CNS Drugs 2012; 26 (4): 297-307

SYSTEMATIC REVIEW 1172-7047/12/0004-0297/$49.95/0

ª 2012 Adis Data Information BV. All rights reserved.

The Role of Antidepressants in the

Management of Fibromyalgia Syndrome
A Systematic Review and Meta-Analysis
Winfried Häuser,1,6 Frederick Wolfe,2,3 Thomas Tölle,4 Nurcan Üc¸eyler5 and Claudia Sommer5
1 Department of Internal Medicine I, Clinical Center Saarbrücken, Saarbrücken, Germany
2 National Data Bank for Rheumatic Diseases, Wichita, KS, USA
3 University of Kansas School of Medicine, Wichita, KS, USA
4 Department of Neurology, University of Würzburg, Würzburg, Germany
5 Department of Neurology, Technische Universität München, München, Germany
6 Department of Psychosomatic Medicine, Technische Universität München, München, Germany

Abstract Background: The role of antidepressants in the management of fibromyalgia

syndrome (FMS) still needs to be determined.
Objective: The objective of this study was to provide a quantitative analysis
(meta-analysis) of the efficacy and harms of antidepressants in the manage-
ment of adult FMS patients.
Data sources: The data sources used were the databases MEDLINE, SCOPUS
and the Cochrane Central Register of Controlled Trials (until December 30,
2010), the reference lists of included articles, and the websites of the US
National Institutes of Health (NIH) and the Pharmaceutical Research and
Manufacturers of America (PhRMA).
Study selection: Studies with a randomized controlled trial (RCT) design
comparing any types of antidepressants with pharmacological placebo or
head-to-head comparisons of different types of antidepressants in FMS
patients were included. RCTs in which antidepressants were combined with
any other defined treatment or antidepressants were tested against anything
but drug placebo were excluded. Patients diagnosed with FMS according to
predefined criteria of any age were included. To be included, studies had to assess
at least one key domain of FMS (pain, sleep, fatigue, health-related quality of life
[HRQOL]) as outcomes of efficacy and report total treatment discontinuation
rates and/or dropout rates due to adverse events as outcomes for harms.
Data extraction: Data were extracted according to protocols of previous
systematic reviews on antidepressants in FMS. Methodology quality was
assessed by the van Tulder score.
Data synthesis: Standardized mean differences (SMD) were calculated for con-
tinuous outcomes by means and standard deviations and relative risks (RR) for
30% pain reduction and total dropout rate for comparisons of antidepressants
with placebo. Examination of the combined results was performed by a random
298 Häuser et al.

effects model. We used Cohen’s categories to evaluate the magnitude of the effect
size, calculated by SMD. Heterogeneity was tested by the I2 statistic.
Thirty-five studies were included in the meta-analysis. The SMDs of seroto-
nin noradrenaline (norepinephrine) reuptake inhibitors (SNRIs) on pain, sleep,
fatigue, depression and HRQOL were significant. Based on Cohen’s categories,
the effect size on pain was small and the ones on sleep, fatigue, depression and
HRQOL were not substantial. 1481/3528 (42.0%) patients with SNRIs and
737/2304 (32.0%) patients with placebo reported a 30% pain reduction (number
needed to treat [NNT] 10.0; 95% CI 8.00, 13.4; I2 = 4%). The RR of dropouts due
to adverse events was 1.83 (95% CI 1.53, 2.18; I2 = 33%).
The SMDs of selective serotonin reuptake inhibitors (SSRIs) on pain, sleep,
depression and HRQOL were significant. Based on Cohen’s categories, the effect
sizes on pain, depression and HRQOL were small and the one on sleep not
substantial. 72/198 (36.4%) patients with SSRIs and 40/194 (20.6%) patients with
placebo reported a 30% pain reduction (NNT 6.3; 95% CI 4.1, 14.1). The RR of
dropouts due to adverse events was 1.60 (95% CI 0.84, 3.04; I2 = 0%).
The SMDs of tricyclic antidepressants (TCAs) on pain, sleep, fatigue and
HRQOL were significant. Based on Cohen’s categories, the effect sizes on pain
and sleep were moderate and the ones on fatigue and HRQOL were small.
140/290 (48.3%) patients with TCAs and 70/252 (27.8%) patients with placebo
reported a 30% pain reduction (NNT 4.9; 95% CI 3.5, 8.0). The RR of dropouts
due to adverse events was 0.84 (95% CI 0.46, 1.52; I2 = 0%).
Conclusions: The TCA amitriptyline and the SNRIs duloxetine and milnaci-
pran are first-line options for the treatment of FMS patients. Physicians and
patients should be realistic about the potential benefits of antidepressants in
FMS. A small number of patients experience a substantial symptom relief
with no or minor adverse effects. However, a remarkable number of patients
dropout of therapy because of intolerable adverse effects or experience only a
small relief of symptoms, which does not outweigh the adverse effects.

1. Background the exclusion of somatic diseases sufficiently ex-

The key symptoms of fibromyalgia syndrome
(FMS) are chronic widespread pain, sleep dis-
turbances/unrefreshed sleep, cognitive dysfunction
and fatigue.[1,2] Patients often report high dis-
ability levels and poor quality of life along with
extensive use of medical care.[3,4] FMS is frequently
associated with other functional somatic symp-
toms such as irritable bowel syndrome,[5] mental
disorders such as anxiety and depressive dis-
orders,[6] and inflammatory rheumatic diseases.[7]
Lacking a specific laboratory test, diagnosis is
established by a history of the key symptoms and
plaining the key symptoms. For the clinical
diagnosis, the 1990 and 2010 American College of
Rheumatology (ACR) criteria[2,8] can be used. In
the past other standardized and recognized cri-
teria have been used to diagnose FMS.
The prevalence of FMS in US adults was es-
timated to be 5 million.[9] The estimated overall
prevalence of FMS was 2.9% in the general pop-
ulation of five European countries.[10]
The definite aetiology of FMS remains unknown.
A biopsychosocial model with interacting somatic,
psychological and social factors predisposing to,
triggering and perpetuating FMS symptoms had

