www.EuropeanJournalPain.com

European Journal of Pain 10 (2006) 185–192

Guidelines for the use of antidepressants in painfulrheumatic conditions

Serge Perrot *, Emmanuel Maheu, Rose-Marie Javier, Alain Eschalier, Anne Coutaux,Manuela LeBars, Philippe Bertin, Bernard Bannwarth, Richard Treves

Cercle d�etude de la douleur en rhumatologie, CEDR, Limoges, France

Received 31 August 2004; accepted 11 March 2005Available online 18 April 2005

Abstract

Objectives: Antidepressants are widely used to treat painful chronic rheumatic conditions but, contrary to neuropathic conditions,little is known about their true analgesic properties and value in these situations. Our group, which focuses on pain in rheumatology,aimed to develop recommendations for the use of antidepressants in rheumatology, based on evidence-based review of publisheddata and expert opinion.Method: We identified relevant drugs and conditions and searched Medline, Embase and Pascal (1966–2003) for relevant publica-tions in a number of European languages. We scored each study for quality, and used an expert consensus approach to formulaterecommendations.Results: We identified 77 studies and 12 meta-analyses and literature review on the use of antidepressant to treat painful rheuma-tological conditions. Forty-nine of these clinical studies were considered valid and were used to develop the recommendations. Whenevidence was lacking we based recommendations on our clinical experience.Conclusions: These recommendations for the treatment of painful rheumatological conditions with antidepressants were devel-oped using evidence-based and expert consensus approaches and are the first of their kind in this field. Our review of the literaturehighlights the need for further, well-designed clinical studies of the use of antidepressants to treat painful rheumatologicalconditions.� 2005 European Federation of Chapters of the International Association for the Study of Pain. Published by Elsevier Ltd. Allrights reserved.

Keywords: Pain; Musculoskeletal; Antidepressants; Treatment; Recommendations

1. Introduction

Pain is the main symptom of many rheumatic condi-tions, inflammatory conditions and degenerativediseases. In many cases, analgesics, non steroidal anti-inflammatory drugs (NSAIDs) or opioids control pain

1090-3801/$32 � 2005 European Federation of Chapters of the International

reserved.

doi:10.1016/j.ejpain.2005.03.004

* Corresponding author. Address: centre de la Douleur, hopitalCochin-Tarnier, 89 Rue d�Assas, 75006 Paris, France. Tel.: +33 1 58 4115 01; fax: +33 1 58 41 15 00.

E-mail address: serge.perrot@cch.ap-hop-paris.fr (S. Perrot).

effectively. However, in some cases, additional treat-ments, such as antidepressants and anticonvulsants,are required (Curatolo and Bogduk, 2001; Sawynoket al., 2001; Watson and Peter, 1994). The prescriptionof antidepressants (Blier and Abbott, 2001; Brysonand Wilde, 1996; Eschalier et al., 1998) is increasingfor many conditions, including fibromyalgia, rheuma-toid arthritis, spondylarthropathies, low back pain andosteoarthritis. Although antidepressants have often beenshown to reduce pain, the results obtained are variable(Barkin and Fawcett, 2000; Lynch, 2001; Onghena and

Association for the Study of Pain. Published by Elsevier Ltd. All rights

186 S. Perrot et al. / European Journal of Pain 10 (2006) 185–192

Van Houdenhove, 1992; Zitman et al., 1992) and manyquestions remain unanswered (Egbunike and Chaffe,1990; Phillip and Fickinger, 1993; Sindrup et al.,1992). For example: Does the analgesic effect dependon the antidepressant effect? What is the optimal dose?When is such treatment appropriate? How long shouldtreatment be continued?

Guidelines are therefore required concerning the useof these drugs. This document reviews the available evi-dence and then proposes a framework for the develop-ment of guidelines, without providing detailed adviceabout doses or formulations. It is designed to be a start-ing point for discussion and to be sufficiently flexible togain practical acceptance in different countries. It aimsto help prescribers (in primary care or specialist settings)to use antidepressants appropriately for the manage-ment of pain in rheumatic conditions.

2. Methods

2.1. The expert panel

This study was carried out by a working group ofnine experts for the CEDR (Cercle d�Etude de la Doul-eur en Rhumatologie) a specific interest group of theFrench Society of Rheumatology that focuses on rheu-matic pain. Seven of the experts involved in this studywere rheumatologists, one of whom was also a pharma-cologist. The other members were a psychiatrist and aneurologist/pharmacologist.

2.2. Search strategy

Relevant treatments and conditions were identifiedby the panel. We then searched the Medline, Embaseand Pascal databases for publications in several Euro-pean languages. Search terms used were: antidepressant,pain, rheumatoid arthritis, osteoarthritis, low back pain,fibromyalgia, fibrositis, rheumatic diseases, spond-ylarthropathy, ankylosing spondylarthritis, and sympa-thetic dystrophy. The search period covered 1966 toAugust 2002. Studies were scored using the EULARguidelines, from 0 to 28 (Pendleton et al., 2000), basedon a checklist. This checklist consists of 27 items distrib-uted between five sub-scales and was developed on thebasis of epidemiological principles, reviews of studydesigns, and existing checklists for the assessment ofrandomised controlled trials. This methodologicalchecklist provides a quality assessment of the reporting,external and internal validity and statistical power ofeach study. All studies are scored 0–1 for 26 questionsand 0–2 for one question, giving a maximum score of28. Power calculations are scored as 1 if present and 0if absent.

