A systematic review on the effectiveness of treatment with antidepressants in fibromyalgia syndrome

Authors

  • Nurcan Üçeyler,

    1. University of Würzburg, Würzburg, Germany
    Search for more papers by this author
  • Winfried Häuser,

    1. Klinikum Saarbrücken, Saarbrücken, Germany
    Search for more papers by this author
    • Dr. Häuser has received consultant fees, speaking fees, and/or honoraria (less than $10,000) from Janssen-Cilag, Eli Lilly, and Mundipharma.

  • Claudia Sommer

    Corresponding author
    1. University of Würzburg, Würzburg, Germany
    • Department of Neurology, University of Würzburg, Josef-Schneider-Strasse 11, 97080 Würzburg, Germany
    Search for more papers by this author
    • Dr. Sommer has received speaking fees (less than $10,000 each) from Eli Lilly and Boehringer-Ingelheim.


Abstract

Objective

To systematically review the efficacy of treatment of fibromyalgia syndrome (FMS) with antidepressants.

Methods

We screened Medline, PsychINFO, SCOPUS, and the Cochrane Library databases (through October 2007) and the reference sections of original studies, meta-analyses, and evidence-based guidelines and recommendations on antidepressants in FMS. Randomized controlled trials (RCTs) on the treatment of FMS with antidepressants were analyzed. Inclusion criteria, study characteristics, quality, and all outcome measures were investigated.

Results

Twenty-six of 167 studies were included. The main outcome variables reviewed were pain, fatigue, sleep, depressiveness, and quality of life. Amitriptyline, studied in 13 RCTs, was efficient in reducing pain with a moderate magnitude of benefit (pain reduction by a mean of 26%, improvement in quality of life by 30%). Selective serotonin reuptake inhibitors (SSRIs) were studied in 12 RCTs, which also showed positive results, except for 2 studies on citalopram and 1 on paroxetine. Three RCTs on the dual serotonin and noradrenaline reuptake inhibitors (SNRIs) duloxetine and milnacipran and 1 of the 2 RCTs using the monoamine oxidase inhibitor moclobemide reported a positive result. The longest study duration was 12 weeks.

Conclusion

Amitriptyline 25–50 mg/day reduces pain, fatigue, and depressiveness in patients with FMS and improves sleep and quality of life. Most SSRIs and the SNRIs duloxetine and milnacipran are probably also effective. Short-term treatment of patients with FMS using amitriptyline or another of the antidepressants that were effective in RCTs can be recommended. Data on long-term efficacy are lacking.

INTRODUCTION

Chronic widespread pain (CWP) is a common disorder with an estimated prevalence of 11% (1). According to the criteria of the American College of Rheumatology (ACR), fibromyalgia syndrome (FMS) is defined as CWP and tenderness in at least 11 of 18 defined tender points (2). Population-based estimates of the prevalence of FMS range from 0.5% to 5.8% (3). CWP and FMS are frequently associated with fatigue, poor sleep, other functional somatic syndromes, mental and physical disorders, and disability and reduced quality of life (4–6). Patients diagnosed with FMS cause high direct (health care use) (7–9) and indirect costs (sick leave, disability pension) (10). Effective treatment options are therefore needed for medical and economic reasons (11).

A variety of pharmacologic and nonpharmacologic treatments are offered to and used by patients diagnosed with FMS (11, 12). More than 500 peer-reviewed articles on the therapy of FMS have been published (13). Yet to date, no treatment has proven effective for the entire scope of symptoms and disabilities associated with FMS (14). Recently, evidence-based guidelines on the management of FMS have been published (15, 16). In evidence-based medicine, systematic reviews and meta-analyses of specific treatment modalities are considered to ensure the highest quality for specific recommendations (17). The strength of these recommendations is, in turn, based on the level of evidence for the presence of a specific treatment effect (17), on considerations such as the clinical relevance and effect size of the outcome measures, and on whether the study population is such that the results may be generalized to the entire patient population (18).

More than 500 randomized controlled trials (RCTs) have tested the effectiveness of various drugs in the treatment of FMS. Based on these, the evidence-based guidelines of the American Pain Society (APS) (16) and the recommendations of the European League Against Rheumatism (EULAR) on the management of FMS (15) gave the highest level of recommendation to antidepressants. None of the older antidepressants has been licensed for the treatment of FMS (the first licensed drug was the anticonvulsant pregabalin), but some have been approved for the treatment of chronic pain, and an application has very recently been granted by the US Food and Drug Administration for duloxetine. The APS guidelines and EULAR review are based on a search for RCTs from 1966 until 2004 and until December 2005, respectively. There have been 2 meta-analyses on antidepressants in the treatment of FMS based on searches of controlled trials from 1966 until 1999 (19) and from 1966 until November 2000 (20).

We saw the need for a current systematic review, because additional RCTs using new classes of antidepressants in FMS have since been published, and moreover, the previous meta-analyses and evidence-based guidelines only reviewed selected outcome variables such as pain, sleep, and mood, and did not include all study end points. Furthermore, no analyses of potential predictors of treatment response and of associations between changes in somatic symptoms and depressiveness have been performed up to now. From an economic point of view, it would also be of interest to know if antidepressants can reduce FMS-related costs (12).

The goals of our systematic review were to expand the overview of RCTs on the treatment of FMS with antidepressants by a new literature search up to October 2007, to outline the results of all outcome parameters of the studies, and to screen the studies for specific patient characteristics that might predict a positive or negative treatment response.

MATERIALS AND METHODS

Setting.

The literature search was part of and expanded upon the development of evidence-based interdisciplinary guidelines for the diagnosis and treatment of FMS on behalf of the Association of the Scientific Medical Societies of Germany coordinated by the German Interdisciplinary Association of Pain. Nine German scientific societies (clinical psychology, general medicine, neurology, orthopedics, pain, psychiatry, psychosomatics, rehabilitation, and rheumatology) were involved in this project.

Searches.

The electronic bibliography databases screened included Medline (1966 to October 2007), PsychINFO (1966 to October 2007), SCOPUS (1980 to October 2007), and the Cochrane Library (1993 to October 2007). The keywords (all languages; limited to “human”) “fibromyalgia,” “fibromyalgia syndrome,” and “chronic widespread pain” were used in combination with the terms “randomized,” “clinical controlled trial,” “randomized controlled trial,” “review,” “metaanalysis,” “antidepressant,” and “antidepressive agents.” In addition, reference sections of original studies, meta-analyses on antidepressants in FMS, and evidence-based recommendations were manually screened independently by the authors.

Inclusion and exclusion criteria.

To be included in our review, studies needed to meet the following criteria: 1) diagnosis of FMS or CWP based on recognized criteria (i.e., ACR &lsqbr;2&rsqbr;, Smythe and Moldofsky [21], Yunus [22]); 2) a controlled study design with a control group that received placebo, usual care, or another well-defined treatment; 3) treatment with antidepressants (tri- and tetracyclic antidepressants [TCAs], selective serotonin reuptake inhibitors [SSRIs], dual serotonin and noradrenaline reuptake inhibitors [SNRIs], and monoamine oxidase inhibitors [MAOIs]); and 4) symptom-specific outcome measures. Although reported effective in a meta-analysis in patients with FMS (20), we did not include studies with cyclobenzaprine, which combines characteristics of an antidepressant and a muscle relaxant. Furthermore, we did not include studies with S-adenosylmethionine because this substance is officially available only in 2 European countries (UK and Italy) as a dietary supplement.

Data extraction.

Each author independently screened the titles and abstracts of potentially eligible articles that were identified by the search strategy detailed above. The full text of the selected studies was examined to determine which studies met the inclusion criteria (see below). Data extraction was carried out independently by the authors, using standard extraction forms. The characteristics of the selected studies including methods, participants, interventions, and outcomes were analyzed according to the user guide of the Review Manager (RevMan 4.2.10) of the Cochrane collaboration (www.cc-ims.net/RevMan). The methodologic quality of the studies was assessed using the following criteria: 1) Jadad score (a validated numerical score ranging 0–5 [23]) and the following parameters based on the Consolidated Standards of Reporting Trials (CONSORT) checklist (24): 2) a priori definition of primary and secondary end points (CONSORT item 5); 3) power calculation (CONSORT item 7); 4) use of adequate multivariate statistical tests; 5) adjustment for multiple testing (CONSORT item 18); and 6) suitability for meta-analysis. Differences in scoring between the authors were discussed until consensus was reached.

Statistical analysis.

All data were analyzed using Winstat for Excel, version 2001.1 (R. Fitch Software, Bad Krozingen Germany). Data derived from descriptive statistical analysis are presented as percentages for categorical variables and as medians and ranges for continuous data. Comparisons of proportions were performed by chi-square test and for continuous variables by nonparametric tests (Mann-Whitney U test, Kruskal-Wallis H-test).