ª 2012 Adis Data Information BV. All rights reserved. CNS Drugs 2012; 26 (4)
Antidepressants in Fibromyalgia Syndrome
been suggested.[11] Cardinal symptoms of FMS may
be due to alterations in central processing of sensory
input, along with aberrations in the endogenous
inhibition of pain. Exposure to physical or psy-
chosocial stressors as outlined above may contrib-
ute to dysfunctional pain processing.[12]
Even though FMS is a ‘‘bitterly controversial
condition’’ among rheumatologists and other med-
ical specialties,[13] numerous pharmacological and
non-pharmacological studies have been conducted
to ameliorate the symptoms of patients diagnosed
with FMS. Besides non-steroidal anti-inflammatory
drugs and opioids, antidepressants are the most
frequently used drugs by FMS patients.[14,15] Anti-
depressants have been recommended by evidence-
based guidelines for the management of FMS.[16-18]
These guidelines partially differ in their evaluation
of the efficacy of antidepressants in FMS.[19]
The serotonin noradrenaline (norepineprhine)
inhibitors duloxetine and milnacipran have been
approved by the US FDA for FMS. Both drugs
were rejected by the European Medical Agency
for the treatment of FMS because their effects
were considered to be too small to be relevant for
patients.[20] Amitriptyline has been approved in
some European countries for the treatment of
chronic pain syndromes.[21] Monoamine oxidase
inhibitors (MAOIs), noradrenaline reuptake in-
hibitors (NRIs) and selective serotonin reuptake
inhibitors (SSRIs) are not approved for FMS in
Europe. All antidepressants except for MAOIs
and NRIs included in this analysis are approved
in the US and most European countries for de-
pressive disorder. The SNRI duloxetine and some
SSRIs are approved in the US and most Europ-
ean countries for generalized anxiety disorder.
Antidepressants are thought to modulate pain
through the central and peripheral nervous sys-
tems. The mechanisms involve noradrenaline
(norepinephrine) and serotonin neurotransmis-
sion, actions on opioid, adrenergic, serotonin,
GABA and NMDA receptors, ion channel activa-
tion and possible effects on inflammatory cytokines.
Effects on peripheral nociceptors, descending
inhibitory pain pathways, central sensitization
and brain areas involved in pain and emotional
processing have been described in basic science
studies.[22,23] However, mechanisms of action of
ª 2012 Adis Data Information BV. All rights reserved.
299
Table I. Types of antidepressants
1. TCAs: imipramine, amitriptyline, clomipramine, desipramine,
dothiepin, ortriptyline, amoxapine, doxepin, protriptyline,
trimipramine, maprotiline
2. MAOIs
a) Irreversible and nonselective classical MAOIs: isocarboxazid,
phenelzine, tranylcypromine
b) Reversible inhibitor of MAOIs: moclobemide
3. SSRIs: citalopram, escitalopram, fluvoxamine, fluoxetine,
paroxetine, sertraline
4. Dual SNRIs: duloxetine, milnacipran, venlafaxine
5. Serotonin 5-HT2 antagonist and reuptake inhibitors (SARIs):
nefazodone, trazodone
6. NaSSAs: mirtazapine
7. NDRIs: bupropion
8. NRIs: reboxetine
9. Other synthetic antidepressants,e.g. melatonergic agonist and
serotonin 5-HT2C antagonists: agomelatine
10. Herbal preparations: St. John’s wort
MAOIs = monoamine oxidase inhibitors; NaSSAs = noradrenergic
and specific serotonergic antidepressants; NDRIs = noradrenaline
and dopamine reuptake inhibitors; NRIs = noradrenaline reuptake
inhibitors; SNRIs = serotonin noradrenaline (norepinephrine)
reuptake inhibitors; SSRIs = selective serotonin reuptake inhibitors;
TCAs = tricyclic antidepressants.
the different types of antidepressants (see table I)
in FMS patients have not been systematically
studied to our knowledge.
This paper presents a systematic review with
quantitative analysis of the efficacy and harms
and a narrative review of evidence-based mech-
anisms of action of antidepressants in the man-
agement of FMS.
2. Methods
Studies with a randomized controlled trial
(RCT) design (parallel or crossover design)
comparing any types of antidepressants with
pharmacological placebo or head-to-head com-
parisons of different types of antidepressants in
FMS patients were included. RCTs in which anti-
depressants were combined with any other de-
fined treatment or antidepressants were tested
against anything but drug placebo (e.g. active
non-pharmacological therapy, physical placebo)
were excluded. Patients diagnosed with FMS
according to predefined criteria of any age were
CNS Drugs 2012; 26 (4)
300
included. Studies should assess at least one key
domain of FMS (pain, sleep, fatigue, health-related
quality of life [HRQOL])[24] as outcomes of effi-
cacy and report total treatment discontinuation
rates and/or dropout rates due to adverse events
(AEs) as outcomes for harms. Discontinuation
rates are considered to be the most consistently
reported estimate of treatment acceptability.[25]
We expanded our searches used for systematic
reviews on antidepressants in FMS[21,26] in the
electronic bibliography databases MEDLINE,
SCOPUS and the Cochrane Central Register of
Controlled Trials (until December 30, 2010) [see
search strategy, Supplemental Digital Content 1
(SDC1), http://links.adisonline.com/CNZ/A13]. We
reviewed the reference lists of included articles. We
searched the websites of the US National Institutes
of Health (NIH) and the Pharmaceutical Re-
search and Manufacturers of America (PhRMA)
for unpublished data.
The methodological details of data extracting
and dealing with missing data have been pre-
viously described.[21] If 30% pain reduction rates
were not reported they were calculated by the
means and SD baseline and final treatment ac-
cording to a validated imputation method.[27]
Methodological quality was assessed by the
van Tulder score (range 0–11).[28] The scoring of
the items was based on the details reported in
publications. Meta-analyses were conducted using
RevMan Analyses software (RevMan 5.1). Stan-
dardized mean differences (SMD) were calculated
by means and SDs for comparison of antide-
pressants with placebo and for intra-group (final
treatment and baseline) effects sizes of antide-
pressants and placebo. Examination of the combin-
ed results was performed by a random effects
model (inverse variance method). SMD used in
Cochrane reviews is the effect size known as
Hedges (adjusted) g. We used Cohen’s categories
to evaluate the magnitude of the effect size, cal-
culated by SMD, with g = 0.0–0.2 = not sub-
stantial effect size, g >0.2–0.5 = small effect size,
g >0.5–0.8 = medium effect size and g >0.8 = large
effect size.[29] Relative risks were calculated for
categorical data by a random effects model (in-
verse variance method). Heterogeneity was tested
using the I2 statistic with I2 values over 50% in-
ª 2012 Adis Data Information BV. All rights reserved.
Häuser et al.
dicating strong heterogeneity.[30] The calculation
of how much the improvement in the active drug
period was attributed to the placebo response was
performed accordingly to a systematic review on
placebo response in antidepressant trials in de-
pression.[31]
3. Results
3.1 Search Results and Study Characteristics
Thirty-seven studies met the inclusion criteria
of which 30 were included in a meta-analysis (see
Prisma flow chart in SDC2, http://links.adis
online.com/CNZ/A14).
RCTs with MAOIs, NRIs, SNRIs, SSRIs and
tricyclic antidepressants (TCAs), but not with
other types of antidepressants in FMS were
found.
We found three studies with MAOIs of which
two studies met the inclusion criteria for a meta-
analysis. MAOIs are not licensed in the US for
FMS or mental disorders. Therefore, no details
of MAOIs in FMS are presented in this paper (see
SDC3, http://links.adisonline.com/CNZ/A15).
We found two studies with the NRI esrebox-
etine in FMS. One study was only available in
database (see SDC3). Pfizer halted the develop-
ment of esreboxetine in FMS. It was considered
unlikely that the drug would provide meaningful
benefit to patients beyond the current standard of
care because of the data available along with current
market dynamics.[32] Therefore, NRIs were not in-
cluded into qualitative and quantitative analysis.
We found 12 RCTs with SNRIs. Five studies
each with duloxetine and milnacipran were in-
cluded into qualitative and quantitative analyses.
One RCT with milnacipran was excluded (double
publication). The results of one study in which
desvenlafaxine was not superior to placebo could
not be included into the analysis because the
outcomes presented were not suited for a meta-
analysis (see SDC4, http://links.adisonline.com/
CNZ/A16). Data from 6070 patients were in-
cluded in the analysis. The study duration ranged
between 12 and 28 weeks. The duloxetine studies
included patients with major depression. All stud-
ies excluded patients with other depressive and
CNS Drugs 2012; 26 (4)
 Antidepressants in Fibromyalgia Syndrome 301

