIAL AND MARKETING PURPOSES.
• P<0.0001 for all.

PRODUCED TO INFORM LOCAL DECISION-MAKING USING THE BEST AVAILABLE EVIDENCE AT THE TIME OF PUBLICATION.
Assessments used: Sex FX scale to assess sexual function (the first 11 out of the 13 item scale address three domains: desire, arousal
and orgasm, whilst the last two items provide a measure of global satisfaction). MADRS, CGI-S and CGI-I scales to assess depression.
SAR: sexually active remitters group: agomelatine n=60 (14 male), venlafaxine n=51 (18 male). Patients in this group had a decrease
at week 10 of 50% or more in sexual activity from baseline and a f
inal MADRS score of ≤12.
SA-FS : sexually active full set – all subjects sexually active at baseline; agomelatine n=103, venlafaxine n=90.

London New Drugs Group APC/DTC Briefing


Page 15
Study Primary
Treatment Results Comments
design endpoint

June 2009
Sexual Acceptability Set (n=87):
Agomelatine

• A SD according to PRSEXDQ-SALSEX at each • This was a phase 2 study in


visit from week 2 to 8: 22.7% agomelatine healthy volunteers, not patients
25mg; 4.8% agomelatine 50mg; 85.7% par- with major depressive episodes.
oxetine, p<0.0001 for each comparison. • The risk of sexual dysfunction was
(8.7% placebo) significantly lower with agome-
• Adjusted risk of SD vs paroxetine: 74% re- latine use than with paroxetine
duction with agomelatine 25mg (RR = 0.26, use. The difference was seen as
Phase 1 ran- 95% CI 0.12, 0.58) and 94% reduction with early as week 2.
domised, agomelatine 50mg (RR=0.06%, 95% CI • The risk of sustained sexual dys-
Sexual dys- 0.01, 0.38).
double-blind, function was significantly higher
function as-
placebo- • Moderate or severe SD: 1 volunteer (4.5%) with paroxetine than with agome-
sessed by
controlled, 4- agomelatine 25mg, 1 volunteer (4.8%) latine.
PRSEXDQ-
group paral- Agomelatine agomelatine 50mg and 13 volunteers • The greatest difference in favour of
SALSEX.
lel-design 8- 25mg or (61.9%) paroxetine, p<0.001, for each com- agomelatine in PRSEXDQ-SALSEX
week study 50mg, parison. No volunteer in the placebo group scores was for delayed orgasm/
Secondary
in healthy paroxetine affected. ejaculation: 9.1% and 4.8% for
endpoints
male volun- 20mg or • Adjusted risk of moderate or severe SD: 93% agomelatine 25mg and 50mg vs.
included
teers (18-30 placebo reduction with agomelatine 25mg (RR=0.07, 81% paroxetine (p<0.0001).
moderate or
years) to (n=23 in 95% CI 0.01, 0.51, p<0.0001), 92% reduc- • Erectile function decreased in the
severe sex-
compare the each group). tion with agomelatine 50mg (RR=0.08, 95% paroxetine but not the agomelatine
ual dysfunc-
effects of CI 0.01, 0.54, p=0.0001). groups.
tion and
agomelatine • Sustained SD: 22.7% agomelatine 25mg, • The most common side effect with
erectile
and paroxet- 4.8% agomelatine 50mg, 81.0% paroxetine agomelatine was somnolence, oc-
function.
ine on sexual and 4.3% placebo. curring in twice as many patients
function.18 • PRSEXDQ-SALSEX score: 0.9±0.2, agome- in each agomelatine group than in
latine 25mg, 0.2±0.9 agomelatine 50mg, the paroxetine group.
5.2±3.6 paroxetine, p<0.0001 for both com- • Although the study was not de-
parisons. (0.5±1.2, placebo). signed to directly compare agome-
• PRSEXDQ-SALSEX individual items: for all latine with placebo, the results
items the results were statistically signifi- suggest that sexual dysfunction
cantly in favour of both agomelatine doses caused by agomelatine was com-
compared with paroxetine, except for parable to that caused by placebo.
‘decreased libido’ agomelatine 25mg.

THIS IS AN NHS DOCUMENT NOT TO BE USED FOR COMMERCIAL AND MARKETING PURPOSES.
PRODUCED TO INFORM LOCAL DECISION-MAKING USING THE BEST AVAILABLE EVIDENCE AT THE TIME OF PUBLICATION.
PRSEXDQ-SALSEX: Psychotrophic-related sexual dysfunction questionnaire. 7 items on sexual function. Sexual dysfunction (SD) was de-
fined as at least one impairment in one of the following four items: decreased libido, delayed orgasm/ejaculation, anorgasmia/no ejacula-
tion (score of ≥1 for each of these three) and erectile dysfunction (score of ≥2). Score 0 = no sexual dysfunction, 15 = maximum sexual
dysfunction. For each item 0=nil or never, 3 = severe or always.
Sexual Acceptability Set: all volunteers having a sexual dysfunction assessment after at least 2 weeks and compliant to the treatment.