2.3. Guideline development

All potentially relevant papers were retrieved andtheir findings were analyzed by all members. The listof clinical questions used to define the scope of the liter-ature search was also used to structure the recommenda-tions together with another set of guidelines (on the useof opioids for the treatment of chronic pain) which hadbeen developed by a similar expert group, including oneof the CEDR panel members (Kalso et al., 2003). Sepa-rate sections were drafted by various participants usingthe evidence from the literature review whereverpossible.

The draft recommendations were circulated amongthe authors for further review and critical evaluation.A Delphi approach was then employed to reach con-sensus and all the recommendations were collated andsent back to the experts, who were asked to rank the10 most important proposals. A final draft of the rec-ommendations was generated, taking into account thecomments and criticisms of the nine panel members.The Delphi method is an exercise in group communi-cation among a panel of geographically dispersed ex-perts. It enables experts to deal systematically with acomplex problem or task. The essence of the tech-nique is fairly straightforward. It comprises a seriesof questionnaires sent by post or e-mail to a pre-selected group of experts. These questionnaires are de-signed to elicit and to develop individual responses tothe problems posed and to enable the experts to refinetheir views as the group�s work progresses in accor-dance with the assigned task. The Delphi methodmakes it possible to overcome the disadvantages ofconventional committee action. The group interactionin Delphi is anonymous, in that the originator ofcomments, forecasts, and the like is not identified.These elements are presented to the group in such away as to suppress any identification. Here are theten steps for the Delphi method:

1. Formation of a team to undertake and to monitora Delphi on a given subject.

2. Selection of one or more panels to participate inthe exercise. Panel members are usually expertsin the area to be investigated.

3. Development of the first-round Delphiquestionnaire.

4. Testing the questionnaire for proper wording (e.g.,ambiguities, vagueness).

5. Transmission of the first questionnaires to thepanelists.

6. Analysis of the first-round responses.7. Preparation of the second-round questionnaires

(and possible testing).8. Transmission of the second-round questionnaires

to the panelists.

S. Perrot et al. / European Journal of Pain 10 (2006) 185–192 187

9. Analysis of the second-round responses (steps 7–9are reiterated as long as many times as required toachieve stable results).

10. Preparation of a report by the analysis team topresent the conclusions of the exercise.

The recommendations were then tested according theAGREE method (AGREE Collaboration, 1999) by twoindependent reviewers and the strength of the evidencesupporting each recommendation was indicated, fromA to D (Shekelle et al., 1999). The purpose of the Ap-praisal of Guidelines Research & Evaluation (AGREE)Instrument is to provide a framework for assessing thequality of clinical practice guidelines. The AGREEInstrument is designed to assess guidelines developedby local, regional, national or international groups.The AGREE Instrument is generic and can be appliedto guidelines in any disease area. The AGREE Instru-ment assesses both the quality of the reporting, andthe quality of some aspects of recommendations. It pro-vides an assessment of the predicted validity of a guide-line, the likelihood that it will achieve its intendedoutcome. It does not assess the impact of a guidelineon patients� outcomes.

3. Results

We identified 137 relevant papers and, of these, se-lected 99 for detailed analysis: 77 were randomised con-trolled studies and 12 were meta-analyses or literaturereview. The results are summarised below.

3.1. Global review of analgesic effects of antidepressants

3.1.1. Analgesic effect of antidepressants

Analysis of 39 studies in various chronic naon-malig-nant painful conditions (Onghena and Van Houdenh-ove, 1992) found that antidepressants had an analgesiceffect, confirmed by Lynch in a meta-analysis of 59 stud-ies (Lynch, 2001). In neuropathic pain, the results areconvincing and have demonstrated differential analgesiceffects regarding the group of the drug: McQuay et al.(1996) and Sindrup and Jensen (1999) found that TCAswere more effective than serotonin-specific reuptakeinhibitors (SSRIs) in relieving neuropathic pain. Inglobal assessment of effects of antidepressants in chronicpain states, most of the authors have concluded that thetricyclic group of antidepressants is analgesic and thatdata regarding selective serotonin reuptake inhibitorsare conflicting (Lynch, 2001). A review of the use ofantidepressants in pain management, including rheu-matic conditions (Smith, 1998) found little reason toreplace TCAs by SSRIs in pain management. In a otherreview comparing some of the newer antidepressantswith TCAs (Ansari, 2000), only paroxetine and citalo-

pram were found to have a positive effect on neuro-pathic pain. Venlafaxine was recently shown to reduceneuropathic pain following chemotherapy for breastcancer (Tasmuth et al., 2002), and may also improvefibromyalgic patients (Sayar et al., 2003). But, exceptin neuropathic conditions and in fibromyalgia, there isno convincing study that really aimed to compare TCAswith SSRIs or other new antidepressants in specificchronic pain states and specifically rheumatological.For most of the authors antidepressant analgesic effectsare independent of their effects on mood (Magni, 1991).

3.2. Evaluation of the dose–response effect in chronic pain

There is no clear evidence for a dose-dependent re-sponse to antidepressant treatment in terms of pain re-lief (McQuay et al., 1993). However, we identified nostudies that dealt specifically with rheumatic pain. Con-flicting data have been obtained concerning the possiblerelationship between concentration and analgesic effect,but current therapeutic plasma concentration rangesseem to give an acceptable response for TCAs. Severalstudies in which TCAs were administered for varioustypes of neuropathic pain reported a relationship be-tween serum drug concentration and analgesic effect(Furlanut et al., 1993; Sindrup et al., 1991). For thenewer antidepressants, also in neuropathic pain, somestudies suggested that plasma concentration and effectare correlated (e.g. for sertraline and paroxetine for neu-ropathic pain) and others reported no such correlation(e.g. for fluoxetine and citalopram for neuropathic pain)(see references in Ansari (2000)).