anxiety disorders. All studies performed an inten-

Median Tulder
score (range)

tion-to-treat (ITT) analysis (see tables I and II).

7.5 (5–8)

7 (4–10)
We found 14 RCTs with SSRIs of which seven

7 (5–9)
met the inclusion criteria for qualitative and
quantitative analyses (two studies each with ci-
talopram and paroxetine and three studies with
(weeks) of studies
Median duration

fluoxetine) [see SDC 5, http://links.adisonline.

15.5 (12–28)

com/CNZ/A17]. 322 patients were included. The

SNRIs = serotonin noradrenaline (norepinephrine) reuptake inhibitors; SSRIs = selective serotonin reuptake inhibitors; TCAs = tricyclic antidepressants.
8 (6–16)

8 (6–24)
study duration ranged between 6 and 16 weeks.
(range)

Half of the studies excluded patients with anxiety

and depressive disorders. 3/7 studies performed
No. of patients

an ITT analysis (see tables I and II).

drug/placebo
3664/2406

We found 21 RCTs with TCAs of which 10

on active

159/163

392/325

with 11 study arms met the inclusion criteria for

qualitative and quantitative analyses (nine stud-
Table II. Study characteristics of controlled trials with antidepressants (excluding antidepressant head-to-head comparisons)

ies with amitriptyline, one study with dothiepine,

Study design (N)

one study with nortriptyline) [see SDC6, http://

Crossover (1),

Crossover (3),
Parallel (10)

links.adisonline.com/CNZ/A18]. 717 patients were

parallel (6)

parallel (8)

included. The study duration ranged between

6 and 24 weeks. Half of the studies excluded
patients with anxiety and depressive disorders.
Two of ten studies performed an ITT analysis
Antidepressants used and dosage

dothiepin 75 mg/d (1), nortriptyline

(see tables I and II).

milnacipran 100–200 mg/d (5)

Amitriptyline 10–50 mg/d (9),

Duloxetine 60–120 mg/d (5),

We found five head-to-head comparisons of

Citalopram 20–40 mg/d (2),

paroxetine 20–60 mg/d (2)

fluoxetine 20–80 mg/d (3),

different classes of antidepressants (see SDC 7,

range (no. of studies)

http://links.adisonline.com/CNZ/A19) that were

included in a qualitative analysis. All these studies
compared amitriptyline with different types of
25 mg/d (1)

SSRIs. A quantitative comparison was not possible

because the outcomes were insufficiently reported
in most studies.
studies/study arms

3.2. Efficacy

The effect sizes (SMD true drug vs placebo =

No. of

10/15

10/11

incremental benefit of true drug over placebo) of

7/7

MAOIs on pain, sleep and fatigue were not sig-

Year publication

nificant, the ones on depression and HRQOL

could not be calculated (details not presented).
2004–2010

1996–2007

1986–2001

The effect sizes of SNRIs on pain, sleep, fa-

(range)

tigue, depression and HRQOL were significant.