London New Drugs Group APC/DTC Briefing


Page 16
This whole cycle is non-rapid eye movement (NREM) sleep, making up 75-85%
Sleep disorders of a normal nights sleep. Each cycle is followed by a discrete episode of REM
sleep, during which time EEG activity quickens and rapid horizontal eye move-
One of the major complaints of patients suffering with depression is poor ments are detected by the electro-oculogram (EOG). There are usually four or
sleep.16 Sleep neurophysiology can be studied using all-night electroencepha- five REM periods across a normal nights sleep, separated by 90-minute inter-
lograms (EEG). Healthy sleep is divided into four stages: stage 1 is the tran- vals.29 Patients with depression can suffer from poor sleep efficiency, de-
sition from drowsiness to light sleep, shown by a slowing of EEG activity; creased slow-wave sleep (SWS), reduced REM latency and increased REM ac-

June 2009
sleep spindles and K-complex waves appear during the onset of the deeper tivity.29
Agomelatine

sleep of stage 2 and stages 3 and 4 (deepest sleep) are defined by desyn-
chronised slow (delta) waves.29

Study Primary
Treatment Results Comments
design endpoint
• Total sleep time increased from 363±60 • At baseline 11 patients reported difficulties in becoming
Open-label minutes at baseline to 392±95 minutes at wide awake in the morning often or very often, and 13
study in day 42 (not statistically significant). patients complained of daytime sleepiness often or very
adults to often. By day 42 these problems were reported in two
• Time to both sleep onset and REM onset
assess the patients.
improved significantly by day 7.
effects of
• No significant changes from baseline in stage • This small study shows that agomelatine can increase
agomelatine
on sleep. 1 or 2 sleep, any aspect of the REM sleep, sleep efficiency and slow-wave sleep, as shown by the
Patients with Sleep cycle the time of sleep onset and the delta power lengthening of stage 3-4 sleep and does not suppress
Agome- ratio (indicating improvement in NREM) by REM sleep, which can be poor in patients with depression,
sleep ap- analysed by
latine day 42. Trend towards an improvement in stage 1 or 2 sleep and time to sleep onset. Dose-
noea, shift poly-
25mg REM latency (uncorrected value) by day 42. dependent suppression of REM sleep has been seen with
workers and somnogra-
(n=15) for tricyclic antidepressants, SSRIs and venlafaxine. It has
post- phy, on six • Statistically significant improvements were
42 days. been thought that the increase in the delta power ratio,
menopausal nights. seen at day 42 in the intra-sleep awakening
women were time, decreasing from 48±31 minutes to seen with e.g. sertraline and amitriptyline, could be due
excluded. 29±27 minutes; the sleep efficiency (total to their REM suppression effect, leading to a delay in the
Patients sleep time: total sleep period ratio), increas- first REM sleep, but this is not the case for agomelatine.
scored at ing from 88% to 93% and stages 3-4 slow- • The authors of this study state that agomelatine improves
least 20 on wave sleep increasing from 15.9% of the the deficiency of the homeostatic system of sleep regula-
the HAMD total sleep period to 19.4%, a mean increase tion, by improving slow wave sleep and activity. Further
scale.16 of 14 minutes. SWS started to improve as controlled studies investigating sleep and daytime alert-
early as day 7. ness would be warranted.
• Data comes from a conference abstract.
• In animal models agomelatine has been shown to resyn-
chronise altered circadian rhythms.
Relative am- • Relative amplitude is based on average activity level dur-
Randomised
plitude (RA) ing the 10 most active and 5 least active hours of a 24
6 week
of rest/ hour period.
study as-
activity cycle, • The mean RA was stable between days 0-7 in the agome-
sessing the Agome- • Sleep latency (how quickly to fall asleep),
sleep latency, latine group and decreased in the sertraline group. The
rest-activity latine 25- changes from baseline: 22.5 to 18.9 min-
sleep effi- mean RA increased in the sertraline group from day 7,
circadian 50mg utes, agomelatine (p<0.0001), vs. 23.48 to
ciency. and from day 14 there was no significant difference be-
rhythm us- (n=154) or 27.75 minutes, sertraline.
ing wrist sertraline tween the groups.
Sleep efficiency (total sleep time: total sleep

THIS IS AN NHS DOCUMENT NOT TO BE USED FOR COMMERCIAL AND MARKETING PURPOSES.
Secondary •
actigraphy in 50-100mg period): 78.85% agomelatine (p<0.0001) • Sleep efficiency increased at each assessment point with

PRODUCED TO INFORM LOCAL DECISION-MAKING USING THE BEST AVAILABLE EVIDENCE AT THE TIME OF PUBLICATION.
endpoint:
patients with (n=159). vs. 76.70% sertraline. agomelatine, yet with sertraline it alternated between
change in
a HAMD decreasing and increasing: a longer study may show no
HAMD score.
score difference between the two treatments. Improvements in
(see Short
≥22.14;15 both measurements were seen as early as after one week
term studies)
of treatment with agomelatine.
• Most commonly reported adverse events were dry mouth,
headache, diarrhoea and fatigue. More patients discon-

London New Drugs Group APC/DTC Briefing


Page 17

tinued sertraline (11.3%) than agomelatine (2.6%).


