Ottawa Panel Evidence-Based Clinical Practice Guidelines for Strengthening Exercises in the Management of Fibromyalgia: Part 2

Background and Purpose: The objective of this study was to create guidelines for the use of strengthening exercises in the management of adult patients (>18 years of age) with fibromyalgia (FM), as defined by the 1990 American College of Rheumatology criteria. Methods: Following Cochrane Collaboration methods, the Ottawa Methods Group found and synthesized evidence from comparative controlled trials and formed the Ottawa Panel, with nominated experts from key stakeholder organizations. The Ottawa Panel then developed criteria for grading the recommendations based on experimental design (I for randomized controlled trials, II for nonrandomized studies) and strength of evidence (A, B, C+, C, D+, D, or D−). From the rigorous literature search, 5 randomized controlled trials were selected. Statistical analysis was based on Cochrane Collaboration methods. Continuous data were calculated with weighted mean differences between the intervention and control groups, and dichotomous data were analyzed with relative risks. Clinical improvement was calculated using absolute benefit and relative difference in change from baseline. Clinical significance was attained when an improvement of 15% relative to a control was found. Results: There were 5 positive recommendations: 2 grade A and 3 grade C+. All 5 were of clinical benefit. Discussion and Conclusion: The Ottawa Panel recommends strengthening exercises for the management of fibromyalgia as a result of the emerging evidence (grades A, B, and C+, although most trials were rated low quality) shown in the literature.

Group found and synthesized evidence from comparative controlled trials and formed the Ottawa Panel, with nominated experts from key stakeholder organizations. The Ottawa Panel then developed criteria for grading the recommendations based on experimental design (I for randomized controlled trials, II for nonrandomized studies) and strength of evidence (A, B, Cϩ, C, Dϩ, D, or DϪ). From the rigorous literature search, 5 randomized controlled trials were selected. Statistical analysis was based on Cochrane Collaboration methods. Continuous data were calculated with weighted mean differences between the intervention and control groups, and dichotomous data were analyzed with relative risks. Clinical improvement was calculated using absolute benefit and relative difference in change from baseline. Clinical significance was attained when an improvement of 15% relative to a control was found.
Results. There were 5 positive recommendations: 2 grade A and 3 grade Cϩ. All 5 were of clinical benefit.