3.3. Onset of action in chronic pain

The analgesic response seems to start before the anti-depressant response. An analgesic response is usuallyobserved within one week of starting treatment, whereasthe antidepressant response usually occurs after the firsttwo weeks (see references in Onghena and Van Hou-denhove (1992)).

3.4. Evaluation of different routes and patterns of

administration

The advantages of the various routes of administra-tion are unclear in humans. Due to a marked first-passeffect, oral bioavailability ranges from 20 to 80% (seereferences in Furlanut et al. (1993)) and genetic poly-morphism may also play a role in the pronounced phar-macokinetic variability observed with these drugs.Parenteral administration overcomes the problem offirst-pass metabolism, and results in high plasma con-centrations. However, apart from a possible indicationin patients unable to swallow, the parenteral route seemsto have no other real advantage, despite reports that this

188 S. Perrot et al. / European Journal of Pain 10 (2006) 185–192

route may accelerate the onset of the therapeutic effect(Pollock et al., 1986).

3.5. Antidepressants side effects

Imipraminic drugs, and TCAs in particular, cause sideeffect in 30–100% of patients treated for painful condi-tions (Eschalier et al., 1998; McQuay et al., 1993). Sideeffects are more frequent in fibromyalgia, occurring in70–95% of cases (Carette et al., 1986; Carette et al.,1994). Side effects parallel TCAs analgesic effects and,in most cases, depend on dose (McQuay et al., 1996). Ra-pid wash-out may lead to severe symptoms, such as nau-sea, vomiting, and trembling (Sindrup et al., 1992).Before starting TCA treatments, the physician shouldcheck for orthostatic hypotension and perform anECG. No recommendations have yet been publishedwith a view to preventing the side effects of TCAs. Theeffects of combined treatment with tramadol should alsobe monitored (Reus and Rawitscher, 2000), due to fre-quent co-utilization in rheumatological conditions.

SSRIs have been demonstrated to be well toleratedand safe. The only reported side effects are abdominalsymptoms at the start of the treatment, and serotoniner-gic syndrome. Side effects are reported in up to 80% ofpatients treated for painful conditions (Bird and Brog-gini, 2000; Norregard et al., 1995; Wolfe et al., 1994)but are clinically relevant in a lower proportion (0–41%) (Jung et al., 1997). This may account for the lowerrate of treatment drop-outs with SSRIs than with TCAs(Bird and Broggini, 2000; Sindrup et al., 1992; UshaRani et al., 1996) in pain studies. In fibromyalgia, SSRIsare well tolerated, about as well as placebo (Cantiniet al., 1994; Norregard et al., 1995; Wolfe et al., 1994),and are therefore more readily prescribed. Furthermore,no tests are required before starting SSRI treatment.Combination with tramadol is also not recommended.

3.6. Use of antidepressants in patients with specific

rheumatological conditions

3.6.1. Fibromyalgia syndrome

We identified 47 studies on the use of antidepressantsin fibromyalgia, including 25 controlled trials, most ofwhich involved tricyclic agents; and there were threemeta-analyses on this topic (Arnold et al., 2000; O�Mal-ley et al., 2000; White and Harth, 1996). The medianquality scores for these papers was 14.5 (range: 8–24).

Despite their widespread use, tricyclic drugs haveonly a moderate effect and only a minority of patientsdisplay sustained, marked improvement (Bennett et al.,1988; Bibolotti et al., 1986; Cantini et al., 1994; Carusoet al., 1987; Fossaluzza and De Vita, 1992; Goldenberget al., 1986, 1996; Hamaty et al., 1989; Hannonen et al.,1998; Heymann et al., 2001; Jaeschke et al., 1991; Qui-mby et al., 1989; Reynolds et al., 1991; Scudds et al.,

1989; Simms et al., 1991). A major placebo effect wasidentified in fibromyalgia studies, and antidepressantshave not been shown to have a lasting effect.

Amitriptyline is the most widely used drug for whichan effect on pain, fatigue, sleep and general conditionshas been demonstrated (Carette et al., 1986; Goldenberget al., 1986, 1996). Amitriptyline and cyclobenzaprinehave a greater effect on sleep disorders and fatigue thanon pain (Carette et al., 1994). Most of the studies re-ported the use of tricyclic drugs at doses lower thanthose used to treat depression, probably due to the sideeffects of these drugs.

Recent studies have assessed the effects of newer anti-depressants, such as citalopram (Andenberg et al., 2000;Norregard et al., 1995) and fluoxetine (Arnold et al.,2002; Cortet et al., 1992; Geller, 1989; Wolfe et al.,1994) or others (Dwight et al., 1998; Olin et al., 1998;Sayar et al., 2003), in patients with fibromyalgia. Inthese cases, the doses used were higher than or similarto those used in depression (Simms et al., 1991). Fewstudies have been carried out on the newer antidepres-sants and the available studies had low quality scoresin many cases. However it seems that although thesedrugs are better tolerated than tricyclic drugs at highdoses, their efficacy is limited (Arnold et al., 2002).