Study design and primary end-
Results Comments
points

June 2009
• The comparison between agomelatine and venlafaxine on
Agomelatine

their effects on sleep was performed only if no significant


difference between the two drugs with respect to the final
• HAMD score: 9.9 ± 6.6, HAMD scores was seen. Full results are in Appendix 2.
Randomised, double-blind 6 week agomelatine vs. 11.0 ± 7.4 • No comment was made by the investigators on the
study compared the effects of venlafaxine. changes from week 1 to the last visit for within-group
agomelatine and venlafaxine on sub- • Response rates: 76.4% scores. Similar improvements were seen for quality of
ject sleep variables. Patients had agomelatine vs. 70.6% venla- sleep score, sleep awakening scores improved to a slightly
HAMD score ≥20.12 faxine. better extent with agomelatine and integrity of behaviour
• LSEQ getting to sleep score: scores (assessing whether patients feel more alert and less
70.5±16.8, agomelatine vs. clumsy after waking up) improved by a greater extent with
Patients were treated with : Agome- 64.1±18.2, p=0.001 for the venlafaxine.
latine 25mg (n=165) or venlafaxine between group difference of • The differences in the HAMD insomnia scores were very
75mg (n=167). 6.36mm. small and may not be reproduced in a larger study.
• No significant difference be- • Interpretation of the sleep diary data is difficult because
Doses were increased at 2 weeks in tween the groups in whether many omissions occurred and the rating by the patient can
non-improved patients Doses were the patients felt more or less be contaminated by cognitive distortion seen in patients
increased in 23 and 17 patients re- drowsy. with depression. Nocturnal awakenings do not consider
spectively. • LSEQ quality of sleep, sleep the ‘micro-awakenings’ that occur frequently in the night in
awakening and integrity of depressed patients that are responsible for a poor quality
Patients could continue with an 18- behaviour scores: between- of sleep.
week extension.6 group differences were statis- • Both agomelatine and venlafaxine improved depression,
tically significant in favour of sleep, daytime alertness and well-being after six weeks of
The Primary Endpoints were: agomelatine. treatment in this patient group. ‘Getting to sleep’ scores
• HAMD insomnia score (max 6 were higher with agomelatine but the improvements were
LSEQ – ‘getting to sleep’ score. from 3 items): decreased by seen to the same extent with venlafaxine. The venlafaxine
Visual analogue scale for daytime similar amount in both dose was not titrated above 150mg/day, so it is possible
sleepiness (ITT population) groups. Difference in final that some patients were not adequately treated. The au-
score statistically significant thors state that any such efficacy benefit might be at the
Pittsburgh Sleep Quality Index ques- and due to difference in fa- expense of poorer tolerability: three times as many pa-
tionnaire vour of agomelatine for early tients withdrew because of venlafaxine side effects (n=22
and middle insomnia. vs. 7). Increasing the venlafaxine dose could have in-
Sleep diary • Sleep diary: improved sleep creased the drop-out rate and biased the results.
latency and reduced number • Longer term studies are needed to determine whether the
HAMD-17 and CGI effects of agomelatine on sleep are sustained.
of awakenings, in both treat-
ment groups. • After an additional 18 weeks, remission rates were higher

THIS IS AN NHS DOCUMENT NOT TO BE USED FOR COMMERCIAL AND MARKETING PURPOSES.
with agomelatine (71.6%) compared with venlafaxine

PRODUCED TO INFORM LOCAL DECISION-MAKING USING THE BEST AVAILABLE EVIDENCE AT THE TIME OF PUBLICATION.
(66.7%). Bias from the initial focus on subjective sleep
parameters in favour of agomelatine cannot be ruled out.6

LSEQ: Leeds Sleep Evaluation Questionnaire, assessed using a visual analogue scale. These are very subjective and are best for compar-
ing changes within individuals, rather than between individuals. Clinical relevance is a 10mm or greater change from baseline. Note all

London New Drugs Group APC/DTC Briefing


Page 18

patients start with a baseline rating of 50 on the LSEQ scale – from this start point improvement or worsening of symptoms is determined.
Agomelatine Page 19

Figure 1: Rating scales

Hamilton Rating Scale for Depression (HAMD)21


• Rates severity of depression.
• Items: Depressed mood; Feelings of guilt; Suicide; Insomnia: Early, Middle and Late;
Work and Activities; Retardation, psychomotor; Agitation; Anxiety (psychological); Anxi-
ety (somatic), Somatic symptoms (gastrointestinal); Somatic symptoms (general); Geni-
tal symptoms; Hypochondriasis; Loss of weight; Insight; Diurnal variation, Depersonal-
isation and derealisation; Paranoid symptoms; Obsessional and compulsive symptoms.
• Scale: 0 (absent) to 4 (most severe symptoms). Maximum score: 66. A score of >18
indicates moderate to severe depression.22

Clinical Global Impression (CGI)23


• Three-item scale to assess treatment response.
• Severity of illness, at the time of assessment: 7 point scale, 1 = normal, 7 = extremely
ill
• Global Improvement – the clinician rates how the patients illness has improved or wors-
ened relative to a baseline state: 7 point scale: 1 = very much improved, 7 = very much
worse Efficacy Index: 4 point scale: none to outweighs therapeutic effect

Montgomery-Asberg Depression Rating Scale (MADRS)24


• Rates severity of depression. Provides an instrument sensitive to minor changes.
• Items include: Apparent sadness; Reported sadness; Inner tension; Reduced sleep; Sui-
cidal thoughts.
• Scale: 0 (absent) to 6 (most severe symptoms).
• Maximum score: 60. Suggested cut-off points: 0-6 (absence of symptoms), 7-19 (mild
depression), 20-34 (moderate depression), 35-60 (severe depression).