Discussion and Conclusion. The Ottawa Panel recommends strengthening
exercises for the management of fibromyalgia as a result of the emerging evidence (grades A, B, and Cϩ, although most trials were rated low quality) shown in the literature.
F ibromyalgia (FM), as defined by the 1990 American College of Rheumatology criteria, is a chronic, generalized, musculoskeletal pain disorder with the presence of at least 11 out of 18 tender-point sites on physical examination. 1 Additional symptoms frequently include fatigue, muscle pain and weakness, memory problems, paresthesia, headaches, depression, and insomnia. 2,3 Fibromyalgia symptoms can become exacerbated from humidity and cold weather, overactivity, and fatigue, 2 but considerable symptom variability exists (ie, frequency and intensity). 4 Fibromyalgia is 2 to 5 times more prevalent than rheumatoid arthritis (RA) 5 and affects 2% to 4% of the American population, the majority of whom are women. 6 Prevalence of FM is likely higher due to misdiagnosis by physicians of other systemic and psychiatric disorders. 5 People of all ages are susceptible to developing FM, but people aged 35 to 50 years are most at risk. 5 According to some researchers, 7 the disability rate among people with FM is as high as 46%. On average, the direct cost (eg, medications, physician visits, imaging and laboratory procedures) and indirect cost (eg, work wage, household work) for a person with FM were CAN $2,298 and CAN $5,035, respectively, over a 6-month period. 8 Successful management of FM remains difficult because little is known of its etiology. 9 Evidencebased clinical practice guidelines (EBCPGs) are precise statements on recommended interventions that are based on scientific literature and include a graded strength of evidence as well as detail on the specific joints affected, outcomes, and length of intervention. The Ottawa Panel, which has published EBCPGs on RA, 10 osteoarthritis (OA), 11 and stroke, 12 collaborated to assess the strength of scientific evidence regarding the efficacy of physical exercise for FM. Other researchers 5,13-15 have published general clinical recommendations for the management of FM; however, these recommendations cannot be regarded as guidelines because they relied, in most cases, on studies that did not follow rigorous methods. Furthermore, all recommendations failed to provide specific statements on the therapeutic exercise utilized and its strength of evidence.
Given the volume of literature found, the Ottawa Panel decided to produce a series of 2 EBCPGs for FM: EBCPGs for aerobic fitness exercises (part 1 of the series) 16 and EBCPGs for strengthening exercises (part 2 of the series). Neither exercise type has been universally effective or wholeheartedly endorsed by all patients with FM. 17 Patients frequently are noted anecdotally to have difficulty performing and sustaining strengthening exercises, which are traditionally a component of musculoskeletal injury training programs. Additionally, 83% of patients with FM do not engage in aerobic exercise, and 65% have below-average aerobic fitness. 18 The EBCPGs in parts 1 and 2 of the series aim to assist patients with FM and health care professionals in identifying the most effective exercise specific to the patients' needs. Exercise has the potential to decrease inactivity and deconditioning, a common cause of pain associated with FM, and to provide multiple beneficial physical and psychological outcomes. The purpose of the part 2 study was to provide effective strengthening exercise guidelines for patients, physiatrists, rheumatologists, physical therapists, occupational therapists, family physicians, kinesiologists, and other health care professionals to assist in the overall management of FM.

Methods
The Ottawa Methods Group, a panel of 9 methodologists with extensive backgrounds in developing EBCPGs, contacted professional associations that specialize in the management of patients with FM to nominate experts with clinical experience. The Ottawa Methods Group chose 9 experts with specialties in rheumatology, physiatry, psychology, psychiatry, occupational therapy, and physical therapy. The Ottawa Methods Group and the chosen experts then formed the Ottawa Panel, who were responsible for the EBCPGs in this report. The Ottawa Methods Group also assembled a research team with expertise in meta-analysis, research methods, and the development and evaluation of EBCPGs.
The Ottawa Methods Group established inclusion criteria for the study design, subject sample, intervention, and outcomes used for conducting separate, systematic literature reviews (Tab. 1). The research team was responsible for reading and analyzing the articles as well as drafting several evidence tables to which the Ottawa Panel made final corrections and reached a consensus.
The EBCPGs were created according to Appraisal of Guidelines Research and Evaluation (AGREE) criteria (www.agreecollaboration.org). 19 The Ottawa Panel graded the recommendations based on their level (I for randomized controlled trials [RCTs], II for nonrandomized studies) and their strength of evidence (A, B, C, Cϩ, D, Dϩ, or D؊). As an illustration, to receive a grade A recommendation, an RCT needed an outcome that was both statistically significant and clinically important (an improvement of Ͼ15% relative to a control, based on panel expertise and empiric results). Table 2 provides a summary of the EBCPG grading system. For further details on Ottawa Panel methods, refer to the EBCPGs for aerobic fitness in the management of fibromyalgia 16 in this issue.

Literature Search
A structured a priori literature search was performed by a library scientist using modified search strategies 20 recommended by the Cochrane Collaboration. 21 The main focus of the search was the methods used and interventions of the study, rather than the outcomes. Potential for bias was minimized through a systematic approach to the search, study selection, and data extraction and synthesis.
This search was further widened through the inclusion of casecontrol, cohort, and nonrandomized studies. The following electronic databases were utilized: MEDLINE, EMBASE (Current Contents), the Cumulative Index to Nursing and Allied Health (CINAHL), AMED, the Cochrane Controlled Trials Register up to December 2006, the registries of the Cochrane Field of Rehabilitation and Related Therapies and the Cochrane Musculoskeletal Group, and the Physiotherapy Evidence and Database (PEDro). The library scientist updated the literature search every 6 months from the first week of October 2004 to the last week of December 2006.