RESULTS

Study selection.

The literature search yielded 167 citations. In the first round of study evaluation, 33 of these articles were selected. The excluded 134 articles contained duplicate publications, review articles, case reports, studies that did not include an antidepressant arm (but neuroleptics or cyclobenzaprine instead), or studies that investigated patients with organic rheumatologic syndromes. In the second round of study evaluation, an additional 7 articles of the 33 initially selected were excluded for the following reasons: no recognized criteria for FMS (25, 26), duplicate publication (27), report of n-of-1 trials (28), and lack of an antidepressant-only arm (combination of antidepressants with cyclobenzaprine [29], cardiovascular fitness training [30], or guided imagery [31] instead). We finally included and reviewed 26 studies investigating the effects of antidepressants on FMS symptoms.

Study design.

Twelve (46%) of the studies were conducted in North America (US, Canada), 1 (4%) study in South America (Brazil), 7 (27%) in Western Europe (Scandinavia, Belgium, France, Italy), and 6 (25%) in Turkey (Table 1). All studies were outpatient based. Most patients (58% of the studies) were recruited from departments of rheumatology. Eight (31%) studies had a multicenter design. Twenty-two (85%) studies used a parallel design; 4 (17%) trials were cross-over studies. TCAs (almost always amitriptyline) were investigated in 13 studies, and were compared with placebo in 9 studies, with other antidepressants in 6 (3 of these with placebo control), and with nonpharmacologic treatment in 1. SSRIs were investigated in 12 studies, and were compared with placebo in 7 studies, with another antidepressant in 3, and with nonpharmacologic treatment in 2. Three RCTs studied SNRIs, 3 studied an MAOI, and 2 studied drug combinations (TCAs and SSRIs, and TCAs and naproxen). Nineteen (73%) trials allowed the additional use of acetaminophen, and 8 (31%) studies allowed the use of acetaminophen with acetylsalicylic acid or nonsteroidal antiinflammatory drugs or codeine. Only 1 (4%) study explicitly prohibited comedication. Six (23%) studies reported a defined dose of the allowed comedication. No study controlled the effects reported for the amount of comedication. The median percentage of patients completing the studies was 81% (range 28–97%) in the treatment groups (including studies with and without a placebo group) and 85% (range 33–100%) in the control groups. The median duration of the studies was 8 weeks (range 2–26 weeks).

Table 1. Study design, study population, methods, and interventions*
Author (ref.), year, country, diagnosis criteriaPatient recruitmentStudy designStudy populationPlacebo groupTreatment groupDuration washout; duration treatment; drugs and dosageNo. of doctor visits; permitted analgesic rescue medicationMean age; mean disease duration; women; white; working status; sick leave
No. screened/randomized (%)Total no./ completing (%)Total no./ completing (%)Total no./ completing (%)
  • *

    TCA = tricyclic/tetracyclic antidepressant; ACR = American College of Rheumatology; RCT = randomized controlled trial; NR = not reported; wks = weeks; yrs = years; FMS = fibromyalgia syndrome; NA = not applicable; SSRI = selective serotonin reuptake inhibitor; TENS = transcutaneous electrical nerve stimulation; NSAIDs = nonsteroidal antiinflammatory drugs; ASA = acetylsalicylic acid; MAOI = monoamine oxidase inhibitors; SNRI = serotonin and noradrenaline reuptake inhibitor.

  • Trials are mentioned more than once if different antidepressants were used in the same trial.

  • No nonpharmacologic cotherapy.

  • §

    Acupuncture and massage not allowed; physical training, warm water baths, and TENS allowed.

  • Both groups received 15 sessions of electromyogram biofeedback.