Figure 2: Sleep assessments12

Leeds Sleep Evaluation Questionnaire


• Quantifies subjective assessments of the effects of drugs on sleep and early morning behav-
iour.
• 10 items grouped into 4 scores, each item quantified by a 100mm visual analogue scale
(VAS). A high score indicates improvement.
• 4 scores evaluate: ease of getting to sleep (3 VAS items), perceived quality of sleep (2 VAS
items), ease of awakening (2 VAS items) and integrity of behaviour after awakening (3 VAS
items).
• There is no baseline evaluation as the LSEQ assesses the changes during treatment relative
to those before treatment.

Visual analogue scale


• For daytime sleepiness and feeling well, completed at each visit.

Pittsburgh Sleep Quality Index questionnaire


• Completed at inclusion to check the comparability of the treatment groups regarding their
status.

Sleep diary
• Completed every morning to the week 3 visit recording: times of light off, wake-up and get-
ting up, sleep onset latency and number of nocturnal awakenings.

HAMD-17 and CGI-I


• For assessing the efficacy of study medication on depression. Responders showed a decrease
of ≥50% in HAMD score relative to baseline.

THIS IS AN NHS DOCUMENT NOT TO BE USED FOR COMMERCIAL AND MARKETING PURPOSES.
PRODUCED TO INFORM LOCAL DECISION-MAKING USING THE BEST AVAILABLE EVIDENCE AT THE TIME OF PUBLICATION.

June 2009 London New Drugs Group APC/DTC Briefing


Agomelatine Page 20

Cost per 100,000 population31

As already stated, NICE recommends an SSRI as first line treatment for moderate to severe depression.
The budget impact model has been based on agomelatine use in second-line or later treatment for de-
pression, with agomelatine taking a 10% share of the current patients who fall into this category. Below
is a breakdown of the cost per 100,000 population if agomelatine is used as a second-line or later treat-
ment for depression, for year 1.

Population: 100,000 Year 1


Incidence of specified second-line or later treatment (all
0.40%, n=397
treatments)
Average number of days on therapy 179
Number of packs 1,918

Second-line or later patients are prescribed:


Venlafaxine 24.6%, n=98
Zispin® 18.6%, n=74
Mirtazapine 29.2%, n=116
Escitalopram 19.5%, n=77
Duloxetine 8.1%, n=32

If agomelatine takes 10% of each second line treatment N = 10 + 7 + 12 + 8 + 3 = 40


Agomelatine price £38.53 per 28 days treatment
Average daily dose 30.9mg
Agomelatine patients in year 1 40 (based on 10% market share)
Agomelatine prescriptions in year 1 192
Cost of agomelatine prescribing £9,124
Cost of agomelatine prescribing including liver function
£9,464
tests

For years 2 and 3, the number of patients treated with agomelatine is expected to rise.

Year 1 Year 2 Year 3

Patients starting agomelatine 40 50 (+25%) 62 (+25%)

Patients taking agomelatine during period 40 69 86

Total cost of prescribing without Agomelatine £381,644 £381,644 £381,644

Plus cost of agomelatine £9,124 £14,379 £17,973

Less reduction in prescribing of other agents £3,572 £5,629 £7,036

Total cost of prescribing with Agomelatine £387,196 £390,394 £392,581

Net increase in prescribing costs £5,552 £8,750 £10,937


Net increase in prescribing costs including liver func-
£5,892 £9,285 £11,607
tion tests
Percentage increase in spend 1.45% 2.29% 2.87%

THIS IS AN NHS DOCUMENT NOT TO BE USED FOR COMMERCIAL AND MARKETING PURPOSES.
PRODUCED TO INFORM LOCAL DECISION-MAKING USING THE BEST AVAILABLE EVIDENCE AT THE TIME OF PUBLICATION.

June 2009 London New Drugs Group APC/DTC Briefing


Agomelatine Page 21

Current costs of antidepressant drugs:

Drug* Dose Cost (Drug Tariff April 2009)32

To be confirmed, anticipated at £30


Agomelatine 25mg, increased to 50mg daily
per 28 days treatment.

75mg, increased to 150-200mg 28x25mg: £0.94


Amitriptyline‡
daily 28x50mg: £1.13

28x20mg: £1.26
Citalopram‡ 20mg, increased to 60mg daily
28x40mg: £1.44

Duloxetine† 60mg daily 28x60mg: £27.72 (Cymbalta)


28x10mg: £14.91 (Cipralex)
Escitalopram‡ 10mg, increased to 20mg daily
28x20mg: £25.20 (Cipralex)

30x20mg: £1.02
Fluoxetine‡ 20mg, increased to 60mg daily
30x60mg: £55.73

50-100mg, increased to 300mg 60x50mg: £6.17


Fluvoxamine‡
daily 30x100mg: £8.22

28x20mg: £2.84
Paroxetine‡ 20mg, increased to 50mg daily
28x30mg: £6.21

28x50mg: £1.36
Sertraline‡ 50mg, increased to 100mg daily
28x100mg: £1.67

56x37.5mg tabs: £21.07 (Efexor)


56x75mg tabs: £35.12 (Efexor)
75mg daily increased to 150mg
Venlafaxine† 28x75mg MR caps: £22.50 (Efexor XL)
daily, to max 375mg daily.
28x150mg MR caps: £37.51 (Efexor
XL)

*generic unless otherwise stated


‡ selective serotonin re-uptake inhibitor (SSRI)
† serotonin and noradrenaline re-uptake inhibitor, SNRI