Study Inclusion/Exclusion
Criteria Type of interventions. Therapeutic exercises related to strengthening were included. Examples of specific interventions that were excluded were surgery of all joints, medications, thermal biofeedback, and exercises combined with an educational program (Tab. 1).
Type of study designs. Comparative controlled studies with comparison groups that assessed strengthening exercises and patients with FM were included: RCTs, controlled clinical trials (CCTs), cohort studies, and case-control studies. Controlled clinical trials are similar to RCTs, except that CCTs are either not randomized or not appropriately randomized. 21 Head-to-head studies (eg, strengthening exercises versus flexibility exercises) also were included.
Studies were excluded if they lacked a comparison group (eg, uncontrolled cohort studies), were case studies, had a dropout rate of Ͼ20%, or had a sample of fewer than 5 participants per group. Abstract-only studies were excluded because they did not have sufficient data for analysis. Studies published in a language other than English or French also were excluded due to time constraints and translation costs (Tab. 1).

Type of participants.
Studies of adult patients (Ͼ18 years of age) with FM, as defined by the 1990 American College of Rheumatology criteria, 1 were included. Participants had to be medically stable and mentally competent. Because FM is a chronic illness, the amount of time since disease onset is not a requirement of FM diagnostic criteria. Although some would argue that there is difficulty in differentiating symptoms of other rheumatologic conditions such as chronic pain syndrome, chronic fatigue syndrome, or myofascial pain syndrome from FM, these rheumatologic conditions were excluded. Studies were excluded if participants had any of the following conditions: (1) cancer or other oncological conditions, (2) cardiac conditions, (3) dermatologic conditions, (4) serious cognitive deficits or severe communication problems, (5) major medical problems that could interfere with the rehabilitation process or incapacitate functional status, or (6) primary psychiatric conditions (Tab. 1).

Ottawa Panel EBPCGs for Strengthening Exercises in the Management of Fibromyalgia
following outcomes: quality of life, pain, fatigue, sleep, global perceived effect, depression, muscle strength (force-generating capacity), endurance, and power. Excluded outcomes were biochemical measures and serum markers (Tab. 1).

Study selection. After receiving
Cochrane process training by the Ottawa Panel, 2 independent reviewers separately evaluated the studies provided by the literature search. Each reviewer drafted a list of included and excluded articles with justifica-tion by applying the inclusion and exclusion criteria created by the Ottawa Panel (Tab. 1). If uncertainty occurred, the reviewer reread the article in question. The level of agreement between the reviewers was tested for interrater reliability (ie, kappa statistic) in a previous Ottawa Panel guidelines article. 12 A senior methodologist and a clinical expert compared the reviewers' lists of included and excluded studies, and a final judgment was made by the Ottawa Panel through consensus.

Data Extraction and Methodological Quality Assessment
Using predetermined extraction forms, the reviewers independently recorded details from the selected articles. Information regarding population characteristics, interventions, study design, allocation concealment, comparative outcomes, and period of measurement were recorded. The methodological quality of the potential studies also was assessed using the Jadad scale, 22 a 5-point scale with reported reliability

Ottawa Panel EBPCGs for Strengthening Exercises in the Management of Fibromyalgia
and validity that assigns 2 points for randomization, 2 points for double blinding, and 1 point for description of participant withdrawals. The senior reviewer was consulted to resolve discrepancies in data extraction and scoring of methodological quality.
Studies with a Jadad scale score of Ն3 are characteristically seen as hav-ing higher methodological quality; however, the Ottawa Panel agreed that trials with a Jadad scale score of Ͻ3 could still be included because the Jadad scale was initially developed for medical interventions (eg, medications, placebos), not exercise interventions. Thus, more consideration was given to the randomization component of the Jadad scale because most exercise trials cannot ob-tain points with the double-blind experiment category (ie, it is not possible to blind participants to an exercise intervention). Quality scores from the Jadad scale were used to interpret the results.
Studies that failed to meet methodological criteria described in the inclusion and exclusion table (Tab. 1) and that were determined by the Ot-

Ottawa Panel EBPCGs for Strengthening Exercises in the Management of Fibromyalgia
tawa Panel to be of poor methodological quality were transferred to the list of excluded studies.

