|
Treatment of fibromyalgia with antidepressants: a meta-analysis |
O'Malley P G, Balden E, Tomkins G, Santoro J, Kroenke K, Jackson J L |
|
|
Authors' objectives To systematically review the efficacy of antidepressants in the treatment of fibromyalgia, and to examine whether this effect is independent of depression.
Searching The authors searched MEDLINE from January 1966 to June 1999, PsycLIT from 1974 to 1998, and EMBASE from 1974 to 1998, using the following text and keywords: 'antidepressive agents', 'serotonin uptake inhibitors', 'monoamine oxidase inhibitors', 'amoxapine', 'clomipramine', 'tramipramine', 'desipramine', 'doxepin', 'imipramine', 'amitriptyline', 'maprotiline', 'nortriptyline', 'protriptyline', 'trazodone', 'nefazodone', 'fluoxetine', 'fluvoxamine', 'paroxetine', 'sertraline', 'femoxetine', 'venlafaxine', 'bupropion', 'citalopram', 'mianserin', 'pizotyline', 'pizotifen', and 'fibromyalgia', 'fibrositis' and 'fibromyositis'.
The authors also searched the Cochrane Controlled Trials Register, the Cochrane Database of Systematic Reviews, and the Federal Research in Progress (FEDRIP) database to identify unpublished literature, and examined the references of reviewed articles. Studies in any language were considered.
Study selection Study designs of evaluations included in the reviewRandomised, placebo-controlled trials (RCTs) including parallel and crossover study designs were considered. The mean sample size was 54 patients (range: 17 - 208) and the mean trial duration was 8.1 weeks (range: 3 - 24 weeks).
Specific interventions included in the reviewAntidepressants, i.e. tricyclics, selective serotonin re-uptake inhibitors and S-adenosylmethionine. Control groups received a placebo.
Participants included in the reviewPatients with fibromyalgia diagnosed using the American College of Rheumatology (ACR) criteria (8 studies) or the Smythe criteria (3 studies).
Outcomes assessed in the reviewFatigue, pain, well-being, and sleep were assessed using self-reported visual analogue scales, whilst the number and severity of trigger points were assessed using physician assessments.
How were decisions on the relevance of primary studies made?Two authors independently performed the selection of papers, and any disagreements were arbitrated by consensus. Inter-rater agreement was assessed using the kappa statistic (k=0.70).
Assessment of study quality The studies were assessed according to the method of Jadad et al. (see Other Publications of Related Interest no.1) using the following criteria: description and appropriateness of randomisation, adequacy of blinding, description of withdrawals and drop-outs, appropriateness of statistical analysis, clear description of inclusion and exclusion criteria, and methodology for assessing adverse treatment effects. Two authors independently performed the quality assessment, and any disagreements were arbitrated by consensus. Inter-rater agreement was assessed using the kappa statistic (k=0.84).
Data extraction The authors do not state how many of the reviewers performed the data extraction. Data were abstracted for the categories of: setting, country of origin, treatment characteristics (dose, duration, follow-up), demographics, number of participants enrolled, assessment of co-morbid psychiatric disease, follow-up losses, adverse effects, and outcomes.
Methods of synthesis How were the studies combined?Summary odds ratios (ORs), risk differences, summary mean symptom scores, and standardised mean differences with 95% confidence intervals (CIs) were calculated using the random-effects model of DerSimonian and Laird (see Other Publications of Related Interest no.2) Publication bias was assessed using the methods of Egger et al. (see Other Publications of Related Interest no.3).
How were differences between studies investigated?Heterogeneity of effect size was assessed using the Mantel-Haenszel method.
Sensitivity analyses were also performed by sequentially dropping individual studies and calculating the resulting summary measures. The effect of year of publication, study design (parallel versus crossover), quality score, and drug class was assessed using meta-regression.
Results of the review Thirteen RCTs (n=640) were included in the review with 15 treatment arms. Ten of the 15 arms studied tricyclic antidepressants.
Quality: included trials had consistently high quality scores (mean 5.6, range: 0 - 8). However, various problems were encountered with the studies, these included: no statement on the method of randomisation, no intention to treat analysis, inadequate assurance that blinding was effective, no method of assessing side-effects, and large losses to follow-up (greater than 20%).
Dichotomous outcome effect sizes (10 trials) were homogeneous (chi- squared 12.49, d.f.=11, p=0.33), and there was evidence of publication bias (p=0.4). The OR for improvement with therapy was 4.2 (95% CI: 2.6, 6.8). The pooled risk difference was 0.25 (95% CI: 0.16, 0.34), and the number-needed-to-treat was 4 (95% CI: 2.9, 6.3).