Reference List

(1) Anon. Servier's Valdoxan receives EU approval for depression. Scrip 2009; No. 3534:20.
(2) Summary of Product Characteristics. Valdoxan 25mg film coated tablets. Servier Laboratories
Limited
(3) Clinical Guideline 23.Depression (amended 27/04/2007). NICE Accessed via: http://
www.nice.org.uk/Guidance/CG23#documents, on 16/12/08.
(4) Depression: the treatment and management of depression in adults. NICE Guidance. Draft for
consultation, February 2009. NICE Accessed via http://www.nice.org.uk/nicemedia/pdf/
DepressionUpdateNICEGuidelineDraft.pdf on 31/03/09.
(5) Howland R. Agomelatine. A novel atypical antidepressant. J Psych Nurse 2007; 45(12):13-17.
(6) CHMP Assessment report for Valdoxan. Committee for Medicinal Products for Human Use
(CHMP) Accessed via: http://www.emea.europa.eu/humandocs/PDFs/EPAR/valdoxan/H-915-
en6.pdf on 18/03/09.
(7) Loo H, Hale A, D'haenen H. Determination of the dose of agomelatine, a melatoninergic agonist
and selective 5-HT2c antagonist, in the treatment of major depressive disorder: a placebo-
controlled dose range study. Int Clin Psychopharmacol 2002; 17(5):239-247.

THIS IS AN NHS DOCUMENT NOT TO BE USED FOR COMMERCIAL AND MARKETING PURPOSES.
PRODUCED TO INFORM LOCAL DECISION-MAKING USING THE BEST AVAILABLE EVIDENCE AT THE TIME OF PUBLICATION.

June 2009 London New Drugs Group APC/DTC Briefing


Agomelatine Page 22

(8) Kennedy SH, Emsley R. Placebo-controlled trial of agomelatine in the treatment of major de-
pressive disorder. Eur Neuropsychopharmacol 2006; 16(2):93-100.
(9) Olie JP, Kasper S. Efficacy of agomelatine, a MT1/MT2 receptor agonist with 5-HT2c antago-
nistic properties in major depressive disorder. International Journal of Neuropsychopharma-
cology 2007; 10(5):661-673.
(10) Goodwin G, Rouillon F, Emsley R Long-term efficacy of agomelatine, a novel antidepressant,
in the prevention of relapse in out-patients with major depressive disorder. Poster presenta-
tion. European College of Neuropsychopharmacology, 21st ECNP Congress. Barcelona, Spain.
30 August - 3 September: 2008
(11) Goodwin GM, Rouillon F, Emsley R Long-term treatment with agomelatine: prevention of re-
lapse in patients with major depressive disorder over 10 months. Poster presentation. Euro-
pean College of Neuropsychopharmacology. 21st ECNP Congress. Barcelona, Spain. 30 Au-
gust - 3 September: 2008
(12) Lemoine P, Guilleminault C, Alvarez E. Improvement in subjective sleep in major depressive
disorder with a novel antidepressant, agomelatine: randomized, double-blind comparison with
venlafaxine. J Clin Psychiatry 2007; 68:1723-1732.
(13) Kennedy SH, Rizvi S, Fulton K et al. A double-blind comparison of sexual functioning, antide-
pressant efficacy and tolerability between agomelatine and venlafaxine XR. J Clin Psychophar-
macol 2008; 28:329-333.
(14) Kasper S, Laigle L, Bayle F Superior antidepressant efficacy of agomelatine vs sertraline: a
randomized, double-blind study. Poster presentation. 21st ECNP Congress, Barcelona, Spain.
30 August - 3 September.: 2008
(15) Kasper S, Wulff K, Laigle L et al. Effect of agomelatine on rest-activity cycle in patients with
major depressive disorder compared to sertraline. 21st ECNP Congress, Barcelona, Spain. 30
August - 3 September.: 2008
(16) Quera Salva M-A, Vanier B, Laredo J et al. Major depressive disorder, sleep EEG and agome-
latine: an open-label study. International Journal of Neuropsychopharmacology 2007; 10
(5):691-696.
(17) Data on file: 045 study Agomelatine vs. fluoxetine: Difference in HAM-D17 change from base-
line to week 8. Reference: 09MKA0147.. Presented at the World Psychiatric Assocation Con-
ference, Florence, Italy. April 2009. Servier Laboratories Limited
(18) Montejo AL, Preito N, Terleira A et al. Better sexual acceptability of agomelatine (25 and
50mg) compared with paroxetine (20mg) in healthy male volunteers. An 8-week, placebo-
controlled study using the PRSEXDQ-SALSEX scale. J Psychopharmacol (Oxf) 2008; 00(00):1-
10.
(19) Note for Guidance on Clinical Investigation of Medicinal Products in the Treatment of Depres-
sion Committee for Medicinal Products for Human Use (CHMP) Accessed via http://
www.emea.europa.eu/pdfs/human/ewp/051897en.pdf on 23/04/09. Ref: CPMP/EWP/518/97
Rev. 1.
(20) Taylor D, Paton C, Kerwin L. Depression & Anxiety. In: The Maudsley Prescribing Guidelines,
9th edition. London: Informa Healthcare., 2007: 177-270
(21) Hamilton Depression Rating Scale (HDRS) Date accessed: 19/11/08 Accessed via: http://
healthnet.umassmed.edu/mhealth/HAMD.pdf
(22) Kirsch I, Deacon BJ, Huedo-Medina TB et al. Initial severity and antidepressant benefits: a
meta-analysis of data submitted to the Food and Drug Administration. PLoS Medicine 2008; 5
(2):260-268.
(23) Clinical Global Impression (CGI). AstraZeneca Accessed on 19/11/08. Accessed via: http://
www.seroquel.info/assessment-tools/CGI?itemId=826546#.
(24) McDowell I. Measuring Health: A guide to rating scales and questionnaires, Third edition. Ox-
ford.: Oxford University Press., 2006: 382-384
(25) Committee for Medicinal Products for Human Use. Summary of Positive Opinion: Valdoxan.
EMEA Accessed via www.emea.europa.eu, on 19/12/08. Doc ref: EMEA/CHMP/575411/2008.