Overall, TCAs appear to be the most effective antide-pressants for the management of pain and other symp-toms in fibromyalgia, even if a large placebo effect wasobserved in many of these studies (Heymann et al.,2001; Lawson, 2002). A symptomatic effect on fibromy-algia is observed even at low doses (Ataoglu et al., 1997).SSRIs are better tolerated but less effective, making itnecessary to increase the dose to obtain significant painrelief.

3.6.2. Low back pain

Seven randomised, controlled trials of good qualityhave been published on this topic. Four reviews haveprecisely examined the analgesic and functional effectsof antidepressants in low back pain (Salerno et al.,2002; Staiger et al., 2003; Turner and Denny, 1993;van Tulder et al., 1997). The median quality score ofclinical papers was 11 (range: 6–15). Analysis of thesestudies suggested that antidepressants were slightly moreeffective than placebo for the relief of low back pain.Antidepressant treatment also tended to improve func-tion and everyday activities, although this trend wasnot statistically significant (Jenkins et al., 1976; Treveset al., 1991). One study indicated that imipramine wasmore effective than placebo, but only in terms of the‘‘numbers of days had to lie down’’ and ‘‘number ofdays with at least some restriction of normal activity’’(Alcoff et al., 1982). Another study (Pheasant et al.,1983) showed that amitryptyline was better than pla-cebo, comparing the use of rescue analgesics in bothtreatment groups. Nortriptyline and maprotiline were

S. Perrot et al. / European Journal of Pain 10 (2006) 185–192 189

significantly more effective than placebo, but nonethe-less had only moderate analgesic effects (Atkinsonet al., 1998; Atkinson et al., 1999; Dickens et al.,2000). In conclusion, tricyclic and tetracyclic antidepres-sants appear to produce moderate symptom reductionsfor patients with chronic low back pain, independentlyof a patient�s depression status. Selective serotonin reup-take inhibitors (SSRIs) do not appear to be beneficial.The effect of antidepressants on health-related qualityof life, mood and functional status is unclear (Staigeret al., 2003; Ward, 1986).

3.6.3. Osteoarthritis and inflammatory rheumatic diseases

We identified 15 randomised controlled trials onosteoarthritis (OA), rheumatoid arthritis (RA), andankylosing spondylitis (AS). Fourteen of these studieswere placebo-controlled and one compared amitrypti-line with paroxetine. Only eight of these 15 trials wereconsidered to meet the minimum standards in termsof methodological quality to demonstrate efficacy (Ashet al., 1999; Bird and Broggini, 2000; Frank et al., 1988;Ganvir et al., 1980; Grace et al., 1985; Koh et al.,1997; Macfarlane et al., 1986; Parker et al., 2003).Thesestudies scored from 13 to 22 on a scale with a maximumof 28 (median quality score: 16.4). In almost all these tri-als, efficacy in the control of pain and symptoms wasindependent of the antidepressant effect (with the excep-tion of Parker et al. (2003) and Sarzi Puttini et al.(1988)), which included depressed patients). Thus, ami-tryptyline, trimipramine, dothiepine and paroxetinemay have analgesic effects in patients with RA, and ami-tryptiline may be effective in reducing symptoms in AS.Analgesic effects of antidepressant were usually detectedafter one week of treatment. Low doses of amitryptiline(10–30 mg daily) may be sufficient to provide an analge-sic effect (Fowler et al., 1977; Hood et al., 2001; Mac-Neill and Dick, 1976; McQuay et al., 1992).

None of the studies specifically dealt with OA; it istherefore difficult to determine whether antidepressantsare beneficial for this condition. Some studies groupedtogether patients with OA, RA and AS, and reporteda small, but significant analgesic effect with TCAs andSSRIs (Gringras, 1976; McDonald Scott, 1969; Thorpeand Marchant-Williams, 1974; Usha Rani et al., 1996).In the study perfomed by Lin et al. (2003) in a largeand diverse population of older adults with arthritis(mostly osteoarthritis) and comorbid depression, bene-fits of improved depression care extended beyond re-duced depressive symptoms and included decreasedpain as well as improved functional status and qualityof life.

3.7. General guidelines for the use of antidepressants

Based on the findings of the 75 studies analyzed, theclinical experience of the expert panel, and previously

published guidelines on the treatment of chronic pain(Kalso et al., 2003) we propose the following guidelinesconcerning the use of antidepressants in painful rheu-matic conditions. The strength of the evidence support-ing each recommendation is indicated, from A to D(Shekelle et al., 1999).

1. Due to their analgesic and antidepressant proper-ties, antidepressants can improve the symptomsand quality of life of patients suffering from pain-ful chronic degenerative and inflammatory osteo-articular and spinal diseases. Their use should beincluded in a global management programmetogether with non-pharmacological approaches. A

2. Antidepressants, especially tricyclic drugs, are rec-ommended as analgesics for fibromyalgia. Theyshould not be first choice analgesic treatment inlow back pain, osteoarthritis and inflammatoryrheumatic painful diseases. A

3. To increase compliance, patients prescribed anti-depressants for analgesic use should be informedof the type of drug, its side effects, the aim of thetreatment and the expected delay until the analge-sic effects begin. D

4. Antidepressants may be prescribed as analgesics innon-depressed patients. The first-choice treatmentshould be a TCA, initiated at a low dose, whichis then increased to the maximal tolerated doseor to the minimal effective dose. A