TCA         
 Amitriptyline         
  Ataoğlu (32), 1997, Turkey, ACR (2)Dept. of Physical Med. and Rehab., single centerRCT, parallelNR68/61 (89.7)No placebo groupParoxetine: 34/32 (94.1), amitriptyline: 34/29 (85.3)2 wks; 6 wks; paroxetine 20 mg/day; amitriptyline 50–100 mg/day4; NRNR; NR; 100%; NR; NR
  Çapacı (33), 2002, Turkey, ACR (2)Dept. of Physical Med. and Rehab., single centerRCT, parallelNRNRNo placebo groupParoxetine: 20/NR, amitriptyline: 20/NRNR; 8 wks; paroxetine 20–40 mg/day; amitriptyline 10–20 mg/day4; NRParoxetine: 45.7 yrs, amitriptyline: 42.1 yrs; NR; 90%; NR; NR
  Carette (44), 1986, Canada, Smythe (21)Dept. of Rheumatol., multicenterRCT, parallelNR70/59 (84.3)36/32 (88.9)34/27 (79.4)3 wks; 9 wks; amitriptyline 10–50 mg/day3; acetaminophen41.8 yrs; 5.9 yrs; 92.6%; NR; NR
  Carette (34), 1994, Canada, ACR (2)Dept. of Rheumatol. and private practices, multicenterRCT, parallelNR208/52 (25)42/NRAmitriptyline: 84/NR, cyclobenzaprine: 82/NR3 wks; 24 wks; amitriptyline 10–50 mg/day; cyclobenzaprine 10–30 mg/day7; acetaminophen44.1 yrs; 5 yrs; 92.9%; NR; 52.4% working, 25% not at work because of FMS, 22.6% not at work for other reasons
  Carette (35), 1995, Canada, ACR (2)Dept. of Rheumatol., single centerRCT, crossoverNR22/20 (90.9)22/20 (90.9)22/20 (90.9)2–4 wks; 2 × 8 wks; amitriptyline 25 mg/day3; acetaminophen43.8 yrs; 6.8 yrs; 95.5.%; NR; 27.3% at work, 22.7% not at work due to FMS, 50.0% not at work for other reasons
  Ginsberg (36), 1996, Belgium, ACR (2)Rheumatol. clinic, multicenterRCT, parallelNR51/46 (90.2)25/22 (88)26/24 (92.3)NR; 8 wks; sustained-release amitriptyline 25 mg/day3; acetaminophen46 yrs; 2.5 yrs; 83%; 92%; NR
  Goldenberg (38), 1986, US, Yunus (22)NR, NRRCT, parallelNR62/58 (93.5)16/14 (87.5)Amitriptyline + naproxen NR (3 dropouts), naproxen NR (1 dropout), amitriptyline NR (1 dropout)3 days; 6 wks; amitriptyline 25 mg/day + naproxen 1,000 mg/day, naproxen 1,000 mg/day, amitriptyline 25 mg/day4; acetaminophen up to 6 × 1,300 mg/day43.8 yrs; 3.5 yrs; 95%; 87.1%; NR
  Goldenberg (37), 1996, US, ACR (2)Rheumatol. center, single centerRCT, crossoverNR31/19 (61.3)NA31/19 (61.3)2 wks; 4 × 6 wks each; amitriptyline 25 mg/day, fluoxetine 20 mg/day, amitriptyline 25 mg/day + fluoxetine 20 mg/day8; acetaminophen43.2 yrs; 6.1 yrs; 90%; 100%; 61.3% working, 35.5% disabled, 3.2% retired
  Gür (39), 2002, Turkey, ACR (2)Dept. of Physical Therapy and Rehab., single centerRCT, parallelNR75/NRNo pharmacol. placebo group (only for laser therapy)25/NR4 wks; 8 wks; amitriptyline 10 mg/dayNR; NR30.1 yrs; 4.4 yrs; 84%; NR; 12% unemployed, 8% employed, 4% retired, 44% homemaker, 24% student, 8% other
  Hannonen (40), 1998, Finland, ACR (2)Rheumatol. centers, multicenterRCT, parallel184/130 (70.6)130/92 (70.8)45/30 (66.7)Moclobemide: 43/30 (69.8); Amitriptyline: 42/32 (76.2)NR; 12 wks; moclobemide 150 mg/day, amitriptyline 12.5 mg/dayNR; acetaminophen up to 4 gm/day49.7 yrs; 8.2 yrs; 100%; NR; NR
  Heymann (41), 2001, Brazil, ACR (2)Rheumatol. center, single centerRCT, parallelNR118/106 (89.8)40/33 (82.5)Amitriptyline: 40/37 (92.5); nortriptyline: 38/36 (94.7)4 wks; 8 wks; amitriptyline 25 mg/day, nortriptyline 25 mg/day2; acetaminophen53.4 yrs; NR; 100%; 65%; NR
  Özerbil (42), 2006, Turkey, ACR (2)Dept. of Rheumatology, single centerRCT, crossoverNR15/NRNRNR2 wks; 2 wks; amitriptyline 25 mg/day, fluoxetine 20 mg/day3; NRNR; NR; NR; NR; NR
  Scudds (43), 1989, Canada, Smythe (21)Dept. of Rheumatology, single centerRCT, crossoverNRNR/3939/36 (92.3)39/36 (92.3)3 wks; 10 wks (4-2-4); amitriptyline 10–50 mg/day4; acetaminophen39.9 yrs; 5.1 yrs; 88.9%; NR; NR
Nortriptyline         
  Heymann (41), 2001, Brazil, ACR (2)Rheumatol. center, single centerRCT, parallelNR118/106 (89.8)40/33 (82.5)Amitriptyline: 40/37 (92.5), nortriptyline: 38/36 (94.7)4 wks; 8 wks; amitriptyline 25 mg/day, nortriptyline 25 mg/day2; acetaminophen53.4 yrs; NR; 100%; 65%; NR
SSRI         
 Paroxetine         
  Ataoğlu (32), 1997, Turkey, ACR (2)Dept. of Physical Med. and Rehab., single centerRCT, parallelNR68/61 (89.7)No placebo groupParoxetine: 34/32 (94.1), amitriptyline: 34/29 (85.3)2 wks; 6 wks; paroxetine 20 mg/day, amitriptyline 50–100 mg/day4; NRNR; NR; 100%; NR; NR
  Çapacı (33), 2002, Turkey, ACR (2)Dept. of Physical Med. and Rehab., single centerRCT, parallelNRNRNo placebo groupParoxetine: 20/NR, amitriptyline: 20/NRNR; 8 wks; paroxetine 20–40 mg/day, amitriptyline 10–20 mg/day4; NRParoxetine: 45.7 yrs, amitriptyline: 42.1 yrs; NR; 90%; NR; NR
  Giordano (45), 1999, Italy, ACRDept. of Physical Therapy, single centerRCT (single blind), parallelNR40/29 (72.5)20/12 (60)20/17 (85)NR; 12 wks; paroxetine 20 mg/day7; none31 yrs; NR; 100%; NR; NR
  Patkar (47), 2007, US, ACR (2)Dept. of Psychiatry, single centerRCT, parallel983/116 (11.8)116/86 (74.1)58/48 (82.8)58/38 (65.5)7 days; 12 wks; paroxetine controlled release 12.5– 62.5 mg/day (mean 39.1 mg/day)12; acetaminophen up to 4 gm/day; ibuprofen up to 1.2 gm/day47.9 yrs; 49% >5 yrs; 94%; NR; NR
  Şencan (46), 2004, Turkey, NRDept. of Rheumatology, single centerRCT, parallelNR67/60 (89.6)NRNRNR; 6 wks; aerobic exercise 3 times per wk, paroxetine 20 mg/day, placebo TENS 3 times per wkNR; acetaminophenNR; 4.7 yrs; 100%; NR; NR
 Fluoxetine         
  Arnold (48), 2002, US, ACR (2)Rheumatol. practices and ads, single centerRCT, parallelNR60/37 (61.7)30/18 (60)30/19 (63.3)NR; 12 wks; fluoxetine 20–80 mg/day5; acetaminophen and NSAIDs46 yrs; 11 yrs; 100%; 90%; partially reported: 67% working, 3% on disability for FMS in the verum group; 73% working, 3% on disability for FMS in the placebo group
  Goldenberg (37), 1996, US, ACR (2)Rheumatol. center, single centerRCT, crossoverNR31/19 (61.3)NA31/19 (61.3)2 wks; 4 × 6 wks each; amitriptyline 25 mg/day, fluoxetine 20 mg/day, amitriptyline 25 mg/day + fluoxetine 20 mg/day8; acetaminophen43.2 yrs; 6.1 yrs; 90%; 100%; 61.3% working, 35.5% disabled, 3.2% retired
  Özerbil (42), 2006, Turkey, ACR (2)Dept. of Rheumatology, single centerRCT, crossoverNR15/NRNRNR2 wks; 2 wks; amitriptyline 25 mg/day, fluoxetine 20 mg/day3; NRNR; NR; NR; NR; NR
  Wolfe (49), 1994, US, ACR (2)Dept. of Rheumatology, single centerRCT, parallelNR42/24 (57.1)21/9 (42.8)21/15 (71.4)1 wk; 6 wks; fluoxetine 20 mg/day6; NSAIDs, acetaminophen48 yrs; 16.1 yrs; 100%; 100%; NR
Citalopram         
  Anderberg (50), 2000, Sweden, ACR (2)Rheumatol. and Rehab. clinics, single centerRCT, parallelNR40/35 (87.5)19/18 (95)21/17 (81)NR; 16 wks; citalopram 20–40 mg/day6; acetaminophen 1 gm/day or ASA 2 gm/day, exceptionally stronger analgesics §48.6 yrs; 11.9 yrs; 100%; NR; most patients employed in service businesses or schools, mostly part time; many received part or full disablement pensions; others on part-time or full-time sick leave
  Nørregaard (51), 1995, Denmark, ACR (2)Dept. of Rheumatology, single centerRCT, parallel150/42 (28)42/33 (78.6)21/21 (100)21/12 (57.1)2 wks; 8 wks; citalopram 20–40 mg/day2; codeine, NSAIDs, acetaminophen48 yrs; 10 yrs; NR; NR; NR
Sertraline         
  Gonzalez-Viejo (52), 2005, France, ACR (2)Rehab. center in general hospital, single centerRCT, parallelNR70/NRNo placebo groupSertraline: 36/NR, ultrasonography and physical therapy: 34/NRNA; 6 months; sertraline 50 mg/day, 15 × ultrasonography and physical therapy3; NR47.5 yrs; NR; NR; NR; 2.9% secretary, 97.1% housewife
MAOI         
 Moclobemide         
  Hannonen (40), 1998, Finland, ACR (2)Rheumatol. centers, multicenterRCT, parallel184/130 (70.6)130/92 (70.8)45/30 (66.7)Moclobemide 43/30 (69.8), amitriptyline 42/32 (76.2)NR; 12 wks; moclobemide 150 mg/day, amitriptyline 12.5 mg/dayNR; acetaminophen up to 4 gm/day49.7 yrs; 8.2 yrs; 100%; NR; NR
  Yavuzer (53), 1998, Turkey, NRDept. of Rheumatology, single centerRCT, parallelNR60/53 (88.3)25/22 (88)28/26 (92.9)3 wks; 6 wks; moclobemide 300 mg/day3; NR33.2 yrs; NR; 58.3%; NR; NR
 Pirlindole         
  Ginsberg (54), 1998, Belgium, ACR (2)NR, multicenterRCT, parallel200/100 (50)100/61 (61)50/28 (56)50/33 (66)2 wks; pirlindole 150 mg/day3; acetaminophen39.7 yrs; 3.6 yrs; 87.9%; NR; NR
SNRI         
 Duloxetine         
  Arnold (55), 2004, US, ACR (2)Research centers, multicenterRCT, parallel555/207 (37.3)207/124 (59.9)103/66 (64.1)104/58 (55.7)1–4 wks; 12 wks; duloxetine 20–120 mg/day10; acetaminophen up to 2 gm/day, ASA up to 325 mg/day49.9 yrs; NR; 88.5%; 88.5%; NR
  Arnold (56), 2005, US, ACR (2)Research centers, multicenterRCT, parallel745/354 (47.5)354/215 (60.7)120/68 (56.7)234/147 (61.5)NR; 12 wks; duloxetine 60 mg or duloxetine 120 mg7; acetaminophen up to 2 gm/day, ASA up to 325 mg/day49.6 yrs; NR; 100%; 89.5%; NR
Milnacipran         
  Vitton (57), 2004, US, ACR (2)NR, multicenterRCT, parallel184/125 (67.9)125/90 (72)28/21 (75)97/69 (71.1)1–4 wks; 12 wks; milnacipran 25– 200 mg/day6; NSAIDs, acetaminophen46–48 yrs; 4–4.3 yrs; 96-98%; 79–89%; NR
Drug combinations         
  Goldenberg (38), 1986, US, Yunus (22)NR, NRRCT, parallelNR62/58 (93.5)16/14 (87.5)Amitriptyline + naproxen NR (3 dropouts), naproxen NR (1 dropout), amitriptyline NR (1 dropout)3 days; 6 wks; amitriptyline 25 mg/day + naproxen 1,000 mg/day, naproxen 1,000 mg/day, amitriptyline 25 mg/day4; acetaminophen up to 6 × 1,300 mg/day43.8 yrs; 3.5 yrs; 95%; 87.1%; NR
  Goldenberg (37), 1996, US, ACR (2)Rheumatol. center, single centerRCT, crossoverNR31/19 (61.3)NA31/19 (61.3)2 wks; 4 × 6 wks each; amitriptyline 25 mg/day, fluoxetine 20 mg/day, amitriptyline 25 mg/day + fluoxetine 20 mg/day8; acetaminophen43.2 yrs; 6.1 yrs; 90%; 100%; NR

Characteristics of the study population.