Continuous outcomes:
the number of trigger points (10 trials) improved by 0.17 standard deviation (SD) units (95% CI: -0.07, +0.42);
pain scores (10 trials) improved by 0.52 SD units (95% CI: 0.21, 0.81);
sleep (8 trials) improved by 0.49 SD units (95% CI: 0.3, 0.69);
fatigue scores (8 trials) improved by 0.39 SD units (95% CI: 0.11, 0.66); and
overall well-being (7 trials) improved by 0.49 SD units (95% CI: 0.18, 0.80).
Tests for heterogeneity showed that pain severity (chi-squared 22.0, d.f.=9, p=0.009), fatigue (chi-squared 14.2, d.f.=8, p=0.08) and well-being (chi-squared 15.5, d.f.=7, p=0.03) effect sizes were heterogeneous, while sleep (chi-squared 4.9, d.f.=8, p=0.8) and trigger points (chi-squared 15.2, d.f.=10, p=0.12) effect sizes were homogeneous. None of the pooled continuous outcomes data showed evidence of publication bias.
Depression: only one study reported a correlation between treatment effect and change in depression scores. Thus, it could not be determined from these studies whether any effect demonstrated was independent of an effect on depression.
Sensitivity analysis: the findings were not overly influenced by any particular study, with the OR varying from 3.2 to 4.9 and the summary mean difference from 0.1 to 1.1 SD units with the sequential exclusion of individual studies from the analysis. Meta-regression showed no significant effect of year of publication (p=0.52), study design (p=0.65), study quality scores (p=0.22) or drug class (p=0.43). Although there was evidence of publication bias (p=0.04) in the dichotomous data, there was no evidence of publication bias in the continuous data.
Authors' conclusions The authors state that antidepressants are efficacious in treating many of the symptoms of fibromyalgia. The myriad facets of fibromyalgia appear to be generally improved with the use of antidepressants. Patients are more than four times as likely to report overall improvement, and reported moderate reductions in individual symptoms, particularly pain. Whether this effect is independent of depression needs further study.
CRD commentary This was a good systematic review. The authors clearly stated the research question and the inclusion and exclusion criteria. The literature search appeared to be thorough, covered several databases and had no language restrictions. The authors also tested for publication bias and found no evidence for additional missed studies. The quality of the included studies was formally assessed and reported, and was discussed further in the review. The authors reported how the articles were selected and who performed the selection and validity assessment, but did not state who performed the data extraction. The data extraction was reported in tables and in the discussion in the text of the review.
The studies were combined in a statistical analysis and heterogeneity was assessed. A random-effects model was then used and reported in the results. Further sensitivity and publication bias analyses were also performed. The conclusions appear to follow from the results.
Implications of the review for practice and research Practice: The authors did not state any implications for practice.
Research: The authors state that further research is needed to assess the relative efficacy of different classes of antidepressants, as well as whether such efficacy is independent of depression.
Funding MacArthur Foundation Initiative.
Bibliographic details O'Malley P G, Balden E, Tomkins G, Santoro J, Kroenke K, Jackson J L. Treatment of fibromyalgia with antidepressants: a meta-analysis. Journal of General Internal Medicine 2000; 15(9): 659-666 Other publications of related interest 1. Jadad AR, Moore RA, Carroll D, Jenkinson C, Reynolds DJ, Gavaghan DJ, et al. Assessing the quality of reports of randomized clinical trials: is blinding necessary? Control Clin Trials 1996;17:1-12. 2. DerSimonian R, Laird N. Meta analysis in clinical trials. Control Clin Trials 1986;7:177-88. 3. Egger M, Smith GD, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ 1997;315:629-34.
This additional published commentary may also be of interest. Jacobs JWG, Geenen R. Review: antidepressants improve symptoms of fibromyalgia. Evid Based Med 2001;6:77.
Indexing Status Subject indexing assigned by NLM MeSH Adult; Aged; Antidepressive Agents /therapeutic use; Depression /complications; Female; Fibromyalgia /classification /drug therapy /psychology; Humans; Male; Middle Aged; Randomized Controlled Trials as Topic; Severity of Illness Index; Treatment Outcome AccessionNumber 12000002036 Date bibliographic record published 28/02/2002 Date abstract record published 28/02/2002 Record Status This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn. |
|
|
|