5. Newer antidepressants with mixed action or spe-cific serotonin reuptake inhibitors should be triedonly if TCAs prove to be ineffective, poorly toler-ated or if they are contraindicated. Another anti-depressant (from the same class or another class)can be tried after failure of the initial antidepres-sant treatment, regardless of whether this failureis related to a lack of efficacy or unbearable sideeffects. C

6. The side effects of antidepressants used as analge-sics are similar to those observed in the treatmentof depression. They may be treated from treatmentinitiation and throughout the treatment period. A

7. If TCAs are prescribed to elderly patients, the phy-sician should monitor blood pressure, cognitionand intestinal transit. B

8. The assessment of treatment efficacy should not belimited to pain evaluation. It should include func-tional evaluation, analgesic consumption, sleep(quality and duration) evaluation, and psycho-sociological assessment, and should be startedafter one week of treatment. A

9. There is no optimal duration of antidepressanttreatment. Antidepressant treatment should bemaintained for at least 4 weeks before beingstopped due to lack of efficacy. Total durationshould be determined with respect to the initial

190 S. Perrot et al. / European Journal of Pain 10 (2006) 185–192

objectives, accepted by both the patients and thephysician, and after careful risk-benefit evaluation.D

10. After 3 to 6 months of remission, the dose may begradually decreased, with regular pain assess-ments. Stopping the treatment too abruptly maylead to nausea, vomiting and trembling. B

A = based on category 1 evidence; B = based on cat-egory 2 evidence or extrapolated recommendation fromcategory 1 evidence; C = based on category 3 evidenceor extrapolated recommendation from category 1 or 2evidence; D: based on category 4 evidence or extrapo-lated from category 2 or 3 evidence (Shekelle et al.,1999).

4. Conclusions

Guidelines should be based on available evidence andthis was the goal of this expert group. However, we soonrealised that little or no evidence was available for manyof the key issues, because of the paucity of appropriateclinical studies and as clinical studies does not reflectthe real clinical situations, e.g. and long-term utilizationin painful chronic situations. The published studies pro-vide little information useful to determine which individ-uals respond to antidepressants with analgesic effects.More studies have focused on fibromyalgia than onother conditions, but although there is a trend towardsthe use antidepressants in fibromyalgia, no clear analge-sic effects have been demonstrated in this situation. Theresults of studies on osteoarthritis, low back pain andrheumatoid arthritis are not very convincing. Therehas also been no comparative study of patients with dif-ferent, specific rheumatological conditions. It is also notyet possible to determine which type of pain will respondmost strongly to treatment with antidepressants. Fur-thermore, it is not possible to define predictive factorsfor analgesic effects: psychological status does not pre-dict analgesic effect and antidepressants do not exhibitmore powerful analgesic effects in depressed patients.

These recommendations have been tested by twoindependent reviewers, according the AGREE method(AGREE Collaboration, 1999), specifically dedicatedfor the assessment of clinical guidelines. Globalassessment was good, and specific assessments of eachthe 6 domains were rated with scores ranging from53% to 100%, thus allowing the spreading of theserecommendations.

These recommendations also gave us the opportunityto define research guidelines. The responsiveness ofmany chronic pain conditions, especially rheumatologi-cal, to antidepressants has not been assessed incontrolled settings and we know very little about thelong-term (months to years) efficacy and adverse effects

of antidepressants. Some clinicians have reported suc-cessful treatment with SSRIs following the failure of tri-cyclics drugs, however no clinical studies have beenperformed in this area. Further studies are needed toinvestigate the role of plasma concentration, the influ-ence of concomitant psychiatric disturbances, and thetype of organic lesions on the analgesic response to anti-depressant effects. Another field for future research isthe possibility of co-administering drugs to increasethe efficacy of antidepressants or to reduce their adverseeffects. Following the development of clinical guidelines(AGREE Collaboration, 1999), we have planned to re-vise these recommendations in a next future, to keep itrelevant, according to most recent clinical findings,and mostly to both the patient�s and physician�sexperience.

References

The AGREE Collaboration. AGREE instrument, www.agreecollabo-ration.org, 1999.

Alcoff J, Jones E, Rust P, Newman R. Controlled trial of imipraminefor chronic low back pain. J Fam Pract 1982;28:841–6.

Andenberg UA, Marteinsdottir I, von Knorring L. Citalopram inpatients with fibromyalgia – a randomised, double-blind, placebo-controlled study. Eur J Pain 2000;4:27–35.

Ansari A. The efficacy of newer antidepressants in the treatment ofchronic pain: a review of current literature. Harvard Rev Psychi-atry 2000;7(5):257–77.

Arnold LM, Keck PE, Welge JA. Antidepressant treatment offibromyalgia. A meta-analysis and review. Psychosomatics2000;41:104–13.

Arnold LM, Hess EV, Hudson JI, Welge JA, Berno SE, Keck Jr PE. Arandomized, placebo-controlled, double-blind, flexible-dose studyof fluoxetine in the treatment of women with fibromyalgia. Am JMed 2002;112:191–7.

Ash G, Dickens CM, Creed FH, Jayson MI, Tomenson B. The effectsof dothiepin on subjects with rheumatoid arthritis and depression.Rheumatology (Oxford) 1999;38:959–67.

Ataoglu S, Ataoglu A, Erdoan F, Sarac J. Comparison of paroxetine,amitriptyline in the treatment of fibromyalgia. Turk J Med Sci1997;27:535–9.