THIS IS AN NHS DOCUMENT NOT TO BE USED FOR COMMERCIAL AND MARKETING PURPOSES.
PRODUCED TO INFORM LOCAL DECISION-MAKING USING THE BEST AVAILABLE EVIDENCE AT THE TIME OF PUBLICATION.

June 2009 London New Drugs Group APC/DTC Briefing


Agomelatine Page 23

(26) Goodwin GM, Rouillon F, Emsley R European College of Neuropsychopharmacology. 21st


ECNP Congress. Barcelona, Spain. 30 August - 3 September: 2008
(27) Summary of Product Characteristics. Seroxat 10mg, 20mg, 30mg tablets, 20mg/10ml oral
suspension. GlaxoSmithKline UK Accessed via http://emc.medicines.org.uk on 12/12/08.
Date of revision of the text: 26/3/08.
(28) Montgomery SA, Kennedy SH, Burrows GD et al. Absence of discontinuation symptoms with
agomelatine and occurrence of discontinuation symptoms with paroxetine: a randomized,
double-blind, placebo-controlled discontinuation study. Int Clin Psychopharmacol 2004; 19
(5):271-280.
(29) Clayton AH, Montejo AL. Major depressive disorder, antidepressants, and sexual dysfunction.
J Clin Psychiatry 2006; 67(S6):33-37.
(30) Thase ME. Depression, sleep and antidepressants. J Clin Psychiatry 1998; 59(4):55-65.
(31) Major Depressive Disorder Budget Impact Model: Agomelatine. June 2009. Servier Laborato-
ries Limited
(32) Department of Health., Welsh Assembly Government. National Health Service England and
Wales. Drug Tariff. April. 2009
(33) Goodwin GM, Rouillon F, Emsley R European College of Neuropsychopharmacology. 21st
ECNP Congress. Barcelona, Spain. 30 August - 3 September: 2008

The LNDG would like to thank Guy Goodwin, Professor of Psychiatry, Waneford Hospital, Oxford;
Karen Harvey, Senior Lead Pharmacist, Mental Health, Charing Cross Hospital; Trudi Hilton, Chief
Pharmacist, West London Mental Health Trust; Svetlana Jokic, Senior Mental Health Pharmacist, Im-
perial College Healthcare, and David Taylor, Professor of Psychopharmacology, Kings College, for
their comments on the initial draft of the review. Servier Laboratories has commented on the final
review.

THIS IS AN NHS DOCUMENT NOT TO BE USED FOR COMMERCIAL AND MARKETING PURPOSES.
PRODUCED TO INFORM LOCAL DECISION-MAKING USING THE BEST AVAILABLE EVIDENCE AT THE TIME OF PUBLICATION.

June 2009 London New Drugs Group APC/DTC Briefing


Agomelatine Page 24

Appendix 1. Agomelatine: results from the clinical studies:


Placebo / active control/
Treatments Endpoints Results Agomelatine
comparator
Agomelatine 25mg Placebo Paroxetine

Agomelatine 1, 56.3%, not sta-


61.5% (25mg),
5, 25mg or pla- Response 46.3% tistically different
p=0.036 vs. placebo
cebo Response: 50% im- from placebo.
Paroxetine 20mg provement in HAMD
(active control) from baseline 30.4% (25mg), 15.4% 25.7%, p<0.05
Remission
for 8 weeks. Remission: HAMD<7 p<0.01 vs. placebo % vs. placebo
(n=711)
Final HAMD scores 12.77 15.34 13.09
Loo et al7
Drug-placebo difference in
2.57 - 2.25
HAMD scores
Response 49.1%, p=0.03 34.3%

Response, severely depressed 48.7%, p≤0.024 30.7%

Remission 20.8% 13.3%

Remission, severely depressed 21.1% 12.0%


Primary: HAMD final
Agomelatine 25- HAMD final score – all patients 14.1 ± 7.7 16.5 ±7.4
score (ITT).
50mg (n=106) or
Secondary: Response: Drug-placebo difference 2.4, p=0.026 -
placebo (n=105)
50% improvement in
for 6 weeks HAMD final score – severely
HAMD from baseline; 14.4 ± 7.9 17.3 ± 7.2
Kennedy et al8 depressed patients
CGI-S and CGI-I.
Drug-placebo difference 2.9 -
CGI-S, final score 3.2±1.3 3.6±1.3
Drug-placebo difference 0.44, p=0.017
CGI-I, final score 2.4±1.1 2.7±1.1
Drug-placebo difference 0.25, p=0.098
Response 54.3%, p=0.003 35.3%