Based on Cohen’s categories, the effect size on
pain was small and the ones on sleep, fatigue,
Antidepressant class

depression and HRQOL were not substantial. Of

patients with SNRIs, 1481/3528 (42.0%) of
patients with placebo 737/2304 (32.0%) reported
SNRIs

SSRIs

a 30% pain reduction (NNT 10.0; [95% CI 8.00,

TCAs

13.4]) [see tables III and IV].

ª 2012 Adis Data Information BV. All rights reserved. CNS Drugs 2012; 26 (4)
302 Häuser et al.

Table III. Patient characteristics

Antidepressant No. of Mean age in active Mean percentage of Mean percentage of No. of patients with
class studies drug group (range) women (range) Caucasians anxiety and depressive
disorders excluded
SNRIs 7 49.4 (47–51) 94.5 (86.5–100) 89 (77–99.5) 10/10a
SSRIs 7 46.4 (42.3–48.6) 96.3 (80–100) Not reported by 4/7
majority of studies
TCAs 10 49 (38–53) 96 (83–100) Not reported by 3/10
majority of studies
a All duloxetine studies included patients with major depression.
SNRIs = serotonin noradrenaline (norepinephrine reuptake inhibitors); SSRIs = selective serotonin reuptake inhibitors; TCAs = tricyclic
antidepressants.

The effects sizes of SSRIs on pain, sleep, de-
pression and HRQOL were significant. Based on
Cohen’s categories, the effect sizes on pain, de-
pression and HRQOL were small and the one on
sleep not substantial. 72/198 (36.4%) patients with
SSRIs and 40/194 (20.6%) patients with placebo
reported a 30% pain reduction (NNT 6.3; 95% CI
4.1, 14.1) [see tables III and IV].
The effect sizes of TCAs on pain, sleep, fatigue
and HRQOL were significant. Based on Cohen’s
categories the effect sizes on pain and sleep were
moderate and the ones on fatigue and HRQOL
were small. 140/290 (48.3%) patients with TCAs
and 70/252 (27.8%) patients with placebo reported a
30% pain reduction (NNT 4.9; 95% CI 3.5, 8.0).
All head-to-head comparison studies of TCAs
versus SSRIs had a low methodological quality.
There were no consistent findings on superiority
of one drug class over another. One study with a
parallel design compared amitriptyline (25–75 mg/d,
20 patients) with venlafaxine (75 mg/d, 20 patients).
After 8 weeks there were no significant differences
between the two drugs in the reduction of pain,
depression and restrictions of HRQOL. Patients
with SSRIs reported less AEs than patients with
amitriptyline. In a 6-week parallel-design study 29
patients received amitriptyline (up to 100 mg/d)
and 32 patients received paroxetine (20 mg/d).
Paroxetine was superior to amitriptyline in redu-
cing pain and sleep disturbances. Less AEs and
dropouts due to AEs were reported in the parox-
etine group. In an 8-week parallel-design trial 20
patients received amitriptyline (20 mg/d) and
paroxetine (40 mg/d). Amitriptyline was superior
to paroxetine in reducing pain and sleep dis-
turbances. The types of AEs differed in both
groups, but not the frequency. In a crossover de-
sign, 15 patients received for 4 weeks each ami-
triptyline 25 mg and fluvoxamine 50 mg. The
drugs did not differ in the reduction of pain and
additional somatic symptoms. Fluvoxamine was
superior to amitriptyline in reducing anxiety. No
details on AEs were given in this study. In a
crossover design, 15 patients received amitripty-
line 25 mg and fluoxetine 20 mg for 2 weeks.
Outcome measures were exercise performance.

Table IV. Efficacy of different classes of antidepressants compared with pharmacological placebo on key domains of fibromyalgia syndrome
Key domain SNRIs SSRIs TCAs
No. of SMD (95% CI) I2 (%) No. of SMD (95% CI) I2 (%) No. of SMD (95% CI) I2 (%)
studies/ studies/ studies/
patients patients patients
Pain 10/6038 -0.23 (-0.29, -0.18) 0 7/322 -0.40 (-0.73, -0.07) 56 9/520 -0.53 (-0.78, -0.29) 44
Sleep 6/4081 -0.07 (-0.16, 0.03) 45 5/185 -0.31 (-0.60, -0.02) 0 7/343 -0.62 (-0.94, -0.31) 44
Fatigue 9/5656 -0.14 (-0.19, -0.08) 0 5/193 -0.17 (-0.46, 0.11) 0 7/343 -0.57 (-0.93, -0.21) 57
Health-related 10/5987 -0.19 (-0.24, -0.14) 0 3/139 -0.47 (-0.84, -0.10) 36 5/322 -0.20 (-0.43, 0.03) 0
quality of life
SMD = standardized mean difference; SNRIs = serotonin noradrenaline (norepinephrine) reuptake inhibitors; SSRIs = selective serotonin
reuptake inhibitors; TCAs = tricycyclic antidepressants.