Twenty-two (85%) studies used the ACR criteria for the diagnosis of FMS (2). Eight (31%) studies excluded patients <18 years of age, 3 (11%) studies excluded patients >65 years. All studies excluded patients with additional somatic diseases and 12 (50%) studies excluded patients with mental disorders or those receiving psychopharmacologic treatment. Three (11%) studies excluded patients because of involvement in disability reviews. The median mean age was 46 years (range 30–53 years) and the median mean duration of symptoms was 6 years (range 4–16 years). The median of the mean percentage of women was 100% (range 58–100%), and the median of the mean percentage of white patients was 89% (65–100%). Only 4 (15%) studies provided data on comorbidities. The percentage of patients with current depressive disorder ranged from 9% to 36%. Seven (27%) studies reported data on working status, and 2 (8%) studies gave information on disability status due to FMS. The median percentage of working patients was 60% (range 52–67%), and the median percentage of patients not working or receiving pension because of FMS was 14% (range 3–25%) (Table 1).

Methodologic quality.

The methodologic quality of the studies varied considerably. Thirteen (50%) studies had a Jadad score ≥4. In 13 (50%) studies, a power calculation was performed. Nine (35%) studies performed an intent-to-treat analysis. Twelve (46%) studies used adequate multivariate statistical tests. One (4%) study adjusted for multiple testing. Seventeen (65%) studies provided data suitable for meta-analysis (Table 2).

Table 2. Study quality, outcome measures, and exclusion criteria*
Author (ref.), year, countryMethod/quality: 1) Jadad score (missing items); 2) power calculation; 3) intent-to-treat analysis; 4) ANOVA/ANCOVA; 5) adjustment multiple testing; 6) data suitable for meta-analysisAll outcome measuresEffects in main outcome measures; positive effects in all outcome measuresImprovement pain verum, %Side effects reported, rate of side effects, dropout due to side effectsExclusion criteria, comorbidity
  • *

    ANOVA = analysis of variance; ANCOVA = analysis of covariance; TCA = tricyclic/tetracyclic antidepressant; rand. meth. = method of randomization not described; blind meth. = method of blinding not described; dropout = dropouts and withdrawals not described; VAS = visual analog scale; HAM-D = Hamilton Depression Rating Scale; QOL = quality of life; NA = not assessed; NR = not reported; BDI = Beck Depression Inventory; FMS = fibromyalgia syndrome; SIP = Sickness Impact profile; HAQ = Health Assessment Questionnaire; AIMS = Arthritis Impact Measurement Scales; MMPI = Minnesota Multiphasic Personality Inventory; FIQ = Fibromyalgia Impact Questionnaire; yrs = years; wks = weeks; NRS = numeric rating scale; HDRS = Hamilton Depression Rating Scale; NHP = Nottingham Health Profile; VO2max = maximum oxygen intake; SSRI = selective serotonin reuptake inhibitor; HAM-A = Hamilton Anxiety Rating Scale; CGI-I = Clinical Global Impression-Improvement; CGI-S = Clinical Global Impression score; SDS = Sheehan Disability Scale; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition; MADRS = Montgomery Asberg Depression Rating Scale; MOS SF-36 = Medical Outcomes Study Short Form 36; MAOI = monoamine oxidase inhibitors; SCL-90-R = Symptom-Checklist-90-Revised; SNRI = serotonin and noradrenaline reuptake inhibitor; BPI = Brief Pain Inventory; BAI = Beck Anxiety Inventory; PGI = Patient Global Impression; QLDS = Quality of Life in Depression Scale.

  • Trials are mentioned more than once if different antidepressants were used in the same trial.

  • Significant positive effect.

  • §

    Assessment by physician.

  • No significant effect.

  • #

    Predefined primary outcome measure.

  • **

    Personal communication: Jadad score = 5; however, for all studies included, only data available in the publication were used for the calculation of the Jadad scores.