Response, severely depressed 49.4%, p=0.0476 34.1%

HAMD final score – all patients 13.9 ± 7.8 17 ± 7.9

Drug-placebo difference 3.1, p=0.002 -


Primary: HAMD final
Agomelatine 25-
score (ITT). HAMD final score – severely
50mg (n=118) or 14.6 ± 7.4 18.1 ± 8.4
Secondary: Response: depressed
placebo (n=120)
50% improvement in Drug-placebo difference – se-
for 6 weeks 3.6, p=0.002
HAMD from baseline, verely depressed
Olie et al9
CGI-S and CGI-I.
CGI-I, final score 2.2±1.2 2.7±1.2

Drug-placebo difference 0.45, p=0.006

CGI-S final score 3.1±1.4 3.6±1.4

Drug-placebo difference 0.50, p=0.006 -

Agomelatine 25- HAMD final score 11.1±7.3 12.7±8.5


50mg (n=247) or Primary: change in
fluoxetine 20- HAMD score from base- Change from baseline -17.3±7.3 -16.0±8.4
40mg (n=257). line. Difference in change from
Data on file.17 1.49, p=0.024
baseline
Agomelatine 25- Response: 50% im- Response after 2 weeks 20%, p=0.027 10.9%
50mg (n=154) or provement in HAMD
Response after 6 weeks 70%, p: not signifi- 61.5%
sertraline 50- from baseline
cant
100mg (n=159) Sleep latency, sleep
for 6 weeks efficiency. HAMD final score 10.3, p=0.031 12.1
Kasper et al14;15 Sleep latency, change From 22.5 to 18.9 From 23.48 to 27.75mins
THIS IS AN NHS DOCUMENT NOT TO BE USED FOR COMMERCIALminsAND MARKETING PURPOSES.
PRODUCED TO INFORM LOCAL DECISION-MAKINGSleep
USINGefficiency 78.85%
THE BEST AVAILABLE EVIDENCE 75.70%.
AT THE TIME OF PUBLICATION

June 2009 London New Drugs Group APC/DTC Briefing


Agomelatine Page 25

Treatments Endpoints Results Agomelatine Placebo / comparator


Agomelatine 25mg After 7 days, discontinu- 7.3±7.1 vs. 3.5±4.1,
(n=167) or paroxet- Number of discon- 3.0±4.2 vs. 4.4±5.7
ing vs. continuing p≤0.001
ine 20mg (n=168) tinuation symp-
for 12 weeks toms, DESS. After 14 days, discontinu-
2.0±2.3 vs. 3.0±3.3 3.0±3.9 vs. 2.8±3.5
Montgomery et al28 ing vs. continuing

Agomelatine 25mg
Relapse rates 25.9%, p=NS 23.5%
for 8 weeks then
HAMD total score,
agomelatine (n=187)
time to relapse.
or placebo (n=180) Relapse rates in pts with 21.3%, p=0.046 at
31.3%
for 26 weeks. EPAR6 more severe depression week 52

Agomelatine 25-
50mg (n=165) or Relapse over 6 months 21.7%, p<0.0001 46.6%
placebo (n=174) for Relapse: HAMD
24 weeks plus 20 ≥16
weeks. Relapse over 10 months 23.9%, p<0.0001 49.9%
Goodwin et al10l33

MADRS final score 10.1 ±4.6 9.8 ± 7.9


Between group difference 0.03
Percentage of pa- ≥1 reduction in Sex FX
tients with 50% or 7.3% 15.7%
scale (SAR)
more decrease in
Agomelatine 50mg
sexual activity from Reduced sex drive (SAR) 3.6%, p=0.007 19.4%
(n=137) or venla-
baseline (SAR
faxine 75-150mg Deterioration in orgasm
group) with ≥1 4.3%, p<0.0001 21.2%
(n=140) for 12 (SAR)
point reduction in
weeks Reduction in desire (full
Sex FX scale. 6%, p<0.0001 16.4%
Kennedy et al13 set)
Full set = all sub-
jects sexually ac- Reduction in orgasm (full
9.1%, p<0.001 18.5%
tive at baseline. set)
Reduced Sex FX scores
8.2%, p<0.0001 15.2%
(full set)
Agomelatine 25mg or 4.5% (25mg) 61.9% (paroxetine)
50mg, paroxetine Moderate or severe SD
4.8% (50mg) 0% (placebo)
20mg or placebo Sexual dysfunction.
(n=23/ 22.7% (25mg) 81.0% (paroxetine)
Sustained SD
group).Montejo18 4.8% (50mg) 4.3% (placebo)
From 363 to 393
Total sleep time change
mins
Variables relating Intra-sleep awakening
Agomelatine 25mg From 48 to 29 mins
to the sleep-wake time change
(n=15) for 6 weeks
cycle, measured by
Quera Salva et al16 Sleep efficiency change From 88% to 93%
polysomnography
Stage 3-4 slow wave From 15.9% to
sleep change 19.4%
Final HAMD score 9.9 ± 6.6 11.0 ± 7.4
Getting to sleep score 70.5±16.9, p=0.001 64.1±18.2
Agomelatine 25-
Primary: ‘getting to Quality of sleep score 72.5±21.4, p=0.021 66.9±22.3
50mg (n=165) or
sleep score’ in
venlafaxine 75- Sleep awakening score 66.9±20.5, p=0.04 62.0±21.8
LSEQ.
150mg (n=167) for 6 Integrity of behaviour
Secondary: LSEQ 66.2±20.1, p=0.024 61.0±20.9
weeks score
subscores, HAMD
Lemoine et al12 HAMD insomnia item
1.4±1.6, p=0.044 1.8±1.7
score
Responders 76.4% 70.6%

CGI: Clinical Global Impression, Severity or Improvement


DESS: Discontinuation Emergent Signs and Symptoms
HAMD: Hamilton rating scale for Depression LSEQ: Leeds Sleep Evaluation Questionnaire, measured on visual analogue scale
MADRS: Montgomery-Asberg Depression Rating Scale
SAR: Sexually active remitters, agomelatine n= 60, venlafaxine n=51

THIS IS AN NHS DOCUMENT NOT TO BE USED FOR COMMERCIAL AND MARKETING PURPOSES.
PRODUCED TO INFORM LOCAL DECISION-MAKING USING THE BEST AVAILABLE EVIDENCE AT THE TIME OF PUBLICATION.

