Dr. Häuser has received consultancies, speaking fees, and/or honoraria (less than $10,000 each) from Lilly, Pfizer, Janssen, and Mundipharma, and a travel grant from Lilly.
Efficacy of multicomponent treatment in fibromyalgia syndrome: A meta-analysis of randomized controlled clinical trials
Version of Record online: 29 JAN 2009
Copyright © 2009 by the American College of Rheumatology
Arthritis Care & Research
Volume 61, Issue 2, pages 216–224, 15 February 2009
How to Cite
Häuser, W., Bernardy, K., Arnold, B., Offenbächer, M. and Schiltenwolf, M. (2009), Efficacy of multicomponent treatment in fibromyalgia syndrome: A meta-analysis of randomized controlled clinical trials. Arthritis & Rheumatism, 61: 216–224. doi: 10.1002/art.24276
- Issue online: 29 JAN 2009
- Version of Record online: 29 JAN 2009
- Manuscript Accepted: 28 OCT 2008
- Manuscript Received: 5 JUL 2008
- German Rheumatology League and the German Fibromyalgia Association
To systematically review the efficacy of multicomponent treatment of fibromyalgia syndrome (FMS).
We screened Medline, PsychINFO, Scopus, and the Cochrane Library (through December 2007), as well as reference sections of original studies, reviews, and evidence-based guidelines. Randomized controlled trials (RCTs) on the multicomponent treatment (at least 1 educational or other psychological therapy with at least 1 exercise therapy) of FMS were analyzed.
We included 9 (of 14) RCTs with 1,119 subjects (median treatment time 24 hours) in the meta-analysis. Effects were summarized using standardized mean differences (SMDs) or weighted mean differences (WMDs). There was strong evidence that multicomponent treatment reduces pain (SMD −0.37; 95% confidence interval [95% CI] −0.62, −0.13), fatigue (WMD −0.85; 95% CI −1.50, −0.20), depressive symptoms (SMD −0.67; 95% CI −1.08, −0.26), and limitations to health-related quality of life (HRQOL) (SMD −0.59; 95% CI −0.90, −0.27) and improves self-efficacy pain (SMD 0.54; 95% CI 0.26, 0.82) and physical fitness (SMD 0.30; 95% CI 0.02, 0.57) at posttreatment. There was no evidence of its efficacy on pain, fatigue, sleep disturbances, depressive symptoms, HRQOL, or self-efficacy pain in the long term. There was strong evidence that positive effects on physical fitness (SMD 0.30; 95% CI 0.09, 0.51) can be maintained in the long term (median followup 7 months).
There is strong evidence that multicomponent treatment has beneficial short-term effects on the key symptoms of FMS. Strategies to maintain the benefits of multicomponent treatment in the long term need to be developed.
Fibromyalgia syndrome (FMS) is defined by the American College of Rheumatology (ACR) as chronic (>3 months), widespread pain (axial plus upper and lower segment plus left- and right-sided pain) and tenderness in at least 11 of 18 tender points (1). Population-based studies estimate that the prevalence of FMS ranges from 0.5– 5.8% (2). FMS is associated with fatigue, poor sleep, other functional somatic syndromes, and mental and physical disorders (3, 4). Patients diagnosed with FMS cause high direct costs (health care use) (5, 6) and indirect costs (sick leave, disability pension) (4). Effective treatment options are therefore needed for medical and economic reasons (7).
Recently, evidence-based guidelines on the management of FMS have been published in order to give patients and physicians an orientation within the continuously growing number of treatment options for FMS. Amitriptyline, duloxetine, and pregabalin; aerobic exercise; balneotherapy and spa therapy, and cognitive–behavioral therapy have been found to be effective monotherapies. These treatment modalities, as well as their combination within a multicomponent therapy assuming synergistic effects, have been recommended by 3 guidelines (8–10).