TCA      
 Amitriptyline      
  Ataoğlu (32), 1997, Turkey1) 2 (rand. meth.; blind meth.; dropout); 2) no; 3) no; 4) no; 5) no; 6) no SDVAS pain, VAS sleep, VAS fatigue, VAS general condition, HAM-D, tender point countPain, sleep, fatigue, depression, QOL NA; 6/6NRAmitriptyline: xerostomia, nausea, dizziness, sweating, constipation, vomiting, headache, sedation, urinary retention, fatigue; paroxetine: xerostomia, nausea, dizziness, sweating, constipation, headache, sedation, insomnia, fatigueCardiovascular, pulmonary, hepatic disease; glaucoma; pregnancy or breast feeding; earlier diagnosis of depression or antidepressant medication; inflammatory rheumatologic disease NR
     93.1% amitriptyline, 37.5% paroxetine 
     14.7% amitriptyline, 5.9% paroxetine 
  Çapacı (33), 2002, Turkey1) 2 (rand. meth.; blind meth.; dropout); 2) no; 3) no; 4) no; 5) yes; 6) no SDVAS fatigue§, VAS disturbed sleep§, VAS nonrefreshed sleep§, VAS morning stiffness§, VAS headache§, VAS paresthesia§, VAS global pain severity§, BDI depression, tender point count, tender point scoreParoxetine: pain, fatigue, sleep, depression, QOL NA; 9/10 Amitriptyline: pain, fatigue, sleep, depression, QOL NA; 9/10NRAmitriptyline: sleepiness, xerostomia, constipation, rash; paroxetine: sleepiness, diarrhea, nausea, dizziness, sexual dysfunction Amitriptyline: up to 30%, paroxetine: up to 40% Amitriptyline: 0%, paroxetine: NRNR NR
  Carette (44), 1986, Canada1) 4 (rand. meth.); 2) yes; 3) no; 4) no; 5) no; 6) yesVAS morning stiffness, VAS pain#, VAS sleep quality, overall assessment of disease, fibrositic point tenderness, total myalgic score#Pain, fatigue NA, sleep, depression NA, QOL NA; 2/631.8Drowsiness, xerostomia 70% amitriptyline, 12% placebo 5.9% amitriptyline, 5.6% placeboAmitriptyline treatment within the preceding year; previous hypersensitivity to amitriptyline; history of glaucoma, urinary retention, ischemic heart disease, arrhythmia, congestive heart failure
      NR
  Carette (34), 1994, Canada1) 5; 2) yes; 3) yes; 4) yes; 5) no; 6) no SDVAS pain, VAS fatigue, VAS sleep, VAS feeling on awakening, VAS morning stiffness, VAS global assessment of FMS, McGill Pain Questionnaire, SIP overall, HAQ, QOL HAQ disability, AIMS depression, AIMS anxiety, MMPI, point tenderness, patient global evaluation, physician global evaluation, total myalgic scorePain, fatigue, sleep, depression, QOL; 5/630Amitriptyline: somnolence, abdominal pain, rash, weight gain; cyclobenzaprine: somnolence, dizziness, abdominal pain, rash, headache 95% amitriptyline, 4.8% placebo 6% amitriptyline, 5% placebo<18 yrs; inflammatory rheumatoid disorder; untreated endocrine, neurologic, infectious, or osseous disorder; previous treatment with amitriptyline or cyclobenzaprine; history of glaucoma, urinary retention, ischemic heart disease, arrhythmia, congestive heart failure NR
  Carette (35), 1995, Canada1) 4 (rand. meth.); 2) no; 3) no; 4) yes; 5) no; 6) yesVAS pain, VAS fatigue, VAS global assessment, VAS sleep, myalgic scorePain, fatigue, sleep, depression NA, QOL NA; 4/528.8NR NR NRNeurologic, infectious, muscular, endocrine, osseous, or inflammatory rheumatoid disorder; history of glaucoma, urinary retention, ischemic heart disease, arrhythmia, congestive heart failure NR
  Ginsberg (36), 1996, Belgium1) 4 (rand. meth.); 2) no; 3) no; 4) yes; 5) no; 6) yesVAS global assessment, VAS pain, VAS sleep disturbance, VAS fatigue, duration of morning stiffness, tender point count, tender point scorePain, fatigue, sleep, depression NA, QOL NA; 9/930.6Xerostomia, digestive symptoms, vertigo, neuropsychic symptoms 29% 4.2%<18 yrs; neurologic, muscular, infectious, endocrine, osseous, inflammatory rheumatic disease; history of glaucoma, urinary retention, myocardial infarction, arrhythmia, congestive heart failure, sleep apnea syndrome NR
  Goldenberg (38), 1986, US1) 3 (rand. meth.; blind meth.); 2) yes; 3) no; 4) no; 5) no; 6) no SDVAS global FMS symptoms, VAS pain, VAS stiffness, VAS fatigue, VAS sleep, tender point scorePain, fatigue, sleep, depression NA, QOL NA; 7/7Amitriptyline 27.8Amitriptyline: xerostomia; naproxen: dyspepsia, diarrhea 12.9% of all patients NRHistory of peptic ulcer, cardiac arrhythmia; inability to be withdrawn from analgesics, antiinflammatory drugs, antidepressants, sleep medication, any other central nervous system–active medication
      NR
  Goldenberg (37), 1996, US1) 5; 2) yes; 3) no; 4) no; 5) no; 6) yesFIQ, VAS pain, VAS global, VAS sleep, BDI, physician VAS, VAS fatigue, VAS refreshed, tender point scoreAmitriptyline; fluoxetine and combination; combination better than monotherapy: pain, fatigue, sleep, depression, QOL; 5/8Amitriptyline 5.9, fluoxetine 14.7, amitriptyline + fluoxetine 36.8NR NR 3.2% fluoxetine, 9.7% fluoxetine + amitriptylineNo current or past history of systemic illness including cardiac, kidney, hematologic, or liver disease; age <18 or >60 yrs; refusal to discontinue all central nervous system–active medication, antiinflammatory drugs, and analgesics other than acetaminophen; major depression NR
  Gür (39), 2002, Turkey1) 2 (rand. meth,: blind meth.; dropout); 2) no; 3) no; 4) yes; 5) no; 6) no pharmacologic placeboNRS pain, NRS sleep disturbance, NRS fatigue, NRS muscle spasm, NRS morning stiffness, NRS skinfold tenderness, HDRS depression, FIQ QOL, tender point countPain, fatigue, sleep, depression, QOL; 8/927.9NR NRRecent or past history of psychiatric disorder, e.g., major depression, alcohol dependence, substance abuse, schizophrenic or paranoid disorder, personality disorders, somatoform disorders; immune deficiency, neurologic, inflammatory, liver, endocrine, or inflammatory diseases; pregnancy NR
  Hannonen (40), 1998, Finland1) 5; 2) yes; 3) yes; 4) yes; 5) no; 6) yesClinical impression of the severity of FM§,#; VAS disability, VAS pain, VAS fatigue, VAS sleep, QOL NHP socialMoclobemide: pain, fatigue, sleep, depression NA, QOL; 4/6 Amitriptyline: pain, fatigue, sleep, depression NA, QOL; 6/6Moclobemide 21, amitriptyline 25Amitriptyline: xerostomia, fatigue; moclobemide: headache, difficulty falling asleep 77% moclobemide, 74% amitriptyline, 80% placebo NR<18 or >65 yrs; severe cardiovascular, pulmonary, hepatic, hematologic, or renal disease; glaucoma; pregnancy; lactation; concomitant use of other psychopharmacologic agents; major depression, drug abuse, psychosis, obsessive-compulsive disorder; excessive alcohol consumption
      NR
  Heymann (41), 2001, Brazil1) 5; 2) yes; 3) no; 4) no; 5) no; 6) yesFIQ QOL, tender point countPain NA, fatigue NA, sleep NA, depression NA, QOL; 2/2NRAmitriptyline: xerostomia, abdominal pain, mouth taste bitter, sleepiness, dizziness, nausea, weight gain, apathy; nortriptyline: xerostomia, abdominal pain, mouth taste bitter, sleepiness, dizziness, nausea, apathy, palpitations, sweating, migraine, memory deficit, diffuse pain 40% amitriptyline, 81.6% nortriptyline, 62.5% placebo NRMales; <18 years; concomitant use of other psychopharmacologic agents; pregnancy; start of physiotherapy concomitant to study initiation; arrhythmia, heart, renal, hepatic disease; glaucoma; urinary retention; hyperthyroidism; chronic inflammatory disease NR
  Özerbil (42), 2006, Turkey1) 4 (dropout); 2) no; 3) no; 4) no; 5) no; 6) outcomes not comparable with other studiesBicycle peak power, bicycle mean power, bicycle fatigue index, bicycle VO2max, bicycle ventilatory anaerobic thresholdPain NA, fatigue (both drugs), sleep NA, depression NA, QOL NA; 3/5; 2/5NRNR NRSystemic lupus erythematosus; rheumatoid arthritis; ankylosing spondylitis, polymyalgia rheumatica, hypothyroidism, inflammatory myopathy, and neuropathy; no current or past history of systemic illness; <20 or >60 yrs; refusal to discontinue all central nervous system–active medication for at least 1 month prior to the study NR
  Scudds (43), 1989, Canada1) 4 (rand. meth.); 2) yes; 3) no; 4) yes; 5) no; 6) yesMcGill Pain Questionnaire, dolorimeter total myalgic score, dolorimeter pain threshold, dolorimeter pain tolerancePain, fatigue NA, sleep NA, depression NA, QOL NA; 2/4NRNR NR Amitriptyline 5.3%, placebo 5.9%History of urinary retention, glaucoma and ischemic heart disease, cardiac arrhythmia and congestive cardiac failure; treatment with amitriptyline within the previous year NR
 Nortriptyline      
  Heymann (41), 2001, Brazil1) 5; 2) yes; 3) no; 4) no; 5) no; 6) yesFIQ QOL, tender point countPain NA, fatigue NA, sleep NA, depression NA, QOL; 2/2NRNortriptyline: xerostomia, abdominal pain, mouth taste bitter, sleepiness, dizziness, nausea, apathy, palpitations, sweating, migraine, memory deficit, diffuse pain; amitriptyline: xerostomia, abdominal pain, mouth taste bitter, sleepiness, dizziness, nausea, weight gain, apathyMales; <18 yrs; concomitant use of other psychopharmacologic agents; pregnancy; start of physiotherapy concomitant to study initiation; arrhythmia, heart, renal, hepatic disease; glaucoma; urinary retention; hyperthyroidism; chronic inflammatory disease NR
     81.6% nortriptyline, 40% amitriptyline, 62.5% placebo 
     NR 
SSRI      
 Paroxetine      
  Ataoğlu (32), 1997, Turkey1) 2 (rand. meth.; blind meth.; dropout); 2) no; 3) no; 4) no; 5) no; 6) no SDVAS pain, VAS sleep, VAS fatigue, VAS general condition, HAM-D, tender point countPain, sleep, fatigue, depression, QOL NA; 6/6NRParoxetine: xerostomia, nausea, dizziness, sweating, constipation, headache, sedation, insomnia, fatigue; amitriptyline: xerostomia, nausea, dizziness, sweating, constipation, vomiting, headache, sedation,Cardiovascular, pulmonary, hepatic disease; glaucoma; pregnancy or breastfeeding; earlier diagnosis of depression or antidepressant medication; inflammatory rheumatologic disease NR
      urinary retention, fatigue 37.5% paroxetine, 93.1%  amitriptyline 5.9% paroxetine, 14.7%  amitriptyline 
  Çapacı (33), 2002, Turkey1) 2 (rand. meth.; blind meth.; dropout); 2) no; 3) no; 4) no; 5) yes; 6) no SDVAS fatigue§, VAS disturbed sleep§, VAS nonrefreshed sleep§, VAS morning stiffness§, VAS headache§, VAS paresthesia§, VAS global pain severity§, BDI depression, tender point count, tender point scoreParoxetine: pain, fatigue, sleep, depression, QOL NA; 9/10 Amitriptyline: pain, fatigue, sleep, depression, QOL NA; 9/10NRParoxetine: sleepiness, diarrhea, nausea, dizziness, sexual dysfunction; amitriptyline: sleepiness, xerostomia, constipation, rash Paroxetine: up to 40%; amitriptyline: up to 30% Paroxetine: NR;  amitriptyline: 0%NR NR
  Giordano (45), 1999, Italy1) 3 (rand. meth.; blind meth.); 2) no; 3) no; 4) no; 5) no; 6) no SDHAM-A, HAM-D, tender point count, tender point score, VAS pain, VAS morning stiffness, VAS mood, VAS sleep disorder, VAS inclination to work, VAS capacity to work, VAS overall condition, VAS efficacy of treatment, VAS tolerability of treatmentPain, fatigue, sleep, depression, QOL NA; 8/8NRNausea, diarrhea, malaise, xerostomia, dyspepsia, headache, sweating, insomnia, palpitation, drowsiness, reduced libido, anxiety 50% paroxetine 15% paroxetine; 0%  placeboSystemic or organic diseases; secondary FMS, inflammatory rheumatoid disorder, pregnancy or breastfeeding NR
  Patkar (47), 2007, US1) 5; 2) yes; 3) yes; 4) yes; 5) no; 6) yesFIQ, CGI-I, CGI-S, SDS, VAS pain, tender point count, tender point indexPain, fatigue, sleep NA, depression, QOL; 2/7NRDrowsiness, dry mouth, female genital disorders, ejaculatory problems, impotence, headaches, sleeplessness, anxiety, nausea, diarrhea, blurred vision In 5% of all patients  (placebo and  paroxetine)Inflammatory disease, unstable medical diseases, psychotic disorders, current depressive or anxiety disorders, substance abuse in the previous 12 months, history of hypersensitivity to paroxetine or paroxetine controlled release, involvement in workers compensation or related litigation, or pregnancy; concomitant medication exclusion included psychotropics, analgesics, muscle relaxants, steroids, and hypnotics NR