Atkinson JH, Slater MA, Williams RA, Zisook S, Patterson TL, GrantI, et al. A placebo-controlled randomised clinical trial of nortrip-tyline for chronic low back pain. Pain 1998;76:287–96.

Atkinson JH, Slater MA, Wahlgreen DR, Williams RA, Zisook S,Pruitt, et al. Effects of noradrenergic and serotoninergic antide-pressants on chronic low back pain. Pain 1999;83:137–45.

Barkin RR, Fawcett J. The management challenges of chronic pain:the role of antidepressants. Am J Therapeut 2000;7:31–47.

Bennett RR, Gatter RR, Campbell SS, Andrews RR, Clark SS,Scarola JA. A comparison of cyclobenzaprine and placebo in themanagement of fibrositis. A double-blind controlled study. Arthri-tis Rheum 1988;31(12):1535–42.

Bibolotti E, Borghi C, Pasculli E, Regoli F, Tavoni A, Baroni L, et al.The management of fibrositis. A double-blind comparison ofmaprotiline (Ludiomil) clorimipramine and placebo. Clin Trials J1986;23:269–80.

Bird H, Broggini M. Paroxetine versus amitriptyline for treatment ofdepression associated with rheumatoid arthritis:a random-ized, double blind, parallel group study. J Rheumatol 2000;27:2791–7.

S. Perrot et al. / European Journal of Pain 10 (2006) 185–192 191

Blier P, Abbott FV. Putative mechanisms of action of antidepressantdrugs in affective and anxiety disorders and pain. J PsychiatryNeurosci 2001;26:37–43.

Bryson HH, Wilde MM. Amitriptyline. A review of its pharmacolog-ical properties and therapeutic use in chronic pain states. DrugsAging 1996;8(6):459–76.

Cantini F, Bellandi F, Niccoli L, Di Munno O. Fluoxetina associata aciclobenzaprina nel trattamento della fibromialgia. Minerva Med1994;85:97–100.

Carette S, McCain GG, Bell DD, Fam AA. Evaluation of amitrip-tyline in primary fibrositis. A double-blinded, placebo-controlledstudy. Arthritis Rheum 1986;29:655–9.

Carette S, Bell MM, John Reynolds W, Haraqui B, McCain GG,Bykerk VV, et al. Comparison of amitriptyline, cyclobenzaprine,and placebo in the treatment of fibromyalgia. Arthritis Rheum1994;37:32–40.

Caruso I, Sarzi Puttini PC, Boccassini L, Santandrea S, Locati M,Volpato R, et al. Double-blind study of dothiepin versus placeboin the treatment of primary fibromyalgia syndrome. J Int Med Res1987;15:154–9.

Cortet B, Houvenagel E, Forzy G, Vincent G, Delcambre B.Evaluation de l�efficacite d�un traitement serotoninergique (chlor-hydrate de fluoxetine). Etude ouverte au cours de la fibromyalgie.Rev Rhum Mal Osteoartic 1992;59:497–500.

Curatolo M, Bogduk N. Pharmacologic pain treatment of musculo-skeletal disorders:current perspectives and future prospects. Clin JPain 2001;17:25–32.

Dickens C, Jayson M, Sutton C, Creed F. The relationship betweenpain and depression in a trial using paroxetine in sufferers ofchronic low back pain. Psychosomatics 2000;41:490–9.

Dwight MM, Arnold LL, O�Brien H, Metzger R, Morris-Park E, KeckPP. An open trial of venlafaxine treatment of fibromyalgia.Psychosomatics 1998;39:14–7.

Egbunike IG, Chaffe BJ. Antidepressants in the management ofchronic syndromes. Pharmacotherapy 1990;10:262–70.

Eschalier A, Ardid D Dubray C. Tricyclic and other antidepressants asanalgesics. In: Jana, editor. Novel aspects of pain management:opi-oids and beyond, 1998.

Fossaluzza V, De Vita S. Combined therapy with cyclobenzaprine andibuprofen in primary fibromyalgia syndrome. Int J Clin Pharm Res1992;XII:99–102.

Fowler PD, MacNeill A, Spencer D, Robinson ET, Dick WC.Imipramine, rheumatoid arthritis and rheumatoid factor. CurrMed Res Opin 1977;5:241–6.

Frank RG, Kashani JH, Parker JC, Beck NC, Brownlee-Duffeck M,Elliott TR, et al. Antidepressant analgesia in rheumatoid arthritis.J Rheumatol 1988;15:1632–8.

Furlanut M, Benetello P, Spina E. Pharmacokinetic optimisation oftricyclic antidepressant therapy. Clin Pharmacokinet1993;24:301–18.

Ganvir P, Beaumont G, Seldrup J. A comparative trial of clomipra-mine and placebo as adjunctive therapy in arthralgia. J Int MedRes 1980;8(Suppl):60–6.

Geller SS. Treatment of fibrositis with fluoxetine hydrochlorhyde(Prozac�). Am J Med 1989;87:594–5.

Goldenberg DD, Felson DD, Dinerman H. A randomised,controlled trial of amitriptyline and naproxene in the treat-ment of patients with fibromyalgia. Arthritis Rheum 1986;29:1371–7.

Goldenberg D, Mayskiy M, Mossey C, Ruthaser R, Schmid C. Arandomized, double-blind crossover trial of fluoxetine and ami-triptyline in the treatment of fibromyalgia. Arthritis Rheum1996;39:1852–9.