There are conflicting results of systematic reviews regarding the efficacy of multicomponent therapy and the grade of recommendation given by evidence-based guidelines. One systematic review, based on a search of controlled trials through 2003, concluded that multicomponent therapy is effective for decreasing pain and FMS impact and for increasing physical fitness (11). Another systematic review, based on a search of randomized controlled trials (RCTs) until January 2006, concluded that the benefits of multicomponent therapy are limited and that positive outcomes disappear in the long run (12). The guideline on the management of FMS by the American Pain Society attributed the highest level of recommendation to multicomponent therapy (9), whereas a multidisciplinary European expert committee recommended only multicomponent therapy based on expert opinion (8).
In the meantime, further controlled studies on multicomponent therapy in FMS have been published that were not included in the reviews by Burckhardt (11) and Koulil (12). To our knowledge, a meta-analysis providing effect sizes of multicomponent therapy has not been previously published. The aims of our review therefore were to expand the search on controlled trials of multicomponent therapy in FMS until December 2007 and to determine the effect sizes of multicomponent therapy on the key symptoms of FMS by a meta-analysis.
MATERIALS AND METHODS
Data sources and searches.
Meta-analysis was performed according to the Quality of Reporting of Meta-analyses guidelines (13). The electronic bibliographic databases screened included Medline (1966 through December 2007), PsycINFO (1966 through December 2007), Scopus (1980 through December 2007) (14), and the Cochrane Library (1993 through December 2007). The keywords (medical subject headings terms) for initial inclusion were “fibromyalgia,” “fibromyalgia syndrome,” and “chronic widespread pain” in combination with the terms “rehabilitation,” “multidisciplinary treatment,” “multimodal treatment,” and “combined modality therapy.” After consulting with the German center of the Cochrane Collaboration, we did not use the highly sensitive search strategy (15). In addition, reference sections of original studies and review papers on nonpharmacologic treatments of FMS were screened manually and independently by 2 authors (WH and KB). No language restrictions were made. We also contacted European experts for additional studies on multicomponent therapy.
To be included in our review, studies were required to meet the following criteria: 1) the multicomponent therapy treatment had to include at least 2 nonpharmacologic therapies (at least 1 educational or other psychological therapy, and at least 1 exercise therapy); 2) the diagnosis of FMS had to be based on recognized criteria; 3) the study had to have a controlled study design with a control group that received no treatment, care as usual, or another well-defined treatment with a lower intensity than multicomponent therapy; 4) the study had to have symptom-specific outcomes of the key symptoms of FMS, such as pain, fatigue, sleep disturbances, depressive symptoms, and health-related quality of life (HRQOL) (16), and/or relevant pain-related psychological domains such as self-efficacy pain (SEP), and/or objective tests of physical fitness; 5) the study had to be published in full paper form (no abstracts); and 6) data had to be suitable for meta-analysis. To create a more complete data set for analysis, we contacted authors of studies with missing data to request the data required for analysis (means and SDs of pretest and posttest data, or SDs of change scores). Studies for which we were unable to obtain missing data were excluded from meta-analysis.
Two reviewers independently screened the titles and abstracts of potentially eligible studies identified by the search strategy detailed above. The full-text articles were then examined independently by 2 reviewers (WH and MO) to determine whether each met the selection criteria. For the preparation of the meta-analysis, 2 of the 4 reviewers independently extracted data (study characteristics, study results) using standard extraction forms. Point estimates for selected variables were extracted and checked by the other 2 reviewers. We used kappa statistics to assess agreement between reviewers. All discrepancies were rechecked and consensus was achieved by discussion.
The van Tulder score (17) using 11 items was applied for assessing the methodologic quality of the analyzed RCTs. Based on the van Tulder score items, we classified RCTs as high-quality (score 8–11), moderate-quality (score 5–7), or low-quality (score 1–4) studies.
Data synthesis and analysis.