  Şencan (46), 2004, Turkey1) 1 (rand. meth.; blind meth.; single blind); 2) no; 3) no; 4) yes; 5) no; 6) no pharmacologic placeboVAS pain, BDI depression, analgesic consumption patients' notes, dolorimeter pain thresholdPain, fatigue NA, sleep NA, depression, QOL NA; 4/451.7NR NR NRTumoral, infectious, metabolic, cardiovascular, endocrine diseases; drug dependency NR
 Fluoxetine      
  Arnold (48), 2002, US1) 4 (rand. meth.); 2) yes; 3) yes; 4) no; 5) no; 6) yesFIQ total and pain score#, FIQ subscores, McGill Pain Questionnaire, pain rating index, tender point count, myalgic scorePain, fatigue, sleep NA, depression, QOL (days felt good); 5/1429.5Nervousness, headache, insomnia, nausea, gastrointestinal reflux, increased/decreased appetite, weight loss, fatigue, sedation, anorgasmia, weakness, decreased coordination and concentration, migraine, worsening pain NR Not differentiated between  verum and placebo<18 yrs; traumatic injury; inflammatory rheumatic diseases; infectious or endocrine-related arthropathy; unstable medical illness; history of seizures, head trauma or stroke; any current mental disorder on DSM-IV axis I; history of hypomania, mania, psychosis, dementia; alcohol or substance dependence during last 6 months; substantial risk of suicide; score ≥10 on 17-item HDRS; antidepressant treatment 2 wks before randomization; previous fluoxetine treatment for FMS NR
  Goldenberg (37), 1996, US1) 5; 2) yes; 3) no; 4) no; 5) no; 6) yesFIQ, VAS pain, VAS global, VAS sleep, BDI, physician VAS, VAS fatigue, VAS refreshed, tender point scoreAmitriptyline; fluoxetine and combination; combination better than monotherapy: pain, fatigue, sleep, depression, QOL; 5/8Amitriptyline 5.9, fluoxetine 14.7, amitriptyline + fluoxetine 36.8NR NR 3.2% fluoxetine; 9.7% fluoxetine + amitriptylineNo current or past history of systemic illness including cardiac, kidney, hematologic, or liver disease; age <18 or >60 yrs; refusal to discontinue all central nervous system–active medication, antiinflammatory drugs, and analgesics other than acetaminophen; major depression NR
  Özerbil (42), 2006, Turkey1) 4 (dropout); 2) no; 3) no; 4) no; 5) no; 6) outcomes not comparable with other studiesBicycle peak power, bicycle mean power, bicycle fatigue index, bicycle VO2max, bicycle ventilatory anaerobic thresholdPain NA, fatigue (both drugs), sleep NA, depression NA, QOL NA; 3/5; 2/5NRNR NRSystemic lupus erythematosus; rheumatoid arthritis; ankylosing spondylitis, polymyalgia rheumatica, hypothyroidism, inflammatory myopathy, and neuropathy; no current or past history of systemic illness; <20 or >60 yrs; refusal to discontinue all central nervous system–active medication for at least 1 month prior to the study
      NR
  Wolfe (49), 1994, US1) 3 (blind meth; dropout); 2) no; 3) yes; 4) no; 5) no; 6) yesVAS sleep, VAS fatigue, hours morning stiffness, VAS pain, VAS global severity, HAQ, AIMS anxiety, AIMS depression, BDI depression, tender point count, tender point score, dolorimetry scorePain, fatigue, sleep, depression, QOL NA; 3/135.8Nausea, rash, headache, bruising Fluoxetine: 81%; placebo:  100% Fluoxetine: 28.6%;  placebo: 57.1%No other significant rheumatic disease NR
 Citalopram      
  Anderberg (50), 2000, Sweden1) 4 (blind meth.; dropout); 2) no; 3) yes; 4) no; 5) NR; 6) yesVAS: pain; FIQ: physical symptoms and functioning; MADRS: psychological symptomsPain, fatigue, sleep, depression, QOL NR; 3/2112.6Xerostomia, nausea, fatigue, insomnia, headache, vertigo, tremor, sweating, sexual side effects, weight gain 100% citalopram; 47.4% 14.3% citalopram; 0%  placeboSevere heart disease (i.e., angina pectoris, myocardial infarction), brain infarction, suicidality, severe depression, psychopharmacotherapy or strong analgesics NR
  Nørregaard (51), 1995, Denmark1) 4 (rand. meth.); 2) yes; 3) no; 4) no; 5) no; 6) yesProportion of responders§; NRS pain, NRS fatigue, NRS sleep, NRS general condition, BDI depression, FIQ physical function, tender point count, muscle strengthPain, fatigue, sleep, depression, QOL; 0/815.9Xerostomia, nausea, vomiting, headache 100% citalopram; 100% 21.4% (not differentiated between verum and placebo)Cardiovascular, pulmonary or hepatic diseases, glaucoma, pregnancy or lactation, earlier diagnosis of endogenous depression or medication with antidepressants or MAOIs; inflammatory rheumatologic disease or other medical diseases 9.5% rheumatic diagnosis; 4.8%  other diagnosis
 Sertraline      
  Gonzalez-Viejo (52), 2005, France1) 2 (rand. meth.; blind meth.); 2) no; 3) no; 4) no; 5) no; 6) no pharmacologic placeboVAS pain, VAS sleep, VAS morning stiffness, MOS SF-36 sleepPain, fatigue NA, sleep, depression NA, QOL NA; 4/458.3NR NR NRHypertension, pregnancy, breastfeeding, use of antidepressants in last 4 wks NR
MAOI      
 Moclobemide      
  Hannonen (40), 1998, Finland1) 5; 2) yes; 3) yes; 4) yes; 5) no; 6) yesClinical impression of the severity of FM§,#; VAS disability, VAS pain, VAS fatigue, VAS sleep, QOL NHP socialMoclobemide: pain, fatigue, sleep, depression NA, QOL; 4/6 Amitriptyline: pain, fatigue, sleep, depression NA, QOL; 6/6Moclobemide 21, amitriptyline 25Moclobemide: headache, difficulty falling asleep; amitriptyline: xerostomia, fatigue 77% moclobemide; 74% amitriptyline; 80% placebo NR<18 or >65 yrs; severe cardiovascular, pulmonary, hepatic, hematologic, or renal disease; glaucoma; pregnancy; lactation; concomitant use of other psychopharmacologic agents; major depression, drug abuse, psychosis, obsessive-compulsive disorder; excessive alcohol consuption NR
  Yavuzer (53), 1998, Turkey1) 1 (rand. meth.; blind meth.; single blind); 2) no; 3) no; 4) no; 5) no; 6) yesNRS pain score, HDRS depression, minutes morning stiffness, tender point countPain, fatigue NA, sleep NA, depression, QOL NA; 4/443.7Hair loss, loss of libido Moclobemide: 14.3%; placebo: 8% NRHistory of traumatic, neurologic, muscular, infectious, osseous, endocrine, or other rheumatic conditions NR
 Pirlindole      
  Ginsberg (54), 1998, Belgium1) 3 (rand. meth.; blind meth.; dropout); 2) no; 3) no; 4) yes; 5) no; 6) yesVAS pain, morning stiffness, tender point score, SCL-90-R, VAS global evaluation investigator, VAS global evaluation patient, VAS fatigue, VAS sleep disturbancePain, fatigue, sleep, depression NA, QOL NA; 5/833.9Depression, dizziness, xerostomia, gastric discomfort, headache, increased appetite, insomnia, irritable bowel, nausea/ vomiting, pain increase, palpitation, sleep during the day, vertigo 13.3%<18 yrs, >75 yrs; inability to give informed consent, pregnancy or lactation, inability to discontinue antidepressants, sleeping medication, antiinflammatory drugs, muscle relaxants, tranquilizers and/or any other central nervous system medication, severe cardiac disease, any other disease sufficient to produce clinical problems during the study, and biochemical or hematologic abnormality clinically significant and likely to interfere with the interpretation of the results NR
SNRI      
 Duloxetine      
  Arnold (55), 2004, US1) 5; 2) yes; 3) yes; 4) yes; 5) no; 6) yesFIQ total and pain score#, FIQ fatigue, FIQ morning tiredness, FIQ stiffness, BPI pain, BPI interference, BDI II depression, BAI anxiety, CGI severity, PGI improvement, tender point count, meanPain, fatigue, sleep, depression, QOL; 17/2427.8Insomnia, xerostomia, constipation 90.4% duloxetine; 74.8% placebo, 17.3% duloxetine; 10.7% placebo<18 yrs; secondary FMS; pain from traumatic injury or structural or regional rheumatic disease; rheumatoid arthritis, inflammatory arthritis, autoimmune disease; unstable medical or psychiatric illness; current dysthymia, which is more
   tender point pain threshold, MOS SF-36, QLDS, SDS    treatment resistant than major depression, primary psychiatric disorder other than major depressive disorder; substance abuse in the last year; history of psychosis; pregnancy or breastfeeding; unacceptable contraception in
       those of childbearing age; involvement in disability reviews that might compromise treatment response; use of an investigational drug within 30 days; prior participation in a study of duloxetine; severe allergic reactions to multiple medications; intolerance to >3 psychoactive drugs or >1 SSRI; and failure to respond to >2 adequate regimens of 2 different classes of antidepressants for depression or FMS Partially reported: major  depression: 35.6%
  Arnold (56), 2005, US1) 3**; 2) yes; 3) yes; 4) yes; 5) no; 6) yesBPI pain severity#, BPI interference, FIQ total score, mean tender point pain threshold, CGI severity, PGI improvement, HDRS depression, MOS SF-36, SDS, QLDSBoth treatment groups: pain, fatigue NA, sleep, depression, QOL; 16/18 duloxetine 60 mg, 17/18 duloxetine 120 mg43.8Dizziness, insomnia, headache, myalgia, nausea, diarrhea, abnormal dreams, depression, anxiety, and emotional disorder; fatigue, depression, arthralgia 92.4% duloxetine 60 mg; 90.5% duloxetine 120 mg; 79.2% placebo 22.2% duloxetine; 11.7% placebo<18 yrs; secondary FMS; pain from traumatic injury or structural or regional rheumatic disease; rheumatoid arthritis, inflammatory arthritis, or autoimmune disease; unstable medical or psychiatric diagnosis other than major depressive disorder, primary anxiety disorder within the past year (specific phobias allowed); substance abuse within the past year; serious suicide risk; pregnancy or breastfeeding; women who, in the investigator's opinion, were treatment refractory or may have been involved in disability reviews that might compromise treatment response; severe allergic reactions to multiple medications; or prior participation in a study of duloxetine Partially reported: 26% current  major depressive disorder
 Milnacipran      
  Vitton (57), 2004, US1) 4 (rand. meth.); 2) yes; 3) yes; 4) yes; 5) no; 6) yesVAS pain, e-diary pain#, Gracely pain intensity scale, Gracely patient global pain status, McGill Pain Questionnaire, present pain intensity, BDI depression, Sleep Jenkins Scale, FIQ QOL, MOS SF-36 QOL, Arizona sexual experience scalePain, fatigue NA, sleep, depression NR, QOL; only partially reported: 3/6NRNo serious side effects, 88% mild or moderate side effects Milnacipran 24.6%;  placebo 4.7%Severe psychiatric illness; significant risk of suicide; alcohol or other drug abuse; history of significant cardiovascular, respiratory, endocrine, genitourinary, liver or kidney disease, autoimmune disease; systemic infection; cancer or current chemotherapy; significant sleep apnea; life expectancy <1 year; active peptic ulcer or inflammatory bowel disease Partially reported: depression 9%
Drug combinations      
  Goldenberg (38), 1986, US1) 3 (rand. meth.; blind meth.); 2) yes; 3) no; 4) no; 5) no; 6) no SDVAS global FMS symptoms, VAS pain, VAS stiffness, VAS fatigue, VAS sleep, tender point scorePain, fatigue, sleep, depression NA, QOL NA; 7/7Amitriptyline 27.8Amitriptyline: xerostomia; naproxen: dyspepsia, diarrhea 12.9% of all patients NRHistory of peptic ulcer, cardiac arrhythmia; inability to be withdrawn from analgesics, antiinflammatory drugs, antidepressants, sleep medication, any other central nervous system—active medication NR
  Goldenberg (37), 1996, US1) 5; 2) yes; 3) no; 4) no; 5) no; 6) yesFIQ, VAS pain, VAS global, VAS sleep, BDI, physician VAS, VAS fatigue, VAS refreshed, tender point scoreAmitriptyline; fluoxetine and combination; combination better than monotherapy: pain, fatigue, sleep, depression, QOL; 5/8Amitriptyline 5.9, fluoxetine 14.7, amitriptyline + fluoxetine 36.8NR NR 3.2% fluoxetine; 9.7% fluoxetine + amitriptylineNo current or past history of systemic illness including cardiac, kidney, hematologic, or liver disease; age <18 or >60 yrs; refusal to discontinue all central nervous system–active medication, antiinflammatory drugs, and analgesics other than acetaminophen; major depression NR