Data Analysis
Statistical analysis was based on Cochrane Collaboration methods. 23 Continuous data were calculated with weighted mean differences (WMDs) between the intervention and control groups. A WMD is "a method of meta-analysis used to combine measures on continuous scales (such as weight), where the mean, standard deviation, and sample size in each group are known." 24 For the present analysis, data could not be pooled because many key study characteristics (eg, population, intervention, control, outcomes) were not comparatively similar. For example, the outcomes (eg, pain) may have been identical across studies that were compared, but the measurement methods used (eg, type of questionnaire) were different. Due to the extent of dissimilarity among studies, WMDs were calculated as opposed to standard mean differences.
Dichotomous data (ie, data that can be divided into 2 categories) were calculated using relative risks. A relative risk is "the ratio of risk in the intervention group to the risk in the control group. The risk (proportion, probability, or rate) is the ratio of people with an event in a group to the total in the group." 24 Data in the present study were illustrated according to Cochrane Collaboration methods 24 (see Figs. 1 and 2 for examples of illustrations of data). The horizontal line in the illustrations of data is the standard deviation of the WMD, and the square represents the WMD between the 2 groups when measured for a particular outcome. When the standard deviation line touches the central vertical line, the confidence interval is 0 and the 2 groups are not statistically different.
Clinical improvement was calculated using absolute benefit and relative difference (RD) in change from baseline. Absolute benefit is the improvement in the treatment group minus the improvement in the control group. Relative difference is absolute benefit divided by the baseline mean (weighted for the intervention and control groups). 12 Clinical significance was attained when an improvement of 15% relative to a control was found. With dichotomous data, the percentage of improvement was calculated as the difference in the percentage of improvement between the intervention and control groups. 25 The 15% value was chosen by the Philadelphia Panel, 24 who are experts in musculoskeletal practice, and was approved by the rheumatology and biostatistics experts of the Ottawa Panel. For greater detail of the statistical analysis, see our previous Ottawa Panel publication. 10

Literature Search
Our literature search strategy found 1,005 articles on various therapeutic exercises (eg, aerobic, fitness, relaxation) and FM. Using the inclusion and exclusion criteria (Tab. 1), 114 articles were found to be potentially relevant, and ultimately 5 studies on strengthening exercises and FM were selected. 26 -30 A strengthening exercise was defined as an isometric, isokinetic, or concentric/eccentric resistance exercise with the purpose of increasing the maximal force generated by a specific muscle or muscle group. 31 A total of 109 studies were excluded (results not shown) for the following reasons: the absence of a control group in 19 studies 32-50 ; the main therapy was fitness exercise in 14 studies 51-64 ; the main therapy was aerobic exercise in 13 studies 65-77 ; 11 studies 78 -88 had insufficient statistical data (ie, mean and standard deviation were not provided or could

Ottawa Panel EBPCGs for Strengthening Exercises in the Management of Fibromyalgia
not be calculated with the information provided); 9 studies 89 -97 were literature reviews; 9 studies 98 -106 had an attrition rate of Ͼ20%; 8 studies 107-114 were abstracts only; 8 studies 104,115-121 included control groups of subjects who were healthy; reports of 4 studies 122-125 were written in a foreign language; 3 studies 126 -128 did not have interventions; 2 studies 129,130 were case reports; and the remaining studies 131-139 were excluded for various other reasons, such as the study was only descriptive. 137

Methodological Quality
Results from the Jadad scale showed that 4 out of the 5 RCTs received low scores (Ͻ3) 26,27,29,30 and, consequently, were of poor methodological quality. The remaining trial 28 was of higher methodological quality, with a score of 3 out of 5. None of the studies received points under the "double-blind" category, and only the study of higher quality 28 received a point for description of participant withdrawals and dropouts.