When researchers reported more than 1 measure for the outcomes assessed, we used a predefined order of preference for entry into the meta-analysis: 1) for pain, a visual analog scale (VAS) of the Fibromyalgia Impact Questionnaire (FIQ), a VAS, or the pain numeric rating scale (NRS); 2) for fatigue, a VAS FIQ or a VAS; 3) for sleep, a VAS FIQ, a VAS, or an NRS; 4) for depression, the Beck Depression Inventory (BDI) or a VAS FIQ; 5) for global well-being (quality of life), the FIQ total score; 6) for self-efficacy pain, the Arthritis Self-Efficacy Scale, the Coping Style Questionnaire, or the Self-Efficacy Scale; and 7) for physical fitness, the Six-Minute Walk Test (meters).
We analyzed intent-to-treat data whenever available. For the comparison of proportions, chi-square test was applied. Nonparametric tests (the Mann-Whitney U test and the Kruskal-Wallis H test) were used for the comparison of continuous variables. Two-sided P values less than or equal to 0.05 were considered significant. Meta-analyses were conducted using RevMan software, version 4.2.10 (18).
Due to the fact that most outcomes of interest were presented as continuous data (mean value and SD or mean changes), we used either the weighted mean difference (WMD) or the standardized mean difference (SMD) as effect measures. The WMD was calculated when the outcome measure in all trials was determined on the same instrument, and the SMD when outcomes were measured by different instruments. To calculate WMD or SMD, we used means and SDs or change scores for each intervention. The SMD used in Cochrane reviews is the effect size known as Hedges' (adjusted) g.
We used I2 statistics to measure heterogeneity between the RCTs. If the I2 value was <50%, a fixed-effect meta-analysis was applied. If the I2 value was ≥50%, a random-effect meta-analysis was used (18).
We used Cohen's categories to evaluate the magnitude of the effect size, calculated by WMD or SMD, where d >0.2–0.5 indicates a small effect size, d >0.5–0.8 indicates a medium effect size, and d >0.8 indicates a large effect size (19).
We used the following modified levels of evidence descriptors to classify the results of the meta-analysis: strong = consistent findings in at least 2 moderate-quality RCTs; moderate = consistent findings in at least 2 low-quality RCTs and/or 1 moderate-quality RCT; limited = findings in 1 low-quality RCT; conflicting = inconsistent findings among multiple RCTs; no evidence = findings in no RCTs (17).
Sensitivity analyses were conducted to determine whether the duration of treatment (<30 hours versus ≥30 hours) and the methodologic quality of each study had a significant effect on the study's results.
Potential publication bias (i.e., the association of publication probability with the statistical significance of study results) was investigated using visual assessment of the funnel plot (plots of effect estimates against sample size) calculated by RevMan software (18). Publication bias may lead to asymmetrical funnel plots (20). Furthermore, we tested the sensitivity of our results to potential unpublished studies using a file-drawer test for meta-analysis. This test determines how many negative studies with an effect size of d = 0.01 would be needed to negate our findings (fail-safe N) (21). If fail-safe N > file-drawer N (5k +10, where k = the number of studies meta-analyzed), the results of the meta-analysis can be regarded as robust against potential reporting bias (22).
Selection of studies.
The literature search produced 340 citations involving FMS, multicomponent therapy, and controlled trials, 14 of which met initial inclusion criteria. Of the 326 excluded articles, 162 did not evaluate multicomponent therapy, 90 were double hits of controlled studies (studies found in at least 2 data sources), 65 were review articles, 8 were uncontrolled trials with multicomponent therapy, and 1 controlled trial was not randomized (23). Additionally, 1 RCT with multicomponent therapy included patients with chronic, widespread pain. After more detailed review of the 13 remaining articles, a further 4 papers were excluded for the following reasons: one because no control group met the inclusion criteria (2 different modalities of multicomponent therapy were compared) (24), and 3 others because means and/or SDs of pretest and posttest data, or SDs of change scores, were not included in the publication, and were not provided by the authors upon request (25–27). Finally, 9 studies met our selection criteria and were included for meta-analysis (28–36) (Figure 1). Interrater reliability for this assessment was κ = 0.92 (Table 1).