Grace EM, Bellamy N, Kassam Y, Buchanan WW. Controlled,double-blind, randomized trial of amitriptyline in relieving artic-ular pain and tenderness in patients with rheumatoid arthritis. CurrMed Res Opin 1985;9:426–9.

Gringras M. A clinical trial of tofranil in rheumatic pain in generalpractice. J Int Med Res 1976;4(Suppl 2):41–9.

HamatyD, Valentine JL, HowardR,HowardCW,Wakefield V, PattenMS. The plasma endorphin, prostaglandin and catecholamineprofile of patients with fibrositis treated with cyclobenzaprine andplacebo: a 5-month study. J Rheumatol Suppl 1989;19:164–8.

Hannonen P, Malminiemi K, Yli-Kerttula U, Isomeri R, Roponen P.A randomized, double-blind, placebo-controlled study of moclobe-mide and amitriptyline in the treatment of fibromyalgia in femaleswithout psychiatric disorder. Br J Rheumatol 1998;37:1279–86.

Heymann RE, Helfenstein M, Feldman D. A double-blind, random-ised, controlled study of amitriptyline, nortriptyline and placebo inpatients with fibromyalgia. An analysis of outcome measures. ClinExp Rheumatol 2001;19:697–702.

Hood SD, Argyropoulos SV, Nutt DJ. Arthritis and serotoninergicantidepressants. J Clin Psychopharmacol 2001;24:458–61.

Jaeschke R, Adachi J, Guyatt G, Keller J, Wong B. Clinical usefulnessof amitriptyline in fibromyalgia:the results of 23 N-of-1 random-ized controlled trials. J Rheumatol 1991;18:447–51.

Jenkins DG, Ebbutt AF, Evans CD. Tofranil in the treatment of lowback pain. J Int Med Res 1976;4:28–40.

Jung AA, Staiger T, Sullivan M. The efficacy of selective serotoninreuptake inhibitors for the management of chronic pain. J GenIntern Med 1997;12:384–9.

Kalso E, Allan L, Dellemijn PL, Faura CC, Ilias WK, Jensen TS,et al. Recommendations for using opioids in chronic non-cancerpain. Eur J Pain 2003;7:381–6.

Koh WW, Pande I, Samuels A, Jones SS, Calin A. Low doseamitriptyline in ankylosing spondylitis: a short term, double blind,placebo controlled study. J Rheumatol 1997;4:2158–21561.

Lawson K. Tricyclic antidepressants and fibromyalgia: what is themechanism of action?. Expert Opin Invest Drugs 2002;11: 1437–45.

Lin EH, Katon W, Von Korff M, Tang L, Williams Jr JW, Kroenke K,et al. Effect of improving depression care on pain and functionaloutcomes among older adults with arthritis: a randomizedcontrolled trial. JAMA 2003;290:2428–9.

Lynch EE. Antidepressants as analgesics: a review of randomisedcontrolled trials. J Psych Neurosci 2001;26:30–6.

Macfarlane JG, Jalali S, Grace EM. Trimipramine in rheumatoidarthritis: a randomized double-blind trial in relieving pain and jointtenderness. Curr Med Res Opin 1986;10(2):89–93.

MacNeill AL, Dick WC. Imipramine and rheumatoid factor. J IntMed Res 1976;4(Suppl 2):23–7.

Magni G. The use of antidepressants in the treatment of chronic pain.Drugs 1991;42:730–48.

McDonald Scott WA. The relief of pain with an antidepressant inarthritis. Practitioner 1969;202:802–7.

McQuay HJ, Carroll D, Glynn CJ. Low-dose amitriptyline in thetreatment of chronic pain. Anesthesia 1992;47:646–52.

McQuay HJ, Carroll D, Glynn CJ. Dose–response for analgesic effectof amitriptyline in chronic pain. Anaesthesia 1993;48:281–5.

McQuay HJ, Tramer M, Nye BA, Carroll D, Wiffen PJ, Moore RA. Asystematic review of antidepressants in neuropathic pain. Pain1996;68:217–27.

Norregard J, Volkmann H, Danneskiold-Samsoe B. A randomizedcontrolled study of citalopram in the treatment of fibromyalgia.Pain 1995;61:445–9.

O�Malley PP, Balden E, Tornkins G, Santoro J, Kroenke K, JacksonJJ. Treatment of fibromyalgia with antidepressants. A meta-analysis. JGIM 2000;15:659–66.

Olin R, Klein R, Berg PA. A randomised double-blind 16-week studyof ritanserin in fibromyalgia syndrome: clinical outcome andanalysis of autoantibodies to serotonin, gangliosides and phospho-lipids. Clin Rheumatol 1998;17:89–94.

Onghena P, Van Houdenhove B. Antidepressant-induced analgesia inchronic non-malignant pain: a meta-analysis of39 placebo-con-trolled studies. Pain 1992;49:2052.

192 S. Perrot et al. / European Journal of Pain 10 (2006) 185–192

Parker JC, Smarr KL, Slaughter JR, Johnston SK, Priesmeyer ML,Hanson KD, et al. Management of depression in rheumatoidarthritis: a combined pharmacologic and cognitive-behavioralapproach. Arthritis Rheum 2003;49:766–77.