ª 2012 Adis Data Information BV. All rights reserved. CNS Drugs 2012; 26 (4)
Antidepressants in Fibromyalgia Syndrome
Table V. Thirty percent pain reduction, acceptability and tolerability of different classes of antidepressants vs pharmacological placebo
Outcome SNRIs SSRIs
No. of RR (95% CI) I2 No. of
studies/ studies/
patients patients
30% pain reduction 10/5832 1.30 (1.20, 1.40) 4 7/392
Acceptability 10//6063 0.98 (0.84, 1.14) 73 7/414
Tolerability 10/6509 1.83 (1.53, 2.18) 33 7/392
a Calculated by imputation method.
RR = relative risk; SNRIs = serotonin noradrenaline (norepinephrine) reuptake inhibitors; SSRIs = selective serotonin reuptake inhibitors;
TCAs = tricyclic antidepressants.
Both drugs improved anaerobic performance. No
details on AEs were given.
3.3. Acceptability, Tolerability and Safety
Only a minority of studies reported the way by
which AEs were assessed (e.g. by spontaneous re-
ports, open or structured questions, standardized
questionnaires). Studies published before 2000 re-
ported a lower number and frequency of AEs than
studies published between 2005 and 2010. AEs in
the placebo arms depended on the AEs of the ac-
tive medication against which the placebo was
compared (details not presented).[33] These findings
are in accordance with the expectation theory of
placebo and nocebo effects.[34] Therefore, the data
on patient-reported AEs of antidepressants in
FMS must be interpreted cautiously.
The acceptability (total dropout rate) did not
differ between all types of antidepressants analysed
and placebo. 743/3647 (21.4%) patients with SNRIs
and 262/2412 (10.9%) patients with placebo drop-
ped out due to AEs (number needed to harm [NNH]
10.5; 94% CI 8.8, 13.0.9) SNRI studies. 14/148
(9.5%) patients with SSRIs and 10/143 (7.0%)
patients with placebo dropped out due to AEs
(NNH 40; 95% CI 19, 66) in SSRI studies. 20/384
(5.2%) patients with TCAs and 20/307 (6.5%)
patients with placebo dropped out to due to AEs
(NNH 76; 95% CI 32, 96) in TCA studies. SNRIs
were less tolerated than placebo, because the drop-
out rate due to AEs of SNRIs was higher than the
one of placebo. The tolerability of SSRIs and TCAs
did not differ from the one of placebo (see table V).
The most frequently reported AEs of SNRIs
were constipation, insomnia, dry mouth, headache,
ª 2012 Adis Data Information BV. All rights reserved.
303
TCAs
RR (95% CI) I2 No. of RR (95% CI) I2
studies/
patients
1.59 (1.01, 2.52)a 0 9/542 1.60 (1.15, 2.24)a 43
1.36 (0.79, 2.36) 46 11/708 0.76 (0.54, 1.07) 0
1.60 (0.84, 3.04) 0 10/691 0.84 (0.46, 1.52) 0
nausea, hyperhidrosis, palpitations and anorexia.
The most frequently reported AEs of SSRIs were
nausea, sleeping problems and sexual disturbances.
The most frequently reported AEs of TCAs were
dry mouth, dizziness and somnolence. Compared
with placebo, the RR of nausea, headache and
dizziness of SNRIs, the RR of sexual dysfunction
of SSRIs and the RR of dry mouth of TCAs was
significant (see table VI). No study reported
deaths of patients attributable to medication or
clinically relevant organ damage.
The contraindications and precautions of SNRIs,
SSRIs and TCAs are listed in table VII.[34-39]
Most notably, all antidepressants increase the risk
of suicide and can lead to severe liver damage.
3.4. Evidence-Based Mechanism of Action
Placebo accounted for 59.8% of the effect of
SNRIs, for 29.3% of SSRIs and for 28.0% of
TCAs on pain. These findings are in line with our
recent analysis of the placebo response rates in
any type of drug therapy in FMS. The placebo
response rate increased over the years. The in-
crease of placebo response rates is associated with
the conduction of large multicentre trials includ-
ing patients of different countries recruited by
media and sponsored by well known pharma-
ceutical companies.[40]
The placebo response is constituted by the pla-
cebo effect (e.g. positive expectations of patients
that the drug will help), natural course of the dis-
ease (e.g. spontaneous improvement), uncontrolled
co-interventions and regression to the mean.[41]
Recent longitudinal studies with patients of
the National Data Bank of Rheumatic Diseases
CNS Drugs 2012; 26 (4)
 304 Häuser et al.

revealed no average clinically meaningful im-

41
I2

0
provement in symptom severity overall in FMS

4.43 (1.18, 16.68)

2.14 (0.23, 20.17)

patients.[42] No RCT with antidepressants in FMS
1.73 (0.49, 6.14)

RR = relative risk; SNRIs = serotonin noradrenaline (norepinephrine) reuptake inhibitors; SSRIs = selective serotonin reuptake inhibitors; TCAs = tricyclic antidepressants.
controlled the effects for the amount of rescue
RR (95% CI)

medication. However, there is no evidence for the

efficacy of the rescue medication used (aceta-
minophen, oxycodone).[17] Therefore, we assume
that the placebo response in FMS trials is mainly
No. of studies/patients

composed by placebo effects.

Two studies with duloxetine performed a path
analysis to test the direct treatment effect on pain
reduction after accounting for the indirect effect
24/255

16/132
TCAs

2/124
through improvement of depressive symptoms.
The indirect treatment effect through improvement
in depressive symptoms accounted for 24.4–30.9%
63
I2
0

(60 mg) and 13.1–17.7% (120 mg) of the total treat-

Table VI. Most frequently reported adverse events of different classes of antidepressants vs pharmacological placebo

1.40 (0.82, 1.66)

0.94 (0.46, 1.68)

9.51 (1.22, 74.0)

ment effect on pain in duloxetine trials including

RR (95% CI)

FMS patients with major depression.[43,44]

Functional magnetic resonance imaging
(fMRI) studies demonstrated that FMS patients
treated with milnacipran exhibited a reduction in
pain sensitivity and a parallel increase in activity
No. of studies/patients

in brain regions implicated in the descending pain

inhibitory pathways compared with placebo-
treated patients.[45]
27/121

33/100
SSRIs

10/79

4. Discussion and Conclusions

24

21

45

Antidepressants are the best-studied drug

I2

group in FMS.[18] Amitriptyline, duloxetine and

2.07 (1.82, 2.37)