|Author, year, country||Tulder score||Mean age, years||Women, %||Race||Exclusion criteria||Study population, n (%)||Control group||Treatment group|
|Screened/ randomized||Randomized/ completing||Total/ completing, n (%)||Treatment, hours||Total/ completing, n (%)||Quality treatment||Duration|
|Cedraschi et al, 2004, Switzerland||4||48.9||93||100% white||Somatic diseases||176/164 (93)||164/129 (79)||80/68 (81)||Waiting list, 0||84/61 (83)||Group: 6x45 minutes swimming pool; 4x45 minutes relaxation; 2x45 minutes low-impact, land-based exercises; 2x90 + 6x45 minutes discussion||6 weeks, 18 hours|
|Gowans et al, 1999, Canada||4||44.3||78||NR||None||NR||45/41 (91)||22/21 (96)||Waiting list, 0||23/20 (86)||Group: 12x1h education, 12x30 minutes walking, side-stepping, arm exercises against water resistance, stretching||12 weeks, 18 hours|
|Hammond and Freeman, 2006, UK||5||48.3||87||NR||Age, somatic diseases, mental disorders||255/183 (72)||183/106 (60)||86/51 (58)||Relaxation, 10||97/53 (55)||Group: education, relaxation, stress management, postural training, stretching, strengthening, tai chi||10 weeks, 20 hours|
|Keel et al, 1998, Switzerland||4||49.0||89||NR||Mental disorders||55/32 (58)||32/27 (84)||14/13 (92)||Relaxation, 15||16/14 (77)||Group: information; instruction in self-control strategies, autogenic training, group discussion, stretching, aerobic exercise||15 weeks, 30 hours|
|King et al, 2002, US||7||46.1||100||NR||Age, somatic diseases||255/169 (66)||169/109 (64)||39/34 (87)||Waiting list, 0||37/35 (94)||Group: aerobic exercise, education, self-management||12 weeks, 24 hours|
|Lemstra and Olszynski, 2005, Canada||6||49.4||84.8||NR||Age, malignan- cies||82/79 (96)||79/72 (91)||36/36 (100)||Treatment as usual, NR||43/36 (93)||Group: 18 hours subaerobic exercise, stretching, light weight training; 4 hours CBT, 2 hours massage||6 weeks, 24 hours|
|Mannerkorpi et al, 2000, Sweden||6||46.0||100||NR||Somatic diseases, mental disorders||NR||69/58 (84)||32/30 (93)||Treatment as usual, NR||37/30 (19)||Group: 6 hours education; 15 hours endurance, flexibility, coordination, relaxation||24 weeks, 21 hours|
|Rooks et al, 2007, US||7||50.0||100||92% white||Age, somatic diseases||356/207 (58)||207/135 (65)||50/27 (54)||Education, 7 weeks, 14 hours||55/38 (69)||32 hours exercise/strength, 14 hours education||16 weeks, 46 hours|
|Zijlstra, 2005, The Netherlands||5||48||95.5||NR||Age, somatic diseases, mental disorders, pending litigation||280/170 (60)||170/134 (79)||86/76 (88)||Treatment as usual, NR||84/58 (31)||Thalassotherapy 18 hours, 7 hours aerobic exercise, 10 hours education||25 weeks, 35 hours|
Eight studies were conducted in an outpatient setting, and one was conducted in a Tunisian holiday spa resort (36). Patients were referred from primary, secondary, and tertiary care settings. One study additionally included patients of a national FMS patient association (36). Five studies were conducted with European patients (28, 30, 31, 34, 36), and the remaining studies took place in the US and Canada.
Education and exercise (a combination of aerobic training, stretching, and strengthening) were applied to the patients in all studies. Elements of cognitive–behavioral therapy such as relaxation and self-management techniques were used in 4 studies (30–33). Two studies additionally offered physical therapy: massage (33) and spa therapy (36). No study provided data on how the intensity of aerobic exercise was determined and controlled. No study provided data on the adherence of patients to multicomponent therapy, on additional home-based training during the study, or on the (dis)continuation of physical exercise and relaxation training by the patients after the end of the study. The median treatment time was 24 hours (range 18–46). Three studies lasted ≥30 hours (31, 35, 36).