Summary of Trials
Five trials (Nϭ150) 26 -30 measured strengthening exercises for the management of FM. Two different exercises were used to measure muscle strength: maximal concentric leg extensor force and maximal isometric knee extensor and flexor force. Using a dynamometer, concentric leg extensor force (in newtons) was measured with the participant sitting with hip and knee joints at 110 and 70 degrees, respectively, and attempting a full knee extension of 180 degrees against resistance. The load was progressively increased after each extension until the participant was not able to perform the exercise correctly. The last acceptable trial with the heaviest load was determined as 1 repetition maximum. Two trials (Nϭ47) 26,30 implemented this measurement.
Another measure of muscular strength utilized was calculating the maximal isometric force (in newtons) of the knee extensors and flexors using a dynamometer. Participants in a seated position with knee joints at 107 degrees were instructed to exert their maximal force as quickly as possible for 3 to 4 seconds. Each participant was given 3 trials, and the best result was used for analysis. Four trials (Nϭ94) 26

Discussion
Through an extensive, systematic review of strengthening exercises, the Ottawa Panel produced 2 EBCPGs, with 5 positive recommendations of clinical importance: 2 grade A and 3 grade Cϩ recommendations. The remaining recommendations were 2 grade C, 1 grade D, and 1 grade Dϩ. More research is needed on the effectiveness of strengthening exercises for pain relief, muscle strength, quality of life, self-efficacy, and sleep quality, which made up the grade C, D, and Dϩ recommendations.
The Ottawa Panel concluded that strengthening exercises are benefi-cial for the overall management of FM. Clinical benefits were shown for muscle strength (maximal concentric leg extensor force, maximal isometric knee extensor force, and maximal concentric leg extensor force), pain relief, physical disability, depression at end of treatment at 21 weeks, and quality of life at end of treatment at 12 weeks. Clinically important benefits without statistical significance were found for muscle strength (maximal isometric knee extensor force 27 and maximal isometric knee flexor force), quality of life at end of treatment at 21 weeks, and depression and anxiety at end of treatment at 12 weeks.
Encouraging results were found for both middle-aged women 26 -28 and elderly women 29,30 (ie, 55ϩ years of age) who had no previous experience with strengthening exercises and trained only twice per week. Patients with FM gained overall improvements in daily physical functioning, perceived fatigue, and mood compared with a control condition. 26,27,29,30 One study 28 showed no statistically significant differences between the experimental (strengthening exercises) and control (flexibility) groups. Using flexibility exercises as a control condition is a possible explanation. Given that both groups were relatively sedentary at baseline (ie, 87% were defined as completely sedentary), any moderate exercise would have led to overall health improvements. 28 Minimum education (eg, in the form of a pamphlet) on FM disease management without reference to specific exercises would be a more appropriate control group intervention.
Another possible explanation for the lack of between-group differences is that participants in the treatment group were not required to progressively increase their exercise intensity. 28 Researchers were concerned that demanding an increase in load (ie, weight) would have aggravated participants' symptoms of muscle soreness and stiffness, possibly resulting in high subject withdrawal. 28,60,61 There are several ways to maintain a low attrition rate without