Pendleton A, Arden N, Dougados M, Doherty M, Bannwarth B, et al.EULAR recommendations for the management of knee osteoar-thritis: report of a task force of the Standing Committee forInternational Clinical Studies Including Therapeutic Trials (ESCI-SIT). Ann Rheum Dis 2000;59:936–44.

Pheasant H, Bursk A, Godfarb J, Azen SS, Weiss JJ, Borelli L.Amitriptyline and chronic low back pain. Spine 1983;8:552–7.

Phillip M, Fickinger M. Psychotropic drugs in the management ofchronic pain syndromes. Pharmacopsychiatry 1993;26:221–34.

Pollock BG, Perel JM, Shostac M, Antelman SM, Brandon B, KupferDJ. Understanding the response lag to tricycles: application ofpulse loading regimens with intravenous clomipramine. Psycho-pharmacol Bull 1986;22:214–9.

Quimby LG, Gratwick GM, Whitney CD, Block SR. A randomizedtrial of cyclobenzaprine for the treatment of fibromyalgia. JRheumatol Suppl 1989;19:140–3.

Reus VV, Rawitscher L. Possible interaction of tramadol andantidepressants. Am J Psychiatry 2000;157:839.

Reynolds WJ, Moldofsky H, Saskin P, Lue FA. The effects ofcyclobenzaprine on sleep physiology and symptoms in patientswith fibromyalgia. J Rheumatol 1991;18:452–4.

Salerno SM, Browning R, Jackson JL. The effect of antidepressanttreatment on chronic back pain. Arch Intern Med 2002;162:19–24.

Sarzi Puttini P, Cazzola M, Boccassini L, Ciniselli G, Santandrea S,Caruso I, et al. A comparison of dothiepin versus placebo in thetreatment of pain in rheumatoid arthritis and the association ofpain with depression. J Int Med Res 1988;16:331–7.

Sawynok J, Esser MM, Reid AR. Antidepresssants as analgesics: anoverview of central and peripheral mechanisms of action. JPsychiatry Neurosci 2001;26(1):21–9.

Sayar K, Aksu G, Ak I, Tosun M. Venlafaxine treatment offibromyalgia. Ann Pharmacother 2003;37:1561–5.

Scudds RA, McCain GA, Rollman GB, Harth M. Improvements inpain responsiveness in patients with fibrositis after successfultreatment with amitriptyline. J Rheumatol Suppl 1989;19:98–103.

Shekelle PG, Woolf SH, Eccles M, Grimshaw J. Clinical guidelines:developing guidelines. BMJ 1999;318:593–6.

Simms RW, Felson DT, Goldenberg DL. Development of preliminarycriteria for response to treatment in fibromyalgia syndrome. JRheumatol 1991;18:1558–63.

Sindrup SS, Grodum E, Gram LL, Beck-Nielsen H. Concentrationresponse relationship of paroxetine treatment in diabetic neurop-athy. A patient-blinded dose escalation study. J Monit 1991;13:408–14.

Sindrup SS, Brosen K, Gram LL. Antidepressant in pain treatment:antidepressant or analgesic effect?. Clin Neuropharm 1992;15:636A–7A.

Sindrup SH, Jensen TS. Efficacy of pharmacological treatments ofneuropathic pain: an update and effect related to mechanism ofdrug action. Pain 1999;83:389–400.

Smith AA. The analgesic effects of selective serotonin reuptakeinhibitors. J Psychopharmacol 1998;12:407–13.

Staiger TO, Gaster B, Sullivan MD, Deyo RA. Systematic review ofantidepressants in the treatment of chronic low back pain. Spine2003;28:2540–5.

Tasmuth T, Hartel B, Kalso E. Venlafaxine in neuropathic painfollowing treatment of breast cancer. Eur J Pain 2002;6:17–24.

Thorpe P, Marchant-Williams R. The role of an antidepressant,dibenzepin (Noveril), in the relief of pain in chronic arthritic states.Med J Aust 1974;1(8):264–6.

Treves R, Montane De La Roque Ph, Dumond JJ, Bertin Ph, ArnaudM, Desproges-Gotteron R. Etude prospective de l�action antalgi-que de la clomipramine versus placebo dans les lombosciatalgiesrebelles(68 cas). Rev Rhum 1991;58:549–52.

Turner JA, Denny MC. Do antidepressant agents relieve chronic lowback pain?. J Fam Pract 1993;37:545–53.

Usha Rani P, Naidu MUR, Prasad VBN, Ramesh Kumar Rao T,Shobha JJ. An evaluation of antidepressants in rheumatic painconditions. Anesth Analg 1996;83:371–5.

van Tulder MW, Koes BW, Bouter LM. Conservative treatment ofacute and chronic nonspecific low back pain. Spine1997;22:2128–56.

Ward NN. Tricyclic antidepressants for chronic low back pain. Spine1986;11:661–5.

Watson C, Peter N. Antidepressant drugs as adjuvant analgesics. JPain Symptom Manage 1994;9:392–405.

White KP, Harth M. An analytical review of 24 controlled clinicaltrials of fibromyalgia syndrome (FMS). Pain 1996;64:211–9.

Wolfe F, Cathey MM, Hawley DJ. A double-blind placebo controlledtrial of fluoxetine in fibromyalgia. Scand J Rheumatol1994;23:255–9.

Zitman FF, Linsse ACG, Edelbroek PM, VanKempen GG. Clinicaleffectiveness of antidepressants and antipsychotics in chronicbenign pain. Clin Neuropharm 1992;15(S1):377A.