Six studies reported the percentage of screened and finally randomized patients (median 66%, range 58–96%). A total of 345 subjects completed multicomponent therapy, and a total of 356 subjects completed control treatment. The median of the sample size for the multicomponent therapy groups was 43 (range 16–97). The median percentage of patients completing the studies was 76% (range 55–94%), and the median percentage of control participants completing the studies was 88% (range 54–100%) (not significant). The median mean age in the studies was 44.7 years (range 44–50). One study involved only women (36); 8 studies examined both men and women, with women being the majority. The median percentage of women in all studies was 96%. Only 2 studies reported the race of the patients (28, 35). The median of the percentage of white patients was 96% (range 92–100%).
FMS was diagnosed in all studies according to the ACR criteria (1). Two studies gave some data on somatic comorbidity (28, 35). No study performed psychiatric interviews or reported the frequency of mental disorders. Five studies defined age-related exclusion criteria (30, 32, 33, 35, 36), 6 studies excluded patients with somatic diseases (28, 32–36), and 4 studies excluded patients with mental disorders (30, 31, 34, 36). One study excluded patients with pending litigation (36).
The most frequent outcome variables based on patient rating were pain, fatigue, sleep, depressive symptoms, and HRQOL. Six studies performed physical tests of muscle strength and/or aerobic capacity (29, 30, 32, 34–36). One study (33) used socioeconomic outcomes such as work status and number of drugs prescribed. Three studies defined a primary outcome measure (HRQOL score) (30, 34, 36). Outcomes were assessed at the end of the treatment in 7 studies. The 3 earliest assessments were after 6 months (28), 4 months (30), and 1 month (36). All but one study (34) performed a followup after the end of treatment. The longest followup was 15 months (33).
The effect sizes on pain, fatigue, and depressed mood are shown in Figures 2, 3, and 4. There is strong evidence (≥3 RCTs with consistent results, and at least 1 moderate-quality study) that multicomponent therapy reduces pain, fatigue, and depressed mood. There is no evidence (1 low-quality study) that multicomponent therapy reduces sleep disturbances (WMD 0.10, 95% confidence interval [95% CI] −0.86, 1.06; P = 0.84). There is strong evidence (≥3 RCTs with consistent results, and at least 1 moderate-quality study) that multicomponent therapy reduces limitations of HRQOL (SMD −0.59, 95% CI −0.90, −0.27; P = 0.002) and improves SEP (SMD 0.54, 95% CI 0.26, 0.82; P = 0.0002) and physical fitness (SMD 0.30, 95% CI 0.02, 0.57; P = 0.03).
First followup (3–4 months).
There were effect sizes <0.2 and nonsignificant tests for overall effect for pain, sleep disturbances, depressed mood, HRQOL, and physical fitness. There was strong evidence (2 moderate-quality studies with consistent findings) that multicomponent therapy improved SEP with a small effect size (SMD 0.47, 95% CI 0.14, 0.80; P = 0.005) (data not shown).
Last-followup (6–12 months).
There were effect sizes <0.2 and nonsignificant tests for overall effect for pain, fatigue, sleep disturbances, depressed mood, HRQOL, and SEP. There was strong evidence (2 moderate-quality studies with consistent findings) that multicomponent therapy led to an improvement of physical fitness with a small effect size (SMD 0.30, 95% CI 0.09, 0.51; P = 0.005) (data not shown).