Outcomes (Table 2).

All antidepressants.

The main outcome variables reviewed were pain, fatigue, sleep, depressiveness, and quality of life. Sixteen (62%) studies assessed pain, fatigue, and sleep using a visual analog scale (VAS). Seven (27%) studies measured depressiveness using the Beck Depression Inventory. Ten (38%) studies investigated quality of life using the Fibromyalgia Impact Questionnaire. Significant improvements were reported as follows: 20 (83%) of 24 for pain; 13 (72%) of 18 for fatigue; 16 (84%) of 19 for sleep; 11 (73%) of 15 for depressiveness; and 8 (73%) of 11 for quality of life. Reduction of pain varied between 6% and 70% (median 29%) and improvement in quality of life varied between 13% and 70% (median 30%). The median percentage of all positive outcome measures was 83% (range 0–100%). Studies with a low (<4) Jadad score reported more positive outcomes than studies with a Jadad score ≥4 (z = −2.9, P = 0.003). The median percentage of reported side effects was 72% (range 14–100%) for antidepressants and 55% (range 5–100%) for placebo. The median dropout rate due to side effects or lack of effectiveness was 13% (range 0–29%) for antidepressants and 7% (range 0–57%) for placebo.

TCA.

Amitriptyline (10–100 mg) was assessed in 13 studies (32–44) and nortriptyline (25 mg/day) in 1 study (41). Reduction of pain intensity was found in 10 (91%) of 11 studies, reduction of fatigue in 8 (80%) of 10, improvement of sleep in 10 (100%) of 10, reduction of depressiveness in 3 (60%) of 5, and improvement of quality of life in 3 (75%) of 4. The study on nortriptyline failed to show an advantage over placebo (41).

SSRI.

Paroxetine (20–62.5 mg/day) was used in 5 studies (32, 33, 45–47), fluoxetine (20–60 mg/day) in 4 (37, 42, 48, 49), citalopram (20–40 mg) in 2 (50, 51), and sertraline (50 mg/day) in 1 (52). Ten (83%) of 12 studies reported an improvement in pain intensity, 5 (50%) of 10 in fatigue, 6 (75%) of 8 in sleep, 7 (70%) of 10 in depressiveness, and 3 (75%) of 4 in quality of life. Two studies failed to show a significant analgesic effect of citalopram 20–40 mg/day versus placebo: a Swedish RCT that investigated 40 patients and a Norwegian RCT that investigated 22 patients for 4 weeks (50, 51).

MAOI.

Moclobemide (300–600 mg/day) was used in 2 studies (40, 53), and pirlindole in one study (54). Reduction in pain was found in 3 (100%) of 3 studies and reduction in depressiveness in 1 (33%) of 3 studies. One (50%) of 2 studies found an improvement in sleep. Quality of life was assessed in 1 study with a negative result.

SNRI.

SNRIs were investigated in 3 studies. Two studies used duloxetine 60 mg/day and 120 mg/day (55, 56) and 1 study used milnacipran 200 mg/day (57). Three (100%) of 3 studies reported pain reduction, 2 (67%) of 3 reported improvement in sleep, 2 (100%) of 2 reported improvement in depressiveness, and 2 (67%) of 3 reported improvement in quality of life. One study found no reduction in fatigue.

Comparisons of antidepressants and drug combinations.

There were no differences in the percentage of positive outcomes between the different classes of antidepressants (H = 1.1, P = 0.8). The percentage of positive outcomes did not differ concerning pain (χ2 = 2.7, P = 0.6), fatigue (χ2 = 3.7, P = 0.4), sleep (χ2 = 4.4, P = 0.3), depressiveness (χ2 = 1.5, P = 0.8), and quality of life (χ2 = 2.3, P = 0.7). Three Turkish RCTs investigated the effect of TCAs versus SSRIs and found conflicting results. In the first study with paroxetine 20 mg/day versus amitriptyline 100 mg/day over 6 weeks, paroxetine was superior to amitriptyline in pain reduction and drug tolerability (32). In the second study lasting 2 months, amitriptyline 20 mg/day was more effective in reducing pain, stiffness, and sleep disturbance than paroxetine 40 mg/day (33). A third trial using amitriptyline 25 mg/day and fluoxetine 20 mg/day over 2 weeks demonstrated similar results for both drugs regarding reduction of sleepiness and improvement in anaerobic exercise performance (42). One US crossover RCT analyzed the effect of fluoxetine 20 mg/day and amitriptyline 25 mg/day as monotherapies and in combination in 62 patients over 6 weeks. Both monotherapies and the combination of the drugs were superior to placebo; the combination of fluoxetine and amitriptyline was more effective than monotherapy (37). In another US RCT, the effect of amitriptyline 25 mg/day was compared with naproxen 1,000 mg/day and with the combination of amitriptyline 25 mg/day and naproxen 1,000 mg/day (38). Amitriptyline monotherapy and in combination with naproxen resulted in similar pain reduction.