1.39 (1.21, 1.59)

1.94 (1.52, 2.48)

4.26 (3.12, 5.82)

milnacipran are, besides pregabalin, first-line

RR (95% CI)

options for the treatment of FMS patients.[18]

This recommendation is based on the following:
amitriptyline, duloxetine and milnacipran are the
best studied (number of studies and patients in-
No. of studies/patients

cluded) antidepressants in FMS. The data on

efficacy are robust even though the incremental
benefit over placebo is small. The potential harms
1120/3341

of these three drugs do not differ substantially

602/3281

356/3178

554/3341
SNRIs

from other antidepressants. Amitriptyline is avail-

able for pain therapy in most countries. Dulox-
etine and milnacipran have been approved for
Dizziness/somnolence

FMS in the US. Amitriptyline, duloxetine and

Sexual dysfunction

milnacipran have been recommended as first-line

treatment options in the most up-to-date evi-
Constipation

Weight gain
Dry mouth
Headache

dence- and interdisciplinary consensus-based

Outcome

Nausea

guidelines on FMS.[18] The review does not fa-

vour amitriptyline, duloxetine and milnacipran

ª 2012 Adis Data Information BV. All rights reserved. CNS Drugs 2012; 26 (4)
Antidepressants in Fibromyalgia Syndrome 305

Table VII. Summary of contraindications and warnings of antidepressants given by the US FDA
Drug Contraindications Warnings and precautions Last update
Duloxetine 1. Concomitant use of MAOIs 1. Suicidality 1/2010
2. Uncontrolled narrow-angle glaucoma 2. Hepatotoxicity
3. Substantial alcohol use or evidence of chronic 3. Orthostatic hypotension and syncope
liver damage 4. Serotonin- or neuroleptic syndrome-like reactions
4. Severe renal impairment 5. Abnormal bleeding
6. Discontinuation syndrome
7. Activation of mania
8. Blood pressure control
9. Hyponatremia
10. Glucose control in diabetes mellitus
11. Slow gastric emptying
12. Urinary hesitation and retention
Milnacpran 1. Concomitant use of MAOIs 1. Suicidality 1/2009
2. Uncontrolled narrow-angle glaucoma 2. Hepatotoxicity
3. Substantial alcohol use or evidence of chronic 3. Serotonin syndrome
liver damage 4. Abnormal bleeding
5. Discontinuation syndrome
6. Elevated blood pressure
7. Urinary hesitation and retention
8. Seizures
SSRIs 1. Concomitant use of MAOIs, thioridazine 1. Suicidality Fluoxetine 4/2011
and pimozide 2. Abnormal bleeding Paroxetine 7/2011
3. Anxiety and insomnia
4. Activation of mania/hypomania
5. Hyponatremia
6. Seizures
TCAs 1. Prior hypersensitization 1. Suicidality Amitriptyline
2. Concomitant use of MAOIs 2. Anxiety and insomnia 1/2010
3. Acute recovery phase following myocardial 3. Activation of mania/hypomania
infarction and schizophrenia
4. Cardiovascular disorders
5. Hyperthyroid patients or those receiving
thyroid medication
6. Elective surgery
7. Elevated or lowered blood sugar
8. Impaired liver function
MAOIs = monoamine oxidase inhibitors; SSRIs = selective serotonin reuptake inhibitors; TCAs = tricyclic antidepressants.

for FMS treatment. The data on TCAs are based
on studies with small sample sizes conducted be-
tween 1986 and 1998 and the data on SNRIs are
based on studies with large sample sizes con-
ducted between 2004 and 2010. The setting of the
studies and the recruitment of the patients chan-
ged over the years. The placebo response rate but
also the nocebo dropout rates increased.[33,40]
Therefore, comparisons of TCAs and SSRIs with
SNRIs are limited.
Our data suggest a higher efficacy and better
tolerability of SSRIs compared with SNRIs.
However, the limited comparability of RCTs as
pointed out above for TCAs and SNRIs is also
valid for the comparison of SNRIs with SSRIs.
The lower placebo response rates in SSRI studies
(36) compared with one of the SNRI studies leads
to more favourable NNT for 30% pain reduction
for SSRIs. The lower nocebo dropout rates in
SSRI studies are associated with a lower NNH
(dropout rate due to AEs) in SSRI studies com-
pared with SNRI studies.
If antidepressants are considered for FMS
therapy, shared decision making of patients and
physicians has been recommended.[18] The choice
of drug should depend on which symptoms of
FMS should be targeted by the drug and on the
potential of the drug to reduce the symptoms.