A comparison of studies with a moderate (van Tulder score 5–7) and low (van Tulder score 4) methodologic quality, as well as with a total duration of ≥30 hours and <30 hours, is presented in Table 2.
|Outcome title||Studies, n||MT participants, n||Statistical method||Effect size (95% CI)||P|
|Moderate methodologic quality (van Tulder score 5–7)|
|Pain||3||102||SMD (Fixed)||−0.44 (−0.72, −0.15)||0.003|
|Fatigue||2||66||WMD (Fixed)||−0.91 (−1.66, −0.15)||0.02|
|Depressed mood||3||102||SMD (Random)||−0.75 (−1.25, −0.24)||0.004|
|Quality of life||3||92||SMD (Random)||−0.59 (−0.90, −0.27)||0.0002|
|Self-efficacy pain||3||92||SMD (Fixed)||0.56 (0.24, 0.87)||0.0005|
|Physical fitness||3||102||SMD (Fixed)||0.29 (−0.02, 0.59)||0.07|
|Low methodologic quality (van Tulder score 4)|
|Pain||2||34||SMD (Fixed)||−0.19 (−0.07, 0.29)||0.45|
|Fatigue||1||20||WMD (Fixed)||−0.70 (−1.98, 0.58)||0.28|
|Sleep||1||14||WMD (Fixed)||0.10 (−0.86, 1.06)||0.84|
|Depressed mood||1||20||WMD (Fixed)||−1.20 (−3.10, −0.70)||0.21|
|Quality of life||ND|
|Physical fitness||1||20||SMD (Fixed)||0.35 (−0.28, 0.97)||0.28|
|Total treatment ≥30 hours|
|Pain||2||52||SMD (Fixed)||−0.36 (−0.78, −0.05)||0.09|
|Fatigue||1||38||WMD (Fixed)||−1.20 (−2.20, −0.20)||0.02|
|Depressed mood||1||38||SMD (Fixed)||−1.07 (−2.73, 0.60)||0.26|
|Quality of life||1||38||SMD (Fixed)||−0.58 (−1.08, −0.07)||0.02|
|Self-efficacy pain||1||38||SMD (Fixed)||0.63 (0.13, 1.14)||0.01|
|Physical fitness||1||38||SMD (Fixed)||0.44 (−0.06, 0.94)||0.09|
|Total treatment <30 hours|
|Pain||3||84||SMD (Fixed)||−0.38 (−0.68, −0.07)||0.02|
|Fatigue||2||48||WMD (Fixed)||−0.59 (−1.45, 0.27)||0.18|
|Depressed mood||3||84||SMD (Fixed)||−0.84 (−1.16, −0.52)||< 0.0001|
|Quality of life||2||54||SMD (Fixed)||−0.60 (−1.00, −0.19)||0.004|
|Self-efficacy pain||3||74||SMD (Fixed)||0.50 (0.16, 0.84)||0.004|
|Physical fitness||3||74||SMD (Fixed)||0.24 (−0.10, 0.57)||0.16|
There is strong evidence that moderate quality studies had small effect sizes on pain, moderate effect sizes on depressed mood, HRQOL, and SEP, and a large effect size on fatigue, with significant tests for overall effect at posttreatment. The test of overall effect for physical fitness was not significant. The tests for overall effect of all low-quality studies at posttreatment were not significant for pain, fatigue, or physical fitness.
Duration of treatment.
There is moderate evidence that studies with a duration of ≥30 hours had a large effect size on fatigue, a small effect size on HRQOL, and a medium effect size on SEP, with significant tests for overall effect. The tests for overall effect were not significant for pain, depressed mood, or physical fitness.
There is strong evidence that studies with a duration of <30 hours had a small effect size on pain, a large effect size on depressed mood, and medium effect sizes on HRQOL and SEP, with significant tests for overall effect. The tests for overall effect were not significant for fatigue or physical fitness.
There was significant heterogeneity between the analyzed studies in most outcome variables. The ranges of the 95% CIs are also indicative of great variations between the studies.
Visual scanning of forest plots for subgroup analysis suggested a random distribution with results in the same direction for most outcomes, indicating that although the study effects differed in size, their results were mostly consistent (data not shown). This is supported by the P tests of significance associated with the meta-analysis of most outcomes: the test for overall effect at posttreatment was significant for all outcomes with the exception of sleep.