Improvement in pain and quality of life.

Twenty-three (88%) studies investigated the improvement of pain in the antidepressant and placebo groups using a VAS. Pain reduction was <50% in 16 (84%) of 20 trials that reported this item in comparison with placebo. In 2 studies, an improvement in pain was reported without specification of the percentage (45, 47). Two studies with low study quality (see Table 2) achieved an improvement in pain reduction ranging between 52% and 58% (46, 52). Even when investigating the same drug, such as amitriptyline, different studies achieved variable results, ranging from 6% (37) to 36% (44). Only 10 (38%) of the 26 studies investigated quality of life. The mean improvement in quality of life in patients with FMS receiving antidepressant treatment was 30% (range 13–74%).

Better-powered studies (≥50 patients completing the study; 8 studies) reported a mean pain improvement of 33.6% (range 21–43.8%), while studies with lower numbers of participants (<50 patients completing; 6 studies) reported a mean pain improvement of only 19% (range 5.8–36.8%) (P = 0.046 by Mann-Whitney U test).

Effect on depression.

The effect of antidepressants on depression and depressive symptoms in the study population was assessed in 15 of the 26 analyzed RCTs; 11 of these studies reported a positive effect. The mean Jadad score of the studies showing no antidepressive effect was 4.8 (range 4–5), while the mean Jadad score of the RCTs with an antidepressive effect was 2.9 (range 1–5). Amitriptyline and paroxetine were used in 4 trials each and were effective in reducing depressive symptoms. Fluoxetine was used in 3 studies, duloxetine in 2, and citalopram and moclobemide in 1 RCT each, showing positive results for depressive symptoms. One study each reported no antidepressive effect of paroxetine (47), fluoxetine and amitriptyline (37), cyclobenzaprine and amitriptyline (34), and citalopram (51).

Association of changes in somatic symptoms and depressiveness.

Six of the 26 reviewed studies investigated the association between changes in somatic symptoms and depressiveness under treatment with antidepressants (Table 2). No correlations were reported between changes in pain and depressiveness for citalopram (50), amitriptyline (45), fluoxetine (48), and fluoxetine plus amitriptyline (37). Giordano et al found a significant correlation between changes in pain and anxiety (45). Only Arnold et al utilized the possibilities of multivariate analyses to quantify the indirect treatment effect on pain through improvement in depressive symptoms, which accounted for 23–39% of the total treatment effect in the duloxetine studies (55, 56).

Predictors of positive treatment outcome.

Only 3 of 26 studies performed subgroup analyses or checked for predictors of a positive treatment outcome. The effects of duloxetine on pain did not differ between FMS patients with and without major depressive disorder. Duloxetine was not effective in male patients (55). Carette et al found no correlation between a positive treatment response to amitriptyline and the Minnesota Multiphasic Personality Inventory, age, disease duration, and working status (34).

Followup.

A followup investigation was performed in only 2 of 26 studies. Carette et al reported that 9 of 19 patients with a positive treatment response were still receiving amitriptyline after 9 months without conducting a reassessment (34). Şencan et al described a reduction in pain and depressiveness 6 months after the discontinuation of paroxetine (46).

Economic outcome parameters.

The effects on indirect health care costs (days of sick leave, return-to-work rate) were not assessed by any study. Only 1 study analyzed direct health costs (self-reported intake of analgesics), which were reduced by paroxetine at the end of the study and at 6-month followup (46).

DISCUSSION

It can be concluded from the available data that the TCA amitriptyline (25–50 mg/day), which is the best-studied antidepressant, reduces pain, fatigue, and depressiveness. Furthermore, it improves sleep and quality of life. The SSRIs fluoxetine (20–40 mg/day), sertraline (50 mg/day), and paroxetine (20 mg/day) reduce pain and depressiveness and improve sleep and quality of life. Citalopram (20–40 mg/day) is not superior to placebo. The SNRIs duloxetine (60–120 mg) and milnacipran (25–200 mg/day) reduce pain and depressiveness and improve sleep and quality of life. The results concerning the effectiveness of the MAOI moclobemide (300–600 mg/day) are conflicting. One study demonstrated an advantage of the combination of TCAs and SSRIs compared with monotherapy. There is no evidence of a superiority of one class of antidepressants. There is preliminary evidence that the effects of TCAs, SSRIs, and SNRIs on pain reduction are mainly independent of the reduction of depressiveness.

There are some major limitations to the recommendation of antidepressants in patients with FMS. First, the short study durations with a mean observation time of 6–12 weeks hampers an adequate statement about long-term efficacy. Up to now, there have only been 2 trials in which patients underwent a followup investigation at later time points, with conflicting results. No studies have been published that investigated the long-term use (>6 months) and side effects of antidepressants. CWP and FMS are usually lifelong disorders that require long-term treatment. The optimal duration of treatment with antidepressants is not unknown.

Second, the mean extent of pain reduction was 26% and of improvement in quality of life was 30% (18% and 16%, respectively, after subtracting the placebo effect), indicating that patients may have had some improvement, but were far from achieving 50% improvement or remission of symptoms.

Third, most studies did not control for consumption, dose, and adverse effects of analgesic rescue medication. Therefore, the influence of this comedication on the study outcomes is not clear, and the reported (side) effects cannot be attributed to antidepressants or placebo alone.

Fourth, all antidepressants, and tricyclic antidepressants in particular, have typical adverse effects (e.g., anticholinergic) that may lead to difficulties in keeping up blinding. When standard placebo is used, these adverse effects might signal to the study participants that they have received the active medication, which might influence the outcome measures.

Fifth, the studies are of very different methodologic quality. Studies with lower quality report more positive outcomes than studies with higher quality.

Sixth, with regard to the exclusion criteria and the characteristics of the studies, there is no evidence for the effectiveness of antidepressants in patients <18 and >70 years of age, in nonwhite patients, in male patients, and in patients with distinct comorbidities (severe general diseases, mental disorders). With regard to the study setting, there is insufficient evidence for the effectiveness of antidepressants in primary care.

Seventh, up to now, no predictors of positive or negative treatment outcome have been identified.

Eighth, there is no evidence for the cost-effectiveness of antidepressants (11).

Finally, serum levels of the antidepressants applied were not checked. The question remains whether the study medication was taken consequently.

Thus, short-term use of antidepressants can be recommended for the symptomatic treatment of pain, fatigue, sleep disturbances, and depressiveness in patients with FMS. Before treatment is initiated, concomitant diseases, possible side effects, patients' preferences, and treatment costs should be considered. Because the evidence for a long-term effect of antidepressants in FMS is still lacking, the effect has to be reevaluated in regular intervals considering benefit for the patient versus side effects, and the drugs should be discontinued if the effect is no longer present.

Despite the large number of studies on the treatment of FMS, many aspects have not or only insufficiently been addressed up to now. First, there is a need for studies of longer duration to investigate the long-term efficacy and side effects of treatment with antidepressants in FMS. In this context, it would also be important to assess the influence of the medication on patients' quality of life and on how many patients return to normal daily life and work while receiving antidepressant treatment. Furthermore, trials analyzing the cost-effectiveness of pharmacologic and nonpharmacologic treatment should be performed. From an economic point of view, it would also be of interest to know if antidepressants can reduce FMS-related costs (12). The identification of positive and negative outcome predictors or FMS subgroups would be helpful to better select drugs for individual patients. It has to be kept in mind that the vast majority of studies investigated a population that did not include nonwhite patients and male patients. A possible subgroup-specific drug response has to be analyzed in future studies. Furthermore, a direct comparison of antidepressant treatment with nonpharmacologic therapies such as psychotherapy and aerobic exercise and a comparison of combinations with nonpharmacologic therapies versus monotherapies are necessary to determine the definite role of antidepressants compared with other treatment options for FMS symptoms.

AUTHOR CONTRIBUTIONS

Dr. Sommer had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study design. Häuser, Sommer.

Acquisition of data. Üçeyler, Häuser, Sommer.

Analysis and interpretation of data. Häuser.

Manuscript preparation. Häuser.

Ancillary