June 2009 London New Drugs Group APC/DTC Briefing


Agomelatine Page 26

Appendix 2: Results of the improvements in subjective sleep study, agomelatine vs. venlafaxine.12

Estimated between group difference


(agomelatine - venlafaxine
Agomelatine Venlafaxine Mean 95% CI P value
LSEQ ‘Getting to sleep’ score (mm)
Week 1 Not stated Not stated 5.40 Not stated 0.007
Week 2 Not stated Not stated 4.92 Not stated 0.017
Week 3 Not stated Not stated 4.86 Not stated 0.009
Week 6 Not stated Not stated 5.57 Not stated 0.006
Last 70.5±16.8 64.1±18.2 6.36 Not stated 0.001
Getting to sleep, last visit scores, mm, ITT and severely depressed populations
Easier / Harder: ITT 78.7±19.8 73.3±20.3 5.40 (2.34) 0.79 to 10.00 0.022
Severely depressed 77.9±19.6 69.4±22.7 8.47 0.004
Quicker / Slower: ITT 77.9±17.4 72.3±20.3 5.58 (2.35) 0.94 to 10.21 0.019
Severely depressed 75.5±21.8 68.2±23.0 7.25 0.020
Felt more / less drowsy: ITT 60.2±25.7 54.3±24.5 5.91 (3.14) -0.28 to 12.09 0.061
Severely depressed 60.0±27.6 54.8±24.5 Not significant
LSEQ quality of sleep score (mm)
Week 1 61.2±19.6 55.7±19.8 5.51 (2.26) 1.06 to 9.96 0.015
Week 2 62.2±20.1 61.3±21.4 0.89 (2.39) -3.82 to 5.60 0.71
Week 3 67.0±20.7 64.9±18.3 2.19 (2.31) -2.25 to 6.73 0.343
Week6 76.3±21.4 71.4±19.8 4.85 (2.37) 0.19 to 9.51 0.041
Last 72.5±21.4 66.9±22.3 5.63 (2.43) 0.85 to 10.41 0.021
Within group change 11.3 11.2
LSEQ sleep awakening score (mm)
Week 1 57.4±18.9 53.8±18.2 3.69 (2.13) -0.50 to 7.88 0.084
Week 2 61.1±18.0 55.5±18.1 5.57 (2.09) 1.45 to 9.69 0.008
Week 3 63.1±20.5 57.5±19.5 5.54 (2.36) 0.89 to 10.19 0.02
Week6 69.4±19.4 64.3±21.4 5.13 (2.54) 0.12 to 10.14 0.045
Last 66.9±20.5 62.0±21.8 4.86 (2.35) 0.23 to 9.49 0.040
Within group change 9.5 8.2
LSEQ intergrity of behaviour score (mm)
Week 1 58.6±17.8 48.4±18.5 10.25 (2.05) 6.21 to 14.29 <0.0001
Week 2 57.8±18.4 55.2±19.2 2.60 (2.15) -1.63 to 6.83 0.227
Week 3 61.0±19.2 59.5±18.5 1.49 (3.61) -2.87 to 5.85 0.501
Week6 68.8±18.1 65.2±18.9 3.61 (2.30) -0.93 to 8.15 0.118
Last 66.2±20.1 61.0±20.9 5.16 (2.28) 0.68 to 9.65 0.024
Within group change 7.6 12.6
HAMD insomnia items score
Baseline 4.6±1.2 4.6±1.4
Last visit 1.4±1.6 1.8±1.7 0.37 0.01 to 0.72 0.044
Early insomnia (max score 2) Not stated Not stated 0.18 Not stated <0.05
Middle insomnia (max score 2) Not stated Not stated 0.17 Not stated <0.05
Early awakening (max score 2) Not stated Not stated 0.02 -0.13 to 0.16 Not stated
Depression scores (all patients)
HAMD total score, last value 9.9±6.6 11.0±7.4 Not significant
Responders 76.4% 70.6% -15.35 to 3.73
Change from baseline excluding 21.3±3.0 to 21.4±2.9 to
sleep items 8.5±5.6 9.2±6.3
CGI-I week 1 3.2±0.8 3.6±0.9 0.39 0.20 to 0.58 <0.0001
CGI-I week 6 1.6±0.7 1.6±0.8 0.32 0.06 to 0.58 0.016

THIS IS AN NHS DOCUMENT NOT TO BE USED FOR COMMERCIAL AND MARKETING PURPOSES.
PRODUCED TO INFORM LOCAL DECISION-MAKING USING THE BEST AVAILABLE EVIDENCE AT THE TIME OF PUBLICATION.

June 2009 London New Drugs Group APC/DTC Briefing