The fail-safe number with dc = 0.01 as the selected criterion value to nullify the average effects at posttreatment on pain was 144, on fatigue was 252, on depressed mood was 264, on HRQOL was 177, on SEP as 212, and on physical fitness was 116. Thus, the fail-safe numbers were larger than Rosenthal's rule of thumb of 35 for pain, 15 for sleep, 30 for depression, 25 for HRQOL, 30 for SEP, and 30 for physical fitness. These results indicate that a publication bias was not evident in the data of this meta-analysis.
The primary outcome for this meta-analysis was to determine the short- and long-term efficacy of multicomponent therapy in FMS. There is strong evidence of the efficacy of multicomponent therapy to reduce some key symptoms of FMS, such as pain, fatigue, and depressed mood, and to improve self-efficacy and physical fitness at posttreatment. There is strong evidence that the positive effects of multicomponent therapy on the key symptoms of FMS decline with time. There is strong evidence that positive effects on SEP and on physical fitness can be detected at followup.
Our results are in line with Burckhart (11), who concluded that multicomponent therapy is effective for decreasing pain and FMS impact, and for increasing physical fitness at the end of treatment. We mainly agree with Koulil (12) that there is no evidence of the efficacy of multicomponent therapy on FMS symptoms in the long run. However, we found positive effects on SEP and physical fitness at followup.
There are several limitations to our study. First, there is no internationally accepted definition of multicomponent therapy. The existing systematic reviews on multicomponent therapy agree that it should include at least 1 educational or other psychological therapy and at least 1 exercise therapy. In Germany, inpatient multicomponent therapy, which is reimbursed by health insurance companies, must include at least 3 different therapy modalities: psychotherapy (cognitive–behavioral therapy) is obligatory; physiotherapy, relaxation training, occupational therapy, aerobic exercise, strength training, sensomotoric training, work hardening (a functional restoration program designed to return a person to work and to improve work abilities), and creative or music therapy can be included. Moreover, a standardized multidisciplinary assessment of the treatment effects and team conferences are necessary to coordinate the different treatment modalities (37). Only 4 of the 10 studies reviewed met the criterion of including at least 3 therapy modalities (28, 30, 31, 36). No study gave data about whether interdisciplinary assessments and team conferences were carried out.
There are no internationally accepted standards for the minimum effective duration of multicomponent therapy for the treatment of FMS. We found a great variety in the duration of the studies, with a maximum length of 46 hours. Taking into consideration the limitations of this meta-analysis due to the small number of studies and participants, we found no relevant differences at posttreatment between studies with total treatment times ≥30 hours and <30 hours. In chronic low back pain, a systematic review concluded that only intensive multicomponent rehabilitation >100 hours with a functional restoration approach improves pain and function. Interventions <30 hours or usual care did not show improvements in clinically relevant outcomes (38). In multicomponent therapy studies on FMS, only 1 nonrandomized controlled study (excluded from meta-analysis) exceeded 100 hours (23). Thus the question arises: were the multicomponent therapy studies included in our meta-analysis underpowered in regard to the duration of treatment, and are they worth the expense?
Busch et al (39) demonstrated that there is moderate evidence that aerobic training (at the intensity recommended for increases in cardiorespiratory fitness) had positive small effect sizes for global well-being and possibly for pain at posttreatment. No multicomponent therapy study that we reviewed reported the intensity of its aerobic training. Moreover, the adherence to all types of exercises described in the included studies was poor. Thus it remains unclear whether the positive effects of multicomponent therapy can be attributed to aerobic exercise, to psychological therapy, or to both modalities.
Three RCTs had to be excluded because the available data were not suited to meta-analysis and were not provided by the authors on request. The results of the 3 studies excluded were in line with the main results of the meta-analysis.
The methods of referral and the study settings included all levels of medical care. However, the applicability of the results to the whole population of FMS patients seeking medical help is limited to female patients living in the US, Canada, and northern and mid-Europe, which are probably mostly white and between 30 and 60 years of age. No statements on the efficacy of multicomponent therapy in children, adolescents, older adults, and non-whites are possible.