ª 2012 Adis Data Information BV. All rights reserved. CNS Drugs 2012; 26 (4)
306
Amitriptyline should be preferred for co-morbid
sleep disturbances and duloxetine for co-morbid
major depression. In addition, the importance of
potential adverse effects to the patient (e.g. sexual
dysfunction by SSRI, weight gain by TCAs) and
contraindications (e.g. SNRIs in case of severe
liver damage) should be considered. Finally, the
costs (amitriptyline is available as a generic and is
cheaper than duloxetine and milnacipran) and
local availability may be an important issue to
consider.
Physicians and patients should be realistic
about the potential benefit of antidepressants in
FMS: antidepressants do not totally relieve all FMS
symptoms. A small number of patients clearly
benefit from treatment with antidepressants. These
patients experience a substantial symptom relief
with no or minor adverse effects. However, a
remarkable number of patients drop out of ther-
apy because of intolerable adverse effects or ex-
perience only a small relief of symptoms that does
not outweigh the adverse effects.
The relief of FMS symptoms by antidepres-
sants can be explained by placebo effects, reduc-
tion of depression and effects on pain processing
including the modulation of the descending pain
control mechanism. Physicians should use place-
bo effects, e.g. by promoting positive (but realis-
tic) expectations on the efficacy and safety of
antidepressants in FMS.
FMS should not be treated by drugs alone.[18,46]
The efficacy of aerobic exercise, balneotherapy and
multicomponent therapy (combination of physical
exercise with psychological therapy) to relieve FMS
symptoms has been demonstrated by systematic
reviews.[47-49] These non-pharmacological treatment
options are safe and well tolerated, and are alter-
natives or combination options to drug therapy.
Acknowledgements
Dr Häuser received honoraria for one educational lecture
from Janssen-Cilag. The National Data Bank for Rheumatic
Diseases, the employer of Dr Frederick Wolfe, received a
grant from Pfizer for conducting a fibromyalgia criteria study.
Dr Tölle received grants within the innovative medicines in-
itiative supported by Pfizer and payments for lectures in-
cluding services on speakers’ bureaus from Pfizer, Eli-Lilly,
Grünenthal and Astellas. Dr Üc¸eyler has no conflicts of in-
ª 2012 Adis Data Information BV. All rights reserved.
Häuser et al.
terest to declare. Dr Sommer received honoraria for educa-
tional lectures from Pfizer and Eli-Lilly and has participated
in advisory boards of these companies.
References
1. Häuser W, Zimmer C, Felde E, et al. What are the key
symptoms of fibromyalgia syndrome? Results of a survey
of the German Fibromyalgia Association [in German].
Schmerz 2008; 22: 176-83
2. Wolfe F, Clauw DJ, Fitzcharles MA, et al. The American
College of Rheumatology preliminary diagnostic criteria
for fibromyalgia and measurement of symptom severity.
Arthritis Care Res (Hoboken) 2010; 62: 600-10
3. Wolfe F, Anderson J, Harkness D, et al. A prospective,
longitudinal, multicenter study of service utilization and
costs in fibromyalgia. Arthritis Rheum 1997; 40: 1553-5
4. Winkelmann A, Perrot S, Schaefer C, et al. Impact of fi-
bromyalgia severity on health economic costs: results from
a European cross-sectional study. Appl Health Econ
Health Policy 2011; 9: 125-36
5. Henningsen P, Zipfel S, Herzog W. Management of func-
tional somatic syndromes. Lancet 2007 Mar 17; 369: 946-55
6. Fietta P, Fietta P, Manganelli P. Fibromyalgia and psy-
chiatric disorders. Acta Biomed 2007; 78: 88-95
7. Weir PT, Harlan GA, Nkoy FL, et al. The incidence of fi-
bromyalgia and its associated comorbidities: a population-
based retrospective cohort study based on International
Classification of Diseases, 9th Revision codes. J Clin
Rheumatol 2006; 12: 124-8
8. Lawrence RC, Felson DT, Helmick CG, et al. Estimates of
the prevalence of arthritis and other rheumatic conditions in
the United States: part II. Arthritis Rheum 2008; 58: 26-35
9. Wolfe F, Smythe HA, Yunus MB, et al. The American
College of Rheumatology 1990 criteria for the classifica-
tion of fibromyalgia: report of the Multicenter Criteria
Committee. Arthritis Rheum 1990; 33: 1863-4
10. Branco JC, Bannwarth B, Failde I, et al. Prevalence of fi-
bromyalgia: a survey in five European countries. Semin
Arthritis Rheum 2010; 39: 448-55
11. Sommer C, Häuser W, Gerhold K, et al. Etiology and pa-
thophysiology of fibromyalgia syndrome and chronic
widespread pain [in German]. Schmerz 2008; 22 (3): 267-82
12. Bradley LA. Pathophysiology of fibromyalgia. Am J Med
2009; 122 (12 Suppl.): S22-30
13. Wolfe F. Fibromyalgia wars. J Rheumatol 2009; 36 (4): 671-8
14. Bennett RM, Jones J, Turk DC, et al. An internet survey of
2,596 people with fibromyalgia. BMC Musculoskelet Dis-
ord 2007 Mar 9; 8: 27
15. Marschall U, Arnold B, Häuser W. Treatment and health-
care costs of fibromyalgia syndrome in Germany: analysis
of the data of the Barmer Health Insurance (BEK) from
2008-2009 [in German]. Schmerz 2011; 25: 402-10
16. Burckhardt CS, Goldenberg D, Crofford L, et al. Guideline
for the management of fibromyalgia syndrome: pain in
adults and children. APS Clinical Practice Guideline Series
No. 4. Glenview (IL): American Pain Society, 2005
17. Carville SF, Arendt-Nielsen S, Bliddal H, et al. EULAR
evidence based recommendations for the management of
fibromyalgia syndrome. Ann Rheum Dis 2008; 67: 536-41
CNS Drugs 2012; 26 (4)
Antidepressants in Fibromyalgia Syndrome
18. Häuser W, Eich W, Herrmann M, et al. Fibromyalgia syn-