Ottawa Panel EBPCGs for Strengthening Exercises in the Management of Fibromyalgia
potentially compromising exercise intensity: personalizing the exercise instruction to patients with FM, using an exercise specialist with clinical FM experience, implementing self-efficacy as a guiding theoretical framework, providing an encouraging and supportive exercise instructor, and administering the exercise plan within relatively close distance of the participants' homes. 28 Additionally, researchers who worked with patients with OA found that, when exercise programs included behavioral components (eg, face-toface visits, social/peer support and positive feedback, patient education, use of pedometer), lower dropout rates and higher adherence rates occurred. 140 -142 Further research is needed on designing exercise programs that patients with FM will find relatively easy to adhere to in the long term.
Most studies showed that patients with FM can successfully engage in an intensive, progressive strengthening program without experiencing an exacerbation of exercise-induced or FM symptoms. 26,28 -30 The function and trainability of muscles in patients with FM were found to be similar to that of participants serving in a "healthy" control condition. 29 Thus, although issues of disability were beyond the scope of this project, it is likely that patients with FM can achieve health benefits from strengthening exercises similar to those achieved by people who are healthy. 29 Furthermore, strength gains in major muscle groups may assist patients in more easily completing aerobic exercises, which also have been shown to alleviate some FM symptoms. 28 Although the causes of FM remain largely unknown, clinicians and researchers continue to advocate a multimodal approach to the management of FM. 55 Medications (eg, to improve sleep quality and mood) and psychosocial interventions (eg, cognitive-behavioral therapy) combined with exercise seem to produce the most positive health effects. 53 Further research is needed to determine the mechanisms involved in how specific muscle strengthening exercises of the lower limbs over 21 weeks have a general benefit on widespread myalgia, fatigue, psychological state, and quality of life. Undocumented psychosocial interventions (eg, unstructured behavioral interventions, positive expectations of the therapists) would need to be addressed as potential confounding factors. Moreover, the long-term therapeutic outcome on psychosocial functioning and, specifically, work capability, should be assessed. 53 Limitations A recurrent problem, as found when studying therapeutic exercises for patients with RA, 10 OA, 11 and stroke, 12 is that researchers often fail to give thorough descriptions of the type of exercise used or its duration, intensity, and frequency. Information on the study sample such as participant level of physical fitness or level of depression (if applicable), both factors that have implications for treatment success, also usually is not provided. A lack of exercise and sample detail presents challenges to researchers who want to replicate a study and to practitioners and other health care professionals who want to prescribe effective strengthening exercises. The Ottawa Panel recommends that researchers provide as much study detail as possible.
The rehabilitation community could recommend that researchers provide more study detail in the form of an appendix (so as to not lengthen the article itself), or they could construct a standardized EBCPG form on study characteristics to be voluntarily filled out by researchers for the potential future development of EBCPGs.
A limitation inherent in the construction of EBCPGs is the occurrence of conflicting evidence. For strengthening exercises versus a control condition, statistical and clinical significance (grade A) was found for maximal isometric knee extensor force in one study, 30 but only clinical significance (grade Cϩ) was found for the same outcome of a different study. 27 The reason for the discrepancy is not known; however, it is most likely the result of different samples, exercise outcomes, exercise interventions, or study designs used, which is why the study data could not be pooled. The heterogeneity of studies selected can lead to the above conflicting evidence, which, consequently, prevents sound conclusions from being drawn. In future studies, the Ottawa Panel will select studies, if possible, that can be pooled more easily.
Finally, another inherent weakness of the EBCPGs was that the recommendations were based on single trials of relatively low methodological quality across multiple outcomes. Four out of the 5 RCTs had a score of less than 3 on the Jadad scale. This finding should be noted when referring to the guidelines, in particular the grade A and Cϩ recommendations.

Implications for Practice
The Ottawa Panel has found emerging evidence to support the use of strengthening exercises as part of the overall management of FM. Most improvements were shown for muscle strength, quality of life, and decreases in depression. However, the lack of detail regarding the strengthening exercises and study sample was a significant weakness. Due to the vast variability in FM symptoms, patients would likely function best with a highly individualized program that incorporates multiple treatment regimens.