No statements on the efficacy of multicomponent therapy in patients with severe somatic disease, including secondary FMS associated with inflammatory rheumatoid disease or patients with severe psychiatric disorder (psychosis, drug dependence), are possible. Because no psychiatric assessment was performed in the studies, no statement can be made as to whether multicomponent therapy is effective in FMS patients with comorbid major depression. Also, no conclusions can be made about patients with a pending litigation for early pension due to FMS, because these patients were either excluded or no subgroup analysis with these patients was performed.
Because medication was not controlled in most studies, there remains some uncertainty whether the effects reported are due only to the multicomponent therapy applied.
Only a few data are available on the socioeconomic outcomes of multicomponent therapy. Only one (33) of 3 RCTs analyzed (30, 31, 33) found a reduction of FMS-related direct costs (visits to doctors, medication). One RCT could not find a change in patients' work status after multicomponent therapy (33).
In summary, it appears difficult to determine the overall benefit of multicomponent therapy from such a limited number of studies. Moreover, the results for some outcomes meta-analyzed are based on only 2 or 3 studies. It is important to note that more RCTs had been conducted only with aerobic exercise and antidepressants than with multicomponent therapy in patients with FMS. A current systematic review with meta-analysis included 34 RCTs with exercise (39). We systematically reviewed 26 RCTs with antidepressants (40). The methodologic problems (e.g., no control for concomitant medication, lack of psychiatric assessment, lack of predefined outcome measures) that we pointed out for multicomponent therapy are also problems in other treatments of FMS (39–41).
Despite these methodologic limitations, the American Pain Society (9) and the German interdisciplinary guideline on the management of FMS (42) gave a strong recommendation to multicomponent therapy. This recommendation is based on the fact that multicomponent therapy and cognitive behavioral therapy are the only treatment modalities with limited evidence of a long-term efficacy on some dimensions of FMS. Multicomponent therapy should be offered to FMS patients with relevant limitations in daily functioning who do not respond to monocomponent pharmacologic or nonpharmacologic treatment (10).
The main problem to solve, in our opinion, is how to maintain the positive effects of multicomponent therapy after the end of treatment. Patients should be encouraged to continue regular aerobic exercises and psychological self-management strategies to maintain the positive results of multicomponent therapy. Multicomponent therapy booster sessions offered by outpatient departments or by self-help organizations to FMS groups could enhance the motivation for life-long aerobic exercise and psychological self-management. In Germany, continuous pool- and land-based exercises as well as support groups are offered by FMS self-help organizations and are reimbursed by health insurance companies (43).
Dr. Häuser 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, Arnold, Schiltenwolf.
Acquisition of data. Häuser, Offenbächer, Schiltenwolf.
Analysis and interpretation of data. Häuser, Bernardy, Arnold, Offenbächer, Schiltenwolf.
Manuscript preparation. Häuser, Arnold, Schiltenwolf.
Statistical analysis. Häuser, Bernardy.
- 9Guideline for the management of fibromyalgia syndrome: pain in adults and children. Glenview (IL): American Pain Society; 2005., , , , , , et al.
- 18The Nordic Cochrane Centre. Review Manager (RevMan): version 4.2. Copenhagen: The Cochrane Collaboration; 2003.
- 19Statistical power analysis for the behavioral sciences. Hillsdale (NJ): Lawrence Erlbaum; 1988..
- 21A fail-safe N for effect size in meta-analysis. J Educ Stat 1983; 8: 157–9..
- 37German Institute of Medical Documentation and Information. Operation and procedure key version 2007. URL: www.dimdi.de/static/de/klassi/prozeduren/ops301/opshtml2007/fr-ops.htm. In German.
- 43Rehabilitationsträger. General agreement on rehabilitation sports and “Funktionstraining.” URL: http://www.bar-frankfurt.de/upload/Rehabilitationssport_314.pdf. In German.