drome: classification, diagnosis, and treatment. Dtsch
Arztebl Int 2009; 106: 383-91
19. Häuser W, Thieme K, Turk DC. Guidelines on the man-
agement of fibromyalgia syndrome: a systematic review.
Eur J Pain 2010; 14: 5-10
20. Briley M. Drugs to treat fibromyalgia: the transatlantic dif-
ference. Curr Opin Inves Drugs 2010; 11: 16-8
21. Häuser W, Petzke F, Üc¸eyler N, et al. Comparative efficacy
and acceptability of amitriptyline, duloxetine and milnaci-
pran in fibromyalgia syndrome: a systematic review with
meta-analysis. Rheumatology (Oxford) 2011; 50: 532-43
22. Verdu B, Decosterd I, Buclin T, et al. Antidepressants for
the treatment of chronic pain. Drugs 2008; 68: 2611-32
23. Chan HN, Fam J, Ng BY. Use of antidepressants in the
treatment of chronic pain. Ann Acad Med Singapore 2009;
38: 974-9
24. Mease PJ, Clauw DJ, Christensen R, et al., OMERACT
Fibromyalgia Working Group. Toward development of a
fibromyalgia responder index and disease activity score:
OMERACT module update. J Rheumatol 2011; 38: 1487-95
25. Cipriani A, Furukawa TA, Salanti G, et al. Comparative
efficacy and acceptability of 12 new-generation antide-
pressants: a multiple-treatments meta-analysis. Lancet
2009; 373: 746-58
26. Häuser W, Bernardy K, Uc¸eyler N, et al. Treatment of
fibromyalgia syndrome with antidepressants: a meta-
analysis. JAMA 2009; 301: 198-209
27. Furukawa TA, Cipriani A, Barbui C, et al. Imputing res-
ponse rates from means and standard deviations in meta-
analyses. Int Clin Psychopharmacol 2005; 20: 49-52
28. van Tulder M, Furlan A, Bombardier C, et al. Updated
method guidelines for systematic reviews in the cochrane
collaboration back review group. Spine 2003; 28: 1290-9
29. Cohen J. Statistical power analysis for the behavioral
sciences. Hillsdale (NJ): Lawrence Erlbaum Associates,
1988
30. Higgins JPT, Green S, editors. Cochrane handbook for
systematic reviews of interventions, version 5.1.0, updated
March 2011 [online]. Available from URL: http://www.
cochrane-handbook.org/ [Accessed 2011 Dec 6]
31. Rief W, Nestoriuc Y, Weiss S, et al. Meta-analysis of the
placebo response in antidepressant trials. J Affect Disord
2009; 118: 1-8
32. Kelly J. Pfizer stops work on esreboxetine for FM [online].
Available from URL: http://www.mskreport.com/articles.
cfm?articleID=3293 [Accessed 2011 Feb 23]
33. Häuser W, Bartram C, Bartram-Wunn E, et al. Systematic
review: adverse events attributable to nocebo in random-
ised controlled drug trials in fibromyalgia syndrome and
painful diabetic peripheral neuropathy. Clin J Pain. Epub
2012 Feb 14
34. Amanzio M, Corazzini LL, Vase L, et al. A systematic re-
view of adverse events in placebo groups of anti-migraine
clinical trials. Pain 2009; 146: 261-9
ª 2012 Adis Data Information BV. All rights reserved.
307
35. Drugs information online. Duloxetine [online]. Available
from URL: http://www.accessdata.fda.gov/drugsatfda_
docs/label/2010/022516lbl.pdf [Accessed 2011 Aug 15]
36. Drugs information online. Milnacipran [online]. Available
from URL: http://www.drugs.com/pro/savella.html [Ac-
cessed 2011 Aug 15]
37. Drugs information online. Fluoxetine [online]. Available
from URL: http://www.drugs.com/pro/fluoxetine.html
[Accessed 2011 Aug 15]
38. Drugs information online. Paroxetine [online]. Available from
URL: http://www.drugs.com/pro/paxil.html [Accessed 2011
Aug 15]
39. Drugs information online. Amitriptyline [online]. Available
from URL: http://www.drugs.com/pro/amitriptyline.html
[Accessed 2011 Aug 15]
40. Häuser W, Bartram-Wunn E, Bartram C, et al. Systematic re-
view: placebo response in drug trials of fibromyalgia syn-
drome and painful peripheral diabetic neuropathy: magnitude
and patient-related predictors. Pain 2011; 152: 1709-17
41. Ernst E, Herxheimer A. The power of placebo. BMJ 1996;
313: 1569-70
42. Walitt B, Fitzcharles MA, Hassett AL, et al. The long-
itudinal outcome of fibromyalgia: a study of 1555 patients.
J Rheumatol 2011; 38: 2238-46
43. Arnold LM, Rosen A, Pritchett YL, et al. A randomized,
double-blind, placebo-controlled trial of duloxetine in the
treatment of women with fibromyalgia with or with out
major depressive disorder. Pain 2005; 119: 5-15
44. Russell IJ, Mease PJ, Smith TR, et al. Efficacy and safety of
duloxetine for treatment of fibromyalgia in patients with or
without major depressive disorder: results from a 6-month,
randomized, double-blind, placebo-controlled, fixed-dose
trial. Pain 2008; 136: 432-44
45. Mainguy Y. Functional magnetic resonance imagery
(fMRI) in fibromyalgia and the response to milnacipran.
Hum Psychopharmacol 2009; 24 Suppl. 1: S19-23
46. Arnold LM, Clauw DJ. Fibromyalgia syndrome: practical
strategies for improving diagnosis and patient outcomes.
Am J Med 2010; 123: S2
47. Häuser W, Klose P, Langhorst J, et al. Efficacy of different
types of aerobic exercise in fibromyalgia syndrome: a
systematic review and meta-analysis of randomised con-
trolled trials. Arthritis Res Ther 2010; 12: R79
48. Langhorst J, Musial F, Klose P, et al. Efficacy of hydro-
therapy in fibromyalgia syndrome: a meta-analysis of
randomized controlled clinical trials. Rheumatol (Oxford)
2009; 48: 1155-9
49. Häuser W, Bernardy K, Offenbächer M, et al. Efficacy of
multicomponent treatment of fibromyalgia syndrome: a
meta-analysis of randomised controlled trials. Arthritis
Rheum 2009; 61: 216-24
Correspondence: Dr Winfried Häuser, MD, Department of
Internal Medicine I, Klinikum Saarbrücken gGmbH Win-
terberg 1, D-66119 Saarbrücken, Germany.
E-mail: whaeuser@klinikum-saarbruecken.de
CNS Drugs 2012; 26 (4)